23. Criticisms of Bailouts Generally

We now turn to five general criticisms of government bailout efforts: (1) taxpayers can suffer losses, (2) bailouts may not work or may be prolonged, (3) bailouts create moral hazard, (4) government decisions over bailout may be political and ad hoc, and (5) bailouts may fail to boost lending activities.

23.1  Taxpayer Loss

Historically bailouts have had varying impact on taxpayers. The Reconstruction Finance Corporation (“RFC”) during the Great Depression had a cumulative profit of $160 million on its capital of $500 million.1 However, the bailout of Continental Illinois National Bank eventually cost the FDIC $1.1 billion2 and taxpayers paid for “$123.8 billion, or 81 percent of the total costs” of the savings and loans crisis.3 As discussed above, TARP’s bank bailout programs will not result in any tax dollar loss; indeed they will be profitable. However, were they profitable enough in light of the risk Treasury took?

Even though bailouts may, in the end, not be costly for taxpayers, one does not know this in advance of the expenditure. At the time TARP was authorized by the Congress, the estimated cost to taxpayers was much higher.4 However, if the objective is to avoid taxpayer losses, one can provide that financial institutions or their investors—or some subset of them—bear any eventual losses through ex post assessments. Section 134 of the Emergency Economic Stabilization Act of 2008, which authorized TARP, provides that the President shall submit a legislative proposal that recoups from the financial industry an amount equal to the shortfall of TARP to ensure that the bailout does not add to the deficit or national debt.5 Pursuant to this Section 134, President Obama in 2013 proposed the “Financial Crisis Responsibility Fee,” which would be imposed on certain financial institutions with $50 billion or more in consolidated assets.6 The President included this fee in his budget proposal for fiscal year 2014, aiming to collect roughly $59 billion in such fees over a ten-year period.7 However, the President’s proposal has never been acted on, and as stated there have been no losses from the CPP program.

At the request of G20 leaders, the IMF in June 2010 proposed two general alternatives to shift the burdens associated with government interventions to the financial sector: (1) a financial stability contribution levied upon financial institutions based on certain attributes such as size and riskiness, or (2) a financial activities tax based on bank profits.8 These proposals were not limited to financing government past losses from interventions, as more fully developed below.

The plan to impose a global tax was subsequently shelved but various European countries have considered similar national legislation.9 In September 2011 the European Commission also proposed a Europe-wide financial transaction tax (FTT).10 In May 2012 the European Parliament passed the FTT11 and in January 2013 approved a plan to allow 11 European Union countries to pursue a FTT.12 Nevertheless, this European FTT has encountered multiple delays and is expected to be implemented no earlier than 2017.13

The imposition of such a “Tobin tax” would supposedly curb what some see as useless financial activities. The proceeds from the FTT, which has received support from French and German leaders but staunch opposition from the United Kingdom,14 would be used to compensate European governments for rescue costs during the financial crisis, but also, more generally, to “contribute to the public finances” during the ongoing European fiscal crisis.15 The FTT has also been justified as a form of compensation for an implicit state guarantee, a roughly-fashioned (but explicit) premium for an implicit bailout guarantee.16 However, the FTT is poorly designed for this purpose, since it taxes all financial transactions indiscriminately, rather than charging those firms that are most likely to be bailed out, based either on their size or the riskiness of their activities.17

There are several concerns with the ex post tax assessment approach to pay for bailouts. First, it is hard to determine how such ex post assessments should be apportioned, such as to banks and nonbanks.18 The challenges are essentially the same as those for an insurance regime funded through ex post assessments, as previously discussed in chapter 12. Second, failing global coordination, banks may be subject to double or even triple taxation.19 Third, there is some concern that such taxes could turn out to be politically unenforceable when the financial crisis is over, but taxpayers should blame politicians and not banks for that outcome.20 Finally, there is the fear that the government could use the opportunity to overtax the banks, as the EU financial transactions tax proposal arguably demonstrates, to achieve other objectives.

23.2  Bailouts May Not Work or Be Prolonged

Before the 2008 financial crisis, the most successful bailout in mature economies was probably the Swedish bank bailout in the early 1990s. The collapse of the property market left several Swedish banks with large quantities of soured real estate loans. The government announced a blanket guarantee of bank debt and took over the major banks. Sweden eventually incurred minimum cost after selling its bank interests several years later.21 However, bailout does not always work. It may carry the risk that one bailout will evolve into multiple efforts to prop up insolvent banks for an extended period of time without any real hope of recovery.

