TY - JOUR
AU - Jin,, Kwon-Yong
AB - Abstract This article analyses the impact of asset and activity diversification on the stability of major financial institutions. Diversification is typically viewed as a positive element in risk management. However, examining recent examples concerning diversified multinational financial institutions and a theoretical model of failure risk facing them, this article demonstrates that under certain conditions, diversification can actually increase systemic risk. Financial conglomerates can be ‘too big to manage’, they can become too similar to each other and susceptible to coordinated failure, and, most importantly, catastrophic losses in one part of the firm can overwhelm the whole firm. Based on this finding, this article proposes a number of mitigation measures to limit intra-firm spillover and to make the resolution of troubled financial institutions smoother. I. INTRODUCTION Perhaps no concept in modern finance theory has had as far reaching impact as diversification has.1 Analogous to the old adage—ʻDon’t put all your eggs in one basketʼ—diversification affords risk-averse investors protection against certain downside risk and has been instrumental in portfolio theory. In particular, the concept of diversification works as follows: a portfolio composed of a single asset is subject to both systematic risk and non-systematic risk.2 Systematic risk, often called market risk, refers to risk that the entire market faces—General Electric and Apple alike—such as the risk of a generalized decline in the economy.3 Non-systematic risk, also called firm-specific risk, refers to risk that is unique to a particular asset; in the case of equity securities, non-systematic risk may include the possibility that the issuer would underperform compared to its peers.4 Thus, a portfolio composed of a single asset faces both risks: the risk that the market as a whole would decline and the risk that the asset itself would underperform compared to the market. However, when the number of assets in a portfolio increases and the non-systematic risk of each asset is uncorrelated, the overall non-systematic risk facing the portfolio decreases. When one asset depreciates due to non-systematic factors, another might appreciate, offsetting the price decline. As a result, sufficient portfolio diversification would reduce non-systematic risk, leaving only systematic (or market) risk. When all volatility is due to non-systematic risk, overall risk and volatility would decrease to miniscule levels with extensive diversification, to the benefit of a risk-averse investor.5 Traditional finance theory tends to view diversification in a positive light.6 Given the risk aversion of the average investor, diversification allows investors to manage their portfolio volatility and risk. Thus, diversification is considered a crucial tool in an investor’s arsenal, one that forms the core of portfolio selection decisions. This positive attitude has expanded beyond portfolio theory, permeating the perception of the stability of financial institutions as well. The markets and the regulators have a tendency to view diversification as a tool that decreases the risk that a particular financial institution faces, thus improving systemic stability. Just as diversification reduces the risk of catastrophic loss for an individual investor, diversification would also reduce the risk of catastrophic loss for a major financial institution.7 This article challenges that assumption. While diversification may serve an important role in mitigating risk for individual investors and financial institutions alike, this article notes five major drawbacks to diversification of financial institutions’ assets and activities. Three of these relate to the increased ex ante probability of failure. Financial conglomerates can become so complex that intra-firm information transmission and internal control mechanisms fail, and as they get more diversified, they can also become more and more similar to one another, with such asset correlation threatening systemic stability. Moreover, diversification can increase the number of potential failure points within a firm and corresponding increase in failure probability of individual financial institutions. In particular, this article’s main contribution to the literature is applying case studies and finance theory to argue that certain assets do have highly asymmetric return structures and that, in the presence of such a return structure, diversification can undermine the stability of a financial conglomerate because catastrophic losses in one part can engulf the whole firm, including other healthy parts. If and when financial conglomerates suffer distress, diversification has repercussions for post-distress resolution as well. First, diversified financial institutions can be more difficult to resolve than undiversified ones, making a disorderly failure more likely. Second, when a government safety net is deployed to prevent such a disorderly failure, diversification increases the likelihood that such government support is not directed in the manner desired by regulators. In order to mitigate these ex ante and ex post destabilizing effects of diversification, this article proposes measures to limit the intra-firm spillover of risk, such as monoline requirement and ringfencing. The rest of this article proceeds as follows. Section II addresses the various types and dimensions of diversification in the financial industry. Section III lays out several examples of regulators’ positive attitudes toward diversification of financial institutions. Section IV then outlines five ways in which diversified financial institutions can be more unstable than their undiversified peers, both from ex ante and ex post perspectives. Section V then discusses monoline requirement and ringfencing as a way to remedy the destabilizing effects of diversification; the Conclusion follows. II. VARIOUS DIMENSIONS OF DIVERSIFICATION IN FINANCIAL INSTITUTIONS Unlike simple diversification in individual stock portfolios, diversification can exist in many different forms in financial institutions, some of which are more relevant to the discussion in this article than others. Thus, this section classifies the various dimensions of diversification in financial institutions and the relevance of each to the discussion. The most obvious way to classify types of diversification is along balance sheet classifications—that is, the asset, liability, and equity-sides. Asset-side diversification can refer to those that affect the composition of and changes in a financial institution’s assets. These can include asset diversification, business line (revenue) diversification, and geographic diversification; the first affects the asset side of a financial institution’s balance sheet directly, and the latter two can affect a financial institution’s assets by flowing through its profits. On the liability and equity side, diversification can refer to the distribution of a financial institution’s sources of and types of equity and debt funding. These can vary along a maturity spectrum (short-term vs long-term), along types of funding (debt vs equity), along types of equity (common vs preferred), and along types of debt (deposit vs wholesale funding).8 The mix of a financial institution’s funding sources—whether a firm funds itself through a limited number of sources or through various sources, whether a firm’s funding sources consist largely of debt funding or equity funding—is indeed an important concern from the policy perspective as well. However, the main concern of this article is with how the return profile of a financial institution’s assets affects the probability of its failure and systemic risk, so discussion of a financial institution’s liability and equity diversification is beyond the scope of this article. Therefore, this article will focus largely on asset-side diversification, including asset and business line diversification. Among the various types of asset-side diversification, the distinction is more technical than substantive. For example, if a financial institution provides financing to a diverse array of debtors, it may be called ‘credit diversification’. If a financial institution holds assets across a number of asset classes, it may be called ‘asset diversification’. And lastly, if a financial institution diversifies its sources of income across a number of business lines—e.g., interest margin, trading income—it may be called ‘business line diversification’. But the dividing line between these different categories can often be blurred, and these categories can overlap as well. If a financial institution holds emerging market fixed income securities, US equities, and commodity derivatives, should this be called asset diversification or business line diversification? To the extent that a financial institution’s assets are a key component of its revenue generation, distinguishing among various types of asset-side diversification can be difficult to do in practice. In fact, diversification may occur in a number of different manners, both direct and indirect, all with similar effects. The most intuitive manner in which a financial institution diversifies its business can encompass holding different assets, generating revenue from different sources, and investing in different geographic areas. However, risk-sharing among financial institutions also results in diversification of a firm’s risk profile. A firm may hold only one asset directly, but if it enters into a swap with another bank whereby it trades a portion of its portfolio risk with the counterparty’s, the firm has effectively diversified its holdings.9 Thus, particularly for the purposes of this article, what is important is that, regardless of the name attached or the manner in which diversification of a financial institution’s business occurs, these types of asset-side diversification all lead to ‘diversification of risk’; the adverse repercussions from a particular source of risk may be smaller, but the number of sources of risk increases. III. THE GOSPEL OF DIVERSIFICATION As noted above, diversification is typically viewed as a positive element in risk management. This section thus lays out the examples in which the industry and/or the regulators adhered to such a ‘gospel of diversification’; that is, the belief that diversification reduces (or at least spreads out) risk. According to this belief, diversification would enhance systemic stability by making it unlikely that risk would be concentrated in any given firm. Many banking and securities regulations, both promulgated before and after the recent financial crisis, are based on this belief and encouraged diversification of financial institutions. 1. Gramm–Leach–Bliley Act and the repeal of the Glass–Steagall Act The most prominent example of the belief that financial conglomerates are more stable than monoline financial institutions is the enactment of the Gramm–Leach–Bliley Act and the repeal of the Glass–Steagall Act in 1999.10 Enacted in 1933 in response to the stock market crash of 1929, the Glass–Steagall Act mandated separation of commercial banking and investment banking activities in four different sections.11 First, Section 16 of the Glass–Steagall Act prescribed that national banks may not purchase or sell securities except for their clients’ account and that they may not underwrite securities issuance.12 Second, Section 20 prohibited Federal Reserve member banks from being affiliated with a firm engaged principally in securities underwriting.13 Third, Section 21 prohibited firms engaged in securities underwriting from receiving deposits like banks.14 Lastly, Section 32 prohibited management interlock between a Federal Reserve member bank and a securities dealer.15 These four provisions split commercial banking from investment banking, reflecting the belief that such a separation would prevent a recurrence of the 1929 stock market crash. Although the Glass–Steagall Act’s separation of commercial banking and investment banking received industry challenge and was chipped away in part in the ensuing decades, it formed the basic framework governing the structure of American financial institutions over the next 60 years.16 However, that framework was ultimately demolished in 1999 with the passage of the Gramm–Leach–Bliley Act, which repealed Sections 20 and 32 of the Glass–Steagall Act and permitted commercial banks to engage in securities underwriting activities.17 In addition, the Gramm–Leach–Bliley Act created ‘financial holding companies’ and permitted bank holding companies meeting certain criteria to elect to become financial holding companies.18 Previously, bank holding companies were restricted to ‘activities … so closely related to the business of banking … as to be a proper incident thereto’.19 For bank holding companies that elect to become financial holding companies, the Gramm–Leach–Bliley Act expanded that scope of permissible activities; financial holding companies could now engage in activities that are ‘financial in nature’, such as securities underwriting and investment advisory activities.20 These provisions significantly expanded the scope of activities that a large commercial bank could engage in, either directly or through its affiliates; the era of ‘broad banking’ was back. One of the primary justifications for the Gramm–Leach–Bliley Act’s enactment was that diversified financial institutions would be more profitable and less susceptible to sudden failures. In particular, the proponents of the Gramm–Leach–Bliley Act focused on how a financial institution’s revenue diversification lowers its risk of failure and improves its safety and soundness. In 1991, the Treasury Department issued a study that pushed for repeal of the Glass–Steagall Act, arguing that when financial institutions are able to offer a wider variety of services and products, the resulting ‘more stable stream of income across financial services firms would contribute to the overall stability of financial markets’.21 Likewise, Treasury Secretary Robert Rubin noted that the Glass–Steagall Act’s restrictions could ‘conceivably impede safety and soundness by limiting revenue diversification’,22 and Congressional deliberation regarding the repeal of the Glass–Steagall Act echoed that sentiment.23 Thus, the enactment of the Gramm–Leach–Bliley Act and the expansion in the scope of permissible activities for bank holding companies reflected the belief that financial conglomerates, by virtue of their revenue diversification, would be subject to less volatile income stream and therefore face lower risk of failure. 2. Basel II’s treatment of operational risk Basel II regulatory framework, first adopted in 2004 (prior to the 2008 financial crisis) but since superseded by the more stringent Basel III framework adopted in 2010, set international standards for minimum bank capital requirements. It consisted of three pillars: Pillar I (minimum capital requirements), Pillar II (supervisory review), and Pillar III (market discipline).24 Pillar I set forth the methods for calculating the minimum capital requirement for banking organizations. Notably, the minimum capital requirement under Pillar I was to be ‘risk-sensitive’, that is, more tailored to each banking organization’s risk profile.25 Pillar II proposed guidelines and principles to be applied when national regulators review and assess whether banking organizations’ capital is adequate. Lastly, Pillar III set forth the disclosure requirement for banking organizations such that market participants can assess the business and risk profiles of those banking organizations; in turn, the market participants, armed with information about banking organizations’ business and risk profiles, can ‘discipline’ those banking organizations with inadequate capital. In setting the ‘risk-sensitive’ capital requirements, Basel II’s Pillar I addressed three distinct risks—credit risk, operational risk, and market risk.26 Credit risk refers to the risk of a credit counterparty failing to comply with its obligations; the simplest example would be that of a borrower failing to make payments on his/her loan.27 Operational risk refers to the ‘risk of loss resulting from inadequate or failed internal processes, people and systems or from external events’.28 Market risk refers to the risk of loss from adverse change in market prices.29 Based on the aggregate risk profile and taking into account each of these three categories, Pillar I of the Basel II framework set forth the applicable minimum capital requirements.30 In calculating the minimum capital requirement for operational risk, the Basel II framework permitted one of three methods: the Basic Indicator Approach, the Standardized Approach, and the Advanced Measurement Approach (AMA). Under the simplest Basic Indicator Approach, a bank must hold a fixed percentage (15%) of its average annual gross income (excluding zero or negative gross income) for the past three years.31 The Standardized Approach was similar in that it was pegged to the average annual gross income, but was more sophisticated than the Basic Indicator Approach in that gross income was broken down into eight different business lines, each of which carried a different multiplier (ranging from 12% to 18%).32 Under the most complex Advanced Measurement Approach, banking organizations used their internal models to calculate the required capital that reflected their respective risk profiles. Thus, unlike the Basic Indicator Approach and the Standardized Approach, the Advanced Measurement Approach was intended to reflect the risk profile of each banking organization more precisely.33 Of particular note was the Advanced Measurement Approach’s treatment of diversification: under this approach, banking organizations could, subject to supervisory approval, incorporate a ‘well-reasoned estimate of diversification benefits … at the group-wide level or at the banking subsidiary level’.34 The assumption underlying this approach was that the operational risk of a banking organization has a subadditive property: that is, the aggregate operational risk of a diversified banking organization is less than the sum of the operational risks of the individual parts of that diversified banking organization. In essence, it was assumed that risks from different parts of a diversified banking organization would offset each other and reduce the aggregate risk. Thus, this assumption of subadditivity and favourable treatment in the calculation of minimum capital requirements under Basel II rewarded those financial institutions with diversified operations. 