# How Management Risk Affects Corporate Debt

How Management Risk Affects Corporate Debt Abstract We evaluate whether management risk, which arises from investors’ uncertainty about management’s added value, affects firms’ default risks and debt pricing. We find that, regardless of the reason for the turnover, CDS, loan, and bond yield spreads increase at the time of management turnover, when management risk is highest, and decline over the first three years of the new CEO’s tenure. The effects increase with prior investor uncertainty about the new management. These results are consistent with the view that management risk affects firms’ default risk. An understanding of management risk yields a number of implications for corporate finance. Received May 15, 2016; editorial decision February 27, 2017 by Editor David Denis. Introduction Boards of directors typically appoint management teams with some knowledge of their particular visions, skill sets, and the types of policies they are likely to adopt. Nonetheless, regardless of any management team’s qualifications, its success is not guaranteed. Sometimes incoming managers adopt a new strategy that ex ante appears to be a good idea, but has unforeseen difficulties once implemented. Other times, managers themselves can be poor fits with the firm, or important firm-specific knowledge does not smoothly transfer to the new management regime. The uncertainty about a management team’s impact on firm value, which comes from both the policies the team is likely to bring to their firms and their ability to implement them personally, is what we call “management risk”. Since default risk reflects not only the likelihood that a firm will have bad luck but also the risk that the firm’s managerial decisions will lead the firm to default, management risk can affect firms’ default risk. Practitioners have long understood the importance of management risk, and regularly characterize it as an important factor affecting a firm’s credit risk.1 However, the academic literature on corporate default risk and the pricing of corporate debt has largely ignored management risk.2 This paper empirically evaluates the importance of management risk in determining a firm’s default risk and the pricing of its debt. We identify the effect of management risk using the idea that uncertainty about future managerial decisions rises around executive turnovers, particularly chief executive officer (CEO) turnovers, and decreases over time while the manager’s actions are observed. When a senior manager departs, there is an immediate increase in the uncertainty about who his replacement will be and the impact the new manager will have on firm value. Part of this uncertainty is resolved when the incoming manager’s identity is revealed, but substantial uncertainty remains about his ability and the quality of match between his managerial approach and the firm. If the ex ante expectation of a manager’s quality is, on average, correct, then there should be no systematic change in the market’s estimate of an average manager’s ability over his tenure in office. What will decline unambiguously, however, is the noise in this estimate, since additional observations of his actions will allow the market to learn more about the manager. Therefore, management risk that arises because of investors’ uncertainty about the manager’s value added should decline with a manager’s tenure. If management risk increases the market’s assessment of a firm’s default probability, then the default risk embedded in the pricing of the firm’s debt should also increase around the time of executive turnover and subsequently decline over the executive’s tenure. Using a sample of primarily S&P 1500 firms between 1992 and 2012, we characterize the way that the risk of a firm’s debt varies with the uncertainty the market is likely to have about its management. The basic pattern is depicted in Figure 1, which illustrates the way CDS spreads change around the time of a CEO change. The announcement of a CEO’s departure is associated with an increase in the firm’s CDS spread, reflecting an increased market assessment of the firm’s default risk. The CDS spread declines when a new CEO arrives, and further declines during the new CEO’s time in office, approximately back to the preturnover level after about three years. Holding other factors constant, the five-year CDS spread is 36 basis points (24% relative to the sample mean) higher when a new CEO takes office than three years into his tenure. Spreads on loans and bond yield spreads also decline following CEO turnovers. Figure 1 View largeDownload slide Average CDS spreads at key events related to CEO turnovers Figure 1 View largeDownload slide Average CDS spreads at key events related to CEO turnovers Identifying the impact of management risk on credit spreads is complicated by the possibility that turnovers tend to occur at times when default risk is high. Ideally, to measure the impact of management risk on debt pricing, one would like to observe a sample in which CEO turnovers occur randomly, a phenomenon that does not occur in reality. Instead, we consider a number of subsamples for which there is unlikely to be a spurious relation between turnovers and spreads, including cases in which the outgoing CEO exits because of death or illness or when the firm is having above average performance. In each case, we find a pattern in spreads that is not economically or statistically significantly different from the pattern observed in the entire sample, suggesting that the results are unlikely to be driven by the endogenous timing of turnovers. The CEO, however, is not the only member of the management team that is relevant for decision-making in the firm. Chief financial officers (CFOs) potentially impact firms’ policies, albeit to a lesser extent than CEOs. Our estimates indicate that, similar to CEOs, spreads on a firm’s CDS and new debt decline over the first three years of its CFO’s tenure, but the magnitude of the decline is 58% smaller than that following CEO turnovers, if the CFO turnover is not accompanied by a CEO turnover. The observed decline in default risk over tenure potentially reflects the resolution of uncertainty about management and hence a decline in management risk. To further evaluate whether this interpretation is the appropriate one, we examine cross-sectional variation in the way that ex ante uncertainty gets resolved across CEOs and firms. Bayesian learning models imply that if the changes in spreads around CEO turnover occur because of changes in management risk, then when ex ante uncertainty about management is higher, spreads should increase more around management turnover and decline faster subsequently. More specifically, management risk concerns the uncertainty about the new manager’s talent and skills, about the strategic decisions that he or she is likely to make, and about the efficiency in transferring control to the new management regime. When the manager is younger and thus has a shorter track record, management risk is likely to be higher. In addition, when the incoming CEO is an outsider who is likely to implement different policies and be subject to greater uncertainty in the leadership transition process, management risk is also likely to be high. In contrast, when there is an “heir apparent” who is groomed for leadership and likely to continue the same policies as his predecessor, management risk is likely to be low. Consistent with these predictions, our estimates suggest that the increase in the CDS spread at the time of the CEO departure announcement, the change in the spread when the incoming CEO takes office, as well as the sensitivity of the spread to the new CEO’s tenure, all depend on the amount of uncertainty there is about the impact of the new management. Furthermore, our estimates suggest that part of the uncertainty about the new management gets resolved between the old CEO’s departure and the new CEO’s arrival. However, the decline in spreads during this transition period is smaller if the new CEO is younger and if the new CEO is an outsider. Presumably less is known about younger CEOs or the quality of match between outsider CEOs and the firms at the time they take office. But once a younger CEO or an outsider CEO does take over, the market learns more about his ability from observing his performance, so the spreads decline faster. In addition, when the CEO has an existing relationship with a lender before he or she takes the current job, the lender is likely to know more about the CEO’s ability and future actions, leading to lower management risk. Consistent with this argument, we find that the sensitivity of interest rates to the CEO’s time in office is 60%-73% lower for loans in which the CEO has a prior relationship with the lender compared to those without such a relationship. This relation holds even if the CEO is an outsider and the relationship was built while he or she worked at a different firm, so this effect is different from the firm-lender relationship. Further, any additional management-induced risk should have a larger impact on the default risk and the pricing of riskier debt than of safer debt. Consistent with this prediction, we find that the firm’s spreads are more sensitive to CEO tenure when it is more highly levered, for term loans and for junior bonds. Overall, the cross-sectional evidence is consistent with the notion that the decline in spreads over executive tenure reflects the resolution of uncertainty about management. Understanding the way management risk affects corporate default risk and the pricing of corporate debt has a number of implications. First, our study identifies an important yet unexplored source of corporate default risk and a potentially important determinant of the pricing of corporate debt. The corporate finance literature on corporate debt pricing has focused on variables intended to capture risks coming from economy-wide factors, and those correlated with the nature of firms’ assets or information environment (see, e.g., van Binsbergen, Graham, and Yang 2010; Mansi, Maxwell, and Miller 2011). A parallel literature in asset pricing models a firm’s credit risk, usually again as a function of economy-wide factors and firms’ assets. However, Collin-Dufresne, Goldstein, and Martin (2001) find that these traditional credit risk factors and liquidity measures fail to explain the bulk part of the credit spread changes. Our analysis suggests that the magnitude of the management risk effect on firms’ CDS spread is comparable to that of firm characteristics, such as leverage and profitability, and is even larger than that of macroeconomic factors, such as credit spread and VIX. Thus, models predicting credit risk could be meaningfully improved by including variables that capture management risk, such as the CEO’s tenure, and CEO’s background, such as CEO age and whether he or she is an heir apparent. Second, management risk can explain why creditors are concerned about management changes and sometimes include a change of management restriction (CMR) clause in debt contracts that gives debtholders control rights in the event of a management change. Akins, DeAngelis, and Gaulin (2016) find that 8.5% of bank loan contracts include a CMR clause and provide evidence suggesting that this clause is more likely to be put in place to protect the interest of lenders than to entrench management. Third, our study highlights the importance of managing the management risk in a firm. Managerial succession planning, policies that improve the efficiency of firm-specific knowledge transfer across leadership, and transparency in managerial policies are practices that can reduce the uncertainty the market has about the firm’s management, and consequently lower the firm’s default risk and the interest rates it will pay on its debt. Consistent with this view, since 2009, the U.S. Securities and Exchange Commission (SEC) has recognized that “CEO succession planning raises a significant policy issue regarding the governance of the corporation” and “one of the board’s key functions is to provide for succession planning so that the company is not adversely affected due to a vacancy in leadership.”3 Fourth, management risk also has implications about the way a firm’s financial management policies may vary over CEO tenure. Since management risk increases the volatility of cash flows, it should increase the demand for precautionary savings. Firms facing higher management risk should have higher cash holdings, and cash holdings should decline with CEO tenure for firms for which management risk is likely to be high. Furthermore, since management risk increases default probabilities and expected distress costs, it should, holding other factors constant, lead firms to use less debt to finance investments when uncertainty about management is likely to be high.4 Fifth, and finally, the effect of management risk on corporate debt pricing can be used to assess the relative value impact of different managerial positions. For example, our estimates suggest that the impact on debt price from the uncertainty about CFO’s value added is about 58% of that from the uncertainty about CEO’s value added, suggesting that CEOs, on average, have a much larger value impact than CFOs.5 In addition, that the difference in the impact of tenure on spreads between insider and outsider CFOs is not significant, whereas the difference between insider and outsider CEOs is significant, suggests that the managerial skills required by the CFO job are more general and transferrable than those required by the CEO job. These results complement prior studies using interview scores or employment history to infer the generality of managerial skills and their value impact (Kaplan, Klebanov, and Sorensen 2012; Custódio, Ferreira, and Matos 2013). 1. Data 1.1 Risk of corporate debt The price of corporate debt is in large part determined by the likelihood that the firm’s future cash flows will be insufficient to cover the promised payments to debtholders. When management’s policies become more uncertain, the firm’s cash flow distribution becomes more dispersed, so the likelihood of default and the loss conditional on default are likely to increase. For this reason, we expect management risk to affect the firm’s default spread. 1.1.1 CDS spread. One way to measure a firm’s default spread is through the CDS spread that is traded on a firm’s debt. The payoff from the CDS contract occurs when the firm defaults on its debt, so the market-clearing price on the CDS contract reflects the market’s expectation that the debt will default and its recovery rate if in default. This default spread is also embedded in the promised yield on a firm’s debt. We measure the promised yield by the interest rate that the firm pays on its new loans, and the yield to maturity on its newly issued bonds. CDS spreads have several advantages for the purpose of measuring management risk. Blanco et al. (2005) document that the CDS and bond yield spreads are close to each other over long intervals, but over short intervals, CDS spreads tend to respond more quickly to changes in credit conditions. In addition, CDS spread data are available at the daily frequency, so these data allow us to measure changes in risk over relatively short intervals. However, many firms do not have CDS contracts traded on their debt, and the CDS data are only available since 2001. To mitigate any potential selection bias, we also conduct analyses using loan spreads, bond spreads and credit ratings, which do not condition on CDS markets. Our CDS data are provided by MarkIt,6 a comprehensive data source that assembles a network of industry-leading partners who contribute information for about 2,600 CDS on a daily basis. Based on the contributed quotes, MarkIt creates a daily composite quote for each CDS contract. We use the five-year spreads in our main specifications because these contracts are the most liquid and constitute over 85 percent of the entire CDS market. For robustness, we also use the one-year and three-year CDS spreads in some specifications. To maintain uniformity in contracts, we only keep CDS quotations for senior unsecured debt, which makes up over 91% of the entire CDS sample in MarkIt, with a modified restructuring (MR) clause and denominated in U.S. dollars.7 The first section of panel A of Table 1 reports the statistics on sample firms’ CDS spreads at the daily frequency over 794 CEOs’ first ten years in office in 486 firms (the CEO sample is described in Section 2.2).8 The average five-year CDS spread in our sample is 152 basis points (median 70). Table 1 Summary statistics A. Loan, bond, and CDS attributes Variables Obs. SD Mean p25 p50 p75 5-year CDS spread 757,434 223.61 152.46 36.52 70.23 165.15 3-year CDS spread 674,057 215.99 100.94 11.93 29.92 84.38 1-year CDS spread 698,214 225.94 129.93 23.86 49.98 127.96 Daily stock price 757,434 37.23 41.09 21.65 35.90 52.93 Recovery rate 757,434 3.08 39.64 39.38 40.00 40.00 Loan spread 11,288 126.87 154.96 50.00 125.00 225.00 Loan maturity 11,288 27.19 43.61 14.00 48.00 60.00 log(loan size) 11,288 1.37 5.46 4.61 5.52 6.35 Number of lenders 11,281 9.68 10.43 4.00 8.00 14.00 Number of covenants 1,965 1.80 4.54 4.00 5.00 6.00 Performance pricing 11,288 0.50 0.51 0.00 1.00 1.00 Refinancing 9,184 0.38 0.83 1.00 1.00 1.00 Secured 7,250 0.49 0.58 0.00 1.00 1.00 Prior CEO-lender relationship 7,718 0.44 0.27 0.00 0.00 1.00 Speculative frade 8,072 0.49 0.39 0.00 0.00 1.00 Term loan 11,288 0.41 0.21 0.00 0.00 0.00 Yield spread 4,849 218.39 201.61 62.89 135.73 269.00 Bond maturity 4,849 126.36 138.86 61.00 120.00 122.00 log(bond size) 4,849 1.29 5.72 5.30 5.84 6.40 Subordinated 4,849 0.34 0.13 0.00 0.00 0.00 Bond rating 4,478 1.23 4.30 4.00 5.00 5.00 A. Loan, bond, and CDS attributes Variables Obs. SD Mean p25 p50 p75 5-year CDS spread 757,434 223.61 152.46 36.52 70.23 165.15 3-year CDS spread 674,057 215.99 100.94 11.93 29.92 84.38 1-year CDS spread 698,214 225.94 129.93 23.86 49.98 127.96 Daily stock price 757,434 37.23 41.09 21.65 35.90 52.93 Recovery rate 757,434 3.08 39.64 39.38 40.00 40.00 Loan spread 11,288 126.87 154.96 50.00 125.00 225.00 Loan maturity 11,288 27.19 43.61 14.00 48.00 60.00 log(loan size) 11,288 1.37 5.46 4.61 5.52 6.35 Number of lenders 11,281 9.68 10.43 4.00 8.00 14.00 Number of covenants 1,965 1.80 4.54 4.00 5.00 6.00 Performance pricing 11,288 0.50 0.51 0.00 1.00 1.00 Refinancing 9,184 0.38 0.83 1.00 1.00 1.00 Secured 7,250 0.49 0.58 0.00 1.00 1.00 Prior CEO-lender relationship 7,718 0.44 0.27 0.00 0.00 1.00 Speculative frade 8,072 0.49 0.39 0.00 0.00 1.00 Term loan 11,288 0.41 0.21 0.00 0.00 0.00 Yield spread 4,849 218.39 201.61 62.89 135.73 269.00 Bond maturity 4,849 126.36 138.86 61.00 120.00 122.00 log(bond size) 4,849 1.29 5.72 5.30 5.84 6.40 Subordinated 4,849 0.34 0.13 0.00 0.00 0.00 Bond rating 4,478 1.23 4.30 4.00 5.00 5.00 B. CEO turnovers Became CEO year # of turnovers in the loan sample # of turnovers in the bond sample # of turnovers in the CDS sample 1992–1996 671 304 90 1997–2001 852 443 262 2002–2006 754 370 280 2007–2010 356 205 168 Total 2,633 1,322 800 B. CEO turnovers Became CEO year # of turnovers in the loan sample # of turnovers in the bond sample # of turnovers in the CDS sample 1992–1996 671 304 90 1997–2001 852 443 262 2002–2006 754 370 280 2007–2010 356 205 168 Total 2,633 1,322 800 C. CEO time in office Obs. Mean p25 Median p75 CEO total time in office (in years) 2,888 5.7 3 5 8 C. CEO time in office Obs. Mean p25 Median p75 CEO total time in office (in years) 2,888 5.7 3 5 8 D. CEO turnover types and CEO characteristics # of turnovers (1) Health/death 87 (2) Health/death/retirement at good performance 126 (3) No mgt. shakeup 465 (4) Good preturnover performance 328 (5) No preturnover run-up in CDS spread 155 (6) Outright forced 203 (7) Non-heir-apparent CEO 2,184 (8) Outsider CEO 739 (9) Young CEO 1,022 D. CEO turnover types and CEO characteristics # of turnovers (1) Health/death 87 (2) Health/death/retirement at good performance 126 (3) No mgt. shakeup 465 (4) Good preturnover performance 328 (5) No preturnover run-up in CDS spread 155 (6) Outright forced 203 (7) Non-heir-apparent CEO 2,184 (8) Outsider CEO 739 (9) Young CEO 1,022 E. Firm and macro-level attributes Variables Obs. SD Mean p25 p50 p75 Credit spread 5,269 44.10 96.23 69.00 85.00 108.00 Term spread 5,255 94.62 119.63 29.00 126.00 206.00 VIX 5,289 8.38 20.42 14.38 18.79 24.02 Log(assets) 25,274 1.86 7.33 6.06 7.30 8.59 Leverage 25,162 0.22 0.24 0.06 0.21 0.35 M/B 24,531 4.50 2.95 1.33 2.06 3.39 ROA 24,345 0.17 0.11 0.07 0.12 0.18 Tangibility 24,272 0.24 0.26 0.07 0.19 0.40 CF volatility 22,108 1.00 0.58 0.25 0.29 0.44 Payout ratio 25,439 0.45 0.21 0.00 0.00 0.30 E. Firm and macro-level attributes Variables Obs. SD Mean p25 p50 p75 Credit spread 5,269 44.10 96.23 69.00 85.00 108.00 Term spread 5,255 94.62 119.63 29.00 126.00 206.00 VIX 5,289 8.38 20.42 14.38 18.79 24.02 Log(assets) 25,274 1.86 7.33 6.06 7.30 8.59 Leverage 25,162 0.22 0.24 0.06 0.21 0.35 M/B 24,531 4.50 2.95 1.33 2.06 3.39 ROA 24,345 0.17 0.11 0.07 0.12 0.18 Tangibility 24,272 0.24 0.26 0.07 0.19 0.40 CF volatility 22,108 1.00 0.58 0.25 0.29 0.44 Payout ratio 25,439 0.45 0.21 0.00 0.00 0.30 Panel A reports the summary statistics of CDS, loan, and bond attributes during the first ten years of CEO tenure, conditional on covariates (in the ten-year spline regressions) nonmissing. Loan-level and bond-level variables, such as the loan spread and yield spread, are calculated when loans are initiated or when bonds are issued for the sample period 1992 to 2012. Panel B reports the distribution of CEO turnovers over time for CEOs in the three samples in panel A. Panel C reports the distribution of CEO’s total time in office (in years) for CEOs in the union of the above three samples. Panel D reports the number of various CEO turnovers based on turnover reason (see Table A1 in the appendix for more details), succession origin, and CEO age at turnover, for the CEO sample in panel B. Panel E reports the summary statistics of firm attributes (yearly, 1992–2012) for all Execucomp firms that had turnovers between 1992 and 2012, as well as the credit market conditions (credit spread, term spread, VIX at the daily level). The CDS variables are measured at the daily frequency for the sample period 2001–2012. Table 1 Summary statistics A. Loan, bond, and CDS attributes Variables Obs. SD Mean p25 p50 p75 5-year CDS spread 757,434 223.61 152.46 36.52 70.23 165.15 3-year CDS spread 674,057 215.99 100.94 11.93 29.92 84.38 1-year CDS spread 698,214 225.94 129.93 23.86 49.98 127.96 Daily stock price 757,434 37.23 41.09 21.65 35.90 52.93 Recovery rate 757,434 3.08 39.64 39.38 40.00 40.00 Loan spread 11,288 126.87 154.96 50.00 125.00 225.00 Loan maturity 11,288 27.19 43.61 14.00 48.00 60.00 log(loan size) 11,288 1.37 5.46 4.61 5.52 6.35 Number of lenders 11,281 9.68 10.43 4.00 8.00 14.00 Number of covenants 1,965 1.80 4.54 4.00 5.00 6.00 Performance pricing 11,288 0.50 0.51 0.00 1.00 1.00 Refinancing 9,184 0.38 0.83 1.00 1.00 1.00 Secured 7,250 0.49 0.58 0.00 1.00 1.00 Prior CEO-lender relationship 7,718 0.44 0.27 0.00 0.00 1.00 Speculative frade 8,072 0.49 0.39 0.00 0.00 1.00 Term loan 11,288 0.41 0.21 0.00 0.00 0.00 Yield spread 4,849 218.39 201.61 62.89 135.73 269.00 Bond maturity 4,849 126.36 138.86 61.00 120.00 122.00 log(bond size) 4,849 1.29 5.72 5.30 5.84 6.40 Subordinated 4,849 0.34 0.13 0.00 0.00 0.00 Bond rating 4,478 1.23 4.30 4.00 5.00 5.00 A. Loan, bond, and CDS attributes Variables Obs. SD Mean p25 p50 p75 5-year CDS spread 757,434 223.61 152.46 36.52 70.23 165.15 3-year CDS spread 674,057 215.99 100.94 11.93 29.92 84.38 1-year CDS spread 698,214 225.94 129.93 23.86 49.98 127.96 Daily stock price 757,434 37.23 41.09 21.65 35.90 52.93 Recovery rate 757,434 3.08 39.64 39.38 40.00 40.00 Loan spread 11,288 126.87 154.96 50.00 125.00 225.00 Loan maturity 11,288 27.19 43.61 14.00 48.00 60.00 log(loan size) 11,288 1.37 5.46 4.61 5.52 6.35 Number of lenders 11,281 9.68 10.43 4.00 8.00 14.00 Number of covenants 1,965 1.80 4.54 4.00 5.00 6.00 Performance pricing 11,288 0.50 0.51 0.00 1.00 1.00 Refinancing 9,184 0.38 0.83 1.00 1.00 1.00 Secured 7,250 0.49 0.58 0.00 1.00 1.00 Prior CEO-lender relationship 7,718 0.44 0.27 0.00 0.00 1.00 Speculative frade 8,072 0.49 0.39 0.00 0.00 1.00 Term loan 11,288 0.41 0.21 0.00 0.00 0.00 Yield spread 4,849 218.39 201.61 62.89 135.73 269.00 Bond maturity 4,849 126.36 138.86 61.00 120.00 122.00 log(bond size) 4,849 1.29 5.72 5.30 5.84 6.40 Subordinated 4,849 0.34 0.13 0.00 0.00 0.00 Bond rating 4,478 1.23 4.30 4.00 5.00 5.00 B. CEO turnovers Became CEO year # of turnovers in the loan sample # of turnovers in the bond sample # of turnovers in the CDS sample 1992–1996 671 304 90 1997–2001 852 443 262 2002–2006 754 370 280 2007–2010 356 205 168 Total 2,633 1,322 800 B. CEO turnovers Became CEO year # of turnovers in the loan sample # of turnovers in the bond sample # of turnovers in the CDS sample 1992–1996 671 304 90 1997–2001 852 443 262 2002–2006 754 370 280 2007–2010 356 205 168 Total 2,633 1,322 800 C. CEO time in office Obs. Mean p25 Median p75 CEO total time in office (in years) 2,888 5.7 3 5 8 C. CEO time in office Obs. Mean p25 Median p75 CEO total time in office (in years) 2,888 5.7 3 5 8 D. CEO turnover types and CEO characteristics # of turnovers (1) Health/death 87 (2) Health/death/retirement at good performance 126 (3) No mgt. shakeup 465 (4) Good preturnover performance 328 (5) No preturnover run-up in CDS spread 155 (6) Outright forced 203 (7) Non-heir-apparent CEO 2,184 (8) Outsider CEO 739 (9) Young CEO 1,022 D. CEO turnover types and CEO characteristics # of turnovers (1) Health/death 87 (2) Health/death/retirement at good performance 126 (3) No mgt. shakeup 465 (4) Good preturnover performance 328 (5) No preturnover run-up in CDS spread 155 (6) Outright forced 203 (7) Non-heir-apparent CEO 2,184 (8) Outsider CEO 739 (9) Young CEO 1,022 E. Firm and macro-level attributes Variables Obs. SD Mean p25 p50 p75 Credit spread 5,269 44.10 96.23 69.00 85.00 108.00 Term spread 5,255 94.62 119.63 29.00 126.00 206.00 VIX 5,289 8.38 20.42 14.38 18.79 24.02 Log(assets) 25,274 1.86 7.33 6.06 7.30 8.59 Leverage 25,162 0.22 0.24 0.06 0.21 0.35 M/B 24,531 4.50 2.95 1.33 2.06 3.39 ROA 24,345 0.17 0.11 0.07 0.12 0.18 Tangibility 24,272 0.24 0.26 0.07 0.19 0.40 CF volatility 22,108 1.00 0.58 0.25 0.29 0.44 Payout ratio 25,439 0.45 0.21 0.00 0.00 0.30 E. Firm and macro-level attributes Variables Obs. SD Mean p25 p50 p75 Credit spread 5,269 44.10 96.23 69.00 85.00 108.00 Term spread 5,255 94.62 119.63 29.00 126.00 206.00 VIX 5,289 8.38 20.42 14.38 18.79 24.02 Log(assets) 25,274 1.86 7.33 6.06 7.30 8.59 Leverage 25,162 0.22 0.24 0.06 0.21 0.35 M/B 24,531 4.50 2.95 1.33 2.06 3.39 ROA 24,345 0.17 0.11 0.07 0.12 0.18 Tangibility 24,272 0.24 0.26 0.07 0.19 0.40 CF volatility 22,108 1.00 0.58 0.25 0.29 0.44 Payout ratio 25,439 0.45 0.21 0.00 0.00 0.30 Panel A reports the summary statistics of CDS, loan, and bond attributes during the first ten years of CEO tenure, conditional on covariates (in the ten-year spline regressions) nonmissing. Loan-level and bond-level variables, such as the loan spread and yield spread, are calculated when loans are initiated or when bonds are issued for the sample period 1992 to 2012. Panel B reports the distribution of CEO turnovers over time for CEOs in the three samples in panel A. Panel C reports the distribution of CEO’s total time in office (in years) for CEOs in the union of the above three samples. Panel D reports the number of various CEO turnovers based on turnover reason (see Table A1 in the appendix for more details), succession origin, and CEO age at turnover, for the CEO sample in panel B. Panel E reports the summary statistics of firm attributes (yearly, 1992–2012) for all Execucomp firms that had turnovers between 1992 and 2012, as well as the credit market conditions (credit spread, term spread, VIX at the daily level). The CDS variables are measured at the daily frequency for the sample period 2001–2012. 