Credit Default Swaps, Agency Problems, and Management Incentives

Recent literature has focused on conflicts between shareholders and creditors in the presence of Credit Default Swaps (CDS), but these papers have generally overlooked potential agency problems arising between shareholders and managers.

CDS are insurance contracts between two parties with contingent payoffs referenced to future credit events of the underlying entity. Credit events that trigger CDS payments are potentially endogenous, since they may occur as outcomes of unobserved interactions between shareholders and creditors. Several studies recently demonstrate that CDS, once issued, may affect bargaining between shareholders and creditors on an ongoing basis. Since CDS provide creditors with a form of default insurance, they could strengthen their bargaining power in debt negotiations, which in turn could generate significant feedback effects on corporate financial policies. Lenders who heavily hedge their debt exposure with CDS become “empty creditors” (Hu and Black, 2008; Bolton and Oehmke, 2011), who may frequently reject debt exchange offers from shareholders in order to avail themselves of voting rights that may attach during bankruptcy. Such feedback effects of CDS on various corporate policies have been extensively documented, including increasing corporate leverage (Saretto and Tookes, 2013), more frequent outright liquidations (Subrahmanyam, Tang, and Wang, 2014), and precautionary corporate liquidity management (Subrahmanyam, Tang, and Wang, 2017) following the onset of CDS trading for the debt of a specific firm. Importantly, these studies rely on the inception of CDS trading as an identification strategy, since the listing of a new CDS contract occurs as a result of decisions taken by third-party market makers who typically have no connection with the firm’s managers.

This paper demonstrates that two types of managerial agency problems may occur as a result of CDS introduction. First, managers may engage in less strategic debt default than optimal, since creditors will bargain harder due to the presence of the CDS insurance contract. Second, creditors may tend to monitor managers with reduced intensity, since they become less concerned about the consequences of non-repayment once they have access to CDS insurance. We present a theoretical model of how CDS affects the incentives for managers to divert assets in the form of perquisites, and we show that agency problems tend to increase through both of these channels. We then use a sample of S&P 1500 firms between 2001−15 to study whether boards change CEO compensation awards after CDS are introduced.

As our key measures for a CEO’s risk-taking and value enhancement incentives, we compute the deltas and vegas for new stock and stock option compensation awards to CEOs of our sample firms. In regression analysis we find that managerial vega increases by 133% from its sample average following the inception of CDS trades. Moreover, we find that both managerial option and share deltas significantly increase after CDS introduction, leading to an increase in overall managerial delta. We find a $41,000 increase in the CEO’s wealth from these newly granted incentive contracts when the underlying stock price increases by one dollar.

Increasing both the CEO’s effective ownership and risk-taking incentives helps to mitigate the new agency problems between managers and shareholders. However, if we focus on the agency problem arising from the extinguishment of the shareholders’ strategic default option, those two contractual adjustments affect the manager differently. The agency cost induced by strategic default is mainly driven by managerial actions that increase the probability of default. Since greater effective ownership only motivates the manager in non-default states, it has little impact on the manager’s motivation to threaten creditors with a default. In contrast, greater risk-taking incentives are effective in this setting. Our empirical evidence is consistent with this conjecture, as we find that for firms with valuable strategic default options, stock option grants (which have nonzero vegas) increase significantly after CDS inception while restricted stock grants (which have zero vega) do not change significantly.

We conduct a robustness test to alleviate concerns about sample selection bias and the influence of confounding endogenous market forces. For example, one could argue that because CDS contracts are not created randomly, our results are driven by omitted firm-level characteristics. We address these selection issues by using an instrumental variable defined at each firm’s main lender level, and therefore, could exclude firm level characteristics as potential confounders that hinder the causal interpretation of our key findings. We show that our results continue to hold after they are re-estimated in an instrumental variables framework.

Our study sheds light on the relative importance of managerial agency problems and incentive contracts when one studies the impact of the empty creditor’s problem in the presence of CDS. While a number of theory papers have identified the potential costs of empty creditors with CDS (Hu and Black, 2008; Bolton and Oehmke, 2011), this agency problem has not been documented empirically. And while many studies find positive impacts of CDS on corporate debt capacity and value (Saretto and Tookes, 2013; Danis and Gamba, 2017), we posit that such net positive consequences of CDS could be driven by important omitted factors, including action by a firm’s board that recognizes potential agency problems and adjusts the CEO’s contract to counteract them. Our findings therefore help explain why the empirical estimate of the cost of the empty creditors’ problem appears to be lower than theoretical predictions. Further work in this area would help set a useful micro-foundation for understanding real effects of CDS.

The complete paper is available for download here.