Safer ratios, riskier portfolios: Banks׳ response to government aid

https://doi.org/10.1016/j.jfineco.2014.03.005Get rights and content

Abstract

Using novel data on bank applications to the Troubled Asset Relief Program (TARP), we study the effect of government assistance on bank risk taking. Bailed-out banks initiate riskier loans and shift assets toward riskier securities after receiving government support. However, this shift in risk occurs mostly within the same asset class and, therefore, remains undetected by regulatory capital ratios, which indicate improved capitalization at bailed-out banks. Consequently, these banks appear safer according to regulatory ratios, but show an increase in volatility and default risk. These findings are robust to controlling for credit demand and account for selection of TARP recipients by exploiting banks׳ geography-based political connections as an instrument for bailout approvals.

Introduction

The financial crisis of 2008–2009 resulted in an unprecedented liquidity shock to financial institutions in the U.S. (Gorton and Metrick, 2012) and abroad (Beltratti and Stulz, 2012). To stabilize the banking system, governments around the world initiated a wave of capital assistance to financial firms. Many economists and regulators argue that this wave altered the perception of the government safety net (Kashyap, Rajan, and Stein, 2008) and created a precedent that will have a profound effect on the future behavior of financial firms. At the forefront of this debate is the effect of the bailout on bank risk taking (Flannery, 2010), since risk taking, coupled with inadequate regulation (Levine, 2012), is often blamed for leading to the crisis in the first place. This debate has broad policy implications, since the relation between government intervention and bank risk taking is at the core of financial system design (Song and Thakor, 2011). This paper studies whether and how the recent bailout affected risk taking in credit origination and investment activities of U.S. banks.

Our empirical analysis exploits an economy-wide liquidity shock during the 2008–2009 financial crisis, which simultaneously affected an unusually large cross-section of firms and resulted in a bailout of hundreds of firms. In particular, we study the effect of the Capital Purchase Program (CPP), which invested $205 billion in U.S. financial institutions, becoming the first and largest initiative of the Troubled Asset Relief Program (TARP). Using hand-collected data on the status of bank applications for federal assistance, we observe both banks׳ decisions to apply for bailout funds and regulators׳ decisions to grant assistance to specific banks. This setting allows us to account for selection of bailed firms and to study the risk taking implications of both bailout approvals and bailout denials. Our risk analysis spans three channels of bank operations: (1) retail lending (mortgages), (2) corporate lending (syndicated loans), and (3) investment activities (financial assets).

Our empirical analysis begins with the retail credit market. By examining both approved and denied loan applications for nearly all residential mortgages in 2006–2010, our empirical strategy distinguishes the supply-side changes in bank credit origination from the demand-side changes in potential borrowers. In difference-in-difference tests, where the first difference is between banks that were granted and denied government assistance, and the second difference is from before to after the bailout, we find no significant effect of CPP on the volume of credit origination at approved banks, compared to their denied peers. We also find no significant change in the distribution of borrowers between approved and denied banks. Our main finding is that after being approved for federal assistance, banks shifted their credit origination toward riskier mortgages. This result holds whether we compare approved banks to denied banks, to non-applicant banks, or to all CPP-eligible banks. In economic terms, we find that relative to banks that were denied federal assistance, approved banks increased their origination rates on riskier mortgage applications (measured by the loan-to-income ratio) by 5.4 percentage points.

Our findings are qualitatively similar for large corporate loans. Using a similar difference-in-difference framework, we find a robust shift by approved banks toward higher-yield, riskier loans. After being approved for federal assistance, banks increased credit issuance to riskier firms, as measured by borrowers׳ cash flow volatility, interest coverage, and asset tangibility, and reduced credit issuance to safer firms. Altogether, our findings for both retail and corporate loans suggest that the bailout was associated with a shift toward higher-yield loans at approved banks rather than an expansion in credit volume.

