Are bank bailouts good?
Following the 2007-2008 financial crisis, governments worldwide provided massive support to distressed financial institutions, either directly through bailouts or indirectly through accommodating fiscal and monetary policies. While the literature generally accepts that such support was essential to prevent a collapse of the global financial system, its long-run effects have been heavily scrutinized for potentially encouraging banks to gamble irresponsibly at the expense of taxpayers and sowing the seeds for the next systemic crisis.
As the recent mini-banking crisis following the bankruptcy of Silicon Valley Bank reignited the debate over bank bailouts, I became curious about the academic research on this topic. In this paper, I examine the effects of bank bailouts, particularly how government bailouts affect banks’ risk-taking incentives. While bailouts create moral hazard at both individual and collective levels, they may also reduce systemic risk by increasing banks’ charter value or by reducing contagion risk across banks. There is still no consensus on which side will prevail. I point to some empirical research and conclude by explaining the need for further research.
Background
In 1983, Douglas Diamond and Philip Dybvig published their seminal paper “Bank Runs, Deposit Insurance, and Liquidity”, where they showed that banks are vulnerable to runs by construction because they fund a portfolio of risky loans and investments with short-term deposits and debts. If many depositors rush to withdraw, the bank may become illiquid. Thus, an individual bank faces two sources of risk. First is portfolio risk which depends on the quality of the bank’s borrowers or investments, which the bank can control through costly monitoring or screening. Second is leverage risk which depends on the amount of cash a bank reserves from
borrowing or investing. Investing in riskier projects or taking on more leverage tends to yield more profits; however, they also increase the probability that the bank may become insolvent at the end of the period.
While a car company going bankrupt is an opportunity for its competitors, a bank going bankrupt after a deposit/funding run is a threat to the entire finance industry, especially if the bank is large and systemically important (“too-big-to-fail”). That is because one bank run can spread through contagion, leading to financial crises on a bigger scale, causing simultaneous difficulties at many other financial institutions, and disrupting many other agents’ ability to borrow (Romer, 2021).
See our interview of Romer: Nobel Prize Winning Economist & Stanford Professor Paul Romer Sits Down With Rebellion Research
The risk of contagion is one of the reasons that makes banks special, and researchers have identified four sources of contagion. Macro contagion and confidence contagion describe how the failure of one bank worsens macroeconomic fundamentals or general borrowers’ confidence, leading to difficulties at other banks (Acharya and Yorulmazer, 2008). Therefore, governments want to prevent any bank run ex ante as much as possible. To do so, they can provide deposit insurance and impose capital requirements or more supervision to restrict portfolio and leverage risks.
Counterparty contagion and fire sale contagion result from interbank exposures. Banks hold various types of claims on one another. When one bank faces a run, hence a risk of failure, the value of its counterpart institutions’ claims on it falls, pushing them into solvency issues as well. Similarly, a distressed bank often needs to sell assets at a loss to meet depositors’ demand (“fire sale”), which lowers asset prices in the market and consequently jeopardizes the balance sheets of other banks (Diamond and Rajan, 2005). Therefore, an overly interconnected financial system is undesirable because it increases the probability of joint failures that the government wants to avoid in the first place.
It is impossible for an individual bank in the modern banking system to fully protect itself from contagion. This is obvious for macro and confidence contagion, but the same holds for fire sale risk and counterparty risk. First, modern banks are inevitably exposed to each other through the interbank market, and these exposures often have to be with the too-big-to-fail banks. Second, even if a bank cuts its exposure to a risky institution, it cannot ensure that its idiosyncratically safe counterparties have done the same (Acemoglu et al., 2012). While banks have some control over their portfolio and leverage risk-taking, the indirect links are beyond any individual bank’s control. That is, the risk of contagion is exogenous to individual banks — it cannot be managed or diversified through the bank’s own operations.
I define the systemic risk of the financial system to be the aggregate idiosyncratic risks from each bank’s portfolio and leverage choices, and the risk of contagion given the failure of some bank in the financial system. Governments sometimes must step in in a crisis to save the financial system from collapsing; at the same time, they must also consider the long-run effects of intervention. The most optimal outcome is to both fight the fire at hand and lower systemic risk of the financial system to prevent future crises.
Moral Hazard
For further reading on ‘Moral Hazard’: Long Term Capital Management LTCM : The Roots of Lehman’s Fall?
