Monetary Policy Implementation and Financial Vulnerability: Evidence from the Overnight Reverse Repurchase Facility

Monetary Policy Implementation and Financial Vulnerability: Evidence from the Overnight Reverse... Abstract In this paper, we examine the Federal Reserve’s newest policy tool, known as the overnight reverse repo (ONRRP) facility, to understand its effects on the repo market. Using exogenous variation in the parameters of the ONRRP, we show that private repo activity is crowded out when money funds invest in the ONRRP. Additionally, we find that the ONRRP increases lenders’ bargaining power, thereby raising borrower funding costs. Lastly, we show that repo borrowers reallocate to repo backed by riskier collateral and borrow more from ONRRP-ineligible asset managers, both of which could increase financial vulnerability due to instability in dealer funding. Received April 27, 2017; editorial decision October 19, 2017 by Editor Stijn Van Nieuwerburgh. In its response to the recent financial crisis, the Federal Reserve (the Fed) intervened heavily in financial markets to implement a series of quantitative easing (QE) programs. Indeed, in the years since the crisis, many central banks resorted to similar measures in an effort to buoy local economies. One consequence of this aggressive central bank intervention has been a rapid expansion of monetary authorities’ asset holdings and presence in financial markets. This unprecedented intervention has brought forth questions regarding its potentially negative effect on financial stability (Borio and Zhu 2012; Yellen 2014; Adrian and Liang 2016). The possible difficulty associated with the removal of this unprecedented accommodation has also raised financial stability concerns (Bernanke 2012). In the case of the Fed, the first major central bank to embark on a tightening cycle after the crisis, the abundance of liquidity prevented a return to the traditional means of monetary policy implementation. Historically, the Fed relied on tight control over the supply of reserves held by the banking system in order to control its main policy rate, known as the federal funds rate. In a reserve-abundant environment, modest changes in the supply of reserves will have no effect on the effective federal funds rate (Ihrig, Meade, and Weinbach 2015). Instead, the creation of a new policy tool known as the overnight reverse repurchase (ONRRP) facility was required to lift short-term rates. In this paper, we conduct the first thorough analysis of the effects of the ONRRP facility on financial markets, and demonstrate that there can be a tradeoff between effective monetary policy implementation and financial stability. Under the ONRRP facility, money market participants including money market funds (MMFs), government-sponsored enterprises (GSEs), banks, and dealers can enter into reverse repurchase agreements with the Fed at a prespecified fixed rate. In these transactions, counterparties lend cash to the Fed overnight, with Treasury securities from the Fed’s portfolio posted as collateral. To gain operational experience with the ONRRP, the Fed began conducting daily tests of the ONRRP facility in 2013. During these tests, the Fed occasionally adjusted parameters of the facility, such as the fixed offer rate and the maximum counterparty bid amount, or “cap.” We analyze the effects of the ONRRP facility by examining the facility’s influence on MMF investment decisions and the systemically important tri-party repo market.1 In particular, we exploit exogenous changes in the ONRRP counterparty cap during the testing phase to identify the displacement of private activity in the repo market as a consequence of Fed intervention. Upon the introduction of the ONRRP, some MMFs faced a constrained investment allocation decision because of the low level of the ONRRP facility’s maximum daily counterparty cap, whereas others were unconstrained by the cap. Future cap hikes precipitated exogenous increases in the ability of initially constrained MMFs to invest in the ONRRP, and provide events around which we construct difference-in-difference estimates to identify the effect of ONRRP participation on both MMFs’ lending decisions and the private repo market. Using a combination of MMF regulatory filings as well as confidential data on tri-party repo market transactions and dealer repo positions, we are able to trace out the effects of the ONRRP on financial markets. We find that by borrowing in the tri-party repo market, the Fed crowds out private MMF repo lending backed by both Treasury and agency collateral, while also reducing MMF investment in bank deposits. However, cash lenders in the tri-party repo market evidently value their dealer relationships highly enough to maintain relationships with all their dealers, a result that is consistent with the importance of relationships in the repo market described in Copeland et al. (2012). In further analysis, we examine the effects of the ONRRP facility on money funds’ bargaining power. We show that an exogenous positive shock to a fund’s ability to invest in the ONRRP facility improves its bargaining power relative to borrowers, thereby increasing private repo rates and, by extension, dealers’ funding costs. This combination of higher rates and lower volumes is consistent with the view that the ONRRP introduced a negative funding supply shock to the private repo market. Lastly, we analyze how repo borrowers respond to these adverse funding shocks. We find that dealers do not decrease their total repo borrowing. Rather, dealers make up for the repo supply lost to the ONRRP by adjusting both their collateral and lender composition. Dealers that are more exposed to MMFs investing in the ONRRP and thus subject to larger adverse funding shocks aim to maintain their relationships with the largest (ONRRP-eligible) asset management complexes by transacting with their non-MMF affiliates. Because non-MMFs lend more against nongovernment collateral repo, dealers must accommodate this transition by expanding their borrowing against riskier collateral and increasing their reliance on risky repo for net financing. To a lesser degree, these dealers also recoup the repo borrowing lost from ONRRP-eligible MMFs by turning to ONRRP-ineligible MMFs, which tend to be smaller and have less stable net assets. The findings reported in this paper indicate that the ONRRP facility can increase vulnerabilities in the financial system through two distinct channels. First, we find that dealers do not decrease total repo issuance in response to the negative funding supply shock caused by the ONRRP, choosing instead to boost their reliance on riskier types of repo borrowing. This adverse impact on dealers’ funding risk leads to heightened financial fragility in this sector and thus an increase in the risks that an episode of financial stress will result in a disruptive funding shock. Second, we find that MMFs withdraw repo funding from their regular dealer counterparties in order to invest with the Fed. Thus, if MMFs become concerned about the actual or perceived safety of their investments (Kacperczyk and Schnabl 2013), they would be likely to invest larger amounts with the Fed at the expense of repo with dealer counterparties. In a severe event, MMFs may even begin to substitute out of other investments, such as commercial paper, in favor of the ONRRP. However, we also find evidence of the importance of borrower-lender relationships in the repo market, which may counteract these adverse effects in severe stress events if MMFs are reluctant to completely sever ties with their dealer counterparties. This study connects three separate strands of literature. First, we contribute to the literature that examines how novel monetary policy programs affect financial markets. Previous studies demonstrate that new policy actions can have a largely favorable impact on prices and quantities in the targeted markets. For instance, Sundaresan and Wang (2008) show that the Fed’s “Y2K options” lowered liquidity premiums in the Treasury market, while Duygan-Bump et al. (2013) show that the Fed’s Asset-Backed Commercial Paper Money Market Mutual Fund Liquidity Facility stabilized MMF asset flows and reduced yields during the financial crisis. However, there is also evidence of potentially unintended consequences of policy actions. Particularly during periods of heavy intervention or acute market strain, the implementation of monetary policy has been shown to crowd out private activity. For example, Brunetti, Filippo, and Harris (2011) present results that suggest ECB intervention during the financial crisis crowded out interbank trades. In additional analysis of ECB actions both before and during the European sovereign debt crisis, de Andoain et al. (2016) show that ECB liquidity injections displaced private activity in the euro area, notably reducing interbank trading. Separately, Kandrac (Forthcoming) documents that Federal Reserve purchases under QE programs displaced private activity in the secondary market for mortgage-backed securities. Consistent with the pattern evident in these studies, we show that the new ONRRP facility crowds out private repo lending supplied by ONRRP-eligible MMFs. Second, our work relates to the growing literature on the financial stability effects of monetary policy (Borio and Zhu 2012; Adrian and Liang 2016), and especially those studies that examine the effects of recent unconventional policies (Stein 2012; Woodford 2016). The general thrust of this literature acknowledges the potential for heightened financial fragility owing to extraordinarily accommodative monetary policy. However, Greenwood, Hanson, and Stein (2015) argue that ONRRP operations can support financial stability by crowding out private money creation and its attendant negative externalities. While we find that counterparties that directly lend to the Fed via the ONRRP do in fact reduce their private repo volumes, dealer repo borrowers evidently make up for the reduction in repo funding by borrowing more from ONRRP-ineligible asset managers against riskier types of collateral. Third, we contribute to the literature that seeks to further understand segments of the increasingly important shadow banking sector including MMFs, securities dealers, and the collateral-backed funding markets that have accompanied the expansion of shadow banking (Sunderam 2014; Di Maggio and Tahbaz-Salehi 2014). In particular, previous studies have documented the growing reliance on the repo market for short-term funding, as well as the role of the repo market in the recent financial crisis (Gorton and Metrick 2012; Copeland, Martin, and Walker 2014; Krishnamurthy, Nagel, and Orlov 2014). Here, we consider a standing monetary policy tool that introduces the Fed as a large, persistent repo borrower willing to expand the supply of a safe asset to a variety of money market participants, and show the resultant effects on private tri-party repo activity. 1. Institutional Background 1.1 The Fed’s ONRRP facility Beginning in 2009, the Fed dramatically expanded its balance sheet through several rounds of asset purchase programs known as QE. Consequently, banks were left with large amounts of excess reserve balances on deposit with the Fed. Given that monetary policy implementation before the crisis relied on altering the amount of scarce reserves to influence short-term rates, the Fed required additional tools to conduct monetary policy in the new environment (Ihrig, Meade, and Weinbach 2015). The primary tool, paying banks interest on excess reserves (IOER), was implemented in October 2008. In theory, banks should be unwilling to lend at less than IOER—the deposit rate that they can earn from the Fed—causing IOER to set a floor on short-term interest rates. However, in the United States, there are many nonbank participants in money markets that do not have access to IOER and are therefore willing to lend at lower rates (Bech and Klee 2011). The difference between IOER and overnight lending rates ostensibly presents an arbitrage opportunity for banks, which could be eliminated through competition among borrowers in the overnight market. However, this arbitrage opportunity was limited by costs of holding reserves, such as minimum capital requirements and a fee charged on total assets to insure bank deposits. To regain control over overnight rates, the Fed responded in two ways: first, by expanding the set of counterparties with which it transacts and second, by introducing the ONRRP facility. As early as October 2009, the Fed was considering using repo transactions as a means of withdrawing policy accommodation, while also expanding the set of eligible repo counterparties to broaden the reach of such transactions (FRBNY 2009). Specifically, repo operations were made available to MMFs, GSEs, and banks, as well as the Fed’s traditional counterparties, primary dealers. Market participants could apply to become a Fed counterparty in several rounds beginning in March 2010.2 The Fed’s eligibility requirements generated important differences between ONRRP eligible and ineligible MMFs. In particular, eligible MMFs tended to be larger, more established funds. This was a result of the requirement that MMFs have at least $\$$ 5 billion in net assets and exist for at least one year to be eligible to participate in the ONRRP. Eligible MMFs also featured more stable assets and repo participation. Between 2011 and 2015, the standard deviation of net assets was 18% for eligible MMFs versus 25% for ineligible MMFs. Likewise, the standard deviation of repo lending volume was 9% for eligible MMFs and 14% for ineligible MMFs. In September 2013, subsequent to the expansion of the Fed’s RRP eligible counterparties, the Fed began conducting regular overnight, fixed-rate capped-allotment reverse repurchase agreements through an extended testing exercise. In these repo transactions, the Fed borrows from a broad set of counterparties on an overnight basis using Treasury securities as collateral.3 At the program’s start, 140 counterparties were eligible to participate in the ONRRP, including MMFs, GSEs, banks, and primary dealers.4 Importantly, the availability of the ONRRP to a broad set of counterparties operating in different market segments ensures that the ONRRP can help establish a firmer floor on short-term rates, in contrast to IOER, which is only offered to depository institutions. As a result of successful testing, the FOMC’s Policy Normalization Principles and Plans, released in September 2014, states the Committee’s intention to use the ONRRP facility as a tool to help control the federal funds rate during the normalization of the stance of monetary policy.5 Moreover, as discussed by the Committee at the November 2016 FOMC meeting, it is possible that there will not be a return to the operating framework that prevailed before the crisis, establishing the ONRRP as a key policy tool indefinitely.6 By offering an outside option to cash investors, the ONRRP can theoretically influence rates even without active participation from repo lenders. Nevertheless, the ONRRP has evidently presented an attractive investment opportunity to investors, especially MMFs, which typically account for over 85% of total take-up. As shown in Figure 1, participation in the facility has often reached sizeable levels. Through 2015, the average total take-up was $\$$ 115.5 billion per day, with a peak of $\$$475 billion in December 2015. One important aspect of the ONRRP facility is that a maximum bid, or cap, is imposed on each individual counterparty. As shown in Figure 2, this maximum has ranged from $\$$500 million at the start of the program to $\$$30 billion through 2017. The maximum individual bid is the primary constraint for participants.7 Changes in the maximum individual bid amount of the ONRRP are key to our identification strategy described in Section 2. Figure 1 View largeDownload slide Investment in the RRP facility This figure shows total daily participation in the RRP facility, including both overnight and term, from its inception through 2015 for MMFs, GSEs, and all others. Source: Federal Reserve Bank of New York, 2013–2015, Reverse Repo Data. Figure 1 View largeDownload slide Investment in the RRP facility This figure shows total daily participation in the RRP facility, including both overnight and term, from its inception through 2015 for MMFs, GSEs, and all others. Source: Federal Reserve Bank of New York, 2013–2015, Reverse Repo Data. Figure 2 View largeDownload slide ONRRP counterparty caps This figure shows the maximum ONRRP bid allowed per counterparty over time. Figure 2 View largeDownload slide ONRRP counterparty caps This figure shows the maximum ONRRP bid allowed per counterparty over time. While the ONRRP has been very effective at controlling short-term interest rates as intended, the presence of a standing reverse repo facility may also have consequences for financial stability that are theoretically unclear. One leading concern regarding the ONRRP is that a standing Fed borrowing facility may increase the likelihood or severity of flight-to-safety “runs” in the private repo market if investors suddenly become reluctant to lend to non-Fed counterparties (Frost et al. 2015). Outside of severe stress events, however, the withdrawal of MMFs from dealers in favor of the Fed could cause dealers to respond by reallocating their repo funding to less stable sources. Such a rise in dealers’ funding risk would increase financial vulnerabilities and heighten the susceptibility of this sector to stress events. Conversely, if the ONRRP facility crowds out privately issued repo created by the shadow-banking sector, it can buttress financial stability by reducing the externalities associated with a large private repo market, which—as demonstrated by the recent financial crisis—is subject to highly disruptive runs (Stein 2012; Carlson et al. 2016; Greenwood, Hanson, and Stein 2016). In addition to crowding out risky private repo, the ONRRP could help reverse the withdrawal of safe assets that occurred during QE by providing an interest-bearing near-money asset to a broad array of counterparties (BIS 2015; Infante 2016; Gorton 2017) and potentially lowering liquidity premiums (Nagel 2016). 1.2 The tri-party repo market In the United States, the repo market is divided into several segments, including the bilateral repo market, the tri-party repo market, and the GCF market, which is a subset of tri-party.8 Given data limitations, the exact size of the repo market and its subcomponents is not attainable. However, it is estimated that the tri-party market accounted for about 40%–45% of the repo market, or about $\$$ 1.5 trillion per day in 2014 and 2015 (Copeland et al. 2014; Baklanova, Copeland, and McCaughrin 2015). The tri-party market has historically relied upon two third-party clearing banks—Bank of New York-Mellon and J.P. Morgan—that settle all repo transactions. Tri-party market transactions typically consist of nondealers, including MMFs, securities lenders, and others, lending to dealers. MMFs conduct most of their repo lending in the tri-party market and account for about a third of daily tri-party volume. Tri-party collateral is specified only by type, rather than by security; that is, it is a general collateral market. Fedwire-eligible collateral, including Treasuries, agency debt, and agency mortgage-backed securities (MBS), account for about 85% of market volume as of 2012 (Copeland et al. 2012). We generally limit our attention to the tri-party repo market in this paper for two reasons. First, all ONRRP operations are conducted in the tri-party market. Second, our primary focus is on the response of MMF repo lending activity, which also takes place over the tri-party platform, to the ONRRP. 2. Identification In general, our aim is to evaluate the effects of the Fed’s ONRRP facility on the private repo market. Because the ONRRP is a standing facility, repo investors (such as MMFs and government-sponsored entities) can choose at will to lend risklessly to the central bank at the facility’s rate rather than to their usual dealer counterparties. Even without actual participation in the facility by ONRRP counterparties, the mere presence of the facility could boost MMF’s bargaining power vis-à-vis dealers, thereby exerting upward pressure on repo and other short-term rates. However, as demonstrated in Figure 1, MMFs evidently found that the terms offered by the ONRRP facility presented an attractive investment opportunity, and did indeed participate in the facility, possibly at the expense of existing counterparties. In Figure 3, we show total MMF repo investment composition, subdivided into Fed (i.e., ONRRP) and private components. Ostensibly, total MMF repo investment remained roughly constant over the last four years, and the entrance of the Fed claimed market share from MMF’s private counterparties. Figure 3 View largeDownload slide Money market fund repo investment This figure shows the monthly private versus Fed decomposition of MMF repo investment backed by Treasury and agency collateral. Source: SEC form N-MFP. Figure 3 View largeDownload slide Money market fund repo investment This figure shows the monthly private versus Fed decomposition of MMF repo investment backed by Treasury and agency collateral. Source: SEC form N-MFP. However, participation in the ONRRP facility is an endogenous outcome of borrower and lender interactions in money markets. Consequently, other factors may have reduced dealers’ willingness or ability to borrow in the repo market. For example, regulatory pressures on dealers including more stringent capital requirements, the introduction of the liquidity coverage ratio, and less permissive rules regarding the netting of repo trades may have reduced dealers’ willingness to borrow in the repo market. In addition, as previously seen in Figure 3, total private tri-party repo declined starting in January 2013, when the supplementary leverage ratio was implemented internationally. As a result of these changes, increases in MMF ONRRP participation may only be coincidentally timed with a reduction in private repo, which could simply reflect developments in the market that are unrelated to the introduction of the ONRRP facility. In other words, it may be that MMFs participate in the ONRRP differentially as a result of their available options in money markets, thereby introducing bias in attempts to estimate a causal effect of ONRRP participation on the investment outcomes of Fed counterparties. To overcome these endogeneity issues and draw causal inference, we exploit exogenous variation in the ONRRP facility parameters, which were occasionally adjusted during the testing phase. In particular, we focus on the changes to the counterparty caps that regulated the maximum daily bid of individual counterparties. As shown in Figure 2 and discussed in the previous section, the counterparty cap was raised in increments over the first year of the ONRRP’s existence. Funds that would have optimally invested more in the ONRRP facility than permitted by the maximum bid amounts were thus forced to invest at the maximum bid. Evidently facing a constrained investment decision, these funds should be expected to boost their ONRRP investment upon an increase in the counterparty cap. Importantly, the maximum bid amounts were raised merely as a normal result of the expansion of the ONRRP tests, and were unrelated to any changes in either the repo market or MMF investment decisions and preferences. Thus, MMFs that were initially constrained by the counterparty cap experienced an exogenous change in their ability to invest in the ONRRP facility. MMFs that were constrained by the initial ONRRP cap at the facility’s inception compose our “treatment” group, as these funds witnessed exogenous increases in their ability to invest in the ONRRP upon future cap increases.9,10 We can then compare treated funds to ONRRP-eligible MMFs that were not constrained by the initial counterparty cap—our “control” group—to generate differences-in-differences (DD) estimates of the effect of an increase in ONRRP participation. Counterparty eligibility for the ONRRP was determined by an application process available to all qualifying funds. Thus, to avoid any possible selection issues emanating from the application decision, we limit our sample to ONRRP-eligible counterparties only. Figure 4 depicts an illustrative example of our identification strategy. Upon introduction of the ONRRP, period 1 on the left side of the figure, the initial counterparty cap does not constrain fund 1, which is able to achieve its optimal investment allocation. Fund 1 is therefore part of the control group. Conversely, funds 2 and 3 face a binding counterparty cap at the inception of the ONRRP. These funds compose our treatment group. In period 2—the beginning of the “post-treatment” period—the counterparty cap has been increased, affording funds 2 and 3 the ability to increase their ONRRP investment, and fund 1 continues to submit below-cap bids. Thus, our treatment and control funds are established at the inception of the ONRRP, and we identify off of subsequent cap changes that present exogenous events around which we can construct DD estimates. Figure 4 View largeDownload slide Identification strategy: An example of control and treatment groups This figure depicts an illustrative example of our identification strategy. In the first period, fund 1 (yellow) does not face a constrained investment allocation decision as a result of the counterparty cap (the dashed line) and is included in the control group. Funds 2 and 3 (red and blue, respectively) are bound by the cap and will compose the treatment group. In the second period, fund 2 now faces an unconstrained investment decision as a result of the increase in the counterparty cap and is thus included in the $$Unbound$$ treatment subgroup. Although fund 3 was able to increase its ONRRP investment as a result of the increase in the counterparty cap in period 2, it continues to face a constrained investment decision and is thus included in the $$Bound$$ treatment subgroup. In period 3, the counterparty cap is again increased, with all funds facing an unconstrained investment decision such that fund 3 joins the $$Unbound$$ treatment subgroup with fund 2. Figure 4 View largeDownload slide Identification strategy: An example of control and treatment groups This figure depicts an illustrative example of our identification strategy. In the first period, fund 1 (yellow) does not face a constrained investment allocation decision as a result of the counterparty cap (the dashed line) and is included in the control group. Funds 2 and 3 (red and blue, respectively) are bound by the cap and will compose the treatment group. In the second period, fund 2 now faces an unconstrained investment decision as a result of the increase in the counterparty cap and is thus included in the $$Unbound$$ treatment subgroup. Although fund 3 was able to increase its ONRRP investment as a result of the increase in the counterparty cap in period 2, it continues to face a constrained investment decision and is thus included in the $$Bound$$ treatment subgroup. In period 3, the counterparty cap is again increased, with all funds facing an unconstrained investment decision such that fund 3 joins the $$Unbound$$ treatment subgroup with fund 2. As portrayed in the figure, in period 2, the maximum allowable bid is increased by an amount large enough to leave fund 2 unconstrained by the new cap, while fund 3 remains constrained by the new, higher cap. Although we consider both funds 2 and 3 to have received treatment as a result of the cap increase, we can differentiate between two treated subgroups: one treatment subgroup transitions from a constrained state to a new constrained state despite the higher cap (e.g., fund 3 in period 2), while the other transitions from a constrained state to an unconstrained state (e.g., fund 2 in period 2). Eventually, all treated funds in our sample enter the unconstrained treatment group, as depicted in period 3, when the counterparty cap is increased to a level allowing fund 3 to achieve an unconstrained investment allocation. Employing such a DD identification strategy addresses many threats to the casual interpretation of our results. First, DD allows us to account for constant differences between funds, as well as any overarching factors that could affect the whole repo market at any point in time. That is, our DD estimation controls for differences between treated and control MMFs and for shared differences between quarter-ends. Moreover, cap increases were numerous, variably sized, and occurred at irregular intervals. This rich variation in our exogenous treatment events means that other explanations of the effects we observe require a time pattern that is well correlated with the timing of cap increases and affect only those funds that were subject to a binding cap at the introduction of the ONRRP (i.e., treated funds). There is also variation in the timing of treatment funds’ transition to the unconstrained treatment group. As a result, any factors confounding the causal interpretation of our estimates must affect only a specific (and changing) subset of treatment funds at specific points in time. In addition, we note that the tri-party repo market introduces frictions that can limit alternate sources of within-fund variation over time, particularly as it concerns each MMF’s set of borrowers. For instance, as described in Copeland et al. (2012), each MMF must execute a master repo agreement with each individual borrower, which lays out important elements of their repo transactions. Before tri-party repo trading can commence, each MMF-dealer pair must additionally execute a unique custodial undertaking agreement with the clearing bank to establish its role as the agent. Lastly, most repo transactions are overnight trades, and many of these are open transactions between a given borrower and lender that are continuously “rolled over” on a day-to-day basis. Thus, MMFs can only transact with those dealers with which they have executed the necessary agreements, and trading conventions in the repo market ensure that MMF-dealer trading relationships are typically persistent. We follow the general DD identification strategy outlined above throughout the remainder of the paper. Summary statistics, for our treatment and control groups follow in Section 3 and the details of the precise DD specifications are provided in Section 4. 3. Data To evaluate the effect of the ONRRP facility on MMFs and their dealer borrowers, we employ one comprehensive public data set of MMF investments, three confidential data sets reported to the Fed that contain detailed information on tri-party repo market transactions and relationships, and a confidential data set that lists the entire repo and reverse repo portfolios of individual securities dealers on a weekly basis.11 First, we use the monthly MMF filings of form N-MFP to the Securities and Exchange Commission, which contain a detailed schedule of portfolio holdings for each MMF on month-ends. These data contain information on each security, including the issuer, security type, maturity date, and volume. In addition, we use three different data sets containing information on the tri-party repo market. The first contains pricing information that allows us to analyze rates in the tri-party repo market. This data set records all transactions facilitated by one of the two tri-party clearing banks in the tri-party repo market, which are reported to the Federal Reserve Bank of New York (FRBNY). Lenders in this data set are provided at the mutual fund complex (“family”) level, rather than the individual fund level. Trade information includes the volume, annualized rate, term to maturity, dealer, lender, and underlying collateral backing each repo transaction. For our purposes, we exclude term trades and retain only those transactions backed by Fedwire-eligible Treasury and agency (including agency MBS and agency debt) collateral. Neither of these filters significantly reduces our sample size, as overnight trades backed by Treasury or agency collateral compose over 70% of the transactions in the data. The second tri-party data set contains dealer position-level information, reported by both clearing banks, providing a complete picture of total repo borrowing in the tri-party market. For each dealer, this data set records the daily amount of repo outstanding for each collateral type. With these data, we are able to examine repo of all tenors and each collateral type to assess how dealers reallocate their funding in response to the change in behavior of their lenders as a result of the ONRRP. It is also possible to calculate the haircut of each repo position from these data, which compares the difference between the value of the collateral and the value of the cash loan normalized by the value of the cash loan. Higher haircuts are imposed when using riskier collateral, with more collateral being required for a given loan amount (Gorton and Metrick 2012). The third tri-party data set is also position-level, but it contains daily outstanding repo between each dealer and lender (with no collateral information). While some small investors are not included, the positions recorded in this data set account for more than three-quarters of total tri-party repo positions. These data allow us to analyze how dealers’ counterparty composition changes as their lenders shift to the ONRRP. Lastly, we compile a data set from the weekly filings of form FR-2004C, which is used to collect information on dealer financing from primary dealers.12 These forms are submitted on a weekly basis, and contain confidential information collected under the Fed’s supervisory authority. These data allow us to examine the entirety of dealers’ repo and reverse repo positions by collateral type and tenor. Dealers’ repo borrowing can reflect funding used for financing a net securities position, or it can reflect so-called “matched book” activity. A dealer’s matched book, which entails offsetting repo and reverse repo positions, increases the balance sheet size but does not materially increase its riskiness. This is because dealers that experience a withdrawal of repo lending can simply refrain from rolling over reverse repos with matching tenors. Conversely, dealers that use repo as a source of net financing for securities holdings expose themselves to risk because of the maturity mismatch between their net securities holdings and their short-term repo funding. The FR-2004C data therefore enable us to examine the extent to which the riskiness of dealers’ funding increases when lenders shift to the ONRRP. For example, if dealers respond to the withdrawal of funding by simply reducing matched book activity (i.e., simultaneously reducing reverse repos), this would not necessarily represent an increase in financial fragility. Summary statistics for both treated (initially constrained) and control (initially unconstrained) MMFs are shown in Table 1. Since most of our analysis is conducted using quarter-end observations when ONRRP participation is typically highest, all values are reported as of June 28, 2013, the last quarter-end before the ONRRP began. Our sample consists of 101 ONRRP-eligible MMFs, of which 39 compose our treatment group, as these funds were constrained by the initial (September 30, 2013) quarter-end counterparty cap of $\$$ 1 billion. Of the treated MMFs, 21% remained constrained by the $\$$3 billion December 2013 cap, 5% were constrained by the $\$$7 billion cap in March 2014, and 4% were constrained by the $\$$10 billion cap in June 2014. The last counterparty cap increase to $\$$30 billion cap in September 2014 effectively removed the constraint for the remaining treatment funds. Table 1 Summary statistics Treated Control Variable (Initially constrained) (Initially unconstrained) Number of funds 39 62 Number of prime funds 20 45 Number of dealers per fund 9.43 8.09 (0.74) (0.57) % of foreign dealers per fund 60.71 61.66 (4.42) (3.14) Weighted CDS spread (%) 0.85 0.83 (0.04) (0.04) HHI 0.18 0.20 (0.03) (0.03) Share of largest dealer 0.26 0.28 (0.03) (0.03) Weighted Treasury/Agency repo rate (bps) 13.38 11.62 (1.23) (0.84) Weighted Treasury repo rate (bps) 10.38 10.02 (0.38) (0.02) AUM ( $\$$, billions) 27.6 14.4$$^{***}$$ (4.05) (2.12) Treasury repo (% of AUM) 15.48 5.97$$^{***}$$ (3.67) (1.70) Agency repo (% of AUM) 11.85 7.98 (2.64) (1.35) CD (% of AUM) 18.25 20.79 (3.21) (2.24) Treasury debt (% of AUM) 14.31 15.08 (2.92) (2.78) Financial CP (% of AUM) 7.57 10.78$$^{*}$$ (1.54) (1.15) Asset-backed CP (% of AUM) 2.01 6.84$$^{***}$$ (0.52) (1.00) Treated Control Variable (Initially constrained) (Initially unconstrained) Number of funds 39 62 Number of prime funds 20 45 Number of dealers per fund 9.43 8.09 (0.74) (0.57) % of foreign dealers per fund 60.71 61.66 (4.42) (3.14) Weighted CDS spread (%) 0.85 0.83 (0.04) (0.04) HHI 0.18 0.20 (0.03) (0.03) Share of largest dealer 0.26 0.28 (0.03) (0.03) Weighted Treasury/Agency repo rate (bps) 13.38 11.62 (1.23) (0.84) Weighted Treasury repo rate (bps) 10.38 10.02 (0.38) (0.02) AUM ( $\$$, billions) 27.6 14.4$$^{***}$$ (4.05) (2.12) Treasury repo (% of AUM) 15.48 5.97$$^{***}$$ (3.67) (1.70) Agency repo (% of AUM) 11.85 7.98 (2.64) (1.35) CD (% of AUM) 18.25 20.79 (3.21) (2.24) Treasury debt (% of AUM) 14.31 15.08 (2.92) (2.78) Financial CP (% of AUM) 7.57 10.78$$^{*}$$ (1.54) (1.15) Asset-backed CP (% of AUM) 2.01 6.84$$^{***}$$ (0.52) (1.00) This table reports summary statistics as of June 28, 2013 (the last quarter-end before the ONRRP began) for treated (initially constrained) and control (initially unconstrained) MMFs in our sample. All dealer-related variables (number of dealers, % foreign, CDS spread, HHI, and share of largest dealer) are based on a sample of only MMFs that have nonzero private repo volume (96 MMFs rather than the full 101). Rates are calculated at a fund complex level, and all other variables are at the individual MMF level. Statistical significance of the difference in averages: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 1 Summary statistics Treated Control Variable (Initially constrained) (Initially unconstrained) Number of funds 39 62 Number of prime funds 20 45 Number of dealers per fund 9.43 8.09 (0.74) (0.57) % of foreign dealers per fund 60.71 61.66 (4.42) (3.14) Weighted CDS spread (%) 0.85 0.83 (0.04) (0.04) HHI 0.18 0.20 (0.03) (0.03) Share of largest dealer 0.26 0.28 (0.03) (0.03) Weighted Treasury/Agency repo rate (bps) 13.38 11.62 (1.23) (0.84) Weighted Treasury repo rate (bps) 10.38 10.02 (0.38) (0.02) AUM ( $\$$, billions) 27.6 14.4$$^{***}$$ (4.05) (2.12) Treasury