We find a similar increase in risk taking by approved banks in their investment activities. After being approved for federal assistance, banks increased their investments in risky securities, such as non-agency mortgage-backed securities, and reduced their allocations to low-risk securities, such as Treasury bonds. For the average bank approved for federal assistance, the total weight of investment securities in bank assets increased by 9.7% after CPP relative to unapproved banks. Moreover, approved banks increased their allocations to risky securities, while, at the same time, reducing their allocations to lower-risk securities relative to unapproved banks. Overall, our analysis at the micro-level indicates a robust increase in risk taking in both lending and investment activities by banks approved for government assistance.

After providing micro-level evidence on the drivers of risk taking, we examine aggregate bank risk. First, we show that federal capital infusions improved capitalization levels of approved banks, with their average Tier-1 capital ratios increasing by approximately 160 basis points relative to unapproved banks. Second, we find that the reduction in leverage at approved banks was more than offset by their shift toward riskier assets. The net effect was a marked increase in the aggregate risk of approved banks compared to observably similar unapproved banks. This result holds robustly whether bank risk is measured by earnings volatility, stock volatility, market beta, or distance to default. For example, after the bailout, approved banks show a 20.9% increase in default risk (measured by the z-score) and a 15.3% increase in beta relative to unapproved banks.

We provide evidence that the shift in risk taking at approved banks is attributable to the treatment effect of government support rather than selection of approved firms. First, we explicitly control for proxies of the declared CPP selection criteria. We also capture any time-invariant heterogeneity between approved and unapproved banks via bank fixed effects. Second, we use propensity score matching of approved and unapproved banks based on firm fundamentals to allow for various functional forms of the relation between bank characteristics and risk. Finally, we use an instrumental variable approach, which relies on banks׳ geography-based political connections as an instrument for bailout approvals. In particular, we show that banks located in election districts of House members who served on key finance subcommittees during the development of CPP were more likely to be bailed out, while being virtually indistinguishable from unconnected banks based on other observable characteristics ex ante. We obtain similar results across these specifications.

We review three non-mutually exclusive explanations for the observed increase in risk at approved banks: (1) government intervention, (2) risk arbitrage, and (3) moral hazard. The first hypothesis—government intervention—posits that the increase in risk taking at approved banks is a consequence of government intervention in bank policies aimed at increasing capital flows into subprime mortgages and mortgage-backed securities. However, to the extent that bailed banks were subject to government regulations, these regulations sought to reduce rather than increase risk taking, for example, by limiting executive pay “to prevent excessive risk taking” and by restricting share repurchases and dividends to prevent asset substitution.

To investigate this hypothesis, we collect data on banks that applied for CPP, were approved, but did not receive CPP funds for various institutional reasons discussed in Section 5.2. We then compare risk taking by this subset of non-recipients to the banks that did receive the money and had similar size, financial condition, and performance at the time of CPP approval. We find a similar increase in risk taking across all banks approved for bailout funds, regardless of whether they received the money and were subject to the subsequent government regulation. As another test of the government intervention hypothesis, we examine changes in bank risk taking after the repayment of CPP capital. We find that the release from government oversight after the repayment of CPP funds has little effect on bank risk taking. Collectively, these results suggest that if government intervention played a role in banks׳ credit and investment policies, it was unlikely the primary driver of risk taking.

The second hypothesis—risk arbitrage—states that some risky assets, such as subprime mortgages and investment securities, were underpriced during the crisis, providing excess profit opportunities with low risk. In this case, CPP capital may have enabled approved banks to exploit these opportunities without an ex post increase in risk. In contrast, we find no evidence that an increase in risk taking at approved banks was followed by superior risk-adjusted returns, as proxied by alpha, the Sharpe ratio, or the information ratio. Rather, the shift toward higher-yield assets was associated with an increase in loan chargeoffs and, if anything, a slight decline in alphas at approved banks. Overall, while the extra capital likely played a role in approved banks׳ investment and lending decisions, these decisions reflected an increase in risk tolerance rather than low-risk arbitrage opportunities.