A bank bailout is government intervention or support to distressed banks. Such support could be direct — capital injections, nationalizations — or more creative, such as accommodating fiscal and monetary expansions (as we’ve seen in 2008) and facilitated acquisitions by healthier banks (as we’ve seen in UBS’ recent acquisition of Credit Suisse). There is some ambiguity. For example, deposit insurance is by definition also a form of government support, and taxpayers’ deposits are used to fund the FDIC (in the U.S.); however, deposit insurance is not usually considered a bailout. For clarity, I use the definition from Fahri
and Tirole and categorize bank bailouts into interest-rate policies and transfer policies. Interest rate policies include various forms of government intervention that lower borrowing costs for banks: lowering the fed funds rate, extending debt guarantees, accepting low-quality assets as collaterals, etc. Transfer policies are governments’ direct transfers that boost the net worth of banks through recapitalizations, purchasing legacy assets at inflated prices, etc. (Fahri and Tirole 2012).
Not only do regulators face an asymmetry of information with transfer policies — they are unsure about which banks are actually distressed — direct transfers also leave rents to incumbent shareholders of distressed banks. Banks are protected by limited liability, so they repay depositors only if they are successful. Incumbent shareholders retain claims on the bank’s future income, which would have been zero if the bank had failed, but becomes positive with a bailout, benefiting the shareholders.
Such bailout rents generate moral hazard. If banks expect to be bailed out by the government in case of crises, the shareholders can shift all downside risk to taxpayers. Then, they have increased moral hazard incentives to invest more recklessly and take on more leverage to maximize profits as much as possible, resulting in a high level of systemic risk in equilibrium.
Interest-rate policies create another layer of moral hazard. As former Chair of the Federal Reserve Ben Bernanke said, adjusting rates to achieve monetary goals is like “using a sledgehammer to kill a mosquito”. Using an imperfectly targeted tool like interest rates to bail out banks inevitably causes unintended consequences. This theory was proposed by Diamond and Rajan. Since a low ex post interest rate offers a very low reward for maintaining liquidity, they argue that if the government is expected to cut rates when liquidity is at a premium, banks will take on more leverage or illiquid loans beforehand, thus bringing about the very states where intervention is needed (Diamond and Rajan, 2009).
In addition to individual moral hazard, more research has been done to show that the expectation of support can also create collective moral hazard. Because interest rates affect the entire economy, there is a significant “fixed cost” to bail out banks through interest-rate policies. Therefore, bailouts are cheaper when a large number of banks are in distress, making policymakers less reluctant to incur the fixed cost associated with adjusting rates. Since support is more likely to come when there are multiple simultaneous failures, banks are incentivized to “herd” by choosing correlated risks or becoming more interconnected (entering into more contracts like loans and derivatives with other banks, etc.), so that they will fail when the largest possible number of other banks are failing.
An overly interconnected banking system is undesirable because it is susceptible to counterparty and fire sale contagion (Acharya and Yorulmazer, 2007). In addition to correlating risks, Fahri and Tirole show that such fixed cost also makes systemic risk-taking a self-perpetuating cycle — banks will take more risks when other banks take more risks. They argue that each bank’s private leverage choice depends on how banks anticipate policy to react to overall maturity mismatch (borrowing short to lend long). This generates strategic complementarities in balance-sheet riskiness choices — it is therefore unwise to play safely when everyone else gambles (Farhi and Tirole, 2012).
“Constructive ambiguity” is often recommended to policymakers to attenuate moral hazard, according to which the central bank should retain some discretion as to when and under what circumstances the government would rescue a distressed bank (Cordella and Levy-Yeyati, 2003). This makes risk management a game of chicken: the government refutes the notion of bailouts until the very last minute, when the financial system is truly on the brink, and then comes to its rescue.
The threat can be made more credible by increasing political obstacles to intervention (such as some provisions from the Dodd-Frank Act) (Dell’Ariccia and Ratnovski, 2012). A risky strategy, but the risk becomes compensated by alleviating ex ante distortion of incentives.
However, constructive ambiguity doesn’t eliminate moral hazard. First, any ex post bailout would undermine the central bank’s credibility by indicating that it is willing to accommodate excessive risk-taking. Second, loose interest rate policies today encourage new maturity mismatches by making short-term debt cheaper, resulting in more leverage and risk correlation, bringing us into the vicious cycle that Diamond and Rajan wrote about.
Systemic Insurance
In contrast to the literature mentioned above, there is another school of thought outside the American academia that argues government bailout can have an opposite effect on bank incentives that balances the moral hazard problem. By preventing contagion and its associated externalities, bailouts could reduce banks’ risk-taking.
This school of thought expands from the literature on government intervention as a means of preventing contagion (Allen and Gale, 2001; Diamond and Rajan, 2005). The key is to consider how the risk externality from contagion affects banks’ incentives. In equilibrium, each bank takes excessive idiosyncratic risk because of limited liability; systematically, banks take too much risk relative to the coordinated solution because no one considers the effects of their risk taking on other banks.