repo (% of AUM) 15.48 5.97$$^{***}$$ (3.67) (1.70) Agency repo (% of AUM) 11.85 7.98 (2.64) (1.35) CD (% of AUM) 18.25 20.79 (3.21) (2.24) Treasury debt (% of AUM) 14.31 15.08 (2.92) (2.78) Financial CP (% of AUM) 7.57 10.78$$^{*}$$ (1.54) (1.15) Asset-backed CP (% of AUM) 2.01 6.84$$^{***}$$ (0.52) (1.00) Treated Control Variable (Initially constrained) (Initially unconstrained) Number of funds 39 62 Number of prime funds 20 45 Number of dealers per fund 9.43 8.09 (0.74) (0.57) % of foreign dealers per fund 60.71 61.66 (4.42) (3.14) Weighted CDS spread (%) 0.85 0.83 (0.04) (0.04) HHI 0.18 0.20 (0.03) (0.03) Share of largest dealer 0.26 0.28 (0.03) (0.03) Weighted Treasury/Agency repo rate (bps) 13.38 11.62 (1.23) (0.84) Weighted Treasury repo rate (bps) 10.38 10.02 (0.38) (0.02) AUM ( $\$$, billions) 27.6 14.4$$^{***}$$ (4.05) (2.12) Treasury repo (% of AUM) 15.48 5.97$$^{***}$$ (3.67) (1.70) Agency repo (% of AUM) 11.85 7.98 (2.64) (1.35) CD (% of AUM) 18.25 20.79 (3.21) (2.24) Treasury debt (% of AUM) 14.31 15.08 (2.92) (2.78) Financial CP (% of AUM) 7.57 10.78$$^{*}$$ (1.54) (1.15) Asset-backed CP (% of AUM) 2.01 6.84$$^{***}$$ (0.52) (1.00) This table reports summary statistics as of June 28, 2013 (the last quarter-end before the ONRRP began) for treated (initially constrained) and control (initially unconstrained) MMFs in our sample. All dealer-related variables (number of dealers, % foreign, CDS spread, HHI, and share of largest dealer) are based on a sample of only MMFs that have nonzero private repo volume (96 MMFs rather than the full 101). Rates are calculated at a fund complex level, and all other variables are at the individual MMF level. Statistical significance of the difference in averages: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. The two types of money funds in our sample, government and prime, each compose about half of the treatment group.13 Government MMFs are restricted to investing in Treasuries, agency debt, and repo backed by either Treasuries or agencies, while prime MMFs can invest in a broader range of assets, including bank deposits, commercial paper, corporate debt, and repo backed by non-Treasury and agency collateral. The control group also contains a mix of prime and government funds, although prime funds account for roughly two-thirds of this cohort. Treatment and control MMFs have no statistically significant differences in their dealer relationships. In particular, they are similar in their average number of dealers, percentage of lending to foreign dealers, riskiness of dealers (as measured by CDS spreads), concentration of lending across dealers (as measured by HHI), and share of the largest dealer.14 We take particular note of the groups’ similarity in their exposure to foreign dealers. Given geographical differences in the implementation of the Basel leverage ratio requirement, foreign dealers contract their repo borrowing on quarter-ends more than domestic dealers.15 Consequently, if constrained funds traded more with large foreign dealers that borrow less on quarter-ends, these funds would invest more in the ONRRP for this reason. Not only does this not seem to be the case given the similar dealer profiles of the two groups, but we also rule out such a borrower-driven explanation by controlling for dealer-time fixed effects in our robustness checks. The next two rows show that the treatment and control groups also receive the same rates in their tri-party repo transactions, both overall and in Treasury collateral repo only. However, the two MMF groups do differ in their size and asset allocation. Specifically, treated funds tend to be larger in terms of assets under management (AUM) and invest a relatively larger share of their assets in both Treasury and, to a lesser extent, agency repo. Control funds are more likely to invest in commercial paper, reflecting the higher proportion of prime funds. It seems reasonable that funds that both have more assets to invest and are more active in the repo market would be larger participants in the ONRRP and therefore more likely to be constrained by the initial counterparty cap, thus composing the treatment group. Separately, we compare the changes in AUM between the treatment and control funds in the weeks around cap increases after which there was at least one constrained fund. If the ability to invest in the ONRRP drives MMF investment flows, AUM may endogenously adjust to cap changes. In Figure 5, we use data from iMoneyNet to depict the weekly differences in the percentage change in AUM between control and treatment funds for each of the two weeks prior to and after the cap changes. No clear pattern is evident, and no statistical significance of the differences is achieved in any instance. In all cases, the percentage change in treatment funds’ AUM relative to control funds is very similar in the week after and week prior to a cap change. Thus, we conclude that neither group of funds witnessed abnormal inflows or outflows as a result of the cap changes. Figure 5 View largeDownload slide Changes in AUM around cap increases This figure depicts the difference in the percentage change in AUM between treatment and control funds for the 4 weeks around cap changes that occurred prior to each quarter-end observation. The point estimate represents the percentage increase (positive) or decrease (negative) in AUM witnessed by treatment funds relative to control (unconstrained) funds. Figure 5 View largeDownload slide Changes in AUM around cap increases This figure depicts the difference in the percentage change in AUM between treatment and control funds for the 4 weeks around cap changes that occurred prior to each quarter-end observation. The point estimate represents the percentage increase (positive) or decrease (negative) in AUM witnessed by treatment funds relative to control (unconstrained) funds. 4. Methods and Results In the subsections below, we describe the specific DD specifications and corresponding results that address four distinct questions regarding the effects of the Fed’s ONRRP facility. In Section 4.1, we examine whether, and to what extent, the ONRRP facility crowds out MMF investment in other asset categories. After demonstrating that the ONRRP causally crowds out private repo activity, Section 4.2 considers the effect of this substitution on MMFs’ dealer relationships. To invest more with the Fed, MMFs may turn away from some counterparties, and transact with fewer dealers. However, we find that the value of lending relationships in the repo market is high enough to prevent MMFs from dropping any of their dealer counterparties. Section 4.3 demonstrates that the ONRRP facility confers bargaining power on MMFs, thereby increasing dealer funding costs. Finally, we examine how dealers respond to the ONRRP-induced adverse funding shock in Section 4.4. We find that dealers that are more exposed to treated MMFs do not decrease their total repo borrowing but instead reallocate their borrowing within the repo market, turning mostly to repo backed by riskier collateral that they increasingly rely upon for net financing. This shift appears to be driven by dealers’ transition to non-MMF asset managers, as well as smaller ONRRP-ineligible MMF lenders. 4.1 Effects of the ONRRP on MMF asset allocation Our first goal is to estimate asset substitution effects in order to determine which money market segments experience a withdrawal of funding when MMFs invest in the ONRRP facility. The data used for this analysis come from the public quarter-end MMF filings of form N-MFP, discussed in Section 3. By aggregating MMF securities holdings from the N-MFP into distinct investment categories, we are able to track fund-level asset allocation over time, which we can then relate to changes in ONRRP participation. For this analysis, we sample the MMF asset portfolio holdings on quarter-end dates between December 2012 and June 2015. Table 2 reports average asset class holdings as a percentage of total AUM for the sample of all ONRRP-eligible MMF counterparties classified as either prime or government-only funds. The column on the left reports MMF’s asset shares as of Q3 2013, a few days after the introduction of the ONRRP facility. Summing the asset class shares reveals that the asset categories we consider account for about 70% of total AUM for MMFs in our sample. By Q3 2014, shown on the right of Table 2, there was a statistically significant reduction in the shares of MMF portfolios invested in Treasury-collateral repo, agency-collateral repo, and Treasury debt. As mentioned earlier, the especially large proportional declines in repo investments (about 43% and 39% for Treasury and agency repo, respectively) came amid regulatory pressures that led to an overall reduction in dealer balance sheets, limiting their participation in the repo market. Even outside of these regulatory issues, other factors causing MMFs’ planned or forced reductions in other asset classes could at least partly determine ONRRP participation. Consequently, an estimate of the substitution effects induced by ONRRP participation requires an identification strategy, such as the one described in Section 2, that relies on an exogenous change in MMFs’ ability or willingness to participate in the facility. Table 2 Money market fund investment shares Asset class 2013 Q3 2014 Q3 Fed ONRRP 3.36 13.27*** (0.37) (1.37) Treasury repo 9.39 5.34** (1.62) (1.00) Agency repo 7.92 4.83** (1.12) (0.73) CD 20.74 22.28 (1.92) (2.02) Treasury debt 14.35 9.83* (2.07) (1.92) Financial CP 9.16 8.80 (0.98) (0.86) Asset-backed CP 4.74 4.79 (0.67) (0.71) N 101 101 Asset class 2013 Q3 2014 Q3 Fed ONRRP 3.36 13.27*** (0.37) (1.37) Treasury repo 9.39 5.34** (1.62) (1.00) Agency repo 7.92 4.83** (1.12) (0.73) CD 20.74 22.28 (1.92) (2.02) Treasury debt 14.35 9.83* (2.07) (1.92) Financial CP 9.16 8.80 (0.98) (0.86) Asset-backed CP 4.74 4.79 (0.67) (0.71) N 101 101 This table reports the average asset holdings as a percentage of total assets under management for the pre-treatment quarter (2013 Q3) and the quarter just after the final increase in the counterparty cap (2014 Q3). Statistical significance of change in mean asset share: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 2 Money market fund investment shares Asset class 2013 Q3 2014 Q3 Fed ONRRP 3.36 13.27*** (0.37) (1.37) Treasury repo 9.39 5.34** (1.62) (1.00) Agency repo 7.92 4.83** (1.12) (0.73) CD 20.74 22.28 (1.92) (2.02) Treasury debt 14.35 9.83* (2.07) (1.92) Financial CP 9.16 8.80 (0.98) (0.86) Asset-backed CP 4.74 4.79 (0.67) (0.71) N 101 101 Asset class 2013 Q3 2014 Q3 Fed ONRRP 3.36 13.27*** (0.37) (1.37) Treasury repo 9.39 5.34** (1.62) (1.00) Agency repo 7.92 4.83** (1.12) (0.73) CD 20.74 22.28 (1.92) (2.02) Treasury debt 14.35 9.83* (2.07) (1.92) Financial CP 9.16 8.80 (0.98) (0.86) Asset-backed CP 4.74 4.79 (0.67) (0.71) N 101 101 This table reports the average asset holdings as a percentage of total assets under management for the pre-treatment quarter (2013 Q3) and the quarter just after the final increase in the counterparty cap (2014 Q3). Statistical significance of change in mean asset share: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. To generate a DD estimate of these substitution effects, we compare the control group of funds that were unconstrained by the introductory ONRRP counterparty cap to a treatment group of funds that faced a binding cap on Q3 2013. Since the cap was increased between each quarter-end, treated funds may no longer face a binding counterparty cap, or they could increase their ONRRP investment to such an extent that they are still bound by the new (higher) cap. Thus, we have two treatment subgroups denoted $$\text{UnboundTreatment}_{it}$$ and $$\text{BoundTreatment}_{it}$$, respectively. Specifically we estimate regressions of the following form: \begin{align} y_{it} &= \delta\cdot(\text{Post}_t\cdot\text{UnboundTreatment}_{it}) + \rho\cdot(\text{Post}_t\cdot\text{BoundTreatment}_{it}) \notag\\ &\quad + \beta_{i}\cdot\text{MMF}_{i} + \gamma_{t}\cdot\text{Quarter}_{t} + \varepsilon_{it}, \end{align} (1) \begin{equation} \text{where} \hspace{15mm} y_{it} = 100\cdot\frac{\text{AssetClass}_{it}}{\text{AUM}_{it}}. \nonumber \end{equation} In Equation (1), $$(\text{Post}_t\cdot\text{UnboundTreatment}_{it})$$ identifies the effect for our main treatment subgroup, taking a value of one for treated funds in the post-treatment period that are not bound by the cap in quarter t. All treated funds eventually enter this group. Thus, $$\delta$$ is the DD estimate of the marginal effect of increased ONRRP participation owing to an increase of the counterparty cap to a point at which it becomes nonbinding and the MMF faces an unconstrained investment decision. However, some treatment funds may continue to face a constraint even after the counterparty cap has been lifted and therefore compose a separate treatment subgroup. These funds also witness a sharp increase in their ability to participate in the ONRRP as a result of the exogenous cap increase, but we make a distinction between these still-bound treated funds as indicated by the $$\text{BoundTreatment}_{it}$$ dummy, which takes a value of one for funds that remained bound by the cap following an increase. Similar to $$\delta$$, $$\rho$$ estimates the marginal effect of ONRRP participation as a result of an increase in the counterparty cap, although $$\rho$$ measures the effect for funds in periods during which they continue to face a binding cap and hence a constrained investment decision. Equation (1) also includes fund fixed effects, as well as a full set of time fixed effects. Because an MMF’s total size may partially determine whether the counterparty cap is binding, we include MMF’s pre-treatment AUM in some specifications as a robustness check, interacted with a “post” dummy that takes a value of one after September 2013.16 Standard errors in Equation (1) are clustered at the fund complex level (Bertrand, Duflo, and Mullainathan 2004). In Figure 6, we show evidence of parallel trends in the asset shares between our control and treatment groups in the pre-treatment period. Although the patterns in Figure 6 suggest the parallel trends assumption is not violated, we nevertheless also include a robustness check that allows us to relax the parallel trends assumption as follows: Figure 6 View largeDownload slide Trends in asset shares This figure depicts the treatment effect for asset shares in the six quarters prior to implementation of the ONRRP. Thus, a positive value corresponds to quarters in which treated funds witness asset shares rising more (or falling less) than the asset share for control funds. A negative value corresponds to quarters in which treated funds witness asset shares rising less (or falling more) than the asset share for control funds. Figure 6 View largeDownload slide Trends in asset shares This figure depicts the treatment effect for asset shares in the six quarters prior to implementation of the ONRRP. Thus, a positive value corresponds to quarters in which treated funds witness asset shares rising more (or falling less) than the asset share for control funds. A negative value corresponds to quarters in which treated funds witness asset shares rising less (or falling more) than the asset share for control funds. \begin{align} y_{it} &= \delta\cdot(\text{Post}_t\cdot\text{UnboundTreatment}_{it}) + \rho\cdot(\text{Post}_t\cdot\text{BoundTreatment}_{it}) \notag\\ & \quad + \lambda_{i}\cdot(\text{MMF}_{i}\cdot t) + \beta_{i}\cdot\text{MMF}_{i} + \gamma_{t}\cdot\text{Quarter}_{t} + \varepsilon_{it}. \end{align} (2) In this specification, we add fund-specific time trends in order to capture any possible divergence in asset share trends between treatment and control funds. Although fund-level trends can weaken our results if the treatment effect emerges gradually over time, MMFs should reallocate their portfolios rapidly in response to cap increases. Additionally, we prefer to include this robustness check in light of the relatively long sample period required by the length of the ONRRP testing phase. Though MMFs’ asset mix does not typically change dramatically on a day-to-day basis, shifts in fund manager preferences or the market environment would cause some funds to adjust their investment mix over time. Table 3 reports results from estimating the specifications described above. Turning to the most basic specification in Column 1, we see that the coefficients on the treatment dummies ($$\text{Unbound}_{it}$$ and $$\text{Bound}_{it}$$) for the Fed RRP dependent variable verify the effect of treatment on MMF ONRRP investment. On average, an increase in the maximum counterparty cap led to treatment MMFs increasing their investment in the ONRRP facility by about 9.0% and 7.0% of assets for our two treatment subgroups. Table 3 Regression results: MMF asset substitution Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 9.01*** 10.05*** 9.14*** 10.10*** (1.86) (1.95) (2.43) (2.18) $$\text{Bound}_{it}$$ 6.99*** 8.59*** 7.28*** 9.20*** (1.63) (1.74) (1.70) (1.70) Treasury repo $$\text{Unbound}_{it}$$ –3.99*** –5.01*** –6.39*** –7.00*** (1.40) (1.62) (1.52) (1.73) $$\text{Bound}_{it}$$ –3.92*** –5.48*** –5.34*** –6.58*** (1.26) (1.52) (1.34) (1.46) Agency repo $$\text{Unbound}_{it}$$ –2.87* –3.09** –2.98** –3.13** (1.47) (1.54) (1.48) (1.54) $$\text{Bound}_{it}$$ –0.31 –0.65 –0.59 –0.88 (1.94) (2.10) (1.71) (1.79) CD $$\text{Unbound}_{it}$$ –2.29*** –2.87*** –1.00 –1.46 (0.88) (0.68) (1.27) (1.20) $$\text{Bound}_{it}$$ –2.38*** –3.27*** –1.62 –2.56* (0.79) (0.78) (1.19) (1.44) Treasury debt $$\text{Unbound}_{it}$$ –0.32 –0.47 0.20 0.29 (1.62) (1.67) (1.93) (2.06) $$\text{Bound}_{it}$$ 1.42 1.18 1.53 1.71 (1.41) (1.45) (1.48) (1.60) Financial CP $$\text{Unbound}_{it}$$ 0.74 0.68 0.40 0.26 (0.91) (0.99) (0.92) (0.92) $$\text{Bound}_{it}$$ –0.11 –0.20 –0.30 –0.59 (0.74) (0.93) (0.95) (0.91) Asset-backed CP $$\text{Unbound}_{it}$$ 0.25 0.28 0.20 0.24 (0.49) (0.46) (0.57) (0.54) $$\text{Bound}_{it}$$ 0.39 0.43 0.33 0.40 (0.45) (0.45) (0.48) (0.45) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ N 1,111 1,111 1,111 1,111 Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 9.01*** 10.05*** 9.14*** 10.10*** (1.86) (1.95) (2.43) (2.18) $$\text{Bound}_{it}$$ 6.99*** 8.59*** 7.28*** 9.20*** (1.63) (1.74) (1.70) (1.70) Treasury repo $$\text{Unbound}_{it}$$ –3.99*** –5.01*** –6.39*** –7.00*** (1.40) (1.62) (1.52) (1.73) $$\text{Bound}_{it}$$ –3.92*** –5.48*** –5.34*** –6.58*** (1.26) (1.52) (1.34) (1.46) Agency repo $$\text{Unbound}_{it}$$ –2.87* –3.09** –2.98** –3.13** (1.47) (1.54) (1.48) (1.54) $$\text{Bound}_{it}$$ –0.31 –0.65 –0.59 –0.88 (1.94) (2.10) (1.71) (1.79) CD $$\text{Unbound}_{it}$$ –2.29*** –2.87*** –1.00 –1.46 (0.88) (0.68) (1.27) (1.20) $$\text{Bound}_{it}$$ –2.38*** –3.27*** –1.62 –2.56* (0.79) (0.78) (1.19) (1.44) Treasury debt $$\text{Unbound}_{it}$$ –0.32 –0.47 0.20 0.29 (1.62) (1.67) (1.93) (2.06) $$\text{Bound}_{it}$$ 1.42 1.18 1.53 1.71 (1.41) (1.45) (1.48) (1.60) Financial CP $$\text{Unbound}_{it}$$ 0.74 0.68 0.40 0.26 (0.91) (0.99) (0.92) (0.92) $$\text{Bound}_{it}$$ –0.11 –0.20 –0.30 –0.59 (0.74) (0.93) (0.95) (0.91) Asset-backed CP $$\text{Unbound}_{it}$$ 0.25 0.28 0.20 0.24 (0.49) (0.46) (0.57) (0.54) $$\text{Bound}_{it}$$ 0.39 0.43 0.33 0.40 (0.45) (0.45) (0.48) (0.45) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ N 1,111 1,111 1,111 1,111 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF asset allocation. $$Unbound_{it}$$ takes a value of one for treated funds that initially experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$Bound_{it}$$ takes a value of one for treated funds that initially experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. Column 1 contains no other controls, and Columns 2 and 4 include AUM as a control for fund size. Columns 3 and 4 control for fund-specific trends. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 3 Regression results: MMF asset substitution Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 9.01*** 10.05*** 9.14*** 10.10*** (1.86) (1.95) (2.43) (2.18) $$\text{Bound}_{it}$$ 6.99*** 8.59*** 7.28*** 9.20*** (1.63) (1.74) (1.70) (1.70) Treasury repo $$\text{Unbound}_{it}$$ –3.99*** –5.01*** –6.39*** –7.00*** (1.40) (1.62) (1.52) (1.73) $$\text{Bound}_{it}$$ –3.92*** –5.48*** –5.34*** –6.58*** (1.26) (1.52) (1.34) (1.46) Agency repo $$\text{Unbound}_{it}$$ –2.87* –3.09** –2.98** –3.13** (1.47) (1.54) (1.48) (1.54) $$\text{Bound}_{it}$$ –0.31 –0.65 –0.59 –0.88 (1.94) (2.10) (1.71) (1.79) CD $$\text{Unbound}_{it}$$ –2.29*** –2.87*** –1.00 –1.46 (0.88) (0.68) (1.27) (1.20) $$\text{Bound}_{it}$$ –2.38*** –3.27*** –1.62 –2.56* (0.79) (0.78) (1.19) (1.44) Treasury debt $$\text{Unbound}_{it}$$ –0.32 –0.47 0.20 0.29 (1.62) (1.67) (1.93) (2.06) $$\text{Bound}_{it}$$ 1.42 1.18 1.53 1.71 (1.41) (1.45) (1.48) (1.60) Financial CP $$\text{Unbound}_{it}$$ 0.74 0.68 0.40 0.26 (0.91) (0.99) (0.92) (0.92) $$\text{Bound}_{it}$$ –0.11 –0.20 –0.30 –0.59 (0.74) (0.93) (0.95) (0.91) Asset-backed CP $$\text{Unbound}_{it}$$ 0.25 0.28 0.20 0.24 (0.49) (0.46) (0.57) (0.54) $$\text{Bound}_{it}$$ 0.39 0.43 0.33 0.40 (0.45) (0.45) (0.48) (0.45) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ N 1,111 1,111 1,111 1,111 Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 9.01*** 10.05*** 9.14*** 10.10*** (1.86) (1.95) (2.43) (2.18) $$\text{Bound}_{it}$$ 6.99*** 8.59*** 7.28*** 9.20*** (1.63) (1.74) (1.70) (1.70) Treasury repo $$\text{Unbound}_{it}$$ –3.99*** –5.01*** –6.39*** –7.00*** (1.40) (1.62) (1.52) (1.73) $$\text{Bound}_{it}$$ –3.92*** –5.48*** –5.34*** –6.58*** (1.26) (1.52) (1.34) (1.46) Agency repo $$\text{Unbound}_{it}$$ –2.87* –3.09** –2.98** –3.13** (1.47) (1.54) (1.48) (1.54) $$\text{Bound}_{it}$$ –0.31 –0.65 –0.59 –0.88 (1.94) (2.10) (1.71) (1.79) CD $$\text{Unbound}_{it}$$ –2.29*** –2.87*** –1.00 –1.46 (0.88) (0.68) (1.27) (1.20) $$\text{Bound}_{it}$$ –2.38*** –3.27*** –1.62 –2.56* (0.79) (0.78) (1.19) (1.44) Treasury debt $$\text{Unbound}_{it}$$ –0.32 –0.47 0.20 0.29 (1.62) (1.67) (1.93) (2.06) $$\text{Bound}_{it}$$ 1.42 1.18 1.53 1.71 (1.41) (1.45) (1.48) (1.60) Financial CP $$\text{Unbound}_{it}$$ 0.74 0.68 0.40 0.26 (0.91) (0.99) (0.92) (0.92) $$\text{Bound}_{it}$$ –0.11 –0.20 –0.30 –0.59 (0.74) (0.93) (0.95) (0.91) Asset-backed CP $$\text{Unbound}_{it}$$ 0.25 0.28 0.20 0.24 (0.49) (0.46) (0.57) (0.54) $$\text{Bound}_{it}$$ 0.39 0.43 0.33 0.40 (0.45) (0.45) (0.48) (0.45) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ N 1,111 1,111 1,111 1,111 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF asset allocation. $$Unbound_{it}$$ takes a value of one for treated funds that initially experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$Bound_{it}$$ takes a value of one for treated funds that initially experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. Column 1 contains no other controls, and Columns 2 and 4 include AUM as a control for fund size. Columns 3 and 4 control for fund-specific trends. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. For the next outcome variable—the share of assets invested in private Treasury-backed repo—we find that MMFs substitute out of private repo transactions in order to invest with the Fed. This asset category is likely the closest substitute for the ONRRP facility, as ONRRP transactions are also repo backed by Treasury collateral. The crowding out of private Treasury repo for both treatment groups confirms the substitutability of these transactions for ONRRP investment. Agency repo investments also fall for MMFs in our main treatment subgroup, but appear insensitive to cap increases when funds remain bound by the new counterparty caps. This suggests that MMFs only substitute away from agency-backed repo after they have first substituted away from other assets. In other words, a fund that remains constrained by a larger counterparty cap will first shift out of Treasury repo, but as the cap is increased enough for the fund to find itself unconstrained, it will eventually begin to shift investment away from agency repo as well. Deposits (CD) show a similar pattern to Treasury repo. The results are evident for both treatment subgroups, suggesting that money funds are willing to move funds out of short-term deposits in order to invest in a reverse repo with the Fed. Summing the coefficient estimates over these four asset categories reveals that the increase in the ONRRP facility for both treatment subgroups is entirely offset by reductions in Treasury repo, agency repo, and deposits. Dividing the point estimates of the effects on these asset classes by the estimated increase in the ONRRP investment yields rates of substitution to the ONRRP from other assets. For example, using the point estimates for the $$\text{UnboundTreatment}_{it}$$ group in our baseline specification reported in Column 1, we find that an increase in ONRRP investment of 1% of AUM corresponds to reductions in Treasury repo, agency repo, and deposits of about 0.45%, 0.30%, and 0.25% of AUM, respectively. Naively applying these figures to the $\$$ 167 billion MMFs invested in Fed RRP on December 31, 2014, implies that private tri-party Treasury collateral repo volume was lower by $\$$75 billion (roughly 11% of the traded volume on that day), and agency collateral repo was lower by $\$$50 billion (or 9% of the traded volume). Reductions in deposits driven by the ONRRP investment represent a far smaller share of the overall deposit market. The remaining asset classes show no statistically significant response to increases in the maximum bid amount, with point estimates that generally lie close to zero. Very similar results are presented in Columns 2 through 4—which add controls for fund size and/or fund-level trends—although the deposits result weakens somewhat in certain specifications. In unreported results, we find that using a weighted least squares estimator yields identical conclusions. Overall, these results show that investment in the Fed’s ONRRP facility comes at the expense of private investments in Treasury repo, agency repo, and deposits. However, the overall effect on financial stability is unclear. On one hand, the funding that is withdrawn from dealers and banks in favor of the ONRRP could potentially be worrisome to a central bank that seeks to minimize its footprint in financial markets. Moreover, the existence of the Fed’s ONRRP facility may increase the likelihood of a shift in risk sentiment spurring a flight-to-quality, with MMFs flocking to the Fed’s ONRRP facility (Frost et al. 2015).17 In such an event, substitution out of other asset classes such as commercial paper could emerge, starving regular funding from an additional class of borrowers with potentially destabilizing effects on financial markets (Kacperczyk and Schnabl 2013). On the other hand, crowding out privately issued repo may ameliorate overall financial stability by reducing the externalities associated with a large private repo market that is prone to runs (Carlson et al. 2016; Frost et al. 2015). As we will later show, however, dealers do not change their total repo borrowing in response to the reduction in MMF lending that we demonstrate here. To further support the causal interpretation of the estimated effects reported above, we conduct two separate placebo tests. Our first test includes a placebo treatment dummy that takes a value of one for treatment group funds in the quarter immediately after they become unconstrained. In the context of the stylized example presented in Figure 4, this dummy would be zero for fund 2 until the third period. For fund 3, this placebo treatment would be zero until period 4 (not shown). Since these funds were unconstrained by the counterparty cap in the period prior to the placebo treatment, we should expect a null result. However, if previously constrained funds were merely in the process of shifting their asset allocation for a reason unrelated to the increases in the ONRRP caps, we would expect to see statistically significant effects similar to our estimated treatment effects. In panel A of Table 4, we report the results of this placebo test for the asset categories for which we found evidence of substitution. Comparing the point estimates for the $$\text{Placebo}_{it}$$ treatment group with the $$\text{Unbound}_{it}$$ treatment group, we can see that funds that become unconstrained do not differentially increase their ONRRP participation in the periods after the cap no longer binds the funds’ investment decision. In other words, the measured effect for the $$\text{Unbound}_{it}$$ treatment group coincides entirely with the increase in the counterparty cap. Therefore, it does not appear that our results are driven by other factors, such as an uneven withdrawal of dealer borrowing that disproportionately affected the funds in our treatment group. Table 4 Regression results: Placebo tests of MMF asset substitution A. Placebo treatment within sample Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 8.40*** 9.52*** 9.56*** 10.43*** (1.78) (1.66) (2.53) (2.41) $$\text{Bound}_{it}$$ 6.82*** 8.43*** 7.79*** 9.60*** (1.50) (1.51) (2.11) (2.25) $$\text{Placebo}_{it}$$ 0.96 0.81 1.53 1.29 (1.89) (1.88) (2.57) (2.55) Treasury repo $$\text{Unbound}_{it}$$ –4.20*** –5.32*** –6.58*** –7.15*** (1.09) (1.40) (1.67) (1.83) $$\text{Bound}_{it}$$ –3.98*** –5.58*** –5.57*** –6.75*** (1.19) (1.46) (1.51) (1.58) $$\text{Placebo}_{it}$$ 0.33 0.48 –0.71 –0.55 (1.15) (1.13) (1.48) (1.49) Agency repo $$\text{Unbound}_{it}$$ –2.71* –2.95* –3.22** –3.34** (1.39) (1.47) (1.54) (1.60) $$\text{Bound}_{it}$$ –0.26 –0.61 –0.88 –1.15 (1.91) (2.08) (1.71) (1.79) $$\text{Placebo}_{it}$$ –0.26 –0.23 –0.89 –0.85 (0.57) (0.56) (0.81) (0.80) CD $$\text{Unbound}_{it}$$ –1.81** –2.44*** –0.95 –1.39 (0.86) (0.66) (1.31) (1.21) $$\text{Bound}_{it}$$ –2.24** –3.15*** –1.56 –2.47* (0.85) (0.84) (1.11) (1.35) $$\text{Placebo}_{it}$$ –0.75 –0.66 0.17 0.29 (0.63) (0.63) (0.56) (0.57) B. Placebo treatment in pre-ONRRP sample Treasury repo $$\text{Placebo}_{it}$$ 0.27 0.05 1.86 1.64 (1.89) (2.18) (1.82) (1.91) Agency repo $$\text{Placebo}_{it}$$ 0.16 –0.03 1.57 1.33 (0.81) (0.95) (1.88) (2.04) CD $$\text{Placebo}_{it}$$ –0.20 0.54 1.43 2.81 (1.49) (1.44) (1.66) (1.70) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ A. Placebo treatment within sample Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 8.40*** 9.52*** 9.56*** 10.43*** (1.78) (1.66) (2.53) (2.41) $$\text{Bound}_{it}$$ 6.82*** 8.43*** 7.79*** 9.60*** (1.50) (1.51) (2.11) (2.25) $$\text{Placebo}_{it}$$ 0.96 0.81 1.53 1.29 (1.89) (1.88) (2.57) (2.55) Treasury repo $$\text{Unbound}_{it}$$ –4.20*** –5.32*** –6.58*** –7.15*** (1.09) (1.40) (1.67) (1.83) $$\text{Bound}_{it}$$ –3.98*** –5.58*** –5.57*** –6.75*** (1.19) (1.46) (1.51) (1.58) $$\text{Placebo}_{it}$$ 0.33 0.48 –0.71 –0.55 (1.15) (1.13) (1.48) (1.49) Agency repo $$\text{Unbound}_{it}$$ –2.71* –2.95* –3.22** –3.34** (1.39) (1.47) (1.54) (1.60) $$\text{Bound}_{it}$$ –0.26 –0.61 –0.88 –1.15 (1.91) (2.08) (1.71) (1.79) $$\text{Placebo}_{it}$$ –0.26 –0.23 –0.89 –0.85 (0.57) (0.56) (0.81) (0.80) CD $$\text{Unbound}_{it}$$ –1.81** –2.44*** –0.95 –1.39 (0.86) (0.66) (1.31) (1.21) $$\text{Bound}_{it}$$ –2.24** –3.15*** –1.56 –2.47* (0.85) (0.84) (1.11) (1.35) $$\text{Placebo}_{it}$$ –0.75 –0.66 0.17 0.29 (0.63) (0.63) (0.56) (0.57) B. Placebo treatment in pre-ONRRP sample Treasury repo $$\text{Placebo}_{it}$$ 0.27 0.05 1.86 1.64 (1.89) (2.18) (1.82) (1.91) Agency repo $$\text{Placebo}_{it}$$ 0.16 –0.03 1.57 1.33 (0.81) (0.95) (1.88) (2.04) CD $$\text{Placebo}_{it}$$ –0.20 0.54 1.43 2.81 (1.49) (1.44) (1.66) (1.70) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF asset allocation, as described in Table 3. In the table above, panel A uses the same sample as Table 3, but includes a placebo treatment variable, $$\text{Placebo}_{it}$$, which takes a value of one for treated funds beginning in the period after these funds become unconstrained. Panel B uses a sample of the same funds for the 11 quarters prior to the introduction of the ONRRP. In panel B, $$\text{Placebo}_{it}$$ takes a value of one for treated funds beginning in December 2011, five quarters after the beginning of the sample. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 4 Regression results: Placebo tests of MMF asset substitution A. Placebo treatment within sample Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 8.40*** 9.52*** 9.56*** 10.43*** (1.78) (1.66) (2.53) (2.41) $$\text{Bound}_{it}$$ 6.82*** 8.43*** 7.79*** 9.60*** (1.50) (1.51) (2.11) (2.25) $$\text{Placebo}_{it}$$ 0.96 0.81 1.53 1.29 (1.89) (1.88) (2.57) (2.55) Treasury repo $$\text{Unbound}_{it}$$ –4.20*** –5.32*** –6.58*** –7.15*** (1.09) (1.40) (1.67) (1.83) $$\text{Bound}_{it}$$ –3.98*** –5.58*** –5.57*** –6.75*** (1.19) (1.46) (1.51) (1.58) $$\text{Placebo}_{it}$$ 0.33 0.48 –0.71 –0.55 (1.15) (1.13) (1.48) (1.49) Agency repo $$\text{Unbound}_{it}$$ –2.71* –2.95* –3.22** –3.34** (1.39) (1.47) (1.54) (1.60) $$\text{Bound}_{it}$$ –0.26 –0.61 –0.88 –1.15 (1.91) (2.08) (1.71) (1.79) $$\text{Placebo}_{it}$$ –0.26 –0.23 –0.89 –0.85 (0.57) (0.56) (0.81) (0.80) CD $$\text{Unbound}_{it}$$ –1.81** –2.44*** –0.95 –1.39 (0.86) (0.66) (1.31) (1.21) $$\text{Bound}_{it}$$ –2.24** –3.15*** –1.56 –2.47* (0.85) (0.84) (1.11) (1.35) $$\text{Placebo}_{it}$$ –0.75 –0.66 0.17 0.29 (0.63) (0.63) (0.56) (0.57) B. Placebo treatment in pre-ONRRP sample Treasury repo $$\text{Placebo}_{it}$$ 0.27 0.05 1.86 1.64 (1.89) (2.18) (1.82) (1.91) Agency repo $$\text{Placebo}_{it}$$ 0.16 –0.03 1.57 1.33 (0.81) (0.95) (1.88) (2.04) CD $$\text{Placebo}_{it}$$ –0.20 0.54 1.43 2.81 (1.49) (1.44) (1.66) (1.70) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ A. Placebo treatment within sample Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 8.40*** 9.52*** 9.56*** 10.43*** (1.78) (1.66) (2.53) (2.41) $$\text{Bound}_{it}$$ 6.82*** 8.43*** 7.79*** 9.60*** (1.50) (1.51) (2.11) (2.25) $$\text{Placebo}_{it}$$ 0.96 0.81 1.53 1.29 (1.89) (1.88) (2.57) (2.55) Treasury repo $$\text{Unbound}_{it}$$ –4.20*** –5.32*** –6.58*** –7.15*** (1.09) (1.40) (1.67) (1.83) $$\text{Bound}_{it}$$ –3.98*** –5.58*** –5.57*** –6.75*** (1.19) (1.46) (1.51) (1.58) $$\text{Placebo}_{it}$$ 0.33 0.48 –0.71 –0.55 (1.15) (1.13) (1.48) (1.49) Agency repo $$\text{Unbound}_{it}$$ –2.71* –2.95* –3.22** –3.34** (1.39) (1.47) (1.54) (1.60) $$\text{Bound}_{it}$$ –0.26 –0.61 –0.88 –1.15 (1.91) (2.08) (1.71) (1.79) $$\text{Placebo}_{it}$$ –0.26 –0.23 –0.89 –0.85 (0.57) (0.56) (0.81) (0.80) CD $$\text{Unbound}_{it}$$ –1.81** –2.44*** –0.95 –1.39 (0.86) (0.66) (1.31) (1.21) $$\text{Bound}_{it}$$ –2.24** –3.15*** –1.56 –2.47* (0.85) (0.84) (1.11) (1.35) $$\text{Placebo}_{it}$$ –0.75 –0.66 0.17 0.29 (0.63) (0.63) (0.56) (0.57) B. Placebo treatment in pre-ONRRP sample Treasury repo $$\text{Placebo}_{it}$$ 0.27 0.05 1.86 1.64 (1.89) (2.18) (1.82) (1.91) Agency repo $$\text{Placebo}_{it}$$ 0.16 –0.03 1.57 1.33 (0.81) (0.95) (1.88) (2.04) CD $$\text{Placebo}_{it}$$ –0.20 0.54 1.43 2.81 (1.49) (1.44) (1.66) (1.70) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF asset allocation, as described in Table 3. In the table above, panel A uses the same sample as Table 3, but includes a placebo treatment variable, $$\text{Placebo}_{it}$$, which takes a value of one for treated funds beginning in the period after these funds become unconstrained. Panel B uses a sample of the same funds for the 11 quarters prior to the introduction of the ONRRP. In panel B, $$\text{Placebo}_{it}$$ takes a value of one for treated funds beginning in December 2011, five quarters after the beginning of the sample. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Panel B of Table 4 presents the results of a second placebo test. In this test, we use an 11 quarter sample period (the same length as that used in our main analysis in Table 3) that ends in June 2013, the quarter immediately prior to the introduction of the ONRRP. In this exercise, our placebo treatment dummy takes a value of one for the treated funds in our sample as of December 2011.18 As demonstrated in panel B, there is no evidence of differential investment in Treasury repo, agency repo, or deposits. Thus, these results confirm that there are no systematic differences in investment patterns between the treatment and control MMFs that are driven by causes other the changing availability of the ONRRP. Moreover, because this sample period spans the period before the introduction of the ONRRP, these results offer further evidence of parallel trends in the asset shares between the two groups of funds. In the remainder of the paper, we solely focus on the effects of the Fed’s ONRRP facility on the repo market. 