1.1.2 Loan spread data We retrieve data for bank loans occurring between 1992 and 2012 from DealScan,9 which is maintained by Thomson Reuters’ Loan Pricing Corporation (LPC).10 We match the borrowers to the firms in our sample using a procedure described by Chava and Roberts (2008).11 The second section of panel A of Table 1 reports loan-level statistics for loans taken by our sample CEOs during the first ten years of their tenure.12 The 2,569 CEOs, from 1,672 firms, initiated 11,288 loans for which DealScan reports nonmissing spreads. To measure the price of bank debt, we use the all-in-drawn spread (AIS) that the borrower pays over LIBOR at the loan origination date,13 winsorized at the top and the bottom 1% of the DealScan sample distribution. The mean of the loan spreads in our sample is 155 basis points, and the median is 125 basis points. We also report summary statistics for other components of the bank loan contracts, such as loan maturity, loan size, number of lenders, number of loan covenants, whether the loan has performance pricing, whether the loan is secured, whether the borrowing company has a speculative grade when the loan was initiated, and whether the loan is classified as “refinancing” by DealScan. Table A1 (see the appendix) reports the detailed variable definitions. 1.1.3 Corporate bond yield spread data The corporate bond data, from 1992 to 2012, are taken from the Mergent Fixed Investment Securities Database (FISD),14 a comprehensive database of publicly offered U.S. bonds. FISD provides details on debt issues and the issuers. The third section of panel A of Table 1 reports statistics for bonds issued during the first ten years of CEO tenure. There are 4,849 public bonds with available data on offering yield, which were issued by 1,299 CEOs from 942 firms.15 To measure the bond yield spread, we use the offering yield of a corporate bond at issue minus the yield of the maturity-matched Treasury bond. We winsorize the spreads at the top and the bottom 1% of the entire FISD sample distribution. When the maturity of the bond for which the spread is calculated does not exactly match the maturity of the available government bonds, we use linear interpolation to estimate the yield of the risk-free benchmark. The average bond yield spread in our sample is 202 basis points (median 136). Summary statistics for other bond characteristics including bond rating, bond maturity, offering size, and whether the bond is subordinated are also reported in panel A of Table 1. 1.2 CEO turnover and tenure We construct a sample of CEOs of large, publicly traded U.S. firms, each of whom became CEO between 1992 and 2010. We use the information on job title, the year becoming CEO, and the CEO annual flag provided in Execucomp to identify CEOs at the firm-year level, from which we identify whether there is a CEO turnover in a firm and year. Panel B of Table 1 describes the distribution of turnovers over time in the loan, bond, and CDS samples. For each CEO, the variable “Tenure (in years)” equals zero for the fiscal year in which the CEO takes office, and increases with each year the CEO is in office. The average CEO’s total time in office (see Table A1 in the appendix for the definition) in our sample, reported in Panel C of Table 1, is 5.7 years and the median is five years. The variable “Tenure (in days)” is constructed similarly, so is set to zero for the day the CEO officially takes office, and increases with each day the CEO is in office. The incoming CEO’s background is likely to be related to the amount of uncertainty about his ability. We identify two dimensions about the CEO’s background that are potentially related to such uncertainty: the CEO’s age and his prior position. The average age of the incoming CEO at the time of turnover in our sample is 51. We thus classify new CEOs who are younger than 51 at the time of turnover as “Young CEOs.” Using information on the time of a CEO “joining company” from Execucomp, supplemented by the data from Boardex, we classify CEOs who have not worked for the firm prior to becoming CEO as “Outsider CEOs,” and others as insider CEOs. We also follow Naveen (2006) and classify “Heir-apparent CEOs” in our sample as executives with the title “president” or “chief operating officer (COO)” prior to becoming CEO. Panel D of Table 1 reports the number of turnovers that involve young CEOs, outsider CEOs, or heir apparent CEOs. 1.3 Other variables To control for other factors that potentially affect the loan, bond, or CDS spreads, we include a set of firm characteristics and credit market conditions in our empirical specifications, mostly following Graham et al. (2008) and van Binsbergen, Graham, and Yang (2010). For credit market conditions, we control for three variables: Credit spread is the difference between the yields of AAA and BAA corporate bonds; Term spread is the difference between the yields of ten-year Treasury bonds and two-year Treasury bonds; and VIX is the CBOE volatility index, which shows the market’s expectation of 30-day volatility. The first section in Table 1, panel E, reports summary statistics for the three variables. We obtain firm-specific variables from Compustat and CRSP, and winsorize them at the top and the bottom 1% of the distribution. The average firm in our sample has book assets of about $${\}$$1.5 billion, book leverage ratio of 0.24, a market-to-book equity ratio of 2.95, an asset tangibility ratio of 0.26, cash flow volatility of 0.58, return on assets (ROA) of 0.11, and dividend payout ratio of 0.21. The second portion of panel E of Table 1 reports summary statistics for these firm-specific measures, as well as other financial variables. Table A1 (see the appendix) presents detailed definitions of all variables. 2. Measuring Variation in Default Risk around Management Changes 2.1 CDS spreads around turnovers To evaluate whether uncertainty about a new CEO’s ability and policies affects the market’s expectation of a firm’s default risk, we first examine the way in which firms’ CDS spreads change over time. Since CDS spreads provide a market-based assessment of the likelihood that the firm will default on its debt at any point in time, the way they vary over a CEO’s tenure measures changes in expected default risk over this period. Figure 1 plots the average CDS spread around key events associated with the evolution of the uncertainty about a new CEO. It is constructed using 374 CEO turnovers of firms with available CDS data, information on the departure announcement dates of the outgoing CEOs and on the dates when the incoming CEOs took office. On the day when the departure is announced, the CDS spread increases by about 55 basis points relative to the average spread in the prior three months, likely reflecting the increase in the uncertainty about the management. The spread subsequently decreases by about 26 basis points when the new CEO takes office, indicating that part of the uncertainty about the management is resolved by the knowledge of the incoming CEO’s identity and possibly his managing style and agenda as well. During his first three years of office, it declines by another 36 basis points. The changes in expected default risk over the relatively short window when the information about the CEO succession is being revealed suggest that the CEO himself has a major impact on the market’s perception of the firm’s default risk. Figure 1 suggests that spreads tend to follow an inverted V-shaped pattern around CEO turnovers. However, whereas the company’s proxy statements usually disclose the date when the CEO takes office,16 it is not always possible to know the exact date when a CEO’s departure is announced and when the identity of his replacement becomes known to the market. For this reason, we focus most of our analysis on the period following new CEO’s appointment, by measuring the way a firm’s managerial-related risk changes over his time in office. We do, however, analyze preturnover changes in spreads in Section 4. 2.2 CDS spreads over CEO tenure To estimate the ways in which CDS spreads are affected by the resolution of uncertainty about the CEO’s ability in the first few years of his tenure, we estimate the following equation: $$\textit{CDS}\_{\textit{Spread}}_t^{\textit{ij}} = f (\textit{Tenure})_t^{\textit{ij}} + \alpha^{\textit{ij}} + \lambda_t + \textit{Controls}_t^i + \epsilon_t^{\textit{ij}}$$ (1) The variable “$$\textit{Tenure}_t^{\textit{ij}}$$” is CEO-$$j$$’s time in office in firm-$$i$$ at time $$t$$. To capture potential nonlinearities in the tenure-spread relation, we use a piecewise-linear (spline) specification that allows the relation to change over time. The variable $$\alpha^{\textit{ij}}$$ is a firm-CEO fixed effect for firm $$i$$ and CEO $$j$$; its inclusion implies that we identify the effect of managerial uncertainty from the time-series variation in CDS spreads within a particular firm-CEO pair. This approach, therefore, controls for any time-invariant differences cross firm-CEO pairs (e.g., the persistent part of its matching quality). The variable $$\lambda_t$$ is the calendar year fixed effect that controls for market-wide factors that affect firm level default risk. The time-varying controls include the debt recovery rate as reported by data distributors, the contemporaneous stock price, other firm-specific financial variables, as well as measures of credit market conditions, such as the aggregate credit spread, term spread, and the VIX index. Table 2 reports estimates of this equation, with CEO tenure measured as the number of days since the CEO officially took office. The specification presented in Column 1 uses a sample consisting of the first ten years of each CEO’s tenure and allows the coefficient on tenure to differ between the first 3 years of tenure and the subsequent years. The estimates of this specification indicate that a firm’s CDS spread declines by 0.031 basis points for each day in the CEO’s first three years in office. Over the first 1,095 calendar days (three years), the total decline amounts to 34 basis points, which is 22% of the sample mean of 152. The speed of decline in the CDS spread becomes statistically insignificant and small in magnitude after the first three years. Table 2 The effect of CEO tenure on CDS spread (1) (2) (3) (4) (5) Years [0,9] Years [0, 2] Years [0, 2], Before 2008 Years [0, 2] Years [0, 2] 5-year CDS spread 5-year CDS spread 5-year CDS spread 3-year CDS spread 1-year CDS spread Tenure –0.031*** (years 0-2) (0.011) Tenure –0.018 (years 3-5) (0.021) Tenure –0.003 (years 6-9) (0.011) Tenure –0.031** –0.044*** –0.058*** –0.080*** (in days) (0.015) (0.016) (0.018) (0.020) Stock price –1.369*** –0.962*** –0.892*** –0.943*** –0.953*** (0.188) (0.244) (0.247) (0.280) (0.299) Recovery rate –12.512*** –9.587*** –8.626*** –15.655*** –20.063*** (1.226) (1.442) (1.531) (1.967) (2.176) Credit spread 0.450*** 0.406*** 0.409*** 0.483*** 0.624*** (0.040) (0.065) (0.082) (0.080) (0.096) Term spread –0.017 –0.042 –0.005 –0.052 –0.008 (0.029) (0.043) (0.048) (0.051) (0.057) VIX 1.655*** 2.225*** 2.584*** 2.182*** 1.996*** (0.138) (0.248) (0.331) (0.290) (0.318) Log(assets) –20.458* 76.013** 52.854 97.048** 125.716*** (12.339) (32.443) (33.414) (38.390) (41.739) Leverage 210.559*** 82.981 40.336 128.798* 100.586 (46.624) (59.360) (59.425) (73.450) (78.873) M/B 9.251 14.190 16.734 23.549 31.336* (6.317) (12.748) (13.566) (17.416) (18.916) ROA –402.824*** –486.699*** –574.895*** –570.857*** –584.902*** (70.904) (115.805) (137.286) (136.150) (146.363) Tangibility 17.383 54.684 –233.555 228.103 259.161 (76.118) (149.272) (149.465) (175.092) (181.409) CF volatility 15.802*** 8.184 6.276 7.929 8.659 (4.426) (7.133) (7.515) (8.204) (9.709) Payout ratio –23.535*** –7.458 –8.204 –13.331** –14.017** (5.095) (4.958) (6.193) (5.901) (6.309) Firm-CEO,Yr FE x x x x x Observations 757,434 301,290 239,516 274,324 266,325 Adj. R$$^{\mathrm{2}}$$ 0.755 0.840 0.829 0.807 0.749 (1) (2) (3) (4) (5) Years [0,9] Years [0, 2] Years [0, 2], Before 2008 Years [0, 2] Years [0, 2] 5-year CDS spread 5-year CDS spread 5-year CDS spread 3-year CDS spread 1-year CDS spread Tenure –0.031*** (years 0-2) (0.011) Tenure –0.018 (years 3-5) (0.021) Tenure –0.003 (years 6-9) (0.011) Tenure –0.031** –0.044*** –0.058*** –0.080*** (in days) (0.015) (0.016) (0.018) (0.020) Stock price –1.369*** –0.962*** –0.892*** –0.943*** –0.953*** (0.188) (0.244) (0.247) (0.280) (0.299) Recovery rate –12.512*** –9.587*** –8.626*** –15.655*** –20.063*** (1.226) (1.442) (1.531) (1.967) (2.176) Credit spread 0.450*** 0.406*** 0.409*** 0.483*** 0.624*** (0.040) (0.065) (0.082) (0.080) (0.096) Term spread –0.017 –0.042 –0.005 –0.052 –0.008 (0.029) (0.043) (0.048) (0.051) (0.057) VIX 1.655*** 2.225*** 2.584*** 2.182*** 1.996*** (0.138) (0.248) (0.331) (0.290) (0.318) Log(assets) –20.458* 76.013** 52.854 97.048** 125.716*** (12.339) (32.443) (33.414) (38.390) (41.739) Leverage 210.559*** 82.981 40.336 128.798* 100.586 (46.624) (59.360) (59.425) (73.450) (78.873) M/B 9.251 14.190 16.734 23.549 31.336* (6.317) (12.748) (13.566) (17.416) (18.916) ROA –402.824*** –486.699*** –574.895*** –570.857*** –584.902*** (70.904) (115.805) (137.286) (136.150) (146.363) Tangibility 17.383 54.684 –233.555 228.103 259.161 (76.118) (149.272) (149.465) (175.092) (181.409) CF volatility 15.802*** 8.184 6.276 7.929 8.659 (4.426) (7.133) (7.515) (8.204) (9.709) Payout ratio –23.535*** –7.458 –8.204 –13.331** –14.017** (5.095) (4.958) (6.193) (5.901) (6.309) Firm-CEO,Yr FE x x x x x Observations 757,434 301,290 239,516 274,324 266,325 Adj. R$$^{\mathrm{2}}$$ 0.755 0.840 0.829 0.807 0.749 This table reports the changes in daily CDS spread over CEO tenure. The sample period is 2001–2012. In Column (1), we use the first ten years of CEO tenure. In Columns (2) to (5), we use the first three years of CEO tenure. In Columns (1) to (3), we use five-year CDS spreads. In Column (3), we include only CEO turnovers before the financial crisis ($$\leqslant$$2007). In Columns (4) and (5), we use three-year and one-year CDS spreads. CEO tenure is measured by days since the CEO takes office. All the control variables are measured contemporaneous to the CDS spreads. Table A1 (see the appendix) reports all variable definitions. Standard errors are clustered at the firm-year level. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. Table 2 The effect of CEO tenure on CDS spread (1) (2) (3) (4) (5) Years [0,9] Years [0, 2] Years [0, 2], Before 2008 Years [0, 2] Years [0, 2] 5-year CDS spread 5-year CDS spread 5-year CDS spread 3-year CDS spread 1-year CDS spread Tenure –0.031*** (years 0-2) (0.011) Tenure –0.018 (years 3-5) (0.021) Tenure –0.003 (years 6-9) (0.011) Tenure –0.031** –0.044*** –0.058*** –0.080*** (in days) (0.015) (0.016) (0.018) (0.020) Stock price –1.369*** –0.962*** –0.892*** –0.943*** –0.953*** (0.188) (0.244) (0.247) (0.280) (0.299) Recovery rate –12.512*** –9.587*** –8.626*** –15.655*** –20.063*** (1.226) (1.442) (1.531) (1.967) (2.176) Credit spread 0.450*** 0.406*** 0.409*** 0.483*** 0.624*** (0.040) (0.065) (0.082) (0.080) (0.096) Term spread –0.017 –0.042 –0.005 –0.052 –0.008 (0.029) (0.043) (0.048) (0.051) (0.057) VIX 1.655*** 2.225*** 2.584*** 2.182*** 1.996*** (0.138) (0.248) (0.331) (0.290) (0.318) Log(assets) –20.458* 76.013** 52.854 97.048** 125.716*** (12.339) (32.443) (33.414) (38.390) (41.739) Leverage 210.559*** 82.981 40.336 128.798* 100.586 (46.624) (59.360) (59.425) (73.450) (78.873) M/B 9.251 14.190 16.734 23.549 31.336* (6.317) (12.748) (13.566) (17.416) (18.916) ROA –402.824*** –486.699*** –574.895*** –570.857*** –584.902*** (70.904) (115.805) (137.286) (136.150) (146.363) Tangibility 17.383 54.684 –233.555 228.103 259.161 (76.118) (149.272) (149.465) (175.092) (181.409) CF volatility 15.802*** 8.184 6.276 7.929 8.659 (4.426) (7.133) (7.515) (8.204) (9.709) Payout ratio –23.535*** –7.458 –8.204 –13.331** –14.017** (5.095) (4.958) (6.193) (5.901) (6.309) Firm-CEO,Yr FE x x x x x Observations 757,434 301,290 239,516 274,324 266,325 Adj. R$$^{\mathrm{2}}$$ 0.755 0.840 0.829 0.807 0.749 (1) (2) (3) (4) (5) Years [0,9] Years [0, 2] Years [0, 2], Before 2008 Years [0, 2] Years [0, 2] 5-year CDS spread 5-year CDS spread 5-year CDS spread 3-year CDS spread 1-year CDS spread Tenure –0.031*** (years 0-2) (0.011) Tenure –0.018 (years 3-5) (0.021) Tenure –0.003 (years 6-9) (0.011) Tenure –0.031** –0.044*** –0.058*** –0.080*** (in days) (0.015) (0.016) (0.018) (0.020) Stock price –1.369*** –0.962*** –0.892*** –0.943*** –0.953*** (0.188) (0.244) (0.247) (0.280) (0.299) Recovery rate –12.512*** –9.587*** –8.626*** –15.655*** –20.063*** (1.226) (1.442) (1.531) (1.967) (2.176) Credit spread 0.450*** 0.406*** 0.409*** 0.483*** 0.624*** (0.040) (0.065) (0.082) (0.080) (0.096) Term spread –0.017 –0.042 –0.005 –0.052 –0.008 (0.029) (0.043) (0.048) (0.051) (0.057) VIX 1.655*** 2.225*** 2.584*** 2.182*** 1.996*** (0.138) (0.248) (0.331) (0.290) (0.318) Log(assets) –20.458* 76.013** 52.854 97.048** 125.716*** (12.339) (32.443) (33.414) (38.390) (41.739) Leverage 210.559*** 82.981 40.336 128.798* 100.586 (46.624) (59.360) (59.425) (73.450) (78.873) M/B 9.251 14.190 16.734 23.549 31.336* (6.317) (12.748) (13.566) (17.416) (18.916) ROA –402.824*** –486.699*** –574.895*** –570.857*** –584.902*** (70.904) (115.805) (137.286) (136.150) (146.363) Tangibility 17.383 54.684 –233.555 228.103 259.161 (76.118) (149.272) (149.465) (175.092) (181.409) CF volatility 15.802*** 8.184 6.276 7.929 8.659 (4.426) (7.133) (7.515) (8.204) (9.709) Payout ratio –23.535*** –7.458 –8.204 –13.331** –14.017** (5.095) (4.958) (6.193) (5.901) (6.309) Firm-CEO,Yr FE x x x x x Observations 757,434 301,290 239,516 274,324 266,325 Adj. R$$^{\mathrm{2}}$$ 0.755 0.840 0.829 0.807 0.749 This table reports the changes in daily CDS spread over CEO tenure. The sample period is 2001–2012. In Column (1), we use the first ten years of CEO tenure. In Columns (2) to (5), we use the first three years of CEO tenure. In Columns (1) to (3), we use five-year CDS spreads. In Column (3), we include only CEO turnovers before the financial crisis ($$\leqslant$$2007). In Columns (4) and (5), we use three-year and one-year CDS spreads. CEO tenure is measured by days since the CEO takes office. All the control variables are measured contemporaneous to the CDS spreads. Table A1 (see the appendix) reports all variable definitions. Standard errors are clustered at the firm-year level. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. Because the relation between CEO tenure and spreads appears to be most pronounced during the first three years of CEO tenure, we re-estimate the spread-tenure relation using just the first three years of tenure in Column 2. The coefficient on tenure during the first three years is identical to that using the longer sample, $$-$$.031, although the coefficients on the other variables in the equation do change to some extent. To provide perspective on the importance of management risk, we compare its impact on CDS spreads with the magnitudes of the effects from other significant macroeconomics or firm-level determinants in Column (1). These variables are credit spread, VIX, leverage rate, and ROA. We compute the effect of a one-standard-deviation increase of each determinant on CDS spread using its coefficient estimate and the within-sample standard deviation in Column (1). The estimated impact of a one standard deviation change in credit spread, VIX, leverage ratio, and ROA are 24.3, 16.9, 32.1, and 31.0 basis points, respectively. These estimates imply that the influence of management risk on spreads is comparable to those of important firm characteristics, and is even larger than measures of macroeconomic risk. One possible concern about the results is that the unusual events occurring during the 2008 financial crisis could be correlated with CDS spreads and also with CEO tenure in our sample. Although we do include year fixed effects and daily macro-level control variables—an addition that should minimize concerns about this issue—we also replicate our results using the pre-2008 turnover subsample. The results using this subsample are presented in Column (3) of Table 2. CDS spreads actually decline faster following turnovers in the pre-2008 subsample. This finding suggests that following turnovers that occurred during the financial crisis period, the CDS spreads declined more slowly over a CEO’s tenure, presumably because of the prolonged heightened credit risk as the result of economy-wide factors, rather than management risk. If uncertainty about management and its resolution is most pronounced in the first three years of a CEO’s tenure, then we expect the spreads on shorter-term CDS contracts to be even more sensitive to the changes in management risk around CEO turnover than longer-term CDS contracts. To test this conjecture, we replace five-year CDS spreads with three-year CDS spreads in Column (4) and one-year CDS spreads in Column (5) of Table 2. Indeed, the CDS spread-tenure sensitivity increases in magnitude to $$-$$0.058 for the three-year spreads and $$-$$0.080 for the one-year spreads. 