A third explanation—moral hazard—posits that a firm׳s approval for federal funds may signal its implicit government protection. According to this view, there is some ex ante probability that a given bank will be bailed out in case of distress. During a financial shock, the bank either receives government protection or is denied it. If there is some consistency in the regulators׳ treatment of banks across time, a bank׳s approval for government support signals an increase in the probability that this bank will be protected again in case of distress. Conversely, if a bank is denied government aid, the probability that this bank will be bailed out in the future goes down. This effect can be particularly significant in the short term, since the government will prefer to avoid the near-term distress of banks it has publicly declared to endorse. As an example of this continued government support, about 21% of CPP recipients were allowed to skip their dividends to the Treasury. Under this view, the bailout may encourage risk taking by protected banks by reducing investors׳ monitoring incentives and increasing moral hazard, as predicted in Acharya and Yorulmazer (2007) and Kashyap, Rajan, and Stein (2008), among others.

Our evidence suggests that moral hazard likely contributed to the increase in risk taking at approved banks. First, the finding that higher risk taking is associated with a signal of government support rather than with the capital injection itself is consistent with the effect of a revised probability of government protection in theoretical work (Mailath and Mester, 1994, Acharya and Yorulmazer, 2007). Second, the cross-sectional evidence aligns well with the predictions from models of moral hazard. In particular, the increase in risk taking is stronger at larger banks, banks that are closer to financial distress, and banks that received multiple signals of government forbearance in the form of skipped dividends. Finally, we find that approved banks increase their risk primarily by investing in assets with a high exposure to common macroeconomic risk, which is also reflected in an increase in banks׳ stock betas. If government protection is more likely in the case of a systematic rather than idiosyncratic shock to a firm, this evidence is consistent with a rational response of protected banks to a revised probability of future government support. This interpretation is also supported by the evaluation of CPP by its chief auditor, the Special Inspector General for the Troubled Asset Relief Program (SIGTARP).1 It is also consistent with the views about a shift in bailed banks׳ risk tolerance expressed by prominent regulators in a testimony to Congress.2

Our article has important policy implications. First, one of the significant recent events was a negative revision of the outlook for long-term U.S. debt by Standard and Poor׳s, followed by a downgrade in August 2011 for the first time since the beginning of ratings in 1860. Among the reasons for a revised outlook cited by the rating agency were the increased risk of U.S. banks and a higher probability of another bailout.3 Our paper identifies potential sources of the increased risk in the financial system and links them to the initial bailout policy and the predictions of academic theory. Second, earlier studies underscore the importance of bank capital for credit origination (Thakor, 1996) and economic growth (Levine, 2005). Our findings suggest an asymmetric response of banks to capital shocks. In particular, while previous research shows that a negative shock to bank capital forces a cut in lending (Berger and Bouwman, 2013), we find that a positive shock to capital need not result in credit expansion, but instead may lead to riskier lending and investments. Finally, though capital requirements are a key instrument in bank regulation (Bernanke and Lown, 1991), we show that banks׳ strategic response to this mechanism erodes its efficacy in monitoring bank risk.

Section snippets

Theoretical motivation and main hypotheses

The government safety net has been long recognized as a cornerstone of the economic system. Its architecture includes social assistance programs, government insurance, and financial regulation. We adopt this broader perspective and begin with a review of theoretical work on government guarantees in general economic settings. We then proceed with a more specific discussion of government guarantees in financial regulation.

The early theoretical work on government guarantees has focused on social

Capital purchase program

The Emergency Economic Stabilization Act (EESA), signed into law on October 3, 2008, created TARP, a system of federal initiatives aimed at stabilizing the financial system. The first and largest of these initiatives was CPP. Initiated on October 14, 2008, this program invested $204.9 billion in 707 firms in 2008–2009.