Since all banks still become affected by contagion externalities. As a result, the high exogenous risk of contagion reduces private return to costly portfolio monitoring and screening, prompting banks to take more risks for higher returns, leading to an inefficiently high-equilibrium probability of a crisis. If governments can remove the threat of an exogenous risk by guaranteeing bailouts, a bank has more incentives to monitor its portfolio prudently, so there is a lower level of overall risk.
This observation was first made by Cordella and Levy-Yeyati. By announcing and committing ex ante to a bailout policy that is contingent upon adverse macroeconomic conditions, government bailouts create a risk-reducing “value effect” that always more than offsets the moral hazard effects of the policy (Cordella and Levy-Yeyati, 2003). The value effect exists because guaranteed bailouts make a bank’s higher probability of survival in an unpreventable contagion, thus raising the value at stake for the bank, which, in turn, increases the bank’s incentives to protect it .
There are two preconditions for the value effect to prevail!
First, the government’s commitment to rescue distressed banks must be explicit, contrary to the conventional constructive ambiguity approach. Second, the macroeconomic shock on which the policy is contingent must be exogenous and not the result of the bank’s own bad decision-making. Without the second condition, Cordella and Levy-Yeyati’s insight still holds. However, the beneficial risk-reducing value effect now weakened, since the moral hazard component becomes more significant.
Dell’Ariccia and Ratnovski extend this model by considering the contingent macroeconomic shocks as exogenous to a bank but endogenous to the financial system as a whole since it depends on risk-taking by all banks (Dell’Ariccia and Ratnovski, 2013). This offers a link between aggregate risk-taking and systemic risk. (The two are not equivalent because contagion has externalities.) They argue ex ante commitment of government support entails a virtuous “systemic insurance” effect. When a bank can fail due to exogenous circumstances, it does not invest as much to protect itself from idiosyncratic risk.
Moreover, this is coherent with Farhi and Tirole’s conclusion that banks take more risk when other banks take more risk (though for a different reason). À la Dell’Ariccia and Ratnovski, since a bank’s success depends on both its effort and the overall stability of the banking system, a government’s commitment to shield banks from contagion risk that they cannot control not only creates a value effect, it also decreases the endogenous systemic risk. In the Dell’Ariccia and Ratnovski model, a higher probability of bailout increases moral hazard since it leaves rents on the table for failing banks.
However, at the same time, it corrects for the externality stemming from the threat of contagion, protecting banks from a risk that they cannot control. Therefore, when the threat of contagion given failure is high, while the rents left to a failing bank are small, the systemic insurance effect prevails over moral hazard costs. More concretely, the systemic insurance effect dominates in a financial system with weak banks and a high probability of contagion, but well designed bank resolution rules that minimize bailout rents.
There are several limitations in these two models. For one, both assume that any announced bailout policy is credible, and neither considers the time inconsistency of policy implementation. In practice, the fixed cost problem may still encourage banks to take correlated risks. These stylized choices can certainly become challenged. But their insight introduces the long-run virtuous effects of bailouts in addition to moral hazard. It then becomes our question of interest how these two forces would interact with each other.
Empirical Research
It is unlikely that there will be one simple answer for which force will prevail, and the answer will most likely resemble the model from Dell’Ariccia and Ratnovski. That is, depending on how one models policy instruments, risk-taking, financial contagion, etc., there exists some set of parameter values (bailout rents, probability of contagion, etc.) with which government bailouts lead to lower/higher systemic risk.
Some economists argue that, at the end of the day, the effect of bailout policies is fundamentally an empirical question. Unfortunately, there is no consistent answer in empirical literature either.
Examining this question is challenging for several reasons. There is no consensus on how to measure systemic risk, so different conclusions might just come from different evaluation metrics. Moreover, we cannot observe the counterfactual of what the bailed-out banks would have been like without bailouts, and it is difficult to separate the effects of bailouts from the effects of other factors, such as the differences among banks.
There are two studies on bank bailouts in Germany that recorded opposite results. Dam and Koetter (2012) showed that a higher probability of being bailed out increases German banks’ risk-taking significantly, reflected by banks’ z-score (probability of insolvency) and risk weighted assets ratio (Dam and Koetter, 2012). In contrast, Berger, Bouwman, Kick, and Schaeck (2016) find that capital support for banks is associated with significant reductions in risk-taking by the same metrics (Berger et al., 2016).