2.3 Alternative interpretations 2.3.1 Endogenous timing of CEO turnover. An alternative interpretation of the declining default risk over CEO tenure is that CEO changes tend to occur at times of relatively high uncertainty that is unrelated to management changes, leading to heightened default risk around CEO changes. The extent to which this alternative interpretation is appropriate depends on turnovers coinciding with non-management-related high uncertainty. A particular concern is that CEO “firings” often occur at times of poor firm or industry performance (Jenter and Kannan 2015) when uncertainty tends to be high and valuation is low and, consequently, debt spreads are high. When the unusual period is over, spreads decline, leading to a pattern similar to that documented above. To address this concern, we follow Pan, Wang, and Weisbach (2015, 2016) and identify several subsamples of turnovers that are unlikely to coincide with heightened credit risk.17 These subsamples are constructed using completely different approaches from one another. If there are significant declines in default risk following multiple subsamples of turnovers that are unlikely to be performance motivated, then it is likely that these changes reflect uncertainty about incoming management’s skills and policies, rather than the circumstances under which management takes office. The first three subsamples we consider are determined by the circumstances of the turnover. First, we consider the group of turnovers caused by an illness or the death of the departing CEO.18 Second, we combine the death/illness subsample with retirements of CEOs older than 65. To mitigate the incidence of “suspicious” retirement announcements that could have actually occurred because of poor performance, we only include retirements for which the firm’s stock performance in the year prior to the turnover is above the sample median (0.3%). The third group consists of turnovers for which there is no change in the top management team, consisting of the top four most highly paid non-CEO executives, other than the CEO in the CEO turnover year. These turnovers are unlikely to be firings, since firings typically involve changes of other top managers in addition to the CEO. The final two subsamples are motivated directly by the alternative hypothesis, which argues that it is the firm’s financial condition that is correlated with the turnovers and also explain the pattern in expected default probabilities. The final two subsamples are explicitly constructed so that the firms are not likely to be in poor financial health, and we would not expect to observe a decline in default probabilities just because of the timing of these turnovers. Our fourth subsample is constructed to be the group of turnovers that are preceded by both good performance (both stock return and ROA above sample median) as well as low idiosyncratic volatility (below sample median), since these turnovers are unlikely to have been motivated by performance. Our fifth subsample consists of turnovers that were not preceded by a significant run-up of default risk as reflected in the firms’ CDS spreads in the prior two years. For each firm with CDS data in our sample, we estimate the time trend in CDS spreads during dates [$$-$$730, $$-$$30], with date 0 being the day when the new CEO takes office, and include turnovers with a negative or insignificantly positive preturnover CDS trend in this subsample. Finally, we use the Factiva news search to identify turnovers that appear to be overtly forced (e.g., Factiva reported that the CEO was forced to leave or exited under pressure). We include this subsample for comparison purposes, to document that forced turnovers are different from the ones in the subsamples we construct that are likely to be not performance motivated. Panel D of Table 1 reports the number of turnovers in each subsample that overlap with either our CDS, loan, or bond data. In Table 3, we report estimates of Equation (1) on subsamples constructed by the likely reason for the turnover of the outgoing CEO. These estimates suggest that the decline in the spreads on the firm’s CDS over CEO tenure occurs regardless of the factors leading to the CEO turnover. The estimated decline in CDS spread in a CEO’s first three years of tenure is significant across subsamples of likely non-performance-driven turnovers (Columns 1–5), as well as in the union of the subsamples in Columns (2) to (5) (Column 6). Table 3 Tenure-CDS spread relations following likely non-performance-driven versus forced turnovers (1) (2) (3) (4) (5) (6) (7) 5-year CDS spread, years [0,2] Death/illness Death/illness/ret. at good perf. No mgt. shakeup Good preturnover Perf. No preturnover run-up Union (2) - (5) Forced Tenure –0.033* –0.039** –0.045** –0.023* –0.046* –0.029** –0.134*** (in days) (0.017) (0.017) (0.021) (0.013) (0.024) (0.014) (0.049) Control Variables x x x x x x x Firm-CEO FE, Year FE x x x x x x x Observations 7,856 18,460 51,885 33,924 96,285 143,526 28,035 # of turnovers 15 40 83 79 152 224 60 Adj. R$$^{\mathrm{2}}$$ 0.901 0.895 0.751 0.715 0.665 0.805 0.909 (1) (2) (3) (4) (5) (6) (7) 5-year CDS spread, years [0,2] Death/illness Death/illness/ret. at good perf. No mgt. shakeup Good preturnover Perf. No preturnover run-up Union (2) - (5) Forced Tenure –0.033* –0.039** –0.045** –0.023* –0.046* –0.029** –0.134*** (in days) (0.017) (0.017) (0.021) (0.013) (0.024) (0.014) (0.049) Control Variables x x x x x x x Firm-CEO FE, Year FE x x x x x x x Observations 7,856 18,460 51,885 33,924 96,285 143,526 28,035 # of turnovers 15 40 83 79 152 224 60 Adj. R$$^{\mathrm{2}}$$ 0.901 0.895 0.751 0.715 0.665 0.805 0.909 This table reports the changes in CDS spread (daily) over the first three years of CEO tenure, like in Column 2 of Table 2, for various turnover subsamples that are likely to be non-performance-driven turnovers and outright forced turnovers. We control for the same set of CDS, firm or macro-level variables used in Table 2, but do not report the coefficients for brevity. We report the number of turnovers for each subsample, in addition to the number of firm-day observations. Table A1 (see the appendix) reports definitions of turnover types. Standard errors are clustered at the firm-year level. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. Table 3 Tenure-CDS spread relations following likely non-performance-driven versus forced turnovers (1) (2) (3) (4) (5) (6) (7) 5-year CDS spread, years [0,2] Death/illness Death/illness/ret. at good perf. No mgt. shakeup Good preturnover Perf. No preturnover run-up Union (2) - (5) Forced Tenure –0.033* –0.039** –0.045** –0.023* –0.046* –0.029** –0.134*** (in days) (0.017) (0.017) (0.021) (0.013) (0.024) (0.014) (0.049) Control Variables x x x x x x x Firm-CEO FE, Year FE x x x x x x x Observations 7,856 18,460 51,885 33,924 96,285 143,526 28,035 # of turnovers 15 40 83 79 152 224 60 Adj. R$$^{\mathrm{2}}$$ 0.901 0.895 0.751 0.715 0.665 0.805 0.909 (1) (2) (3) (4) (5) (6) (7) 5-year CDS spread, years [0,2] Death/illness Death/illness/ret. at good perf. No mgt. shakeup Good preturnover Perf. No preturnover run-up Union (2) - (5) Forced Tenure –0.033* –0.039** –0.045** –0.023* –0.046* –0.029** –0.134*** (in days) (0.017) (0.017) (0.021) (0.013) (0.024) (0.014) (0.049) Control Variables x x x x x x x Firm-CEO FE, Year FE x x x x x x x Observations 7,856 18,460 51,885 33,924 96,285 143,526 28,035 # of turnovers 15 40 83 79 152 224 60 Adj. R$$^{\mathrm{2}}$$ 0.901 0.895 0.751 0.715 0.665 0.805 0.909 This table reports the changes in CDS spread (daily) over the first three years of CEO tenure, like in Column 2 of Table 2, for various turnover subsamples that are likely to be non-performance-driven turnovers and outright forced turnovers. We control for the same set of CDS, firm or macro-level variables used in Table 2, but do not report the coefficients for brevity. We report the number of turnovers for each subsample, in addition to the number of firm-day observations. Table A1 (see the appendix) reports definitions of turnover types. Standard errors are clustered at the firm-year level. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. The subsample of CEO turnovers due to death or illness of the prior CEO is likely to be closest to random CEO turnovers. This subsample contains only 15 turnovers, corresponding to about 2.6% of the trading days used in the full sample, which reduces the precision of estimates. However, the coefficient for the health/death subsample ($$-$$.033) is comparable to that for the full sample in Column (2) of Table 2, ($$-$$.031) and is still significantly different from zero at the 10% level. That the coefficients from the subsample in which the turnover timing is most likely to be unrelated to firm performance is very close to the coefficient from the full sample suggests that it is unlikely that the negative coefficient on tenure in the full sample is generated by the endogenous timing of CEO turnover. When there is a new CEO, regardless of the reason for the outgoing CEO’s departure, there appears to be uncertainty about the value added from the new CEO. This uncertainty could come from his or her underlying ability or with the quality of the match between the new strategies that he or she will implement and the firm. In contrast, for the outright forced turnover sample (Column 7), the estimated decline in CDS spread is about 0.134 basis point per day, which is about four times as large as it is for the other turnovers, and is statistically significantly different from the coefficient for the full sample. Since outright forced turnovers tend to follow poor firm performance and high volatility, the estimated larger decline could reflect relatively high uncertainty at the time of the turnover about both the firm’s fundamentals and the new CEO. To complement our analysis above based on firm-specific preturnover conditions, we further conduct a placebo test with our sample firms’ industry peers to assess the extent to which the industry performance around CEO turnovers contributes to the spread-tenure relation that we document. Following Jenter and Kanaan (2015), we classify industries using the Fama-French 48 industry classification. For each CEO-firm pair in Table 2, Column (2), we use its Tenure (in days) as a pseudo tenure variable for a firm that: (1) is in the same industry-year, (2) in Execucomp, (3) without CEO turnovers from two years before until two years after, and (4) has CDS data. For about 14% firm-CEO pairs in the sample, there is only one peer firm that satisfies these conditions. When there is more than one qualifying peer firm, we randomly select one. Then we study these peer firms’ change in CDS spreads over their pseudo CEO tenure. We simulate the pseudo sample this way 100 times, and estimate the same specification in Table 2, Column (2), on each pseudo sample. The estimated coefficient on the (pseudo) tenure variable ranges from $$-$$0.016 to $$+$$0.027, with a mean of 0.003, and is not statistically significant from zero in each of the 100 simulations. Thus, the placebo results do not support the alternative hypothesis that the spread-tenure relation is due to CEO turnovers simply coinciding with heightened industry risk that drives up the CDS spreads. 2.3.2 Past performance uncertainty Another alternative interpretation of our results is that CEO turnovers tend to uncover prior management’s errors or even fraud. If the market anticipates a higher likelihood of poor prior performance being revealed, CDS spreads could increase at the time of CEO turnover. As such uncertainty gets resolved over time, the expected probability of error correction or fraud detection should decline, lowering CDS spreads. To evaluate whether the declining spreads we observe are a function of uncertainty about the past rather than the future, we collect data from Audit Analytics on financial restatements. We include restatements due to both accounting errors and intentional fraud. We calculate the likelihood of a restatement in each year relative to CEO turnover and present these probabilities in the Internet Appendix Table IA.1. The results in panel A of this table suggest that the likelihood of a restatement does decline in the first three years of CEO tenure, from 7.41% in the year of the turnover to 3.5% two years later. This pattern is consistent with the idea that part of the uncertainty could reflect the likelihood of bad news about the prior management being revealed. However, this pattern seems to be particularly pronounced following forced CEO turnovers (12.79% in the turnover year compared to 2.82% two years later), but is not observed following the turnovers that we classify as “likely non-performance-related” turnovers. If the revelation of poor prior performance were causing the observed patterns in CDS spreads, we should also observe CDS spreads declining following those “likely non-performance-related” turnovers. Therefore, we believe that the uncertainty about prior management’s mistakes being revealed is unlikely to be a main driver of our results. This conclusion is also confirmed by results in panel B of Table IA.1, which show that controlling for restatements in our baseline regressions does not change the estimated spread-tenure relation at all. 2.4 CEO tenure and interest rates on firms’ debt 2.4.1 Loan spreads over CEO tenure If declining management risk is the reason for the decrease in CDS spreads over tenure, then this declining management risk should also affect the promised interest rates on the firms’ debt. To test this hypothesis, we estimate the way that the spreads on newly-initiated loans vary over CEO tenure, using the specification in Equation (1) with CDS spreads replaced with the “all in drawn” spreads on the loans. When we estimate this equation, we add a number of controls for loan characteristics. In particular, we include the loan size, maturity, number of lenders, and dummy variables indicating whether the loan uses performance pricing, loan purposes, and tranche types. During the first three years of CEOs’ tenure, firms initiate 1.6 loans, on average (median is one loan), per year. Among firms that took loans, about 34% of new CEOs’ firms only took one loan over the first three years of the CEO’s tenure. To avoid having to eliminate these observations, we use firm fixed effects instead of firm-CEO fixed effects in the loan equations and control for firm life cycle effects by including firm age into the equation. Finally, since loan spreads are observable only when a new loan is initiated, we use annual rather than daily data, and measure CEO tenure by the number of years since the CEO takes office. So the resultant coefficients have to be accordingly adjusted to compare magnitudes across specifications. These estimates are reported in panel A of Table 4. Column (1) documents that loan spreads decrease with CEO tenure. The speed at which loan spreads decrease declines over time, with the fastest decrease occurring in the first three years. The estimated coefficients imply that loan spreads decline 6.7 basis points per year, amounting to about 20 basis points over the three-year period. Column (2) reports the estimates over the CEO’s first three years in office for the subsample of CEOs who stay in office for at least three years, which are of a similar magnitude to those reported in Column (1). Column (3) contains estimates using the subsample of likely non-performance-driven turnovers (the union of rows (2)-(5) in panel D of Table 1); these estimates imply that following these turnovers, the loan spread a firm pays declines by 5.5 basis points per year.19 Table 4 Borrowing rates over CEO tenure A. Loan spread Years [0,9] Years [0,2] (1) (2) (3) (4) Likely non-performance-driven turnovers Refinancing Tenure (years 0-2) –6.741*** (1.742) Tenure (years 3-5) –0.524 (1.225) Tenure (years 6-9) –0.562 (1.076) Tenure (in years) –6.842*** –5.483* –8.048*** (2.148) (3.233) (2.700) Credit spread 0.365*** 0.391*** 0.632*** 0.395*** (0.060) (0.118) (0.187) (0.141) Term spread 0.182*** 0.165** 0.315** 0.121 (0.037) (0.071) (0.127) (0.085) log(debt maturity) –4.761 –2.515 6.109 –12.857 (3.502) (5.753) (8.207) (8.243) log(debt size) –12.407*** –12.759*** –4.355 –10.792*** (1.586) (2.500) (3.818) (3.205) Performance pricing –17.359*** –24.389*** –19.852** –30.967*** (2.752) (4.857) (8.249) (6.457) Firm age –3.758 –0.859 5.539 –8.253 (3.341) (5.191) (15.382) (7.301) Tranche type and loan purpose x x x x Firm-level controls x x x x Firm and year F.E. x x x x Observations 11,288 5,273 1,499 3,450 Adj. R$$^{\mathrm{2}}$$ 0.700 0.723 0.785 0.753 B. Bond yield spreads and rating over CEO tenure Years [0,9] Years [0,2] (1) (2) (3) (4) Bond spread Bond spread Bond spread, likely non-performance-driven turnovers Bond rating Tenure –7.884** (years 0-2) (3.883) Tenure –2.743 (years 3-5) (4.121) Tenure –0.741 (years 6-9) (2.721) Tenure (in years) –7.452* –16.373* 0.054* (4.485) (9.888) (0.032) Credit spread 1.221*** 1.367*** 1.442*** 0.013 (0.109) (0.111) (0.205) (0.063) Term spread 0.155* 0.115** 0.202 –0.035*** (0.089) (0.058) (0.139) (0.012) log(debt maturity) –18.414* –17.005 26.143** 0.006 (9.863) (12.664) (12.743) (0.018) log(debt size) –22.328 –28.037 –6.591 –0.008 (14.846) (21.700) (10.013) (0.016) Firm-level controls x x x x Firm F.E. and year F.E. x x x x Observations 4,849 2,114 742 1,909 Adj. R$$^{\mathrm{2}}$$ 0.553 0.599 0.708 0.898 A. Loan spread Years [0,9] Years [0,2] (1) (2) (3) (4) Likely non-performance-driven turnovers Refinancing Tenure (years 0-2) –6.741*** (1.742) Tenure (years 3-5) –0.524 (1.225) Tenure (years 6-9) –0.562 (1.076) Tenure (in years) –6.842*** –5.483* –8.048*** (2.148) (3.233) (2.700) Credit spread 0.365*** 0.391*** 0.632*** 0.395*** (0.060) (0.118) (0.187) (0.141) Term spread 0.182*** 0.165** 0.315** 0.121 (0.037) (0.071) (0.127) (0.085) log(debt maturity) –4.761 –2.515 6.109 –12.857 (3.502) (5.753) (8.207) (8.243) log(debt size) –12.407*** –12.759*** –4.355 –10.792*** (1.586) (2.500) (3.818) (3.205) Performance pricing –17.359*** –24.389*** –19.852** –30.967*** (2.752) (4.857) (8.249) (6.457) Firm age –3.758 –0.859 5.539 –8.253 (3.341) (5.191) (15.382) (7.301) Tranche type and loan purpose x x x x Firm-level controls x x x x Firm and year F.E. x x x x Observations 11,288 5,273 1,499 3,450 Adj. R$$^{\mathrm{2}}$$ 0.700 0.723 0.785 0.753 B. Bond yield spreads and rating over CEO tenure Years [0,9] Years [0,2] (1) (2) (3) (4) Bond spread Bond spread Bond spread, likely non-performance-driven turnovers Bond rating Tenure –7.884** (years 0-2) (3.883) Tenure –2.743 (years 3-5) (4.121) Tenure –0.741 (years 6-9) (2.721) Tenure (in years) –7.452* –16.373* 0.054* (4.485) (9.888) (0.032) Credit spread 1.221*** 1.367*** 1.442*** 0.013 (0.109) (0.111) (0.205) (0.063) Term spread 0.155* 0.115** 0.202 –0.035*** (0.089) (0.058) (0.139) (0.012) log(debt maturity) –18.414* –17.005 26.143** 0.006 (9.863) (12.664) (12.743) (0.018) log(debt size) –22.328 –28.037 –6.591 –0.008 (14.846) (21.700) (10.013) (0.016) Firm-level controls x x x x Firm F.E. and year F.E. x x x x Observations 4,849 2,114 742 1,909 Adj. R$$^{\mathrm{2}}$$ 0.553 0.599 0.708 0.898 Panel A reports the changes in loan spread at origination over CEO tenure. Panel A, Column (1), uses piecewise linear regressions, for all the CEOs from year 0 (turnover year) to year 9. Columns (2) to (4) report the results for the first three years of CEOs’ tenure. Further, Column (3) reports the results for 658 likely non-performance-driven turnovers only (union of the four types of turnovers in Columns (2)-(5) in panel D of Table 1 in the loan sample). Column (4) reports the results for loans that are classified as “refinancing” by DealScan. Panel B reports the changes in bond yield spread and bond rating at issuance over CEO tenure. Panel B, Column (1) uses piecewise linear regressions, for all the CEOs from year 0 (turnover year) to year 9. Columns (2) and (3) report the change in bond yield spreads for the first three years of CEOs’ tenure. In particular, Column (3) reports the results for 280 likely non-performance-driven turnovers only (union of the four types of turnovers in columns (2)-(5) in Panel D of Table 1 in the bond sample). Unlike the other three columns, Column (4) reports the change in bond rating over the first three years of CEO tenure. We control for macro-level credit conditions, bond characteristics, and lagged firm-level characteristics (firm age, firm age, firm size, leverage, M/B, ROA, Tangibility, cash flow volatility, payout ratio). Standard errors are clustered at the firm level. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. Table 4 Borrowing rates over CEO tenure A. Loan spread Years [0,9] Years [0,2] (1) (2) (3) (4) Likely non-performance-driven turnovers Refinancing Tenure (years 0-2) –6.741*** (1.742) Tenure (years 3-5) –0.524 (1.225) Tenure (years 6-9) –0.562 (1.076) Tenure (in years) –6.842*** –5.483* –8.048*** (2.148) (3.233) (2.700) Credit spread 0.365*** 0.391*** 0.632*** 0.395*** (0.060) (0.118) (0.187) (0.141) Term spread 0.182*** 0.165** 0.315** 0.121 (0.037) (0.071) (0.127) (0.085) log(debt maturity) –4.761 –2.515 6.109 –12.857 (3.502) (5.753) (8.207) (8.243) log(debt size) –12.407*** –12.759*** –4.355 –10.792*** (1.586) (2.500) (3.818) (3.205) Performance pricing –17.359*** –24.389*** –19.852** –30.967*** (2.752) (4.857) (8.249) (6.457) Firm age –3.758 –0.859 5.539 –8.253 (3.341) (5.191) (15.382) (7.301) Tranche type and loan purpose x x x x Firm-level controls x x x x Firm and year F.E. x x x x Observations 11,288 5,273 1,499 3,450 Adj. R$$^{\mathrm{2}}$$ 0.700 0.723 0.785 0.753 B. Bond yield spreads and rating over CEO tenure Years [0,9] Years [0,2] (1) (2) (3) (4) Bond spread Bond spread Bond spread, likely non-performance-driven turnovers Bond rating Tenure –7.884** (years 0-2) (3.883) Tenure –2.743 (years 3-5) (4.121) Tenure –0.741 (years 6-9) (2.721) Tenure (in years) –7.452* –16.373* 0.054* (4.485) (9.888) (0.032) Credit spread 1.221*** 1.367*** 1.442*** 0.013 (0.109) (0.111) (0.205) (0.063) Term spread 0.155* 0.115** 0.202 –0.035*** (0.089) (0.058) (0.139) (0.012) log(debt maturity) –18.414* –17.005 26.143** 0.006 (9.863) (12.664) (12.743) (0.018) log(debt size) –22.328 –28.037 –6.591 –0.008 (14.846) (21.700) (10.013) (0.016) Firm-level controls x x x x Firm F.E. and year F.E. x x x x Observations 4,849 2,114 742 1,909 Adj. R$$^{\mathrm{2}}$$ 0.553 0.599 0.708 0.898 A. Loan spread Years [0,9] Years [0,2] (1) (2) (3) (4) Likely non-performance-driven turnovers Refinancing Tenure (years 0-2) –6.741*** (1.742) Tenure (years 3-5) –0.524 (1.225) Tenure (years 6-9) –0.562 (1.076) Tenure (in years) –6.842*** –5.483* –8.048*** (2.148) (3.233) (2.700) Credit spread 0.365*** 0.391*** 0.632*** 0.395*** (0.060) (0.118) (0.187) (0.141) Term spread 0.182*** 0.165** 0.315** 0.121 (0.037) (0.071) (0.127) (0.085) log(debt maturity) –4.761 –2.515 6.109 –12.857 (3.502) (5.753) (8.207) (8.243) log(debt size) –12.407*** –12.759*** –4.355 –10.792*** (1.586) (2.500) (3.818) (3.205) Performance pricing –17.359*** –24.389*** –19.852** –30.967*** (2.752) (4.857) (8.249) (6.457) Firm age –3.758 –0.859 5.539 –8.253 (3.341) (5.191) (15.382) (7.301) Tranche type and loan purpose x x x x Firm-level controls x x x x Firm and year F.E. x x x x Observations 11,288 5,273 1,499 3,450 Adj. R$$^{\mathrm{2}}$$ 0.700 0.723 0.785 0.753 B. Bond yield spreads and rating over CEO tenure Years [0,9] Years [0,2] (1) (2) (3) (4) Bond spread Bond spread Bond spread, likely non-performance-driven turnovers Bond rating Tenure –7.884** (years 0-2) (3.883) Tenure –2.743 (years 3-5) (4.121) Tenure –0.741 (years 6-9) (2.721) Tenure (in years) –7.452* –16.373* 0.054* (4.485) (9.888) (0.032) Credit spread 1.221*** 1.367*** 1.442*** 0.013 (0.109) (0.111) (0.205) (0.063) Term spread 0.155* 0.115** 0.202 –0.035*** (0.089) (0.058) (0.139) (0.012) log(debt maturity) –18.414* –17.005 26.143** 0.006 (9.863) (12.664) (12.743) (0.018) log(debt size) –22.328 –28.037 –6.591 –0.008 (14.846) (21.700) (10.013) (0.016) Firm-level controls x x x x Firm F.E. and year F.E. x x x x Observations 4,849 2,114 742 1,909 Adj. R$$^{\mathrm{2}}$$ 0.553 0.599 0.708 0.898 Panel A reports the changes in loan spread at origination over CEO tenure. Panel A, Column (1), uses piecewise linear regressions, for all the CEOs from year 0 (turnover year) to year 9. Columns (2) to (4) report the results for the first three years of CEOs’ tenure. Further, Column (3) reports the results for 658 likely non-performance-driven turnovers only (union of the four types of turnovers in Columns (2)-(5) in panel D of Table 1 in the loan sample). Column (4) reports the results for loans that are classified as “refinancing” by DealScan. Panel B reports the changes in bond yield spread and bond rating at issuance over CEO tenure. Panel B, Column (1) uses piecewise linear regressions, for all the CEOs from year 0 (turnover year) to year 9. Columns (2) and (3) report the change in bond yield spreads for the first three years of CEOs’ tenure. In particular, Column (3) reports the results for 280 likely non-performance-driven turnovers only (union of the four types of turnovers in columns (2)-(5) in Panel D of Table 1 in the bond sample). Unlike the other three columns, Column (4) reports the change in bond rating over the first three years of CEO tenure. We control for macro-level credit conditions, bond characteristics, and lagged firm-level characteristics (firm age, firm age, firm size, leverage, M/B, ROA, Tangibility, cash flow volatility, payout ratio). Standard errors are clustered at the firm level. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. One difference between the results using CDS spreads and loan spreads is that we observe a firm’s CDS spread every day but observe its loan spreads only when a firm initiates a new loan. Therefore, it is possible that the results could be influenced by the endogenous timing of the loan initiations. One approach to assessing the importance of this potential bias is to consider the set of firms for which the loan represents a refinance, rather than a new capital raising, using DealScan’s classification (the variable “Refinancing Indicator”) to identify refinancing. The equation estimated on refinancing loans only is presented in Column (4) of panel A of Table 4. These estimates are similar to those estimated on the entire sample, and indicate that there is a statistically significant decrease in loan spread of 8.0 basis points on refinancing loans in each of the first three years of CEO tenure.20 Although the pattern in the loan spread is similar to that in the CDS spread around CEO turnovers, the magnitude of the decline in spread is much smaller in the loan sample (21 basis points in three years) than in the CDS sample (34 basis points). The difference in these estimated effects potentially occurs because firms in the CDS sample are, on average, much larger than those in the loan sample, with the CDS sample having average book assets equal to $${\}$$10.1 billion, whereas the loan sample is equal to $${\}$$2.48 billion. Larger firms tend to be more transparent because of more disclosure, more analyst coverage, and more media coverage. This greater transparency likely facilitates market learning about the new management, leading to a faster resolution of uncertainty and a faster decline of spreads over CEO tenure. In fact, in an untabulated test, we find that firms with book assets in the top quartile of the loan sample distribution have an estimated 36-basis points decline in loan spread in the first three years of a CEO’s tenure, close to the decline in CDS spread. 2.4.2 Bond yield spreads over CEO tenure The major alternative corporate debt instrument to a bank loan is a corporate bond. Since management risk should affect the pricing of all corporate liabilities, we also test the hypothesis that the perceived default risk and hence promised yields on issuances of corporate bonds decline with the tenure of firms’ CEOs. To do so, we estimate equations similar to those for bank loans using the promised yield spread on a firm’s corporate bonds as our dependent variable. Panel B of Table 4 presents estimates of this equation, again using CEO tenure measured by the number of years since the CEO takes office. These estimates indicate that CEO tenure has a similar effect on promised bond yields as it does on CDS spreads and loan spreads. Column (1) presents the spline specification using the entire sample. The estimates in this column imply that, as with spreads on CDS and loans, bond yield spreads decline in a convex manner over the CEO’s tenure. Column (2) restricts the sample to the first three years of tenure for CEOs who remain on the job that entire period, and reports a negative effect of tenure on spreads. The coefficients on tenure in these equations are $$-$$7.9 and $$-$$7.5, implying that over the three-year period, yield spreads decline around 23 basis points. Finally, Column (3) further restricts the sample to those CEOs following the “likely non-performance-related” turnovers (the union of rows (2)-(5) in panel D of Table 1) and again finds a negative relation between tenure and spreads. 2.4.3 Credit ratings. Given that spreads appear to decline with CEO tenure, one would expect that firms’ credit ratings would increase with tenure to reflect this lower credit risk. To evaluate this hypothesis, we replace firms’ bond yield spreads with their numerical credit rating (following Jiang et al. 2012; see Table A1 in the appendix for definitions of all variables) in Column (4) of panel B in Table 4. In this specification, the coefficient on tenure is positive and is statistically significant (although only at the 10% level). This result suggests that while a CEO’s tenure lengthens, the firm’s bond ratings improve. This finding is consistent with the notion that management risk declines with tenure, improving their credit ratings and lowering their spreads. 2.5 Uncertainty about management teams: The role of the CFO The analysis to this point has focused on the way that the uncertainty about incoming CEOs’ abilities and policies affects firms’ default risks. The underlying assumption is that the CEO plays an important decision-making role in the firm, so that when the person occupying this position changes, policies can change. However, the CEO is only one member of the management team. Presumably, when top managers other than the CEO change, there is also an increase in uncertainty about future policies, although potentially a smaller one than when the CEO changes. One important member in the senior management team is the Chief Financial Officer (CFO). We examine whether a change of CFO has a similar effect on the perceived default risk as does a change in the CEO. We focus on the CFO rather than other members of the top management team because US firms almost always have one and only one individual with that title, so it is straightforward to identify changes in the individual holding that position. Although there is an extensive literature that discusses the general importance of CEOs, we know little about the relative importance of CFOs, as well as the differences in the skills required for these two jobs. The effect of management risk on corporate debt pricing can be used to quantify the importance of learning about managerial ability for different managerial roles. We collect CFO turnover data from corporate news announcements in the Capital IQ database from 2001 to 2009.21 This process leads to a sample of 1,320 CFO turnovers in 1,079 firms during the 2001–2009 period with CDS or loan or bond spread information, summarized in panel A of Table 5. In this sample, the average CFO spends 4.3 years in office and the median is 3.8 years. In addition, outsider succession is more common in the CFO sample (35%) than in the CEO sample (29%), consistent with data reported by Mian (2001). Table 5 CFO turnovers and default risk A. Summary statistics of CFO turnover and CFO tenure # of turnovers, 2001–2009 CFO turnovers 1,320 CFO turnovers not accompanied by CEO turnovers within $$\pm$$ 6 months 761 CFO turnovers accompanied by CEO turnovers within $$\pm$$ 6 months 559 CFO turnovers due to death/health/retirement at good performance 79 CFO turnovers due to death/health/retirement at good performance, 53 $$\quad$$ and not accompanied by CEO turnovers within $$\pm$$ 6 months Outsider CFO succession 464 A. Summary statistics of CFO turnover and CFO tenure # of turnovers, 2001–2009 CFO turnovers 1,320 CFO turnovers not accompanied by CEO turnovers within $$\pm$$ 6 months 761 CFO turnovers accompanied by CEO turnovers within $$\pm$$ 6 months 559 CFO turnovers due to death/health/retirement at good performance 79 CFO turnovers due to death/health/retirement at good performance, 53 $$\quad$$ and not accompanied by CEO turnovers within $$\pm$$ 6 months Outsider CFO succession 464 CFO time in office Obs. Mean 25th percentile Median 75th percentile CFO total time in office (in years) 1,320 4.25 2 3.83 6 CFO time in office Obs. Mean 25th percentile Median 75th percentile CFO total time in office (in years) 1,320 4.25 2 3.83 6 B. Uncertainty about CFO and CDS spread (1) (2) (3) (4) (5) Years [0, 9] Years [0,2] Accompanied by CEO turnovers within $$\pm$$ 6 months Not accompanied by CEO turnovers within $$\pm$$ 6 months Health/death/retirement at good perf., not accompanied by by CEO turnovers 5-year CDS spread CFO tenure –0.029*** (years 0-2) (0.009) CFO tenure –0.006 (years 3-5) (0.009) CFO tenure –0.002 (years 6-9) (0.011) Tenure –0.031** –0.057** –0.019* –0.022** (in days) (0.014) (0.028) (0.010) (0.010) Controls x x x x x Firm-CFO F.E. and year F.E. x x x x x Obs. 591,048 370,269 109,747 260,522 15,581 Adj. R$$^{\mathrm{2}}$$ 0.814 0.849 0.841 0.860 0.901 B. Uncertainty about CFO and CDS spread (1) (2) (3) (4) (5) Years [0, 9] Years [0,2] Accompanied by CEO turnovers within $$\pm$$ 6 months Not accompanied by CEO turnovers within $$\pm$$ 6 months Health/death/retirement at good perf., not accompanied by by CEO turnovers 5-year CDS spread CFO tenure –0.029*** (years 0-2) (0.009) CFO tenure –0.006 (years 3-5) (0.009) CFO tenure –0.002 (years 6-9) (0.011) Tenure –0.031** –0.057** –0.019* –0.022** (in days) (0.014) (0.028) (0.010) (0.010) Controls x x x x x Firm-CFO F.E. and year F.E. x x x x x Obs. 591,048 370,269 109,747 260,522 15,581 Adj. R$$^{\mathrm{2}}$$ 0.814 0.849 0.841 0.860 0.901 C. Loan spread and bond yield spread during first three years of a CFO’s tenure (1) (2) Loan spread Yield spread Tenure (in years) –5.767* –6.590* (3.099) (3.445) Credit Spread 0.436*** 1.296*** (0.146) (0.268) Term Spread 0.092 0.624** (0.137) (0.260) log(debt maturity) 9.498 –125.159*** (12.138) (38.527) log(debt size) –11.322** –42.916*** (4.435) (10.514) Performance pricing –27.580*** (10.274) Tranche type and loan purpose x x Firm-level controls x x Firm F.E. and year F.E. x x Observations 2,691 1,573 Adj. R$$^{\mathrm{2}}$$ 0.766 0.701 C. Loan spread and bond yield spread during first three years of a CFO’s tenure (1) (2) Loan spread Yield spread Tenure (in years) –5.767* –6.590* (3.099) (3.445) Credit Spread 0.436*** 1.296*** (0.146) (0.268) Term Spread 0.092 0.624** (0.137) (0.260) log(debt maturity) 9.498 –125.159*** (12.138) (38.527) log(debt size) –11.322** –42.916*** (4.435) (10.514) Performance pricing –27.580*** (10.274) Tranche type and loan purpose x x Firm-level controls x x Firm F.E. and year F.E. x x Observations 2,691 1,573 Adj. R$$^{\mathrm{2}}$$ 0.766 0.701 The CFO turnover data are assembled based on the news announcements from 2001 to 2009 in Capital IQ database. Panel A reports the summary statistics of CFO turnover and total time in office (in years) for CFOs whose employers took loans or issued bonds or are in the CDS sample during their first 10 years of tenure, conditional on covariates nonmissing. In the sample with “CFO turnovers not accompanied by CEO turnovers within $$\pm$$ 6 months,” there is no CEO turnover in the 6 months before or in the 6 months after a CFO turnover. “Outsider CFO Succession” means that the new CFO comes from outside the company. Panel B reports changes in a firm’s CDS spread during its CFO’s tenure. The CFO tenure is measured by the number of days since the CFO takes office. Column (1) reports the results using a piecewise linear specification for all CFOs from year 0 to year 9. Columns (2) to (5) report the results for the first three years. Panel C reports changes in a firm’s loan/bond spread during the first three years of its CFO’s tenure. In both panels B and C, we include the usual set of controls used in the CDS, loan, and bond baseline regressions (Tables 2A, 4A, and 4B), but for brevity we do not report the coefficient estimates of some of those variables. Table A1 (see the appendix) reports all variable definitions. Standard errors are clustered at the firm-year level in Panel B, and at the firm level in the loan and bond regressions in panel C. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. Table 5 CFO turnovers and default risk A. Summary statistics of CFO turnover and CFO tenure # of turnovers, 2001–2009 CFO turnovers 1,320 CFO turnovers not accompanied by CEO turnovers within $$\pm$$ 6 months 761 CFO turnovers accompanied by CEO turnovers within $$\pm$$ 6 months 559 CFO turnovers due to death/health/retirement at good performance 79 CFO turnovers due to death/health/retirement at good performance, 53 $$\quad$$ and not accompanied by CEO turnovers within $$\pm$$ 6 months Outsider CFO succession 464 A. Summary statistics of CFO turnover and CFO tenure # of turnovers, 2001–2009 CFO turnovers 1,320 CFO turnovers not accompanied by CEO turnovers within $$\pm$$ 6 months 761 CFO turnovers accompanied by CEO turnovers within $$\pm$$ 6 months 559 CFO turnovers due to death/health/retirement at good performance 79 CFO turnovers due to death/health/retirement at good performance, 53 $$\quad$$ and not accompanied by CEO turnovers within $$\pm$$ 6 months Outsider CFO succession 464 CFO time in office Obs. Mean 25th percentile Median 75th percentile CFO total time in office (in years) 1,320 4.25 2 3.83 6 CFO time in office Obs. Mean 25th percentile Median 75th percentile CFO total time in office (in years) 1,320 4.25 2 3.83 6 B. Uncertainty about CFO and CDS spread (1) (2) (3) (4) (5) Years [0, 9] Years [0,2] Accompanied by CEO turnovers within $$\pm$$ 6 months Not accompanied by CEO turnovers within $$\pm$$ 6 months Health/death/retirement at good perf., not accompanied by by CEO turnovers 5-year CDS spread CFO tenure –0.029*** (years 0-2) (0.009) CFO tenure –0.006 (years 3-5) (0.009) CFO tenure –0.002 (years 6-9) (0.011) Tenure –0.031** –0.057** –0.019* –0.022** (in days) (0.014) (0.028) (0.010) (0.010) Controls x x x x x Firm-CFO F.E. and year F.E. x x x x x Obs. 591,048 370,269 109,747 260,522 15,581 Adj. R$$^{\mathrm{2}}$$ 0.814 0.849 0.841 0.860 0.901 B. Uncertainty about CFO and CDS spread (1) (2) (3) (4) (5) Years [0, 9] Years [0,2] Accompanied by CEO turnovers within $$\pm$$ 6 months Not accompanied by CEO turnovers within $$\pm$$ 6 months Health/death/retirement at good perf., not accompanied by by CEO turnovers 5-year CDS spread CFO tenure –0.029*** (years 0-2) (0.009) CFO tenure –0.006 (years 3-5) (0.009) CFO tenure –0.002 (years 6-9) (0.011) Tenure –0.031** –0.057** –0.019* –0.022** (in days) (0.014) (0.028) (0.010) (0.010) Controls x x x x x Firm-CFO F.E. and year F.E. x x x x x Obs. 591,048 370,269 109,747 260,522 15,581 Adj. R$$^{\mathrm{2}}$$ 0.814 0.849 0.841 0.860 0.901 C. Loan spread and bond yield spread during first three years of a CFO’s tenure (1) (2) Loan spread Yield spread Tenure (in years) –5.767* –6.590* (3.099) (3.445) Credit Spread 0.436*** 1.296*** (0.146) (0.268) Term Spread 0.092 0.624** (0.137) (0.260) log(debt maturity) 9.498 –125.159*** (12.138) (38.527) log(debt size) –11.322** –42.916*** (4.435) (10.514) Performance pricing –27.580*** (10.274) Tranche type and loan purpose x x Firm-level controls x x Firm F.E. and year F.E. x x Observations 2,691 1,573 Adj. R$$^{\mathrm{2}}$$ 0.766 0.701 C. Loan spread and bond yield spread during first three years of a CFO’s tenure (1) (2) Loan spread Yield spread Tenure (in years) –5.767* –6.590* (3.099) (3.445) Credit Spread 0.436*** 1.296*** (0.146) (0.268) Term Spread 0.092 0.624** (0.137) (0.260) log(debt maturity) 9.498 –125.159*** (12.138) (38.527) log(debt size) –11.322** –42.916*** (4.435) (10.514) Performance pricing –27.580*** (10.274) Tranche type and loan purpose x x Firm-level controls x x Firm F.E. and year F.E. x x Observations 2,691 1,573 Adj. R$$^{\mathrm{2}}$$ 0.766 0.701 The CFO turnover data are assembled based on the news announcements from 2001 to 2009 in Capital IQ database. Panel A reports the summary statistics of CFO turnover and total time in office (in years) for CFOs whose employers took loans or issued bonds or are in the CDS sample during their first 10 years of tenure, conditional on covariates nonmissing. In the sample with “CFO turnovers not accompanied by CEO turnovers within $$\pm$$ 6 months,” there is no CEO turnover in the 6 months before or in the 6 months after a CFO turnover. “Outsider CFO Succession” means that the new CFO comes from outside the company. Panel B reports changes in a firm’s CDS spread during its CFO’s tenure. The CFO tenure is measured by the number of days since the CFO takes office. Column (1) reports the results using a piecewise linear specification for all CFOs from year 0 to year 9. Columns (2) to (5) report the results for the first three years. Panel C reports changes in a firm’s loan/bond spread during the first three years of its CFO’s tenure. In both panels B and C, we include the usual set of controls used in the CDS, loan, and bond baseline regressions (Tables 2A, 4A, and 4B), but for brevity we do not report the coefficient estimates of some of those variables. Table A1 (see the appendix) reports all variable definitions. Standard errors are clustered at the firm-year level in Panel B, and at the firm level in the loan and bond regressions in panel C. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. In panel B of Table 5, we estimate the relation between the firm’s CDS spread and the CFO’s time in office, using the same specification used in Table 2. Column (1) contains estimates implying that the CDS spread declines by a statistically significant 0.029 basis points per day during the first three years of a new CFO’s tenure. During the subsequent three years, the decline is just 0.006 basis points per day, which is not statistically significantly different from zero. Column (2) restricts the sample to the first 3 years of a CFO’s tenure, and finds a decline in CDS spreads of 0.031 basis points per day. An issue in interpreting these results is that many CFO turnovers coincide with CEO turnovers; declines in spreads following these cases likely reflect uncertainty about both managers, and possibly others as well if there is a large change in the top management team. For this reason, we re-estimate this equation on the subsample of CFO turnovers accompanied by a CEO change within six months before or after the time of the CFO change (Column (3), 220 turnovers), as well as on the subsample for which there was not a CEO change within this one-year period (Column (4), 520 turnovers). In each subsample, the firm’s CDS spread significantly declines with CFO’s tenure, but the magnitude of the estimated coefficient on tenure is much smaller on the subsample of CFO changes without CEO changes ($$-$$0.019 in Column (4)) than for the subsample in which there is both a CEO change and a CFO change ($$-$$0.057 in Column (3)). The difference between the two estimates is statistically significant at the 5% level. For further comparison, in the subsample for which there is a CEO change without a CFO or other top management change, the estimated coefficient on tenure is $$-$$0.045, which is between these two estimates (see Table 3, Column (3)), and is significantly more negative than the coefficient on tenure for the subsample of CFO changes not accompanied by CEO changes ($$-$$0.019). Since the CEO is the most important officer, uncertainty about his ability affects firms the most, but uncertainty about the CFO matters as well. Not surprisingly, when both officers change at the same time, the decline in spreads with tenure is the largest, most likely because these cases are associated with the greatest managerial uncertainty at turnover.22 As with CEO turnovers, another possible explanation for the spread/tenure relation is that both the CFO turnover and the higher default risk could be driven by factors unrelated to management risk, which is more likely when a turnover is due to poor firm performance. For this reason, we consider the subsample of 31 CFO turnovers that are not accompanied by CEO turnovers and follow the death, illness, or retirement of the departing CFOs when the firm is performing well. The estimates using this subsample are presented in Column (5) of Table 5. The results indicate that in this subsample, we still observe a significant CDS spread decline over the CFO’s first three years of tenure. The magnitude of the decline (0.022 basis points per day) is close to the estimate with all CFOs turnovers not accompanied by CEO turnovers (0.019 basis points per day). Consequently, it does not appear that the observed decline in default risk over the first 3 years of a CFO’s tenure occurs because of the management changes occurring at times of high uncertainty unrelated to management. Instead, the results suggest that uncertainty about his ability or future actions generates incremental default risk. In panel C of Table 5, we examine the relation between the firm’s loan and bond yield spreads and the CFO’s tenure. We find that the loan spread tends to decline by 5.8 basis points per year and the bond yield spread tends to decline by 6.6 basis points per year in the CFO’s first three years in office. These declines are statistically significant and similar in magnitude to those reported in Table 4 following CEO turnovers. Overall, the results in Tables 2–5 suggest that there is a substantial, statistically significant decrease in the default risk of a firm’s debt over the CEO’s and the CFO’s tenures, reflected by the firm’s CDS spread, the spreads on its bank loans, and the yield on its corporate bonds. The decrease is fastest in the executives’ first three years in office. This decline does not appear to come from executive turnovers occurring in periods when non-management-related uncertainty is high. Like it would for other sources of investor uncertainty about the firm, the market raises its estimate of the firm’s default risk when management’s ability or policies are unknown. 3. Cross-Sectional Differences in the Change in Default Risk Cross-sectional variation in the sensitivity of spreads to executive turnover and tenure provides a way to confirm that the decline in spreads over CEO tenure does in fact reflect the resolution of managerial uncertainty. In particular, if the increase in spreads following management changes reflects uncertainty about the ability and policies of the new management, then when this uncertainty is higher, there should be a larger increase in spreads around the time of the turnover. In addition, there should be a larger subsequent decline, while investor uncertainty resolves. 