The Japanese “lost decade” is a prominent example of a prolonged bailout. Until very recently large and small Japanese banks have been saddled with bad loans since the collapse of its stock and real property markets in 1990.22 Through the 1990s the government purchased nonperforming loans (NPL) from banks and tried other rescue measures. However, it continued to delay in recognizing the full scale of the NPL problem by endorsing questionable accounting practices due to the high social cost that would follow from corporate bankruptcies and the political consequences of admitting large losses.23 In 1998, when the level of NPL became extremely high, the government purchased 1.8 trillion yen ($16 billion) in subordinated debt and preferred shares in 21 major banks that were undercapitalized. The bailout failed to stabilize the market and the government nationalized two major banks, the Long-Term Credit Bank of Japan and Nippon Credit Bank, followed by another injection of 7.5 trillion yen ($71 billion) in 15 banks the next year.24 Bailouts continued thereafter, including the injection of 1.96 trillion yen (around $19 billion) in Resona in 2003. The government also consolidated the biggest banks during this period to create some of the largest banks in the world.

The Japanese financial system finally began to stabilize after its economy started to recover and the government began to address the NPL problem seriously around 2003.25 By 2006 banks appeared to be able to repay public funds within a few years. The largest ones have already done so,26 but regional banks continue to struggle.27 Some nationalized banks were sold to nonbanking and foreign owners.28

The Japanese experience demonstrates that fundamental problems with banking systems may only be prolonged, but not resolved, through bailouts. Bailouts may be the beginning and not the end of financial recovery. Bailing out banks without dealing with the NPL problem was throwing good money after bad. The Congressional Oversight Panel in early 2009 agreed that CPP and other bank investment programs would work based on the key assumption that the financial crisis was in large part the product of temporary liquidity constraints resulting from nonfunctioning markets for troubled assets.29 In other words, bailouts should only be used to curb risk contagion and stabilize the market so that the government could have breathing time to implement other cleanup efforts such as proactive write-downs, reform of banking practices and gradual sales of assets.

23.3  Creation of Moral Hazard and Too-Big-to-Fail Competitive Advantage

The fear of moral hazard is the strongest argument against government bailouts. Both individual firms and the market may have perverse incentives if they know the government will come to the rescue. The consequence of this moral hazard is that firms will take on more risk than would otherwise be optimal because risk taking becomes a one-sided bet.30 Investors, especially debt investors, may have less incentive to monitor the performance of the firm. As the firm becomes too big to fail, meaning presumptively entitled to a bailout, it may also enjoy an unfair competitive advantage over other firms because its cost of financing could be cheaper.31 These concerns are similar to the ones regarding insurance or guarantees.

Summers identifies an excessive fear of moral hazard among a group he calls “moral hazard fundamentalists,” and suggests these fundamentalists are misguided in three respects as they analogize the moral hazard of bailout to the moral hazards raised by insurance: first, individual actors in the financial world may underestimate the role of contagion and the benefits that their own insurance will have on other actors; thus they will tend to underinsure (suggesting the free market may not be an adequate way to address the risks they pose). Second, institutions may fail simply because of a loss of confidence, rather than because of increased risk taking. In these cases, the possibility of bailout can help to avoid panic and contagion. However, confidence (of short-term creditors) is better addressed through the lender of last resort. Finally, unlike insurance, bailout can actually leave taxpayers better off, for example, when a government program like TARP is potentially profitable. For all these reasons, policy should not be developed simply on the basis of “avoiding moral hazard,” but rather must take into account contagion, potential liquidity runs and the benefits of quelling panic, and the potential costs and benefits of a bailout to taxpayers.32

In my view, however, bailouts (public capital injections) are not the answer to contagion; this can be avoided by a strong lender of last resort and guarantees. Nor are bailouts necessary to avoid single large bank failures (this may be different in countries where one or two banks dominate an economy), assuming there are effective resolution procedures and the preservation of critical functions. Bailouts like TARP should be reserved in the United States for correlation related systemic risk when multiple large banks are rendered insolvent by a common external event. There is no denying that bailouts increase moral hazard and put taxpayers at risk, so they should only be used when there is no feasible alternative. Even when necessary bailouts should be designed to minimize moral hazard.

Government bailouts should wipe out existing common shareholders, thus helping reduce moral hazard; at the very least the common should be heavily and permanently diluted through the creation of new common. Government bailouts could also be combined with some hit on longer term creditors (e.g., through bail-ins), as discussed in chapter 16. These longer term creditors should therefore have incentive to monitor risk.33 Indeed, since bond losses are tightly linked to ratings, and rating agencies are usually not certain that a given financial institution will be bailed out, they may downgrade poorly capitalized banks, a threat also reducing incentives of managers to take undue risk. Rating agencies obviously consider implicit government backing in its rating decisions. In 2011 S&P downgraded Bank of America, Citigroup, Goldman Sachs, Morgan Stanley, JPMorgan, and Wells Fargo.34 In September 2011, Moody’s also downgraded Citigroup, Bank of America and Wells Fargo. Both rating agencies that believed the government to be less likely to assist these banks going forward than during the financial crisis when contagion risk was high.35 Moody’s also considered the enactment of Dodd–Frank Act as showing the government’s intent to impose losses on bondholders in future crises.36 In 2012, citing the “clear intent of government around the world to reduce support for creditors,” Moody’s downgraded five of the six largest US banks: JPMorgan, Bank of America, Citigroup, Goldman Sachs, and Morgan Stanley.37 More recently, in November 2013, Moody’s further downgraded JPMorgan, Goldman Sachs, Morgan Stanley, and Bank of New York Mellon.38 A managing director at the rating agency explained that “rather than relying on public funds to bail out one of these institutions, we expect that bank holding company creditors will be bailed-in and thereby shoulder much of the burden to help recapitalize a failing bank.”39