4.1.1 Robustness using trade-level data In this subsection, we establish the robustness of the substitution effects of the ONRRP outlined above for both Treasury and agency-backed repo. It may be the case that the substitution results achieved in section 4.1 are merely driven by differences in demand for repo funding by dealers that borrowed from MMFs in our treatment group. For example, treated funds may trade with dealers that are winding down their repo positions more rapidly than the control funds’ dealers. Similarly, treated funds may also trade more heavily with certain foreign dealers, who, as discussed above, withdraw more heavily from the repo market on quarter-ends.19 Therefore, treated funds may be forced to invest in the ONRRP more than their control fund counterparts in response to differences in borrower demand, which could bias our estimated treatment effects. However, it is important to note that treated and control funds trading with different sets of dealers is not necessarily sufficient for our results to be biased. Rather, treated funds’ dealers would also need to disproportionately pull back from the repo market after the introduction of the ONRRP and in concert with the timing of the increases in the ONRRP cap. To address these concerns, we construct a panel of borrower-lender pairs using the data available in the N-MFP reports. By identifying the dealer counterparty to each repo transaction, we are able to then control for the borrower-quarter specific demand for repo (Khwaja and Mian 2008). Thus, estimating variants of the following regression specification on quarter-ends allows us to address the potential threats to the causal estimates that stem from the behavior of repo borrowers. \begin{align} y_{ijt} &= \delta\cdot(\text{Post}_t\cdot\text{UnboundTreatment}_{it}) + \rho\cdot(\text{Post}_t\cdot\text{BoundTreatment}_{it}) \notag \\ & \quad +\lambda_{ij}\cdot(\text{MMF-Dealer}_{ij}\cdot t) + \phi_{jt}\cdot(\text{Dealer}_{j}\cdot\text{Quarter}_{t}) \notag\\ &\quad + \beta_{ij}\cdot\text{MMF-Dealer}_{ij} + \varepsilon_{ijt} \end{align} (3) In Equation (3), $$y_{ijt}$$ represents the asset share of the day t repo activity between MMF i and dealer j. Here, we are now able to control for trends in activity between individual trading partners $$(\text{MMF-Dealer}_{ij}\cdot t)$$, as well as changes in dealer-specific borrowing demand in each quarter $$(\text{Dealer}_{j}\cdot\text{Quarter}_{t})$$. Consequently, we are able to measure the difference in private repo lending after a cap increase between a control fund and a treated fund for a given date, holding the borrower constant. As shown in Table 5, our conclusions remain unchanged even when accounting for the extent to which dealer behavior can explain the results. Comparing the sign and significance for both Treasury repo (panel A) and agency repo (panel B) to the results in Table 3, we see the findings are nearly identical. Of course, the magnitude of the coefficients is somewhat lower, as we are now estimating the effects of treatment on relationship-level asset substitution. These results strongly support our earlier interpretation that MMFs in fact substitute out of private repo in favor of the ONRRP. Table 5 Regression results: MMF asset substitution at the relationship level A. Treasury repo (1) (2) (3) (4) $$\text{Unbound}_{it}$$ –0.39*** –0.39*** –0.39** –0.36** (0.13) (0.15) (0.16) (0.18) $$\text{Bound}_{it}$$ –0.32** –0.31** –0.33*** –0.31*** (0.12) (0.12) (0.12) (0.11) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 8,360 8,360 8,360 8,360 Adj. R$$^2$$ 0.45 0.48 0.60 0.61 B. Agency repo $$\text{Unbound}_{it}$$ –0.29** –0.33** –0.38*** –0.40*** (0.12) (0.15) (0.14) (0.14) $$\text{Bound}_{it}$$ 0.00 –0.01 –0.07 –0.07 (0.17) (0.21) (0.14) (0.16) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 9,251 9,251 9,251 9,251 Adj. R$$^2$$ 0.49 0.51 0.66 0.67 A. Treasury repo (1) (2) (3) (4) $$\text{Unbound}_{it}$$ –0.39*** –0.39*** –0.39** –0.36** (0.13) (0.15) (0.16) (0.18) $$\text{Bound}_{it}$$ –0.32** –0.31** –0.33*** –0.31*** (0.12) (0.12) (0.12) (0.11) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 8,360 8,360 8,360 8,360 Adj. R$$^2$$ 0.45 0.48 0.60 0.61 B. Agency repo $$\text{Unbound}_{it}$$ –0.29** –0.33** –0.38*** –0.40*** (0.12) (0.15) (0.14) (0.14) $$\text{Bound}_{it}$$ 0.00 –0.01 –0.07 –0.07 (0.17) (0.21) (0.14) (0.16) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 9,251 9,251 9,251 9,251 Adj. R$$^2$$ 0.49 0.51 0.66 0.67 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF repo trades with dealers. $$\textit{Unbound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$\textit{Bound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. All specifications include relationship fixed effects. Columns 1 and 3 contain quarter fixed effects, and Columns 2 and 4 include dealer*time fixed effects. Columns 3 and 4 additionally control for trends specific to each unique dealer-fund pair. Standard errors are clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 5 Regression results: MMF asset substitution at the relationship level A. Treasury repo (1) (2) (3) (4) $$\text{Unbound}_{it}$$ –0.39*** –0.39*** –0.39** –0.36** (0.13) (0.15) (0.16) (0.18) $$\text{Bound}_{it}$$ –0.32** –0.31** –0.33*** –0.31*** (0.12) (0.12) (0.12) (0.11) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 8,360 8,360 8,360 8,360 Adj. R$$^2$$ 0.45 0.48 0.60 0.61 B. Agency repo $$\text{Unbound}_{it}$$ –0.29** –0.33** –0.38*** –0.40*** (0.12) (0.15) (0.14) (0.14) $$\text{Bound}_{it}$$ 0.00 –0.01 –0.07 –0.07 (0.17) (0.21) (0.14) (0.16) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 9,251 9,251 9,251 9,251 Adj. R$$^2$$ 0.49 0.51 0.66 0.67 A. Treasury repo (1) (2) (3) (4) $$\text{Unbound}_{it}$$ –0.39*** –0.39*** –0.39** –0.36** (0.13) (0.15) (0.16) (0.18) $$\text{Bound}_{it}$$ –0.32** –0.31** –0.33*** –0.31*** (0.12) (0.12) (0.12) (0.11) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 8,360 8,360 8,360 8,360 Adj. R$$^2$$ 0.45 0.48 0.60 0.61 B. Agency repo $$\text{Unbound}_{it}$$ –0.29** –0.33** –0.38*** –0.40*** (0.12) (0.15) (0.14) (0.14) $$\text{Bound}_{it}$$ 0.00 –0.01 –0.07 –0.07 (0.17) (0.21) (0.14) (0.16) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 9,251 9,251 9,251 9,251 Adj. R$$^2$$ 0.49 0.51 0.66 0.67 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF repo trades with dealers. $$\textit{Unbound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$\textit{Bound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. All specifications include relationship fixed effects. Columns 1 and 3 contain quarter fixed effects, and Columns 2 and 4 include dealer*time fixed effects. Columns 3 and 4 additionally control for trends specific to each unique dealer-fund pair. Standard errors are clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. 4.2 Effects of the ONRRP on trading relationships Next, we consider how the changes in MMF asset allocation seen in the previous section affect funds’ relationships with dealers in the tri-party repo market. We are primarily interested in whether MMFs transition to smaller trading networks as a result of the ONRRP, possibly increasing financial fragility, and whether MMFs’ differentially substitute away from certain dealers (e.g., those that are riskier counterparties). We again look at quarter-end MMF repo activity from the N-MFP data between December 2012 and June 2015. In particular, we focus on four measures of MMF lending relationships: number of dealers, weighted CDS spread, Herfindahl index (HHI), and share of lending to the largest dealer. Number of dealers is simply the total number of dealers that an MMF trades with on a given quarter-end date. As funds substitute into the ONRRP and away from private repo, they may do so either by dropping certain counterparties or by simply reducing their volume with one or several of their existing counterparties. Weighted CDS spread measures the volume-weighted CDS spread of all J dealers with which fund i trades.20 It is defined as \begin{equation} \text{Weighted CDS Spread}_{it} = \frac{\sum\limits_{j=1}^{J}\text{CDS Spread}_{jt}\cdot\text{Repo Volume}_{ijt}}{\text{Repo Volume}_{it}}. \end{equation} (4) Funds that invest more with the Fed through the ONRRP may favor some dealer counterparties over others and prefer to reduce trading activity with less-favored counterparties. In particular, MMFs may substitute away from low-risk dealers since they are the most direct substitute for the Fed, which is a risk-free counterparty. Conversely, they could substitute away from their higher-risk dealers in an effort to reduce their overall risk profile, although, as documented by Krishnamurthy, Nagel, and Orlov (2014), repo funding with Treasury and agency collateral appears less sensitive to dealers’ perceived default risk. HHI measures concentration among a funds’ borrowers excluding the Fed and is defined as \begin{equation} \text{HHI}_{it} = \sum\limits_{j=1}^{J}\bigg[\frac{\text{Repo Volume}_{ijt}}{\text{Repo Volume}_{it}}\bigg]^{2}. \end{equation} (5) HHI ranges from 1/J to 1, with lower values representing less concentration among dealers whereas higher values correspond to more highly concentrated lending. Lastly, we examine the share of lending to a fund’s largest dealer, calculated simply as the percentage of each MMF’s total repo volume that is transacted with the dealer to which it lends most heavily. These last two measures are used to assess whether MMFs differentially substitute away from dealers based on the volume traded between the pair. MMFs may be less likely to substitute away from those dealers with which they have a strong relationship, as measured by volume. For each of the four measures of dealer relationships, we consider MMFs’ Treasury- and agency-collateral repo activity separately. There may be stronger results for Treasury repo since, as seen in the previous section, the substitution effects are stronger for Treasury repo. For each of our measures of counterparty relationships, we present evidence of parallel trends in Figure 7, and proceed by estimating the following regression: Figure 7 View largeDownload slide Trends in measures of counterparty relationships This figure presents the treatment effect for measures of MMF counterparty relationships in the six quarters prior to implementation of the ONRRP. Thus, a positive value corresponds to quarters in which treated funds witness counterparty relationship measures rising more (or falling less) than the measure for control funds. A negative value corresponds to quarters in which treated funds witness counterparty relationship measures rising less (or falling more) than that of control funds. Figure 7 View largeDownload slide Trends in measures of counterparty relationships This figure presents the treatment effect for measures of MMF counterparty relationships in the six quarters prior to implementation of the ONRRP. Thus, a positive value corresponds to quarters in which treated funds witness counterparty relationship measures rising more (or falling less) than the measure for control funds. A negative value corresponds to quarters in which treated funds witness counterparty relationship measures rising less (or falling more) than that of control funds. \begin{align} y_{it} &= \delta\cdot(\text{Post}_t\cdot\text{UnboundTreatment}_{it}) + \rho\cdot(\text{Post}_t\cdot\text{BoundTreatment}_{it}) \notag\\ &\quad + \lambda_{i}\cdot(\text{MMF}_{i}\cdot t) + \beta_{i}\cdot\text{MMF}_{i} + \gamma_{t}\cdot\text{Quarter}_{t} + \varepsilon_{it}, \end{align} (6) \begin{align*} \text{where}\hspace{15mm} y_{it} &\in \left\{\text{Number of Dealers}_{it},\text{Weighted CDS Spread}_{it},\right.\\ &\qquad\left.\text{HHI}_{it},\text{Share Largest Dealer}_{it}\right\}. \end{align*} The results are shown in Table 6. There are no statistically significant differences in either the number of dealers or the weighted CDS spread. This suggests that MMFs continue trading with all of their original counterparties and do not differentially substitute away from or toward riskier dealers, consistent with the patterns shown in Krishnamurthy, Nagel, and Orlov (2014). In unreported results, we confirm that MMFs are not likely to sever dealer relationships by estimating the probability of a trade between the unique dealer-fund trading pairs in our sample. We find that, after a counterparty cap increase, initially bound treatment funds are no less likely to conduct a trade with one of their trading partners than funds that were never constrained by the cap. Table 6 Regression results: MMF-dealer relationships A. No fund trends Number of dealers Weighted CDS spread HHI Share largest dealer (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ –0.07 –0.91 0.10 0.01 –0.04 –0.01 –0.02 –0.002 (0.43) (0.84) (0.08) (0.06) (0.05) (0.05) (0.05) (0.05) $$\text{Bound}_{it}$$ 0.31 0.61 0.13 –0.005 –0.08 –0.12* –0.06 –0.11* (0.50) (0.68) (0.08) (0.06) (0.10) (0.06) (0.09) (0.06) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.94 0.93 0.90 0.94 0.89 0.87 0.92 0.91 B: With fund trends (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ 0.25 –0.34 0.04 0.05 –0.16* –0.09 –0.12 –0.07 (0.69) (0.68) (0.10) (0.11) (0.08) (0.07) (0.07) (0.06) $$\text{Bound}_{it}$$ 0.45 0.84 0.08 0.01 –0.13 –0.16** –0.10 –0.14** (0.58) (0.82) (0.08) (0.07) (0.10) (0.07) (0.09) (0.07) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.95 0.95 0.92 0.96 0.91 0.90 0.94 0.93 A. No fund trends Number of dealers Weighted CDS spread HHI Share largest dealer (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ –0.07 –0.91 0.10 0.01 –0.04 –0.01 –0.02 –0.002 (0.43) (0.84) (0.08) (0.06) (0.05) (0.05) (0.05) (0.05) $$\text{Bound}_{it}$$ 0.31 0.61 0.13 –0.005 –0.08 –0.12* –0.06 –0.11* (0.50) (0.68) (0.08) (0.06) (0.10) (0.06) (0.09) (0.06) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.94 0.93 0.90 0.94 0.89 0.87 0.92 0.91 B: With fund trends (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ 0.25 –0.34 0.04 0.05 –0.16* –0.09 –0.12 –0.07 (0.69) (0.68) (0.10) (0.11) (0.08) (0.07) (0.07) (0.06) $$\text{Bound}_{it}$$ 0.45 0.84 0.08 0.01 –0.13 –0.16** –0.10 –0.14** (0.58) (0.82) (0.08) (0.07) (0.10) (0.07) (0.09) (0.07) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.95 0.95 0.92 0.96 0.91 0.90 0.94 0.93 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF relationships in the repo market. $$\textit{Unbound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$\textit{Bound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. Dependent variables include the number of dealers per fund, the weighted CDS spread of dealers that trade with each fund, the Herfindahl-Hirschman index (HHI) for each fund (calculated using volume transacted with each dealer) and the percentage of volume traded with the fund’s largest dealer. For each dependent variable, Column 1 includes Treasury collateral repo transactions, and Column 2 includes agency collateral repo. Panel A excludes fund-level trends from the specification, whereas the specifications in panel B include fund trends. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 6 Regression results: MMF-dealer relationships A. No fund trends Number of dealers Weighted CDS spread HHI Share largest dealer (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ –0.07 –0.91 0.10 0.01 –0.04 –0.01 –0.02 –0.002 (0.43) (0.84) (0.08) (0.06) (0.05) (0.05) (0.05) (0.05) $$\text{Bound}_{it}$$ 0.31 0.61 0.13 –0.005 –0.08 –0.12* –0.06 –0.11* (0.50) (0.68) (0.08) (0.06) (0.10) (0.06) (0.09) (0.06) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.94 0.93 0.90 0.94 0.89 0.87 0.92 0.91 B: With fund trends (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ 0.25 –0.34 0.04 0.05 –0.16* –0.09 –0.12 –0.07 (0.69) (0.68) (0.10) (0.11) (0.08) (0.07) (0.07) (0.06) $$\text{Bound}_{it}$$ 0.45 0.84 0.08 0.01 –0.13 –0.16** –0.10 –0.14** (0.58) (0.82) (0.08) (0.07) (0.10) (0.07) (0.09) (0.07) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.95 0.95 0.92 0.96 0.91 0.90 0.94 0.93 A. No fund trends Number of dealers Weighted CDS spread HHI Share largest dealer (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ –0.07 –0.91 0.10 0.01 –0.04 –0.01 –0.02 –0.002 (0.43) (0.84) (0.08) (0.06) (0.05) (0.05) (0.05) (0.05) $$\text{Bound}_{it}$$ 0.31 0.61 0.13 –0.005 –0.08 –0.12* –0.06 –0.11* (0.50) (0.68) (0.08) (0.06) (0.10) (0.06) (0.09) (0.06) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.94 0.93 0.90 0.94 0.89 0.87 0.92 0.91 B: With fund trends (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ 0.25 –0.34 0.04 0.05 –0.16* –0.09 –0.12 –0.07 (0.69) (0.68) (0.10) (0.11) (0.08) (0.07) (0.07) (0.06) $$\text{Bound}_{it}$$ 0.45 0.84 0.08 0.01 –0.13 –0.16** –0.10 –0.14** (0.58) (0.82) (0.08) (0.07) (0.10) (0.07) (0.09) (0.07) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.95 0.95 0.92 0.96 0.91 0.90 0.94 0.93 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF relationships in the repo market. $$\textit{Unbound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$\textit{Bound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. Dependent variables include the number of dealers per fund, the weighted CDS spread of dealers that trade with each fund, the Herfindahl-Hirschman index (HHI) for each fund (calculated using volume transacted with each dealer) and the percentage of volume traded with the fund’s largest dealer. For each dependent variable, Column 1 includes Treasury collateral repo transactions, and Column 2 includes agency collateral repo. Panel A excludes fund-level trends from the specification, whereas the specifications in panel B include fund trends. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. However, there is some statistical significance in the regressions for HHI and the share of the largest dealer, especially when fund trends are included in panel B, with consistently negative point estimates throughout. Thus, funds’ dealer-concentration HHIs fall as funds become less constrained by the ONRRP cap. In Section 4.1, we demonstrated that funds’ repo volume decreases in response to the ONRRP, while we show here that the number of repo borrowers does not change. Therefore, the HHI results suggest that MMFs appear to be substituting away from those counterparties with which they trade the most, while preserving dealer relationships even for those dealers that account for only a small share of total trading volume. This interpretation is consistent with the decrease in the share of lending to a fund’s largest counterparty shown in the rightmost columns of Table 6. Thus, MMF participation in the ONRRP evidently results in a more even distribution of MMF’s remaining private repo investment. In total, these results suggest that lending relationships are very important in the repo market and MMFs have a desire to maintain their relationships with dealers. In particular, funds do not drop any of their dealer counterparties and do not differentially substitute away from dealers according to their risk of default. Rather, MMFs adjust trading volume across their existing counterparties by substituting away from their largest counterparties to some degree. Importantly, the results of this analysis suggest that, despite the ONRRP disintermediating private repo volume, it does not significantly affect the relationship structure of ONRRP-eligible MMFs in the tri-party repo market. Of course, these results are local to substitution effects stemming from increases in the counterparty caps only, and relationships may in fact be severed by ONRRP participation in times of severe financial stress. On the other hand, the importance of trading relationships in repo markets that we identify here may contribute to the resiliency of the repo market as a whole if it reduces the willingness of lenders to completely divest from their regular counterparties. 4.3 Effects of the ONRRP on bargaining power in the repo market In our next exercise, we aim to identify the effects of the ONRRP on prevailing rates in the repo market. By providing repo lenders with a credible outside option, MMFs eligible for the ONRRP should be put in a more advantageous bargaining position vis-à-vis dealers. Of course, funds that are already investing in the ONRRP facility at the counterparty cap do not possess an option to invest a marginal dollar in the ONRRP facility at the expense of dealer-provided repo. Therefore, we again exploit the exogenous increases in counterparty caps, because these increases should bestow additional bargaining power on MMFs that were previously investing the maximum-allowable amount at the ONRRP. The data used for this analysis come from the confidential transaction-level tri-party data collected by the FRBNY discussed in Section 3. Repo activity is aggregated to unique MMF family-dealer-collateral triples. Since there may be many trades on a given day between MMFs and their dealers, we calculate both the weighted average rate between the MMF and dealer for a given collateral type, as well as the volume-weighted 25th and 75th percentile rates. To identify the effect of the potentially increased bargaining power MMFs command in the private repo market, we again appeal to a DD strategy. Here, though, we take advantage of the high frequency data and treat each cap increase as its own separate event, examining the days around each increase. Specifically, we identify fund families in which at least one MMF was bound by the counterparty cap on the day before a cap increase. The first four counterparty cap increases, which occurred on September 27, 2013, December 23, 2013, January 30, 2014, and March 5, 2014, witnessed at least one bound fund family on the day before the change.21,22 Multiple fund families contain MMFs that were constrained by the counterparty cap in all but the final date (March 5, 2014), when only a single fund family faced a binding constraint immediately prior to the cap increase. We then compare changes in rates for previously bound MMFs to previously unbound funds on the day of a counterparty cap increase by estimating the following regression: \begin{align} \text{RepoRate}_{ijct} &= \delta\cdot(\text{Post}_{t}\cdot\text{BoundTreatment}_{it-1})\notag \\ &\quad +\beta_{ijc}\cdot(\text{MMF Family-Dealer-Collateral}_{ijc}) + \gamma_{t}\cdot\text{Day}_{t} + \varepsilon_{ijct}. \end{align} (7) In Equation (7), $$\delta$$ is the DD estimate of the effect of a cap increase on previously bound (treatment group) funds’ weighted average repo rate with a given dealer for a given type of collateral. For simplicity, and because most funds did not face a binding counterparty cap on the day of the cap increases, we consider only a single treatment group that is composed of fund families that were bound the day prior to the increase.23 If MMFs’ bargaining power sufficiently increases upon counterparty cap increases, $$\delta$$ would be expected to be positive, indicating an increase in the rate that MMFs can command when presented with the outside option. Similarly, $$\delta$$ could also be positive if money funds simply shift lower-rate private transactions to the ONRRP facility, though the ability of dealers to pay a rate below the ONRRP indicates a differential in bargaining power that favors dealers. Figure 8 demonstrates parallel trends in the daily average repo rates earned by constrained and unconstrained fund families in the days leading up to the changes in the ONRRP maximum bid amount. There is no apparent divergence in advance of any of the cap changes in our sample. Figure 8 View largeDownload slide Trends in rates on private repo transactions This figure presents treatment effects for the weighted-average rates in the days preceding each counterparty cap change. Thus, a positive value corresponds to days in which treated funds witness average rates rising more (or falling less) than the rate for control funds. A negative value corresponds to days in which treated funds witness average rates rising less (or falling more) than rates for control funds. Figure 8 View largeDownload slide Trends in rates on private repo transactions This figure presents treatment effects for the weighted-average rates in the days preceding each counterparty cap change. Thus, a positive value corresponds to days in which treated funds witness average rates rising more (or falling less) than the rate for control funds. A negative value corresponds to days in which treated funds witness average rates rising less (or falling more) than rates for control funds. The left side of Table 7 reports the estimate of $$\delta$$ from a regression with the weighted average repo rate used as the dependent variable. Panel A reports the results around the first cap change only, with the subsequent panels adding observations around additional cap increases, as indicated. Specification (1) includes only the day before and the day of the counterparty cap increase, and shows a robustly positive effect of an increase in ONRRP participation on the rates that previously bound funds command in private repo transactions. The coefficient of 0.26 in panel A implies that the first cap increase led to an increase in previously bound funds’ average repo rate of 0.26 basis points. Although this increase appears small, the average tri-party repo rate (including both Treasury and agency collateral) was only 3 basis points on the day before the cap increase, according to the publicly available BNY Mellon Tri-Party Repo Indices. The result that the ONRRP appears to increase repo rates by offering an outside option to money funds is consistent with the findings in Han and Nikolaou (2016). Table 7 Regression results: Repo rates A. 1st cap increase Weighted average rate Memo: 25th percentile rate Memo: 75th percentile rate (1) (2) (3) (1) (2) (3) (1) (2) (3) $$\text{Bound}_{t-1}$$ 0.26** 0.23*** 0.23** 0.32*** 0.44** 0.47** 0.26** 0.21 0.20 (0.10) (0.08) (0.09) (0.10) (0.16) (0.20) (0.12) (0.13) (0.15) N 177 533 444 177 533 444 177 533 444 Adj. R$$^2$$ 0.99 0.98 0.98 0.99 0.88 0.86 0.99 0.97 0.97 B. 1st and 2nd cap increase $$\text{Bound}_{t-1}$$ 0.26*** 0.47** 0.53** 0.30*** 0.62* 0.70* 0.25*** 0.51** 0.57** (0.07) (0.19) (0.23) (0.09) (0.31) (0.38) (0.07) (0.22) (0.27) N 345 1,043 869 345 1,043 869 345 1,043 869 Adj. R$$^2$$ 0.98 0.84 0.82 0.98 0.79 0.76 0.98 0.80 0.77 C: 1st, 2nd, and 3rd cap increase $$\text{Bound}_{t-1}$$ 0.13* 0.35*** 0.39*** 0.14* 0.43*** 0.49*** 0.11* 0.35*** 0.40** (0.06) (0.09) (0.11) (0.07) (0.14) (0.17) (0.06) (0.12) (0.15) N 497 1,510 1,258 497 1,510 1,258 497 1,510 1,258 Adj. R$$^2$$ 0.98 0.86 0.84 0.98 0.82 0.79 0.98 0.83 0.80 D: 1st, 2nd, 3rd, and 4th cap increase $$\text{Bound}_{t-1}$$ 0.18* 0.31* 0.36* 0.23* 0.40* 0.46 0.20 0.34 0.39 (0.10) (0.17) (0.20) (0.12) (0.23) (0.27) (0.21) (0.13) (0.25) N 643 1,943 1,618 643 1,943 1,618 643 1,943 1,618 Adj. R$$^2$$ 0.90 0.81 0.78 0.89 0.77 0.74 0.89 0.77 0.73 A. 1st cap increase Weighted average rate Memo: 25th percentile rate Memo: 75th percentile rate (1) (2) (3) (1) (2) (3) (1) (2) (3) $$\text{Bound}_{t-1}$$ 0.26** 0.23*** 0.23** 0.32*** 0.44** 0.47** 0.26** 0.21 0.20 (0.10) (0.08) (0.09) (0.10) (0.16) (0.20) (0.12) (0.13) (0.15) N 177 533 444 177 533 444 177 533 444 Adj. R$$^2$$ 0.99 0.98 0.98 0.99 0.88 0.86 0.99 0.97 0.97 B. 1st and 2nd cap increase $$\text{Bound}_{t-1}$$ 0.26*** 0.47** 0.53** 0.30*** 0.62* 0.70* 0.25*** 0.51** 0.57** (0.07) (0.19) (0.23) (0.09) (0.31) (0.38) (0.07) (0.22) (0.27) N 345 1,043 869 345 1,043 869 345 1,043 869 Adj. R$$^2$$ 0.98 0.84 0.82 0.98 0.79 0.76 0.98 0.80 0.77 C: 1st, 2nd, and 3rd cap increase $$\text{Bound}_{t-1}$$ 0.13* 0.35*** 0.39*** 0.14* 0.43*** 0.49*** 0.11* 0.35*** 0.40** (0.06) (0.09) (0.11) (0.07) (0.14) (0.17) (0.06) (0.12) (0.15) N 497 1,510 1,258 497 1,510 1,258 497 1,510 1,258 Adj. R$$^2$$ 0.98 0.86 0.84 0.98 0.82 0.79 0.98 0.83 0.80 D: 1st, 2nd, 3rd, and 4th cap increase $$\text{Bound}_{t-1}$$ 0.18* 0.31* 0.36* 0.23* 0.40* 0.46 0.20 0.34 0.39 (0.10) (0.17) (0.20) (0.12) (0.23) (0.27) (0.21) (0.13) (0.25) N 643 1,943 1,618 643 1,943 1,618 643 1,943 1,618 Adj. R$$^2$$ 0.90 0.81 0.78 0.89 0.77 0.74 0.89 0.77 0.73 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF repo rates. $$\text{Bound}_{t-1}$$ takes a value of one for fund complexes that had at least one fund facing a binding cap on the day before an increase. Panel A includes data from the first cap increase—September 27, 2013—only. Panels B through D add data from days around the three subsequent cap increases: December 23, 2013, January 30, 2014, and March 5, 2014. Column 1 includes a 2-day sample window spanning the day before and day of a cap increase. Column 2 uses a 6-day sample window including the 5 days prior to the cap increase. Column 3 uses the same sample as Column 2, but drops the day before the cap change (the announcement day). Results are reported for fund complexes’ volume-weighted average rate and the volume-weighted 25th and 75th percentile rates, as indicated. All specifications include fund complex and day fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 7 Regression results: Repo rates A. 1st cap increase Weighted average rate Memo: 25th percentile rate Memo: 75th percentile rate (1) (2) (3) (1) (2) (3) (1) (2) (3) $$\text{Bound}_{t-1}$$ 0.26** 0.23*** 0.23** 0.32*** 0.44** 0.47** 0.26** 0.21 0.20 (0.10) (0.08) (0.09) (0.10) (0.16) (0.20) (0.12) (0.13) (0.15) N 177 533 444 177 533 444 177 533 444 Adj. R$$^2$$ 0.99 0.98 0.98 0.99 0.88 0.86 0.99 0.97 0.97 B. 1st and 2nd cap increase $$\text{Bound}_{t-1}$$ 0.26*** 0.47** 0.53** 0.30*** 0.62* 0.70* 0.25*** 0.51** 0.57** (0.07) (0.19) (0.23) (0.09) (0.31) (0.38) (0.07) (0.22) (0.27) N 345 1,043 869 345 1,043 869 345 1,043 869 Adj. R$$^2$$ 0.98 0.84 0.82 0.98 0.79 0.76 0.98 0.80 0.77 C: 1st, 2nd, and 3rd cap increase $$\text{Bound}_{t-1}$$ 0.13* 0.35*** 0.39*** 0.14* 0.43*** 0.49*** 0.11* 0.35*** 0.40** (0.06) (0.09) (0.11) (0.07) (0.14) (0.17) (0.06) (0.12) (0.15) N 497 1,510 1,258 497 1,510 1,258 497 1,510 1,258 Adj. R$$^2$$ 0.98 0.86 0.84 0.98 0.82 0.79 0.98 0.83 0.80 D: 1st, 2nd, 3rd, and 4th cap increase $$\text{Bound}_{t-1}$$ 0.18* 0.31* 0.36* 0.23* 0.40* 0.46 0.20 0.34 0.39 (0.10) (0.17) (0.20) (0.12) (0.23) (0.27) (0.21) (0.13) (0.25) N 643 1,943 1,618 643 1,943 1,618 643 1,943 1,618 Adj. R$$^2$$ 0.90 0.81 0.78 0.89 0.77 0.74 0.89 0.77 0.73 A. 1st cap increase Weighted average rate Memo: 25th percentile rate Memo: 75th percentile rate (1) (2) (3) (1) (2) (3) (1) (2) (3) $$\text{Bound}_{t-1}$$ 0.26** 0.23*** 0.23** 0.32*** 0.44** 0.47** 0.26** 0.21 0.20 (0.10) (0.08) (0.09) (0.10) (0.16) (0.20) (0.12) (0.13) (0.15) N 177 533 444 177 533 444 177 533 444 Adj. R$$^2$$ 0.99 0.98 0.98 0.99 0.88 0.86 0.99 0.97 0.97 B. 1st and 2nd cap increase $$\text{Bound}_{t-1}$$ 0.26*** 0.47** 0.53** 0.30*** 0.62* 0.70* 0.25*** 0.51** 0.57** (0.07) (0.19) (0.23) (0.09) (0.31) (0.38) (0.07) (0.22) (0.27) N 345 1,043 869 345 1,043 869 345 1,043 869 Adj. R$$^2$$ 0.98 0.84 0.82 0.98 0.79 0.76 0.98 0.80 0.77 C: 1st, 2nd, and 3rd cap increase $$\text{Bound}_{t-1}$$ 0.13* 0.35*** 0.39*** 0.14* 0.43*** 0.49*** 0.11* 0.35*** 0.40** (0.06) (0.09) (0.11) (0.07) (0.14) (0.17) (0.06) (0.12) (0.15) N 497 1,510 1,258 497 1,510 1,258 497 1,510 1,258 Adj. R$$^2$$ 0.98 0.86 0.84 0.98 0.82 0.79 0.98 0.83 0.80 D: 1st, 2nd, 3rd, and 4th cap increase $$\text{Bound}_{t-1}$$ 0.18* 0.31* 0.36* 0.23* 0.40* 0.46 0.20 0.34 0.39 (0.10) (0.17) (0.20) (0.12) (0.23) (0.27) (0.21) (0.13) (0.25) N 643 1,943 1,618 643 1,943 1,618 643 1,943 1,618 Adj. R$$^2$$ 0.90 0.81 0.78 0.89 0.77 0.74 0.89 0.77 0.73 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF repo rates. $$\text{Bound}_{t-1}$$ takes a value of one for fund complexes that had at least one fund facing a binding cap on the day before an increase. Panel A includes data from the first cap increase—September 27, 2013—only. Panels B through D add data from days around the three subsequent cap increases: December 23, 2013, January 30, 2014, and March 5, 2014. Column 1 includes a 2-day sample window spanning the day before and day of a cap increase. Column 2 uses a 6-day sample window including the 5 days prior to the cap increase. Column 3 uses the same sample as Column 2, but drops the day before the cap change (the announcement day). Results are reported for fund complexes’ volume-weighted average rate and the volume-weighted 25th and 75th percentile rates, as indicated. All specifications include fund complex and day fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. The second column of Table 7—labeled (2)—extends the pretreatment period to include the five trading days prior to the change in the counterparty cap, for a total of six days, and shows a very similar pattern upon an increase in the maximum bid. To avoid any potentially confounding effects of the announcement of the cap increase, which occurred 1 business day before each cap increase, specification (3) excludes the announcement day from the 6-day window used in Column 2. Although the announcements were made during the trading day prior to the cap increase, they generally occurred after most private repo activity had taken place. Thus, it is unsurprising that the five day results excluding the announcement day are consistent with the six day sample presented in Column 2. Panels B through D show that these results persist when the sample is expanded to include the days around other cap changes. Including the final cap increase in the sample (panel D) produces less precise estimates, but this is likely attributable to the fact that most complexes did not face a binding cap during this episode, and thus there are very few trades associated with treated funds. In the memorandum items in the middle and right side of Table 7, we report the effect of the cap increase on the weighted 25th and 75th percentile of fund families’ rate distribution. As expected, the cap increase has a somewhat larger effect on lower-rate trades. In total, the results reported in Table 7 are consistent with the hypothesis that the outside option presented by the ONRRP increases repo lenders’ bargaining position vis-à-vis borrowers.24 Therefore, as a result of the ONRRP facility, dealers not only witnessed a reduction in the supply of funds as shown in section 4.1, but also faced a contemporaneous increase in their funding costs. Consistent with our interpretation, this combination characterizes a reduction in the supply of private repo investment. 4.4 Effects of the ONRRP on private repo borrowers Having shown that repo lenders substitute away from private repo when they invest more in the ONRRP, we now turn to the question of how dealers respond to this adverse funding shock and what additional implications this may have for financial stability. As dealers face more competition for repo funding from the ONRRP, financial stability could be affected in a number of ways. In response to the withdrawal of lending by ONRRP-eligible MMFs, dealers may decrease total repo volume, possibly reducing financial fragility by limiting the externalities associated with extensive private money creation in the repo market. Conversely, dealers could maintain their level of repo funding in the wake of the negative ONRRP supply shock by shifting their composition of counterparties and/or collateral. Such a response could potentially increase financial fragility if dealers increase their borrowing from smaller, less stable lenders. A shift to riskier collateral repo would also have financial stability implications for two reasons. First, repo backed by riskier collateral types appears to be more susceptible to runs during a financial crisis. For instance, Krishnamurthy, Nagel, and Orlov (2014) and Copeland, Martin, and Walker (2014) show that the use of riskier collateral in the tri-party repo market dropped off significantly during the recent financial crisis. In contrast, Treasury repo was more stable. Therefore, if dealers increase their reliance on repo backed by other collateral types, it may make a run in the repo market both more likely and more pronounced.25 Second, to enter into repo transactions backed by riskier collateral types, dealers must either own or borrow those assets. Such a shift toward riskier types of repo funding could therefore lead in turn to a shift toward a riskier overall asset portfolio among dealers, thereby increasing financial fragility (Parlatore 2016). The final portion of our analysis thus proceeds with an examination of the dealer borrowing response to the ONRRP-induced withdrawal of privately supplied repo funding in two steps. First, we consider the effects on the mix and quality of collateral that backs dealer repo borrowing using our dealer/collateral position-level repo data set. Additionally, we use the FR-2004C schedule of all dealer repo positions to determine if dealers begin using riskier types of repo for net financing. Second, we examine the effects of the ONRRP on dealers’ composition of lenders using our dealer/lender position-level repo data set. As before, we identify off of exogenous cap increases in the period after the introduction of the ONRRP facility using a DD framework. However, we no longer have a binary treatment variable available for this analysis. Instead, we generate a continuous treatment variable that measures dealers’ ex ante exposure to constrained (treatment) MMFs. In particular, we define the constrained share as of the September 2013 introduction of the ONRRP for dealer $$j$$ as \begin{equation} \text{Constrained}_{j} = \frac{\text{Repo Volume with Constrained MMFs}_{j}}{\text{Total Repo Volume}_{j}}. \end{equation} (8) $$\text{Constrained}_{j}$$ has a mean of 0.087 (i.e., 8.7% of total repo borrowing comes from constrained MMFs) with a standard deviation of 0.12 and a range of 0 to 0.4. We determine which dealers trade with constrained MMFs using the N-MFP data, as before. This information is then merged into both of our position-level data sets on tri-party repo positions, as well as the FR-2004C data. Using the tri-party data sets allows us to capture the entirety of dealer borrowing in the tri-party market rather than just the trades with MMFs that we observe in the N-MFP. Except where noted, we again sample dealer repo borrowing on quarter-end dates between December 2012 and June 2015. 