3.1 CEO background and prior uncertainty about the CEO Different types of CEO successions and different types of CEOs are likely to be associated with different amounts of uncertainty about the new management. For example, the existence of an “heir apparent” usually indicates a well-anticipated succession, a smoother transition of management control, and an incoming CEO of known ability and a continuation of the prior CEO’s policies. In contrast, appointments of outsider CEOs are likely to lead to more uncertainty about the leadership transition process and the quality of the match with the new firm or future policies. Consequently, appointment of outsider CEOs are likely to be associated with more uncertainty about future cash flows. In addition, the market will tend to know less about younger managers who are appointed to be CEOs, since they tend to have shorter job histories and less visibility than older incoming CEOs. For this reason, the market is likely to have a more diffuse prior about younger incoming CEOs than older ones. 3.1.1 Evidence from prior to the arrival of the new CEO Although we have mostly focused on the declines in risk subsequent to the appointment of a new CEO, Figure 1 suggests that a number of components contribute to the change in CDS spreads around a CEO turnover: CDS spreads increase at the announcement of a CEO’s departure, decline to some extent by the time the new CEO takes office, and further decline after the CEO takes office. If the changes in CDS spreads are caused by uncertainty about the new management, then the magnitude of each of these changes should depend on the amount of uncertainty there is about the incoming management. We test this hypothesis and present the results in Table 6. Table 6 Cross-sectional differences in the spread-tenure relation A. The rise in CDS spread at the departure announcement (1) CDS(Departure)-CDS(Predeparture) Firm has heir –29.753*** (10.404) Death/illness 15.494* (8.115) Forced 3.803 (17.264) % change in price –1.608* (0.819) Year F.E. x Observations 427 Adj. R$$^{\mathrm{2}}$$ 0.080 A. The rise in CDS spread at the departure announcement (1) CDS(Departure)-CDS(Predeparture) Firm has heir –29.753*** (10.404) Death/illness 15.494* (8.115) Forced 3.803 (17.264) % change in price –1.608* (0.819) Year F.E. x Observations 427 Adj. R$$^{\mathrm{2}}$$ 0.080 B. The decrease in CDS spread from departure to inauguration CDS(Departure)-CDS(Predeparture) CDS(Inauguration)-CDS(Departure) % of the initial increase Obs Full sample 44.95 –20.852 46.39 284 Outsider new CEO 61.588 –16.957 27.53 73 Insider new CEO 37.023 –22.312 60.27 209 Diff. btw. out- and insiders 24.565*** 5.355** Young CEO (< 50) 53.098 –18.1189 34.12 70 Old CEO (> $$=$$50) 41.697 –21.6145 51.84 210 Diff. btw. young and old 11.401** 3.496** B. The decrease in CDS spread from departure to inauguration CDS(Departure)-CDS(Predeparture) CDS(Inauguration)-CDS(Departure) % of the initial increase Obs Full sample 44.95 –20.852 46.39 284 Outsider new CEO 61.588 –16.957 27.53 73 Insider new CEO 37.023 –22.312 60.27 209 Diff. btw. out- and insiders 24.565*** 5.355** Young CEO (< 50) 53.098 –18.1189 34.12 70 Old CEO (> $$=$$50) 41.697 –21.6145 51.84 210 Diff. btw. young and old 11.401** 3.496** C. Differences in prior uncertainty about management and post-turnover decline in CDS spread (1) (2) (3) (4) 5-year CDS spread Tenure (in days) –0.006 –0.018 –0.024 –0.034*** (0.019) (0.015) (0.015) (0.010) Tenure*Non-heir-apparent CEO –0.035** (0.016) Tenure*Outsider CEO –0.077*** (0.022) Tenure*Young CEO –0.026* (0.015) Tenure*Outsider CFO –0.004 (0.016) Stock price –0.938*** –0.946*** –0.963*** –0.967*** (0.243) (0.243) (0.242) (0.231) Recovery rate –10.767*** –9.550*** –9.596*** –11.431*** (1.427) (1.429) (1.441) (1.496) Credit spread 0.415*** 0.404*** 0.403*** 0.447*** (0.067) (0.065) (0.065) (0.057) Term spread –0.043 –0.042 –0.043 –0.092** (0.043) (0.043) (0.043) (0.039) VIX 2.093*** 2.223*** 2.220*** 2.129*** (0.260) (0.251) (0.248) (0.213) Log(assets) 53.731* 67.754** 76.044** –10.371 (31.747) (33.086) (32.324) (22.573) Leverage 95.384 69.743 86.207 121.877* (60.555) (58.707) (59.383) (62.730) M/B 24.399** 6.767 14.401 –0.500 (9.830) (13.123) (12.790) (0.979) ROA –478.799*** –470.723*** –487.562*** –462.426*** (120.744) (118.082) (116.266) (110.726) Tangibility 138.923 41.373 51.120 115.704 (142.312) (151.427) (147.104) (96.746) CF volatility 6.091 8.166 7.819 34.937*** (7.182) (7.190) (7.150) (7.356) Payout ratio –6.995 –8.583* –7.657 –12.464** (5.030) (4.955) (4.927) (5.659) Firm-CEO (CFO) F.E. and Year F.E. x x x x Observations 278,688 296,951 300,715 370,269 Adjusted R$$^2$$ 0.841 0.840 0.840 0.849 C. Differences in prior uncertainty about management and post-turnover decline in CDS spread (1) (2) (3) (4) 5-year CDS spread Tenure (in days) –0.006 –0.018 –0.024 –0.034*** (0.019) (0.015) (0.015) (0.010) Tenure*Non-heir-apparent CEO –0.035** (0.016) Tenure*Outsider CEO –0.077*** (0.022) Tenure*Young CEO –0.026* (0.015) Tenure*Outsider CFO –0.004 (0.016) Stock price –0.938*** –0.946*** –0.963*** –0.967*** (0.243) (0.243) (0.242) (0.231) Recovery rate –10.767*** –9.550*** –9.596*** –11.431*** (1.427) (1.429) (1.441) (1.496) Credit spread 0.415*** 0.404*** 0.403*** 0.447*** (0.067) (0.065) (0.065) (0.057) Term spread –0.043 –0.042 –0.043 –0.092** (0.043) (0.043) (0.043) (0.039) VIX 2.093*** 2.223*** 2.220*** 2.129*** (0.260) (0.251) (0.248) (0.213) Log(assets) 53.731* 67.754** 76.044** –10.371 (31.747) (33.086) (32.324) (22.573) Leverage 95.384 69.743 86.207 121.877* (60.555) (58.707) (59.383) (62.730) M/B 24.399** 6.767 14.401 –0.500 (9.830) (13.123) (12.790) (0.979) ROA –478.799*** –470.723*** –487.562*** –462.426*** (120.744) (118.082) (116.266) (110.726) Tangibility 138.923 41.373 51.120 115.704 (142.312) (151.427) (147.104) (96.746) CF volatility 6.091 8.166 7.819 34.937*** (7.182) (7.190) (7.150) (7.356) Payout ratio –6.995 –8.583* –7.657 –12.464** (5.030) (4.955) (4.927) (5.659) Firm-CEO (CFO) F.E. and Year F.E. x x x x Observations 278,688 296,951 300,715 370,269 Adjusted R$$^2$$ 0.841 0.840 0.840 0.849 D. Good preturnover performance (1) (2) (3) 5-year CDS spread Tenure (in days) –0.014 –0.011 –0.011 (0.018) (0.013) (0.014) Tenure*Non-heir-apparent CEO –0.028** (0.014) Tenure*Outsider CEO –0.033** (0.015) Tenure*Young CEO –0.026* (0.013) Stock price –0.102 –0.145 –0.122 (0.082) (0.094) (0.081) Recovery rate –0.919 –1.532* –0.785 (1.451) (0.903) (1.286) Credit spread 0.148*** 0.111 0.148*** (0.042) (0.087) (0.044) Term spread –0.019 –0.011 –0.005 (0.044) (0.043) (0.039) VIX 0.412*** 0.997** 0.361*** (0.149) (0.488) (0.130) Log(assets) 12.965 57.645 –25.101 (28.515) (36.762) (36.466) Leverage –9.061 115.377 22.526 (53.467) (79.285) (52.113) M/B 13.757 19.393* 5.282 (8.542) (10.787) (7.905) ROA –193.550*** –103.046 –182.149*** (50.774) (77.716) (56.384) Tangibility 391.664*** 277.055** 280.421** (145.236) (118.918) (130.739) CF volatility 1.611 3.733** 1.421 (1.805) (1.623) (1.669) Payout ratio 16.755*** 3.844 15.041*** (4.727) (9.938) (4.069) Firm-CEO F.E. and Year F.E. x x x Observations 31,120 32,642 32,642 Adjusted R$$^2$$ 0.725 0.717 0.720 D. Good preturnover performance (1) (2) (3) 5-year CDS spread Tenure (in days) –0.014 –0.011 –0.011 (0.018) (0.013) (0.014) Tenure*Non-heir-apparent CEO –0.028** (0.014) Tenure*Outsider CEO –0.033** (0.015) Tenure*Young CEO –0.026* (0.013) Stock price –0.102 –0.145 –0.122 (0.082) (0.094) (0.081) Recovery rate –0.919 –1.532* –0.785 (1.451) (0.903) (1.286) Credit spread 0.148*** 0.111 0.148*** (0.042) (0.087) (0.044) Term spread –0.019 –0.011 –0.005 (0.044) (0.043) (0.039) VIX 0.412*** 0.997** 0.361*** (0.149) (0.488) (0.130) Log(assets) 12.965 57.645 –25.101 (28.515) (36.762) (36.466) Leverage –9.061 115.377 22.526 (53.467) (79.285) (52.113) M/B 13.757 19.393* 5.282 (8.542) (10.787) (7.905) ROA –193.550*** –103.046 –182.149*** (50.774) (77.716) (56.384) Tangibility 391.664*** 277.055** 280.421** (145.236) (118.918) (130.739) CF volatility 1.611 3.733** 1.421 (1.805) (1.623) (1.669) Payout ratio 16.755*** 3.844 15.041*** (4.727) (9.938) (4.069) Firm-CEO F.E. and Year F.E. x x x Observations 31,120 32,642 32,642 Adjusted R$$^2$$ 0.725 0.717 0.720 Panel A reports the effects of firms having a designated heir apparent (“Firm has heir” equals to one), turnovers due to the departing CEO’s death/illness, and outright forced turnovers on the increase in CDS spread between predeparture (the average level from month -3 to month -1) and the departure announcement date. We control for the percentage change in stock price over the same period, as well as the year fixed effects. Panel B reports the average drop in CDS spread between the previous CEOs’ departure announcement days and new CEOs’ inauguration days, for the whole sample, firms with outsider versus insider new CEOs, and firms with young versus old incoming CEOs. “Young CEO” is a CEO who is less than 50 years old when taking office. We also report the percentage of this drop as a fraction of the initial spread increase at departure, the number of observations for each sample, as well as the differences between various pairs of subsamples. Panel C reports the effect of prior uncertainty about the new CEO (or new CFO) on changes in CDS spread during the first three years of the CEO’s (or CFO’s) tenure. “Non-heir-apparent CEO” indicates that the CEO was not an heir apparent before becoming the CEO. “Outsider CEO (CFO)” indicates that the CEO (CFO) comes from outside the company. “Young CEO” is a CEO who is less than 50 years old when taking office. Panel D reports the effect of prior uncertainty about the new CEO on changes in CDS spread during the first three years of the CEO’s tenure, for the sample of CEO turnovers following good firm performance only (i.e., Column (4) in Table 3, panel A). Table A1 (see the appendix) reports all variable definitions. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. Table 6 Cross-sectional differences in the spread-tenure relation A. The rise in CDS spread at the departure announcement (1) CDS(Departure)-CDS(Predeparture) Firm has heir –29.753*** (10.404) Death/illness 15.494* (8.115) Forced 3.803 (17.264) % change in price –1.608* (0.819) Year F.E. x Observations 427 Adj. R$$^{\mathrm{2}}$$ 0.080 A. The rise in CDS spread at the departure announcement (1) CDS(Departure)-CDS(Predeparture) Firm has heir –29.753*** (10.404) Death/illness 15.494* (8.115) Forced 3.803 (17.264) % change in price –1.608* (0.819) Year F.E. x Observations 427 Adj. R$$^{\mathrm{2}}$$ 0.080 B. The decrease in CDS spread from departure to inauguration CDS(Departure)-CDS(Predeparture) CDS(Inauguration)-CDS(Departure) % of the initial increase Obs Full sample 44.95 –20.852 46.39 284 Outsider new CEO 61.588 –16.957 27.53 73 Insider new CEO 37.023 –22.312 60.27 209 Diff. btw. out- and insiders 24.565*** 5.355** Young CEO (< 50) 53.098 –18.1189 34.12 70 Old CEO (> $$=$$50) 41.697 –21.6145 51.84 210 Diff. btw. young and old 11.401** 3.496** B. The decrease in CDS spread from departure to inauguration CDS(Departure)-CDS(Predeparture) CDS(Inauguration)-CDS(Departure) % of the initial increase Obs Full sample 44.95 –20.852 46.39 284 Outsider new CEO 61.588 –16.957 27.53 73 Insider new CEO 37.023 –22.312 60.27 209 Diff. btw. out- and insiders 24.565*** 5.355** Young CEO (< 50) 53.098 –18.1189 34.12 70 Old CEO (> $$=$$50) 41.697 –21.6145 51.84 210 Diff. btw. young and old 11.401** 3.496** C. Differences in prior uncertainty about management and post-turnover decline in CDS spread (1) (2) (3) (4) 5-year CDS spread Tenure (in days) –0.006 –0.018 –0.024 –0.034*** (0.019) (0.015) (0.015) (0.010) Tenure*Non-heir-apparent CEO –0.035** (0.016) Tenure*Outsider CEO –0.077*** (0.022) Tenure*Young CEO –0.026* (0.015) Tenure*Outsider CFO –0.004 (0.016) Stock price –0.938*** –0.946*** –0.963*** –0.967*** (0.243) (0.243) (0.242) (0.231) Recovery rate –10.767*** –9.550*** –9.596*** –11.431*** (1.427) (1.429) (1.441) (1.496) Credit spread 0.415*** 0.404*** 0.403*** 0.447*** (0.067) (0.065) (0.065) (0.057) Term spread –0.043 –0.042 –0.043 –0.092** (0.043) (0.043) (0.043) (0.039) VIX 2.093*** 2.223*** 2.220*** 2.129*** (0.260) (0.251) (0.248) (0.213) Log(assets) 53.731* 67.754** 76.044** –10.371 (31.747) (33.086) (32.324) (22.573) Leverage 95.384 69.743 86.207 121.877* (60.555) (58.707) (59.383) (62.730) M/B 24.399** 6.767 14.401 –0.500 (9.830) (13.123) (12.790) (0.979) ROA –478.799*** –470.723*** –487.562*** –462.426*** (120.744) (118.082) (116.266) (110.726) Tangibility 138.923 41.373 51.120 115.704 (142.312) (151.427) (147.104) (96.746) CF volatility 6.091 8.166 7.819 34.937*** (7.182) (7.190) (7.150) (7.356) Payout ratio –6.995 –8.583* –7.657 –12.464** (5.030) (4.955) (4.927) (5.659) Firm-CEO (CFO) F.E. and Year F.E. x x x x Observations 278,688 296,951 300,715 370,269 Adjusted R$$^2$$ 0.841 0.840 0.840 0.849 C. Differences in prior uncertainty about management and post-turnover decline in CDS spread (1) (2) (3) (4) 5-year CDS spread Tenure (in days) –0.006 –0.018 –0.024 –0.034*** (0.019) (0.015) (0.015) (0.010) Tenure*Non-heir-apparent CEO –0.035** (0.016) Tenure*Outsider CEO –0.077*** (0.022) Tenure*Young CEO –0.026* (0.015) Tenure*Outsider CFO –0.004 (0.016) Stock price –0.938*** –0.946*** –0.963*** –0.967*** (0.243) (0.243) (0.242) (0.231) Recovery rate –10.767*** –9.550*** –9.596*** –11.431*** (1.427) (1.429) (1.441) (1.496) Credit spread 0.415*** 0.404*** 0.403*** 0.447*** (0.067) (0.065) (0.065) (0.057) Term spread –0.043 –0.042 –0.043 –0.092** (0.043) (0.043) (0.043) (0.039) VIX 2.093*** 2.223*** 2.220*** 2.129*** (0.260) (0.251) (0.248) (0.213) Log(assets) 53.731* 67.754** 76.044** –10.371 (31.747) (33.086) (32.324) (22.573) Leverage 95.384 69.743 86.207 121.877* (60.555) (58.707) (59.383) (62.730) M/B 24.399** 6.767 14.401 –0.500 (9.830) (13.123) (12.790) (0.979) ROA –478.799*** –470.723*** –487.562*** –462.426*** (120.744) (118.082) (116.266) (110.726) Tangibility 138.923 41.373 51.120 115.704 (142.312) (151.427) (147.104) (96.746) CF volatility 6.091 8.166 7.819 34.937*** (7.182) (7.190) (7.150) (7.356) Payout ratio –6.995 –8.583* –7.657 –12.464** (5.030) (4.955) (4.927) (5.659) Firm-CEO (CFO) F.E. and Year F.E. x x x x Observations 278,688 296,951 300,715 370,269 Adjusted R$$^2$$ 0.841 0.840 0.840 0.849 D. Good preturnover performance (1) (2) (3) 5-year CDS spread Tenure (in days) –0.014 –0.011 –0.011 (0.018) (0.013) (0.014) Tenure*Non-heir-apparent CEO –0.028** (0.014) Tenure*Outsider CEO –0.033** (0.015) Tenure*Young CEO –0.026* (0.013) Stock price –0.102 –0.145 –0.122 (0.082) (0.094) (0.081) Recovery rate –0.919 –1.532* –0.785 (1.451) (0.903) (1.286) Credit spread 0.148*** 0.111 0.148*** (0.042) (0.087) (0.044) Term spread –0.019 –0.011 –0.005 (0.044) (0.043) (0.039) VIX 0.412*** 0.997** 0.361*** (0.149) (0.488) (0.130) Log(assets) 12.965 57.645 –25.101 (28.515) (36.762) (36.466) Leverage –9.061 115.377 22.526 (53.467) (79.285) (52.113) M/B 13.757 19.393* 5.282 (8.542) (10.787) (7.905) ROA –193.550*** –103.046 –182.149*** (50.774) (77.716) (56.384) Tangibility 391.664*** 277.055** 280.421** (145.236) (118.918) (130.739) CF volatility 1.611 3.733** 1.421 (1.805) (1.623) (1.669) Payout ratio 16.755*** 3.844 15.041*** (4.727) (9.938) (4.069) Firm-CEO F.E. and Year F.E. x x x Observations 31,120 32,642 32,642 Adjusted R$$^2$$ 0.725 0.717 0.720 D. Good preturnover performance (1) (2) (3) 5-year CDS spread Tenure (in days) –0.014 –0.011 –0.011 (0.018) (0.013) (0.014) Tenure*Non-heir-apparent CEO –0.028** (0.014) Tenure*Outsider CEO –0.033** (0.015) Tenure*Young CEO –0.026* (0.013) Stock price –0.102 –0.145 –0.122 (0.082) (0.094) (0.081) Recovery rate –0.919 –1.532* –0.785 (1.451) (0.903) (1.286) Credit spread 0.148*** 0.111 0.148*** (0.042) (0.087) (0.044) Term spread –0.019 –0.011 –0.005 (0.044) (0.043) (0.039) VIX 0.412*** 0.997** 0.361*** (0.149) (0.488) (0.130) Log(assets) 12.965 57.645 –25.101 (28.515) (36.762) (36.466) Leverage –9.061 115.377 22.526 (53.467) (79.285) (52.113) M/B 13.757 19.393* 5.282 (8.542) (10.787) (7.905) ROA –193.550*** –103.046 –182.149*** (50.774) (77.716) (56.384) Tangibility 391.664*** 277.055** 280.421** (145.236) (118.918) (130.739) CF volatility 1.611 3.733** 1.421 (1.805) (1.623) (1.669) Payout ratio 16.755*** 3.844 15.041*** (4.727) (9.938) (4.069) Firm-CEO F.E. and Year F.E. x x x Observations 31,120 32,642 32,642 Adjusted R$$^2$$ 0.725 0.717 0.720 Panel A reports the effects of firms having a designated heir apparent (“Firm has heir” equals to one), turnovers due to the departing CEO’s death/illness, and outright forced turnovers on the increase in CDS spread between predeparture (the average level from month -3 to month -1) and the departure announcement date. We control for the percentage change in stock price over the same period, as well as the year fixed effects. Panel B reports the average drop in CDS spread between the previous CEOs’ departure announcement days and new CEOs’ inauguration days, for the whole sample, firms with outsider versus insider new CEOs, and firms with young versus old incoming CEOs. “Young CEO” is a CEO who is less than 50 years old when taking office. We also report the percentage of this drop as a fraction of the initial spread increase at departure, the number of observations for each sample, as well as the differences between various pairs of subsamples. Panel C reports the effect of prior uncertainty about the new CEO (or new CFO) on changes in CDS spread during the first three years of the CEO’s (or CFO’s) tenure. “Non-heir-apparent CEO” indicates that the CEO was not an heir apparent before becoming the CEO. “Outsider CEO (CFO)” indicates that the CEO (CFO) comes from outside the company. “Young CEO” is a CEO who is less than 50 years old when taking office. Panel D reports the effect of prior uncertainty about the new CEO on changes in CDS spread during the first three years of the CEO’s tenure, for the sample of CEO turnovers following good firm performance only (i.e., Column (4) in Table 3, panel A). Table A1 (see the appendix) reports all variable definitions. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. In panel A of Table 6, we examine how the increase in CDS spread at the time of the CEO departure announcement depends on the prior uncertainty about the new management, using the 427 CEO turnovers in our sample with CDS spread, stock price data and identifiable CEO departure announcement dates. For these firms, the CDS spread is an average of about 40 basis points higher at the departure announcement relative to the average in the three months prior to the announcement month. However, there is significant cross-sectional heterogeneity in the magnitude of the increase. Turnovers for which the firm has a designated heir apparent presumably have lower uncertainty about the incoming management. Consistent with this idea, we find that the estimated increase in CDS spreads at CEO departure announcement for these turnovers is 30 basis points lower than in turnovers with no designated successors, after controlling for the contemporaneous stock price changes. CEO departures occurring because of the death or illness of the CEO often occur unexpectedly. Therefore, although these turnovers are unlikely to occur at times of heightened uncertainty about the firms’ fundamentals, they are likely to be associated with high uncertainty about the incoming management. In the cases of CEOs departing because of death or illness, the CDS spread increases by an average of 15 basis points more than in a non-death/illness-related turnover. In contrast, forced turnovers tend to occur at times of high uncertainty about the firm’s fundamentals. However, the average increase in CDS spread is only 3.8 basis points higher than turnovers we classify as nonforced at the time of the departure announcement, and the difference is not statistically significant. One potential reason for the lack of a difference for forced turnovers is that in these cases CDS spreads are likely to have already been relatively high in the months prior to the announcement of the forced CEO turnover. Panel B of Table 6 presents statistics on the change in CDS spread between the departure of the outgoing CEO and the arrival of the incoming CEO. Since part of the uncertainty about the new management (e.g., the new CEO’s identity and possibly his managing style and tentative plan) could get resolved during this period of time, we expect the firm’s CDS spread to decrease. In the 284 CEO turnovers for which the announcement of the outgoing CEO’s departure and the arrival of the new CEO occur on different dates, the average CDS spread declines by 21 basis points between the departure of the outgoing CEO and the arrival of the new CEO, which is 46% of the initial rise in spread (45 basis points) at the CEO departure announcement. If the initial increase in spread reflects the total additional uncertainty brought on by the imminent arrival of new management, then the estimates suggest that close to half of such uncertainty gets revolved by the time the new CEO takes office. Cross-sectionally, the amount of uncertainty about the new management that can be resolved by the time the new management arrives in office should depend on the characteristics of the incoming management. For example, when the CEO is hired from outside, the uncertainty about the quality of match between the firm and the new CEO could remain high even after the new CEO’s identity and his tentative plan is revealed, leading to a smaller fraction of uncertainty being resolved at the time of his arrival. In contrast, when the firm hires an insider, more uncertainty about his fit and future strategies could be resolved by the time he or she takes office. In our sample, when the replacement is an outsider, the CDS spread declines by an average of 17 basis points between the outgoing CEO’s departure and the replacement’s arrival, which equals 28% of the initial spread increase. In contrast, when the incoming CEO is an insider, the spread declines by 22 basis points, or 60% of the initial rise in this subsample. Similarly, we find that for younger new CEOs (younger than 50 when taking office), only 34% of the initial spread rise is reversed by the arrival time of the new CEO, whereas for older new CEOs, the fraction is 52%. These differences in the magnitudes of the intermediate CDS spread changes across turnovers suggest that the pattern of increases and subsequent decreases in CDS spreads around the time of CEO turnover is likely because of the market’s uncertainty about the leadership transition process and the CEO’s fit with the firm and his future actions. 3.1.2 Evidence from after CEO turnover The results in panels A and B of Table 6 suggest that uncertainty about a new CEO’s ability when he or she takes office is lower for heir-apparent CEOs, and higher for outsider CEOs and younger CEOs. Learning models, such as that of Pastor and Veronesi (2003), predict that learning should be faster when prior uncertainty about ability is higher (see Hermalin and Weisbach 2017 for applications of this idea to governance). A consequence of a faster learning speed is that after turnover, perceived default risk (and thus CDS spreads) should decline at a higher rate when there is more prior uncertainty about the CEO’s ability. In panel C of Table 6, we examine whether the sensitivity of CDS spreads to tenure increases with the uncertainty about the incoming CEO at the time he or she takes office. Column (1) includes a term interacting tenure with a dummy variable indicating whether the new CEO was not an “heir apparent.” The results suggest that the spread-tenure sensitivity is largely concentrated in non-heir-apparent CEOs. For heir-apparent CEOs, the spread-tenure slope is negative but close to zero, indicating little ex ante uncertainty about them when they take office. In Column (2), we compare the spread-tenure sensitivity between insider new CEOs and outsider new CEOs. The effect of tenure on CDS spread for outsider CEOs ($$-$$0.077) is substantially and significantly larger than that for insider CEOs. Similarly, in Column (3) we find that the CDS spreads are significantly more sensitive to the tenure of young CEOs (under 50 years old) than older CEOs. All these results suggest that the firm’s CDS spread is more sensitive to CEO tenure when there is higher prior uncertainty about the CEO. A potentially important distinction between CEO and CFO is that a CFO’s skills, such as experience with financial reporting, tax, and making accounting judgments, are typically more general and transferrable across firms than a CEO’s skills.23 The generality of CFOs’ skills could be one reason why we observe more outsider successions for CFOs than for CEOs. Comparing the market learning processes for insider and outsider CFOs can also shed light on the degree of generality of managerial skills required by the CFO job. In Column (4) of Table 6, panel C, we find that the CDS spread to tenure sensitivity is not significantly different between insider and outsider CFOs, suggesting that the prior uncertainty about a new CFO’s ability is similar regardless of his succession origin. This finding is different from what we report for insider and outsider CEOs, consistent with the idea that CFO skills are more general than CEO skills. When comparing the effects of different types of incoming CEOs, it is possible that the endogenous nature in the firm-CEO match could raise concerns about our interpretation of the results based on managerial risk. On the one hand, new managers are likely to implement, and in fact be chosen because they would implement, different strategies from the previous managers. These new strategies create uncertainty, leading to management risk. On the other hand, the preturnover firm condition could influence both the choice of the new CEO and the post-turnover changes in CDS spreads. This is the endogenous turnover timing issue that we address in Section 3.3. For example, outsider new CEOs could be associated with larger changes in strategies, creating more managerial risk. But at the same time, outsider CEO succession tend to follow poor firm performance and potentially heightened CDS spreads. Thus, to make the point that outsider CEOs create more managerial risk, we need to address the turnover timing issue. We do so by integrating the analysis in Section 3.3 into our analysis here. Specifically, we re-estimate panel C of Table 6 only for turnovers with good preturnover firm performance as defined in Column (4) of Table 3. Panel D reports the results. We find that outsider new CEOs are still related to a significantly steeper decline of CDS spreads over tenure than insider new CEOs in this subsample (Column (2)), suggesting that the outsider CEO effect in panel C is not completely driven by the endogenous timing of outsider CEO succession. However, the magnitude of the outsider CEO effect is significantly smaller than that in panel C ($$-$$0.033 versus $$-$$0.077), suggesting that the endogenous timing leads to an overestimation of the outsider CEO effect in panel C. Interestingly, the magnitudes of the heir-apparent effect (Column (1)) and the young CEO effect (Column (3)) on the spread-tenure relation do not change significantly between panels C and D, suggesting that the choice between an heir-apparent CEO and a non-heir-apparent one or between a younger CEO and an older one is less subject to the endogenous turnover timing issue. In summary, the cross-sectional evidence from both the CEO turnover process and the post-turnover time in office suggest that a firm’s default risk and CDS spread react to changes in the amount of uncertainty about the new management. Even though we cannot completely remove the endogeneity in the timing, as well as the choice of new CEOs, we believe that it is reasonable to conclude from our tests that alternative interpretations, such as endogenous CEO turnover timing or changes in firms’ fundamentals, do not appear to explain these cross-sectional findings. 