There are many studies on the effects of one specific bailout, the U.S. Troubled Assets Relief Program (TARP), designed to prevent a financial collapse. Additionally, reduce systemic risk during the Great Recession. Results are mixed from researchers that investigate how TARP changed banks’ loaning patterns (Duchin and Sosyura, 2014; Black and Hazelwood, 2013; Berger et al., 2017), as well as those from researchers that study TARP’s effects on banks’ credit supply (Bassett and Demiralp, 2014; Duchin and Sosyura, 2014; Berger et al., 2017). There is one research from Berger, Roman and Sedunov (2020) — the most comprehensive to date — that is worth highlighting.
Furthermore, they considered both sides of the theoretical literature and developed a rigorous examination method. Quite surprisingly, they found that TARP “significantly reduced contributions to systemic risk, particularly for larger and safer banks located in better local economies” (Berger et al., 2020). Their results proved robust to additional tests, including accounting for potential endogeneity and selection bias. Their findings provide great insights into how to design bailouts most optimally and which banks might be the best bailout targets.
Conclusion
Financial crisis is a recurring phenomenon, and the consequences of bailouts will affect the financial system for a considerable time afterward. Therefore, understanding the different effects of bailouts helps policymakers know their options when future financial crises occur. As I’ve showed in this paper, bailouts affect systemic risk through different channels (leverage risk, portfolio risk, and the risk of contagion given failure). Knowledge of which channels are most effective is valuable to inform policymakers how to use their tools to achieve welfare-improving government intervention. Like Dell’Ariccia and Ratnovski concluded, governments should not focus on avoiding bailouts, but on making policies effective. To them, the most optimal policy leaves banks with as little rent as possible. To Diamond and Rajan or Farhi and Tirole, liquidity infusion or liquidity requirements are the most benign. Many questions are left for future research.
Are bank bailouts good? Do Bank Bailouts Increase or Reduce Systemic Risk? Written by Luxy Sun
References
Acemoglu, D.; Ozdaglar, A.; Tahbaz-Salehi, A.; 2012, “Systemic Risk and Stability in Financial Networks.” MIT.
Acharya, V., Yorulmazer, T., 2007. “Too many to fail — An Analysis of Time-inconsistency in Bank Closure Policies.” Journal of Financial Intermediation: 16, 1-31.
Acharya, V., Yorulmazer, T., 2008. “Cash-in-the-Market Pricing and Optimal Resolution of Bank Failures.” Review of Financial Studies: 21, 2705-2742.
Allen, F., Gale, D., 2001. “Financial Contagion.” Journal of Political Economy: 108, 1-33. Bassett, W.F., Demiralp, S., 2014. “Government Support of Banks and Bank Lending.” Board of Governors of the Federal Reserve System Working Paper.
Berger, A. N., Bouwman C. H. S., Kick T. K., Schaeck K., 2016. “Bank Risk Taking and Liquidity Creation Following Regulatory Interventions and Capital Support.” Journal of Financial Intermediation: 26, 115-141.
Berger, A. N., Makaew, T., Roman, R. A., 2017. “Do Borrowers Benefit from Bank Bailouts during Financial Crises? The Effects of TARP on Loan Contract Terms.” Working Paper, University of South Carolina.
Berger, A. N., Roman, R. A., Sedunov, J., 2020. “Did TARP reduce or increase systemic risk? The effects of government aid on financial system stability.” Journal of Financial Intermediation:Volume 43.
Black, L., Hazelwood, L., 2013. “The Effect of TARP on Bank Risk-taking.” Journal of Financial Stability: 9, 790-803.
Cordella, T., Yeyati, E. L., 2003. “Bank Bailouts: Moral Hazard versus Value Effect.” Journal of Financial Intermediation: 12, 300-330.
Dam, L., Koetter, M., 2012. Bank Bailouts and Moral Hazard: Empirical Evidence from Germany. Review of Financial Studies: 25, 2343-2380.
Dell’Ariccia, G., Ratnovski, L., 2012. “Bailouts, Contagion, and Bank Risk-Taking.” Society for Economic Dynamics, 2012 Meeting Papers 133.
Dell’Ariccia, G., Ratnovski, L., 2013. “Bailouts and Systemic Insurance.” International Monetary Fund Working Papers, 2013/233.
Diamond, D. W., Rajan, R. G., 2005. “Liquidity Shortages and Banking Crises.” Journal of Finance: 60, 615-47.
Diamond, D. W., Rajan, R. G., 2009. “Illiquidity and Interest Rate Policy.” National Bureau of Economic Research.
Duchin R., Sosyura, D., 2014. “Safer Ratios, Riskier Portfolios: Banks’ Response to Government Aid.” Journal of Financial Economics: 113, 1-28.
Farhi, E., Tirole, J., 2012. “Collective Moral Hazard, Maturity Mismatch, and Systemic Bailouts.” American Economic Review: 102, 60-93.
Romer, D., 2021. Advanced Macroeconomics, McGraw-Hill Education.