4.4.1 Repo collateral quality To determine how the amount and composition of dealer repo is affected by MMFs’ investment with the Fed, we consider four measures of dealer repo borrowing: the log of repo volume, both in aggregate and collateralized with either Treasury, agency, or other securities. Specifically, we estimate the following regression: \begin{equation} \begin{split} \text{y}_{jt} &= \delta\cdot(\text{Post}_{t}\cdot\text{Constrained}_{j}) + \beta_{j}\cdot\text{Dealer}_{j} + \gamma_{t}\cdot\text{Quarter}_{t} + \varepsilon_{jt}. \end{split} \end{equation} (9) In this specification, $${Post}_{t}$$ takes a value of one for all quarter-ends after September 2013, and $$\delta$$ is our primary coefficient of interest, measuring the sensitivity of dealer repo activity to their ex ante exposure to MMFs that increase their use of the ONRRP during the post-treatment period as a result of the exogenous cap increases. The results of this analysis are shown in panel A of Table 8. In the first column, we see that total repo borrowing is essentially unchanged after the ONRRP cap increases. Therefore, we do not find evidence that the ONRRP reduces dealers’ total repo borrowing. Examining the collateral composition of dealer repo borrowing, we see in the second column that Treasury repo borrowing is essentially unchanged. Evidently, dealers replace the Treasury-backed repo withdrawn by ONRRP-eligible MMFs by trading with other counterparties. Agency repo—reported in the third column—appears to fall, though this result does not achieve statistical significance. Evidently, dealers replace the Treasury-backed repo withdrawn by ONRRP-eligible MMFs by trading with other counterparties. Finally, other repo volume is significantly higher, as seen in the final column. For context, the $$\delta$$ value of 0.71 for other repo implies that for a 1 percentage point increase in repo borrowing from initially constrained ONRRP-eligible MMFs ($${\$}$$80 million for the median dealer), riskier tri-party repo borrowing against nongovernment collateral increases $${\$}$$14 million for the median dealer. Table 8 Regression results: Dealers’ collateral response to the adverse funding shock A. Dealer repo volume by collateral Total repo volume Treasury repo volume Agency repo volume Other repo volume $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.03 0.15 –0.40 0.71* (0.29) (0.43) (0.48) (0.40) N 654 499 558 429 Adj. R$$^2$$ 0.99 0.98 0.99 0.98 A. Dealer repo volume by collateral Total repo volume Treasury repo volume Agency repo volume Other repo volume $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.03 0.15 –0.40 0.71* (0.29) (0.43) (0.48) (0.40) N 654 499 558 429 Adj. R$$^2$$ 0.99 0.98 0.99 0.98 B. Haircuts 50th percentile 75th percentile 90th percentile 95th percentile $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.002 0.02 0.04** 0.03** (0.004) (0.01) (0.02) (0.01) N 654 654 654 654 Adj. R$$^2$$ 0.98 0.95 0.93 0.91 B. Haircuts 50th percentile 75th percentile 90th percentile 95th percentile $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.002 0.02 0.04** 0.03** (0.004) (0.01) (0.02) (0.01) N 654 654 654 654 Adj. R$$^2$$ 0.98 0.95 0.93 0.91 C. Net other repo financing All nongovernment collateral Corporate debt Other miscellaneous collateral $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.63*** 1.59*** 0.31** (0.22) (0.38) (0.15) N 1,628 1,508 1,405 Adj. R$$^2$$ 0.63 0.60 0.69 C. Net other repo financing All nongovernment collateral Corporate debt Other miscellaneous collateral $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.63*** 1.59*** 0.31** (0.22) (0.38) (0.15) N 1,628 1,508 1,405 Adj. R$$^2$$ 0.63 0.60 0.69 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on dealer repo activity. Post is equal to 1 for all time periods beginning in Q4 2013. Constrained is equal to a dealer’s dollar volume of repo borrowed from money funds that are constrained by the ONRRP cap as of Q3 2013 quarter-end, divided by the dealer’s total repo volume. In panel A, the dependent variables are measured in log volumes. Dependent variables in panel B are dealer-specific volume-weighted percentile haircuts. The dependent variables in panel C are the shares of repo of different collateral types, as indicated, that are used for net financing. Data used in panel C are measured at the weekly frequency. All specifications in panels A through C include dealer and time fixed effects. Standard errors are clustered at the dealer level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 8 Regression results: Dealers’ collateral response to the adverse funding shock A. Dealer repo volume by collateral Total repo volume Treasury repo volume Agency repo volume Other repo volume $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.03 0.15 –0.40 0.71* (0.29) (0.43) (0.48) (0.40) N 654 499 558 429 Adj. R$$^2$$ 0.99 0.98 0.99 0.98 A. Dealer repo volume by collateral Total repo volume Treasury repo volume Agency repo volume Other repo volume $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.03 0.15 –0.40 0.71* (0.29) (0.43) (0.48) (0.40) N 654 499 558 429 Adj. R$$^2$$ 0.99 0.98 0.99 0.98 B. Haircuts 50th percentile 75th percentile 90th percentile 95th percentile $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.002 0.02 0.04** 0.03** (0.004) (0.01) (0.02) (0.01) N 654 654 654 654 Adj. R$$^2$$ 0.98 0.95 0.93 0.91 B. Haircuts 50th percentile 75th percentile 90th percentile 95th percentile $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.002 0.02 0.04** 0.03** (0.004) (0.01) (0.02) (0.01) N 654 654 654 654 Adj. R$$^2$$ 0.98 0.95 0.93 0.91 C. Net other repo financing All nongovernment collateral Corporate debt Other miscellaneous collateral $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.63*** 1.59*** 0.31** (0.22) (0.38) (0.15) N 1,628 1,508 1,405 Adj. R$$^2$$ 0.63 0.60 0.69 C. Net other repo financing All nongovernment collateral Corporate debt Other miscellaneous collateral $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.63*** 1.59*** 0.31** (0.22) (0.38) (0.15) N 1,628 1,508 1,405 Adj. R$$^2$$ 0.63 0.60 0.69 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on dealer repo activity. Post is equal to 1 for all time periods beginning in Q4 2013. Constrained is equal to a dealer’s dollar volume of repo borrowed from money funds that are constrained by the ONRRP cap as of Q3 2013 quarter-end, divided by the dealer’s total repo volume. In panel A, the dependent variables are measured in log volumes. Dependent variables in panel B are dealer-specific volume-weighted percentile haircuts. The dependent variables in panel C are the shares of repo of different collateral types, as indicated, that are used for net financing. Data used in panel C are measured at the weekly frequency. All specifications in panels A through C include dealer and time fixed effects. Standard errors are clustered at the dealer level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. This increase in risky repo borrowing can also be demonstrated by comparing changes in the value of collateral that dealers must deliver to secure a repo loan of a given size. The difference between the collateral value and the loan amount divided by the loan amount is known as the haircut, with riskier collateral subject to higher haircuts. For example, while the haircut on Treasury repo is around 2%, the haircut on equity repo is around 8%. To assess whether dealers’ average haircut increases as they lose funding to the ONRRP, we estimate an alternative version of Equation (9) in which $${y}_{jt}$$ is dealer j’s volume-weighted haircut across all repo positions for selected percentiles listed in panel B of Table 8. There is no change in the average (data not shown) or median haircut. However, there is a significant increase in the haircut at the 90th and 95th percentiles of the distribution. The point estimate of 0.04 is economically significant, as it implies that just a 1 percentage point increase in repo volume borrowed from initially constrained ONRRP-eligible MMFs translates to a 1% increase in the average 90th percentile haircut of 4.7%. As dealers shift to repo backed by nongovernment collateral, their haircut requirements are also increasing, showing that they have taken on more risk. The move to a riskier repo position by those dealers borrowing from MMFs that invest in the ONRRP would be most concerning if this repo is used as a net source of funding for securities with longer maturities. If risky repo is used to fund dealers’ net securities position, a stress event that limited repo financing backed by risky collateral could lead to fire sales of those securities. If dealers instead simply acquired the riskier collateral through reverse repos as part of their matched book activity, the net exposure to the dealer would be very limited, as the transactions would essentially offset. In the event that repo funding secured by risky collateral dried up, the dealer could simply decline to roll over the matching reverse repo transactions. Using the FR-2004C data that contain primary dealers’ tri-party and bilateral repo and reverse repo activity by collateral type, we are able to examine whether net financing with riskier repo increases in response to the expansion of the ONRRP. Specifically, we estimate Equation (9) on weekly FR-2004C data in which $${y}_{jt}$$ is defined as follows for dealer j’s non-Treasury or agency collateral:26 \begin{equation} y_{jt} = \frac{\text{other repo}_{jt}-\text{other reverse repo}_{jt}}{\text{other repo}_{jt}}. \end{equation} (10) Therefore, $${y}_{jt}$$ measures the share of repo borrowing backed by risky collateral that is used for net financing as opposed to matched book activity. Given the relative flightiness of other repo, higher values of $${y}_{jt}$$ represent an increase in fragility. The first column of panel C in Table 8 shows that dealers that lost more repo funding to the ONRRP do indeed boost their net financing using collateral that is not government-backed. The point estimate implies that a 1 percentage point increase in repo volume borrowed from initially constrained MMFs resulted in an additional $${\$}$$75 million of net funding with repo backed by risky collateral, on average. This represents nearly 1% of average net financing using other repo. The second and third columns of panel C show that this result also holds for different subtypes of nongovernment collateral. However, the marginal effect is larger for the relatively safe collateral like corporate bonds (average haircut of 6.6%) than for the riskiest collateral like private CMOs (average haircut of 9.8%), which are included in the miscellaneous category. 4.4.2 Repo counterparties To better understand the increase in dealers’ other repo borrowing, we now turn to an analysis of the change in lender composition. We divide repo lenders into 3 categories: ONRRP-eligible MMFs, ONRRP-ineligible MMFs, and non-MMFs.27 We then estimate alternate versions of Equation (9) in which $${y}_{jt}$$ is the share of total repo volume borrowed from each of these groups of lenders. Panel A of Table 9 reports the results from this exercise. Table 9 Regression results: Dealers’ counterparty response to the adverse funding shock A. Counterparty composition of dealer repo Eligible MMF share Ineligible MMF share Non-MMF share Memo: Weighted $$\sigma$$(AUM) $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.30*** 0.08** 0.22** 0.10* (0.08) (0.03) (0.09) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.96 0.80 0.99 0.92 A. Counterparty composition of dealer repo Eligible MMF share Ineligible MMF share Non-MMF share Memo: Weighted $$\sigma$$(AUM) $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.30*** 0.08** 0.22** 0.10* (0.08) (0.03) (0.09) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.96 0.80 0.99 0.92 B. Non-MMF counterparty composition of dealer repo Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.16*** 0.02 –0.10 0.05 (0.04) (0.01) (0.11) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.77 0.98 0.83 0.96 B. Non-MMF counterparty composition of dealer repo Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.16*** 0.02 –0.10 0.05 (0.04) (0.01) (0.11) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.77 0.98 0.83 0.96 C. Non-MMF counterparty composition of dealer repo at the relationship level Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.05** 0.02 0.11 0.05 (0.02) (0.03) (0.10) (0.04) N 1,700 2,082 681 968 Adj. R$$^2$$ 0.74 0.86 0.95 0.21 C. Non-MMF counterparty composition of dealer repo at the relationship level Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.05** 0.02 0.11 0.05 (0.02) (0.03) (0.10) (0.04) N 1,700 2,082 681 968 Adj. R$$^2$$ 0.74 0.86 0.95 0.21 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on dealer counterparties in the repo market. Post is equal to 1 for all time periods beginning in Q4 2013. Constrained is equal to a dealer’s dollar volume of repo borrowed from money funds that are constrained by the ONRRP cap as of Q3 2013 quarter-end, divided by the dealer’s total repo volume. In each panel, the dependent variables are the share of total repo volume borrowed from a particular counterparty type. All specifications in panels A through C include dealer fixed effects. Panels A and B include time fixed effects, and panel C contains lender-time fixed effects. Standard errors are clustered at the dealer level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 9 Regression results: Dealers’ counterparty response to the adverse funding shock A. Counterparty composition of dealer repo Eligible MMF share Ineligible MMF share Non-MMF share Memo: Weighted $$\sigma$$(AUM) $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.30*** 0.08** 0.22** 0.10* (0.08) (0.03) (0.09) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.96 0.80 0.99 0.92 A. Counterparty composition of dealer repo Eligible MMF share Ineligible MMF share Non-MMF share Memo: Weighted $$\sigma$$(AUM) $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.30*** 0.08** 0.22** 0.10* (0.08) (0.03) (0.09) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.96 0.80 0.99 0.92 B. Non-MMF counterparty composition of dealer repo Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.16*** 0.02 –0.10 0.05 (0.04) (0.01) (0.11) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.77 0.98 0.83 0.96 B. Non-MMF counterparty composition of dealer repo Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.16*** 0.02 –0.10 0.05 (0.04) (0.01) (0.11) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.77 0.98 0.83 0.96 C. Non-MMF counterparty composition of dealer repo at the relationship level Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.05** 0.02 0.11 0.05 (0.02) (0.03) (0.10) (0.04) N 1,700 2,082 681 968 Adj. R$$^2$$ 0.74 0.86 0.95 0.21 C. Non-MMF counterparty composition of dealer repo at the relationship level Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.05** 0.02 0.11 0.05 (0.02) (0.03) (0.10) (0.04) N 1,700 2,082 681 968 Adj. R$$^2$$ 0.74 0.86 0.95 0.21 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on dealer counterparties in the repo market. Post is equal to 1 for all time periods beginning in Q4 2013. Constrained is equal to a dealer’s dollar volume of repo borrowed from money funds that are constrained by the ONRRP cap as of Q3 2013 quarter-end, divided by the dealer’s total repo volume. In each panel, the dependent variables are the share of total repo volume borrowed from a particular counterparty type. All specifications in panels A through C include dealer fixed effects. Panels A and B include time fixed effects, and panel C contains lender-time fixed effects. Standard errors are clustered at the dealer level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. As expected, the first column reveals a clear decline in repo trading with ONRRP-eligible MMFs, as these are precisely the funds that substitute to the ONRRP. There is a small increase in borrowing from MMFs that are ineligible for the ONRRP, as shown in the second column.28 In Column 3, we see that much of the lost MMF funding is recouped by borrowing from non-MMF lenders (recall that total repo borrowing is unchanged). This result is intuitive in the context of the results found in panel A of Table 8, because non-MMFs are the primary lenders in the market for other collateral repo. The point estimates imply that a 1 percentage point increase in repo volume borrowed from initially constrained MMFs translates to a 0.3 percentage point decrease in borrowing from eligible MMFs, a 0.08 percentage point increase in borrowing from ineligible MMFs, and a 0.22 percentage point increase from non-MMFs. ONRRP-ineligible MMFs have fewer assets under management than eligible funds on average such that the dealers’ move from eligible to ineligible MMFs implies a shift to smaller lenders. However, dealers may choose to increase borrowing from only the most stable ineligible MMFs, making their relative size less of a concern. To test this, we measure the percentage standard deviation of detrended AUM for each MMF from 2011-2014 and compute a volume-weighted average of this standard deviation for each dealer. We then use this volume-weighted standard deviation of counterparty AUM as a dependent variable in Equation (9), and report the results in the memorandum item in the last column of panel A of Table 9. We find that after ONRRP cap increases, dealers trade with more asset-volatile MMFs, suggesting a move to a less stable source of funding for dealers. Since dealers are predominately substituting to non-MMFs, we turn now to an analysis of the increase in funding from non-MMFs. Using our data set on tri-party repo positions between dealers and lenders, we observe total dealer repo positions with individual lenders, which are then aggregated by lender type. Specifically, we focus on total borrowing from asset managers affiliated with ONRRP-eligible MMFs, asset managers affiliated only with ONRRP-ineligible MMFs, primary dealer affiliates, and custodian banks. Asset managers include mutual funds, hedge funds, and other asset managers. Custodian banks include nonprimary dealer custodians, namely Bank of New York-Mellon and State Street. In total, repo borrowing from these lenders accounts for over 90% of total tri-party repo volume. We then estimate alternate versions of Equation (9) in which $${y}_{jt}$$ is the share of repo volume borrowed from each group of non-MMF lenders. The results of this analysis are reported in panel B of Table 9. When dealers shift to non-MMF counterparties, they transact more with asset managers affiliated with at least one eligible ONRRP counterparty (shown in the first column). There is not a significant change in borrowing from asset managers without any eligible counterparties, primary dealer affiliates, or custodian banks (Columns 2 through 4 of panel B). Evidently, dealers endeavor to maintain their relationships with the largest (ONRRP-eligible) asset manager complexes by transacting with their non-MMF affiliates. Because non-MMFs lend more against other collateral repo, however, dealers must accommodate this transition by expanding their borrowing against riskier collateral, consistent with the findings discussed in the previous subsection and detailed in Table 8. If constrained dealers have relationships with asset managers that are increasing their other collateral repo lending at the expense of collateral types held by their MMFs for reasons unrelated to the ONRRP, this could invalidate the interpretation of our results. To rule out a lender-driven explanation of our results, we construct a repo credit registry that contains total volume between each dealer-lender pair. We can then estimate a regression that includes lender-time fixed effects as follows: \begin{align} y_{ijt} &= \delta\cdot(\text{Post}_{t}\cdot\text{Constrained}_{j}) + \phi_{it}\cdot(\text{Lender}_{i}\cdot\text{Quarter}_{t}) \notag\\ &\quad + \beta_{ij}\cdot\text{Lender-Dealer}_{ij} + \varepsilon_{ijt}. \end{align} (11) In Equation 11, $$y_{ijt}$$ represents the share of the total repo volume of dealer $$j$$ borrowed from lender $$i$$. The key coefficient $$\delta$$ then measures the borrowing response of dealers that differ in their constrained share, holding lender and quarter constant. The results are shown in panel C of Table 9. Consistent with the results in panel B, we see that constrained dealers shift to asset managers affiliated with at least one eligible ONRRP counterparty (shown in the first column). There is no significant change in repo borrowing from other lender types. These results validate the finding that dealers seek to maintain their relationships with the largest asset managers by switching to affiliated non-MMFs when MMFs withdraw repo funding due to the ONRRP. These non-MMF managers, however, are more likely to lend against a mix of non-Treasury and agency securities that mirrors their existing asset portfolios. 4.4.3 Discussion of dealer responses Overall, these results indicate that the ONRRP poses some risks to financial stability. Dealers that are more exposed to initially constrained MMFs that subsequently withdraw private repo funding in favor of the ONRRP do not reduce their total repo volume. Thus, one potential benefit of the ONRRP—that it could reduce the size of the run-prone repo market (Carlson et al. 2016)—has evidently not materialized. Rather, dealers make up for any lending lost to the ONRRP by switching to a different collateral and lender mix. Specifically, dealers increase their repo borrowing from non-MMFs, leading to an increase in borrowing against riskier collateral and a greater reliance on riskier repo for net financing. To a lesser degree, these dealers also shift to more asset-volatile, ONRRP-ineligible MMF counterparties. While the economic magnitudes of the effects we observe in response to the relatively small funding shock are not large enough to pose an immediate threat to financial stability, the results demonstrate how dealers respond to funding supply shocks that are introduced by monetary policy implementation. In the event of severe financial stress, a flight of repo lenders to the ONRRP could trigger a sharp decline in private repo supply. Our results show that dealers are more susceptible to such a shock for two reasons. First, dealers shift to repo backed by less-safe collateral types, which is far less reliable during episodes of financial stress (Krishnamurthy, Nagel, and Orlov 2014; Copeland, Martin, and Walker 2014). Second, a move by dealers to substitute ONRRP-eligible MMFs with smaller and more volatile ineligible MMFs could also have implications for financial stability. Although ineligible MMFs cannot run to the ONRRP in times of financial stress, their investors may still withdraw cash, forcing the fund to exit the private repo market and starving dealers of this source of funding. Both of these dealer responses therefore suggest that monetary policy implementation can increase the likelihood and severity of financial disruption in the repo market. In response to a flight to the ONRRP, the Fed could potentially limit the ONRRP by imposing stricter caps or lowering the ONRRP rate. However, this precisely captures the tradeoff between effective monetary policy implementation and financial stability. To improve financial stability by restricting the ONRRP, the Fed would need to relinquish its control of short-term interest rates. Furthermore, there is an additional tradeoff between the stability of dealers and the stability of MMFs. By offering reliable access to a safe asset, the ONRRP improves the safety of the MMF sector. If MMFs did not have the backstop of the ONRRP, their investors may run, leading to a comparable decline in private repo funding. Given these tradeoffs, the Fed may be unwilling to curtail the ONRRP in times of stress. Although the ONRRP effectively controls rates and provides an important backstop for MMFs during times of financial strain, it could also increase the likelihood or severity of financial instability emanating from dealers, as we have shown. Efforts to stem this instability by limiting the ONRRP could destabilize the MMF sector and undermine the Fed’s ability to manage the policy rate. 5. Conclusion In this paper, we conduct an analysis of the Fed’s regular intervention in the repo market through the ONRRP facility, the Fed’s newest monetary policy tool. Specifically, we exploit exogenous changes in an MMF’s ability to invest in the Fed’s ONRRP facility to identify substitution away from private repo transactions and into repo investment with the Fed. Further analysis of the pattern of MMF substitution shows that, rather than severing trades with certain borrowers entirely, money funds withdraw from their dealer counterparties in a roughly even fashion, albeit with somewhat more withdrawal from their largest borrowers. This pattern of substitution likely reflects MMFs’ desire to preserve existing lending relationships, highlighting the importance of these relationships in the repo market. Additionally, we use confidential data on trades in the tri-party repo market to show that the ONRRP facility bestows additional bargaining power on repo lenders with the option of investing with the Fed. When MMFs are able to invest more in the ONRRP, rates on their private repo transactions increase. Thus, we demonstrate that the presence of the Fed as a borrower in the repo market not only saps repo funding from the market, but also leads to higher dealer funding costs. This dynamic can potentially increase the likelihood of future runs in the repo market if MMFs move away from private repo in favor of Fed repo in the event of financial turmoil, with repo borrowers forced to pay increasingly higher rates. Therefore, it is important to understand how repo borrowers respond to an adverse funding shock. We find that, although dealers lose funding from ONRRP-eligible funds that invest more in the ONRRP, they are able to fill that void by reallocating their repo portfolio and thus the ONRRP does not lead to a decrease in total dealer repo volume. Rather, dealers substitute away from Treasury and agency collateral repo toward repo backed by riskier assets, and increase their reliance on risky repo as a source of net financing. This new funding mix is evidently the result of dealers’ shifting borrowing away from ONRRP-eligible MMFs and toward non-MMF asset managers and ONRRP-ineligible MMFs, which have less stable assets and repo investment. Overall, this study demonstrates how the implementation of monetary policy—particularly when it relies on a large presence in financial markets—can disrupt private activity in funding markets. Our results illuminate an important tradeoff between effective monetary policy implementation and financial stability. Many central banks have enlarged their footprint in financial markets as a consequence of the drastic expansion of the scale and scope of monetary intervention in financial markets since the recent crisis. As we have begun to show in this paper, the financial repercussions engendered by this intervention can have potentially far-reaching implications for private trading activity, collateral assets, and financial stability. We are grateful for helpful comments from Stijn Van Nieuwerburgh (the editor); two anonymous referees; Sriya Anbil, Jim Clouse, Jane Ihrig, Elizabeth Klee, Marco Macchiavelli, Bernd Schlusche, and Rebecca Zarutskie; members of the Money Markets Workgroup for the Federal Reserve System’s Long-Run Monetary Policy Implementation Framework; conference participants at the Australasian Finance and Banking Conference, the World Finance Conference, the IBEFA Summer Meeting, and the FMA European Conference; and seminar participants at the Federal Reserve Board, the Federal Reserve Bank of New York, and the International Monetary Fund. We thank Michelle Bongard for excellent research assistance. The views expressed in this paper are solely the responsibility of the authors and should not be interpreted as reflecting the views of the Board of Governors of the Federal Reserve System or of anyone else associated with the Federal Reserve System. Footnotes 1 The tri-party repo market is one segment of the total repo market. For more information, see Section 1.2. 2 See www.newyorkfed.org/markets/rrp_eligibility_criteria.html for a description of the current eligibility requirements for MMFs, GSEs, and banks. 3 See Frost et al. (2015) for a complete discussion of the design of the ONRRP. 4 Subsequently, other counterparties became eligible. See www.newyorkfed.org/markets/expanded_counterparties.html for a full list of the current eligible counterparties. 5 See www.federalreserve.gov/newsevents/press/monetary/20140917c.htm. 6 See “Minutes of the Federal Open Market Committee November 1-2, 2016,” https://www.federalreserve.gov/monetarypolicy/files/fomcminutes20161102.pdf. 7 The ONRRP also had an aggregate cap of $\$$300 billion between September 22, 2014 and December 16, 2015. Before and after that period, there was no aggregate cap. However, the cap was reached only once, on September 30, 2014. 8 In the bilateral market, cash providers and cash borrowers trade directly. Bilateral repo typically consists of interdealer trading or dealers lending to hedge funds and others. Conversely, the General Collateral Finance (GCF) market is a blind-brokered interdealer market that is a subset of the broader tri-party market. For a complete description of the U.S. repo market, see Copeland et al. (2012). 9 Our treatment group is defined as those funds that were constrained by the $\$$1 billion cap on the first quarter-end after the start of the ONRRP (September 30, 2013). $\$$1 billion was technically the second cap, as the cap was $\$$500 million for the first 4 days of the ONRRP. However, because we are using quarter-end data for most of our analysis, we use the cap at the first quarter-end ( $\$$1 billion) to sort funds into treatment and control groups. 10 In results presented in Section 4, we show that treatment funds did in fact significantly increase their ONRRP investment after an increase in the maximum bid amount. 11 Repo and reverse repo are two sides the same transaction; cash borrowers are said to engage in repo transactions, while cash lenders commit to reverse repo transactions. Unlike most institutions, dealers engage in both types of transactions given their role as intermediaries between different market segments. 12 Primary dealers represent a subset of the dealers that we observe in the tri-party repo market. However, most of the largest dealers are primary dealers and a majority of the dealers that trade in the tri-party market are primary dealers. 13 We exclude the relatively few eligible funds classified as “tax-exempt.” 14 See Section 4.2 for definitions of weighted CDS spreads and HHI. 15 Countries differ on the time frame over which the leverage ratio is calculated. In the United States, the leverage ratio is based on a daily average across a quarter, while in the eurozone, it is based on the value on the quarter-end day only. Other foreign regions have intermediate calculation time frames. The treatment and control funds also have no difference in the percentage of their lending specifically to eurozone dealers (data not shown). 16 We note, however, that size of an MMF is not an absolute predictor of whether a fund found the counterparty cap binding. In fact, 8 of the largest 20 MMFs in our sample (including the largest fund) did not bid at the $\$$1 billion counterparty cap on September 30, 2013. Our results are nearly identical if we simply control for AUM by including the lagged value for all funds in all time periods. 17 The total investment in the ONRRP could be constrained by the designated aggregate and counterparty caps, which could be changed in the future. 18 This placebo treatment date is chosen to correspond to the first post-treatment period of our baseline results (December 2013). However, the results of this exercise are not sensitive to the choice of hypothetical treatment date. 19 However, as noted in Section 3, treated and control funds have similar relationships with foreign dealers. 20 CDS data are obtained from Markit. 21 In each case, the announcement of the increase in the maximum counterparty bid amount was made on the business day prior to the change. 22 Because the cap was increased monotonically, trades by previously bound funds generally declined over time as a share of total trades. After collapsing trades to family-dealer-collateral triples, we find that the percentage of trades attributable to the previously bound treatment cohort were, respectively, 23%, 28%, 15%, and 9%. 23 Though power is decreased, we find very similar results when separating the treatment group into funds that were unconstrained on the day of the cap increase and funds that remained bound by the cap. 24 The finding that Fed intervention can increase repo rates accords with Fleming, Hrung, and Keane (2010), who show that the Term Securities Lending Facility (TSLF)—a temporary emergency response to the developing financial crisis in 2008—resulted in higher repo rates. However, the TSLF boosted repo rates by mitigating acute shortages of Treasury collateral, whereas the ONRRP evidently works through increased bargaining power. 25 Other repo includes repo backed by assets such as asset backed securities, private label mortgage-backed securities, corporate bonds, and equity. 26 Because of changes in FR-2004C reporting requirements, we are unable to use the sample period employed earlier. The sample period for this analysis begins in April 2013. 27 Non-MMF lenders in the tri-party repo market include asset managers, primary dealer-affiliated institutions, securities lenders, banks (including custodian banks), pension funds, insurance companies, and others. 28 In addition to the low point estimate, we note that base effects also imply a small increase, as there is less borrowing from ineligible MMFs relative to other types of lenders. References Adrian, T., and Liang. N. 2016 . Monetary policy, financial conditions, and financial stability. Staff Report. Center for Economic Policy Research . Google Scholar CrossRef Search ADS Baklanova, V., Copeland, A. M. and McCaughrin. R. 2015 . Reference guide to US repo and securities lending markets. Working Paper, FRBNY . Bech, M., and Klee. E. 2011 . The mechanics of a graceful exit: Interest on reserves and segmentation in the federal funds market. Journal of Monetary Economics 58 : 415 – 31 . Google Scholar CrossRef Search ADS Bernanke, B. S. 2012 . Monetary policy since the onset of the crisis. Speech at the Federal Reserve Bank of Kansas City Economic Symposium, Jackson Hole, WY, August 31 . http://www.federalreserve.gov/newsevents/speech/bernanke20120831a.htm. Bertrand, M., Duflo, E. and Mullainathan. S. 2004 . How much should we trust differences-in-differences estimates? Quarterly Journal of Economics 119 : 249 – 75 . Google Scholar CrossRef Search ADS BIS. 2015 . Central bank operating frameworks and collateral markets. Report. Committee on the global financial system markets committee. Bank for International Settlements . Borio, C., and Zhu. H. 2012 . Capital regulation, risk-taking and monetary policy: A missing link in the transmission mechanism? Journal of Financial Stability 8 : 236 – 51 . Google Scholar CrossRef Search ADS Brunetti, C., Filippo, M. D. and Harris. J. H. 2011 . Effects of central bank intervention on the interbank market during the subprime crisis. Review of Financial Studies 24 : 2053 – 83 . Google Scholar CrossRef Search ADS Carlson, M., Duygan-Bump, B. Natalucci, F. Nelson, W. R. Ochoa, M. Stein, J. and den Heuvel. S. V. 2016 . The demand for short-term, safe assets and financial stability: Some evidence and implications for central bank policies. International Journal of Central Banking 12 : 307 – 33 . Copeland, A., Davis, I. LeSueur, E. and Martin. A. 2014 . Lifting the veil on the U.S. bilateral repo market. Liberty Street Economics , July 9 . http://libertystreeteconomics.newyorkfed.org/2014/07/lifting-the-veil-on-the-us-bilateral-repo-market.html. Copeland, A., Duffie, D. Martin, A. and McLaughlin. S. 2012 . Key mechanics of the U.S. tri-party repo market. FRBNY Economic Policy Review 18 : 17 – 28 . Copeland, A., Martin, A. and Walker. M. 2014 . Repo runs: evidence from the tri-party repo market. Journal of Finance 69 : 2343 – 80 . Google Scholar CrossRef Search ADS de Andoain, C. G., Heider, F. Hoerova, M. and Manganelli. S. 2016 . Lending-of-last-resort is as lending-of-last-resort does: Central bank liquidity provision and interbank market functioning in the euro area. Journal of Financial Intermediation 28 : 32 – 47 . Google Scholar CrossRef Search ADS Di Maggio, M., and Tahbaz-Salehi. A. 2014 . Financial intermediation networks. Research Paper , Columbia Business School . Google Scholar CrossRef Search ADS Duygan-Bump, B., Parkinson, P. Rosengren, E. Suarez, G. A. and Willen. P. 2013 . How effective were the Federal Reserve emergency liquidity facilities? Evidence from the asset-backed commercial paper money market mutual fund liquidity facility. Journal of Finance 68 : 715 – 37 . Google Scholar CrossRef Search ADS Fleming, M. J., Hrung, W. B. and Keane. F. M. 2010 . Repo market effects of the term securities lending facility. American Economic Review 100 : 591 – 6 . Google Scholar CrossRef Search ADS FRBNY. 2009 . Operating policy: Statement regarding reverse repurchase agreements. https://www.newyorkfed.org/markets/operating_policy_091019.html. Frost, J., Logan, L. Martin, A. McCabe, P. Natalucci, F. and Remache. J. 2015 . Overnight RRP operations as a monetary policy tool: Some design considerations. Working Paper, Finance and Economics Discussion Series . Gorton, G. B. 2017 . The history and economics of safe assets. Annual Review of Economics 9 : 547 – 86 . Google Scholar CrossRef Search ADS Gorton, G. B., and Metrick. A. 2012 . Securitized banking and the run on repo. Journal of Financial Economics 104 : 425 – 51 . Google Scholar CrossRef Search ADS Greenwood, R., Hanson, S. G. and Stein. J. C. 2015 . A comparative-advantage approach to government debt maturity. Journal of Finance 70 : 1683 – 722 . Google Scholar CrossRef Search ADS Greenwood, R., Hanson, S. G. and Stein. J. C. 2016 . The federal reserve’s balance sheet as a financial stability tool. Jackson Hole Symposium : Federal Reserve Bank of Kansas City . Han, S., and Nikolaou. K. 2016 . Trading relationships in the OTC market for secured claims: Evidence from triparty repos. Board of Governors of the Federal Reserve System . Ihrig, J., Meade, E. and Weinbach. G. 2015 . Rewriting monetary policy 101: What’s the fed’s preferred post-crisis approach to raising interest rates? Journal of Economic Perspectives 29 : 177 – 98 . Google Scholar CrossRef Search ADS Infante, S. 2016 . Private money creation with safe assets and term premia. Working Paper, Finance and Economics Discussion Series . Google Scholar CrossRef Search ADS Kacperczyk, M., and Schnabl. P. 2013 . How safe are money market funds? Quarterly Journal of Economics 128 : 1073 – 122 . Google Scholar CrossRef Search ADS Kandrac, J. Forthcoming. The costs of quantitative easing: Liquidity and market functioning effects of Federal Reserve MBS purchases. International Journal of Central Banking . Khwaja, A. I., and Mian. A. 2008 . Tracing the impact of bank liquidity shocks: Evidence from an emerging market. American Economic Review 98 : 1413 – 42 . Google Scholar CrossRef Search ADS Krishnamurthy, A., Nagel, S. and Orlov. D. 2014 . Sizing up repo. Journal of Finance 69 : 2381 – 417 . Google Scholar CrossRef Search ADS Nagel, S. 2016 . The liquidity premium of near-money assets. Quarterly Journal of Economics 131 : 1927 – 71 . Google Scholar CrossRef Search ADS Parlatore, C. 2016 . Fragility in money market funds: Sponsor support and regulation. Journal of Financial Economics 121 : 595 – 623 . Google Scholar CrossRef Search ADS Stein, J. C. 2012 . Monetary policy as financial stability regulation. Quarterly Journal of Economics 127 : 57 – 95 . Google Scholar CrossRef Search ADS Sundaresan, S., and Wang. Z. 2008 . Y2K options and the liquidity premium in Treasury markets. Review of Financial Studies 22 : 1021 – 56 . Google Scholar CrossRef Search ADS Sunderam, A. 2014 . Money creation and the shadow banking system. Review of Financial Studies 28 : 939 – 77 . Google Scholar CrossRef Search ADS Woodford, M. 2016 . Quantitative easing and financial stability. Working Paper , NBER . Yellen, J. L. 2014 . Monetary policy and financial stability. Speech at the Michel Camdessus Central Banking Lecture, International Monetary Fund, Washington, DC , July 2 . https://www.federalreserve.gov/newsevents/speech/yellen20140702a.htm. Published by Oxford University Press on behalf of The Society for Financial Studies 2017. This work is written by US Government employees and is in the public domain in the US. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Review of Financial Studies Oxford University Press

Monetary Policy Implementation and Financial Vulnerability: Evidence from the Overnight Reverse Repurchase Facility

Loading next page...
 
/lp/ou_press/monetary-policy-implementation-and-financial-vulnerability-evidence-0dhk5foRDo
Publisher
Oxford University Press
Copyright
Published by Oxford University Press on behalf of The Society for Financial Studies 2017. This work is written by US Government employees and is in the public domain in the US.
ISSN
0893-9454
eISSN
1465-7368
D.O.I.