3.1.3 Prior relationships with lenders We have presumed to this point that all suppliers of debt capital have access to the same information about the firm’s management, so that all have the same assessment of the firm’s executives’ abilities or policies at each point in time. However, in the loan market for example, it is possible that some lenders do know better about the manager’s ability than other lenders, because of the prior lending relationship. This informational advantage underlies the literature on relationship banking, which suggests that a long-term relationship between firms and lenders reduces asymmetric information and consequently the spreads that firms pay on loans.24 If part of the asymmetric information that contributes to the spread differences between relationship and non-relationship-based loans is about the management of the borrowing firm, then the existence of a personal relationship between the manager and the lender should reduce this information asymmetry. Consistent with this idea, Karolyi (Forthcoming) finds that firms are more likely to choose the lenders that have a personal relationship with their new executives after executive turnovers. To the extent that such a personal relationship lowers the amount of prior uncertainty about the new management from the perspective of a lender, we consider the possibility that it reduces the sensitivity of loan spreads to the manager’s time in office. To test this prediction, we rely on DealScan data, together with information on executive job changes from Execucomp, from which we can measure whether a CEO worked for a firm that previously took a loan from a particular lender. We construct an indicator variable, “Prior CEO-lender relationship,” that equals one if at least one of the lead bank(s) of the current new loan was a lead bank in a loan of the CEO’s employer in the five years before he or she became the CEO of the current firm, and zero otherwise. We expect such a prior relationship to reduce the lender’s initial uncertainty about the new CEO’s ability, leading to lower sensitivity of spreads to the new CEO’s tenure. The interpretation of this variable depends on whether a CEO was an internal or external hire; for internal hires, the prior relationship would exist whenever the current firm had taken a loan with the lender, whereas for an external hire, it would exist if his prior firm had taken the loan. Panel A of Table 1 reports summary statistics of this variable. We estimate the ways in which the effect of tenure on loan spreads varies with previous lending relationships in Column (1) of Table 7, focusing on the first three years of a CEO’s tenure. The estimated direct effect of CEO tenure on loan spread is $$-$$8.928 and is statistically significant, and the interaction effect between CEO tenure and prior CEO-lender relationship is 6.530 and also is statistically significant. These estimates imply that when there is no prior relationship, the spread declines by about 9 basis points per year of CEO tenure. However, the existence of a prior relationship between the new CEO and the lead bank(s) reduces the spread-tenure sensitivity by about 73% ($$=$$6.530/8.928). Table 7 Prior lending relationship and spread-tenure relations (1) (2) (3) (4) Full CEO sample Outsider CEOs Full CFO sample Outsider CEOs Loan spread Tenure (in years) –8.928*** –18.957*** –14.680* –16.531* (2.708) (6.775) (7.822) (8.445) Tenure (in years)*Prior CEO 6.530* 12.960* 8.914* 6.555* $$\quad$$ (CFO)-lender relationship (3.418) (6.693) (5.123) (3.371) Prior CEO (CFO)-lender relationship –8.768 –18.882 –7.604 –10.101 (7.162) (16.308) (11.552) (23.907) log(debt maturity) –2.504 3.243 7.588 9.707 (5.747) (10.871) (12.131) (17.540) log(debt size) –12.359*** –14.796*** –11.815*** –11.031* (2.500) (5.484) (4.572) (6.258) Performance pricing –23.955*** –26.658*** –24.458** –38.710** (4.933) (10.296) (10.841) (15.609) Tranche type and loan purpose x x x x Firm-level controls x x x x Firm F.E. and year F.E. x x x x Observations 5,266 1,284 2,652 1,057 Adjusted R$$^2$$ 0.724 0.752 0.754 0.750 (1) (2) (3) (4) Full CEO sample Outsider CEOs Full CFO sample Outsider CEOs Loan spread Tenure (in years) –8.928*** –18.957*** –14.680* –16.531* (2.708) (6.775) (7.822) (8.445) Tenure (in years)*Prior CEO 6.530* 12.960* 8.914* 6.555* $$\quad$$ (CFO)-lender relationship (3.418) (6.693) (5.123) (3.371) Prior CEO (CFO)-lender relationship –8.768 –18.882 –7.604 –10.101 (7.162) (16.308) (11.552) (23.907) log(debt maturity) –2.504 3.243 7.588 9.707 (5.747) (10.871) (12.131) (17.540) log(debt size) –12.359*** –14.796*** –11.815*** –11.031* (2.500) (5.484) (4.572) (6.258) Performance pricing –23.955*** –26.658*** –24.458** –38.710** (4.933) (10.296) (10.841) (15.609) Tranche type and loan purpose x x x x Firm-level controls x x x x Firm F.E. and year F.E. x x x x Observations 5,266 1,284 2,652 1,057 Adjusted R$$^2$$ 0.724 0.752 0.754 0.750 This table reports the effect of relationship lending on changes in loan spreads during the first three year of CEO’s or CFO’s tenure. “Prior CEO (or CFO)-lender relationship” is a dummy variable that equals one if the lead bank(s) of the current new loan was a lead bank in a loan of the CEO’s employer in the five years before he or she became the CEO of the current firm, and zero otherwise. We include the firm-year-level controls used in Table 4, panel A, but for brevity we do not report the coefficient estimates of some of those variables. Table A1 (see the appendix) reports all variable definitions. Standard errors are clustered at the firm level. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. Table 7 Prior lending relationship and spread-tenure relations (1) (2) (3) (4) Full CEO sample Outsider CEOs Full CFO sample Outsider CEOs Loan spread Tenure (in years) –8.928*** –18.957*** –14.680* –16.531* (2.708) (6.775) (7.822) (8.445) Tenure (in years)*Prior CEO 6.530* 12.960* 8.914* 6.555* $$\quad$$ (CFO)-lender relationship (3.418) (6.693) (5.123) (3.371) Prior CEO (CFO)-lender relationship –8.768 –18.882 –7.604 –10.101 (7.162) (16.308) (11.552) (23.907) log(debt maturity) –2.504 3.243 7.588 9.707 (5.747) (10.871) (12.131) (17.540) log(debt size) –12.359*** –14.796*** –11.815*** –11.031* (2.500) (5.484) (4.572) (6.258) Performance pricing –23.955*** –26.658*** –24.458** –38.710** (4.933) (10.296) (10.841) (15.609) Tranche type and loan purpose x x x x Firm-level controls x x x x Firm F.E. and year F.E. x x x x Observations 5,266 1,284 2,652 1,057 Adjusted R$$^2$$ 0.724 0.752 0.754 0.750 (1) (2) (3) (4) Full CEO sample Outsider CEOs Full CFO sample Outsider CEOs Loan spread Tenure (in years) –8.928*** –18.957*** –14.680* –16.531* (2.708) (6.775) (7.822) (8.445) Tenure (in years)*Prior CEO 6.530* 12.960* 8.914* 6.555* $$\quad$$ (CFO)-lender relationship (3.418) (6.693) (5.123) (3.371) Prior CEO (CFO)-lender relationship –8.768 –18.882 –7.604 –10.101 (7.162) (16.308) (11.552) (23.907) log(debt maturity) –2.504 3.243 7.588 9.707 (5.747) (10.871) (12.131) (17.540) log(debt size) –12.359*** –14.796*** –11.815*** –11.031* (2.500) (5.484) (4.572) (6.258) Performance pricing –23.955*** –26.658*** –24.458** –38.710** (4.933) (10.296) (10.841) (15.609) Tranche type and loan purpose x x x x Firm-level controls x x x x Firm F.E. and year F.E. x x x x Observations 5,266 1,284 2,652 1,057 Adjusted R$$^2$$ 0.724 0.752 0.754 0.750 This table reports the effect of relationship lending on changes in loan spreads during the first three year of CEO’s or CFO’s tenure. “Prior CEO (or CFO)-lender relationship” is a dummy variable that equals one if the lead bank(s) of the current new loan was a lead bank in a loan of the CEO’s employer in the five years before he or she became the CEO of the current firm, and zero otherwise. We include the firm-year-level controls used in Table 4, panel A, but for brevity we do not report the coefficient estimates of some of those variables. Table A1 (see the appendix) reports all variable definitions. Standard errors are clustered at the firm level. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. The estimates using our full sample pool CEOs who were internal hires together with those who were external hires. For each case, the prior lending relationship likely resolves some uncertainty perceived by the lenders. However, when the new CEO is an internal hire, we cannot distinguish CEO-lender relationships from firm-lender relationships, whereas when the new CEO is an external hire, we can. To isolate the extent to which the personal relationship with the lender leads to lower uncertainty perceived by the lender about the management, we re-estimate this equation on the subsample of CEOs who were hired from outside the firm. Column (2) reports the estimates for the subsample of outsider CEOs. The direct effect of tenure is much larger in absolute magnitude here than in Column (1) ($$-$$18.957 vs. $$-$$8.928), probably because there is more uncertainty about outsider CEOs. The interaction effect is 12.960, which implies that if the incoming outsider CEO has worked with the lender before joining the current company, then the loan spread is 68% ($$=$$12.960/18.957) less sensitive to the new CEO’s time in office. Thus, a personal relationship between a CEO and a lender, even if it occurs prior to the CEO joining his current firm, leads to less uncertainty perceived by a lender, and consequently lower sensitivity of spreads to CEO tenure. In Columns (3) and (4), we repeat this exercise for CFOs. Similar to the estimates for CEOs, the existence of a prior relationship between the CFO and the lender(s) significantly reduces the sensitivity of the firm’s loan spreads to the CFO’s time in office by 61% for the full sample of CFOs ($$=$$8.914/14.680) and 40% for outsider CFOs ($$=$$6.555/16.531). Overall, the results in Table 6 and 7 suggest that the CDS spread and loan spread to tenure relations are affected by the amount of prior uncertainty about the new CEO. This pattern is consistent with the argument that the declining interest rate over CEO tenure is driven by the decrease in the amount of uncertainty about the new management over time. 3.3 Risk of the debt claim Another cross-sectional prediction is that the effect of management uncertainty should be larger when the debt is more risky. When the firm is closer to default, the incremental effect of any additional risk on default probabilities is higher. For this reason, default risk should be more sensitive to CEO tenure for speculative grade issuers than for investment grade issuers, for highly levered issuers than for moderately levered issuers, and for subordinated debt than for senior debt. We evaluate these predictions in Table 8. In Column (1), we re-estimate the CDS equation from Table 2, Column (2), but also include interaction terms between tenure and a dummy variable indicating whether the firm has a speculative credit rating (below BBB-) at the time of the turnover, and between tenure and a dummy variable that equals one if the firm’s leverage ratio is in the top quartile of the sample distribution (above 36%) at the time of turnover.25 The coefficient on the interaction with “Speculative Grade” is small and insignificant, but the coefficient on the interaction with high leverage is negative, large in magnitude, and significantly different from zero. The sensitivity of CDS spread to CEO tenure in highly leveraged firms is more than twice the level in moderately leveraged firms. Table 8 Risk of the debt claim and the spread-tenure relation (1) (2) (3) CDS spread Loan spread Yield spread Tenure (in days) –0.023* (0.013) Tenure (in days)*Speculative –0.012 (0.021) Tenure (in days)*Highly levered –0.041* (0.023) Tenure (in years) –0.460 –4.580* (2.489) (2.430) Tenure (in years)*Speculative –5.438 –9.328 (5.819) (7.454) Tenure (in years)*Highly levered –6.832* –6.090* (3.939) (3.138) Tenure (in years)*Term loan –11.053* (5.858) Tenure (in years)*Subordinated –8.864* (4.548) Speculative 110.975*** 79.415** (12.732) (34.190) Highly levered 21.590** 54.055*** (8.588) (19.065) Term loan 52.108** (20.545) Subordinated 16.976 (23.567) Firm-level controls x x x Loan- or bond-level controls x x Firm-CEO F.E. and Year F.E. x Firm F.E. and year F.E. x x Observations 261,757 3,673 1,962 Adjusted R$$^2$$ 0.844 0.775 0.605 (1) (2) (3) CDS spread Loan spread Yield spread Tenure (in days) –0.023* (0.013) Tenure (in days)*Speculative –0.012 (0.021) Tenure (in days)*Highly levered –0.041* (0.023) Tenure (in years) –0.460 –4.580* (2.489) (2.430) Tenure (in years)*Speculative –5.438 –9.328 (5.819) (7.454) Tenure (in years)*Highly levered –6.832* –6.090* (3.939) (3.138) Tenure (in years)*Term loan –11.053* (5.858) Tenure (in years)*Subordinated –8.864* (4.548) Speculative 110.975*** 79.415** (12.732) (34.190) Highly levered 21.590** 54.055*** (8.588) (19.065) Term loan 52.108** (20.545) Subordinated 16.976 (23.567) Firm-level controls x x x Loan- or bond-level controls x x Firm-CEO F.E. and Year F.E. x Firm F.E. and year F.E. x x Observations 261,757 3,673 1,962 Adjusted R$$^2$$ 0.844 0.775 0.605 This table contrasts the changes in the CDS spread, loan spread or bond yield spread over the first three years of the CEO’s tenure for riskier debt claims and less risky debt claims. “Speculative grade” indicates a firm with credit rating below BBB- in the turnover year. “Highly levered” indicates borrowers with leverage ratio in the top quartile (0.36) in the turnover year. “Subordinated” indicates junior bonds. “Term loan” indicates (all types of) term loans. We include the usual set of firm, loan, or bond level controls, but for brevity we do not report the coefficient estimates of those variables. Table A1 (see the appendix) reports all variable definitions. Standard errors are clustered at the firm-year level in Column (1), and at the firm level in Column (2) and Column (3). *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. Table 8 Risk of the debt claim and the spread-tenure relation (1) (2) (3) CDS spread Loan spread Yield spread Tenure (in days) –0.023* (0.013) Tenure (in days)*Speculative –0.012 (0.021) Tenure (in days)*Highly levered –0.041* (0.023) Tenure (in years) –0.460 –4.580* (2.489) (2.430) Tenure (in years)*Speculative –5.438 –9.328 (5.819) (7.454) Tenure (in years)*Highly levered –6.832* –6.090* (3.939) (3.138) Tenure (in years)*Term loan –11.053* (5.858) Tenure (in years)*Subordinated –8.864* (4.548) Speculative 110.975*** 79.415** (12.732) (34.190) Highly levered 21.590** 54.055*** (8.588) (19.065) Term loan 52.108** (20.545) Subordinated 16.976 (23.567) Firm-level controls x x x Loan- or bond-level controls x x Firm-CEO F.E. and Year F.E. x Firm F.E. and year F.E. x x Observations 261,757 3,673 1,962 Adjusted R$$^2$$ 0.844 0.775 0.605 (1) (2) (3) CDS spread Loan spread Yield spread Tenure (in days) –0.023* (0.013) Tenure (in days)*Speculative –0.012 (0.021) Tenure (in days)*Highly levered –0.041* (0.023) Tenure (in years) –0.460 –4.580* (2.489) (2.430) Tenure (in years)*Speculative –5.438 –9.328 (5.819) (7.454) Tenure (in years)*Highly levered –6.832* –6.090* (3.939) (3.138) Tenure (in years)*Term loan –11.053* (5.858) Tenure (in years)*Subordinated –8.864* (4.548) Speculative 110.975*** 79.415** (12.732) (34.190) Highly levered 21.590** 54.055*** (8.588) (19.065) Term loan 52.108** (20.545) Subordinated 16.976 (23.567) Firm-level controls x x x Loan- or bond-level controls x x Firm-CEO F.E. and Year F.E. x Firm F.E. and year F.E. x x Observations 261,757 3,673 1,962 Adjusted R$$^2$$ 0.844 0.775 0.605 This table contrasts the changes in the CDS spread, loan spread or bond yield spread over the first three years of the CEO’s tenure for riskier debt claims and less risky debt claims. “Speculative grade” indicates a firm with credit rating below BBB- in the turnover year. “Highly levered” indicates borrowers with leverage ratio in the top quartile (0.36) in the turnover year. “Subordinated” indicates junior bonds. “Term loan” indicates (all types of) term loans. We include the usual set of firm, loan, or bond level controls, but for brevity we do not report the coefficient estimates of those variables. Table A1 (see the appendix) reports all variable definitions. Standard errors are clustered at the firm-year level in Column (1), and at the firm level in Column (2) and Column (3). *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. In Columns (2) and (3), we estimate a similar equation for the loan and bond samples. For the loan equation, we also compare term loans and lines of credit, since term loans are more risky for banks than are lines of credit. Similarly, for the bond equation, we compare subordinated bonds and senior bonds.26, 27 The estimates presented in Column (2) document that loan spreads are significantly more sensitive to CEO tenure for highly leveraged firms and for term loans. Similarly, Column (3) shows that the bond yield spreads are significantly more sensitive to CEO tenure when firms are highly leveraged and bonds are more junior. Overall, tenure-spread sensitivities appear to be higher when debt is more risky.28 In summary, the cross-sectional evidence about the relation between the spreads on the firm’s debt and its CEO’s time in office is consistent with the view that the observed decline in spreads over tenure is driven by a reduction in uncertainty about management. The default risk implicit in the pricing of the firm’s debt appear to be more sensitive to CEO tenure when there is higher prior uncertainty about the CEO’s ability, when there is no prior relationship between the new CEO and the lender, and when the debt claim is more risky. 4. Conclusion A central feature of financial markets is that the promised yield on a firm’s debt increases with the market’s perception of the firm’s risk. This risk occurs because of factors that affect the value of the firm’s underlying assets and also because of uncertainty about how these assets will be managed. Consequently, a rational market should incorporate managerial-generated uncertainty into its assessment of a firm’s risk when pricing its securities. Holding constant a firm’s fundamental risks, when there is more uncertainty about a new management team’s abilities or its future choices of actions relative to those of the departing management, creditors should increase the interest rates they charge the firm. Practitioners and capital markets clearly understand the importance of management risk, but the topic has not yet found its way into finance textbooks or the academic literature. We study the ways in which investor uncertainty about management affects firms’ default risk and, consequently, the pricing of their debt. Our results suggest that risk should not be viewed monolithically: A firm’s risk comes from many sources, including both the fundamentals of its business and its choice of management team. Uncertainty about management is likely to be highest when a management team is new and is likely to decrease over time, while the new management’s talent, strategies, and the outcomes of these strategies, become better known to the market. Our empirical analysis suggests that CDS spreads on a firm’s debt, loan spreads at origination, and the bond yield spreads at issuance are all significantly higher when the firm’s CEO and/or CFO are new in office than when management team has been in office for three years. This pattern persists regardless of whether the turnover occurred for non-performance-related reasons. The sensitivity of promised rates to tenure increases with the prior uncertainty about the incoming manager, where higher prior uncertainty is characterized by the incoming manager not being an heir apparent, being an outsider, being younger, or not having a prior relationship with the firm’s lenders. These results are consistent with the view that management risk is an important component of default risk. The impact of management risk on corporate debt has a number of implications. First, management risk should affect the way that academics and practitioners model credit risk. Such models could be meaningfully improved by explicitly incorporating managerial characteristics likely to be associated with the market’s uncertainty about the policies the management will adopt. Second, management risk is likely to have implications for the way in which firms’ financial management policies change over a CEO’s tenure, including changes in firms’ cash management and investment financing. Third, because existing creditors are hurt at times of CEO succession, it provides an explanation for why creditors would sometimes seek additional control rights in the event of a management change, for example, in the form of “change of management restriction” (CMR) clauses. Fourth, and finally, that the increase in credit risk at times of turnover appears to be related to the uncertainty about the incoming management’s value-added suggests that policies that reduce such uncertainty can help to lower a firm’s credit risk. Future research that identifies specific sources of management risk could shed important light both on how credit risk models could be improved and on corporate policy implications for the existence of management risk. Acknowledgements We would like to thank Shan Ge, Tyler Jensen, Abby Kim, Dongxu Li, Xingzhou Li, Keeseon Nam, Xi Wu, and Julian Zhang for excellent research assistance. We thank participants in presentations at Arizona, Beijing University, CKGSB, Fullerton, George Washington, London Business School, University of Minnesota, Ohio State University, University of Oregon, Southern Methodist University, University of Southern California, University of Texas-Dallas, University of Utah, Villanova University, 2015 Western Finance Association Meeting, 2015 Annual Conference on Financial Economics and Accounting, and 2015 FMA Asia Annual Meeting. In addition, Benjamin Bennett, Jeff Coles, Michael Cooper, Naveen Daniel, David Denis, Harry DeAngelo, Isil Erel, Steve Karolyi, Sigitas Karpavicius, John Matsusaka, Oded Palmon, Miriam Schwartz-Ziv, Berk Sensoy, Henri Servaes, Anil Shivdasani, Léa Stern, Luke Taylor, Jun Yang, Xiaoyun Yu, Lu Zhang, and two anonymous referees provided very helpful suggestions. Yihui Pan, Tracy Wang, and Michael Weisbach are Fellows of the Risk Institute at the Fisher College of Business, Ohio State University, and acknowledge the Center’s support for this research. Supplementary data can be found on The Review of Financial Studies web site. Appendix Table A1 Variable definitions Variable Definition Loan spread (in basis points) All-in-drawn spread (AIS) over LIBOR at the origination date, from the current pricing file. Winsorized at 1% in the DealScan/Compustat-merged database Loan maturity (in months) A calculation of how long (in months) the facility will be active from signing date to expiration date, from the facility file Loan size (in $${\}$$ millions) The amount of the facility, from the facility data set Secured An indicator variable that equals one if the loan is secured, from the facility file Number of lenders Total number of lenders in a loan, from the lender file Number of loan covenants The total number of covenants in six categories (therefore this variable ranges from zero to six): equity sweep, debt sweep, asset sweep, financial, dividend, and secured, following Bradley and Roberts (2015) Performance pricing A loan feature that ties the interest rate of the loan to an indicator (e.g., leverage, interest coverage ratio) of the firm’s performance, from the performance pricing file Loan type Type of the loan (facility): term loan, revolver, etc. Loan purpose Purpose of the loan (facility): takeover, working capital, debt repayment, etc. Lead bank Following Bharath et al. (2007), we focus on lead bank(s) in the syndicate in relationship lending. Any lender characterized as “lead arranger,” “lead bank,” or “lead manager,” or who has an allocation of more than 90% of the total committed amount to the facility is characterized as a lead bank. Any bank that is described as a “participant” is not a leading bank Prior CEO (CFO)-lender relationship A binary variable that equals one if the lead bank(s) of the current new loan was a lead bank in a loan of the CEO’s employer in the five years before he or she became the CEO of the current firm, and zero otherwise Dummy (loan initiation or bond issuance) An indicator variable that equals one if the firm takes at least one loan or issued one bond in the fiscal year Yield spread (in basis point) Offering yield spread. The difference between the offering yield at issuance and the yield of the benchmark treasury bond, calculated only for fixed coupon bond (about 78% of the Mergent