Bernardo, Talley, and Welch, using a model that focuses exclusively on moral hazard, show that bailouts can be welfare-enhancing if (1) they are only used sparingly, where social externalities are large and subsidies are small; (2) the government eliminates incumbent owners, board, and managers to improve a priori incentives; and (3) the bailout is funded through redistributive taxes on healthy firms rather than forcing recipients to repay in the future, as the government has already fully expropriated existing owners and managers.40 With regard to the problem that bank managers may take excessive risks in exchange for large bonuses and then leave before the ship sinks, even a no-bailout rule could not correct the problem and the solution may lie elsewhere, such as compensation reforms and clawback requirements. Goodhart and Avgouleas argue that a bail-in may “be much superior to [a] bailout in the case of idiosyncratic failure.”41 However, “bail-in regimes will not eradicate the need for injection of public funds where there is a threat of systemic collapse.”42

23.4  Bailout Decisions May Be Political and Ad Hoc

Some have claimed that the use of TARP funds was determined based on political rather than actual systemic risk grounds.43 One prominent example was the bailout of GMAC, which showed that the presence of systemic risk was not a condition for granting public assistance because GMAC’s rescue had little or nothing to do with mitigating the financial crisis.44 Some claim that some companies are simply “too connected to fail” due to their executives’ extensive connections with the key decision makers within the federal government.45 Further the assistance given to small banks under TARP had little to do with concern with systemic risk.

Public confidence in the bailout effort can be seriously damaged if it is perceived by the public that the government did not follow any clearly articulated goals and principles in making important decisions.46 This was particularly the case in the pre-TARP period. Bailouts often seemed to be ad hoc responses to an impending crisis—“when a major financial institution got into trouble, the Treasury Department and the Federal Reserve would engineer a bailout over the weekend and announce that everything was fine on Monday.”47 This is another reason why it may be better to have a standing program rather than responding ad hoc.

23.5  Bailouts May Fail to Boost Lending Activities

Bailout efforts often serve the twin purposes of stabilizing the financial system and alleviating the adverse impacts on the real economy caused by the collapse of the lending market. Obviously there is some inherent tension between the two purposes—if the financial system fails largely due to the failure of businesses in the real economy, as in the case of Japan, extending credit to these failed businesses would simply generate new bad debt. However, if the real economy could have maintained its good shape under normal lending conditions and the crisis is caused by a failure within the financial system itself, as many consider to be the case in the recent crisis, a government bailout of the financial system should increase business lending, grow the real economy, boost the financial performance of the banks and in turn facilitate the government’s exit from its investments.

Although general lending conditions have significantly improved since the peak of the financial crisis in 2008, many blamed TARP for failing to revive the real economy. To start, critics of TARP often point to the flaw in CPP’s design. When the UK government made equity investments in RBS and Lloyds, there were explicit contractual requirements that they maintain their level of lending at pre-crisis levels.48 On the contrary, no similar requirement was imposed by US Treasury for CPP participants. To be fair, the FDIC has instructed the banks it regulates to monitor the use of TARP funds as well as the use of money raised with FDIC debt guarantees.49 In addition an interagency statement urges all banking organizations to make loans to creditworthy borrowers.50

Treasury was also faulted for failing to implement proper measures to monitor the actual use of TARP funds, prompting the Congressional Oversight Panel to repeatedly ask Treasury where the money went.51 Treasury’s response was that money was fungible so it was impossible to correlate the TARP funds with specific uses of funds.52 As pointed out by the Panel, Treasury’s claims were challenged by a survey of the Special Inspector General for the Troubled Asset Relief Program (SIGTARP) demonstrating that banks could provide meaningful information on their use of TARP funds without much difficulty.53 The Panel also pointed out that some banks voluntarily disclosed information on the use of their TARP funds in public filings.54 But all said, it is likely the low level of lending was dictated by low demand, not just low supply, due to the general economic stagnation brought on by the financial crisis.

I would conclude that despite the criticisms, which may be largely justified, under dire circumstances, like those that existed in October 2008, bailouts may be the only tenable solution. Simultaneous bail-ins of several large financial institutions, assuming such bail-ins would work perfectly, could still leave the financial system so weak that it could not finance the recovery of the real economy. In any event, the bailout option should exist if needed. A standing TARP’s design should benefit from a diagnosis of how the old TARP was found wanting and could be improved. We turn to that issue in the next chapter.