To apply for CPP funds, a qualifying financial institution (QFI)a domestic bank, bank holding company, savings association, or savings and loan holding companysubmitted a

Baseline evidence on retail lending

In this section, we study the effect of CPP on credit origination and risk taking in the mortgage market. We begin with a difference-in-difference model of credit origination, where the first difference is from before to after CPP, and the second difference is between approved banks and various control groups: denied banks, non-applicant banks, or all eligible banks. We continue with a matched sample analysis and instrumental variable regressions.

To isolate banks׳ active lending decisions from

Additional evidence and possible explanations

In this section, we provide cross-sectional evidence on bank risk taking, examine several explanations for our results, and discuss robustness tests. Throughout the rest of the paper, we estimate our tests using the three methods discussed above: baseline difference-in-difference model, matched samples, and IV regressions. Our conclusions are similar across these tests. For brevity, we report baseline difference-in-difference results in the paper and offer evidence from the two other methods in

Extensions

In this section, we extend our analysis by studying the effect of CPP on two other channels of bank operations: (1) corporate credit and (2) portfolio investments. While we believe that the richness of data in the mortgage market provides the cleanest empirical setting, we offer these additional tests as complementary evidence.

Bank risk

In this section, we study how changes in bank credit policies and portfolio investments after CPP affected aggregate bank risk. Since, broadly defined, the two primary sources of bank risk are leverage and asset composition, we first examine the effect of CPP on capital ratios and continue with evidence on aggregate risk.

Conclusion

This paper has investigated the effect of government assistance on bank risk taking. While we do not find a significant effect of government assistance on the aggregate credit supply, our results suggest a considerable effect on the risk of originated loans. After being approved for federal funds, CPP participants issue riskier loans and increase capital allocations to riskier, higher-yield securities, as compared to banks that were denied federal funds. A fraction of CPP funding is also used

References (87)

  • G. Gorton et al.

    Securitized banking and the run on repo

    Journal of Financial Economics

    (2012)
  • V. Ivashina

    Asymmetric information effects on loan spreads

    Journal of Financial Economics

    (2009)
  • L. Laeven et al.

    Bank governance, regulation, and risk taking

    Journal of Financial Economics

    (2009)
  • R. Levine

    Finance and growth: theory and evidence

  • G. Mailath et al.

    A positive analysis of bank closure

    Journal of Financial Intermediation

    (1994)
  • R. Merton

    An analytic derivation of the cost of deposit insurance and loan guarantees

    Journal of Banking and Finance

    (1977)
  • M. Puri et al.

    Global retail lending in the aftermath of the US financial crisis: distinguishing between supply and demand effects

    Journal of Financial Economics

    (2011)
  • P. Sapienza

    The effects of government ownership on bank lending

    Journal of Financial Economics

    (2004)
  • D. Tang et al.

    Market conditions, default risk and credit spreads

    Journal of Banking and Finance

    (2010)
  • P. Veronesi et al.

    Paulson׳s gift

    Journal of Financial Economics

    (2010)
  • V. Acharya et al.

    A pyrrhic victory? Bank bailouts and sovereign credit risk

    Journal of Finance

    (2014)
  • D. Bayazitova et al.

    Assessing TARP

    Review of Financial Studies

    (2012)
  • A. Berger et al.

    Bank liquidity creation

    Review of Financial Studies

    (2009)
  • Berger, A., Bouwman, C., Kick, T., Schaeck, K., 2012. Bank risk taking and liquidity creation following regulatory...
  • A. Berger et al.

    Bank competition and financial stability

    Journal of Financial Services Research

    (2009)
  • Berger A., Roman R., 2013. Did TARP banks get competitive advantages? Unpublished working paper. University of South...
  • B. Bernanke et al.

    The credit crunch

    Brookings Papers on Economic Activity

    (1991)
  • M. Blume et al.

    The declining credit quality of U.S. corporate debt: myth or reality?

    Journal of Finance

    (1998)
  • J. Boyd et al.

    The theory of bank risk taking and competition revisited

    Journal of Finance

    (2005)
  • C. Brown et al.

    Too many to fail? Evidence of regulatory forbearance when the banking sector is weak

    Review of Financial Studies

    (2011)
  • J. Bulow et al.