sample) Bond size (in $${\}$$ millions) Offering amount, the par value of debt at issuance (in $${\}$$ millions) Bond maturity (in months) Maturity date – offering date (in months) Subordinated An indicator variable that equals one if the bond is junior, junior subordinate, subordinate, senior subordinate, and zero otherwise (senior or senior secured) Bond rating Following Jiang et al. (2012), we use the following numerical bond rating: 7 corresponds to AAA; 6 corresponds to AA- to AA$$+$$; 5 corresponds to A- to A$$+$$; 4 corresponds to BBB- to BBB$$+$$; 3 corresponds to BB- to BB$$+$$; 2 corresponds to B$$-$$ to B$$+$$; and 1 corresponds to C-level ratings CDS spread (in basis points) The amount paid by the Protection Buyer to the Protection Seller, typically denominated in basis points, with an annualized quote but paid quarterly. We use the five-year spreads because these contracts are the most liquid and constitute over 85% of the entire CDS market. To maintain uniformity in contracts, we only keep CDS quotations for senior unsecured debt with a modified restructuring (MR) clause and denominated in U.S. dollars CDS(Departure)-CDS(Pre-departure) The rise in CDS spreads between the average CDS spreads from three months before the departure announcement (month -3 to month -1; with month 0 being the departure month) and the departure announcement day CDS(Inauguration)-CDS(Departure) The drop in CDS spreads between the dates of the previous CEO’s departure announcement and when the new CEO takes office Recovery Rate (in percentage) Reported by data contributors. Most pricing methodologies estimate recovery rates in a very simplistic way: a percentage is assigned to the seniority of the debt of a company. For investment grade issuers, recovery is generally assumed to be 40% (as the probability of default is low, the recovery rate is at best an estimate). For distressed issuers however, where the probability of default is higher, recovery tends to be more precisely defined Credit Spread (in basis point) The difference between AAA corporate bond yield and BAA corporate bond yield (data source: Federal Reserve Board of Governors) measured in the month prior to loan initiation Term Spread (in basis point) The difference between the ten-year Treasury yield and the two-year Treasury yield (source: Federal Reserve Board of Governors) measured in the month prior to loan initiation VIX (in percentage) CBOE volatility index, which shows the market’s expectation of 30-day volatility. It is constructed using the implied volatilities of a wide range of S&P 500 index options. This volatility is meant to be forward looking and is calculated from both calls and puts Firm age Age of the firm since IPO, using the first day appear in CRSP (or the IPO date in Compustat if missing), constructed for each firm-year Log(assets) Logarithm of the total book assets (assets are measured in $${\}$$ millions) Leverage (Long-term debt $$+$$ debt in current liabilities)/total assets $$M/B$$ Market value of equity (closing price at the fiscal year end times shares outstanding) divided by book value of equity $$Q$$ (Market value of equity $$+$$ the book value of total debt)/book value of total assets ROA Earnings before interest, tax, and depreciation scaled by the total book assets Tangibility Net property, plant, and equipment/total assets CF volatility Residual volatility of the AR(1) process of ROE, following Pastor and Veronesi (2003) Payout ratio (Dividend/earnings) per share Speculative grade An indicator variable that equals one if the firm has a rating below BBB-, and zero otherwise (investment grade) Highly levered An indicator variable that equals one if leverage is greater than 36% (corresponds to the 75% of the leverage distribution, as well as the mean of the speculative grade firms) Total time in office Equals zero if the CEO (or CFO) came into and exited the office in the same year; one if he or she exited the year after becoming CEO (or CFO), etc. Management ability CEO’s ability, relative to their industry peers, in transforming corporate resources to revenues (see Demerjian, Lev, and McVay 2012 for details). We use the second stage residual of regressing the raw ability scores on firm-level characteristics Outsider CEO An indicator that equals one if the CEO is hired from outside (i.e., never worked for the firm prior to becoming its CEO) Heir apparent CEO An indicator variable that equals one if the new CEO was an heir apparent. An executive with the title “president” or “chief operating officer (COO)” or both, who is distinct from the CEO and the chairman is designated as the “heir apparent” Firm has heir An indicator variable that equals one if the firm has an heir apparent that year, and zero otherwise Young CEO An indicator variable that equals one if the CEO who was younger than 50 when taking office Turnovers due to health or illness Include cases in which (1) news searches revealed that the CEO departure was related to a health condition or death (from Fee, Hadlock, and Pierce 2013) or (2) turnover reason provided in Execucomp is “deceased” Turnovers due to retirement of departing CEO This sample includes turnovers where (1) news searches revealed that the CEO departure was related to a health condition or death (from Fee, Hadlock, and Pierce 2013), (2) turnover reason provided in Execucomp is “deceased,” and (3) departing CEOs older than 65 years. We exclude the “suspicious” retirements by focusing on retirements at good performance. This means that the cumulative monthly stock return during the 12 months before the new CEO’s inauguration month (see the variable definition for Cum. return month [-12,-1] below) is greater than sample median (0.3%) No management shakeup CEO turnovers not accompanied by management (top-four highest paid non-CEO executives) changes during the turnover year and the year after turnover Cum. return month [-12,-1] Cumulative stock return during the 12 months before the inauguration month Median monthly IVOL month[-12,-1] The median of the monthly idiosyncratic volatility during the 12 months before the inauguration month Good preturnover performance Turnovers that satisfy the following three conditions: (1) the median of the monthly idiosyncratic volatility during the 12 months before the inauguration month (see the variable definition for Median IVOL month [-12,-1] above) is less or equal to 6.62% (sample median in Table 2, panel B, Column (1)); (2) the cumulative monthly stock return during the 12 months before the inauguration month (see the variable definition for Cum. return month [-12,-1] above) is no less than 0.3% (sample median in Table 2, panel B, Column (1)); and (3) the ROA in the fiscal year prior to the inauguration month is no less than 12.7% (sample median in Table 2, panel B, Column (1)). ROA is defined as the earnings before interest, tax, and depreciation scaled by the beginning of fiscal year total book assets No preturnover run-up in CDS spread To capture the change in the CDS spread before turnover, we run firm-CEO specific regressions of daily CDS spread on event days [-730, -30], with day 0 being the day when the CEO takes office. We require at least 250 trading day data on CDS spread. Turnovers with nonpositive (or insignificant) tenure-time slope are classified as not preceded by an increase in the CDS spread Outright forced Outright forced turnovers include the “overtly forced” group from Fee, Hadlock, and Pierce (2013) with cases for which news searches indicated that the CEO was forced to leave or exited under pressure Variable Definition Loan spread (in basis points) All-in-drawn spread (AIS) over LIBOR at the origination date, from the current pricing file. Winsorized at 1% in the DealScan/Compustat-merged database Loan maturity (in months) A calculation of how long (in months) the facility will be active from signing date to expiration date, from the facility file Loan size (in $${\}$$ millions) The amount of the facility, from the facility data set Secured An indicator variable that equals one if the loan is secured, from the facility file Number of lenders Total number of lenders in a loan, from the lender file Number of loan covenants The total number of covenants in six categories (therefore this variable ranges from zero to six): equity sweep, debt sweep, asset sweep, financial, dividend, and secured, following Bradley and Roberts (2015) Performance pricing A loan feature that ties the interest rate of the loan to an indicator (e.g., leverage, interest coverage ratio) of the firm’s performance, from the performance pricing file Loan type Type of the loan (facility): term loan, revolver, etc. Loan purpose Purpose of the loan (facility): takeover, working capital, debt repayment, etc. Lead bank Following Bharath et al. (2007), we focus on lead bank(s) in the syndicate in relationship lending. Any lender characterized as “lead arranger,” “lead bank,” or “lead manager,” or who has an allocation of more than 90% of the total committed amount to the facility is characterized as a lead bank. Any bank that is described as a “participant” is not a leading bank Prior CEO (CFO)-lender relationship A binary variable that equals one if the lead bank(s) of the current new loan was a lead bank in a loan of the CEO’s employer in the five years before he or she became the CEO of the current firm, and zero otherwise Dummy (loan initiation or bond issuance) An indicator variable that equals one if the firm takes at least one loan or issued one bond in the fiscal year Yield spread (in basis point) Offering yield spread. The difference between the offering yield at issuance and the yield of the benchmark treasury bond, calculated only for fixed coupon bond (about 78% of the Mergent sample) Bond size (in $${\}$$ millions) Offering amount, the par value of debt at issuance (in $${\}$$ millions) Bond maturity (in months) Maturity date – offering date (in months) Subordinated An indicator variable that equals one if the bond is junior, junior subordinate, subordinate, senior subordinate, and zero otherwise (senior or senior secured) Bond rating Following Jiang et al. (2012), we use the following numerical bond rating: 7 corresponds to AAA; 6 corresponds to AA- to AA$$+$$; 5 corresponds to A- to A$$+$$; 4 corresponds to BBB- to BBB$$+$$; 3 corresponds to BB- to BB$$+$$; 2 corresponds to B$$-$$ to B$$+$$; and 1 corresponds to C-level ratings CDS spread (in basis points) The amount paid by the Protection Buyer to the Protection Seller, typically denominated in basis points, with an annualized quote but paid quarterly. We use the five-year spreads because these contracts are the most liquid and constitute over 85% of the entire CDS market. To maintain uniformity in contracts, we only keep CDS quotations for senior unsecured debt with a modified restructuring (MR) clause and denominated in U.S. dollars CDS(Departure)-CDS(Pre-departure) The rise in CDS spreads between the average CDS spreads from three months before the departure announcement (month -3 to month -1; with month 0 being the departure month) and the departure announcement day CDS(Inauguration)-CDS(Departure) The drop in CDS spreads between the dates of the previous CEO’s departure announcement and when the new CEO takes office Recovery Rate (in percentage) Reported by data contributors. Most pricing methodologies estimate recovery rates in a very simplistic way: a percentage is assigned to the seniority of the debt of a company. For investment grade issuers, recovery is generally assumed to be 40% (as the probability of default is low, the recovery rate is at best an estimate). For distressed issuers however, where the probability of default is higher, recovery tends to be more precisely defined Credit Spread (in basis point) The difference between AAA corporate bond yield and BAA corporate bond yield (data source: Federal Reserve Board of Governors) measured in the month prior to loan initiation Term Spread (in basis point) The difference between the ten-year Treasury yield and the two-year Treasury yield (source: Federal Reserve Board of Governors) measured in the month prior to loan initiation VIX (in percentage) CBOE volatility index, which shows the market’s expectation of 30-day volatility. It is constructed using the implied volatilities of a wide range of S&P 500 index options. This volatility is meant to be forward looking and is calculated from both calls and puts Firm age Age of the firm since IPO, using the first day appear in CRSP (or the IPO date in Compustat if missing), constructed for each firm-year Log(assets) Logarithm of the total book assets (assets are measured in $${\}$$ millions) Leverage (Long-term debt $$+$$ debt in current liabilities)/total assets $$M/B$$ Market value of equity (closing price at the fiscal year end times shares outstanding) divided by book value of equity $$Q$$ (Market value of equity $$+$$ the book value of total debt)/book value of total assets ROA Earnings before interest, tax, and depreciation scaled by the total book assets Tangibility Net property, plant, and equipment/total assets CF volatility Residual volatility of the AR(1) process of ROE, following Pastor and Veronesi (2003) Payout ratio (Dividend/earnings) per share Speculative grade An indicator variable that equals one if the firm has a rating below BBB-, and zero otherwise (investment grade) Highly levered An indicator variable that equals one if leverage is greater than 36% (corresponds to the 75% of the leverage distribution, as well as the mean of the speculative grade firms) Total time in office Equals zero if the CEO (or CFO) came into and exited the office in the same year; one if he or she exited the year after becoming CEO (or CFO), etc. Management ability CEO’s ability, relative to their industry peers, in transforming corporate resources to revenues (see Demerjian, Lev, and McVay 2012 for details). We use the second stage residual of regressing the raw ability scores on firm-level characteristics Outsider CEO An indicator that equals one if the CEO is hired from outside (i.e., never worked for the firm prior to becoming its CEO) Heir apparent CEO An indicator variable that equals one if the new CEO was an heir apparent. An executive with the title “president” or “chief operating officer (COO)” or both, who is distinct from the CEO and the chairman is designated as the “heir apparent” Firm has heir An indicator variable that equals one if the firm has an heir apparent that year, and zero otherwise Young CEO An indicator variable that equals one if the CEO who was younger than 50 when taking office Turnovers due to health or illness Include cases in which (1) news searches revealed that the CEO departure was related to a health condition or death (from Fee, Hadlock, and Pierce 2013) or (2) turnover reason provided in Execucomp is “deceased” Turnovers due to retirement of departing CEO This sample includes turnovers where (1) news searches revealed that the CEO departure was related to a health condition or death (from Fee, Hadlock, and Pierce 2013), (2) turnover reason provided in Execucomp is “deceased,” and (3) departing CEOs older than 65 years. We exclude the “suspicious” retirements by focusing on retirements at good performance. This means that the cumulative monthly stock return during the 12 months before the new CEO’s inauguration month (see the variable definition for Cum. return month [-12,-1] below) is greater than sample median (0.3%) No management shakeup CEO turnovers not accompanied by management (top-four highest paid non-CEO executives) changes during the turnover year and the year after turnover Cum. return month [-12,-1] Cumulative stock return during the 12 months before the inauguration month Median monthly IVOL month[-12,-1] The median of the monthly idiosyncratic volatility during the 12 months before the inauguration month Good preturnover performance Turnovers that satisfy the following three conditions: (1) the median of the monthly idiosyncratic volatility during the 12 months before the inauguration month (see the variable definition for Median IVOL month [-12,-1] above) is less or equal to 6.62% (sample median in Table 2, panel B, Column (1)); (2) the cumulative monthly stock return during the 12 months before the inauguration month (see the variable definition for Cum. return month [-12,-1] above) is no less than 0.3% (sample median in Table 2, panel B, Column (1)); and (3) the ROA in the fiscal year prior to the inauguration month is no less than 12.7% (sample median in Table 2, panel B, Column (1)). ROA is defined as the earnings before interest, tax, and depreciation scaled by the beginning of fiscal year total book assets No preturnover run-up in CDS spread To capture the change in the CDS spread before turnover, we run firm-CEO specific regressions of daily CDS spread on event days [-730, -30], with day 0 being the day when the CEO takes office. We require at least 250 trading day data on CDS spread. Turnovers with nonpositive (or insignificant) tenure-time slope are classified as not preceded by an increase in the CDS spread Outright forced Outright forced turnovers include the “overtly forced” group from Fee, Hadlock, and Pierce (2013) with cases for which news searches indicated that the CEO was forced to leave or exited under pressure Table A1 Variable definitions Variable Definition Loan spread (in basis points) All-in-drawn spread (AIS) over LIBOR at the origination date, from the current pricing file. Winsorized at 1% in the DealScan/Compustat-merged database Loan maturity (in months) A calculation of how long (in months) the facility will be active from signing date to expiration date, from the facility file Loan size (in $${\}$$ millions) The amount of the facility, from the facility data set Secured An indicator variable that equals one if the loan is secured, from the facility file Number of lenders Total number of lenders in a loan, from the lender file Number of loan covenants The total number of covenants in six categories (therefore this variable ranges from zero to six): equity sweep, debt sweep, asset sweep, financial, dividend, and secured, following Bradley and Roberts (2015) Performance pricing A loan feature that ties the interest rate of the loan to an indicator (e.g., leverage, interest coverage ratio) of the firm’s performance, from the performance pricing file Loan type Type of the loan (facility): term loan, revolver, etc. Loan purpose Purpose of the loan (facility): takeover, working capital, debt repayment, etc. Lead bank Following Bharath et al. (2007), we focus on lead bank(s) in the syndicate in relationship lending. Any lender characterized as “lead arranger,” “lead bank,” or “lead manager,” or who has an allocation of more than 90% of the total committed amount to the facility is characterized as a lead bank. Any bank that is described as a “participant” is not a leading bank Prior CEO (CFO)-lender relationship A binary variable that equals one if the lead bank(s) of the current new loan was a lead bank in a loan of the CEO’s employer in the five years before he or she became the CEO of the current firm, and zero otherwise Dummy (loan initiation or bond issuance) An indicator variable that equals one if the firm takes at least one loan or issued one bond in the fiscal year Yield spread (in basis point) Offering yield spread. The difference between the offering yield at issuance and the yield of the benchmark treasury bond, calculated only for fixed coupon bond (about 78% of the Mergent sample) Bond size (in $${\}$$ millions) Offering amount, the par value of debt at issuance (in $${\}$$ millions) Bond maturity (in months) Maturity date – offering date (in months) Subordinated An indicator variable that equals one if the bond is junior, junior subordinate, subordinate, senior subordinate, and zero otherwise (senior or senior secured) Bond rating Following Jiang et al. (2012), we use the following numerical bond rating: 7 corresponds to AAA; 6 corresponds to AA- to AA$$+$$; 5 corresponds to A- to A$$+$$; 4 corresponds to BBB- to BBB$$+$$; 3 corresponds to BB- to BB$$+$$; 2 corresponds to B$$-$$ to B$$+$$; and 1 corresponds to C-level ratings CDS spread (in basis points) The amount paid by the Protection Buyer to the Protection Seller, typically denominated in basis points, with an annualized quote but paid quarterly. We use the five-year spreads because these contracts are the most liquid and constitute over 85% of the entire CDS market. To maintain uniformity in contracts, we only keep CDS quotations for senior unsecured debt with a modified restructuring (MR) clause and denominated in U.S. dollars CDS(Departure)-CDS(Pre-departure) The rise in CDS spreads between the average CDS spreads from three months before the departure announcement (month -3 to month -1; with month 0 being the departure month) and the departure announcement day CDS(Inauguration)-CDS(Departure) The drop in CDS spreads between the dates of the previous CEO’s departure announcement and when the new CEO takes office Recovery Rate (in percentage) Reported by data contributors. Most pricing methodologies estimate recovery rates in a very simplistic way: a percentage is assigned to the seniority of the debt of a company. For investment grade issuers, recovery is generally assumed to be 40% (as the probability of default is low, the recovery rate is at best an estimate). For distressed issuers however, where the probability of default is higher, recovery tends to be more precisely defined Credit Spread (in basis point) The difference between AAA corporate bond yield and BAA corporate bond yield (data source: Federal Reserve Board of Governors) measured in the month prior to loan initiation Term Spread (in basis point) The difference between the ten-year Treasury yield and the two-year Treasury yield (source: Federal Reserve Board of Governors) measured in the month prior to loan initiation VIX (in percentage) CBOE volatility index, which shows the market’s expectation of 30-day volatility. It is constructed using the implied volatilities of a wide range of S&P 500 index options. This volatility is meant to be forward looking and is calculated from both calls and puts Firm age Age of the firm since IPO, using the first day appear in CRSP (or the IPO date in Compustat if missing), constructed for each firm-year Log(assets) Logarithm of the total book assets (assets are measured in $${\}$$ millions) Leverage (Long-term debt $$+$$ debt in current liabilities)/total assets $$M/B$$ Market value of equity (closing price at the fiscal year end times shares outstanding) divided by book value of equity $$Q$$ (Market value of equity $$+$$ the book value of total debt)/book value of total assets ROA Earnings before interest, tax, and depreciation scaled by the total book assets Tangibility Net property, plant, and equipment/total assets CF volatility Residual volatility of the AR(1) process of ROE, following Pastor and Veronesi (2003) Payout ratio (Dividend/earnings) per share Speculative grade An indicator variable that equals one if the firm has a rating below BBB-, and zero otherwise (investment grade) Highly levered An indicator variable that equals one if leverage is greater than 36% (corresponds to the 75% of the leverage distribution, as well as the mean of the speculative grade firms) Total time in office Equals zero if the CEO (or CFO) came into and exited the office in the same year; one if he or she exited the year after becoming CEO (or CFO), etc. Management ability CEO’s ability, relative to their industry peers, in transforming corporate resources to revenues (see Demerjian, Lev, and McVay 2012 for details). We use the second stage residual of regressing the raw ability scores on firm-level characteristics Outsider CEO An indicator that equals one if the CEO is hired from outside (i.e., never worked for the firm prior to becoming its CEO) Heir apparent CEO An indicator variable that equals one if the new CEO was an heir apparent. An executive with the title “president” or “chief operating officer (COO)” or both, who is distinct from the CEO and the chairman is designated as the “heir apparent” Firm has heir An indicator variable that equals one if the firm has an heir apparent that year, and zero otherwise Young CEO An indicator variable that equals one if the CEO who was younger than 50 when taking office Turnovers due to health or illness Include cases in which (1) news searches revealed that the CEO departure was related to a health condition or death (from Fee, Hadlock, and Pierce 2013) or (2) turnover reason provided in Execucomp is “deceased” Turnovers due to retirement of departing CEO This sample includes turnovers where (1) news searches revealed that the CEO departure was related to a health condition or death (from Fee, Hadlock, and Pierce 2013), (2) turnover reason provided in Execucomp is “deceased,” and (3) departing CEOs