3. Single counterparty exposure limit for systemically important financial institutions The recent financial crisis and the failure of Bear Stearns (although it was ‘saved’ through its merger with JPMorgan Chase) and Lehman Brothers highlighted the importance of heightened prudential regulation for large financial institutions whose failure could threaten the entire financial system. In response, the Dodd–Frank Act established the Financial Stability Oversight Council (FSOC), composed of the heads of various federal banking and securities regulators.35 The Dodd–Frank Act tasked the FSOC with a mission to identify and guard against risks to the financial stability of the United States and, as part of this mission, authorized designation of systemically important financial institutions (SIFIs).36 These SIFIs are subject to heightened prudential regulation and supervision by the Federal Reserve, including higher capital and leverage requirements.37 Pursuant to its Section 165 obligations, therefore, the Federal Reserve has proposed numerous new regulations applicable to SIFIs, such as risk-based capital and leverage requirements, liquidity standards, and requirements for overall risk management.38 Most importantly for the purpose of this section, the Federal Reserve also adopted single counterparty exposure limits.39 Under these limits, a covered company’s net exposure to a single counterparty is capped at 25% of its tier 1 capital.40 In addition, bank holding companies identified as ʻglobal systemically important bank holding companiesʼ face even more stringent single counterparty exposure limits; their net credit exposure to any single major counterparty is capped at 15% of tier 1 capital.41 As a result, these single counterparty exposure limits for large financial institutions force these institutions to diversify their credit exposure. The Federal Reserve’s rationale for proposing the single counterparty exposure limits was preventing a shock in one part of the financial system from spreading into other parts and turning into a system-wide catastrophe. It noted that during the recent crisis, interconnectedness of major financial institutions served to propagate adverse shocks throughout the financial system, as one institution’s failure and inability to meet its obligation meant that its counterparty could not meet its own obligations either.42 Therefore, the Federal Reserve implied that preventing concentration of credit exposure and mandating asset diversification would spread out risk and enhance stability of the financial system by preventing one SIFI’s failure from snowballing into a systemic threat. 4. Summary The above examples all reveal a common pattern. First, banking and securities regulators often believe that exposure to a diversified set of risks would lower the probability of failure of a regulated financial institution compared to exposure to a small number of risk sources. Whether it be diversification of permissible activities (as in the Gramm–Leach–Bliley Act) or of counterparty credit exposure (as in single counterparty exposure limits for systemically important financial institutions), banking and securities regulators tend to view diversification in a very favourable light. Second, as a result of this belief, many regulations steer financial institutions toward diversification of activities and assets. Thus, given the pervasiveness of the belief that diversification reduces systemic risk, it is important to now consider whether this belief is indeed justified. IV. DESTABILIZING EFFECTS OF DIVERSIFICATION Section III laid out the status quo of the relationship between diversification and banking regulation, and this section questions whether that status quo is indeed optimal from a systemic risk perspective. In particular, this section will address five potentially destabilizing effects of diversification: first, strains on internal control mechanism of diversified financial institutions; second, dangerous asset correlation among systemically important financial institutions and increase in joint failure probability; third, increase in the probability of failure of individual financial institutions as a result of asymmetric asset return profiles; fourth, difficulty of resolving a large, diversified financial institution without causing systemic failure; and lastly, unjustified expansion of the federal safety net. The first point discusses from an organizational perspective why diversification can undermine the stability and risk management of a financial firm. The second and third points show that diversification can actually increase, rather than decrease, the failure probability of financial institutions. The fourth and fifth points then address the risk that the failure of a diversified financial institution would be disorderly and that a federal safety net deployed to prevent such a disorderly failure may expand beyond the policymakers’ intended scope. These five adverse effects of diversification imply that diversification may not improve the stability of financial institutions, both individually and collectively. 1. Strains on internal control The first adverse effect of diversification inside a major financial institution is that such diversification can put a significant burden on the institution’s internal control mechanism, leading to laxity in risk and compliance control. In many ways, the internal control mechanism of a major financial institution can be considered an essential asset. Any financial institution, however small, must have sufficient internal control mechanisms to ensure that its risks are well managed and its operations are conducted in compliance with relevant law. Without such an internal control mechanism, the financial institution takes on significant risk: the risk that a rogue employee breaches applicable laws and regulations (compliance failure), the risk that its credit exposure makes it vulnerable to sudden failure (risk management failure), and so on. As a financial institution increases in size, its internal control mechanism must correspondingly increase in size and complexity as well. At the most basic level, the compliance and risk management staff headcount must increase to manage the firm’s expanding operations. Even beyond the headcount, the sophistication of the internal control mechanism must keep pace with the sophistication of the firm’s operations; after all, a diversified multinational financial institution faces qualitatively different risks compared to a small regional commercial bank. However, as a financial institution increases in size and complexity to become a multinational behemoth, it reaches a point at which its internal control mechanism cannot keep pace with its expanding size, breadth, and complexity of operations. At this point, regardless of the headcount of the compliance and risk management staff, the firm’s internal controls can fail, exposing the firm to compliance and risk management threats. One factor contributing to this result is that top management is a fixed asset. A firm can increase the size of its internal control staff, but the information processed by the internal control team must also be processed by the firm’s top management. Unfortunately, a firm cannot simply increase the number of its top officers and directors to meet the informational demands on its top management. The Chief Executive Officer of a diversified multinational financial institution must process information that is much more complicated and much more voluminous than the Chief Executive Officer of a small regional bank must, and such a burden on top management can lead to cracks in internal control. This idea that top management is subject to bounded rationality and may not fully comprehend the risks facing the executive’s financial institution is in essence the reverse of the idea that organizational loss of control across successive hierarchies sets a limit to firm size.43 Under this idea, subordinates are only able to fulfil a fraction of his/her immediate superior’s intentions (loss of control), and the accumulation of this loss of control sets the maximum number of hierarchies in an organization and therefore the organizational size.44 The Williamson theory focused on the loss of control as a result of incomplete communication from superior to subordinate, but in a risk management context, this loss of control can work in reverse in the form of incomplete communication and/or incomplete processing of communication from subordinate to superior. As risk management information is delivered from the lowest ladders of the organizational hierarchy—the front lines—to a financial institution’s top executives, bounded rationality requires simplification and trimming of information, which acts as ‘loss of control’. As this upward loss of control accumulates, the risk management practices of a hierarchical financial institution can become unstable. Given this upward loss of control framework, why would a diversified financial institution be more susceptible to upward loss of control than an undiversified financial institution of the same size and hierarchy? To resolve this question, I argue that diversification itself contributes to an increasing loss of control. When an executive receives information from multiple sources that are qualitatively similar to one another, that information can be aggregated without significant loss. Moreover, that executive may develop expertise in processing information flowing upward, further reducing the loss of information. On the other hand, when an executive receives information from multiple sources that are qualitatively distinct from one another, those multiple pieces of information must be crammed into a uniform framework, which inevitably requires simplification and trimming, and the executive may also be required to process information from an area that he/she has little expertise in. For example, as insurance giant AIG’s credit default swap business boomed, its executives had incomplete information on the burgeoning business, such as the exact terms of the credit default swaps that AIG was writing and selling; this meant that AIG’s senior executives were caught unprepared when its credit default swap business began to burden its liquidity position.45 Thus, even with a similar level of hierarchy and size, diversification can contribute to loss of information flowing upward, further undermining a firm’s risk management. This idea of loss of control from lower levels of a financial institution’s hierarchy to the upper levels can be generalized. A diversified financial institution is an agglomeration of different—but interconnected—segments and units, each of which can pose risk to the firm’s stability. Moreover, sound risk management of a diversified financial institution requires seamless and adequate sharing of information from one division to another; if the risk management division of a financial institution does not receive accurate information from its trading divisions, for example, the firm’s risk management practices will not reflect its risks correctly. But not all types of information can be transmitted accurately and fully from one part of a financial institution to another. Some types of information are ʻhardʼ—information that is quantifiable and/or otherwise objective. Other types of information are ʻsoftʼ—information that is more subjective and therefore less verifiable. It has been shown that in the context of financial institutions, hard information is better transmitted in a large, hierarchical organization, while soft information cannot; as a result, soft information is more suitable for decentralized organizations, whereas hard information is more suitable for centralized, hierarchical organizations.46 Thus, for sound enterprise-wide risk management in a diversified financial institution, which requires that information be passed on efficiently from all parts of the firm to its risk management division and vice versa, information must be ‘hardened’.47 The subjective realities of a financial institution’s diversified business must be compressed and depicted such that the depictions can be relayed across division boundaries. Yet, as financial institutions grow, they can become ‘too complex to depict’.48 The depiction tools at the disposal of major financial institutions—accounting tools, for example—are often insufficient to fully and accurately depict the true characteristics of complex financial instruments.49 For example, on a day-to-day basis, executives often rely on summarized information, such as value-at-risk measurements, to monitor their businesses, which are mere ‘depictions’ of the actual reality.50 As a result of this ‘too complex to depict’ problem, there is a limit to how reliable ‘hardened’ information can be, leading to gaps in information transmission and sound risk management. Worse, aside from the issue of complexity and limits of depiction tools, in any large organization—with financial institutions being no exception—there can be siloing of information that leads to diseconomies of scale in internal control. In a multinational financial empire, it is difficult to ensure free flow of information among different divisions and departments,51 and internal control is no exception. As a financial institution increases in size and complexity, the internal control department may sound the alarm about a particular activity’s riskiness, but that alarm may not be heard by other parts of the organization. Likewise, traders may not be communicating fully and openly with their managers, the managers may not be communicating fully and openly with the firm’s top management, and the top management may not be communicating fully and openly with the board of directors. One part of the financial institution may know the true extent of the risks of a particular business line or financial instrument, but the institution as a whole may not.52 In some cases, this siloing may even be intentional.53 As a financial institution increases in size and complexity, the risk of failure in information transmission increases as well. Large, diversified financial institutions are not monolithic creatures but organizations that hire hundreds of thousands of employees across hundreds of businesses and divisions in different countries. As a result of this size, diversity, and complexity, even within the firm, information may not flow freely from one part to another, and even the information that is transmitted may not be entirely accurate and adequate. This limitation to intra-firm information flow can undermine the firm’s enterprise risk management and internal control, and as the internal control system of a major financial institution becomes stretched thin, it is less able to guard against compliance and risk management failures.54 a. The Banamex scandal The prime example of this strain on internal control and risk management in a large, diversified financial institution is the 2014 fraud scandal involving multinational financial giant Citigroup’s Mexican subsidiary, Banamex (short for Banco Nacional de Mexico). In early 2014, it was revealed that Banamex, one of the largest banks in Mexico, had suffered losses of as much as $400 million as a result of fraud by one of its borrowers, forcing Citi to cut its 2013 fourth quarter earnings by $235 million.55 Banamex had extended $585 million in short term credit and $33 million in letters of credit and direct loans to Oceanografia SA de CV, a Mexican oil services company. Oceanografia was a significant supplier to Pemex, Mexico’s state-owned petroleum company, and its $585 million in short term credit was secured by its accounts receivable.56 However, when Citigroup and Banamex reviewed Oceanografia’s accounts receivable, Pemex informed them that ‘significant portion of the accounts receivable … were fraudulent and that the valid receivables were substantially less than … $585 million’.57 Banamex estimated that valid receivables were only worth $185 million, and as a result, it suffered $400 million in losses, forcing its parent Citigroup to restate its 2013 earnings. The Banamex scandal revealed substantial deficiencies in the Mexican bank’s internal control mechanism. Banamex employees, instead of demanding official invoices as proof of the accounts receivable, were satisfied with progress reports and work orders.58 In addition, Banamex managers and employees did not even bother to confirm