10.1093/rfs/hhx141
Publisher site
See Article on Publisher Site

Abstract

Abstract In this paper, we examine the Federal Reserve’s newest policy tool, known as the overnight reverse repo (ONRRP) facility, to understand its effects on the repo market. Using exogenous variation in the parameters of the ONRRP, we show that private repo activity is crowded out when money funds invest in the ONRRP. Additionally, we find that the ONRRP increases lenders’ bargaining power, thereby raising borrower funding costs. Lastly, we show that repo borrowers reallocate to repo backed by riskier collateral and borrow more from ONRRP-ineligible asset managers, both of which could increase financial vulnerability due to instability in dealer funding. Received April 27, 2017; editorial decision October 19, 2017 by Editor Stijn Van Nieuwerburgh. In its response to the recent financial crisis, the Federal Reserve (the Fed) intervened heavily in financial markets to implement a series of quantitative easing (QE) programs. Indeed, in the years since the crisis, many central banks resorted to similar measures in an effort to buoy local economies. One consequence of this aggressive central bank intervention has been a rapid expansion of monetary authorities’ asset holdings and presence in financial markets. This unprecedented intervention has brought forth questions regarding its potentially negative effect on financial stability (Borio and Zhu 2012; Yellen 2014; Adrian and Liang 2016). The possible difficulty associated with the removal of this unprecedented accommodation has also raised financial stability concerns (Bernanke 2012). In the case of the Fed, the first major central bank to embark on a tightening cycle after the crisis, the abundance of liquidity prevented a return to the traditional means of monetary policy implementation. Historically, the Fed relied on tight control over the supply of reserves held by the banking system in order to control its main policy rate, known as the federal funds rate. In a reserve-abundant environment, modest changes in the supply of reserves will have no effect on the effective federal funds rate (Ihrig, Meade, and Weinbach 2015). Instead, the creation of a new policy tool known as the overnight reverse repurchase (ONRRP) facility was required to lift short-term rates. In this paper, we conduct the first thorough analysis of the effects of the ONRRP facility on financial markets, and demonstrate that there can be a tradeoff between effective monetary policy implementation and financial stability. Under the ONRRP facility, money market participants including money market funds (MMFs), government-sponsored enterprises (GSEs), banks, and dealers can enter into reverse repurchase agreements with the Fed at a prespecified fixed rate. In these transactions, counterparties lend cash to the Fed overnight, with Treasury securities from the Fed’s portfolio posted as collateral. To gain operational experience with the ONRRP, the Fed began conducting daily tests of the ONRRP facility in 2013. During these tests, the Fed occasionally adjusted parameters of the facility, such as the fixed offer rate and the maximum counterparty bid amount, or “cap.” We analyze the effects of the ONRRP facility by examining the facility’s influence on MMF investment decisions and the systemically important tri-party repo market.1 In particular, we exploit exogenous changes in the ONRRP counterparty cap during the testing phase to identify the displacement of private activity in the repo market as a consequence of Fed intervention. Upon the introduction of the ONRRP, some MMFs faced a constrained investment allocation decision because of the low level of the ONRRP facility’s maximum daily counterparty cap, whereas others were unconstrained by the cap. Future cap hikes precipitated exogenous increases in the ability of initially constrained MMFs to invest in the ONRRP, and provide events around which we construct difference-in-difference estimates to identify the effect of ONRRP participation on both MMFs’ lending decisions and the private repo market. Using a combination of MMF regulatory filings as well as confidential data on tri-party repo market transactions and dealer repo positions, we are able to trace out the effects of the ONRRP on financial markets. We find that by borrowing in the tri-party repo market, the Fed crowds out private MMF repo lending backed by both Treasury and agency collateral, while also reducing MMF investment in bank deposits. However, cash lenders in the tri-party repo market evidently value their dealer relationships highly enough to maintain relationships with all their dealers, a result that is consistent with the importance of relationships in the repo market described in Copeland et al. (2012). In further analysis, we examine the effects of the ONRRP facility on money funds’ bargaining power. We show that an exogenous positive shock to a fund’s ability to invest in the ONRRP facility improves its bargaining power relative to borrowers, thereby increasing private repo rates and, by extension, dealers’ funding costs. This combination of higher rates and lower volumes is consistent with the view that the ONRRP introduced a negative funding supply shock to the private repo market. Lastly, we analyze how repo borrowers respond to these adverse funding shocks. We find that dealers do not decrease their total repo borrowing. Rather, dealers make up for the repo supply lost to the ONRRP by adjusting both their collateral and lender composition. Dealers that are more exposed to MMFs investing in the ONRRP and thus subject to larger adverse funding shocks aim to maintain their relationships with the largest (ONRRP-eligible) asset management complexes by transacting with their non-MMF affiliates. Because non-MMFs lend more against nongovernment collateral repo, dealers must accommodate this transition by expanding their borrowing against riskier collateral and increasing their reliance on risky repo for net financing. To a lesser degree, these dealers also recoup the repo borrowing lost from ONRRP-eligible MMFs by turning to ONRRP-ineligible MMFs, which tend to be smaller and have less stable net assets. The findings reported in this paper indicate that the ONRRP facility can increase vulnerabilities in the financial system through two distinct channels. First, we find that dealers do not decrease total repo issuance in response to the negative funding supply shock caused by the ONRRP, choosing instead to boost their reliance on riskier types of repo borrowing. This adverse impact on dealers’ funding risk leads to heightened financial fragility in this sector and thus an increase in the risks that an episode of financial stress will result in a disruptive funding shock. Second, we find that MMFs withdraw repo funding from their regular dealer counterparties in order to invest with the Fed. Thus, if MMFs become concerned about the actual or perceived safety of their investments (Kacperczyk and Schnabl 2013), they would be likely to invest larger amounts with the Fed at the expense of repo with dealer counterparties. In a severe event, MMFs may even begin to substitute out of other investments, such as commercial paper, in favor of the ONRRP. However, we also find evidence of the importance of borrower-lender relationships in the repo market, which may counteract these adverse effects in severe stress events if MMFs are reluctant to completely sever ties with their dealer counterparties. This study connects three separate strands of literature. First, we contribute to the literature that examines how novel monetary policy programs affect financial markets. Previous studies demonstrate that new policy actions can have a largely favorable impact on prices and quantities in the targeted markets. For instance, Sundaresan and Wang (2008) show that the Fed’s “Y2K options” lowered liquidity premiums in the Treasury market, while Duygan-Bump et al. (2013) show that the Fed’s Asset-Backed Commercial Paper Money Market Mutual Fund Liquidity Facility stabilized MMF asset flows and reduced yields during the financial crisis. However, there is also evidence of potentially unintended consequences of policy actions. Particularly during periods of heavy intervention or acute market strain, the implementation of monetary policy has been shown to crowd out private activity. For example, Brunetti, Filippo, and Harris (2011) present results that suggest ECB intervention during the financial crisis crowded out interbank trades. In additional analysis of ECB actions both before and during the European sovereign debt crisis, de Andoain et al. (2016) show that ECB liquidity injections displaced private activity in the euro area, notably reducing interbank trading. Separately, Kandrac (Forthcoming) documents that Federal Reserve purchases under QE programs displaced private activity in the secondary market for mortgage-backed securities. Consistent with the pattern evident in these studies, we show that the new ONRRP facility crowds out private repo lending supplied by ONRRP-eligible MMFs. Second, our work relates to the growing literature on the financial stability effects of monetary policy (Borio and Zhu 2012; Adrian and Liang 2016), and especially those studies that examine the effects of recent unconventional policies (Stein 2012; Woodford 2016). The general thrust of this literature acknowledges the potential for heightened financial fragility owing to extraordinarily accommodative monetary policy. However, Greenwood, Hanson, and Stein (2015) argue that ONRRP operations can support financial stability by crowding out private money creation and its attendant negative externalities. While we find that counterparties that directly lend to the Fed via the ONRRP do in fact reduce their private repo volumes, dealer repo borrowers evidently make up for the reduction in repo funding by borrowing more from ONRRP-ineligible asset managers against riskier types of collateral. Third, we contribute to the literature that seeks to further understand segments of the increasingly important shadow banking sector including MMFs, securities dealers, and the collateral-backed funding markets that have accompanied the expansion of shadow banking (Sunderam 2014; Di Maggio and Tahbaz-Salehi 2014). In particular, previous studies have documented the growing reliance on the repo market for short-term funding, as well as the role of the repo market in the recent financial crisis (Gorton and Metrick 2012; Copeland, Martin, and Walker 2014; Krishnamurthy, Nagel, and Orlov 2014). Here, we consider a standing monetary policy tool that introduces the Fed as a large, persistent repo borrower willing to expand the supply of a safe asset to a variety of money market participants, and show the resultant effects on private tri-party repo activity. 1. Institutional Background 1.1 The Fed’s ONRRP facility Beginning in 2009, the Fed dramatically expanded its balance sheet through several rounds of asset purchase programs known as QE. Consequently, banks were left with large amounts of excess reserve balances on deposit with the Fed. Given that monetary policy implementation before the crisis relied on altering the amount of scarce reserves to influence short-term rates, the Fed required additional tools to conduct monetary policy in the new environment (Ihrig, Meade, and Weinbach 2015). The primary tool, paying banks interest on excess reserves (IOER), was implemented in October 2008. In theory, banks should be unwilling to lend at less than IOER—the deposit rate that they can earn from the Fed—causing IOER to set a floor on short-term interest rates. However, in the United States, there are many nonbank participants in money markets that do not have access to IOER and are therefore willing to lend at lower rates (Bech and Klee 2011). The difference between IOER and overnight lending rates ostensibly presents an arbitrage opportunity for banks, which could be eliminated through competition among borrowers in the overnight market. However, this arbitrage opportunity was limited by costs of holding reserves, such as minimum capital requirements and a fee charged on total assets to insure bank deposits. To regain control over overnight rates, the Fed responded in two ways: first, by expanding the set of counterparties with which it transacts and second, by introducing the ONRRP facility. As early as October 2009, the Fed was considering using repo transactions as a means of withdrawing policy accommodation, while also expanding the set of eligible repo counterparties to broaden the reach of such transactions (FRBNY 2009). Specifically, repo operations were made available to MMFs, GSEs, and banks, as well as the Fed’s traditional counterparties, primary dealers. Market participants could apply to become a Fed counterparty in several rounds beginning in March 2010.2 The Fed’s eligibility requirements generated important differences between ONRRP eligible and ineligible MMFs. In particular, eligible MMFs tended to be larger, more established funds. This was a result of the requirement that MMFs have at least $\$$ 5 billion in net assets and exist for at least one year to be eligible to participate in the ONRRP. Eligible MMFs also featured more stable assets and repo participation. Between 2011 and 2015, the standard deviation of net assets was 18% for eligible MMFs versus 25% for ineligible MMFs. Likewise, the standard deviation of repo lending volume was 9% for eligible MMFs and 14% for ineligible MMFs. In September 2013, subsequent to the expansion of the Fed’s RRP eligible counterparties, the Fed began conducting regular overnight, fixed-rate capped-allotment reverse repurchase agreements through an extended testing exercise. In these repo transactions, the Fed borrows from a broad set of counterparties on an overnight basis using Treasury securities as collateral.3 At the program’s start, 140 counterparties were eligible to participate in the ONRRP, including MMFs, GSEs, banks, and primary dealers.4 Importantly, the availability of the ONRRP to a broad set of counterparties operating in different market segments ensures that the ONRRP can help establish a firmer floor on short-term rates, in contrast to IOER, which is only offered to depository institutions. As a result of successful testing, the FOMC’s Policy Normalization Principles and Plans, released in September 2014, states the Committee’s intention to use the ONRRP facility as a tool to help control the federal funds rate during the normalization of the stance of monetary policy.5 Moreover, as discussed by the Committee at the November 2016 FOMC meeting, it is possible that there will not be a return to the operating framework that prevailed before the crisis, establishing the ONRRP as a key policy tool indefinitely.6 By offering an outside option to cash investors, the ONRRP can theoretically influence rates even without active participation from repo lenders. Nevertheless, the ONRRP has evidently presented an attractive investment opportunity to investors, especially MMFs, which typically account for over 85% of total take-up. As shown in Figure 1, participation in the facility has often reached sizeable levels. Through 2015, the average total take-up was $\$$ 115.5 billion per day, with a peak of $\$$475 billion in December 2015. One important aspect of the ONRRP facility is that a maximum bid, or cap, is imposed on each individual counterparty. As shown in Figure 2, this maximum has ranged from $\$$500 million at the start of the program to $\$$30 billion through 2017. The maximum individual bid is the primary constraint for participants.7 Changes in the maximum individual bid amount of the ONRRP are key to our identification strategy described in Section 2. Figure 1 View largeDownload slide Investment in the RRP facility This figure shows total daily participation in the RRP facility, including both overnight and term, from its inception through 2015 for MMFs, GSEs, and all others. Source: Federal Reserve Bank of New York, 2013–2015, Reverse Repo Data. Figure 1 View largeDownload slide Investment in the RRP facility This figure shows total daily participation in the RRP facility, including both overnight and term, from its inception through 2015 for MMFs, GSEs, and all others. Source: Federal Reserve Bank of New York, 2013–2015, Reverse Repo Data. Figure 2 View largeDownload slide ONRRP counterparty caps This figure shows the maximum ONRRP bid allowed per counterparty over time. Figure 2 View largeDownload slide ONRRP counterparty caps This figure shows the maximum ONRRP bid allowed per counterparty over time. While the ONRRP has been very effective at controlling short-term interest rates as intended, the presence of a standing reverse repo facility may also have consequences for financial stability that are theoretically unclear. One leading concern regarding the ONRRP is that a standing Fed borrowing facility may increase the likelihood or severity of flight-to-safety “runs” in the private repo market if investors suddenly become reluctant to lend to non-Fed counterparties (Frost et al. 2015). Outside of severe stress events, however, the withdrawal of MMFs from dealers in favor of the Fed could cause dealers to respond by reallocating their repo funding to less stable sources. Such a rise in dealers’ funding risk would increase financial vulnerabilities and heighten the susceptibility of this sector to stress events. Conversely, if the ONRRP facility crowds out privately issued repo created by the shadow-banking sector, it can buttress financial stability by reducing the externalities associated with a large private repo market, which—as demonstrated by the recent financial crisis—is subject to highly disruptive runs (Stein 2012; Carlson et al. 2016; Greenwood, Hanson, and Stein 2016). In addition to crowding out risky private repo, the ONRRP could help reverse the withdrawal of safe assets that occurred during QE by providing an interest-bearing near-money asset to a broad array of counterparties (BIS 2015; Infante 2016; Gorton 2017) and potentially lowering liquidity premiums (Nagel 2016). 1.2 The tri-party repo market In the United States, the repo market is divided into several segments, including the bilateral repo market, the tri-party repo market, and the GCF market, which is a subset of tri-party.8 Given data limitations, the exact size of the repo market and its subcomponents is not attainable. However, it is estimated that the tri-party market accounted for about 40%–45% of the repo market, or about $\$$ 1.5 trillion per day in 2014 and 2015 (Copeland et al. 2014; Baklanova, Copeland, and McCaughrin 2015). The tri-party market has historically relied upon two third-party clearing banks—Bank of New York-Mellon and J.P. Morgan—that settle all repo transactions. Tri-party market transactions typically consist of nondealers, including MMFs, securities lenders, and others, lending to dealers. MMFs conduct most of their repo lending in the tri-party market and account for about a third of daily tri-party volume. Tri-party collateral is specified only by type, rather than by security; that is, it is a general collateral market. Fedwire-eligible collateral, including Treasuries, agency debt, and agency mortgage-backed securities (MBS), account for about 85% of market volume as of 2012 (Copeland et al. 2012). We generally limit our attention to the tri-party repo market in this paper for two reasons. First, all ONRRP operations are conducted in the tri-party market. Second, our primary focus is on the response of MMF repo lending activity, which also takes place over the tri-party platform, to the ONRRP. 2. Identification In general, our aim is to evaluate the effects of the Fed’s ONRRP facility on the private repo market. Because the ONRRP is a standing facility, repo investors (such as MMFs and government-sponsored entities) can choose at will to lend risklessly to the central bank at the facility’s rate rather than to their usual dealer counterparties. Even without actual participation in the facility by ONRRP counterparties, the mere presence of the facility could boost MMF’s bargaining power vis-à-vis dealers, thereby exerting upward pressure on repo and other short-term rates. However, as demonstrated in Figure 1, MMFs evidently found that the terms offered by the ONRRP facility presented an attractive investment opportunity, and did indeed participate in the facility, possibly at the expense of existing counterparties. In Figure 3, we show total MMF repo investment composition, subdivided into Fed (i.e., ONRRP) and private components. Ostensibly, total MMF repo investment remained roughly constant over the last four years, and the entrance of the Fed claimed market share from MMF’s private counterparties. Figure 3 View largeDownload slide Money market fund repo investment This figure shows the monthly private versus Fed decomposition of MMF repo investment backed by Treasury and agency collateral. Source: SEC form N-MFP. Figure 3 View largeDownload slide Money market fund repo investment This figure shows the monthly private versus Fed decomposition of MMF repo investment backed by Treasury and agency collateral. Source: SEC form N-MFP. However, participation in the ONRRP facility is an endogenous outcome of borrower and lender interactions in money markets. Consequently, other factors may have reduced dealers’ willingness or ability to borrow in the repo market. For example, regulatory pressures on dealers including more stringent capital requirements, the introduction of the liquidity coverage ratio, and less permissive rules regarding the netting of repo trades may have reduced dealers’ willingness to borrow in the repo market. In addition, as previously seen in Figure 3, total private tri-party repo declined starting in January 2013, when the supplementary leverage ratio was implemented internationally. As a result of these changes, increases in MMF ONRRP participation may only be coincidentally timed with a reduction in private repo, which could simply reflect developments in the market that are unrelated to the introduction of the ONRRP facility. In other words, it may be that MMFs participate in the ONRRP differentially as a result of their available options in money markets, thereby introducing bias in attempts to estimate a causal effect of ONRRP participation on the investment outcomes of Fed counterparties. To overcome these endogeneity issues and draw causal inference, we exploit exogenous variation in the ONRRP facility parameters, which were occasionally adjusted during the testing phase. In particular, we focus on the changes to the counterparty caps that regulated the maximum daily bid of individual counterparties. As shown in Figure 2 and discussed in the previous section, the counterparty cap was raised in increments over the first year of the ONRRP’s existence. Funds that would have optimally invested more in the ONRRP facility than permitted by the maximum bid amounts were thus forced to invest at the maximum bid. Evidently facing a constrained investment decision, these funds should be expected to boost their ONRRP investment upon an increase in the counterparty cap. Importantly, the maximum bid amounts were raised merely as a normal result of the expansion of the ONRRP tests, and were unrelated to any changes in either the repo market or MMF investment decisions and preferences. Thus, MMFs that were initially constrained by the counterparty cap experienced an exogenous change in their ability to invest in the ONRRP facility. MMFs that were constrained by the initial ONRRP cap at the facility’s inception compose our “treatment” group, as these funds witnessed exogenous increases in their ability to invest in the ONRRP upon future cap increases.9,10 We can then compare treated funds to ONRRP-eligible MMFs that were not constrained by the initial counterparty cap—our “control” group—to generate differences-in-differences (DD) estimates of the effect of an increase in ONRRP participation. Counterparty eligibility for the ONRRP was determined by an application process available to all qualifying funds. Thus, to avoid any possible selection issues emanating from the application decision, we limit our sample to ONRRP-eligible counterparties only. Figure 4 depicts an illustrative example of our identification strategy. Upon introduction of the ONRRP, period 1 on the left side of the figure, the initial counterparty cap does not constrain fund 1, which is able to achieve its optimal investment allocation. Fund 1 is therefore part of the control group. Conversely, funds 2 and 3 face a binding counterparty cap at the inception of the ONRRP. These funds compose our treatment group. In period 2—the beginning of the “post-treatment” period—the counterparty cap has been increased, affording funds 2 and 3 the ability to increase their ONRRP investment, and fund 1 continues to submit below-cap bids. Thus, our treatment and control funds are established at the inception of the ONRRP, and we identify off of subsequent cap changes that present exogenous events around which we can construct DD estimates. Figure 4 View largeDownload slide Identification strategy: An example of control and treatment groups This figure depicts an illustrative example of our identification strategy. In the first period, fund 1 (yellow) does not face a constrained investment allocation decision as a result of the counterparty cap (the dashed line) and is included in the control group. Funds 2 and 3 (red and blue, respectively) are bound by the cap and will compose the treatment group. In the second period, fund 2 now faces an unconstrained investment decision as a result of the increase in the counterparty cap and is thus included in the $$Unbound$$ treatment subgroup. Although fund 3 was able to increase its ONRRP investment as a result of the increase in the counterparty cap in period 2, it continues to face a constrained investment decision and is thus included in the $$Bound$$ treatment subgroup. In period 3, the counterparty cap is again increased, with all funds facing an unconstrained investment decision such that fund 3 joins the $$Unbound$$ treatment subgroup with fund 2. Figure 4 View largeDownload slide Identification strategy: An example of control and treatment groups This figure depicts an illustrative example of our identification strategy. In the first period, fund 1 (yellow) does not face a constrained investment allocation decision as a result of the counterparty cap (the dashed line) and is included in the control group. Funds 2 and 3 (red and blue, respectively) are bound by the cap and will compose the treatment group. In the second period, fund 2 now faces an unconstrained investment decision as a result of the increase in the counterparty cap and is thus included in the $$Unbound$$ treatment subgroup. Although fund 3 was able to increase its ONRRP investment as a result of the increase in the counterparty cap in period 2, it continues to face a constrained investment decision and is thus included in the $$Bound$$ treatment subgroup. In period 3, the counterparty cap is again increased, with all funds facing an unconstrained investment decision such that fund 3 joins the $$Unbound$$ treatment subgroup with fund 2. As portrayed in the figure, in period 2, the maximum allowable bid is increased by an amount large enough to leave fund 2 unconstrained by the new cap, while fund 3 remains constrained by the new, higher cap. Although we consider both funds 2 and 3 to have received treatment as a result of the cap increase, we can differentiate between two treated subgroups: one treatment subgroup transitions from a constrained state to a new constrained state despite the higher cap (e.g., fund 3 in period 2), while the other transitions from a constrained state to an unconstrained state (e.g., fund 2 in period 2). Eventually, all treated funds in our sample enter the unconstrained treatment group, as depicted in period 3, when the counterparty cap is increased to a level allowing fund 3 to achieve an unconstrained investment allocation. Employing such a DD identification strategy addresses many threats to the casual interpretation of our results. First, DD allows us to account for constant differences between funds, as well as any overarching factors that could affect the whole repo market at any point in time. That is, our DD estimation controls for differences between treated and control MMFs and for shared differences between quarter-ends. Moreover, cap increases were numerous, variably sized, and occurred at irregular intervals. This rich variation in our exogenous treatment events means that other explanations of the effects we observe require a time pattern that is well correlated with the timing of cap increases and affect only those funds that were subject to a binding cap at the introduction of the ONRRP (i.e., treated funds). There is also variation in the timing of treatment funds’ transition to the unconstrained treatment group. As a result, any factors confounding the causal interpretation of our estimates must affect only a specific (and changing) subset of treatment funds at specific points in time. In addition, we note that the tri-party repo market introduces frictions that can limit alternate sources of within-fund variation over time, particularly as it concerns each MMF’s set of borrowers. For instance, as described in Copeland et al. (2012), each MMF must execute a master repo agreement with each individual borrower, which lays out important elements of their repo transactions. Before tri-party repo trading can commence, each MMF-dealer pair must additionally execute a unique custodial undertaking agreement with the clearing bank to establish its role as the agent. Lastly, most repo transactions are overnight trades, and many of these are open transactions between a given borrower and lender that are continuously “rolled over” on a day-to-day basis. Thus, MMFs can only transact with those dealers with which they have executed the necessary agreements, and trading conventions in the repo market ensure that MMF-dealer trading relationships are typically persistent. We follow the general DD identification strategy outlined above throughout the remainder of the paper. Summary statistics, for our treatment and control groups follow in Section 3 and the details of the precise DD specifications are provided in Section 4. 3. Data To evaluate the effect of the ONRRP facility on MMFs and their dealer borrowers, we employ one comprehensive public data set of MMF investments, three confidential data sets reported to the Fed that contain detailed information on tri-party repo market transactions and relationships, and a confidential data set that lists the entire repo and reverse repo portfolios of individual securities dealers on a weekly basis.11 First, we use the monthly MMF filings of form N-MFP to the Securities and Exchange Commission, which contain a detailed schedule of portfolio holdings for each MMF on month-ends. These data contain information on each security, including the issuer, security type, maturity date, and volume. In addition, we use three different data sets containing information on the tri-party repo market. The first contains pricing information that allows us to analyze rates in the tri-party repo market. This data set records all transactions facilitated by one of the two tri-party clearing banks in the tri-party repo market, which are reported to the Federal Reserve Bank of New York (FRBNY). Lenders in this data set are provided at the mutual fund complex (“family”) level, rather than the individual fund level. Trade information includes the volume, annualized rate, term to maturity, dealer, lender, and underlying collateral backing each repo transaction. For our purposes, we exclude term trades and retain only those transactions backed by Fedwire-eligible Treasury and agency (including agency MBS and agency debt) collateral. Neither of these filters significantly reduces our sample size, as overnight trades backed by Treasury or agency collateral compose over 70% of the transactions in the data. The second tri-party data set contains dealer position-level information, reported by both clearing banks, providing a complete picture of total repo borrowing in the tri-party market. For each dealer, this data set records the daily amount of repo outstanding for each collateral type. With these data, we are able to examine repo of all tenors and each collateral type to assess how dealers reallocate their funding in response to the change in behavior of their lenders as a result of the ONRRP. It is also possible to calculate the haircut of each repo position from these data, which compares the difference between the value of the collateral and the value of the cash loan normalized by the value of the cash loan. Higher haircuts are imposed when using riskier collateral, with more collateral being required for a given loan amount (Gorton and Metrick 2012). The third tri-party data set is also position-level, but it contains daily outstanding repo between each dealer and lender (with no collateral information). While some small investors are not included, the positions recorded in this data set account for more than three-quarters of total tri-party repo positions. These data allow us to analyze how dealers’ counterparty composition changes as their lenders shift to the ONRRP. Lastly, we compile a data set from the weekly filings of form FR-2004C, which is used to collect information on dealer financing from primary dealers.12 These forms are submitted on a weekly basis, and contain confidential information collected under the Fed’s supervisory authority. These data allow us to examine the entirety of dealers’ repo and reverse repo positions by collateral type and tenor. Dealers’ repo borrowing can reflect funding used for financing a net securities position, or it can reflect so-called “matched book” activity. A dealer’s matched book, which entails offsetting repo and reverse repo positions, increases the balance sheet size but does not materially increase its riskiness. This is because dealers that experience a withdrawal of repo lending can simply refrain from rolling over reverse repos with matching tenors. Conversely, dealers that use repo as a source of net financing for securities holdings expose themselves to risk because of the maturity mismatch between their net securities holdings and their short-term repo funding. The FR-2004C data therefore enable us to examine the extent to which the riskiness of dealers’ funding increases when lenders shift to the ONRRP. For example, if dealers respond to the withdrawal of funding by simply reducing matched book activity (i.e., simultaneously reducing reverse repos), this would not necessarily represent an increase in financial fragility. Summary statistics for both treated (initially constrained) and control (initially unconstrained) MMFs are shown in Table 1. Since most of our analysis is conducted using quarter-end observations when ONRRP participation is typically highest, all values are reported as of June 28, 2013, the last quarter-end before the ONRRP began. Our sample consists of 101 ONRRP-eligible MMFs, of which 39 compose our treatment group, as these funds were constrained by the initial (September 30, 2013) quarter-end counterparty cap of $\$$ 1 billion. Of the treated MMFs, 21% remained constrained by the $\$$3 billion December 2013 cap, 5% were constrained by the $\$$7 billion cap in March 2014, and 4% were constrained by the $\$$10 billion cap in June 2014. The last counterparty cap increase to $\$$30 billion cap in September 2014 effectively removed the constraint for the remaining treatment funds. Table 1 Summary statistics Treated Control Variable (Initially constrained) (Initially unconstrained) Number of funds 39 62 Number of prime funds 20 45 Number of dealers per fund 9.43 8.09 (0.74) (0.57) % of foreign dealers per fund 60.71 61.66 (4.42) (3.14) Weighted CDS spread (%) 0.85 0.83 (0.04) (0.04) HHI 0.18 0.20 (0.03) (0.03) Share of largest dealer 0.26 0.28 (0.03) (0.03) Weighted Treasury/Agency repo rate (bps) 13.38 11.62 (1.23) (0.84) Weighted Treasury repo rate (bps) 10.38 10.02 (0.38) (0.02) AUM ( $\$$, billions) 27.6 14.4$$^{***}$$ (4.05) (2.12) Treasury repo (% of AUM) 15.48 5.97$$^{***}$$ (3.67) (1.70) Agency repo (% of AUM) 11.85 7.98 (2.64) (1.35) CD (% of AUM) 18.25 20.79 (3.21) (2.24) Treasury debt (% of AUM) 14.31 15.08 (2.92) (2.78) Financial CP (% of AUM) 7.57 10.78$$^{*}$$ (1.54) (1.15) Asset-backed CP (% of AUM) 2.01 6.84$$^{***}$$ (0.52) (1.00) Treated Control Variable (Initially constrained) (Initially unconstrained) Number of funds 39 62 Number of prime funds 20 45 Number of dealers per fund 9.43 8.09 (0.74) (0.57) % of foreign dealers per fund 60.71 61.66 (4.42) (3.14) Weighted CDS spread (%) 0.85 0.83 (0.04) (0.04) HHI 0.18 0.20 (0.03) (0.03) Share of largest dealer 0.26 0.28 (0.03) (0.03) Weighted Treasury/Agency repo rate (bps) 13.38 11.62 (1.23) (0.84) Weighted Treasury repo rate (bps) 10.38 10.02 (0.38) (0.02) AUM ( $\$$, billions) 27.6 14.4$$^{***}$$ (4.05) (2.12) Treasury repo (% of AUM) 15.48 5.97$$^{***}$$ (3.67) (1.70) Agency repo (% of AUM) 11.85 7.98 (2.64) (1.35) CD (% of AUM) 18.25 20.79 (3.21) (2.24) Treasury debt (% of AUM) 14.31 15.08 (2.92) (2.78) Financial CP (% of AUM) 7.57 10.78$$^{*}$$ (1.54) (1.15) Asset-backed CP (% of AUM) 2.01 6.84$$^{***}$$ (0.52) (1.00) This table reports summary statistics as of June 28, 2013 (the last quarter-end before the ONRRP began) for treated (initially constrained) and control (initially unconstrained) MMFs in our sample. All dealer-related variables (number of dealers, % foreign, CDS spread, HHI, and share of largest dealer) are based on a sample of only MMFs that have nonzero private repo volume (96 MMFs rather than the full 101). Rates are calculated at a fund complex level, and all other variables are at the individual MMF level. Statistical significance of the difference in averages: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 1 Summary statistics Treated Control Variable (Initially constrained) (Initially unconstrained) Number of funds 39 62 Number of prime funds 20 45 Number of dealers per fund 9.43 8.09 (0.74) (0.57) % of foreign dealers per fund 60.71 61.66 (4.42) (3.14) Weighted CDS spread (%) 0.85 0.83 (0.04) (0.04) HHI 0.18 0.20 (0.03) (0.03) Share of largest dealer 0.26 0.28 (0.03) (0.03) Weighted Treasury/Agency repo rate (bps) 13.38 11.62 (1.23) (0.84) Weighted Treasury repo rate (bps) 10.38 10.02 (0.38) (0.02) AUM ( $\$$, billions) 27.6 14.4$$^{***}$$ (4.05) (2.12) Treasury repo (% of AUM) 15.48 5.97$$^{***}$$ (3.67) (1.70) Agency repo (% of AUM) 11.85 7.98 (2.64) (1.35) CD (% of AUM) 18.25 20.79 (3.21) (2.24) Treasury debt (% of AUM) 14.31 15.08 (2.92) (2.78) Financial CP (% of AUM) 7.57 10.78$$^{*}$$ (1.54) (1.15) Asset-backed CP (% of AUM) 2.01 6.84$$^{***}$$ (0.52) (1.00) Treated Control Variable (Initially constrained) (Initially unconstrained) Number of funds 39 62 Number of prime funds 20 45 Number of dealers per fund 9.43 8.09 (0.74) (0.57) % of foreign dealers per fund 60.71 61.66 (4.42) (3.14) Weighted CDS spread (%) 0.85 0.83 (0.04) (0.04) HHI 0.18 0.20 (0.03) (0.03) Share of largest dealer 0.26 0.28 (0.03) (0.03) Weighted Treasury/Agency repo rate (bps) 13.38 11.62 (1.23) (0.84) Weighted Treasury repo rate (bps) 10.38 10.02 (0.38) (0.02) AUM ( $\$$, billions) 27.6 14.4$$^{***}$$ (4.05) (2.12) Treasury repo (% of AUM) 15.48 5.97$$^{***}$$ (3.67) (1.70) Agency repo (% of AUM) 11.85 7.98 (2.64) (1.35) CD (% of AUM) 18.25 20.79 (3.21) (2.24) Treasury debt (% of AUM) 14.31 15.08 (2.92) (2.78) Financial CP (% of AUM) 7.57 10.78$$^{*}$$ (1.54) (1.15) Asset-backed CP (% of AUM) 2.01 6.84$$^{***}$$ (0.52) (1.00) This table reports summary statistics as of June 28, 2013 (the last quarter-end before the ONRRP began) for treated (initially constrained) and control (initially unconstrained) MMFs in our sample. All dealer-related variables (number of dealers, % foreign, CDS spread, HHI, and share of largest dealer) are based on a sample of only MMFs that have nonzero private repo volume (96 MMFs rather than the full 101). Rates are calculated at a fund complex level, and all other variables are at the individual MMF level. Statistical significance of the difference in averages: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. The two types of money funds in our sample, government and prime, each compose about half of the treatment group.13 Government MMFs are restricted to investing in Treasuries, agency debt, and repo backed by either Treasuries or agencies, while prime MMFs can invest in a broader range of assets, including bank deposits, commercial paper, corporate debt, and repo backed by non-Treasury and agency collateral. The control group also contains a mix of prime and government funds, although prime funds account for roughly two-thirds of this cohort. Treatment and control MMFs have no statistically significant differences in their dealer relationships. In particular, they are similar in their average number of dealers, percentage of lending to foreign dealers, riskiness of dealers (as measured by CDS spreads), concentration of lending across dealers (as measured by HHI), and share of the largest dealer.14 We take particular note of the groups’ similarity in their exposure to foreign dealers. Given geographical differences in the implementation of the Basel leverage ratio requirement, foreign dealers contract their repo borrowing on quarter-ends more than domestic dealers.15 Consequently, if constrained funds traded more with large foreign dealers that borrow less on quarter-ends, these funds would invest more in the ONRRP for this reason. Not only does this not seem to be the case given the similar dealer profiles of the two groups, but we also rule out such a borrower-driven explanation by controlling for dealer-time fixed effects in our robustness checks. The next two rows show that the treatment and control groups also receive the same rates in their tri-party repo transactions, both overall and in Treasury collateral repo only. However, the two MMF groups do differ in their size and asset allocation. Specifically, treated funds tend to be larger in terms of assets under management (AUM) and invest a relatively larger share of their assets in both Treasury and, to a lesser extent, agency repo. Control funds are more likely to invest in commercial paper, reflecting the higher proportion of prime funds. It seems reasonable that funds that both have more assets to invest and are more active in the repo market would be larger participants in the ONRRP and therefore more likely to be constrained by the initial counterparty cap, thus composing the treatment group. Separately, we compare the changes in AUM between the treatment and control funds in the weeks around cap increases after which there was at least one constrained fund. If the ability to invest in the ONRRP drives MMF investment flows, AUM may endogenously adjust to cap changes. In Figure 5, we use data from iMoneyNet to depict the weekly differences in the percentage change in AUM between control and treatment funds for each of the two weeks prior to and after the cap changes. No clear pattern is evident, and no statistical significance of the differences is achieved in any instance. In all cases, the percentage change in treatment funds’ AUM relative to control funds is very similar in the week after and week prior to a cap change. Thus, we conclude that neither group of funds witnessed abnormal inflows or outflows as a result of the cap changes. Figure 5 View largeDownload slide Changes in AUM around cap increases This figure depicts the difference in the percentage change in AUM between treatment and control funds for the 4 weeks around cap changes that occurred prior to each quarter-end observation. The point estimate represents the percentage increase (positive) or decrease (negative) in AUM witnessed by treatment funds relative to control (unconstrained) funds. Figure 5 View largeDownload slide Changes in AUM around cap increases This figure depicts the difference in the percentage change in AUM between treatment and control funds for the 4 weeks around cap changes that occurred prior to each quarter-end observation. The point estimate represents the percentage increase (positive) or decrease (negative) in AUM witnessed by treatment funds relative to control (unconstrained) funds. 