Deposit Insurance: Savior or Subsidy?

The insurance of bank liabilities began as an American experiment in a handful of states during the early-to-mid 19^th century. The early state liability insurance systems disappeared by the 1860s, but a second wave of systems arose in the early 20^th and U.S. federal deposit insurance was enacted in 1933. Worldwide, bank liability insurance remained a unique (and controversial) policy choice of the United States until the late 1950s, but it spread rapidly throughout the world in recent decades. Today it is a nearly ubiquitous feature of banking regulation endorsed by influential cross-border institutions such as the International Monetary Fund, the World Bank, and the European Union.

In our paper, Deposit Insurance: Theories and Facts, which was recently made publicly available on SSRN, we review the economic and political theories that seek to explain bank liability insurance and juxtaposes those theories against the worldwide experience with bank liability insurance—including an analysis of the factors that led to its passage and expansion as well as an analysis of its performance.

Broadly speaking, there are two theoretical approaches to explaining the creation and expansion of deposit insurance: an economic approach grounded in potential efficiency gains from limiting bank runs (i.e., the public interest motivation) and a political approach grounded in the rising power of special interest groups that favored insurance as a means to access subsidies (i.e., the private interest motivation).

Economic theories show how liability insurance may improve the efficient management of the banking system by reducing systemic liquidity risk. Government-provided deposit insurance eliminates the incentive for depositors to run a bank in anticipation of other depositors doing so. Despite that potential advantage, economic theories also recognize there are costs of enacting liability insurance. Models of principal-agent conflicts such as Calomiris and Kahn (1991) show that demandable debt contracts allow depositors to subject banks to “market discipline” that rewards good behavior and punishes bad. Because insurance reduces the incentive for market discipline, it may increase fundamental insolvency risk as a consequence of greater conscious risk taking by bankers (i.e., moral hazard) or through an increase in the proportion of bankers who are incompetent managers (i.e., adverse selection). Therefore, whether bank liability insurance, on balance, reduces or increases risk in the banking system is an empirical question. Economic theories of liability insurance only make sense on economic grounds if the gains from liquidity risk reduction tend to exceed the moral-hazard or adverse-selection costs from reduced market discipline.

Political theory provides a separate theoretical basis for bank liability insurance. Political bargaining models identify circumstances under which the interests of particular groups within society (i.e., the beneficiaries of passing liability insurance) may succeed in securing its passage, even though liability insurance may be inefficient (Stigler 1971; Peltzman 1976; Becker 1983). Political models seek to explain why liability insurance may be chosen to favor certain groups in society even when it imposes large costs on society in the form of higher systemic risk for banks. In this context, liability insurance needs to be understood as part of an equilibrium political bargain achieved by a winning political coalition, and consequently, its function may vary across countries as the result of the differing political functions that it plays in different regimes and contexts.

Our review considers empirical evidence about which factors have been shown to be instrumental in creating bank liability insurance, as well as evidence about the consequences of passing insurance—that is, whether insurance has improved (or worsened) the stability of banking systems. We find that political theories are much more consistent with both sets of evidence.

First, the historical push for liability insurance in the United States came from a coalition of small rural bankers and landowning farmers. The recent global wave of legislation creating and expanding insurance also can be traced to exogenous political influences. For example, deposit insurance is more common in countries with a more contestable political system (proxied, among other ways, by polity score) and with larger and more under-capitalized banks, as shown by Demirgüç-Kunt, Kane, and Laeven (2008). Modern multilateral institutions such as the IMF, the World Bank, and the EU also have played a large role in encouraging adoption of deposit insurance or expansion of its generosity.

Second, the expansion of deposit insurance is associated with reductions in banking system stability. For the early 20^th century U.S. deposit insurance systems, Calomiris and Jaremski (2016) show that insured banks were able to attract deposits away from uninsured banks that were subject to market discipline. Insured banks were able to do so despite their increasing default risk profile. Similarly, empirical studies have uniformly found that modern deposit insurance (1) encourages greater risk-taking, (2) reduces market discipline, and (3) negatively affects the growth of the financial system. Therefore, although insurance is justified economically as a means of limiting liquidity risk, its adverse effect on fundamental risk taking by banks typically dominates the reduction in liquidity risk, resulting in greater overall banking instability.

The political theories of liability insurance point to a major political advantage: it provides an effective means for a government to supply hard-to-trace subsidies to particular classes of bank borrowers. In autocracies, insurance generally has been used to favor influential borrowers, which typically include industrial firms that participate in “crony” networks. In democracies, insurance has been especially useful in favoring mortgage borrowers with political clout. For example, Calomiris and Haber (2014) show that protected U.S. banks faced precisely this tradeoff in the 1990s and 2000s as the combination of liability insurance and other legislation worked together to encourage massive amounts of mortgage lending by banks and other protected intermediaries.

There is also substantial evidence that prudential regulation and supervision are subject to politicization that can undermine their ability to rein in risk taking by protected banks. Regulatory failure is often a predictable consequence of the political bargains that give rise both to safety nets and prudential regulations. Barth, Caprio and Levine (2006) find that prudential requirements have no identifiable effects on systemic risk, and the extent of ineffective prudential rules is greater in more corrupt countries. These and other results suggest that complex regulation is motivated more by the desire to create opportunities for bribery than by its effectiveness in limiting excessive risk taking. Brown and Dinc (2005) find that the regulatory recognition of bank losses and government interventions to close insolvent banks are unlikely to occur in an election year so as to avoid declines in credit availability.

We also analyze the relative desirability of alternative government policies that protect bank liabilities under some circumstances, such as times of high systemic risk (which we label “limited and conditional protection”). We contrast limited and conditional protection with unconditional liability insurance and with unconditional laissez faire policy. Acharya and Thakor (2016) show that, as a matter of theory, limited and conditional protection generally is superior to either of the alternatives. However, like the failure of effective prudential regulation, the failure to adopt limited and conditional protection—despite its historically proven record of success—has a simple political explanation: liability insurance is designed to create subsidies, not to limit systemic risk. As Yogi Berra might have said, if contemporary liability insurance systems had been designed to limit systemic risk, they wouldn’t have been.

The full paper is available for download here.