the validity of the accounts receivable with Pemex, an incredible laxity in risk management considering the size of the loans made to Oceanografia.59 After the scandal, Mexican regulators noted ‘weaknesses in Banamex’s internal controls[,] errors in loan origination and administration[,] and deficiencies in contracts, risk administration and internal audit functions’.60 What’s worse, the Banamex scandal also showed the strains on Citigroup’s—in addition to Banamex’s—internal controls as it grew to become a multinational financial giant. Banamex was acquired by Citigroup in 2001 and is the biggest unit in Citigroup’s Latin American business, but its full integration into Citigroup’s internal control system had continually been delayed, even a decade after its acquisition.61 Banamex resisted its parent’s efforts to upgrade Banamex’s internal controls to match Citigroup’s overall standards, leading to laxity in internal control. For example, a number of failures in information transmission between different divisions of Citigroup and Banamex were found in the aftermath of the fraud. Credits of the type extended to Oceanografia were provided through two banking divisions within Citigroup: Consumer and Commercial Bank (CCB) and Institutional Clients Group (ICG).62 CCB provided credit to smaller clients, while ICG provided credits to much larger clients. As Banamex’s relationship with Oceanografia grew, in 2012, Oceanografia was transferred from CCB to ICG and designated a target client.63 At the time of the transfer, CCB had assessed Oceanografia negatively, including potential reputational risk, and was actually looking to reduce its exposure to Oceanografia.64 However, CCB’s negative assessment of Oceanografia as a borrower was not—and was not required to be, under Citigroup’s policies at the time—relayed to ICG when the client was transferred, and as a result of this gap in communication, while CCB had sought to reduce Banamex’s exposure to Oceanografia, ICG sought to increase Banamex’s exposure to the same client.65 There were other warning signs, including a publicly available report that Oceanografia had defrauded another Mexican bank of $30 million and another Citigroup subsidiary refusing to open a bank account for an Oceanografia executive citing potential reputational risks, but these pieces of information were not shared upward with Citigroup’s Treasury and Trade Services (TTS) business, which oversaw the credit program.66 While different parts of Citigroup were aware of the dangers and risks concerning Oceanografia, the warning signs were lost in the gap between various Citigroup divisions. Furthermore, Citigroup had a number of internal policies and procedures, many of which were not followed by its subsidiaries and divisions, evidencing vertical loss of control. According to Citigroup’s internal policies, the credit extended to Oceanografia based on its receivables from Pemex, since the ultimate obligation rested with Oceanografia, should have been classified as credit risk to Oceanografia (so-called ‘seller centric’ credit risk).67 However, partly as a result of assertions made by the employees responsible for the credit extension, the credit was improperly classified as credit risk to Pemex (so-called ‘buyer centric’ credit risk), enabling Banamex to increase its credit line substantially since Pemex was a more creditworthy client.68 Likewise, Citigroup had a Global Deal Review (GDR) process in effect to ensure that ICG’s capital was deployed appropriately, but the Oceanografia credit was submitted to GDR only three times.69 Thus, Citigroup’s policies, in effect, were not followed by the front line divisions, indicating the degree of ‘loss of control’ across hierarchies in a financial conglomerate like Citigroup. Oceanografia’s fraud against Banamex was not an unpredictable event. There were a number of warning signs, including CCB’s own negative assessment of Oceanografia, but these were relayed neither to ICG, the division newly in charge of Oceanografia, nor to higher level (TTS) at Citigroup. As predicted above, this gap in information transmission contributed to a weakness in Citigroup’s internal control and illustrates the increased possibility of risk management and compliance failure as a financial institution diversifies and becomes larger, more complex, and more geographically dispersed. 2. Increase in joint failure probability The second source of systemic risk arising out of diversification is the likelihood of coordinated failure among the largest, most diversified financial institutions. When all financial institutions in a system are undiversified—take the extreme case of each financial institution holding only one asset—an idiosyncratic shock to one asset can easily lead to the failure of the individual institution holding that specific asset. However, since there is little asset correlation among financial institutions, the probability of a widespread failure of multiple financial institutions as a result of the idiosyncratic shock is minimal. For a simplified example, assume the following: first, there are three financial institutions in the system; second, there are three assets in the system (e.g., bonds, stocks, and commodities) that are uncorrelated; and lastly, each institution only holds one asset respectively. In this example of extremely undiversified financial institutions, if there is an idiosyncratic shock adversely affecting stock prices, the financial institution holding stocks would fail, but none of the remaining two financial institutions would be affected (since they hold bonds and commodities, respectively). While an isolated failure of a single financial institution poses consumer protection and other related concerns, such a failure is hardly a systemic risk problem. In contrast, when all financial institutions are diversified, the failure probability of any given individual institution may be lower, but the probability of joint failure can increase. As financial institutions diversify, their portfolios would gradually resemble each other, and when an adverse shock hits one of the institutions, that shock would likely affect other institutions in the system as well. Revisiting the simplified example from the previous paragraph, now assume that the asset portfolio of each of the three financial institutions in the system is composed of one-third stocks, one-third bonds, and one-third commodities. In this scenario, if there is an isolated shock in the stock market that is adverse enough, all institutions in the system will fail. Moreover, when financial institutions diversify through loss mutualization—e.g., providing insurance to each other—the direct linkage between multiple financial institutions means that when one firm incurs a significant loss, that loss may directly impair the balance sheet of other firms in the system, putting them in danger of failure as well.70 Such a simultaneous failure would lead to a systemic crisis, similar to the one the United States experienced in 2008. If, in an undiversified system, one institution fails by a large margin and another remains robust, forcing diversification and mixing the two institutions would lead to both institutions failing by a small margin (but failing nonetheless).71 As financial institutions diversify and begin to resemble one another, even if the common shock does not bring them all down at the same time, the appearance of similarity may endanger them as well. Financial institutions, since they rely heavily on short-term funding, depend on the decisions of their creditors. Under normal circumstances, creditors—depositors, holders of commercial paper, and other lenders—do not have concerns about the solvency of the debtor financial institution and therefore have little reason to cut off their funding (e.g., by refusing to roll over their credit). When one financial institution is subject to an adverse shock and fails, that failure puts creditors on alert. Particularly if there is imperfect information about the solvency of the financial institutions in the system, creditors are left ‘guesstimating’ the solvency of other financial institutions still standing based on the profile of the firm that failed.72 As a result, creditors are more likely to pull funding from financial institutions that resemble the failed firm, even if the surviving financial institutions are actually solvent. Once their funding evaporates, the surviving financial institutions are forced to liquidate their assets rapidly, impairing their balance sheet and potentially (and likely) leading to their failure. Thus, as financial institutions diversify and become similar to one another, they become more informationally linked.73 Failure of one institution can then lead to runs at other institutions, creating a series of failures. This increase in joint failure probability can be particularly problematic for two reasons. First, the welfare reduction from coordinated failure of multiple financial institutions can far outweigh the welfare reduction from idiosyncratic failure of a single financial institution. When one firm fails in the system and others remain healthy, the failed firm can be resolved relatively easily, as the funding market remains robust. If necessary, the healthy firms in the system can absorb the failed firm’s balance sheet. In contrast, when multiple firms fail at the same time, simultaneous liquidation of their balance sheets can depress the market price for their assets rapidly, further exacerbating their crises, and there may be considerable difficulty finding buyers for the troubled financial institutions as well. As a result, the coordinated failure of multiple financial institutions will be harder to contain. Second, financial institutions internalize the costs of their own failure but not the costs of other firms’ failures. To the extent that there is an ‘optimal’ level of diversification, the fact that the increase in joint failure probability is an externality means that each financial institution would over-diversify, since the costs they internalize—of their own failure—are reduced and the systemic costs they do not internalize increase.74 3. ‘Weakest link’ and the increase in individual failure probability The last section assumed—too simplistically, as it will be shown—that diversification reduces the individual failure probability of a specific financial institution but increases the joint failure probability of the financial system as a whole. This section, however, challenges even the first part of that assumption: that diversification reduces the individual probability of failure. The breakdown of that assumption would lead to the troubling conclusion that diversification may not succeed at reducing either the individual probability of failure or the joint probability of failure and throw into doubt the belief that diversification enhances systemic stability. The starting point of this analysis is the intuitive notion that ‘a chain is only as strong as its weakest link’. A similar notion applies to financial institutions as well; one part of a financial institution can incur losses that engulf the whole firm and lead to firmwide failure, even if all other parts of that institution are healthy. As a corollary, a financial institution becomes more vulnerable to sudden failure as it diversifies its assets and activities, and possible failure points become more numerous. a. Model of a financial institution with a diversified portfolio To illustrate the possibility that diversification can increase a financial institution’s probability of failure, we present a simplified model of a financial institution with assets that have highly asymmetric returns. First, assume that there is only one possible type of investment asset in the economy. One unit of that asset can yield the following payoff: gain of $100 with 99% probability (‘good scenario’) or loss of $9,000 with 1% probability (‘bad scenario’). Thus, the expected payoff of one unit of this asset is positive $9. Second, assume that an investor can split his/her investment in this asset into multiple fractional units and that the returns of each fractional unit of the asset are uncorrelated. Thus, an investor can choose to invest in one whole unit of the asset, two one-half units of the asset, three one-third units of the asset, and so on. When, for example, the investor chooses to invest in two one-half units of the asset, the returns of each half unit are uncorrelated to each other. Lastly, there is one financial institution in the system, and that financial institution can invest up to one unit of the asset.75 Thus, the only decision that the financial institution can make is how to allocate its investments: whether to invest in one whole unit of the asset, two one-half units of the asset, three one-third units of the asset, and so on. This financial institution is at the lower bound of its capital requirement, so if it suffers any loss, it will automatically fail. With these assumptions, we note an ironic result. As the firm increases the number of its uncorrelated investments, its failure probability actually increases. When it invests in one whole unit of the asset, it will fail if the bad scenario occurs, so its failure probability is 1%. When it invests in two one-half units of the asset, it will fail if the bad scenario occurs in either asset, so its failure probability is 1.99%.76 Similarly, when it invests in three one-third units of the asset, it will fail if the bad scenario occurs in any of the three assets, so its failure probability again increases to 2.9701%. If the number of positions increases to five, the probability of failure goes up to 4.90%, and at ten positions, the probability of failure increases to 9.56%. This counter-intuitive result arises because of the highly asymmetric payoff profile of the investment asset. The payoff in the bad scenario is so negative that if the bad scenario occurs with respect to any one fractional asset, the financial institution automatically fails, regardless of the payoff from other fractional assets.77 The benefit of diversification is that when one asset decreases in value, returns of other assets will offset the negative return and reduce the probability of a catastrophic loss. However, in this case, that benefit of diversification is lost because no good scenario can offset any bad scenario. Adverse scenario in any one asset is catastrophic. Thus, as the firm diversifies and increases the number of its investment positions, its failure probability can actually increase. b. Examples of asymmetric return profiles Of course, the counterintuitive result in the model presented above depends on the asymmetry of investment returns for the asset. If the asset’s payoff profile is more symmetric, diversification would not increase the failure probability of the financial institution. In the financial institution context, however, it is not difficult to find examples of assets and businesses with highly asymmetric returns. In this section, we survey various examples of such assets and businesses. The most basic form of a business with asymmetric return is selling an option. An option contract permits the buyer the right, but not the obligation, to purchase from or sell to the seller of the option contract the underlying asset at a fixed price (the so-called ‘strike price’) during a predetermined period.78 In exchange for this right, the buyer pays the seller a ‘premium’; in return for this premium, the seller of the option must fulfil his/her corresponding obligation to sell to or purchase from the option buyer the underlying asset at the strike price.79 An option contract giving the option buyer the right to purchase the underlying asset at the strike price is a ‘call option’, whereas an option contract giving the option buyer the right to sell the underlying asset at the strike price is a ‘put option’. Furthermore, an option contract whose strike price is below the current price (for a call option) or above the current price (for a put option)—that is, an option contract whose call or put right is worth exercising—is called ‘in-the-money’; the reverse (an option contract whose call or put right is currently not worth exercising) is called ‘out-of-the-money’. It is intuitive how the seller of an option faces an asymmetric return profile, particularly if the option is out-of-the-money.80 The seller’s return is at most the premium he/she gets for selling the option contract. For an out-of-the-money option, the premium can be quite small since the probability of the buyer being able to exercise the option is low. In this case, with high probability the option will expire out-of-the-money, and the seller will earn the premium without having to fulfil his/her obligation to the option buyer. On the other hand, if the price of the underlying asset moves significantly in an unfavourable direction for the seller—e.g., for a call option, the underlying asset’s price skyrockets—the seller’s loss can balloon, potentially without limit.81 Thus, the seller of an out-of-the-money option faces a low positive return with high probability and a high negative return with low probability.82 More broadly speaking, most forms of insurance also have an asymmetric return profile as well. The insurer’s upside is capped at the insurance premium, which is often small compared to the amount of the covered loss. With high probability the covered loss will not arise, and the insurer will earn a small income. However, with a small probability, the covered loss