4. Methods and Results In the subsections below, we describe the specific DD specifications and corresponding results that address four distinct questions regarding the effects of the Fed’s ONRRP facility. In Section 4.1, we examine whether, and to what extent, the ONRRP facility crowds out MMF investment in other asset categories. After demonstrating that the ONRRP causally crowds out private repo activity, Section 4.2 considers the effect of this substitution on MMFs’ dealer relationships. To invest more with the Fed, MMFs may turn away from some counterparties, and transact with fewer dealers. However, we find that the value of lending relationships in the repo market is high enough to prevent MMFs from dropping any of their dealer counterparties. Section 4.3 demonstrates that the ONRRP facility confers bargaining power on MMFs, thereby increasing dealer funding costs. Finally, we examine how dealers respond to the ONRRP-induced adverse funding shock in Section 4.4. We find that dealers that are more exposed to treated MMFs do not decrease their total repo borrowing but instead reallocate their borrowing within the repo market, turning mostly to repo backed by riskier collateral that they increasingly rely upon for net financing. This shift appears to be driven by dealers’ transition to non-MMF asset managers, as well as smaller ONRRP-ineligible MMF lenders. 4.1 Effects of the ONRRP on MMF asset allocation Our first goal is to estimate asset substitution effects in order to determine which money market segments experience a withdrawal of funding when MMFs invest in the ONRRP facility. The data used for this analysis come from the public quarter-end MMF filings of form N-MFP, discussed in Section 3. By aggregating MMF securities holdings from the N-MFP into distinct investment categories, we are able to track fund-level asset allocation over time, which we can then relate to changes in ONRRP participation. For this analysis, we sample the MMF asset portfolio holdings on quarter-end dates between December 2012 and June 2015. Table 2 reports average asset class holdings as a percentage of total AUM for the sample of all ONRRP-eligible MMF counterparties classified as either prime or government-only funds. The column on the left reports MMF’s asset shares as of Q3 2013, a few days after the introduction of the ONRRP facility. Summing the asset class shares reveals that the asset categories we consider account for about 70% of total AUM for MMFs in our sample. By Q3 2014, shown on the right of Table 2, there was a statistically significant reduction in the shares of MMF portfolios invested in Treasury-collateral repo, agency-collateral repo, and Treasury debt. As mentioned earlier, the especially large proportional declines in repo investments (about 43% and 39% for Treasury and agency repo, respectively) came amid regulatory pressures that led to an overall reduction in dealer balance sheets, limiting their participation in the repo market. Even outside of these regulatory issues, other factors causing MMFs’ planned or forced reductions in other asset classes could at least partly determine ONRRP participation. Consequently, an estimate of the substitution effects induced by ONRRP participation requires an identification strategy, such as the one described in Section 2, that relies on an exogenous change in MMFs’ ability or willingness to participate in the facility. Table 2 Money market fund investment shares Asset class 2013 Q3 2014 Q3 Fed ONRRP 3.36 13.27*** (0.37) (1.37) Treasury repo 9.39 5.34** (1.62) (1.00) Agency repo 7.92 4.83** (1.12) (0.73) CD 20.74 22.28 (1.92) (2.02) Treasury debt 14.35 9.83* (2.07) (1.92) Financial CP 9.16 8.80 (0.98) (0.86) Asset-backed CP 4.74 4.79 (0.67) (0.71) N 101 101 Asset class 2013 Q3 2014 Q3 Fed ONRRP 3.36 13.27*** (0.37) (1.37) Treasury repo 9.39 5.34** (1.62) (1.00) Agency repo 7.92 4.83** (1.12) (0.73) CD 20.74 22.28 (1.92) (2.02) Treasury debt 14.35 9.83* (2.07) (1.92) Financial CP 9.16 8.80 (0.98) (0.86) Asset-backed CP 4.74 4.79 (0.67) (0.71) N 101 101 This table reports the average asset holdings as a percentage of total assets under management for the pre-treatment quarter (2013 Q3) and the quarter just after the final increase in the counterparty cap (2014 Q3). Statistical significance of change in mean asset share: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 2 Money market fund investment shares Asset class 2013 Q3 2014 Q3 Fed ONRRP 3.36 13.27*** (0.37) (1.37) Treasury repo 9.39 5.34** (1.62) (1.00) Agency repo 7.92 4.83** (1.12) (0.73) CD 20.74 22.28 (1.92) (2.02) Treasury debt 14.35 9.83* (2.07) (1.92) Financial CP 9.16 8.80 (0.98) (0.86) Asset-backed CP 4.74 4.79 (0.67) (0.71) N 101 101 Asset class 2013 Q3 2014 Q3 Fed ONRRP 3.36 13.27*** (0.37) (1.37) Treasury repo 9.39 5.34** (1.62) (1.00) Agency repo 7.92 4.83** (1.12) (0.73) CD 20.74 22.28 (1.92) (2.02) Treasury debt 14.35 9.83* (2.07) (1.92) Financial CP 9.16 8.80 (0.98) (0.86) Asset-backed CP 4.74 4.79 (0.67) (0.71) N 101 101 This table reports the average asset holdings as a percentage of total assets under management for the pre-treatment quarter (2013 Q3) and the quarter just after the final increase in the counterparty cap (2014 Q3). Statistical significance of change in mean asset share: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. To generate a DD estimate of these substitution effects, we compare the control group of funds that were unconstrained by the introductory ONRRP counterparty cap to a treatment group of funds that faced a binding cap on Q3 2013. Since the cap was increased between each quarter-end, treated funds may no longer face a binding counterparty cap, or they could increase their ONRRP investment to such an extent that they are still bound by the new (higher) cap. Thus, we have two treatment subgroups denoted $$\text{UnboundTreatment}_{it}$$ and $$\text{BoundTreatment}_{it}$$, respectively. Specifically we estimate regressions of the following form: \begin{align} y_{it} &= \delta\cdot(\text{Post}_t\cdot\text{UnboundTreatment}_{it}) + \rho\cdot(\text{Post}_t\cdot\text{BoundTreatment}_{it}) \notag\\ &\quad + \beta_{i}\cdot\text{MMF}_{i} + \gamma_{t}\cdot\text{Quarter}_{t} + \varepsilon_{it}, \end{align} (1) \begin{equation} \text{where} \hspace{15mm} y_{it} = 100\cdot\frac{\text{AssetClass}_{it}}{\text{AUM}_{it}}. \nonumber \end{equation} In Equation (1), $$(\text{Post}_t\cdot\text{UnboundTreatment}_{it})$$ identifies the effect for our main treatment subgroup, taking a value of one for treated funds in the post-treatment period that are not bound by the cap in quarter t. All treated funds eventually enter this group. Thus, $$\delta$$ is the DD estimate of the marginal effect of increased ONRRP participation owing to an increase of the counterparty cap to a point at which it becomes nonbinding and the MMF faces an unconstrained investment decision. However, some treatment funds may continue to face a constraint even after the counterparty cap has been lifted and therefore compose a separate treatment subgroup. These funds also witness a sharp increase in their ability to participate in the ONRRP as a result of the exogenous cap increase, but we make a distinction between these still-bound treated funds as indicated by the $$\text{BoundTreatment}_{it}$$ dummy, which takes a value of one for funds that remained bound by the cap following an increase. Similar to $$\delta$$, $$\rho$$ estimates the marginal effect of ONRRP participation as a result of an increase in the counterparty cap, although $$\rho$$ measures the effect for funds in periods during which they continue to face a binding cap and hence a constrained investment decision. Equation (1) also includes fund fixed effects, as well as a full set of time fixed effects. Because an MMF’s total size may partially determine whether the counterparty cap is binding, we include MMF’s pre-treatment AUM in some specifications as a robustness check, interacted with a “post” dummy that takes a value of one after September 2013.16 Standard errors in Equation (1) are clustered at the fund complex level (Bertrand, Duflo, and Mullainathan 2004). In Figure 6, we show evidence of parallel trends in the asset shares between our control and treatment groups in the pre-treatment period. Although the patterns in Figure 6 suggest the parallel trends assumption is not violated, we nevertheless also include a robustness check that allows us to relax the parallel trends assumption as follows: Figure 6 View largeDownload slide Trends in asset shares This figure depicts the treatment effect for asset shares in the six quarters prior to implementation of the ONRRP. Thus, a positive value corresponds to quarters in which treated funds witness asset shares rising more (or falling less) than the asset share for control funds. A negative value corresponds to quarters in which treated funds witness asset shares rising less (or falling more) than the asset share for control funds. Figure 6 View largeDownload slide Trends in asset shares This figure depicts the treatment effect for asset shares in the six quarters prior to implementation of the ONRRP. Thus, a positive value corresponds to quarters in which treated funds witness asset shares rising more (or falling less) than the asset share for control funds. A negative value corresponds to quarters in which treated funds witness asset shares rising less (or falling more) than the asset share for control funds. \begin{align} y_{it} &= \delta\cdot(\text{Post}_t\cdot\text{UnboundTreatment}_{it}) + \rho\cdot(\text{Post}_t\cdot\text{BoundTreatment}_{it}) \notag\\ & \quad + \lambda_{i}\cdot(\text{MMF}_{i}\cdot t) + \beta_{i}\cdot\text{MMF}_{i} + \gamma_{t}\cdot\text{Quarter}_{t} + \varepsilon_{it}. \end{align} (2) In this specification, we add fund-specific time trends in order to capture any possible divergence in asset share trends between treatment and control funds. Although fund-level trends can weaken our results if the treatment effect emerges gradually over time, MMFs should reallocate their portfolios rapidly in response to cap increases. Additionally, we prefer to include this robustness check in light of the relatively long sample period required by the length of the ONRRP testing phase. Though MMFs’ asset mix does not typically change dramatically on a day-to-day basis, shifts in fund manager preferences or the market environment would cause some funds to adjust their investment mix over time. Table 3 reports results from estimating the specifications described above. Turning to the most basic specification in Column 1, we see that the coefficients on the treatment dummies ($$\text{Unbound}_{it}$$ and $$\text{Bound}_{it}$$) for the Fed RRP dependent variable verify the effect of treatment on MMF ONRRP investment. On average, an increase in the maximum counterparty cap led to treatment MMFs increasing their investment in the ONRRP facility by about 9.0% and 7.0% of assets for our two treatment subgroups. Table 3 Regression results: MMF asset substitution Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 9.01*** 10.05*** 9.14*** 10.10*** (1.86) (1.95) (2.43) (2.18) $$\text{Bound}_{it}$$ 6.99*** 8.59*** 7.28*** 9.20*** (1.63) (1.74) (1.70) (1.70) Treasury repo $$\text{Unbound}_{it}$$ –3.99*** –5.01*** –6.39*** –7.00*** (1.40) (1.62) (1.52) (1.73) $$\text{Bound}_{it}$$ –3.92*** –5.48*** –5.34*** –6.58*** (1.26) (1.52) (1.34) (1.46) Agency repo $$\text{Unbound}_{it}$$ –2.87* –3.09** –2.98** –3.13** (1.47) (1.54) (1.48) (1.54) $$\text{Bound}_{it}$$ –0.31 –0.65 –0.59 –0.88 (1.94) (2.10) (1.71) (1.79) CD $$\text{Unbound}_{it}$$ –2.29*** –2.87*** –1.00 –1.46 (0.88) (0.68) (1.27) (1.20) $$\text{Bound}_{it}$$ –2.38*** –3.27*** –1.62 –2.56* (0.79) (0.78) (1.19) (1.44) Treasury debt $$\text{Unbound}_{it}$$ –0.32 –0.47 0.20 0.29 (1.62) (1.67) (1.93) (2.06) $$\text{Bound}_{it}$$ 1.42 1.18 1.53 1.71 (1.41) (1.45) (1.48) (1.60) Financial CP $$\text{Unbound}_{it}$$ 0.74 0.68 0.40 0.26 (0.91) (0.99) (0.92) (0.92) $$\text{Bound}_{it}$$ –0.11 –0.20 –0.30 –0.59 (0.74) (0.93) (0.95) (0.91) Asset-backed CP $$\text{Unbound}_{it}$$ 0.25 0.28 0.20 0.24 (0.49) (0.46) (0.57) (0.54) $$\text{Bound}_{it}$$ 0.39 0.43 0.33 0.40 (0.45) (0.45) (0.48) (0.45) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ N 1,111 1,111 1,111 1,111 Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 9.01*** 10.05*** 9.14*** 10.10*** (1.86) (1.95) (2.43) (2.18) $$\text{Bound}_{it}$$ 6.99*** 8.59*** 7.28*** 9.20*** (1.63) (1.74) (1.70) (1.70) Treasury repo $$\text{Unbound}_{it}$$ –3.99*** –5.01*** –6.39*** –7.00*** (1.40) (1.62) (1.52) (1.73) $$\text{Bound}_{it}$$ –3.92*** –5.48*** –5.34*** –6.58*** (1.26) (1.52) (1.34) (1.46) Agency repo $$\text{Unbound}_{it}$$ –2.87* –3.09** –2.98** –3.13** (1.47) (1.54) (1.48) (1.54) $$\text{Bound}_{it}$$ –0.31 –0.65 –0.59 –0.88 (1.94) (2.10) (1.71) (1.79) CD $$\text{Unbound}_{it}$$ –2.29*** –2.87*** –1.00 –1.46 (0.88) (0.68) (1.27) (1.20) $$\text{Bound}_{it}$$ –2.38*** –3.27*** –1.62 –2.56* (0.79) (0.78) (1.19) (1.44) Treasury debt $$\text{Unbound}_{it}$$ –0.32 –0.47 0.20 0.29 (1.62) (1.67) (1.93) (2.06) $$\text{Bound}_{it}$$ 1.42 1.18 1.53 1.71 (1.41) (1.45) (1.48) (1.60) Financial CP $$\text{Unbound}_{it}$$ 0.74 0.68 0.40 0.26 (0.91) (0.99) (0.92) (0.92) $$\text{Bound}_{it}$$ –0.11 –0.20 –0.30 –0.59 (0.74) (0.93) (0.95) (0.91) Asset-backed CP $$\text{Unbound}_{it}$$ 0.25 0.28 0.20 0.24 (0.49) (0.46) (0.57) (0.54) $$\text{Bound}_{it}$$ 0.39 0.43 0.33 0.40 (0.45) (0.45) (0.48) (0.45) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ N 1,111 1,111 1,111 1,111 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF asset allocation. $$Unbound_{it}$$ takes a value of one for treated funds that initially experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$Bound_{it}$$ takes a value of one for treated funds that initially experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. Column 1 contains no other controls, and Columns 2 and 4 include AUM as a control for fund size. Columns 3 and 4 control for fund-specific trends. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 3 Regression results: MMF asset substitution Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 9.01*** 10.05*** 9.14*** 10.10*** (1.86) (1.95) (2.43) (2.18) $$\text{Bound}_{it}$$ 6.99*** 8.59*** 7.28*** 9.20*** (1.63) (1.74) (1.70) (1.70) Treasury repo $$\text{Unbound}_{it}$$ –3.99*** –5.01*** –6.39*** –7.00*** (1.40) (1.62) (1.52) (1.73) $$\text{Bound}_{it}$$ –3.92*** –5.48*** –5.34*** –6.58*** (1.26) (1.52) (1.34) (1.46) Agency repo $$\text{Unbound}_{it}$$ –2.87* –3.09** –2.98** –3.13** (1.47) (1.54) (1.48) (1.54) $$\text{Bound}_{it}$$ –0.31 –0.65 –0.59 –0.88 (1.94) (2.10) (1.71) (1.79) CD $$\text{Unbound}_{it}$$ –2.29*** –2.87*** –1.00 –1.46 (0.88) (0.68) (1.27) (1.20) $$\text{Bound}_{it}$$ –2.38*** –3.27*** –1.62 –2.56* (0.79) (0.78) (1.19) (1.44) Treasury debt $$\text{Unbound}_{it}$$ –0.32 –0.47 0.20 0.29 (1.62) (1.67) (1.93) (2.06) $$\text{Bound}_{it}$$ 1.42 1.18 1.53 1.71 (1.41) (1.45) (1.48) (1.60) Financial CP $$\text{Unbound}_{it}$$ 0.74 0.68 0.40 0.26 (0.91) (0.99) (0.92) (0.92) $$\text{Bound}_{it}$$ –0.11 –0.20 –0.30 –0.59 (0.74) (0.93) (0.95) (0.91) Asset-backed CP $$\text{Unbound}_{it}$$ 0.25 0.28 0.20 0.24 (0.49) (0.46) (0.57) (0.54) $$\text{Bound}_{it}$$ 0.39 0.43 0.33 0.40 (0.45) (0.45) (0.48) (0.45) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ N 1,111 1,111 1,111 1,111 Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 9.01*** 10.05*** 9.14*** 10.10*** (1.86) (1.95) (2.43) (2.18) $$\text{Bound}_{it}$$ 6.99*** 8.59*** 7.28*** 9.20*** (1.63) (1.74) (1.70) (1.70) Treasury repo $$\text{Unbound}_{it}$$ –3.99*** –5.01*** –6.39*** –7.00*** (1.40) (1.62) (1.52) (1.73) $$\text{Bound}_{it}$$ –3.92*** –5.48*** –5.34*** –6.58*** (1.26) (1.52) (1.34) (1.46) Agency repo $$\text{Unbound}_{it}$$ –2.87* –3.09** –2.98** –3.13** (1.47) (1.54) (1.48) (1.54) $$\text{Bound}_{it}$$ –0.31 –0.65 –0.59 –0.88 (1.94) (2.10) (1.71) (1.79) CD $$\text{Unbound}_{it}$$ –2.29*** –2.87*** –1.00 –1.46 (0.88) (0.68) (1.27) (1.20) $$\text{Bound}_{it}$$ –2.38*** –3.27*** –1.62 –2.56* (0.79) (0.78) (1.19) (1.44) Treasury debt $$\text{Unbound}_{it}$$ –0.32 –0.47 0.20 0.29 (1.62) (1.67) (1.93) (2.06) $$\text{Bound}_{it}$$ 1.42 1.18 1.53 1.71 (1.41) (1.45) (1.48) (1.60) Financial CP $$\text{Unbound}_{it}$$ 0.74 0.68 0.40 0.26 (0.91) (0.99) (0.92) (0.92) $$\text{Bound}_{it}$$ –0.11 –0.20 –0.30 –0.59 (0.74) (0.93) (0.95) (0.91) Asset-backed CP $$\text{Unbound}_{it}$$ 0.25 0.28 0.20 0.24 (0.49) (0.46) (0.57) (0.54) $$\text{Bound}_{it}$$ 0.39 0.43 0.33 0.40 (0.45) (0.45) (0.48) (0.45) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ N 1,111 1,111 1,111 1,111 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF asset allocation. $$Unbound_{it}$$ takes a value of one for treated funds that initially experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$Bound_{it}$$ takes a value of one for treated funds that initially experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. Column 1 contains no other controls, and Columns 2 and 4 include AUM as a control for fund size. Columns 3 and 4 control for fund-specific trends. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. For the next outcome variable—the share of assets invested in private Treasury-backed repo—we find that MMFs substitute out of private repo transactions in order to invest with the Fed. This asset category is likely the closest substitute for the ONRRP facility, as ONRRP transactions are also repo backed by Treasury collateral. The crowding out of private Treasury repo for both treatment groups confirms the substitutability of these transactions for ONRRP investment. Agency repo investments also fall for MMFs in our main treatment subgroup, but appear insensitive to cap increases when funds remain bound by the new counterparty caps. This suggests that MMFs only substitute away from agency-backed repo after they have first substituted away from other assets. In other words, a fund that remains constrained by a larger counterparty cap will first shift out of Treasury repo, but as the cap is increased enough for the fund to find itself unconstrained, it will eventually begin to shift investment away from agency repo as well. Deposits (CD) show a similar pattern to Treasury repo. The results are evident for both treatment subgroups, suggesting that money funds are willing to move funds out of short-term deposits in order to invest in a reverse repo with the Fed. Summing the coefficient estimates over these four asset categories reveals that the increase in the ONRRP facility for both treatment subgroups is entirely offset by reductions in Treasury repo, agency repo, and deposits. Dividing the point estimates of the effects on these asset classes by the estimated increase in the ONRRP investment yields rates of substitution to the ONRRP from other assets. For example, using the point estimates for the $$\text{UnboundTreatment}_{it}$$ group in our baseline specification reported in Column 1, we find that an increase in ONRRP investment of 1% of AUM corresponds to reductions in Treasury repo, agency repo, and deposits of about 0.45%, 0.30%, and 0.25% of AUM, respectively. Naively applying these figures to the $\$$ 167 billion MMFs invested in Fed RRP on December 31, 2014, implies that private tri-party Treasury collateral repo volume was lower by $\$$75 billion (roughly 11% of the traded volume on that day), and agency collateral repo was lower by $\$$50 billion (or 9% of the traded volume). Reductions in deposits driven by the ONRRP investment represent a far smaller share of the overall deposit market. The remaining asset classes show no statistically significant response to increases in the maximum bid amount, with point estimates that generally lie close to zero. Very similar results are presented in Columns 2 through 4—which add controls for fund size and/or fund-level trends—although the deposits result weakens somewhat in certain specifications. In unreported results, we find that using a weighted least squares estimator yields identical conclusions. Overall, these results show that investment in the Fed’s ONRRP facility comes at the expense of private investments in Treasury repo, agency repo, and deposits. However, the overall effect on financial stability is unclear. On one hand, the funding that is withdrawn from dealers and banks in favor of the ONRRP could potentially be worrisome to a central bank that seeks to minimize its footprint in financial markets. Moreover, the existence of the Fed’s ONRRP facility may increase the likelihood of a shift in risk sentiment spurring a flight-to-quality, with MMFs flocking to the Fed’s ONRRP facility (Frost et al. 2015).17 In such an event, substitution out of other asset classes such as commercial paper could emerge, starving regular funding from an additional class of borrowers with potentially destabilizing effects on financial markets (Kacperczyk and Schnabl 2013). On the other hand, crowding out privately issued repo may ameliorate overall financial stability by reducing the externalities associated with a large private repo market that is prone to runs (Carlson et al. 2016; Frost et al. 2015). As we will later show, however, dealers do not change their total repo borrowing in response to the reduction in MMF lending that we demonstrate here. To further support the causal interpretation of the estimated effects reported above, we conduct two separate placebo tests. Our first test includes a placebo treatment dummy that takes a value of one for treatment group funds in the quarter immediately after they become unconstrained. In the context of the stylized example presented in Figure 4, this dummy would be zero for fund 2 until the third period. For fund 3, this placebo treatment would be zero until period 4 (not shown). Since these funds were unconstrained by the counterparty cap in the period prior to the placebo treatment, we should expect a null result. However, if previously constrained funds were merely in the process of shifting their asset allocation for a reason unrelated to the increases in the ONRRP caps, we would expect to see statistically significant effects similar to our estimated treatment effects. In panel A of Table 4, we report the results of this placebo test for the asset categories for which we found evidence of substitution. Comparing the point estimates for the $$\text{Placebo}_{it}$$ treatment group with the $$\text{Unbound}_{it}$$ treatment group, we can see that funds that become unconstrained do not differentially increase their ONRRP participation in the periods after the cap no longer binds the funds’ investment decision. In other words, the measured effect for the $$\text{Unbound}_{it}$$ treatment group coincides entirely with the increase in the counterparty cap. Therefore, it does not appear that our results are driven by other factors, such as an uneven withdrawal of dealer borrowing that disproportionately affected the funds in our treatment group. Table 4 Regression results: Placebo tests of MMF asset substitution A. Placebo treatment within sample Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 8.40*** 9.52*** 9.56*** 10.43*** (1.78) (1.66) (2.53) (2.41) $$\text{Bound}_{it}$$ 6.82*** 8.43*** 7.79*** 9.60*** (1.50) (1.51) (2.11) (2.25) $$\text{Placebo}_{it}$$ 0.96 0.81 1.53 1.29 (1.89) (1.88) (2.57) (2.55) Treasury repo $$\text{Unbound}_{it}$$ –4.20*** –5.32*** –6.58*** –7.15*** (1.09) (1.40) (1.67) (1.83) $$\text{Bound}_{it}$$ –3.98*** –5.58*** –5.57*** –6.75*** (1.19) (1.46) (1.51) (1.58) $$\text{Placebo}_{it}$$ 0.33 0.48 –0.71 –0.55 (1.15) (1.13) (1.48) (1.49) Agency repo $$\text{Unbound}_{it}$$ –2.71* –2.95* –3.22** –3.34** (1.39) (1.47) (1.54) (1.60) $$\text{Bound}_{it}$$ –0.26 –0.61 –0.88 –1.15 (1.91) (2.08) (1.71) (1.79) $$\text{Placebo}_{it}$$ –0.26 –0.23 –0.89 –0.85 (0.57) (0.56) (0.81) (0.80) CD $$\text{Unbound}_{it}$$ –1.81** –2.44*** –0.95 –1.39 (0.86) (0.66) (1.31) (1.21) $$\text{Bound}_{it}$$ –2.24** –3.15*** –1.56 –2.47* (0.85) (0.84) (1.11) (1.35) $$\text{Placebo}_{it}$$ –0.75 –0.66 0.17 0.29 (0.63) (0.63) (0.56) (0.57) B. Placebo treatment in pre-ONRRP sample Treasury repo $$\text{Placebo}_{it}$$ 0.27 0.05 1.86 1.64 (1.89) (2.18) (1.82) (1.91) Agency repo $$\text{Placebo}_{it}$$ 0.16 –0.03 1.57 1.33 (0.81) (0.95) (1.88) (2.04) CD $$\text{Placebo}_{it}$$ –0.20 0.54 1.43 2.81 (1.49) (1.44) (1.66) (1.70) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ A. Placebo treatment within sample Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 8.40*** 9.52*** 9.56*** 10.43*** (1.78) (1.66) (2.53) (2.41) $$\text{Bound}_{it}$$ 6.82*** 8.43*** 7.79*** 9.60*** (1.50) (1.51) (2.11) (2.25) $$\text{Placebo}_{it}$$ 0.96 0.81 1.53 1.29 (1.89) (1.88) (2.57) (2.55) Treasury repo $$\text{Unbound}_{it}$$ –4.20*** –5.32*** –6.58*** –7.15*** (1.09) (1.40) (1.67) (1.83) $$\text{Bound}_{it}$$ –3.98*** –5.58*** –5.57*** –6.75*** (1.19) (1.46) (1.51) (1.58) $$\text{Placebo}_{it}$$ 0.33 0.48 –0.71 –0.55 (1.15) (1.13) (1.48) (1.49) Agency repo $$\text{Unbound}_{it}$$ –2.71* –2.95* –3.22** –3.34** (1.39) (1.47) (1.54) (1.60) $$\text{Bound}_{it}$$ –0.26 –0.61 –0.88 –1.15 (1.91) (2.08) (1.71) (1.79) $$\text{Placebo}_{it}$$ –0.26 –0.23 –0.89 –0.85 (0.57) (0.56) (0.81) (0.80) CD $$\text{Unbound}_{it}$$ –1.81** –2.44*** –0.95 –1.39 (0.86) (0.66) (1.31) (1.21) $$\text{Bound}_{it}$$ –2.24** –3.15*** –1.56 –2.47* (0.85) (0.84) (1.11) (1.35) $$\text{Placebo}_{it}$$ –0.75 –0.66 0.17 0.29 (0.63) (0.63) (0.56) (0.57) B. Placebo treatment in pre-ONRRP sample Treasury repo $$\text{Placebo}_{it}$$ 0.27 0.05 1.86 1.64 (1.89) (2.18) (1.82) (1.91) Agency repo $$\text{Placebo}_{it}$$ 0.16 –0.03 1.57 1.33 (0.81) (0.95) (1.88) (2.04) CD $$\text{Placebo}_{it}$$ –0.20 0.54 1.43 2.81 (1.49) (1.44) (1.66) (1.70) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF asset allocation, as described in Table 3. In the table above, panel A uses the same sample as Table 3, but includes a placebo treatment variable, $$\text{Placebo}_{it}$$, which takes a value of one for treated funds beginning in the period after these funds become unconstrained. Panel B uses a sample of the same funds for the 11 quarters prior to the introduction of the ONRRP. In panel B, $$\text{Placebo}_{it}$$ takes a value of one for treated funds beginning in December 2011, five quarters after the beginning of the sample. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 4 Regression results: Placebo tests of MMF asset substitution A. Placebo treatment within sample Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 8.40*** 9.52*** 9.56*** 10.43*** (1.78) (1.66) (2.53) (2.41) $$\text{Bound}_{it}$$ 6.82*** 8.43*** 7.79*** 9.60*** (1.50) (1.51) (2.11) (2.25) $$\text{Placebo}_{it}$$ 0.96 0.81 1.53 1.29 (1.89) (1.88) (2.57) (2.55) Treasury repo $$\text{Unbound}_{it}$$ –4.20*** –5.32*** –6.58*** –7.15*** (1.09) (1.40) (1.67) (1.83) $$\text{Bound}_{it}$$ –3.98*** –5.58*** –5.57*** –6.75*** (1.19) (1.46) (1.51) (1.58) $$\text{Placebo}_{it}$$ 0.33 0.48 –0.71 –0.55 (1.15) (1.13) (1.48) (1.49) Agency repo $$\text{Unbound}_{it}$$ –2.71* –2.95* –3.22** –3.34** (1.39) (1.47) (1.54) (1.60) $$\text{Bound}_{it}$$ –0.26 –0.61 –0.88 –1.15 (1.91) (2.08) (1.71) (1.79) $$\text{Placebo}_{it}$$ –0.26 –0.23 –0.89 –0.85 (0.57) (0.56) (0.81) (0.80) CD $$\text{Unbound}_{it}$$ –1.81** –2.44*** –0.95 –1.39 (0.86) (0.66) (1.31) (1.21) $$\text{Bound}_{it}$$ –2.24** –3.15*** –1.56 –2.47* (0.85) (0.84) (1.11) (1.35) $$\text{Placebo}_{it}$$ –0.75 –0.66 0.17 0.29 (0.63) (0.63) (0.56) (0.57) B. Placebo treatment in pre-ONRRP sample Treasury repo $$\text{Placebo}_{it}$$ 0.27 0.05 1.86 1.64 (1.89) (2.18) (1.82) (1.91) Agency repo $$\text{Placebo}_{it}$$ 0.16 –0.03 1.57 1.33 (0.81) (0.95) (1.88) (2.04) CD $$\text{Placebo}_{it}$$ –0.20 0.54 1.43 2.81 (1.49) (1.44) (1.66) (1.70) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ A. Placebo treatment within sample Dependent variable Treatment group (1) (2) (3) (4) Fed ONRRP $$\text{Unbound}_{it}$$ 8.40*** 9.52*** 9.56*** 10.43*** (1.78) (1.66) (2.53) (2.41) $$\text{Bound}_{it}$$ 6.82*** 8.43*** 7.79*** 9.60*** (1.50) (1.51) (2.11) (2.25) $$\text{Placebo}_{it}$$ 0.96 0.81 1.53 1.29 (1.89) (1.88) (2.57) (2.55) Treasury repo $$\text{Unbound}_{it}$$ –4.20*** –5.32*** –6.58*** –7.15*** (1.09) (1.40) (1.67) (1.83) $$\text{Bound}_{it}$$ –3.98*** –5.58*** –5.57*** –6.75*** (1.19) (1.46) (1.51) (1.58) $$\text{Placebo}_{it}$$ 0.33 0.48 –0.71 –0.55 (1.15) (1.13) (1.48) (1.49) Agency repo $$\text{Unbound}_{it}$$ –2.71* –2.95* –3.22** –3.34** (1.39) (1.47) (1.54) (1.60) $$\text{Bound}_{it}$$ –0.26 –0.61 –0.88 –1.15 (1.91) (2.08) (1.71) (1.79) $$\text{Placebo}_{it}$$ –0.26 –0.23 –0.89 –0.85 (0.57) (0.56) (0.81) (0.80) CD $$\text{Unbound}_{it}$$ –1.81** –2.44*** –0.95 –1.39 (0.86) (0.66) (1.31) (1.21) $$\text{Bound}_{it}$$ –2.24** –3.15*** –1.56 –2.47* (0.85) (0.84) (1.11) (1.35) $$\text{Placebo}_{it}$$ –0.75 –0.66 0.17 0.29 (0.63) (0.63) (0.56) (0.57) B. Placebo treatment in pre-ONRRP sample Treasury repo $$\text{Placebo}_{it}$$ 0.27 0.05 1.86 1.64 (1.89) (2.18) (1.82) (1.91) Agency repo $$\text{Placebo}_{it}$$ 0.16 –0.03 1.57 1.33 (0.81) (0.95) (1.88) (2.04) CD $$\text{Placebo}_{it}$$ –0.20 0.54 1.43 2.81 (1.49) (1.44) (1.66) (1.70) Initial AUM*Post — ✓ — ✓ Fund trends — — ✓ ✓ This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF asset allocation, as described in Table 3. In the table above, panel A uses the same sample as Table 3, but includes a placebo treatment variable, $$\text{Placebo}_{it}$$, which takes a value of one for treated funds beginning in the period after these funds become unconstrained. Panel B uses a sample of the same funds for the 11 quarters prior to the introduction of the ONRRP. In panel B, $$\text{Placebo}_{it}$$ takes a value of one for treated funds beginning in December 2011, five quarters after the beginning of the sample. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Panel B of Table 4 presents the results of a second placebo test. In this test, we use an 11 quarter sample period (the same length as that used in our main analysis in Table 3) that ends in June 2013, the quarter immediately prior to the introduction of the ONRRP. In this exercise, our placebo treatment dummy takes a value of one for the treated funds in our sample as of December 2011.18 As demonstrated in panel B, there is no evidence of differential investment in Treasury repo, agency repo, or deposits. Thus, these results confirm that there are no systematic differences in investment patterns between the treatment and control MMFs that are driven by causes other the changing availability of the ONRRP. Moreover, because this sample period spans the period before the introduction of the ONRRP, these results offer further evidence of parallel trends in the asset shares between the two groups of funds. In the remainder of the paper, we solely focus on the effects of the Fed’s ONRRP facility on the repo market. 