The “shadow financial system” of government guarantees

We are witnessing the massive use—perhaps even abuse—of guarantees, which suggests that policy makers consider them a “free lunch” that permits them to bypass budgetary scrutiny in the US and other countries. My objective in this article is not to rail against guarantees, but rather to inform about the consequences of their use and to advocate for their transparent and judicious use in a very narrow set of instances.

One blunt instrument used in previous financial crises is government blanket guarantees. Such guarantees for depositors and creditors were introduced in East Asia (and subsequently in Ecuador and Turkey) to restore confidence in and protect the stability of the banking system. While the guarantees stabilized the situation, they limited the subsequent options for dealing with financial distress. Equally, the guarantees created a sense of complacency equivalent to injecting morphine in a patient, in the resolution process. While the morphine takes away the pain it does not address the underlying problem. It only delays the restructuring, while increasing the costs. The lessons of experience from the Asia crisis suggest that blanket guarantees can have adverse consequences for the stability of the financial system and the speed of economic recovery and that these consequences can weigh as heavily on the system as the fiscal costs.

Are we about to repeat the same mistakes? The answer is yes. Ironically, the guarantees have been used excessively in the United States and other countries as a substitute for on-balance-sheet government funding by the same policy makers who rallied to criticize the “shadow banking system.” The policy makers have created a “shadow budget system” of their own, with guarantees that are neither scrutinized nor allocated by the Congress. The extensive use of guarantees is only now coming to light in the case of AIG. The “stealth” use of government blanket guarantees in the recent crisis has not received the scrutiny it deserves from legislators, policy analysts, and academic researchers.

How big are the guarantee programs? It is not entirely clear that there is a complete accounting of the guarantees that were dispensed so freely over the last 18 months, only bits and pieces of evidence from fragmented sources. According to Bloomberg News the commitments are extremely large and potentially very costly. In the United States alone, government guarantees account for 79% of GDP. Bloomberg tabulated the data from Federal Reserve, Treasury, and FDIC and regulatory officials, the US government has provided more than $11.6 trillion to date in government guarantees on behalf of American taxpayers (24 February 2009) (http://www.bloomberg.com/apps/news?pid=newsarchive&sid=aZchK__XUF84). Were the government guarantees priced in order to instill market discipline, as Kotlikoff and Mehrling propose in a recent FT Economist Forum article? Not likely. The evidence points to ad hoc “weekend” improvisation in the cases of Citi and AIG.

What will be the ultimate fiscal cost of this commitment? We do not know, but we can offer an indicative range and it is high, upward of $1.5 trillion. In Thailand, where the guarantees covered the liabilities of the banking system, Idanna Kaplan-Appio, (“Estimating the Value of Implicit Government Guarantees to Thai Banks.” Review of International Economics 10 (1): 26–35, 2002) estimated the contingent liability of the Thai government by modeling the costs as a put option on the value of bank assets. It estimated the value of the government guarantee was nearly 15 percent of the banking system’s liabilities. Therefore, the potential costs of the guarantees are by no means negligible. While the institutional foundations of the US financial system are stronger, the scope of the crisis is global and the instruments far more complex than in the Asia crisis and a 15 percent cost is reasonable. In short, the US was committed to $1.76 trillion governmental expenditure without any congressional budgetary scrutiny.

For example, the U.S. Temporary Liquidity Guarantee Program (TLGP) provides debt guarantees for newly issued senior unsecured debt. The Transaction Account Guarantee Program provides guarantees for funds in non-interest-bearing transaction accounts. The entities eligible for the TLGP are insured depository institutions; U.S. bank-holding companies that control at least one subsidiary that is an insured depository institution; U.S. savings and loan holding companies that control at least one subsidiary that is an insured depository institution; or any other affiliate of an insured depository institution that the Federal Deposit Insurance Corporation (FDIC) designates. But the blanket guarantees extend farther than the TLGP and the Bloomberg article offers a detailed breakdown.

Guarantees are used extensively by other governments. The Royal Bank of Scotland’s Overview on Guarantee Schemes, third edition, dated 30 January 2009, provides an informed overview of bank debt guarantee schemes around the globe. According to this report, numerous countries have established guarantee schemes in recent months: Australia, Austria, Canada, Denmark, Finland, France (Dexia S.A.), Germany (NORD/LB), Greece, Ireland, Italy, Korea, the Netherlands, New Zealand, Portugal, Spain, Sweden, the United Kingdom, and the United States. Jurisdictions as diverse as Austria, France, Australia, and New Zealand permit the issuance of five-year government-guaranteed banking paper. Other jurisdictions, including the Netherlands and Demark, have or are planning to extend the maturities on government-guaranteed paper, and the United Kingdom will allow rollovers up to April 2014. In Japan, the reopening of Japan’s samurai bond market is helped by government guarantees on bond issues.

What are the risks of the guarantees? The various guarantee schemes provided by the governments of the United States, European countries, and the rest of the world have considerable risks associated with them. Yet, they have hardly received scrutiny by policy analysts and academics. An exception is a recent article by Ed Kane, “Safety-net Subsidies Keep ‘Toxic’ Assets Illiquid” Ed points out that guarantees encourage “zombie” institutions to hang onto worthless toxic assets on the remote chance that the market for toxic assets will recover.

What are the findings in the academic literature? The theoretical literature is unequivocal regarding the moral hazard associated with blanket guarantees. It points out that governments limit their policy options by implementing blanket guarantees that extend forbearance. Moreover, the fiscal costs of a crisis are not predetermined. If the underlying problems get swept under the rug, the “silent” crisis lingers on, and the costs escalate. In the absence of a resolution, the recovery is delayed. In my empirical research on the topic I found that the blanket guarantees variable is robust to any specification, including controlling for endogeneity and the depth of the crisis. Blanket guarantees have an adverse effect on fiscal costs, while lengthening the duration of the crisis and the GDP loss. Other academic literature finds that accommodative policies, reflected in blanket guarantees and other forms of forbearance add to the fiscal cost of banking crises but do not accelerate the speed of recovery. Much of the variation in fiscal costs is explained by poor policy measures, such as muddling through. The literature favors a stricter response to crisis resolution.

The puzzle of why governments continue to use blanket guarantees in crisis after crisis, despite the universal understanding that they entail high contingent costs and create moral hazard problems, is easy to explain. Governments use blanket guarantees to stabilize sizable systemic financial crises in the absence of the institutional and political will or the fiscal headroom needed to address the problems head-on. The findings are intuitively plausible if we revert back to the morphine example.

In conclusion, the academic literature offers robust statistical evidence that blanket guarantees increase the fiscal costs, prolong the duration of a crisis, and increase the GDP loss. Policy makers are advised to use blanket government guarantees sparingly in mapping the next phases of the resolution program. Congress is advised to hold the administration accountable for the use of guarantees by ensuring that the guarantees are priced properly and their costs are disclosed explicitly.