will arise, and the insurer will incur a large loss. One of the contributing factors to the 2008 financial crisis, credit default swaps (CDSs), illustrates this asymmetry. A CDS is a derivative contract whose buyer pays the seller a regular ‘spread’. In return, when a ‘credit event’ occurs with respect to the reference obligation, the seller of a CDS contract is obligated to pay the buyer a set amount to make the buyer whole.83 Typically the credit event can include bankruptcy, default, and other failures to make payment on the reference obligation (corporate bond, mortgage-backed securities, etc.). Upon the occurrence of the credit event, the CDS contract can be physically settled (the buyer delivers the impaired reference obligation to the seller, and the seller pays the par amount to the buyer) or cash settled (the seller pays the par value minus the recovery value to the buyer).84 In essence, the seller sells default insurance on the reference obligation to the buyer. Particularly when market conditions are calm and the reference obligation is deemed safe, the CDS spread is typically a small fraction of the notional amount.85 In these circumstances, the credit event is expected not to materialize, so the CDS seller will earn the small spread with a low probability that it will have to make the payment on the reference obligation. When the credit event does materialize, however, the seller’s payout can dwarf the spreads that the seller earned; as will be discussed below, this catastrophic scenario contributed to the near-failure of insurance giant AIG. In addition to CDSs, asset-backed commercial paper (ABCP) programs are another contributing factor to the 2008 financial crisis that exhibit the asymmetric return profile. At the centre of an ABCP program is a conduit (special purpose vehicle) set up by a sponsor (typically a bank or other financial institution). The conduit issues commercial paper and, with the proceeds, purchases assets such as mortgages, auto loans, and receivables.86 The sponsor administers the conduit and receives fee income for its services.87 In particular, ABCP programs, with assets with longer maturity than liabilities, were subject to the risk of runs by its commercial paper (CP) holders; the CP holders may refuse to roll over their financing, and new holders may not be found. To stave off this risk, the sponsors would provide liquidity support in various forms, including a liquidity put, in exchange for a small fee.88 With a liquidity put, when the CP holders do not roll over their CPs and new buyers cannot be found, the sponsor commits to step in and buy the CPs.89 Thus, the sponsors’ liquidity put can be considered a form of private deposit insurance. However, the problematic feature of this liquidity support is that when the sponsor is required to step in and provide liquidity—when market liquidity dries up—the sponsor is also in its most vulnerable position. Indeed, as the financial crisis unfolded and ABCP conduits began suffering widespread runs,90 those liquidity puts added further strain on their sponsors’ balance sheets.91 The resulting liquidity drain and balance sheet impairment could force the sponsor to start a fire sale of its assets, escalating its losses. For a small upside of less than a percent, the sponsor faces a chance—albeit small—of impaired assets burdening its balance sheet at a time when it is most vulnerable. These examples highlight that it is not uncommon to find assets and businesses in the financial sector with asymmetric return profile. In most circumstances, the catastrophic event will not materialize, and the financial institution will reap a small profit steadily. However, with a small probability, the financial institution will incur a large loss, sometimes enough to destabilize the whole firm. As a financial institution grows bigger and more diversified, these potential failure points increase, and the probability of a destabilizing loss can increase as well. c. Asymmetry of return in finance theory With the presence of highly leveraged financial institutions, asymmetry of return can be extended beyond those immediate examples noted above. The starting point of the analysis is the Shleifer–Vishny model of arbitrage.92 In the Shleifer–Vishny model, there is an arbitrage opportunity that a certain number of informed arbitrageurs know about: currently, the price of a certain asset is different from its true value, but sooner or later that price will certainly converge to the asset’s true value.93 Only the informed arbitrageurs know the asset’s true value, while uninformed noise traders do not know the asset’s true value until the price converges to the asset’s true value.94 The arbitrageurs get their funding from outside investors, who can allocate their capital between the arbitrageurs and other investment opportunities (think of the arbitrageurs as hedge fund managers).95 However, because of the presence of uninformed noise traders, in the short run, the price of the asset can deviate further away from the asset’s true value. If that scenario occurs, the arbitrageurs would face mark-to-market losses, and investors would flee. Facing an investor run, the arbitrageurs would be forced to liquidate their arbitrage positions, unable to stay until the price converges to true value. We can apply the Shleifer–Vishny model here as well. For example, consider a riskless arbitrage opportunity: the price of an asset is certain to converge to its true value. A leveraged financial institution takes advantage of this opportunity and engages in arbitrage. If the asset price converges to true value quickly—as it frequently does—the financial institution reaps a small positive return (since the price deviation from the true value is often small), albeit with very high probability. However, if, as it sometimes does, the asset price deviates further away from the true value, the leveraged financial institution will suffer mark-to-market losses. It may become undercapitalized, and its lenders may pull their funding. In such a scenario, the arbitrageur would face a liquidity crisis and be forced to exit its position at fire sale prices, suffering large losses in the process (although the probability of such a liquidity crisis is low). Thus, the payoff to the leveraged financial institution engaging in arbitrage trading would resemble the asymmetric payoff noted above: a small positive return with high probability and a large negative return with low probability. We can extend the model even further. Many financial institutions engage in maturity transformation: that is, they borrow with short-term liabilities and invest in long-term assets, creating a mismatch between assets and liabilities.96 In the vast majority of cases, the lenders to such financial institutions (whether it be depositors or repo counterparties) roll over their funding, allowing the institutions to stay in their investments and earn positive returns. However, in a small number of cases, the lenders can lose their confidence and pull their funding. With maturity transformation, the effect of the lenders’ loss of confidence is disastrous, as the financial institution engaged in maturity transformation cannot liquidate its assets easily. Therefore, it will be forced to sell those long-term assets at a substantial discount, incurring large losses and replicating the same asymmetric payoff profile. What is more notable is that this result can arise even if the underlying assets have more symmetric return profiles; the loss of financing can elongate the left tail, so the combination of symmetric return asset and maturity mismatch can lead to asymmetric return profile overall. d. Breakdown of welfare equivalence: discrete costs of banking failure As noted above, diversification with asymmetric return profile can increase the probability of banking failure. At the same time, it is true that in most cases diversification reduces the potential losses on a financial institution’s balance sheet. Thus, a diversified financial institution may fail more often but will fail by a smaller margin, and there should be equivalence of welfare overall. However, this equivalence only holds true if there is no discrete cost of banking failure. If there are discrete costs of banking failure, the scenario with the higher probability of failure will lead to a suboptimal outcome even if the failures are by a smaller margin. This section will outline various discrete costs of banking failure and show why diversification can lead to a suboptimal outcome. Prior to the financial institution’s failure itself, as the firm’s financial health spirals downward, short-term lenders to the firm must make a decision whether to roll over their financing or cut off their funding. If they receive information that indicates that the firm will not be able to honour all of its commitments or that their claims may be caught in bankruptcy proceedings, the creditors will run. As long as the expected recovery rate is less than 100%, whether it is slightly or significantly below full repayment does not matter; what matters instead is that, to each individual lender, pulling funding is the dominant strategy compared to rolling over their funding. This sudden evaporation of credit leads the liquidity-troubled financial institution to engage in a fire sale of assets to meet its liquidity needs, destroying asset value in the immediate days before the firm’s actual failure. Once a financial institution fails, a number of discrete costs arise by virtue of its failure. Most directly, there are expenses of bankruptcy (or other applicable resolution regime): legal, accounting, and other professional costs. Particularly for a lengthy resolution of a complex financial institution, these costs can be substantial. The total administrative expenses of Lehman Brothers’ failure amounted to $7.3 billion, consisting of $5.9 billion of Chapter 11 bankruptcy expenses and $1.4 billion of Securities Investor Protection Act (SIPA) resolution proceeding expenses.97 Lehman’s Chapter 11 bankruptcy expenses alone amounted to 2% of its assets.98 Similarly, the SIPA resolution of MF Global, a brokerage firm that failed in 2011, cost $312 million, or nearly 4% of total customer claims.99 In addition, sudden, unplanned unwinding of a large financial institution’s asset and liability positions can lead to tremendous destruction of value. Financial assets that are worth more together may be liquidated in pieces, destroying their combined value. For example, two offsetting derivative positions may be worth more together to a risk-averse investor than separately.100 Moreover, rebuilding these positions in a time of stressed financial markets can be difficult and costly. Thus, haphazard unwinding of a large financial institution’s balance sheet can destroy the value inherent in such a ‘combination’ of assets.101 Outside of the firm itself, there are additional costs to the immediate stakeholders—creditors, customers, and employees. Creditors (including custody clients) stuck in a lengthy resolution proceeding suffer not only because of the time value of money but also because their assets are stuck in an illiquid form and cannot be otherwise deployed. For example, when Lehman Brothers failed and its prime brokerage customers (hedge funds) found their assets stuck, the stocks held by Lehman customer hedge funds became more illiquid and the prime brokerage customers became more susceptible to failure.102 Likewise, the loss of longstanding client relationships can hurt the clients themselves as well; one study showed that Lehman’s equity underwriting clients suffered an abnormal negative return of 5% after Lehman’s bankruptcy, or $23 billion.103 Another victim of financial institution failure, its employees, can suffer layoffs, and even when they are hired again, they often face lower wages and/or cultural clashes.104 Lastly, there are the costs to the broader financial markets and to the real economy, which often dwarf the direct costs of a large banking failure. As noted above, large, diversified financial institutions are informationally linked; the failure of one diversified financial institution, particularly in times of market stress, can lead creditors of other financial institutions to run as well, precipitating their failure too.105 The adverse effects of a large banking failure on credit provision are also well-documented. Not only is the failed financial institution unable to make new credit extensions, in a stressed market, financial institutions still standing face two pressures: their creditors pull their short-term credit, and their borrowers expand their credit line usage.106 These pressures restrict new credit extensions, even on the part of surviving financial institutions as well, with significant repercussions for the real economy. The costs outlined above do not depend on the margin of failure but the occurrence of failure itself. For instance, the administrative costs of resolution and illiquidity costs to creditors and customers depend on the complexity and length of resolution, rather than the recovery rate. Even if the recovery rate is high—i.e., the firm failed by a small margin—if the resolution is complex, it will take longer, significant administrative costs will be incurred, and creditors and customers will have their assets tied up longer. Therefore, costs of financial institution distress ‘jump’ at the moment of formal failure of the firm, and failing more often by a smaller margin and failing less often by a larger margin are not equivalent outcomes. By increasing the frequency of failure, diversification can lead to a suboptimal outcome, even if it reduces the margin of failure in most instances. e. AIG and AIGFP: intra-firm spillover in action One piece of evidence that a financial institution is only as strong as its weakest link is the near-demise of the insurance giant AIG during the 2008 financial crisis. One of the largest insurance companies in the world, AIG was involved in a diverse array of businesses, such as general insurance, life insurance, retirement services, financial services, and asset management.107 Even within each operating segment, AIG was involved in various businesses; for example, its financial services segment included aircraft leasing, consumer finance, and CDS underwriting.108 One of AIG’s subsidiaries in the financial services sphere was AIG Financial Products, commonly called AIGFP. AIGFP was a major dealer in the derivatives market and notably had a large business selling CDSs guaranteeing the performance of various debt obligations, including collateralized debt obligations (CDOs).109 AIGFP would receive a steady stream of payments from the holders of its CDSs in return for the credit risk protection it offered.110 Moreover, AIG was not required to post collateral at the time these CDS contracts were written but agreed to post it if the reference securities dropped in value or if AIG itself was downgraded.111 Over the years prior to the 2008 financial crisis, AIGFP had become one of the more profitable subsidiaries of AIG, but it was only one part of the insurance giant. In 2005, of AIG’s total revenue of $108 billion and operating income of $15.2 billion, AIGFP’s operation contributed $3.3 billion of revenue and $2.7 billion of operating income, or 3% of total revenue and 18% of operating income, respectively.112 When the financial crisis hit in 2008, the collateral requirement from CDS contracts written and sold by AIGFP led to a catastrophic liquidity shortage for AIG. As the reference securities of AIGFP’s CDS contracts dropped in value, AIG was required to post ever-increasing collateral. In June 2008, AIG faced a collateral call of $15.7 billion, and by September, it was required to post $23.4 billion in collateral.113 Moreover, on 15 September 2008, the day Lehman Brothers filed for bankruptcy, AIG was downgraded by Standard & Poor’s (S&P), Moody’s, and Fitch, and as a result, it was required to post $13 billion in additional collateral.114 As AIG guaranteed AIGFP’s liabilities, AIGFP’s problems did not stop at the subsidiary’s corporate borders and threatened to overwhelm the parent company as well.115 Even though AIG’s core insurance business remained healthy, if AIG failed as a result of AIGFP’s liabilities, the core insurance business would not escape unscathed either. AIG’s failure would then cut off vital insurance and other financial services from hundreds of thousands of businesses and tens of millions of Americans.116 The concern over interruption of vital services played a role in the Treasury Department and the Federal Reserve’s decision to bail out AIG.117 Before the market turmoil began in 2007, for the five years from 2002 to 2006, AIGFP contributed operating income of $3 billion, or 4% of AIG’s total operating income of $71 billion during the same five-year period.118 In 2007 and 2008, AIGFP’s operating income took an extreme nosedive, with a 2007 operating loss of $10.6 billion and a 2008 operating loss of $40.5 billion; in 2008 alone, AIGFP suffered an operating loss close to 70 times its average operating income from 2002 to 2006 and close to three times AIG’s average operating income for the same five-year period.119 Not only did AIGFP’s losses dwarf its prior year income, but as a result of this extreme asymmetry, AIGFP also threatened the viability of AIG as a whole. AIG’s experience thus illustrates that in the presence of an asymmetric return profile, a small part of a large financial institution can cause losses that can