4.1.1 Robustness using trade-level data In this subsection, we establish the robustness of the substitution effects of the ONRRP outlined above for both Treasury and agency-backed repo. It may be the case that the substitution results achieved in section 4.1 are merely driven by differences in demand for repo funding by dealers that borrowed from MMFs in our treatment group. For example, treated funds may trade with dealers that are winding down their repo positions more rapidly than the control funds’ dealers. Similarly, treated funds may also trade more heavily with certain foreign dealers, who, as discussed above, withdraw more heavily from the repo market on quarter-ends.19 Therefore, treated funds may be forced to invest in the ONRRP more than their control fund counterparts in response to differences in borrower demand, which could bias our estimated treatment effects. However, it is important to note that treated and control funds trading with different sets of dealers is not necessarily sufficient for our results to be biased. Rather, treated funds’ dealers would also need to disproportionately pull back from the repo market after the introduction of the ONRRP and in concert with the timing of the increases in the ONRRP cap. To address these concerns, we construct a panel of borrower-lender pairs using the data available in the N-MFP reports. By identifying the dealer counterparty to each repo transaction, we are able to then control for the borrower-quarter specific demand for repo (Khwaja and Mian 2008). Thus, estimating variants of the following regression specification on quarter-ends allows us to address the potential threats to the causal estimates that stem from the behavior of repo borrowers. \begin{align} y_{ijt} &= \delta\cdot(\text{Post}_t\cdot\text{UnboundTreatment}_{it}) + \rho\cdot(\text{Post}_t\cdot\text{BoundTreatment}_{it}) \notag \\ & \quad +\lambda_{ij}\cdot(\text{MMF-Dealer}_{ij}\cdot t) + \phi_{jt}\cdot(\text{Dealer}_{j}\cdot\text{Quarter}_{t}) \notag\\ &\quad + \beta_{ij}\cdot\text{MMF-Dealer}_{ij} + \varepsilon_{ijt} \end{align} (3) In Equation (3), $$y_{ijt}$$ represents the asset share of the day t repo activity between MMF i and dealer j. Here, we are now able to control for trends in activity between individual trading partners $$(\text{MMF-Dealer}_{ij}\cdot t)$$, as well as changes in dealer-specific borrowing demand in each quarter $$(\text{Dealer}_{j}\cdot\text{Quarter}_{t})$$. Consequently, we are able to measure the difference in private repo lending after a cap increase between a control fund and a treated fund for a given date, holding the borrower constant. As shown in Table 5, our conclusions remain unchanged even when accounting for the extent to which dealer behavior can explain the results. Comparing the sign and significance for both Treasury repo (panel A) and agency repo (panel B) to the results in Table 3, we see the findings are nearly identical. Of course, the magnitude of the coefficients is somewhat lower, as we are now estimating the effects of treatment on relationship-level asset substitution. These results strongly support our earlier interpretation that MMFs in fact substitute out of private repo in favor of the ONRRP. Table 5 Regression results: MMF asset substitution at the relationship level A. Treasury repo (1) (2) (3) (4) $$\text{Unbound}_{it}$$ –0.39*** –0.39*** –0.39** –0.36** (0.13) (0.15) (0.16) (0.18) $$\text{Bound}_{it}$$ –0.32** –0.31** –0.33*** –0.31*** (0.12) (0.12) (0.12) (0.11) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 8,360 8,360 8,360 8,360 Adj. R$$^2$$ 0.45 0.48 0.60 0.61 B. Agency repo $$\text{Unbound}_{it}$$ –0.29** –0.33** –0.38*** –0.40*** (0.12) (0.15) (0.14) (0.14) $$\text{Bound}_{it}$$ 0.00 –0.01 –0.07 –0.07 (0.17) (0.21) (0.14) (0.16) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 9,251 9,251 9,251 9,251 Adj. R$$^2$$ 0.49 0.51 0.66 0.67 A. Treasury repo (1) (2) (3) (4) $$\text{Unbound}_{it}$$ –0.39*** –0.39*** –0.39** –0.36** (0.13) (0.15) (0.16) (0.18) $$\text{Bound}_{it}$$ –0.32** –0.31** –0.33*** –0.31*** (0.12) (0.12) (0.12) (0.11) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 8,360 8,360 8,360 8,360 Adj. R$$^2$$ 0.45 0.48 0.60 0.61 B. Agency repo $$\text{Unbound}_{it}$$ –0.29** –0.33** –0.38*** –0.40*** (0.12) (0.15) (0.14) (0.14) $$\text{Bound}_{it}$$ 0.00 –0.01 –0.07 –0.07 (0.17) (0.21) (0.14) (0.16) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 9,251 9,251 9,251 9,251 Adj. R$$^2$$ 0.49 0.51 0.66 0.67 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF repo trades with dealers. $$\textit{Unbound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$\textit{Bound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. All specifications include relationship fixed effects. Columns 1 and 3 contain quarter fixed effects, and Columns 2 and 4 include dealer*time fixed effects. Columns 3 and 4 additionally control for trends specific to each unique dealer-fund pair. Standard errors are clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 5 Regression results: MMF asset substitution at the relationship level A. Treasury repo (1) (2) (3) (4) $$\text{Unbound}_{it}$$ –0.39*** –0.39*** –0.39** –0.36** (0.13) (0.15) (0.16) (0.18) $$\text{Bound}_{it}$$ –0.32** –0.31** –0.33*** –0.31*** (0.12) (0.12) (0.12) (0.11) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 8,360 8,360 8,360 8,360 Adj. R$$^2$$ 0.45 0.48 0.60 0.61 B. Agency repo $$\text{Unbound}_{it}$$ –0.29** –0.33** –0.38*** –0.40*** (0.12) (0.15) (0.14) (0.14) $$\text{Bound}_{it}$$ 0.00 –0.01 –0.07 –0.07 (0.17) (0.21) (0.14) (0.16) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 9,251 9,251 9,251 9,251 Adj. R$$^2$$ 0.49 0.51 0.66 0.67 A. Treasury repo (1) (2) (3) (4) $$\text{Unbound}_{it}$$ –0.39*** –0.39*** –0.39** –0.36** (0.13) (0.15) (0.16) (0.18) $$\text{Bound}_{it}$$ –0.32** –0.31** –0.33*** –0.31*** (0.12) (0.12) (0.12) (0.11) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 8,360 8,360 8,360 8,360 Adj. R$$^2$$ 0.45 0.48 0.60 0.61 B. Agency repo $$\text{Unbound}_{it}$$ –0.29** –0.33** –0.38*** –0.40*** (0.12) (0.15) (0.14) (0.14) $$\text{Bound}_{it}$$ 0.00 –0.01 –0.07 –0.07 (0.17) (0.21) (0.14) (0.16) Dealer*Time FEs — ✓ — ✓ Dealer-fund trends — — ✓ ✓ Observations 9,251 9,251 9,251 9,251 Adj. R$$^2$$ 0.49 0.51 0.66 0.67 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF repo trades with dealers. $$\textit{Unbound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$\textit{Bound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. All specifications include relationship fixed effects. Columns 1 and 3 contain quarter fixed effects, and Columns 2 and 4 include dealer*time fixed effects. Columns 3 and 4 additionally control for trends specific to each unique dealer-fund pair. Standard errors are clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. 4.2 Effects of the ONRRP on trading relationships Next, we consider how the changes in MMF asset allocation seen in the previous section affect funds’ relationships with dealers in the tri-party repo market. We are primarily interested in whether MMFs transition to smaller trading networks as a result of the ONRRP, possibly increasing financial fragility, and whether MMFs’ differentially substitute away from certain dealers (e.g., those that are riskier counterparties). We again look at quarter-end MMF repo activity from the N-MFP data between December 2012 and June 2015. In particular, we focus on four measures of MMF lending relationships: number of dealers, weighted CDS spread, Herfindahl index (HHI), and share of lending to the largest dealer. Number of dealers is simply the total number of dealers that an MMF trades with on a given quarter-end date. As funds substitute into the ONRRP and away from private repo, they may do so either by dropping certain counterparties or by simply reducing their volume with one or several of their existing counterparties. Weighted CDS spread measures the volume-weighted CDS spread of all J dealers with which fund i trades.20 It is defined as \begin{equation} \text{Weighted CDS Spread}_{it} = \frac{\sum\limits_{j=1}^{J}\text{CDS Spread}_{jt}\cdot\text{Repo Volume}_{ijt}}{\text{Repo Volume}_{it}}. \end{equation} (4) Funds that invest more with the Fed through the ONRRP may favor some dealer counterparties over others and prefer to reduce trading activity with less-favored counterparties. In particular, MMFs may substitute away from low-risk dealers since they are the most direct substitute for the Fed, which is a risk-free counterparty. Conversely, they could substitute away from their higher-risk dealers in an effort to reduce their overall risk profile, although, as documented by Krishnamurthy, Nagel, and Orlov (2014), repo funding with Treasury and agency collateral appears less sensitive to dealers’ perceived default risk. HHI measures concentration among a funds’ borrowers excluding the Fed and is defined as \begin{equation} \text{HHI}_{it} = \sum\limits_{j=1}^{J}\bigg[\frac{\text{Repo Volume}_{ijt}}{\text{Repo Volume}_{it}}\bigg]^{2}. \end{equation} (5) HHI ranges from 1/J to 1, with lower values representing less concentration among dealers whereas higher values correspond to more highly concentrated lending. Lastly, we examine the share of lending to a fund’s largest dealer, calculated simply as the percentage of each MMF’s total repo volume that is transacted with the dealer to which it lends most heavily. These last two measures are used to assess whether MMFs differentially substitute away from dealers based on the volume traded between the pair. MMFs may be less likely to substitute away from those dealers with which they have a strong relationship, as measured by volume. For each of the four measures of dealer relationships, we consider MMFs’ Treasury- and agency-collateral repo activity separately. There may be stronger results for Treasury repo since, as seen in the previous section, the substitution effects are stronger for Treasury repo. For each of our measures of counterparty relationships, we present evidence of parallel trends in Figure 7, and proceed by estimating the following regression: Figure 7 View largeDownload slide Trends in measures of counterparty relationships This figure presents the treatment effect for measures of MMF counterparty relationships in the six quarters prior to implementation of the ONRRP. Thus, a positive value corresponds to quarters in which treated funds witness counterparty relationship measures rising more (or falling less) than the measure for control funds. A negative value corresponds to quarters in which treated funds witness counterparty relationship measures rising less (or falling more) than that of control funds. Figure 7 View largeDownload slide Trends in measures of counterparty relationships This figure presents the treatment effect for measures of MMF counterparty relationships in the six quarters prior to implementation of the ONRRP. Thus, a positive value corresponds to quarters in which treated funds witness counterparty relationship measures rising more (or falling less) than the measure for control funds. A negative value corresponds to quarters in which treated funds witness counterparty relationship measures rising less (or falling more) than that of control funds. \begin{align} y_{it} &= \delta\cdot(\text{Post}_t\cdot\text{UnboundTreatment}_{it}) + \rho\cdot(\text{Post}_t\cdot\text{BoundTreatment}_{it}) \notag\\ &\quad + \lambda_{i}\cdot(\text{MMF}_{i}\cdot t) + \beta_{i}\cdot\text{MMF}_{i} + \gamma_{t}\cdot\text{Quarter}_{t} + \varepsilon_{it}, \end{align} (6) \begin{align*} \text{where}\hspace{15mm} y_{it} &\in \left\{\text{Number of Dealers}_{it},\text{Weighted CDS Spread}_{it},\right.\\ &\qquad\left.\text{HHI}_{it},\text{Share Largest Dealer}_{it}\right\}. \end{align*} The results are shown in Table 6. There are no statistically significant differences in either the number of dealers or the weighted CDS spread. This suggests that MMFs continue trading with all of their original counterparties and do not differentially substitute away from or toward riskier dealers, consistent with the patterns shown in Krishnamurthy, Nagel, and Orlov (2014). In unreported results, we confirm that MMFs are not likely to sever dealer relationships by estimating the probability of a trade between the unique dealer-fund trading pairs in our sample. We find that, after a counterparty cap increase, initially bound treatment funds are no less likely to conduct a trade with one of their trading partners than funds that were never constrained by the cap. Table 6 Regression results: MMF-dealer relationships A. No fund trends Number of dealers Weighted CDS spread HHI Share largest dealer (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ –0.07 –0.91 0.10 0.01 –0.04 –0.01 –0.02 –0.002 (0.43) (0.84) (0.08) (0.06) (0.05) (0.05) (0.05) (0.05) $$\text{Bound}_{it}$$ 0.31 0.61 0.13 –0.005 –0.08 –0.12* –0.06 –0.11* (0.50) (0.68) (0.08) (0.06) (0.10) (0.06) (0.09) (0.06) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.94 0.93 0.90 0.94 0.89 0.87 0.92 0.91 B: With fund trends (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ 0.25 –0.34 0.04 0.05 –0.16* –0.09 –0.12 –0.07 (0.69) (0.68) (0.10) (0.11) (0.08) (0.07) (0.07) (0.06) $$\text{Bound}_{it}$$ 0.45 0.84 0.08 0.01 –0.13 –0.16** –0.10 –0.14** (0.58) (0.82) (0.08) (0.07) (0.10) (0.07) (0.09) (0.07) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.95 0.95 0.92 0.96 0.91 0.90 0.94 0.93 A. No fund trends Number of dealers Weighted CDS spread HHI Share largest dealer (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ –0.07 –0.91 0.10 0.01 –0.04 –0.01 –0.02 –0.002 (0.43) (0.84) (0.08) (0.06) (0.05) (0.05) (0.05) (0.05) $$\text{Bound}_{it}$$ 0.31 0.61 0.13 –0.005 –0.08 –0.12* –0.06 –0.11* (0.50) (0.68) (0.08) (0.06) (0.10) (0.06) (0.09) (0.06) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.94 0.93 0.90 0.94 0.89 0.87 0.92 0.91 B: With fund trends (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ 0.25 –0.34 0.04 0.05 –0.16* –0.09 –0.12 –0.07 (0.69) (0.68) (0.10) (0.11) (0.08) (0.07) (0.07) (0.06) $$\text{Bound}_{it}$$ 0.45 0.84 0.08 0.01 –0.13 –0.16** –0.10 –0.14** (0.58) (0.82) (0.08) (0.07) (0.10) (0.07) (0.09) (0.07) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.95 0.95 0.92 0.96 0.91 0.90 0.94 0.93 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF relationships in the repo market. $$\textit{Unbound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$\textit{Bound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. Dependent variables include the number of dealers per fund, the weighted CDS spread of dealers that trade with each fund, the Herfindahl-Hirschman index (HHI) for each fund (calculated using volume transacted with each dealer) and the percentage of volume traded with the fund’s largest dealer. For each dependent variable, Column 1 includes Treasury collateral repo transactions, and Column 2 includes agency collateral repo. Panel A excludes fund-level trends from the specification, whereas the specifications in panel B include fund trends. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 6 Regression results: MMF-dealer relationships A. No fund trends Number of dealers Weighted CDS spread HHI Share largest dealer (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ –0.07 –0.91 0.10 0.01 –0.04 –0.01 –0.02 –0.002 (0.43) (0.84) (0.08) (0.06) (0.05) (0.05) (0.05) (0.05) $$\text{Bound}_{it}$$ 0.31 0.61 0.13 –0.005 –0.08 –0.12* –0.06 –0.11* (0.50) (0.68) (0.08) (0.06) (0.10) (0.06) (0.09) (0.06) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.94 0.93 0.90 0.94 0.89 0.87 0.92 0.91 B: With fund trends (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ 0.25 –0.34 0.04 0.05 –0.16* –0.09 –0.12 –0.07 (0.69) (0.68) (0.10) (0.11) (0.08) (0.07) (0.07) (0.06) $$\text{Bound}_{it}$$ 0.45 0.84 0.08 0.01 –0.13 –0.16** –0.10 –0.14** (0.58) (0.82) (0.08) (0.07) (0.10) (0.07) (0.09) (0.07) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.95 0.95 0.92 0.96 0.91 0.90 0.94 0.93 A. No fund trends Number of dealers Weighted CDS spread HHI Share largest dealer (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ –0.07 –0.91 0.10 0.01 –0.04 –0.01 –0.02 –0.002 (0.43) (0.84) (0.08) (0.06) (0.05) (0.05) (0.05) (0.05) $$\text{Bound}_{it}$$ 0.31 0.61 0.13 –0.005 –0.08 –0.12* –0.06 –0.11* (0.50) (0.68) (0.08) (0.06) (0.10) (0.06) (0.09) (0.06) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.94 0.93 0.90 0.94 0.89 0.87 0.92 0.91 B: With fund trends (1) (2) (1) (2) (1) (2) (1) (2) $$\text{Unbound}_{it}$$ 0.25 –0.34 0.04 0.05 –0.16* –0.09 –0.12 –0.07 (0.69) (0.68) (0.10) (0.11) (0.08) (0.07) (0.07) (0.06) $$\text{Bound}_{it}$$ 0.45 0.84 0.08 0.01 –0.13 –0.16** –0.10 –0.14** (0.58) (0.82) (0.08) (0.07) (0.10) (0.07) (0.09) (0.07) N 748 789 625 730 748 789 748 789 Adj. R$$^2$$ 0.95 0.95 0.92 0.96 0.91 0.90 0.94 0.93 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF relationships in the repo market. $$\textit{Unbound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap, but, as a result of subsequent cap increases, no longer find the counterparty cap binding. Similarly, $$\textit{Bound}_{it}$$ takes a value of one for treated funds that previously experienced a binding counterparty cap and continue to face a binding cap despite subsequent increases. Dependent variables include the number of dealers per fund, the weighted CDS spread of dealers that trade with each fund, the Herfindahl-Hirschman index (HHI) for each fund (calculated using volume transacted with each dealer) and the percentage of volume traded with the fund’s largest dealer. For each dependent variable, Column 1 includes Treasury collateral repo transactions, and Column 2 includes agency collateral repo. Panel A excludes fund-level trends from the specification, whereas the specifications in panel B include fund trends. All specifications include fund and quarter fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. However, there is some statistical significance in the regressions for HHI and the share of the largest dealer, especially when fund trends are included in panel B, with consistently negative point estimates throughout. Thus, funds’ dealer-concentration HHIs fall as funds become less constrained by the ONRRP cap. In Section 4.1, we demonstrated that funds’ repo volume decreases in response to the ONRRP, while we show here that the number of repo borrowers does not change. Therefore, the HHI results suggest that MMFs appear to be substituting away from those counterparties with which they trade the most, while preserving dealer relationships even for those dealers that account for only a small share of total trading volume. This interpretation is consistent with the decrease in the share of lending to a fund’s largest counterparty shown in the rightmost columns of Table 6. Thus, MMF participation in the ONRRP evidently results in a more even distribution of MMF’s remaining private repo investment. In total, these results suggest that lending relationships are very important in the repo market and MMFs have a desire to maintain their relationships with dealers. In particular, funds do not drop any of their dealer counterparties and do not differentially substitute away from dealers according to their risk of default. Rather, MMFs adjust trading volume across their existing counterparties by substituting away from their largest counterparties to some degree. Importantly, the results of this analysis suggest that, despite the ONRRP disintermediating private repo volume, it does not significantly affect the relationship structure of ONRRP-eligible MMFs in the tri-party repo market. Of course, these results are local to substitution effects stemming from increases in the counterparty caps only, and relationships may in fact be severed by ONRRP participation in times of severe financial stress. On the other hand, the importance of trading relationships in repo markets that we identify here may contribute to the resiliency of the repo market as a whole if it reduces the willingness of lenders to completely divest from their regular counterparties. 4.3 Effects of the ONRRP on bargaining power in the repo market In our next exercise, we aim to identify the effects of the ONRRP on prevailing rates in the repo market. By providing repo lenders with a credible outside option, MMFs eligible for the ONRRP should be put in a more advantageous bargaining position vis-à-vis dealers. Of course, funds that are already investing in the ONRRP facility at the counterparty cap do not possess an option to invest a marginal dollar in the ONRRP facility at the expense of dealer-provided repo. Therefore, we again exploit the exogenous increases in counterparty caps, because these increases should bestow additional bargaining power on MMFs that were previously investing the maximum-allowable amount at the ONRRP. The data used for this analysis come from the confidential transaction-level tri-party data collected by the FRBNY discussed in Section 3. Repo activity is aggregated to unique MMF family-dealer-collateral triples. Since there may be many trades on a given day between MMFs and their dealers, we calculate both the weighted average rate between the MMF and dealer for a given collateral type, as well as the volume-weighted 25th and 75th percentile rates. To identify the effect of the potentially increased bargaining power MMFs command in the private repo market, we again appeal to a DD strategy. Here, though, we take advantage of the high frequency data and treat each cap increase as its own separate event, examining the days around each increase. Specifically, we identify fund families in which at least one MMF was bound by the counterparty cap on the day before a cap increase. The first four counterparty cap increases, which occurred on September 27, 2013, December 23, 2013, January 30, 2014, and March 5, 2014, witnessed at least one bound fund family on the day before the change.21,22 Multiple fund families contain MMFs that were constrained by the counterparty cap in all but the final date (March 5, 2014), when only a single fund family faced a binding constraint immediately prior to the cap increase. We then compare changes in rates for previously bound MMFs to previously unbound funds on the day of a counterparty cap increase by estimating the following regression: \begin{align} \text{RepoRate}_{ijct} &= \delta\cdot(\text{Post}_{t}\cdot\text{BoundTreatment}_{it-1})\notag \\ &\quad +\beta_{ijc}\cdot(\text{MMF Family-Dealer-Collateral}_{ijc}) + \gamma_{t}\cdot\text{Day}_{t} + \varepsilon_{ijct}. \end{align} (7) In Equation (7), $$\delta$$ is the DD estimate of the effect of a cap increase on previously bound (treatment group) funds’ weighted average repo rate with a given dealer for a given type of collateral. For simplicity, and because most funds did not face a binding counterparty cap on the day of the cap increases, we consider only a single treatment group that is composed of fund families that were bound the day prior to the increase.23 If MMFs’ bargaining power sufficiently increases upon counterparty cap increases, $$\delta$$ would be expected to be positive, indicating an increase in the rate that MMFs can command when presented with the outside option. Similarly, $$\delta$$ could also be positive if money funds simply shift lower-rate private transactions to the ONRRP facility, though the ability of dealers to pay a rate below the ONRRP indicates a differential in bargaining power that favors dealers. Figure 8 demonstrates parallel trends in the daily average repo rates earned by constrained and unconstrained fund families in the days leading up to the changes in the ONRRP maximum bid amount. There is no apparent divergence in advance of any of the cap changes in our sample. Figure 8 View largeDownload slide Trends in rates on private repo transactions This figure presents treatment effects for the weighted-average rates in the days preceding each counterparty cap change. Thus, a positive value corresponds to days in which treated funds witness average rates rising more (or falling less) than the rate for control funds. A negative value corresponds to days in which treated funds witness average rates rising less (or falling more) than rates for control funds. Figure 8 View largeDownload slide Trends in rates on private repo transactions This figure presents treatment effects for the weighted-average rates in the days preceding each counterparty cap change. Thus, a positive value corresponds to days in which treated funds witness average rates rising more (or falling less) than the rate for control funds. A negative value corresponds to days in which treated funds witness average rates rising less (or falling more) than rates for control funds. The left side of Table 7 reports the estimate of $$\delta$$ from a regression with the weighted average repo rate used as the dependent variable. Panel A reports the results around the first cap change only, with the subsequent panels adding observations around additional cap increases, as indicated. Specification (1) includes only the day before and the day of the counterparty cap increase, and shows a robustly positive effect of an increase in ONRRP participation on the rates that previously bound funds command in private repo transactions. The coefficient of 0.26 in panel A implies that the first cap increase led to an increase in previously bound funds’ average repo rate of 0.26 basis points. Although this increase appears small, the average tri-party repo rate (including both Treasury and agency collateral) was only 3 basis points on the day before the cap increase, according to the publicly available BNY Mellon Tri-Party Repo Indices. The result that the ONRRP appears to increase repo rates by offering an outside option to money funds is consistent with the findings in Han and Nikolaou (2016). Table 7 Regression results: Repo rates A. 1st cap increase Weighted average rate Memo: 25th percentile rate Memo: 75th percentile rate (1) (2) (3) (1) (2) (3) (1) (2) (3) $$\text{Bound}_{t-1}$$ 0.26** 0.23*** 0.23** 0.32*** 0.44** 0.47** 0.26** 0.21 0.20 (0.10) (0.08) (0.09) (0.10) (0.16) (0.20) (0.12) (0.13) (0.15) N 177 533 444 177 533 444 177 533 444 Adj. R$$^2$$ 0.99 0.98 0.98 0.99 0.88 0.86 0.99 0.97 0.97 B. 1st and 2nd cap increase $$\text{Bound}_{t-1}$$ 0.26*** 0.47** 0.53** 0.30*** 0.62* 0.70* 0.25*** 0.51** 0.57** (0.07) (0.19) (0.23) (0.09) (0.31) (0.38) (0.07) (0.22) (0.27) N 345 1,043 869 345 1,043 869 345 1,043 869 Adj. R$$^2$$ 0.98 0.84 0.82 0.98 0.79 0.76 0.98 0.80 0.77 C: 1st, 2nd, and 3rd cap increase $$\text{Bound}_{t-1}$$ 0.13* 0.35*** 0.39*** 0.14* 0.43*** 0.49*** 0.11* 0.35*** 0.40** (0.06) (0.09) (0.11) (0.07) (0.14) (0.17) (0.06) (0.12) (0.15) N 497 1,510 1,258 497 1,510 1,258 497 1,510 1,258 Adj. R$$^2$$ 0.98 0.86 0.84 0.98 0.82 0.79 0.98 0.83 0.80 D: 1st, 2nd, 3rd, and 4th cap increase $$\text{Bound}_{t-1}$$ 0.18* 0.31* 0.36* 0.23* 0.40* 0.46 0.20 0.34 0.39 (0.10) (0.17) (0.20) (0.12) (0.23) (0.27) (0.21) (0.13) (0.25) N 643 1,943 1,618 643 1,943 1,618 643 1,943 1,618 Adj. R$$^2$$ 0.90 0.81 0.78 0.89 0.77 0.74 0.89 0.77 0.73 A. 1st cap increase Weighted average rate Memo: 25th percentile rate Memo: 75th percentile rate (1) (2) (3) (1) (2) (3) (1) (2) (3) $$\text{Bound}_{t-1}$$ 0.26** 0.23*** 0.23** 0.32*** 0.44** 0.47** 0.26** 0.21 0.20 (0.10) (0.08) (0.09) (0.10) (0.16) (0.20) (0.12) (0.13) (0.15) N 177 533 444 177 533 444 177 533 444 Adj. R$$^2$$ 0.99 0.98 0.98 0.99 0.88 0.86 0.99 0.97 0.97 B. 1st and 2nd cap increase $$\text{Bound}_{t-1}$$ 0.26*** 0.47** 0.53** 0.30*** 0.62* 0.70* 0.25*** 0.51** 0.57** (0.07) (0.19) (0.23) (0.09) (0.31) (0.38) (0.07) (0.22) (0.27) N 345 1,043 869 345 1,043 869 345 1,043 869 Adj. R$$^2$$ 0.98 0.84 0.82 0.98 0.79 0.76 0.98 0.80 0.77 C: 1st, 2nd, and 3rd cap increase $$\text{Bound}_{t-1}$$ 0.13* 0.35*** 0.39*** 0.14* 0.43*** 0.49*** 0.11* 0.35*** 0.40** (0.06) (0.09) (0.11) (0.07) (0.14) (0.17) (0.06) (0.12) (0.15) N 497 1,510 1,258 497 1,510 1,258 497 1,510 1,258 Adj. R$$^2$$ 0.98 0.86 0.84 0.98 0.82 0.79 0.98 0.83 0.80 D: 1st, 2nd, 3rd, and 4th cap increase $$\text{Bound}_{t-1}$$ 0.18* 0.31* 0.36* 0.23* 0.40* 0.46 0.20 0.34 0.39 (0.10) (0.17) (0.20) (0.12) (0.23) (0.27) (0.21) (0.13) (0.25) N 643 1,943 1,618 643 1,943 1,618 643 1,943 1,618 Adj. R$$^2$$ 0.90 0.81 0.78 0.89 0.77 0.74 0.89 0.77 0.73 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF repo rates. $$\text{Bound}_{t-1}$$ takes a value of one for fund complexes that had at least one fund facing a binding cap on the day before an increase. Panel A includes data from the first cap increase—September 27, 2013—only. Panels B through D add data from days around the three subsequent cap increases: December 23, 2013, January 30, 2014, and March 5, 2014. Column 1 includes a 2-day sample window spanning the day before and day of a cap increase. Column 2 uses a 6-day sample window including the 5 days prior to the cap increase. Column 3 uses the same sample as Column 2, but drops the day before the cap change (the announcement day). Results are reported for fund complexes’ volume-weighted average rate and the volume-weighted 25th and 75th percentile rates, as indicated. All specifications include fund complex and day fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 7 Regression results: Repo rates A. 1st cap increase Weighted average rate Memo: 25th percentile rate Memo: 75th percentile rate (1) (2) (3) (1) (2) (3) (1) (2) (3) $$\text{Bound}_{t-1}$$ 0.26** 0.23*** 0.23** 0.32*** 0.44** 0.47** 0.26** 0.21 0.20 (0.10) (0.08) (0.09) (0.10) (0.16) (0.20) (0.12) (0.13) (0.15) N 177 533 444 177 533 444 177 533 444 Adj. R$$^2$$ 0.99 0.98 0.98 0.99 0.88 0.86 0.99 0.97 0.97 B. 1st and 2nd cap increase $$\text{Bound}_{t-1}$$ 0.26*** 0.47** 0.53** 0.30*** 0.62* 0.70* 0.25*** 0.51** 0.57** (0.07) (0.19) (0.23) (0.09) (0.31) (0.38) (0.07) (0.22) (0.27) N 345 1,043 869 345 1,043 869 345 1,043 869 Adj. R$$^2$$ 0.98 0.84 0.82 0.98 0.79 0.76 0.98 0.80 0.77 C: 1st, 2nd, and 3rd cap increase $$\text{Bound}_{t-1}$$ 0.13* 0.35*** 0.39*** 0.14* 0.43*** 0.49*** 0.11* 0.35*** 0.40** (0.06) (0.09) (0.11) (0.07) (0.14) (0.17) (0.06) (0.12) (0.15) N 497 1,510 1,258 497 1,510 1,258 497 1,510 1,258 Adj. R$$^2$$ 0.98 0.86 0.84 0.98 0.82 0.79 0.98 0.83 0.80 D: 1st, 2nd, 3rd, and 4th cap increase $$\text{Bound}_{t-1}$$ 0.18* 0.31* 0.36* 0.23* 0.40* 0.46 0.20 0.34 0.39 (0.10) (0.17) (0.20) (0.12) (0.23) (0.27) (0.21) (0.13) (0.25) N 643 1,943 1,618 643 1,943 1,618 643 1,943 1,618 Adj. R$$^2$$ 0.90 0.81 0.78 0.89 0.77 0.74 0.89 0.77 0.73 A. 1st cap increase Weighted average rate Memo: 25th percentile rate Memo: 75th percentile rate (1) (2) (3) (1) (2) (3) (1) (2) (3) $$\text{Bound}_{t-1}$$ 0.26** 0.23*** 0.23** 0.32*** 0.44** 0.47** 0.26** 0.21 0.20 (0.10) (0.08) (0.09) (0.10) (0.16) (0.20) (0.12) (0.13) (0.15) N 177 533 444 177 533 444 177 533 444 Adj. R$$^2$$ 0.99 0.98 0.98 0.99 0.88 0.86 0.99 0.97 0.97 B. 1st and 2nd cap increase $$\text{Bound}_{t-1}$$ 0.26*** 0.47** 0.53** 0.30*** 0.62* 0.70* 0.25*** 0.51** 0.57** (0.07) (0.19) (0.23) (0.09) (0.31) (0.38) (0.07) (0.22) (0.27) N 345 1,043 869 345 1,043 869 345 1,043 869 Adj. R$$^2$$ 0.98 0.84 0.82 0.98 0.79 0.76 0.98 0.80 0.77 C: 1st, 2nd, and 3rd cap increase $$\text{Bound}_{t-1}$$ 0.13* 0.35*** 0.39*** 0.14* 0.43*** 0.49*** 0.11* 0.35*** 0.40** (0.06) (0.09) (0.11) (0.07) (0.14) (0.17) (0.06) (0.12) (0.15) N 497 1,510 1,258 497 1,510 1,258 497 1,510 1,258 Adj. R$$^2$$ 0.98 0.86 0.84 0.98 0.82 0.79 0.98 0.83 0.80 D: 1st, 2nd, 3rd, and 4th cap increase $$\text{Bound}_{t-1}$$ 0.18* 0.31* 0.36* 0.23* 0.40* 0.46 0.20 0.34 0.39 (0.10) (0.17) (0.20) (0.12) (0.23) (0.27) (0.21) (0.13) (0.25) N 643 1,943 1,618 643 1,943 1,618 643 1,943 1,618 Adj. R$$^2$$ 0.90 0.81 0.78 0.89 0.77 0.74 0.89 0.77 0.73 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on MMF repo rates. $$\text{Bound}_{t-1}$$ takes a value of one for fund complexes that had at least one fund facing a binding cap on the day before an increase. Panel A includes data from the first cap increase—September 27, 2013—only. Panels B through D add data from days around the three subsequent cap increases: December 23, 2013, January 30, 2014, and March 5, 2014. Column 1 includes a 2-day sample window spanning the day before and day of a cap increase. Column 2 uses a 6-day sample window including the 5 days prior to the cap increase. Column 3 uses the same sample as Column 2, but drops the day before the cap change (the announcement day). Results are reported for fund complexes’ volume-weighted average rate and the volume-weighted 25th and 75th percentile rates, as indicated. All specifications include fund complex and day fixed effects, with standard errors clustered at the fund complex level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. The second column of Table 7—labeled (2)—extends the pretreatment period to include the five trading days prior to the change in the counterparty cap, for a total of six days, and shows a very similar pattern upon an increase in the maximum bid. To avoid any potentially confounding effects of the announcement of the cap increase, which occurred 1 business day before each cap increase, specification (3) excludes the announcement day from the 6-day window used in Column 2. Although the announcements were made during the trading day prior to the cap increase, they generally occurred after most private repo activity had taken place. Thus, it is unsurprising that the five day results excluding the announcement day are consistent with the six day sample presented in Column 2. Panels B through D show that these results persist when the sample is expanded to include the days around other cap changes. Including the final cap increase in the sample (panel D) produces less precise estimates, but this is likely attributable to the fact that most complexes did not face a binding cap during this episode, and thus there are very few trades associated with treated funds. In the memorandum items in the middle and right side of Table 7, we report the effect of the cap increase on the weighted 25th and 75th percentile of fund families’ rate distribution. As expected, the cap increase has a somewhat larger effect on lower-rate trades. In total, the results reported in Table 7 are consistent with the hypothesis that the outside option presented by the ONRRP increases repo lenders’ bargaining position vis-à-vis borrowers.24 Therefore, as a result of the ONRRP facility, dealers not only witnessed a reduction in the supply of funds as shown in section 4.1, but also faced a contemporaneous increase in their funding costs. Consistent with our interpretation, this combination characterizes a reduction in the supply of private repo investment. 4.4 Effects of the ONRRP on private repo borrowers Having shown that repo lenders substitute away from private repo when they invest more in the ONRRP, we now turn to the question of how dealers respond to this adverse funding shock and what additional implications this may have for financial stability. As dealers face more competition for repo funding from the ONRRP, financial stability could be affected in a number of ways. In response to the withdrawal of lending by ONRRP-eligible MMFs, dealers may decrease total repo volume, possibly reducing financial fragility by limiting the externalities associated with extensive private money creation in the repo market. Conversely, dealers could maintain their level of repo funding in the wake of the negative ONRRP supply shock by shifting their composition of counterparties and/or collateral. Such a response could potentially increase financial fragility if dealers increase their borrowing from smaller, less stable lenders. A shift to riskier collateral repo would also have financial stability implications for two reasons. First, repo backed by riskier collateral types appears to be more susceptible to runs during a financial crisis. For instance, Krishnamurthy, Nagel, and Orlov (2014) and Copeland, Martin, and Walker (2014) show that the use of riskier collateral in the tri-party repo market dropped off significantly during the recent financial crisis. In contrast, Treasury repo was more stable. Therefore, if dealers increase their reliance on repo backed by other collateral types, it may make a run in the repo market both more likely and more pronounced.25 Second, to enter into repo transactions backed by riskier collateral types, dealers must either own or borrow those assets. Such a shift toward riskier types of repo funding could therefore lead in turn to a shift toward a riskier overall asset portfolio among dealers, thereby increasing financial fragility (Parlatore 2016). The final portion of our analysis thus proceeds with an examination of the dealer borrowing response to the ONRRP-induced withdrawal of privately supplied repo funding in two steps. First, we consider the effects on the mix and quality of collateral that backs dealer repo borrowing using our dealer/collateral position-level repo data set. Additionally, we use the FR-2004C schedule of all dealer repo positions to determine if dealers begin using riskier types of repo for net financing. Second, we examine the effects of the ONRRP on dealers’ composition of lenders using our dealer/lender position-level repo data set. As before, we identify off of exogenous cap increases in the period after the introduction of the ONRRP facility using a DD framework. However, we no longer have a binary treatment variable available for this analysis. Instead, we generate a continuous treatment variable that measures dealers’ ex ante exposure to constrained (treatment) MMFs. In particular, we define the constrained share as of the September 2013 introduction of the ONRRP for dealer $$j$$ as \begin{equation} \text{Constrained}_{j} = \frac{\text{Repo Volume with Constrained MMFs}_{j}}{\text{Total Repo Volume}_{j}}. \end{equation} (8) $$\text{Constrained}_{j}$$ has a mean of 0.087 (i.e., 8.7% of total repo borrowing comes from constrained MMFs) with a standard deviation of 0.12 and a range of 0 to 0.4. We determine which dealers trade with constrained MMFs using the N-MFP data, as before. This information is then merged into both of our position-level data sets on tri-party repo positions, as well as the FR-2004C data. Using the tri-party data sets allows us to capture the entirety of dealer borrowing in the tri-party market rather than just the trades with MMFs that we observe in the N-MFP. Except where noted, we again sample dealer repo borrowing on quarter-end dates between December 2012 and June 2015. 