older than 65 years. We exclude the “suspicious” retirements by focusing on retirements at good performance. This means that the cumulative monthly stock return during the 12 months before the new CEO’s inauguration month (see the variable definition for Cum. return month [-12,-1] below) is greater than sample median (0.3%) No management shakeup CEO turnovers not accompanied by management (top-four highest paid non-CEO executives) changes during the turnover year and the year after turnover Cum. return month [-12,-1] Cumulative stock return during the 12 months before the inauguration month Median monthly IVOL month[-12,-1] The median of the monthly idiosyncratic volatility during the 12 months before the inauguration month Good preturnover performance Turnovers that satisfy the following three conditions: (1) the median of the monthly idiosyncratic volatility during the 12 months before the inauguration month (see the variable definition for Median IVOL month [-12,-1] above) is less or equal to 6.62% (sample median in Table 2, panel B, Column (1)); (2) the cumulative monthly stock return during the 12 months before the inauguration month (see the variable definition for Cum. return month [-12,-1] above) is no less than 0.3% (sample median in Table 2, panel B, Column (1)); and (3) the ROA in the fiscal year prior to the inauguration month is no less than 12.7% (sample median in Table 2, panel B, Column (1)). ROA is defined as the earnings before interest, tax, and depreciation scaled by the beginning of fiscal year total book assets No preturnover run-up in CDS spread To capture the change in the CDS spread before turnover, we run firm-CEO specific regressions of daily CDS spread on event days [-730, -30], with day 0 being the day when the CEO takes office. We require at least 250 trading day data on CDS spread. Turnovers with nonpositive (or insignificant) tenure-time slope are classified as not preceded by an increase in the CDS spread Outright forced Outright forced turnovers include the “overtly forced” group from Fee, Hadlock, and Pierce (2013) with cases for which news searches indicated that the CEO was forced to leave or exited under pressure Variable Definition Loan spread (in basis points) All-in-drawn spread (AIS) over LIBOR at the origination date, from the current pricing file. Winsorized at 1% in the DealScan/Compustat-merged database Loan maturity (in months) A calculation of how long (in months) the facility will be active from signing date to expiration date, from the facility file Loan size (in $${\}$$ millions) The amount of the facility, from the facility data set Secured An indicator variable that equals one if the loan is secured, from the facility file Number of lenders Total number of lenders in a loan, from the lender file Number of loan covenants The total number of covenants in six categories (therefore this variable ranges from zero to six): equity sweep, debt sweep, asset sweep, financial, dividend, and secured, following Bradley and Roberts (2015) Performance pricing A loan feature that ties the interest rate of the loan to an indicator (e.g., leverage, interest coverage ratio) of the firm’s performance, from the performance pricing file Loan type Type of the loan (facility): term loan, revolver, etc. Loan purpose Purpose of the loan (facility): takeover, working capital, debt repayment, etc. Lead bank Following Bharath et al. (2007), we focus on lead bank(s) in the syndicate in relationship lending. Any lender characterized as “lead arranger,” “lead bank,” or “lead manager,” or who has an allocation of more than 90% of the total committed amount to the facility is characterized as a lead bank. Any bank that is described as a “participant” is not a leading bank Prior CEO (CFO)-lender relationship A binary variable that equals one if the lead bank(s) of the current new loan was a lead bank in a loan of the CEO’s employer in the five years before he or she became the CEO of the current firm, and zero otherwise Dummy (loan initiation or bond issuance) An indicator variable that equals one if the firm takes at least one loan or issued one bond in the fiscal year Yield spread (in basis point) Offering yield spread. The difference between the offering yield at issuance and the yield of the benchmark treasury bond, calculated only for fixed coupon bond (about 78% of the Mergent sample) Bond size (in $${\}$$ millions) Offering amount, the par value of debt at issuance (in $${\}$$ millions) Bond maturity (in months) Maturity date – offering date (in months) Subordinated An indicator variable that equals one if the bond is junior, junior subordinate, subordinate, senior subordinate, and zero otherwise (senior or senior secured) Bond rating Following Jiang et al. (2012), we use the following numerical bond rating: 7 corresponds to AAA; 6 corresponds to AA- to AA$$+$$; 5 corresponds to A- to A$$+$$; 4 corresponds to BBB- to BBB$$+$$; 3 corresponds to BB- to BB$$+$$; 2 corresponds to B$$-$$ to B$$+$$; and 1 corresponds to C-level ratings CDS spread (in basis points) The amount paid by the Protection Buyer to the Protection Seller, typically denominated in basis points, with an annualized quote but paid quarterly. We use the five-year spreads because these contracts are the most liquid and constitute over 85% of the entire CDS market. To maintain uniformity in contracts, we only keep CDS quotations for senior unsecured debt with a modified restructuring (MR) clause and denominated in U.S. dollars CDS(Departure)-CDS(Pre-departure) The rise in CDS spreads between the average CDS spreads from three months before the departure announcement (month -3 to month -1; with month 0 being the departure month) and the departure announcement day CDS(Inauguration)-CDS(Departure) The drop in CDS spreads between the dates of the previous CEO’s departure announcement and when the new CEO takes office Recovery Rate (in percentage) Reported by data contributors. Most pricing methodologies estimate recovery rates in a very simplistic way: a percentage is assigned to the seniority of the debt of a company. For investment grade issuers, recovery is generally assumed to be 40% (as the probability of default is low, the recovery rate is at best an estimate). For distressed issuers however, where the probability of default is higher, recovery tends to be more precisely defined Credit Spread (in basis point) The difference between AAA corporate bond yield and BAA corporate bond yield (data source: Federal Reserve Board of Governors) measured in the month prior to loan initiation Term Spread (in basis point) The difference between the ten-year Treasury yield and the two-year Treasury yield (source: Federal Reserve Board of Governors) measured in the month prior to loan initiation VIX (in percentage) CBOE volatility index, which shows the market’s expectation of 30-day volatility. It is constructed using the implied volatilities of a wide range of S&P 500 index options. This volatility is meant to be forward looking and is calculated from both calls and puts Firm age Age of the firm since IPO, using the first day appear in CRSP (or the IPO date in Compustat if missing), constructed for each firm-year Log(assets) Logarithm of the total book assets (assets are measured in $${\}$$ millions) Leverage (Long-term debt $$+$$ debt in current liabilities)/total assets $$M/B$$ Market value of equity (closing price at the fiscal year end times shares outstanding) divided by book value of equity $$Q$$ (Market value of equity $$+$$ the book value of total debt)/book value of total assets ROA Earnings before interest, tax, and depreciation scaled by the total book assets Tangibility Net property, plant, and equipment/total assets CF volatility Residual volatility of the AR(1) process of ROE, following Pastor and Veronesi (2003) Payout ratio (Dividend/earnings) per share Speculative grade An indicator variable that equals one if the firm has a rating below BBB-, and zero otherwise (investment grade) Highly levered An indicator variable that equals one if leverage is greater than 36% (corresponds to the 75% of the leverage distribution, as well as the mean of the speculative grade firms) Total time in office Equals zero if the CEO (or CFO) came into and exited the office in the same year; one if he or she exited the year after becoming CEO (or CFO), etc. Management ability CEO’s ability, relative to their industry peers, in transforming corporate resources to revenues (see Demerjian, Lev, and McVay 2012 for details). We use the second stage residual of regressing the raw ability scores on firm-level characteristics Outsider CEO An indicator that equals one if the CEO is hired from outside (i.e., never worked for the firm prior to becoming its CEO) Heir apparent CEO An indicator variable that equals one if the new CEO was an heir apparent. An executive with the title “president” or “chief operating officer (COO)” or both, who is distinct from the CEO and the chairman is designated as the “heir apparent” Firm has heir An indicator variable that equals one if the firm has an heir apparent that year, and zero otherwise Young CEO An indicator variable that equals one if the CEO who was younger than 50 when taking office Turnovers due to health or illness Include cases in which (1) news searches revealed that the CEO departure was related to a health condition or death (from Fee, Hadlock, and Pierce 2013) or (2) turnover reason provided in Execucomp is “deceased” Turnovers due to retirement of departing CEO This sample includes turnovers where (1) news searches revealed that the CEO departure was related to a health condition or death (from Fee, Hadlock, and Pierce 2013), (2) turnover reason provided in Execucomp is “deceased,” and (3) departing CEOs older than 65 years. We exclude the “suspicious” retirements by focusing on retirements at good performance. This means that the cumulative monthly stock return during the 12 months before the new CEO’s inauguration month (see the variable definition for Cum. return month [-12,-1] below) is greater than sample median (0.3%) No management shakeup CEO turnovers not accompanied by management (top-four highest paid non-CEO executives) changes during the turnover year and the year after turnover Cum. return month [-12,-1] Cumulative stock return during the 12 months before the inauguration month Median monthly IVOL month[-12,-1] The median of the monthly idiosyncratic volatility during the 12 months before the inauguration month Good preturnover performance Turnovers that satisfy the following three conditions: (1) the median of the monthly idiosyncratic volatility during the 12 months before the inauguration month (see the variable definition for Median IVOL month [-12,-1] above) is less or equal to 6.62% (sample median in Table 2, panel B, Column (1)); (2) the cumulative monthly stock return during the 12 months before the inauguration month (see the variable definition for Cum. return month [-12,-1] above) is no less than 0.3% (sample median in Table 2, panel B, Column (1)); and (3) the ROA in the fiscal year prior to the inauguration month is no less than 12.7% (sample median in Table 2, panel B, Column (1)). ROA is defined as the earnings before interest, tax, and depreciation scaled by the beginning of fiscal year total book assets No preturnover run-up in CDS spread To capture the change in the CDS spread before turnover, we run firm-CEO specific regressions of daily CDS spread on event days [-730, -30], with day 0 being the day when the CEO takes office. We require at least 250 trading day data on CDS spread. Turnovers with nonpositive (or insignificant) tenure-time slope are classified as not preceded by an increase in the CDS spread Outright forced Outright forced turnovers include the “overtly forced” group from Fee, Hadlock, and Pierce (2013) with cases for which news searches indicated that the CEO was forced to leave or exited under pressure Footnotes 1 For example, a special document circulated by Moody’s about corporate governance claims: “[T]here is inherent transition risk in any CEO change and we therefore look to evaluate any changes to strategic initiatives or financial policies that differ from previous expectations, and whether credit metrics or liquidity deteriorates as a result.” See Plath (2008). 2 As an indication of the lack of attention given to the topic, the term “management risk” does not appear in the indexes of the three leading MBA textbooks in finance by Brealey, Myers, and Allen (2015), Berk and DeMarzo (2014), and Ross et al. (2015). 3 The SEC’s (2009) Division of Corporate Finance Legal Bulletin 14E states: “a company generally may not rely on Rule 14a-8(i)(7) to exclude a proposal that focuses on CEO succession planning.” 4 An earlier version of this paper contained preliminary empirical results consistent with these predictions. 5 A recent study by Stern (2015) uses a similar approach to compare the relative importance of different types of directors. 6 See https://wrds-web.wharton.upenn.edu/wrds/ds/markit/cds (accessed on January 15, 2016) 7 The modified restructuring clause was introduced in the ISDA standard contract in 2001. This clause limits the scope of opportunistic behavior by sellers in the event of restructuring agreement to deliverable obligations with maturity of 30 or fewer months. This clause applies to the majority of quoted CDS for North American entities. 8 This sample conditions on all the covariates in our baseline specification (in Table 2, Column (1)) to be nonmissing. 9 See https://wrds-web.wharton.upenn.edu/wrds/ds/dealscan (accessed on August 21, 2013) 10 The data are primarily gathered from SEC filings, and the remainder is from direct research by LPC through contacts with borrowers, lenders, and the credit industry at large. Strahan (1999) provides a detailed description of the DealScan database. 11 See http://wrds-web.wharton.upenn.edu/wrds/ds/linkingtable/index.cfm (accessed on August 21, 2013). 12 This sample conditions on all the covariates in Table 4, panel A, Column (1), to be nonmissing. 13 This measure adds to the borrowing spread any annual fees the firms pay to the lenders. 14 See https://wrds-web.wharton.upenn.edu/wrds/ds/fisd (accessed on September 9, 2013) 15 This sample conditions on all the covariates in Table 4, panel B, Column (1), to be nonmissing. 16 Execucomp reports the dates for which the new CEOs officially took office for about 90% of the CEO changes it covers. 17 See Jenter and Lewellen (2014) for the argument that poor performance motivates more CEO turnovers than is typically assumed. Regardless of the underlying motivations for the turnovers, each of the subsamples we consider is highly unlikely to be associated with times of unusually high credit risk. 18 We combine CEO turnover announcements in Capital IQ’s Key Developments with Factiva news search to identify a subsample of such turnovers. We thank Edward Fee, Charles Hadlock, and Joshua Pierce for kindly providing us with the classification of illness, death-related, and outright forced turnovers between 1990 and 2006 used by Fee, Hadlock, and Pierce (2013). 19 Because of the relatively small number of observations with loan data, we pool the different subsamples of likely non-performance-related turnovers together to calculate the estimates reported in Column (3). 20 Besides the loan spread, we also observe information on other nonprice terms of the loan contracts, such as the loan maturity, loan size, number of lenders, whether the loan is secured, and the number of loan covenants. The results are reported in Table IA.2 of the Internet Appendix. The main difference in nonprice terms over CEO tenure is that bank loans that originate earlier tend to have significantly shorter maturities than those that originate later. 21 We do not use Execucomp to identify CFO turnovers because there is no reliable indicator for CFOs in the pre-2007 data and there is no information on the time that a new CFO takes office. For the firm-years between 2007 and 2009, we have verified that the two data sources are consistent in 86% of observations. 22 This finding is consistent with the findings in Bennedsen, Pérez-González, and Wolfenzon (forthcoming) based on managers’ hospitalization records that CEOs are more important to firm value than other top executives. 23 Ralph Bender, CFO of the Manship Media Group, for example, suggests that a successful CFO should be a technical generalist, rather than specializing in one area: “The key to being a successful CFO is not so much knowing everything, but knowing a little bit about a lot of things, trying to stay abreast of these things” (see Lamoreaux 2009). 24 This literature began with Rajan’s (1992) analysis. Petersen and Rajan (1994, 1995), Berger and Udell (1995), Schenone (2010), Bharath et al. (2007, 2011), and Karolyi (Forthcoming) all document that relationship-based loans have lower spreads than otherwise identical loans in which there is not a prior relationship between the firm and its lender. These studies also suggest that the loan market is competitive enough that the benefit from reduced information asymmetry is at least partly passed onto the borrower. 25 In the firms with speculative grades in our loan sample, the majority (more than 96%) have an issuer credit rating between BB- and B, so the vast majority of our sample firms are not in default. 26 Almost all of the loans in our sample are senior, so we cannot consider the effect of seniority using the loan sample. 27 The other differences in the specifications between the columns come from the features of the different markets. For firms with a traded CDS, we have daily values for the CDS, so we estimate our equation using daily data. With daily data, we choose to include firm-CEO fixed effects, so the direct effects of speculative grade and high leverage cannot be estimated since these firm characteristics are measured at the time of each turnover. Using the loan and bond data, we only have one observation for each time a firm takes out a loan or issues a bond, so we measure tenure in years and use firm fixed effects rather than firm-CEO fixed effects, and the direct effects of firm characteristics, such as speculative grade and high leverage, can be estimated because a firm may have multiple turnovers. 28 In an untabulated test, we also find that when a firm’s debt is more risky, there is a larger increase in the firm’s CDS spread at its CEO departure announcement. This finding is consistent with the result in Table 8 that management risk is higher for firms with riskier debt claims. References Akins, B., De Angelis, D. and Gaulin. M. 2016 . Debt contracting on management. Working Paper. Bennedsen, M., Pérez-González, F. and Wolfenzon. D. Forthcoming. Do CEOs matter: Evidence from hospitalization events. Journal of Finance . Berger, A. N., and Udell. G. F. 1995. Relationship lending and lines of credit in small firm finance. Journal of Business 68 : 351 – 81 . Google Scholar CrossRef Search ADS Berk, J., and DeMarzo. P. 2014 . Corporate finance . Boston : Pearson . Blanco, R., Brennan, S. and Marsh. I. W. 2005 . An empirical analysis of the dynamic relationship between investmentgrade bonds and credit default swaps. Journal of Finance 60 : 2255 – 81 . Bharath, S., Dahiya, S. Saunders, A. and Srinivasan. A. 2007 . So what do I get? The bank’s view of lending relationship. Journal of Financial Economics 85 : 368 – 419 . Google Scholar CrossRef Search ADS Bharath, S., Dahiya, S. Saunders, A. and Srinivasan. A. 2011 . Lending relationships and loan contract terms. Review of Financial Studies 24 : 1141 – 203 . Google Scholar CrossRef Search ADS Bradley, M., and Roberts. M. 2015 . The structure and pricing of debt covenants. Quarterly Journal of Finance 5 : 1 – 37 . Google Scholar CrossRef Search ADS Brealey, R., Myers, S. and Allen. F. 2015 . Principles of corporate finance . New York : McGraw Hill. Chava, S., and Roberts. M. 2008 . How does financing impact investment? The role of debt covenants. Journal of Finance 63 : 2085 – 121 . Google Scholar CrossRef Search ADS Collin-Dufresne, P., Goldstein, R. S. and Martin. J. S. 2001 . The determinants of credit spread changes. Journal of Finance 56 : 2177 – 207 . Google Scholar CrossRef Search ADS Custódio, C., Ferreira, M. and Matos. P. 2013 . Generalists vs. specialists: Lifetime work experience and chief executive officer pay. Journal of Financial Economics 108 : 471 – 92 . Google Scholar CrossRef Search ADS Demerjian, P., Lev, B. and McVay. S. 2012 . Quantifying managerial ability: A new measure and validity tests. Management Science 58 : 1229 – 48 . Google Scholar CrossRef Search ADS Fee, C. E., Hadlock, C. J. and Pierce. J. R. 2013 . Managers who lack style: Evidence from exogenous CEO changes. Review of Financial Studies 26 : 567 – 601 . Google Scholar CrossRef Search ADS Graham, J. R., Li, S. and Qiu. J. 2008 . Corporate misreporting and bank loan contracting. Journal of Financial Economics 89 : 44 – 61 . Google Scholar CrossRef Search ADS Hermalin, B. E., and Weisbach. M. S. 2017 . Assessing managerial quality: Implications for corporate governance. In Handbook of the economics of corporate governance , eds. Hermalin B. and Weisbach. M. Amsterdam : North Holland. Jenter, D., and Kanaan. F. 2015 . CEO turnover and relative performance evaluation. Journal of Finance 70 : 2155 – 83 . Google Scholar CrossRef Search ADS Jenter, D., and Lewellen. K. 2014 . Performance-induced CEO turnover. Working Paper. Jiang, J., Stanford, M. H. and Xie. Y. 2012 . Does it matter who pays for bond ratings? Historical evidence. Journal of Financial Economics 105 : 607 – 21 . Google Scholar CrossRef Search ADS Kaplan, S. N., Klebanov, M. M. and Sorensen. M. 2012 . Which CEO characteristics and abilities matter? Journal of Finance 67 : 973 – 1007 . Google Scholar CrossRef Search ADS Karolyi, S. A. Forthcoming. Personal lending relationships. Journal of Finance . Lamoreaux, M. G. 2009 . CFO 101, five prerequisites. Journal of Accountancy 208 : 3 – 5 . Mansi, S., Maxwell, W. and Miller. D. 2011 . Analyst forecast characteristics and the cost of debt. Review of Accounting Studies 16 : 116 – 42 . Google Scholar CrossRef Search ADS Mian, S. 2001 . On the choice and replacement of chief financial officers. Journal of Financial Economics 60 : 143 – 75 . Google Scholar CrossRef Search ADS Naveen, L. 2006 . Organizational complexity and succession planning. Journal of Financial and Quantitative Analysis 41 : 661 – 83 . Google Scholar CrossRef Search ADS Pan, Y., Wang, T. Y. and Weisbach. M. S. 2015 . Learning about CEO ability and stock return volatility. Review of Financial Studies 28 : 1623 – 66 . Google Scholar CrossRef Search ADS Pan, Y., Wang, T. Y. and Weisbach. M. S. 2016 . CEO investment cycles. Review of Financial Studies 29 : 2955 – 99 . Google Scholar CrossRef Search ADS Pastor, L., and Veronesi. P. 2003 . Stock valuation and learning about profitability. Journal of Finance 58 : 1749 – 89 . Google Scholar CrossRef Search ADS Petersen, M., and Rajan. R. 1994 . The benefits of firm-creditor relationships: Evidence from small business data. Journal of Finance 49 : 3 – 37 . Google Scholar CrossRef Search ADS Petersen, M., and Rajan. R. 1995 . The effect of credit market competition on lending relationships. Quarterly Journal of Economics 110 : 407 – 43 . Google Scholar CrossRef Search ADS Plath, C. 2008 . Analyzing credit and governance implications of management succession planning. Working Paper. Rajan, R. 1992 . Insiders and outsiders: The choice between informed and arm’s-length debt. Journal of Finance 47 : 1367 – 400 . Google Scholar CrossRef Search ADS Ross, S., Westerfield, R. Jaffe, J. and Jordan. B. 2015 . Corporate finance . New York : Irwin McGraw-Hill. Schenone, C. 2010 . Lending relationships and information rents: Do banks exploit their information advantage? Review of Financial Studies 23 : 1149 – 99 . Google Scholar CrossRef Search ADS Strahan, P. 1999 . Borrower risk and the price and nonprice terms of bank loans. Working Paper. Stern, L. 2015 . A learning based approach to evaluating boards of directors. Working Paper. Security Exchange Commission (SEC). 2009 . Shareholder proposals. Staff Legal Bulletin No. 14E (CF), Division of Corporation Finance, Washington, DC. https://www.sec.gov/interps/legal/cfslb14e.htm ( accessed on February 24, 2017 ). van Binsbergen, J. H., Graham, J. R. and Yang. J. 2010 . The cost of debt. Journal of Finance 65 : 2089 – 136 . Google Scholar CrossRef Search ADS © The Author 2017. Published by Oxford University Press on behalf of The Society for Financial Studies. All rights reserved. For permissions, please e-mail: 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/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Review of Financial Studies Oxford University Press

# How Management Risk Affects Corporate Debt

, Volume 31 (9) – Sep 1, 2018
41 pages