destabilize the whole group. As a corollary, as these potential ‘weak links’ increase, the probability of failure of a large, diversified financial institution can increase as well. f. Long-Term Capital Management: picking up a nickel in front of a bulldozer The potential dangers of diversification in the presence of an asymmetric payoff structure can also be illustrated in the demise of Long-Term Capital Management (LTCM), a hedge fund widely known for its meteoric rise and equally meteoric failure. Founded in 1994, LTCM focused on absolute return strategies, such as bond arbitrage. For example, LTCM would look for two assets whose prices should be the same but were not. It would arbitrage the difference by buying the cheaper asset and shorting the more expensive asset and waiting for the price differential to disappear. In theory, the strategy posed minimal risk, as the price differential would disappear sooner or later. LTCM was involved in arbitrage opportunities in all corners of the financial markets, from US domestic merger situations120 to interest-only bonds121 to Italian government bonds.122 Furthermore, LTCM was involved in as many as 7,600 positions, and its partners believed that such diversification would protect LTCM from downside risk.123 Because these price differentials were often small, LTCM leveraged its equity capital to make a substantial return on its arbitrage investments. At the end of 1995, it was leveraged 28 to 1, greatly magnifying its returns.124 From 1994 to 1997, LTCM was the darling of the hedge fund world. It posted impressive returns, even after fees were taken into account. However, in August 1998, the unthinkable event happened: Russia declared a moratorium on its government debt, shaking the belief that nuclear powers did not default on their debt.125 Russia’s moratorium sent investors running for the safest bonds, increasing the price differential between safer and riskier bonds. LTCM, which had invested under the thesis that the differential between safer and riskier bonds would narrow, suffered massive losses; in August alone, it lost $1.9 billion, or 45% of its capital.126 LTCM’s ultimate demise soon followed, and it was bailed out by a consortium of Wall Street banks coordinated by the Federal Reserve. One commentator called LTCM’s investment strategy ‘picking up nickels in front of bulldozers’.127 Such a characterization is very helpful for our analysis and provides a support for why diversification with highly asymmetric payoff structures can be dangerous: picking up nickels in front of stopped bulldozers is almost always riskless, but when the strategy is repeated enough, in one of those many instances the bulldozer might move forward, with deadly consequences. 4. Increased difficulty of resolution Sections IV.1 to IV.3 showed the increased possibility of a major financial institution failure, with the attendant systemic risk concerns. This section will show that diversification can have adverse effects on the post-failure resolution of a systemically important financial institution as well. Thus, diversification can not only increase the probability of failure of a systemically important financial institution but also increase the probability of disorderly failure of a systemically important financial institution, similar to the Chapter 11 bankruptcy of Lehman Brothers that pushed the global financial system into a crisis in September 2008. It appears intuitive that more diversified financial institutions, by virtue of their size, would be more difficult to resolve in the case of failure than their less diversified peers. However, even controlling for size, diversification can act as a roadblock to the sale of a troubled financial institution, obstructing a quick pathway to rescue. Sale to a third party has long been the default resolution method of many banking regulators. For example, when Bear Stearns teetered on the edge of bankruptcy in March 2008, the Federal Reserve arranged its sale to JPMorgan Chase.128 Likewise, in September 2008, when Merrill Lynch ran into financial trouble, it was sold to Bank of America,129 and in October 2008, Wachovia was sold in a similar fashion to Wells Fargo.130 Whereas the piecemeal wind-down of a large financial institution can increase uncertainty and fear in the market—as creditors would fear a drawn-out resolution process and potential losses and attempt to run in order to avoid being trapped in liquidation—a sale to a stronger financial institution can be accomplished in a simple manner. However, a diversified financial institution may be a less appropriate target for a quick sale, as the business lines of the target and the acquirer may not match. Traditionally, financial institutions are more inclined to acquire targets operating in the same industry or geographic area. For example, between 1990 and 1999, of 7,304 acquisitions in the financial sector in 13 advanced countries, 5,260 (72%) were transactions for which the acquirer and the target operated in the same country and in the same industry.131 Intra-industry, cross-border transactions accounted for another 1,131 transactions (15%), and cross-industry transactions accounted for only 13% of the total.132 In the United States, the comparison is even more lopsided; of 4,094 financial sector transactions involving an American acquirer between 1990 and 1999, 3,795 (93%) were intra-industry transactions.133 Furthermore, particularly in the United States, regulations governing financial institutions often limit the scope of activities that financial institutions may engage in, restricting the ability of financial firms to make cross-industry acquisitions.134 As a result, a diversified financial institution may be less appealing to an acquirer because the target institution has a great likelihood of having business lines or assets that the acquirer does not wish to get involved in. A monoline broker-dealer may wish to acquire a diversified financial institution, but the target’s non-broker-dealer assets may kill the potential acquisition. Similarly, on an asset level, holdings of undesirable assets may deter potential acquirers from buying the target intact. Of course, one may point to several examples of business line or asset acquisitions to argue that partial acquisition of a diversified financial institution is workable as an alternative. After Lehman Brothers filed for bankruptcy, Barclays bought just Lehman’s North American investment banking and capital markets business out of bankruptcy.135 Moreover, in late September 2008, when Wachovia teetered on the edge of failure, Citigroup agreed to buy Wachovia’s banking subsidiaries while leaving Wachovia’s retail brokerage, asset management, and certain wealth management businesses behind.136 In these cases, the acquirer selected and acquired only certain parts of the troubled financial institution’s business. But while these acquisitions may be ‘workable’, they are far from smooth. In the case of Lehman, the hasty sale of its North American investment banking and capital markets business to Barclays meant that the parameters of the sale were not clearly defined. Disputes among Barclays, Lehman, and Lehman’s clearing agent regarding which Lehman collateral Barclays had agreed to purchase threatened to derail the acquisition.137 Moreover, Barclays unexpectedly refused to assume some brokerage accounts, including prime brokerage accounts, and these had to be transferred through a complex SIPA proceeding.138 Similarly, when Citigroup agreed to acquire Wachovia’s banking assets, Wachovia executives were concerned about the continued survival of the retail brokerage, asset management, and wealth management businesses left behind.139 In the depths of a liquidity crisis, these concerns could prompt creditors and customers of entities and accounts that will be or may be left behind to run, making a rescue futile. An alternative approach to the resolution of a diversified financial institution is the Orderly Liquidation Authority under Title II of the Dodd-Frank Act.140 In place of traditional bankruptcy (or other liquidation regime that may be applicable), the FDIC, acting as receiver of the failed financial institution, would resolve the firm, with special provisions, such as a stay on the termination of qualified financial contracts, to help the resolution occur smoothly.141 Furthermore, the FDIC has articulated its ‘Single Point of Entry’ strategy for implementing the Orderly Liquidation Authority; under this strategy, the FDIC would take only the top tier holding company into resolution, while leaving the operating subsidiaries to conduct business as ordinary.142 While the Orderly Liquidation Authority may be touted as an alternative to the sale of a financial institution as a resolution method, it should be characterized more as a last resort method than a perfect one. Moreover, a number of concerns have been raised regarding its implementation, and it has yet been tried in practice.143 In particular, when a financial institution is diversified across multiple business lines and multiple countries, two problems arise. First, the dividing line between these multiple business lines may be blurred, and there may be uncertainty about the solvency of each part of the conglomerate as a result.144 Second, there can be unanticipated ex post ringfencing of assets by national regulators when successful resolution requires large cross-jurisdictional transfers.145 Orderly resolution of a financial institution typically requires a quick sale to a third party or predictable, consistent winding down. A financial conglomerate can be a less attractive target for a quick sale, and it can also be more difficult to wind down using the Dodd-Frank Act’s Orderly Liquidation Authority. Given these obstacles, the failure of a diversified financial institution, rather than a monoline one, is likely to be more disruptive and poses greater systemic risk threats. 5. Unjustified expansion of the federal safety net As noted in Sections IV.1 to IV.3, diversification can increase the probability of failure of a financial institution, and Section IV.4 showed how such failure can be disorderly. With the spectre of disorderly failure of a financial conglomerate, there is a corresponding possibility that a government backstop may be necessary. The backstop may be institutionalized (e.g., discount window lending for depository institutions) or ad hoc (e.g., bailouts of financial institutions not eligible or not suitable for existing support programmes). Diversification, however, can have a negative impact on implementation of the government safety net as well. As a threshold matter, the increased likelihood of a disorderly failure inevitably brings increased need for a federal backstop. The Dodd-Frank Act implements some measures to make an ad hoc bailout less necessary, such as a resolution plan submission requirement and Orderly Liquidation Authority, but the efficacy of these measures is hardly certain.146 Moreover, to the extent that a financial subsidiary that provides a critical market function is coupled with another financial subsidiary that poses high risk to the overall conglomerate, in the event of a financial distress arising out of the latter’s losses, the regulators may be forced to bail out the whole conglomerate (including the high risk financial subsidiary) in order to save the portion of the conglomerate providing a critical function.147 After a federal safety net is deployed to prevent a financial conglomerate’s disorderly failure, there still remains a further problem that the government support may expand beyond its originally intended scope. When regulators provide assistance to one part of a financial conglomerate, the recipient of that assistance may funnel that support to other parts of the firm (e.g., through upstream, downstream, or cross-stream guarantees or through other affiliate transactions), sometimes without the explicit blessing of the regulators. For example, during the 2008 financial crisis, many bank holding companies that received federal assistance were criticized for funneling only a small part of such capital infusion to their bank subsidiaries as was expected.148 There is only so much control that the regulators can wield regarding how the federal support is ultimately utilized, leaving open the possibility that federal assistance may not be used in the manner desired by the regulators. Of course, this problem exists for undiversified financial institutions as well, but the presence of multiple business lines increases the possibility that liquidity and capital from government assistance would ‘leak’ from one part of the firm to another.149 To the extent that the support is originally injected into a bank subsidiary (particularly since banks have access to a number of institutionalized safety nets), Section 23A of the Federal Reserve Act provides some protection against such leakage by restricting banks’ transactions with their affiliates.150 When the support is injected at the parent level (as may be the case for ad hoc bailouts), there is no similar statutory safeguard. The regulators may provide support to the parent company expecting it to act as a ‘source of strength’ for its subsidiaries, but the parent may act as a source of strength for a different part of the firm than the regulators had intended. This situation is particularly problematic as it creates an uneven playing field between a financial company that is a part of a financial conglomerate and one that is not. If they are both not directly eligible for a federal safety net, a financial company that is a part of a financial conglomerate can indirectly tap into the support available to other parts of the conglomerate, consequently with lower risk of failure and lower cost of funding than a company that is not a part of a financial conglomerate but otherwise similar, giving the subsidiary an unfair advantage. 6. Summary The abovementioned five downsides of diversification throw into question whether diversification is an absolute good. While in many scenarios diversification can serve a valuable purpose—spreading out risk, reducing the probability of a catastrophic loss—in the context of financial institutions, diversification can have the opposite effect. First, it can lead to strains in internal control, increasing the probability of compliance or risk management failure. Second, as all financial institutions in a system diversify, they become more and more similar to each other, increasing the probability of joint failure and a systemic crisis. Third, with a highly asymmetric payoff structure, diversification can actually raise the individual failure probability of a financial institution by increasing the number of potential failure points. Diversified financial institutions can fail more often by a smaller margin, and given that a number of discrete costs arise as a result of failure itself (rather than the margin of failure), failing more often by a smaller margin is a suboptimal result. Fourth, from an ex post perspective, failure of a diversified financial institution can be more disorderly, raising systemic risk concerns. And lastly, diversification can increase the frequency of federal backstops to prevent disorderly failure of a financial conglomerate, and the scope of such federal backstops may expand beyond the regulators’ intentions. Unlike other industries, financial institutions are characterized by a number of peculiarities, including potential for systemic failure, asymmetry of return profile, maturity transformation, and need for quick resolution upon failure, and these peculiarities combine to produce the counter-intuitive result that diversified financial institutions can be more unstable than their undiversified counterparts.151 V. MAKING FINANCIAL CONGLOMERATES SAFER 1. Monoline requirement for financial institutions With Section IV’s discussion of the destabilizing characteristics of financial conglomerates, the most simplistic remedy would be breaking up financial conglomerates and/or imposing a monoline requirement on especially risky activities. In particular, a monoline requirement has long been used in the insurance industry and can provide useful guidance. In insurance regulation, ‘monoline’ refers to insurers that provide only one line of insurance coverage, as opposed to providing coverage for a number of different classes, such as life insurance and automobile insurance. The most prominent type of a monoline insurer is a financial guaranty insurer, which insures policyholders against default of financial instruments, including municipal bonds; perceived as riskier than other types of insurance, financial guaranty insurers are typically required to be monoline.152 In fact, the term ‘monoline insurer’, although it can refer to any insurer that provides only one type of insurance, has become synonymous with ‘financial guaranty insurer’. Originally, financial guaranty insurers focused on insuring municipal bonds. With the rise of securitization, however, financial guaranty insurers moved into insuring structured finance products, such as collateralized debt obligations and asset-backed securities, and by the time the 2008 financial crisis hit, financial guaranty insurers were subject to significant structured finance exposure.153 As a result of the 2008 financial crisis, financial guaranty insurers saw mounting losses on their structured finance insurance businesses, and one financial guaranty insurer after another lost their AAA credit rating.154 As