4.4.1 Repo collateral quality To determine how the amount and composition of dealer repo is affected by MMFs’ investment with the Fed, we consider four measures of dealer repo borrowing: the log of repo volume, both in aggregate and collateralized with either Treasury, agency, or other securities. Specifically, we estimate the following regression: \begin{equation} \begin{split} \text{y}_{jt} &= \delta\cdot(\text{Post}_{t}\cdot\text{Constrained}_{j}) + \beta_{j}\cdot\text{Dealer}_{j} + \gamma_{t}\cdot\text{Quarter}_{t} + \varepsilon_{jt}. \end{split} \end{equation} (9) In this specification, $${Post}_{t}$$ takes a value of one for all quarter-ends after September 2013, and $$\delta$$ is our primary coefficient of interest, measuring the sensitivity of dealer repo activity to their ex ante exposure to MMFs that increase their use of the ONRRP during the post-treatment period as a result of the exogenous cap increases. The results of this analysis are shown in panel A of Table 8. In the first column, we see that total repo borrowing is essentially unchanged after the ONRRP cap increases. Therefore, we do not find evidence that the ONRRP reduces dealers’ total repo borrowing. Examining the collateral composition of dealer repo borrowing, we see in the second column that Treasury repo borrowing is essentially unchanged. Evidently, dealers replace the Treasury-backed repo withdrawn by ONRRP-eligible MMFs by trading with other counterparties. Agency repo—reported in the third column—appears to fall, though this result does not achieve statistical significance. Evidently, dealers replace the Treasury-backed repo withdrawn by ONRRP-eligible MMFs by trading with other counterparties. Finally, other repo volume is significantly higher, as seen in the final column. For context, the $$\delta$$ value of 0.71 for other repo implies that for a 1 percentage point increase in repo borrowing from initially constrained ONRRP-eligible MMFs ($${\$}$$80 million for the median dealer), riskier tri-party repo borrowing against nongovernment collateral increases $${\$}$$14 million for the median dealer. Table 8 Regression results: Dealers’ collateral response to the adverse funding shock A. Dealer repo volume by collateral Total repo volume Treasury repo volume Agency repo volume Other repo volume $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.03 0.15 –0.40 0.71* (0.29) (0.43) (0.48) (0.40) N 654 499 558 429 Adj. R$$^2$$ 0.99 0.98 0.99 0.98 A. Dealer repo volume by collateral Total repo volume Treasury repo volume Agency repo volume Other repo volume $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.03 0.15 –0.40 0.71* (0.29) (0.43) (0.48) (0.40) N 654 499 558 429 Adj. R$$^2$$ 0.99 0.98 0.99 0.98 B. Haircuts 50th percentile 75th percentile 90th percentile 95th percentile $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.002 0.02 0.04** 0.03** (0.004) (0.01) (0.02) (0.01) N 654 654 654 654 Adj. R$$^2$$ 0.98 0.95 0.93 0.91 B. Haircuts 50th percentile 75th percentile 90th percentile 95th percentile $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.002 0.02 0.04** 0.03** (0.004) (0.01) (0.02) (0.01) N 654 654 654 654 Adj. R$$^2$$ 0.98 0.95 0.93 0.91 C. Net other repo financing All nongovernment collateral Corporate debt Other miscellaneous collateral $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.63*** 1.59*** 0.31** (0.22) (0.38) (0.15) N 1,628 1,508 1,405 Adj. R$$^2$$ 0.63 0.60 0.69 C. Net other repo financing All nongovernment collateral Corporate debt Other miscellaneous collateral $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.63*** 1.59*** 0.31** (0.22) (0.38) (0.15) N 1,628 1,508 1,405 Adj. R$$^2$$ 0.63 0.60 0.69 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on dealer repo activity. Post is equal to 1 for all time periods beginning in Q4 2013. Constrained is equal to a dealer’s dollar volume of repo borrowed from money funds that are constrained by the ONRRP cap as of Q3 2013 quarter-end, divided by the dealer’s total repo volume. In panel A, the dependent variables are measured in log volumes. Dependent variables in panel B are dealer-specific volume-weighted percentile haircuts. The dependent variables in panel C are the shares of repo of different collateral types, as indicated, that are used for net financing. Data used in panel C are measured at the weekly frequency. All specifications in panels A through C include dealer and time fixed effects. Standard errors are clustered at the dealer level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 8 Regression results: Dealers’ collateral response to the adverse funding shock A. Dealer repo volume by collateral Total repo volume Treasury repo volume Agency repo volume Other repo volume $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.03 0.15 –0.40 0.71* (0.29) (0.43) (0.48) (0.40) N 654 499 558 429 Adj. R$$^2$$ 0.99 0.98 0.99 0.98 A. Dealer repo volume by collateral Total repo volume Treasury repo volume Agency repo volume Other repo volume $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.03 0.15 –0.40 0.71* (0.29) (0.43) (0.48) (0.40) N 654 499 558 429 Adj. R$$^2$$ 0.99 0.98 0.99 0.98 B. Haircuts 50th percentile 75th percentile 90th percentile 95th percentile $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.002 0.02 0.04** 0.03** (0.004) (0.01) (0.02) (0.01) N 654 654 654 654 Adj. R$$^2$$ 0.98 0.95 0.93 0.91 B. Haircuts 50th percentile 75th percentile 90th percentile 95th percentile $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.002 0.02 0.04** 0.03** (0.004) (0.01) (0.02) (0.01) N 654 654 654 654 Adj. R$$^2$$ 0.98 0.95 0.93 0.91 C. Net other repo financing All nongovernment collateral Corporate debt Other miscellaneous collateral $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.63*** 1.59*** 0.31** (0.22) (0.38) (0.15) N 1,628 1,508 1,405 Adj. R$$^2$$ 0.63 0.60 0.69 C. Net other repo financing All nongovernment collateral Corporate debt Other miscellaneous collateral $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.63*** 1.59*** 0.31** (0.22) (0.38) (0.15) N 1,628 1,508 1,405 Adj. R$$^2$$ 0.63 0.60 0.69 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on dealer repo activity. Post is equal to 1 for all time periods beginning in Q4 2013. Constrained is equal to a dealer’s dollar volume of repo borrowed from money funds that are constrained by the ONRRP cap as of Q3 2013 quarter-end, divided by the dealer’s total repo volume. In panel A, the dependent variables are measured in log volumes. Dependent variables in panel B are dealer-specific volume-weighted percentile haircuts. The dependent variables in panel C are the shares of repo of different collateral types, as indicated, that are used for net financing. Data used in panel C are measured at the weekly frequency. All specifications in panels A through C include dealer and time fixed effects. Standard errors are clustered at the dealer level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. This increase in risky repo borrowing can also be demonstrated by comparing changes in the value of collateral that dealers must deliver to secure a repo loan of a given size. The difference between the collateral value and the loan amount divided by the loan amount is known as the haircut, with riskier collateral subject to higher haircuts. For example, while the haircut on Treasury repo is around 2%, the haircut on equity repo is around 8%. To assess whether dealers’ average haircut increases as they lose funding to the ONRRP, we estimate an alternative version of Equation (9) in which $${y}_{jt}$$ is dealer j’s volume-weighted haircut across all repo positions for selected percentiles listed in panel B of Table 8. There is no change in the average (data not shown) or median haircut. However, there is a significant increase in the haircut at the 90th and 95th percentiles of the distribution. The point estimate of 0.04 is economically significant, as it implies that just a 1 percentage point increase in repo volume borrowed from initially constrained ONRRP-eligible MMFs translates to a 1% increase in the average 90th percentile haircut of 4.7%. As dealers shift to repo backed by nongovernment collateral, their haircut requirements are also increasing, showing that they have taken on more risk. The move to a riskier repo position by those dealers borrowing from MMFs that invest in the ONRRP would be most concerning if this repo is used as a net source of funding for securities with longer maturities. If risky repo is used to fund dealers’ net securities position, a stress event that limited repo financing backed by risky collateral could lead to fire sales of those securities. If dealers instead simply acquired the riskier collateral through reverse repos as part of their matched book activity, the net exposure to the dealer would be very limited, as the transactions would essentially offset. In the event that repo funding secured by risky collateral dried up, the dealer could simply decline to roll over the matching reverse repo transactions. Using the FR-2004C data that contain primary dealers’ tri-party and bilateral repo and reverse repo activity by collateral type, we are able to examine whether net financing with riskier repo increases in response to the expansion of the ONRRP. Specifically, we estimate Equation (9) on weekly FR-2004C data in which $${y}_{jt}$$ is defined as follows for dealer j’s non-Treasury or agency collateral:26 \begin{equation} y_{jt} = \frac{\text{other repo}_{jt}-\text{other reverse repo}_{jt}}{\text{other repo}_{jt}}. \end{equation} (10) Therefore, $${y}_{jt}$$ measures the share of repo borrowing backed by risky collateral that is used for net financing as opposed to matched book activity. Given the relative flightiness of other repo, higher values of $${y}_{jt}$$ represent an increase in fragility. The first column of panel C in Table 8 shows that dealers that lost more repo funding to the ONRRP do indeed boost their net financing using collateral that is not government-backed. The point estimate implies that a 1 percentage point increase in repo volume borrowed from initially constrained MMFs resulted in an additional $${\$}$$75 million of net funding with repo backed by risky collateral, on average. This represents nearly 1% of average net financing using other repo. The second and third columns of panel C show that this result also holds for different subtypes of nongovernment collateral. However, the marginal effect is larger for the relatively safe collateral like corporate bonds (average haircut of 6.6%) than for the riskiest collateral like private CMOs (average haircut of 9.8%), which are included in the miscellaneous category. 4.4.2 Repo counterparties To better understand the increase in dealers’ other repo borrowing, we now turn to an analysis of the change in lender composition. We divide repo lenders into 3 categories: ONRRP-eligible MMFs, ONRRP-ineligible MMFs, and non-MMFs.27 We then estimate alternate versions of Equation (9) in which $${y}_{jt}$$ is the share of total repo volume borrowed from each of these groups of lenders. Panel A of Table 9 reports the results from this exercise. Table 9 Regression results: Dealers’ counterparty response to the adverse funding shock A. Counterparty composition of dealer repo Eligible MMF share Ineligible MMF share Non-MMF share Memo: Weighted $$\sigma$$(AUM) $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.30*** 0.08** 0.22** 0.10* (0.08) (0.03) (0.09) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.96 0.80 0.99 0.92 A. Counterparty composition of dealer repo Eligible MMF share Ineligible MMF share Non-MMF share Memo: Weighted $$\sigma$$(AUM) $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.30*** 0.08** 0.22** 0.10* (0.08) (0.03) (0.09) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.96 0.80 0.99 0.92 B. Non-MMF counterparty composition of dealer repo Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.16*** 0.02 –0.10 0.05 (0.04) (0.01) (0.11) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.77 0.98 0.83 0.96 B. Non-MMF counterparty composition of dealer repo Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.16*** 0.02 –0.10 0.05 (0.04) (0.01) (0.11) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.77 0.98 0.83 0.96 C. Non-MMF counterparty composition of dealer repo at the relationship level Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.05** 0.02 0.11 0.05 (0.02) (0.03) (0.10) (0.04) N 1,700 2,082 681 968 Adj. R$$^2$$ 0.74 0.86 0.95 0.21 C. Non-MMF counterparty composition of dealer repo at the relationship level Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.05** 0.02 0.11 0.05 (0.02) (0.03) (0.10) (0.04) N 1,700 2,082 681 968 Adj. R$$^2$$ 0.74 0.86 0.95 0.21 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on dealer counterparties in the repo market. Post is equal to 1 for all time periods beginning in Q4 2013. Constrained is equal to a dealer’s dollar volume of repo borrowed from money funds that are constrained by the ONRRP cap as of Q3 2013 quarter-end, divided by the dealer’s total repo volume. In each panel, the dependent variables are the share of total repo volume borrowed from a particular counterparty type. All specifications in panels A through C include dealer fixed effects. Panels A and B include time fixed effects, and panel C contains lender-time fixed effects. Standard errors are clustered at the dealer level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. Table 9 Regression results: Dealers’ counterparty response to the adverse funding shock A. Counterparty composition of dealer repo Eligible MMF share Ineligible MMF share Non-MMF share Memo: Weighted $$\sigma$$(AUM) $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.30*** 0.08** 0.22** 0.10* (0.08) (0.03) (0.09) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.96 0.80 0.99 0.92 A. Counterparty composition of dealer repo Eligible MMF share Ineligible MMF share Non-MMF share Memo: Weighted $$\sigma$$(AUM) $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ –0.30*** 0.08** 0.22** 0.10* (0.08) (0.03) (0.09) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.96 0.80 0.99 0.92 B. Non-MMF counterparty composition of dealer repo Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.16*** 0.02 –0.10 0.05 (0.04) (0.01) (0.11) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.77 0.98 0.83 0.96 B. Non-MMF counterparty composition of dealer repo Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.16*** 0.02 –0.10 0.05 (0.04) (0.01) (0.11) (0.06) N 654 654 654 654 Adj. R$$^2$$ 0.77 0.98 0.83 0.96 C. Non-MMF counterparty composition of dealer repo at the relationship level Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.05** 0.02 0.11 0.05 (0.02) (0.03) (0.10) (0.04) N 1,700 2,082 681 968 Adj. R$$^2$$ 0.74 0.86 0.95 0.21 C. Non-MMF counterparty composition of dealer repo at the relationship level Eligible asset manager share Ineligible asset manager share Primary dealer share Custodian share $$\text{Post}_{t}\cdot\text{Constrained}_{j}$$ 0.05** 0.02 0.11 0.05 (0.02) (0.03) (0.10) (0.04) N 1,700 2,082 681 968 Adj. R$$^2$$ 0.74 0.86 0.95 0.21 This table reports DD estimates of the effect of an increase in the ONRRP counterparty cap on dealer counterparties in the repo market. Post is equal to 1 for all time periods beginning in Q4 2013. Constrained is equal to a dealer’s dollar volume of repo borrowed from money funds that are constrained by the ONRRP cap as of Q3 2013 quarter-end, divided by the dealer’s total repo volume. In each panel, the dependent variables are the share of total repo volume borrowed from a particular counterparty type. All specifications in panels A through C include dealer fixed effects. Panels A and B include time fixed effects, and panel C contains lender-time fixed effects. Standard errors are clustered at the dealer level. Statistical significance: *** p $$\leq$$ .01, ** p $$\leq$$ .05, * p $$\leq$$ .10. As expected, the first column reveals a clear decline in repo trading with ONRRP-eligible MMFs, as these are precisely the funds that substitute to the ONRRP. There is a small increase in borrowing from MMFs that are ineligible for the ONRRP, as shown in the second column.28 In Column 3, we see that much of the lost MMF funding is recouped by borrowing from non-MMF lenders (recall that total repo borrowing is unchanged). This result is intuitive in the context of the results found in panel A of Table 8, because non-MMFs are the primary lenders in the market for other collateral repo. The point estimates imply that a 1 percentage point increase in repo volume borrowed from initially constrained MMFs translates to a 0.3 percentage point decrease in borrowing from eligible MMFs, a 0.08 percentage point increase in borrowing from ineligible MMFs, and a 0.22 percentage point increase from non-MMFs. ONRRP-ineligible MMFs have fewer assets under management than eligible funds on average such that the dealers’ move from eligible to ineligible MMFs implies a shift to smaller lenders. However, dealers may choose to increase borrowing from only the most stable ineligible MMFs, making their relative size less of a concern. To test this, we measure the percentage standard deviation of detrended AUM for each MMF from 2011-2014 and compute a volume-weighted average of this standard deviation for each dealer. We then use this volume-weighted standard deviation of counterparty AUM as a dependent variable in Equation (9), and report the results in the memorandum item in the last column of panel A of Table 9. We find that after ONRRP cap increases, dealers trade with more asset-volatile MMFs, suggesting a move to a less stable source of funding for dealers. Since dealers are predominately substituting to non-MMFs, we turn now to an analysis of the increase in funding from non-MMFs. Using our data set on tri-party repo positions between dealers and lenders, we observe total dealer repo positions with individual lenders, which are then aggregated by lender type. Specifically, we focus on total borrowing from asset managers affiliated with ONRRP-eligible MMFs, asset managers affiliated only with ONRRP-ineligible MMFs, primary dealer affiliates, and custodian banks. Asset managers include mutual funds, hedge funds, and other asset managers. Custodian banks include nonprimary dealer custodians, namely Bank of New York-Mellon and State Street. In total, repo borrowing from these lenders accounts for over 90% of total tri-party repo volume. We then estimate alternate versions of Equation (9) in which $${y}_{jt}$$ is the share of repo volume borrowed from each group of non-MMF lenders. The results of this analysis are reported in panel B of Table 9. When dealers shift to non-MMF counterparties, they transact more with asset managers affiliated with at least one eligible ONRRP counterparty (shown in the first column). There is not a significant change in borrowing from asset managers without any eligible counterparties, primary dealer affiliates, or custodian banks (Columns 2 through 4 of panel B). Evidently, dealers endeavor to maintain their relationships with the largest (ONRRP-eligible) asset manager complexes by transacting with their non-MMF affiliates. Because non-MMFs lend more against other collateral repo, however, dealers must accommodate this transition by expanding their borrowing against riskier collateral, consistent with the findings discussed in the previous subsection and detailed in Table 8. If constrained dealers have relationships with asset managers that are increasing their other collateral repo lending at the expense of collateral types held by their MMFs for reasons unrelated to the ONRRP, this could invalidate the interpretation of our results. To rule out a lender-driven explanation of our results, we construct a repo credit registry that contains total volume between each dealer-lender pair. We can then estimate a regression that includes lender-time fixed effects as follows: \begin{align} y_{ijt} &= \delta\cdot(\text{Post}_{t}\cdot\text{Constrained}_{j}) + \phi_{it}\cdot(\text{Lender}_{i}\cdot\text{Quarter}_{t}) \notag\\ &\quad + \beta_{ij}\cdot\text{Lender-Dealer}_{ij} + \varepsilon_{ijt}. \end{align} (11) In Equation 11, $$y_{ijt}$$ represents the share of the total repo volume of dealer $$j$$ borrowed from lender $$i$$. The key coefficient $$\delta$$ then measures the borrowing response of dealers that differ in their constrained share, holding lender and quarter constant. The results are shown in panel C of Table 9. Consistent with the results in panel B, we see that constrained dealers shift to asset managers affiliated with at least one eligible ONRRP counterparty (shown in the first column). There is no significant change in repo borrowing from other lender types. These results validate the finding that dealers seek to maintain their relationships with the largest asset managers by switching to affiliated non-MMFs when MMFs withdraw repo funding due to the ONRRP. These non-MMF managers, however, are more likely to lend against a mix of non-Treasury and agency securities that mirrors their existing asset portfolios. 4.4.3 Discussion of dealer responses Overall, these results indicate that the ONRRP poses some risks to financial stability. Dealers that are more exposed to initially constrained MMFs that subsequently withdraw private repo funding in favor of the ONRRP do not reduce their total repo volume. Thus, one potential benefit of the ONRRP—that it could reduce the size of the run-prone repo market (Carlson et al. 2016)—has evidently not materialized. Rather, dealers make up for any lending lost to the ONRRP by switching to a different collateral and lender mix. Specifically, dealers increase their repo borrowing from non-MMFs, leading to an increase in borrowing against riskier collateral and a greater reliance on riskier repo for net financing. To a lesser degree, these dealers also shift to more asset-volatile, ONRRP-ineligible MMF counterparties. While the economic magnitudes of the effects we observe in response to the relatively small funding shock are not large enough to pose an immediate threat to financial stability, the results demonstrate how dealers respond to funding supply shocks that are introduced by monetary policy implementation. In the event of severe financial stress, a flight of repo lenders to the ONRRP could trigger a sharp decline in private repo supply. Our results show that dealers are more susceptible to such a shock for two reasons. First, dealers shift to repo backed by less-safe collateral types, which is far less reliable during episodes of financial stress (Krishnamurthy, Nagel, and Orlov 2014; Copeland, Martin, and Walker 2014). Second, a move by dealers to substitute ONRRP-eligible MMFs with smaller and more volatile ineligible MMFs could also have implications for financial stability. Although ineligible MMFs cannot run to the ONRRP in times of financial stress, their investors may still withdraw cash, forcing the fund to exit the private repo market and starving dealers of this source of funding. Both of these dealer responses therefore suggest that monetary policy implementation can increase the likelihood and severity of financial disruption in the repo market. In response to a flight to the ONRRP, the Fed could potentially limit the ONRRP by imposing stricter caps or lowering the ONRRP rate. However, this precisely captures the tradeoff between effective monetary policy implementation and financial stability. To improve financial stability by restricting the ONRRP, the Fed would need to relinquish its control of short-term interest rates. Furthermore, there is an additional tradeoff between the stability of dealers and the stability of MMFs. By offering reliable access to a safe asset, the ONRRP improves the safety of the MMF sector. If MMFs did not have the backstop of the ONRRP, their investors may run, leading to a comparable decline in private repo funding. Given these tradeoffs, the Fed may be unwilling to curtail the ONRRP in times of stress. Although the ONRRP effectively controls rates and provides an important backstop for MMFs during times of financial strain, it could also increase the likelihood or severity of financial instability emanating from dealers, as we have shown. Efforts to stem this instability by limiting the ONRRP could destabilize the MMF sector and undermine the Fed’s ability to manage the policy rate. 5. Conclusion In this paper, we conduct an analysis of the Fed’s regular intervention in the repo market through the ONRRP facility, the Fed’s newest monetary policy tool. Specifically, we exploit exogenous changes in an MMF’s ability to invest in the Fed’s ONRRP facility to identify substitution away from private repo transactions and into repo investment with the Fed. Further analysis of the pattern of MMF substitution shows that, rather than severing trades with certain borrowers entirely, money funds withdraw from their dealer counterparties in a roughly even fashion, albeit with somewhat more withdrawal from their largest borrowers. This pattern of substitution likely reflects MMFs’ desire to preserve existing lending relationships, highlighting the importance of these relationships in the repo market. Additionally, we use confidential data on trades in the tri-party repo market to show that the ONRRP facility bestows additional bargaining power on repo lenders with the option of investing with the Fed. When MMFs are able to invest more in the ONRRP, rates on their private repo transactions increase. Thus, we demonstrate that the presence of the Fed as a borrower in the repo market not only saps repo funding from the market, but also leads to higher dealer funding costs. This dynamic can potentially increase the likelihood of future runs in the repo market if MMFs move away from private repo in favor of Fed repo in the event of financial turmoil, with repo borrowers forced to pay increasingly higher rates. Therefore, it is important to understand how repo borrowers respond to an adverse funding shock. We find that, although dealers lose funding from ONRRP-eligible funds that invest more in the ONRRP, they are able to fill that void by reallocating their repo portfolio and thus the ONRRP does not lead to a decrease in total dealer repo volume. Rather, dealers substitute away from Treasury and agency collateral repo toward repo backed by riskier assets, and increase their reliance on risky repo as a source of net financing. This new funding mix is evidently the result of dealers’ shifting borrowing away from ONRRP-eligible MMFs and toward non-MMF asset managers and ONRRP-ineligible MMFs, which have less stable assets and repo investment. Overall, this study demonstrates how the implementation of monetary policy—particularly when it relies on a large presence in financial markets—can disrupt private activity in funding markets. Our results illuminate an important tradeoff between effective monetary policy implementation and financial stability. Many central banks have enlarged their footprint in financial markets as a consequence of the drastic expansion of the scale and scope of monetary intervention in financial markets since the recent crisis. As we have begun to show in this paper, the financial repercussions engendered by this intervention can have potentially far-reaching implications for private trading activity, collateral assets, and financial stability. We are grateful for helpful comments from Stijn Van Nieuwerburgh (the editor); two anonymous referees; Sriya Anbil, Jim Clouse, Jane Ihrig, Elizabeth Klee, Marco Macchiavelli, Bernd Schlusche, and Rebecca Zarutskie; members of the Money Markets Workgroup for the Federal Reserve System’s Long-Run Monetary Policy Implementation Framework; conference participants at the Australasian Finance and Banking Conference, the World Finance Conference, the IBEFA Summer Meeting, and the FMA European Conference; and seminar participants at the Federal Reserve Board, the Federal Reserve Bank of New York, and the International Monetary Fund. We thank Michelle Bongard for excellent research assistance. The views expressed in this paper are solely the responsibility of the authors and should not be interpreted as reflecting the views of the Board of Governors of the Federal Reserve System or of anyone else associated with the Federal Reserve System. Footnotes 1 The tri-party repo market is one segment of the total repo market. For more information, see Section 1.2. 2 See www.newyorkfed.org/markets/rrp_eligibility_criteria.html for a description of the current eligibility requirements for MMFs, GSEs, and banks. 3 See Frost et al. (2015) for a complete discussion of the design of the ONRRP. 4 Subsequently, other counterparties became eligible. See www.newyorkfed.org/markets/expanded_counterparties.html for a full list of the current eligible counterparties. 5 See www.federalreserve.gov/newsevents/press/monetary/20140917c.htm. 6 See “Minutes of the Federal Open Market Committee November 1-2, 2016,” https://www.federalreserve.gov/monetarypolicy/files/fomcminutes20161102.pdf. 7 The ONRRP also had an aggregate cap of $\$$300 billion between September 22, 2014 and December 16, 2015. Before and after that period, there was no aggregate cap. However, the cap was reached only once, on September 30, 2014. 8 In the bilateral market, cash providers and cash borrowers trade directly. Bilateral repo typically consists of interdealer trading or dealers lending to hedge funds and others. Conversely, the General Collateral Finance (GCF) market is a blind-brokered interdealer market that is a subset of the broader tri-party market. For a complete description of the U.S. repo market, see Copeland et al. (2012). 9 Our treatment group is defined as those funds that were constrained by the $\$$1 billion cap on the first quarter-end after the start of the ONRRP (September 30, 2013). $\$$1 billion was technically the second cap, as the cap was $\$$500 million for the first 4 days of the ONRRP. However, because we are using quarter-end data for most of our analysis, we use the cap at the first quarter-end ( $\$$1 billion) to sort funds into treatment and control groups. 10 In results presented in Section 4, we show that treatment funds did in fact significantly increase their ONRRP investment after an increase in the maximum bid amount. 11 Repo and reverse repo are two sides the same transaction; cash borrowers are said to engage in repo transactions, while cash lenders commit to reverse repo transactions. Unlike most institutions, dealers engage in both types of transactions given their role as intermediaries between different market segments. 12 Primary dealers represent a subset of the dealers that we observe in the tri-party repo market. However, most of the largest dealers are primary dealers and a majority of the dealers that trade in the tri-party market are primary dealers. 13 We exclude the relatively few eligible funds classified as “tax-exempt.” 14 See Section 4.2 for definitions of weighted CDS spreads and HHI. 15 Countries differ on the time frame over which the leverage ratio is calculated. In the United States, the leverage ratio is based on a daily average across a quarter, while in the eurozone, it is based on the value on the quarter-end day only. Other foreign regions have intermediate calculation time frames. The treatment and control funds also have no difference in the percentage of their lending specifically to eurozone dealers (data not shown). 16 We note, however, that size of an MMF is not an absolute predictor of whether a fund found the counterparty cap binding. In fact, 8 of the largest 20 MMFs in our sample (including the largest fund) did not bid at the $\$$1 billion counterparty cap on September 30, 2013. Our results are nearly identical if we simply control for AUM by including the lagged value for all funds in all time periods. 17 The total investment in the ONRRP could be constrained by the designated aggregate and counterparty caps, which could be changed in the future. 18 This placebo treatment date is chosen to correspond to the first post-treatment period of our baseline results (December 2013). However, the results of this exercise are not sensitive to the choice of hypothetical treatment date. 19 However, as noted in Section 3, treated and control funds have similar relationships with foreign dealers. 20 CDS data are obtained from Markit. 21 In each case, the announcement of the increase in the maximum counterparty bid amount was made on the business day prior to the change. 22 Because the cap was increased monotonically, trades by previously bound funds generally declined over time as a share of total trades. After collapsing trades to family-dealer-collateral triples, we find that the percentage of trades attributable to the previously bound treatment cohort were, respectively, 23%, 28%, 15%, and 9%. 23 Though power is decreased, we find very similar results when separating the treatment group into funds that were unconstrained on the day of the cap increase and funds that remained bound by the cap. 24 The finding that Fed intervention can increase repo rates accords with Fleming, Hrung, and Keane (2010), who show that the Term Securities Lending Facility (TSLF)—a temporary emergency response to the developing financial crisis in 2008—resulted in higher repo rates. However, the TSLF boosted repo rates by mitigating acute shortages of Treasury collateral, whereas the ONRRP evidently works through increased bargaining power. 25 Other repo includes repo backed by assets such as asset backed securities, private label mortgage-backed securities, corporate bonds, and equity. 26 Because of changes in FR-2004C reporting requirements, we are unable to use the sample period employed earlier. The sample period for this analysis begins in April 2013. 27 Non-MMF lenders in the tri-party repo market include asset managers, primary dealer-affiliated institutions, securities lenders, banks (including custodian banks), pension funds, insurance companies, and others. 28 In addition to the low point estimate, we note that base effects also imply a small increase, as there is less borrowing from ineligible MMFs relative to other types of lenders. References Adrian, T., and Liang. N. 2016 . Monetary policy, financial conditions, and financial stability. Staff Report. Center for Economic Policy Research . Google Scholar CrossRef Search ADS Baklanova, V., Copeland, A. M. and McCaughrin. R. 2015 . Reference guide to US repo and securities lending markets. Working Paper, FRBNY . Bech, M., and Klee. E. 2011 . The mechanics of a graceful exit: Interest on reserves and segmentation in the federal funds market. Journal of Monetary Economics 58 : 415 – 31 . Google Scholar CrossRef Search ADS Bernanke, B. S. 2012 . Monetary policy since the onset of the crisis. Speech at the Federal Reserve Bank of Kansas City Economic Symposium, Jackson Hole, WY, August 31 . http://www.federalreserve.gov/newsevents/speech/bernanke20120831a.htm. Bertrand, M., Duflo, E. and Mullainathan. S. 2004 . How much should we trust differences-in-differences estimates? Quarterly Journal of Economics 119 : 249 – 75 . Google Scholar CrossRef Search ADS BIS. 2015 . Central bank operating frameworks and collateral markets. Report. Committee on the global financial system markets committee. Bank for International Settlements . Borio, C., and Zhu. H. 2012 . Capital regulation, risk-taking and monetary policy: A missing link in the transmission mechanism? Journal of Financial Stability 8 : 236 – 51 . Google Scholar CrossRef Search ADS Brunetti, C., Filippo, M. D. and Harris. J. H. 2011 . Effects of central bank intervention on the interbank market during the subprime crisis. Review of Financial Studies 24 : 2053 – 83 . Google Scholar CrossRef Search ADS Carlson, M., Duygan-Bump, B. Natalucci, F. Nelson, W. R. Ochoa, M. Stein, J. and den Heuvel. S. V. 2016 . The demand for short-term, safe assets and financial stability: Some evidence and implications for central bank policies. International Journal of Central Banking 12 : 307 – 33 . Copeland, A., Davis, I. LeSueur, E. and Martin. A. 2014 . Lifting the veil on the U.S. bilateral repo market. Liberty Street Economics , July 9 . http://libertystreeteconomics.newyorkfed.org/2014/07/lifting-the-veil-on-the-us-bilateral-repo-market.html. Copeland, A., Duffie, D. Martin, A. and McLaughlin. S. 2012 . Key mechanics of the U.S. tri-party repo market. FRBNY Economic Policy Review 18 : 17 – 28 . Copeland, A., Martin, A. and Walker. M. 2014 . Repo runs: evidence from the tri-party repo market. Journal of Finance 69 : 2343 – 80 . Google Scholar CrossRef Search ADS de Andoain, C. G., Heider, F. Hoerova, M. and Manganelli. S. 2016 . Lending-of-last-resort is as lending-of-last-resort does: Central bank liquidity provision and interbank market functioning in the euro area. Journal of Financial Intermediation 28 : 32 – 47 . Google Scholar CrossRef Search ADS Di Maggio, M., and Tahbaz-Salehi. A. 2014 . Financial intermediation networks. Research Paper , Columbia Business School . Google Scholar CrossRef Search ADS Duygan-Bump, B., Parkinson, P. Rosengren, E. Suarez, G. A. and Willen. P. 2013 . How effective were the Federal Reserve emergency liquidity facilities? Evidence from the asset-backed commercial paper money market mutual fund liquidity facility. Journal of Finance 68 : 715 – 37 . Google Scholar CrossRef Search ADS Fleming, M. J., Hrung, W. B. and Keane. F. M. 2010 . Repo market effects of the term securities lending facility. American Economic Review 100 : 591 – 6 . Google Scholar CrossRef Search ADS FRBNY. 2009 . Operating policy: Statement regarding reverse repurchase agreements. https://www.newyorkfed.org/markets/operating_policy_091019.html. Frost, J., Logan, L. Martin, A. McCabe, P. Natalucci, F. and Remache. J. 2015 . Overnight RRP operations as a monetary policy tool: Some design considerations. Working Paper, Finance and Economics Discussion Series . Gorton, G. B. 2017 . The history and economics of safe assets. Annual Review of Economics 9 : 547 – 86 . Google Scholar CrossRef Search ADS Gorton, G. B., and Metrick. A. 2012 . Securitized banking and the run on repo. Journal of Financial Economics 104 : 425 – 51 . Google Scholar CrossRef Search ADS Greenwood, R., Hanson, S. G. and Stein. J. C. 2015 . A comparative-advantage approach to government debt maturity. Journal of Finance 70 : 1683 – 722 . Google Scholar CrossRef Search ADS Greenwood, R., Hanson, S. G. and Stein. J. C. 2016 . The federal reserve’s balance sheet as a financial stability tool. Jackson Hole Symposium : Federal Reserve Bank of Kansas City . Han, S., and Nikolaou. K. 2016 . Trading relationships in the OTC market for secured claims: Evidence from triparty repos. Board of Governors of the Federal Reserve System . Ihrig, J., Meade, E. and Weinbach. G. 2015 . Rewriting monetary policy 101: What’s the fed’s preferred post-crisis approach to raising interest rates? Journal of Economic Perspectives 29 : 177 – 98 . Google Scholar CrossRef Search ADS Infante, S. 2016 . Private money creation with safe assets and term premia. Working Paper, Finance and Economics Discussion Series . Google Scholar CrossRef Search ADS Kacperczyk, M., and Schnabl. P. 2013 . How safe are money market funds? Quarterly Journal of Economics 128 : 1073 – 122 . Google Scholar CrossRef Search ADS Kandrac, J. Forthcoming. The costs of quantitative easing: Liquidity and market functioning effects of Federal Reserve MBS purchases. International Journal of Central Banking . Khwaja, A. I., and Mian. A. 2008 . Tracing the impact of bank liquidity shocks: Evidence from an emerging market. American Economic Review 98 : 1413 – 42 . Google Scholar CrossRef Search ADS Krishnamurthy, A., Nagel, S. and Orlov. D. 2014 . Sizing up repo. Journal of Finance 69 : 2381 – 417 . Google Scholar CrossRef Search ADS Nagel, S. 2016 . The liquidity premium of near-money assets. Quarterly Journal of Economics 131 : 1927 – 71 . Google Scholar CrossRef Search ADS Parlatore, C. 2016 . Fragility in money market funds: Sponsor support and regulation. Journal of Financial Economics 121 : 595 – 623 . Google Scholar CrossRef Search ADS Stein, J. C. 2012 . Monetary policy as financial stability regulation. Quarterly Journal of Economics 127 : 57 – 95 . Google Scholar CrossRef Search ADS Sundaresan, S., and Wang. Z. 2008 . Y2K options and the liquidity premium in Treasury markets. Review of Financial Studies 22 : 1021 – 56 . Google Scholar CrossRef Search ADS Sunderam, A. 2014 . Money creation and the shadow banking system. Review of Financial Studies 28 : 939 – 77 . Google Scholar CrossRef Search ADS Woodford, M. 2016 . Quantitative easing and financial stability. Working Paper , NBER . Yellen, J. L. 2014 . Monetary policy and financial stability. Speech at the Michel Camdessus Central Banking Lecture, International Monetary Fund, Washington, DC , July 2 . https://www.federalreserve.gov/newsevents/speech/yellen20140702a.htm. Published by Oxford University Press on behalf of The Society for Financial Studies 2017. This work is written by US Government employees and is in the public domain in the US.

Journal

The Review of Financial StudiesOxford University Press

Published: Dec 18, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off