    A constant recontracting model of sovereign debt

    Journal of Political Economy

    (1989)
  • K. Burdett et al.

    Unemployment insurance and short-time compensation: the effects on layoffs, hours per worker, and wages

    Journal of Political Economy

    (1989)
  • C. Calomiris et al.

    The last trillion-dollar commitment: the destruction of Fannie Mae and Freddie Mac

    Journal of Structured Finance

    (2009)
  • Campbell, J., Cocco, J., 2011. A model of mortgage default. Unpublished working paper. Harvard University and London...
  • M. Campello

    Internal capital markets in financial conglomerates: evidence from small bank responses to monetary policy

    Journal of Finance

    (2002)
  • I. Cheng et al.

    The hazards of debt: rollover freezes, incentives, and bailouts

    Review of Financial Studies

    (2012)
  • L. Cohen et al.

    Do powerful politicians cause corporate downsizing?

    Journal of Political Economy

    (2011)
  • G. Dell׳Ariccia et al.

    Credit booms and lending standards: evidence from the subprime mortgage market

    Journal of Money, Credit and Banking

    (2012)
  • R. Demsetz et al.

    Banks with something to lose: the disciplinary role of franchise value

    Economic Policy Review, Federal Reserve Bank of New York

    (1996)
  • Y. Demyanyk et al.

    Understanding the subprime mortgage crisis

    Review of Financial Studies

    (2011)
  • Douglas, A., Huang, A., Vetzal, K., 2012. Cash flow volatility and corporate bond yield spreads. Unpublished working...
  • R. Ehrenberg et al.

    Unemployment insurance, duration of unemployment, and subsequent wage gain

    American Economic Review

    (1976)
  • M. Faccio et al.

    Political connections and corporate bailouts

    Journal of Finance

    (2006)
  • Cited by (367)

    • The longer-term impact of TARP on banks’ default risk

      2024, Quarterly Review of Economics and Finance
    • Bank market power and supervisory enforcement actions

      2024, International Review of Financial Analysis
    View all citing articles on Scopus

    We thank the Mitsui Life Financial Center at the University of Michigan and the Millstein Center for Corporate Governance at Yale University for financial support. We gratefully acknowledge the helpful comments from an anonymous referee, Sumit Agarwal, Andrea Beltratti, Christa Bouwman, Andrew Ellul, Charles Hadlock, Vasso Ioannidou, Chris James, Augustin Landier, Gyöngyi Lóránth, Mitchell Petersen, Tigran Poghosyan, N.R. Prabhala, and James Vickery, as well as conference participants at the 2013 Western Finance Association (WFA) Annual Meeting, the 2013 Symposium on Financial Institutions and Financial Stability at the University of California at Davis, the 2013 Conference on Financing the Recovery after the Crisis at Bocconi University, the 2013 WU Gutmann Center Symposium on Sovereign Credit Risk and Asset Management, the 2012 NYU Credit Risk Conference, the 2012 Adam Smith Corporate Finance Conference at Oxford University, the 2012 Journal of Accounting Research Pre-Conference at the University of Chicago, the 2012 Singapore International Conference on Finance, the 2012 CEPR Conference on Finance and the Real Economy, the 2012 Financial Stability Conference at Tilburg University, the 2012 IBEFA Annual Meeting, the 2011 Financial Intermediation Research Society (FIRS) Annual Meeting, the 2011 FDIC Banking Research Conference, the 2011 FinLawMetrics Conference at Bocconi University, and the 2011 Michigan Finance and Economics Conference and seminar participants at the Board of Governors of the Federal Reserve System, Emory University, Hong Kong University of Science and Technology, Michigan State University, Norwegian Business School, Norwegian School of Economics, the University of Hong Kong, the University of Illinois at Urbana-Champaign, the University of Illinois at Chicago, the University of Maryland, the University of Michigan, the University of Washington, and Vanderbilt University.

    View full text