these insurers lost their top-tier credit ratings, the loss of confidence spread over to the municipal bonds they insured, leading to a freeze-up of the municipal bond market as well.155 Because of this freeze-up, some commentators point to financial guaranty insurers as a contributing factor to the 2008 financial crisis. The performance of financial guaranty insurers, however, actually supports the proposition that a monoline requirement can help reduce systemic risk. Of course, it is widely acknowledged that the regulation of financial guaranty insurers prior to the 2008 financial crisis was inadequate.156 In contrast, the monoline requirement helped limit the spillover effect from the flawed substantive regulation of financial guaranty insurers.157 Had financial guaranty insurers been allowed to engage in other insurance activities, it is possible that their structured finance losses would have overwhelmed their non-structured finance businesses as well, leading to the interruption of vital insurance services. The freeze-up in the municipal bond market is a testament to the dangers of coupling different types of insurance; if municipal bond insurers were restricted from engaging in structured finance insurance and vice versa, the losses in the structured finance market would not have spread over to the municipal bond market as well. This analysis can be applied to financial institutions more broadly. Under a monoline requirement, financial institutions engaged in a particularly risky activity—such as those with asymmetric return profiles—would be required to engage only in such activity and prohibited from branching out to other business lines. If the catastrophic loss scenario does materialize with respect to these activities, the monoline firm itself would fail, but the losses would stop at the firm’s borders. a. Harmonizing a monoline requirement with existing regulatory framework One possible way in which this monoline requirement can be implemented is through the current US regulatory structure governing systemically important financial institutions. As noted above, the Dodd-Frank Act established the FSOC, responsible for identifying risks to and responding to emerging threats to the financial stability of the United States.158 As part of its mandate, the FSOC is authorized to designate certain non-bank financial companies as systemically important and subject to heightened supervision if ‘material financial distress [of such a company] or the nature, scope, size, scale, concentration, interconnectedness, or mix of the activities of the [company] could pose a threat to the financial stability of the United States’.159 In fulfilling its mandate, the FSOC originally used an entity-based approach, in which it designated a number of non-bank financial companies as systemically important and subjected them to enhanced prudential supervision. However, in 2019, the FSOC announced that it would switch from an entity-based approach to an activities-based approach.160 Under this new approach, the FSOC would identify an activity or a product (rather than an entity) that poses a risk to financial stability and work with relevant regulators to reduce or eliminate that threat.161 If the cooperation with relevant regulators fails to address that threat to financial stability, the FSOC may then designate one or more non-bank financial companies as systemically important, leaving an entity-based approach as a backup option.162 With this revised activities-based approach, the FSOC can focus on activities/business lines that pose an unacceptably high tail risk. As a transitional matter, the FSOC can take broad-based measures designed to compensate for the tail risk posed by such activity, including enhanced risk-management requirements. Such requirements may impose costs that effectively induce the conglomerates engaged in these activities to spin them out as monoline entities. When the heightened requirements on such high-risk activities fail to induce conglomerates to spin them out, the FSOC can then opt for the entity-based approach to take measures designed to mitigate such tail risk, including a divestment requirement.163 Thus, the FSOC’s newly unveiled activities-based approach provides a framework in which the monoline requirement for activities/business lines with high tail risk can be implemented. 2. Ringfencing An alternative approach that is less drastic but still achieves a similar result is ringfencing of risky activities within a financial conglomerate. Ringfencing is defined as ‘legally deconstructing a firm in order to more optimally reallocate and reduce risk’.164 Some commentators view physical separation as one form of ringfencing—thus, ringfencing includes (but is not limited to) full separation of certain activities and businesses from the rest of the firm. Under this definition, the breakup of a large financial institution into separate entities would still be ‘ringfencing’. Others view ringfencing as separation short of a full breakup. With this definition, ringfenced entities still remain part of the same firm, but are operated as if they are standalone entities. As a result, they are ‘legally’ deconstructed. Section V.1 has already addressed full breakup and monoline requirement, and here we take the latter definition of ringfencing as the legal, but otherwise not complete, deconstruction of a firm. Particularly after the 2008 financial crisis, ringfencing has been in vogue. For instance, as part of the Enhanced Prudential Standards (EPS) mandated by the Dodd-Frank Act, the Federal Reserve now requires foreign banking organizations (FBOs) with $50 billion or more in US non-branch assets to set up an intermediate holding company.165 Such FBOs must hold their entire ownership interest in US subsidiaries through the US intermediate holding company,166 and the intermediate holding company is subject to capital planning, stress testing, liquidity, risk management, and other EPS requirements.167 In essence, the Federal Reserve requires that large FBOs geographically ringfence their US operations to minimize the adverse impact of FBOs on the US financial system’s stability. In addition to geographic ringfencing, a number of jurisdictions have pushed for activity-based ringfencing in recent years. Section 716 of the Dodd-Frank Act, the so-called ‘swap pushout rule’, requires banks to move swaps to their non-bank affiliates and to comply with affiliate transaction restrictions of Sections 23A and 23B of the Federal Reserve Act in order to remain eligible for federal assistance, effectively ringfencing bank holding companies’ swap operations from their bank subsidiaries.168 Furthermore, the Volcker Rule in the United States prohibits banks and their affiliates from proprietary trading and from acquiring or retaining any equity, partnership, or other ownership interest in or sponsoring a hedge fund or a private equity fund.169 It therefore can be considered a form of ringfencing, although more of the type that requires full separation. In the United Kingdom, the Independent Commission on Banking (commonly called the Vickers Commission), established in 2010 to study structural reforms to the British banking sector, recommended that retail activities of large banking organizations be ringfenced, in order to protect them from risky activities conducted in wholesale/investment banking divisions of such organizations.170 Taking the Vickers Commission’s recommendations, the Financial Services (Banking Reform) Act 2013, followed by additional secondary legislation, implemented the ringfencing requirement.171 The ringfencing requirement applies to banks with average core deposit of more than £25 billion, with some types of speciality financial institutions exempt.172 ‘Core activities’ (the business of accepting deposits) must be placed in a ringfenced entity, and ‘excluded activities’ (dealing in investments as principal) must be placed outside the ringfenced entity.173 Activities that fall under neither category may be placed either inside or outside the ringfence.174 With the activities separated, the ringfenced entity must have a control and governance structure that permits it to make independent decisions, must have sufficient capital and liquidity, and must maintain arm’s-length terms in its dealings with and cannot be financially dependent on other parts of the financial group.175 The European Union has pushed for a similar activity-based ringfencing measure as well. Based on the recommendations of the High-Level Expert Group on Reforming the Structure of the EU Banking Sector (commonly called the Liikanen Group), the European Commission proposed a ringfencing requirement for its depository institutions in 2014.176 First, under the EU ringfencing proposal, covered banks and their affiliates would be prohibited from engaging in proprietary trading and from acquiring or retaining interest in alternative investment funds (hedge funds, etc.).177 Second, the applicable banking regulator of a covered bank must review its trading activities, and if it finds that there is ‘a threat to the financial stability of the [covered bank] or to the [EU] financial system’, it must require that the covered bank not engage in such trading activities.178 Should the financial group wish to continue carrying out such activities, they must be conducted by ‘a group entity that is legally, economically and operationally separate … from the [covered bank]’.179 The covered bank and the trading entity must be sufficiently separated; they must issue their own debt, the covered bank cannot own equity interest in the trading entity, management must be separated, and the covered bank faces intra-group exposure limit.180 These restrictions are designed such that even when the trading entity fails, the covered bank can continue its operations.181 While a number of regulators have pushed for ringfencing of financial institutions, both based on geography and activity, the merits of ringfencing have been the subject of considerable controversy. On the positive side, some commentators note ringfencing’s potential to reduce systemic risk,182 to ensure continued provision of publicly beneficial activities,183 to limit greater consolidation,184 or to correct market failure.185 In contrast, on the negative side, commentators note that ringfencing by multiple national regulators could trap capital and reduce the benefits of a shared reserve186 and that even after ringfencing, the entity engaged in risky activity can still fail and cause systemic risk concerns.187 Assessing these benefits and costs, I argue that activity-based ringfencing can be a prudent measure to ensure stability of diversified financial institutions. When a financial institution is engaged in activities that have a very long left tail, ringfencing can limit the spillover effect from a catastrophic scenario materializing in those activities. Even when the losses are sufficient to overwhelm the whole financial group, ringfencing means that the troubled subsidiary would fail by a large margin but the remainder of the firm would survive, rather than the entire firm failing. The failure of a single part of a financial conglomerate is more palatable than the failure of the entire conglomerate. If the conglomerate is engaged in providing services vital to the financial system, it can continue to provide those services despite the failure of one part of the firm. Moreover, by creating modularity, ringfencing makes resolution of troubled financial firms easier; the failed subsidiary can be sold or resolved much more smoothly than the whole conglomerate can. To the extent that systemic risk created by failure of a financial conglomerate is a market failure—the conglomerate does not internalize the costs of its own failure, such as interruption of vital services—ringfencing can correct that market failure. VI. CONCLUSION This article has argued that while regulators have a tendency to view diversification as an absolute good, it is far from so. While, in many cases, diversification of assets and activities can have risk management benefits, it is equally as important to acknowledge diversification’s many shortcomings and limits. From a governance perspective, diversification also means that the financial institution is becoming larger and more sophisticated, and such an expansion can lead to gaps in information transmission and risk management. From a systemic risk perspective, diversified financial institutions can resemble each other, and such a system would be vulnerable to a sudden, coordinated failure of a large number of financial institutions. From an individual firm’s perspective, with a highly asymmetric return profile, losses in one business can overwhelm the entire conglomerate and lead to failure of otherwise healthy parts of the firm as well; this intra-firm spillover becomes more likely as the number of potential failure points within a firm increases, and as a result, diversification can increase the firm’s probability of failure. In particular, this article demonstrates two key factors underlying the last insight that diversification can lead to a suboptimal outcome by raising the individual probability of failure of financial institutions: first, that asymmetric return profiles are not only common in practice but also derive from maturity transformation inherent in many financial institutions’ business; and second, that while diversified financial institutions may fail by a smaller margin (but more often), because of discrete costs of financial distress, welfare equivalence does not hold and more frequent failures by a smaller margin are suboptimal compared to less frequent failures by a larger margin. Furthermore, once the firm does fail, resolution of a diversified financial institution can be much more difficult and disorderly compared to the resolution of its monoline counterpart. A financial conglomerate can be more difficult to resolve through a sale to a third party, and implementation of the Dodd-Frank Act’s Orderly Liquidation Authority may be more difficult as well. If a government safety net is deployed to prevent such a disorderly failure, diversification of a financial institution makes it more likely that some of the government support would not be utilized in the manner desired or intended by the regulators. With these insights, this article proposes monoline requirement and ringfencing as two potential mitigation techniques. While the two measures are formally different—ringfencing is ‘legally deconstructing’ a firm, while monoline requirement is ‘actually deconstructing’ a firm—they pursue the same goal of limiting the intra-firm spillover of losses. Should catastrophic losses arise in one part of a financial conglomerate, the losses can be contained in that particular subsidiary without affecting the otherwise healthy parts of the firm. The troubled part of the firm can also be segregated and resolved in a much more efficient manner than the whole diversified firm can. Of course, diversification can and often does serve a valuable purpose in risk management. This article’s purpose is not to argue that diversification of a financial institution’s business and assets should be discarded as a risk management concept. Instead, the main purpose is to delineate the circumstances under which diversification can undermine the stability of financial institutions, both collectively and individually, and propose solutions to mitigate these destabilizing effects. A financial conglomerate can be considered a rope; if any part breaks, then the rope itself breaks. A modularized financial conglomerate, on the other hand, can be considered a link chain; even if one link in the chain breaks, that link can be discarded and the rest of the chain reconnected, preserving the integrity of the system. Footnotes 1 CFA Institute, Corporate Finance and Portfolio Management (2013) 296. 2 Zvi Bodie, Alex Kane, and Alan J Marcus, Investments (7th edn, McGraw-Hill 2008) 206. 3 Ibid. 4 Ibid. 5 Ibid. 6 Rustam Ibragimov, Dwight Jaffee, and Johan Walden, ‘Diversification Disasters’ (2011) 99 Journal of Financial Economics 333, 334. 7 See Section III for examples of these perceptions of diversification. 8 Of course, a financial institution’s choice of revenue stream and business line, which we have classified as asset-side, can and do affect the liability side as well. For example, acting as a counterparty on a credit default swap (CDS), while a revenue-generating activity, imposes potential liabilities. 9 Wolf Wagner, ‘Diversification at Financial Institutions and Systemic Crises’ (2010) 19 Journal of Financial Intermediation 373, 374 (‘Various forms of financial integration (such as interbank liquidity insurance or bank mergers) are comparable to diversification in that they also make systemic failures more likely’); Franklin Allen, Ana Babus and Elena Carletti, ‘Asset Commonality, Debt Maturity and Systemic Risk’ (2012) 104 Journal of Financial Economics 519, 519 (‘The emergence of financial instruments in the form of credit default swaps and similar products has improved the possibility for financial institutions to diversify risk’). 10 Gramm–Leach–Bliley Act, Pub L No 106-102, 113 Stat 1338 (1999). 11 Glass–Steagall Act (Banking Act of 1933), Pub L No 73-66, 48 Stat 162 (1933). 12 Ibid s 16 (‘The business of dealing in investment securities by [a national banking association] shall be limited to purchasing and selling such securities without recourse, solely upon the order, and for the account of, customers, and in no case for its own account, and [the national banking association] shall not underwrite any issue of securities’). 13 Ibid s 20 (‘No member bank shall be affiliated … with any corporation, association, business trust, or other similar organization engaged principally in the issue, flotation, underwriting, public sale, or distribution … of stocks, bonds, debentures, notes, or other securities’). 14 Ibid s 21 (‘It shall be unlawful … [for] any person, firm, corporation, association, business trust, or other similar organization, engaged in the business of issuing, underwriting, selling, or distributing … stocks, bonds, debentures, notes, or other securities, to engage at the same time … in the business of receiving deposits’). 15 Ibid s 32 (‘No officer or director of any [Federal Reserve member bank] shall be an officer, director, or manager of any corporation, partnership, or unincorporated association engaged primarily in the business of purchasing, selling, or negotiating securities’). 16 Some of the regulatory changes that relaxed the Glass–Steagall Act’s prohibition included (1) permitting banks to sell mortgage pass-through certificates in a public offering (thus allowing securitization of mortgage assets) and (2) permitting bank holding companies to establish subsidiaries (so-called ‘Section 20 subsidiaries’) that engaged in limited securities underwriting (these subsidiaries could derive up to 25% of their revenues from ineligible securities underwriting activities). With respect to banks’ public sale of mortgage pass-through certificates, see Susan M Golden, ‘Collateralized Mortgage Obligations: Probing the Limits of National Bank Powers Under the Glass–Steagall Act’ (1987) 36 Catholic University Law Review 1025, 1037; OCC Interpretive Letter No 388, 1987 WL 154126 (16 June 1987) (interpreting that a public offering of mortgage pass-through certificates ‘involves a sale of bank assets, which is fully permitted under the national banking laws’). With respect to Section 20 subsidiaries of bank holding companies, see James R Barth, R Dan Brumbaugh and James A Wilcox, ‘The Repeal of Glass–Steagall and the Advent of Broad Banking’ (2000) 14(2) Journal of Economic Perspectives 191, 196; Lawrence J White, ‘The Gramm–Leach–Bliley Act of 1999: A Bridge Too Far? Or Not Far Enough?’ (2010) 43 Suffolk Law Review 937, 940; Arthur E Wilmarth, Jr, ‘The Road to Repeal of the Glass–Steagall Act’ (2017) 17 Wake Forest Journal of Business and Intellectual Property Law 443, 470. Professor Wilmarth’s article provides a comprehensive overview of industry pressure to relax the Glass–Steagall Act’s prohibitions and gradual dismantling of the separation of commercial and investment banking. 17 Gramm–Leach–Bliley Act, Pub L No 106-102, s 101, 113 Stat 1338, 1341 (1999). It is important to note that the separation of investment banking and commercial banking was only one part of the Glass–Steagall Act and that the Gramm–Leach–Bliley Act left other parts of the Glass–Steagall Act intact. However, since the focus of this section is on the separation of investment banking and commercial banking, this section will use the term ‘Glass–Steagall Act’ largely to refer to the prohibition on commercial banks’ investment banking activities. 18 Ibid s 103. 19 Bank Holding Company Act of 1956, Pub L No 84-511, s 4(c), 70 Stat 133, 136–37. 20 Gramm–Leach–Bliley Act s 103. 21 US Department of the Treasury, ‘Modernizing the Financial System: Recommendations for Safer, More Competitive Banks’ (1991) XVIII-32. 22 Wilmarth (n 16) 516. 23 H R Rep No 104-127, pt 2, at 3 (1995) (reflecting the Congressional Budget Office’s view that ‘the bill could reduce potential risk to the [Deposit Insurance Fund] by allowing banks to diversify their sources of income and by helping banks to be more competitive in the world’s financial markets’); S Rep No 106-44, at 4, 41 (noting that ‘[financial institutions’] inability to diversify actually increases risks to the financial system’ and that ‘diversifying income sources also could result in lower overall risks for banks’). 24 Basel Committee on Banking Supervision, International Convergence of Capital Measurement and Capital Standards (Bank for International Settlements 2006) 6 (hereinafter Basel II). 25 Ibid 2 (‘In developing the revised Framework, the Committee has sought to arrive at significantly more risk-sensitive capital requirements that are conceptually sound and at the same time pay due regard to particular features of the present supervisory and accounting systems in individual member countries. It believes that this objective has been achieved’). 26 Ibid 12. 27 Emily Jones and Alexandra O Zeitz, ‘The Limits of Globalizing Basel Banking Standards’ (2017) 3 Journal of Financial Regulation 89, 94. 28 Basel II (n 24) 144. 29 Ibid 157. 30 To be more precise, minimum capital requirement was calculated as the sum of (1) 8% of risk-weighted assets for credit risk, (2) minimum capital requirement for operational risk, and (3) minimum capital requirement for market risk. Ibid 12. 31 Ibid 144. 32 Ibid 145. 33 Jan Lubbe and Flippie Snyman, ‘The Advanced Measurement Approach for Banks’ (2010) 33 IFC Bulletin 141, 141. 34 Basel II (n 24) 148. 35 Dodd–Frank Wall Street Reform and Consumer Protection Act, Pub L No 111-203, s 111, 124 Stat 1376, 1392–93 (2010). The FSOC is chaired by the Secretary of the Treasury, and its 10 voting members are: the Treasury Secretary, Chair of the Board of Governors of the Federal Reserve System, the Comptroller of the Currency, the Director of the Consumer Financial Protection Bureau, the Chair of the Securities and Exchange Commission, the Chair of the Federal Deposit Insurance Corporation, the Chair of the Commodity Futures Trading Commission, the Director of the Federal Housing Finance Agency, the Chair of the National Credit Union Administration, and an independent member (appointed by the President) with insurance expertise. Ibid. 36 Ibid s 113. 37 Ibid s 165. 38 Enhanced Prudential Standards (Regulation YY), 12 CFR pt 252 (2018). 39 Ibid subpt H. 40 Ibid s 252.72(a). A ‘covered company’ means any bank holding company (other than foreign banking organizations and their US intermediate holding companies) with total consolidated assets over $250 billion or identified as a global systemically important bank holding company. Ibid s 252.70(a)(2)(i). Foreign banking organizations are subject to similar limits in subpart Q of Regulation YY. Ibid s 252.172(a). 41 Ibid s 252.72(b). 42 Enhanced Prudential Standards and Early Remediation Requirements for Covered Companies, 77 Fed Reg 594, 612 (5 January 2012). 43 Oliver E Williamson, ‘Hierarchical Control and Optimum Firm Size’ (1967) 75 Journal of Political Economy 123; Guillermo A Calvo and Stanislaw Wellisz, ‘Supervision, Loss of Control, and the Optimum Size of the Firm’ (1978) 86 Journal of Political Economy 943. 44 Williamson (n 43) 135 (‘For any given span of control … an irreducible minimum degree of control loss results from the simple serial reproduction distortion that occurs in communicating across successive hierarchical levels’). 45 Financial Crisis Inquiry Commission, Financial Crisis Inquiry Report (PublicAffair 2011) 266 (‘The initial collateral call was a shock to AIG’s senior executives, most of whom had not even known that the credit default swaps with Goldman contained collateral call provisions’). 46 Jeremy C Stein, ‘Information Production and Capital Allocation: Decentralized Versus Hierarchical Firms’ (2002) 57 Journal of Finance 1891, 1893. 47 One can make the argument that risk management can be decentralized and therefore does not require hard information. However, fully decentralized risk management has two significant problems. First, there is no ‘second line of defense’ that verifies whether risk management at each division level is sound. Second, risks identified by each division may not reflect the risks facing the organization as a whole. Furthermore, in practice, large diversified financial institutions rely on central risk management as supervisor of division-by-division risk management. JP Morgan Chase & Co, Annual Report (Form 10-K) 77 (27 February 2018) (‘The [Independent Risk Management] function is independent of the businesses and forms “the second line of defense”. The IRM function sets and oversees various standards for the risk governance framework, including risk policy, identification, measurement, assessment, testing, limit setting, monitoring and reporting, and conducts independent challenge of adherence to such standards’); Goldman Sachs Group, Annual Report (Form 10-K) 79 (26 February 2018). 48 Henry T C Hu, ‘Too Complex to Depict? Innovation, “Pure Information,” and the SEC Disclosure Paradigm’ (2012) 90 Texas Law Review 1601; Henry T C Hu, ‘Disclosure Universes and Modes of Information: Banks, Innovation, and Divergent Regulatory Quests’ (2014) 31 Yale Journal on Regulation 565. Professor Hu mostly focused on how complex financial institutions are ‘too complex to depict’ to their shareholders and regulators, but this ‘too complex to depict’ problem applies to information exchange between different divisions of a complex financial institution as well. 49 Ibid 619–23. 50 Hu, ‘Too Complex to Depict?’ (n 48) 1669. Professor Hu does note that this ‘too complex to depict’ problem can be remediated by resorting to ‘transfer mode’. Instead of ‘depicting’ (and inevitably simplifying) the state of the financial institution, the source of a particular piece of information would ‘transfer’ the pure information to those demanding that information. However, at least with respect to intra-firm risk management, as a firm diversifies, this ‘transfer mode’ would require immense computational and cognitive capability on the part of a firm’s executives and risk managers. Whether such management practice is possible is questionable. Even the broader market participants may find the analysis of such voluminous information daunting. Steven L Schwarcz, ‘Regulating Complexity in Financial Markets’ (2009) 87 Wash University Law Review 211, 221 (‘Complexity can deprive investors and other market participants of the understanding needed for markets to operate effectively. Even if all information about a complex structure is disclosed, complexity increases the amount of information that must be analyzed in order to value the investment with a degree of certainty. This additional analysis entails higher cost’). 51 Hu, ‘Disclosure Universes and Modes of Information’ (n 48) 626–27 (‘Major money center banks are large, complex organizations that span the globe. In all such organizations, private or governmental, some “siloing” of information is inevitable’). 52 Ibid. 53 For an example of this intentional siloing, see Maurice R Greenberg and Lawrence A Cunningham, The AIG Story (Wiley 2013) 234 (‘[AIG’s] new internal control and risk systems apparently impaired information flow from [AIG Financial Products (FP)] to corporate headquarters. A recently hired AIG internal auditor, Joseph St. Denis … grew concerned in September 2007 when FP received a large collateral call … But as St. Denis tried to investigate, [Joseph Cassano, head of FP] reportedly discouraged him and blocked his access to report up the corporate chain to AIG’s senior management or its board’). 54 There is empirical evidence that diversified financial institutions have lower market value compared to their undiversified peers. While many causes have been proposed, such as increased shareholder-manager agency problem in a diversified financial institution, this potential for failure of internal control may play a factor as well. Luc Laeven and Ross Levine, ‘Is There a Diversification Discount in Financial Conglomerates?’ (2007) 85 Journal of Financial Economics 331; Martin R Goetz, Luc Laeven and Ross Levine, ‘Identifying the Valuation Effects and Agency Costs of Corporate Diversification: Evidence from the Geographic Diversification of U.S. Banks’ (2013) 26 Review of Financial Studies 1787, 1815. 55 Saabira Chaudhuri, Amy Guthrie and Shayndi Raice, ‘Citigroup Takes $400 Million Hit, Alleging Fraud in Mexico’ The Wall Street Journal (28 February 2014) <www.wsj.com/news/articles/SB10001424052702304071004579410863489379726> last accessed 11 March 2019. 56 Philip van Doorn, ‘How Did the Fraud Against Citi’s Banamex Work?’ The Street (28 February 2014) <www.thestreet.com/story/12462191/1/how-did-the-fraud-against-citis-banamex-work.html> last accessed 11 March 2019. 57 Ibid. 58 Amy Guthrie and Juan Montes, ‘Mexican Regulators Identify Critical Errors in Banamex’s Oceanografia Dealings’ The Wall Street Journal (18 June 2014) <www.wsj.com/articles/mexican-regulators-identify-critical-errors-in-banamexs-oceanografia-dealings-1403131238> last accessed 11 March 2019. 59 Ibid. 60 Ibid. 61 Dakin Campbell, ‘Banamex Said to Snub Citigroup Oversight as Fraud Mounted’ Bloomberg (7 March 2014) <www.bloomberg.com/news/articles/2014-03-08/banamex-said-to-snub-citigroup-oversight-as-fraud-mounted> last accessed 11 March 2019. 62 Citigroup Global Mkts and Citigroup Inc, Exchange Act Release No 83859, Investment Advisers Act Release No 4986 (16 August 2018) 3. 63 Ibid 4. 64 Ibid. 65 Ibid. 66 Ibid. 67 Ibid 5–6. 68 Ibid 6. There is some indication that conflict of interest played a role in this misclassification, since employees in charge of the Oceanografia credit would gain if the credit line was increased. Ibid. 69 Ibid 8. 70 Ibragimov, Jaffee and Walden (n 6) 333; Stefano Battiston and others, ‘Liaisons Dangereuses: Increasing Connectivity, Risk Sharing, and Systemic Risk’ (2012) 36 Journal of Economic Dynamics and Control 1121, 1122. 71 Wagner (n 9) 373–74. 72 This uncertainty can make investors more risk adverse and can lead to a panic. Schwarcz (n 50) 225. 73 Allen, Babus and Carletti (n 9) 531. 74 Wagner (n 9) 379. 75 Since this section only deals with individual failure probability, we do not consider the interaction among multiple financial institutions in a system and assume, for simplicity, that there is only one financial institution. 76 0.01% probability that the bad scenario would occur with both assets plus 1.98% probability that the bad scenario would occur with only one asset. 77 This assertion holds true unless the number of fractional assets exceeds 901. 78 Pierre Vernimmen and others, Corporate Finance: Theory and Practice (5th edn, Wiley 2017) 404. 79 Ibid 405. 80 Ibid 406. 81 For a call option, technically there is no limit to the seller’s potential loss, since the price of the underlying asset can increase indefinitely. For a put option, the zero lower bound on the underlying asset’s price caps the seller’s potential loss. 82 If an option is in-the-money, the return profile is more symmetric, but the fact that the seller’s upside is limited to the premium continues to hold true. 83 Arvind Rajan, Glen McDermott and Ratul Roy, The Structured Credit Handbook (Wiley 2007) 17; Pimco, ‘Understanding Investing: Credit Default Swaps’ (Pacific Investment Management Company, 2017) <www.pimco.co.uk/en-gb/resources/education/understanding-credit-default-swaps> last accessed 11 March 2019. 84 Rajan, McDermott and Roy (n 83) 24. 85 International Monetary Fund, Global Financial Stability Report: Containing Systemic Risks and Restoring Financial Soundness (2008) 9. 86 Bank of America Global Capital Management, Asset-Backed Commercial Paper: A Primer (2011) 1; Moody’s Investor Services, The Fundamentals of Asset-Backed Commercial Paper (2003) 11. 87 Ibid 8. 88 For example, in the years prior to the 2008 financial crisis, Citigroup charged 10 to 20 basis points for its liquidity support. Financial Crisis Inquiry Commission (n 45) 138. 89 John Armour and others, Principles of Financial Regulation (Oxford University Press 2016) 440. 90 One study estimates that 40% of ABCP programs suffered a run in the early stages of the financial crisis. Daniel Covitz, Nellie Liang and Gustavo A Suarez, ‘The Evolution of a Financial Crisis: Collapse of the Asset-Backed Commercial Paper Market’ (2013) 68 Journal of Finance 815, 828. 91 Financial Crisis Inquiry Commission (n 45) 138–39. 92 Andrei Shleifer and Robert W Vishny, ‘The Limits of Arbitrage’ (1997) 52 Journal of Finance 35. 93 Ibid 38. 94 Ibid. 95 Ibid. 96 Sandra C Krieger, ‘Reducing Systemic Risk in Shadow Maturity Transformation’ (Remarks at the Global Association of Risk Professionals 12th Annual Risk Management Convention, New York City, 8 March 2011) <www.newyorkfed.org/newsevents/speeches/2011/kri110308.html> last accessed 11 March 2019. 97 Erin Denison, Michael Fleming and Asani Sarkar, ‘Lehman’s Bankruptcy Expenses’ Federal Reserve Bank of New York Liberty Street Economics (15 January 2019) <libertystreeteconomics.newyorkfed.org/2019/01/lehmans-bankruptcy-expenses.html> last accessed 11 March 2019. 98 Ibid. 99 Securities Investor Protection Corporation, Annual Report (2014) 33. 100 Mark J Roe and Stephen D Adams, ‘Restructuring Failed Financial Firms in Bankruptcy: Selling Lehman’s Derivative Portfolio’ (2015) 32 Yale Journal on Regulation 363, 375–76. 101 Ibid. 102 George O Aragon and Philip E Strahan, ‘Hedge Funds as Liquidity Providers: Evidence from the Lehman Bankruptcy’ (2012) 103 Journal of Financial Economics 570, 585–86. 103 Chitru S Fernando, Anthony D May and William L Megginson, ‘The Value of Investment Banking Relationships: Evidence from the Collapse of Lehman Brothers’ (2012) 67 Journal of Finance 235, 237. 104 For example, after Barclays Capital bought Lehman’s North American investment banking and trading business out of bankruptcy and took on 10,000 employees, within three months it laid off 3,500 US employees. ‘Barclays Capital Says Lehman Staff Told of Job Cuts’ Reuters (18 December 2008) <www.reuters.com/article/us-barclayscapital/update-1-barclays-capital-says-lehman-staff-told-of-job-cuts-idUSLI34437620081218> last accessed 11 March 2019. 105 See n 73. 106 Victoria Ivashina and David Scharfstein, ‘Banking Lending During the Financial Crisis of 2008’ (2010) 97 Journal of Financial Economics 319, 319–20. 107 American International Group, Annual Report (Form 10-K) 3 (28 February 2008) (hereinafter AIG 2007 Annual Report). 108 Ibid. 109 Financial Crisis Inquiry Commission (n 45) 140. 110 For a more detailed explanation of CDS, see Section IV.3.b. 111 Financial Crisis Inquiry Commission (n 45) 141. 112 American International Group, Annual Report (Form 10-K) 27, 54 (16 March 2006) (hereinafter AIG 2005 Annual Report). 113 Financial Crisis Inquiry Commission (n 45) 344. 114 Ibid 349. 115 AIG 2007 Annual Report (n 107) 179 (‘AIG has issued unconditional guarantees with respect to the prompt payment, when due, of all present and future payment obligations and liabilities of AIGFP arising from transactions entered into by AIGFP’). 116 David Wessel, In Fed We Trust (Crown Business 2009) 192 (‘AIG’s businesses also touched hundreds of millions of American companies and households. In addition to its foundering hedge fund, AIG sold insurance to more than 100,000 big and small companies, pension plans, and municipalities, and sold insurance or managed retirement plans for thirty million Americans’). 117 Financial Crisis Inquiry Commission (n 45) 350; Alan S Blinder, After the Music Stopped: The Financial Crisis, The Response, and the Work Ahead (Penguin 2013) 137. 118 American International Group, Annual Report (Form 10-K) 13, 22 (31 March 2003); AIG 2005 Annual Report (n 112) 54, 74; American International Group, Annual Report (Form 10-K) 63, 120 (1 March 2007). 119 American International Group, Annual Report (Form 10-K) 116 (2 March 2009). 120 Roger Lowenstein, When Genius Failed (Random House 2001) 100–02. 121 Ibid 52–53. 122 Ibid 56–57. 123 Ibid 36, 114. 124 Ibid 78. 125 Ibid 144. 126 Ibid 159. 127 Ibid 102. 128 Financial Crisis Inquiry Commission (n 45) 290. 129 Ibid 335, 382–84. 130 Wells Fargo & Co, ‘Wells Fargo, Wachovia Agree to Merge’ (Press Release, 3 October 2008). 131 Group of Ten, Report on Consolidation in the Financial Sector (2001) 335. 132 Ibid. 133 Ibid 353. 134 Such regulations include 12 USC s 1843 (2012) (restricting a bank holding company’s ability to own nonbanking companies); 12 USC s 24 (Seventh) (enumerating the powers of a national bank). 135 Michael J Fleming and Asani Sarkar, ‘The Failure Resolution of Lehman Brothers’ [December 2014] Federal Reserve Bank of New York Economic Policy Rev 175, 180, 182. 136 Citigroup, ‘Citi and Wachovia Reach Agreement-in-Principle for Citi to Acquire Wachovia’s Banking Operations in an FDIC-Assisted Transaction’ (Press Release, 29 September 2008). After Citigroup’s agreement-in-principle with Wachovia was announced, however, Wells Fargo made a competing offer and prevailed over Citigroup. Two years after the acquisition, Wells Fargo paid Citigroup $100 million to resolve the claim that the Wells Fargo-Wachovia deal violated Wachovia’s agreement-in-principle with Citigroup. Eric Dash, ‘Wells Fargo Pays Citi $100 Million for Wachovia Claims’ (New York Times Dealbook, 19 November 2010) <dealbook.nytimes.com/2010/11/19/wells-to-pay-citi-100-million-for-wachovia-claims> last accessed 11 March 2019. 137 Fleming and Sarkar (n 135) 182. 138 Ibid 199. 139 Frank A Hirsch and Joseph A Dowdy, ‘Whither Wachovia? Wells Fargo Wins the Battle for the Storied North Carolina Banking Institution’ (2009) 13 North Carolina Banking Institute 167, 178. 140 Dodd-Frank Wall Street Reform and Consumer Protection Act, Pub L No 111-203, title 2, 124 Stat 1376, 1442–1520 (2010). 141 Ibid s 210. 142 Resolution of Systemically Important Financial Institutions: The Single Point of Entry Strategy, 78 Fed Reg 76,614 (18 December 2013). For further overview of the Dodd-Frank Act’s Orderly Liquidation Authority and Single Point of Entry Strategy, see Kwon-Yong Jin, ‘How to Eat an Elephant: Corporate Group Structure of Systemically Important Financial Institutions, Orderly Liquidation Authority, and Single Point of Entry Resolution’ (2015) 124 Yale Law Journal 1746, 1753–59. 143 Howell E Jackson and Stephanie Massman, ‘The Resolution of Distressed Financial Conglomerates’ (2017) 3 Russell Sage Foundation Journal of the Social Sciences 48; Jin (n 142); David A Skeel, ‘Single Point of Entry and the Bankruptcy Alternative’ in Martin Neil Baily and John B Taylor (eds), Across the Great Divide: New Perspectives on the Financial Crisis (Hoover Press 2014). 144 Jin (n 142) 1772–77. 145 Patrick Bolton and Martin Oehmke, ‘Bank Resolution and the Structure of Global Banks’ (2019) 32 Review of Financial Studies 2384 (‘When the cross-jurisdictional transfer required for a successful SPOE resolution is too large, regulators may prefer to ring-fence assets in their own jurisdiction, thereby preventing the required transfers. The planned SPOE resolution breaks down, leading to a disorderly liquidation or a tax-funded bailout’). 146 See n 143. 147 AIG, which combined an insurance business for millions of Americans with its high-risk AIGFP subsidiary, provides a classic example of a bailout of a conglomerate to save a critical market function. See Section IV.3.e. Furthermore, recognizing this possibility that potential interruption of a critical function can create systemic risk and force a bailout, in designating a non-bank financial company as systemically important, the FSOC considers whether such a company provides a ‘critical function or service’. Authority to Require Supervision and Regulation of Certain Nonbank Financial Companies, 77 Fed Reg 21,637, 21,657 (11 April 2012) (‘The Council has identified the following channels as most likely to facilitate the transmission of the negative effects of a nonbank financial company’s material financial distress or activities to other financial firms and markets: … A nonbank financial company is no longer able or willing to provide a critical function or service that is relied upon by market participants and for which there are no ready substitutes’). When AIG was designated a systemically important financial institution in 2013, the FSOC cited AIG’s commercial insurance portfolio (and particularly its specialized insurance portfolio) as a key factor, and when GE Capital was similarly designated, the FSOC cited GE Capital’s role in providing credit in commercial lending/leasing, consumer revolving credit, and aviation financing. ‘Basis of the Financial Stability Oversight Council’s Final Determination Regarding American International Group, Inc.’ (Financial Stability Oversight Council, 2013) <https://www.treasury.gov/initiatives/fsoc/designations/Documents/Basis%20of%20Final%20Determination%20Regarding%20American%20International%20Group,%20Inc.pdf> last accessed 12 May 2019; ‘Basis of the Financial Stability Oversight Council’s Final Determination Regarding General Electric Capital Corporation, Inc.’ (Financial Stability Oversight Council, 2013) <https://www.treasury.gov/initiatives/fsoc/designations/Documents/Basis%20of%20Final%20Determination%20Regarding%20General%20Electric%20Capital%20Corporation,%20Inc.pdf> last accessed 12 May 2019. 148 John Coates and David Scharfstein, ‘Lowering the Cost of Bank Recapitalization’ (2009) 26 Yale Journal on Regulation 373, 377 (‘Our analysis of call reports of the lead banks of the four largest BHCs to receive the first $90 billion of TARP investments shows that as of the end of 2008, less than $15 billion had been downstreamed to the banks as equity capital’). 149 There is also a related situation in which the regulators intentionally use institutionalized safety net (such as those available for depository institutions) to support other parts of the financial conglomerate that are not directly eligible for such safety net. One such example would be a two-part transaction in which the regulators provide short-term liquidity to a bank, and the bank in turn provides the liquidity to its non-bank affiliates, with the regulators’ blessing. Saule T Omarova, ‘From Gramm-Leach-Bliley Act to Dodd-Frank: The Unfulfilled Promise of Section 23A of the Federal Reserve Act’ (2011) 89 North Carolina Law Review 1683, 1730–33 (describing a 2007 Federal Reserve Board decision to open a special discount window lending facility for depository institutions and to permit a number of major banks to extend credit to their securities broker-dealer affiliates, effectively funneling liquidity from the discount window to such broker-dealer affiliates). In this situation, the federal support is used in the manner desired by the regulators but not in the manner originally intended by the applicable statute. Given that the federal support is still being used in the regulators’ desired direction, this situation may be less problematic, but the regulators’ deviating away from the original statutory intent is subject to criticism and concern. 150 Even in such a case, one may question whether Section 23A is sufficiently strong as a firewall, given the Federal Reserve Board, the Office of the Comptroller of the Currency, and the Federal Deposit Insurance Corporation’s (albeit not unlimited) power to grant exemptions. 12 USC s 371c(f)(2) (2018); Omarova (n 149) 1766–67. 151 Empirical evidence suggests that banks’ non-interest income increases systemic risk related to their tail risk and interconnectedness, and that the effect on banks’ systemic risk related to interconnectedness is particularly pronounced for larger banks. Markus K Brunnermeier, Gang Dong and Darius Palia, ‘Banks’ Non-Interest Income and Systemic Risk’ (2019) Working Paper, 28, 32 <https://www.ssrn.com/abstract=3328890> last accessed 11 May 2019. This is consistent with the potential ex ante downsides of diversification noted in this article. The potential for failure of internal control and spillover effects from high-risk business lines would increase financial institutions’ tail risk. Asset correlation would increase financial institutions’ interconnectedness risk, but given that small financial institutions may not be able to diversify sufficiently to resemble one another, we would expect the asset correlation/similarity risk to be most pronounced for larger institutions. 152 For example, see NY Ins Law s 6902 (2018). See also Daniel Schwarcz and Steven L Schwarcz, ‘Regulating Systemic Risk in Insurance’ (2014) 81 University of Chicago Law Review 1569, 1586–87. 153 Tima Tilek-uulu Moldogaziev, ‘The Collapse of the Municipal Bond Insurance Market: How Did We Get Here and Is There Life for the Monoline Industry Beyond the Great Recession?’ (2013) 25 Journal of Public Budgeting, Accounting & Financial Management 199, 221 (‘By 2007 the insurance exposure of monoline firms to structured finance risks in their portfolios was anywhere between 20% (MBIA and FGIC) to over 80% (ACA)’). For example, at the end of 2007, Ambac, one of the largest financial guaranty insurers in the United States, had guaranteed $524 billion par value of financial instruments, with $281 billion (53%) in public finance, $171 billion (33%) in structured finance, and $72 billion (14%) in international finance. Ambac Financial Group, Inc, Annual Report (Form 10-K) 16 (29 February 2008). 154 Sebastian Schich, ‘Challenges Related to Financial Guarantee Insurance’ (OECD Financial Market Trends 2008) 101 <www.oecd.org/finance/financial-markets/45500761.pdf> last accessed 11 March 2019; ‘Ambac Loses AAA Rating from Fitch; On Watch at S&P’ (CNBC, 18 January 2008) <www.cnbc.com/id/22727053> last accessed 11 March 2019; Dena Aubin, ‘Fitch Cuts “AAA” Rating of FGIC Insurance Unit’ Reuters (31 January 2008) <www.reuters.com/article/us-fgic-rating-fitch-idUSN3023531520080130> last accessed 11 March 2019; Alistair Barr, ‘Ambac, MBIA Lose AAA Ratings from S&P’ MarketWatch (5 June 2008) <www.marketwatch.com/story/ambac-mbia-lose-aaa-ratings-from-sp> last accessed 11 March 2019. 155 Schwarcz and Schwarcz (n 152) 1587. 156 Ibid 1591. 157 Ibid. 158 Dodd-Frank Wall Street Reform and Consumer Protection Act, Pub L No 111-203, s 112, 124 Stat 1376, 1394–95 (2010). 159 Ibid s 113. 160 Authority to Require Supervision and Regulation of Certain Nonbank Financial Companies, 84 Fed Reg 9028 (13 March 2019). 161 Ibid 9039–40 (‘The Council will prioritize its efforts to identify, assess, and address potential risks and threats to U.S. financial stability through a process that emphasizes an activities-based approach, and will pursue entity-specific determinations under section 113 of the Dodd-Frank Act only if a potential risk or threat cannot be addressed through an activities-based approach’). 162 Ibid 9041 (‘If the Council’s collaboration and engagement with the relevant financial regulatory agencies does not adequately address a potential threat identified by the Council—or if a potential threat to U.S. financial stability is outside the jurisdiction or authority of financial regulatory agencies—and if the potential threat identified by the Council is one that could be addressed by a Council determination regarding one or more companies, the Council may evaluate one or more nonbank financial companies for an entity-specific determination under section 113 of the Dodd-Frank Act’). 163 The FSOC is explicitly authorized to take such measures up to and including divestment under the Dodd-Frank Act. Dodd-Frank Wall Street Reform and Consumer Protection Act, Pub L No 111-203, s 121, 124 Stat 1376, 1410 (2010). 164 Steven L Schwarcz, ‘Ring-fencing’ (2013) 87 Southern California Law Review 69, 72. 165 Enhanced Prudential Standards (Regulation YY), 12 CFR s 252.153(a) (2018). 166 Ibid s 252.153(b). 167 Ibid s 252.153(e). With the passage of Economic Growth, Regulatory Relief, and Consumer Protection Act, Pub L No 115-174 (2018), which raised the asset threshold for application of enhanced prudential standards from $50 billion to $250 billion, it is possible that the asset threshold for the intermediate holding company requirement would be raised as well. 168 Dodd-Frank Wall Street Reform and Consumer Protection Act, Pub L No 111-203, s 716, 124 Stat 1376, 1648 (2010). 169 Ibid s 619(a)(1). 170 Independent Commission on Banking, Final Report: Recommendations (2011) 35–36. 171 Financial Services (Banking Reform) Act 2013. 172 Financial Services and Markets Act 2000 (Ring-fenced Bodies and Core Activities) Order 2014, SI 2014/1960, art 12(1). ‘Core deposit’ generally refers to deposits from individuals (excluding certain wealthy individuals) and small businesses. Ibid art 2(2). 173 Financial Services (Banking Reform) Act 2013, s 4; Financial Services and Markets Act 2000 (Excluded Activities and Prohibitions) Order 2014, SI 2014/2080, art 4. The Excluded Activities and Prohibitions Order also specifies exposures (e.g., financial institution exposure with certain exceptions) prohibited for ringfenced entities. Ibid art 14. 174 Katie Britton and others, ‘Ring-fencing: What It Is and How Will It Affect Banks and Their Customers?’ [2016 Q4] Bank of England Quarterly Bulletin 164, 166. 175 Ibid 167. 176 High Level Expert Group on Reforming the Structure of the EU Banking Sector, Final Report (2012); European Commission, ‘Proposal for a Regulation on Structural Measures Improving the Resilience of EU Credit Institutions’ COM (2014) 43 final. The EU ringfencing proposal, however, was withdrawn in 2017, as no agreement on the proposal was ‘foreseeable’. European Commission, ‘Annex 4 to the Commission Work Programme 2018: An Agenda for a More United, Strong and More Democratic Europe’ COM (2017) 650 final. 177 COM (2014) 43 (n 176) art 6. See also Winthrop N Brown, ‘With This Ring, I Thee Fence: How Europe’s Ringfencing Proposal Compares with U.S. Ringfencing Measures’ (2014) 45 Georgetown Journal of International Law 1029, 1034. 178 COM (2014) 43 (n 176) arts 9, 10(3). 179 Ibid art 13. 180 Ibid art 13, 14(2). 181 Ibid art 13(4). 182 Arthur E Wilmarth, Jr, ‘Narrow Banking as a Structural Remedy for the Problem of Systemic Risk: A Comment on Professor Schwarcz’s Ring-fencing’ (2014) 88 Southern California Law Review 1, 7–8. 183 Schwarcz (n 164) 106–07. 184 Alexander Coley, ‘U.S. Regulation of Cross-Border Banks: Is It Time to Embrace Balkanization in Global Finance?’ (2016) 56 Virginia Journal of International Law 701, 736–37. 185 Schwarcz (n 164) 84–95. 186 Wilson Ervin, ‘Understanding “Ring-fencing” and How It Could Make Banking Riskier’ (Brookings Institution, 7 February 2018) <www.brookings.edu/research/understanding-ring-fencing-and-how-it-could-make-banking-riskier> last accessed 11 March 2019. 187 Schwarcz (n 164) 99. Author notes The views expressed in this article do not necessarily represent the views of the Government of the Republic of Korea or of any of the author's prior employers. Gabriel Rosenberg provided valuable guidance and feedback, and anonymous referees also provided immensely helpful comments. © The Author(s) 2019. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - The Myth of Diversification: The Destabilizing Impact of Diversification on Financial Institutions
JF - Journal of Financial Regulation
DO - 10.1093/jfr/fjz007
DA - 2019-12-31
UR - https://www.deepdyve.com/lp/oxford-university-press/the-myth-of-diversification-the-destabilizing-impact-of-hog0430nNS
SP - 179
VL - 5
IS - 2
DP - DeepDyve
ER -