April 21, 2011 Reading Time: 5 minutes

The problem of bank runs is probably the most serious concern in monetary economics. It is clear that if all clients claim back their depositors together the bank will fail. This situation, it is argued, puts the banks in an unstable position where their clients may try to run each other to get their deposits back before other clients. This self-enforcing mechanisms provokes banks to go bankrupt, a negative shock in money supply and this becomes a problem to the economy. It is not so clear, however, that a) bank runs are always a bad outcome, b) that banks need a lender of last resort or deposit insurance because they are inherently unstable and c) that bank runs were a real problem in the economy according to the empirical evidence. Let’s see them in turn.

Are all bank runs a bad outcome?

The fact that a bank may go bankrupt if their clients call back their deposits is not necessarily a problem in the same way we don’t say it is an economic problem if an unprofitable company goes bankrupt. Just as firms that cannot cover their cost of opportunity should shut down and reallocate their resources to more useful uses, a non-solvent bank should be put out of business before it keeps accumulating losses. If a bank run is triggered by the insolvency of the bank, rather then the other way around, then the bank run is not different to the situation where creditors of a firm withdrawing their support to a non profitable business. It is when solvent banks face a bank run that puts them out of business when this becomes a problem.

However, if there is no monopolist issuer and banks have to compete in the market issuing their own and distinguishable notes, a run to a bank and a run to the whole system became two different things. If the initial situation, or assumption, was that clients want to save in bank accounts, then a run to a bank should have the effect of moving deposits from one bank to another. What the market experiences is a change in its structure of deposits, but not in its aggregate deposits. If we implicitly assume a change in the preferences of the individuals to be away from banks, then that change in preferences needs to be explained as well, if not the run to the system, rather than to an insolvent bank, remains unexplained.

Are banks inherently unstable?

Where does the argument that bank runs are a serious problem find such strong support? One of the main drivers on banking instability rests on the Diamond-Dybvig (1983) model (DD model). The DD model is a game with an unstable good equilibrium prone to be disturbed by any random effect (i.e. sunspots) and pushed to a bad equilibrium where there is a bank run. The DD bank, being so unstable, begs the question of how banks managed survived before the era of central banks as lenders of last resort. But besides this point, there are a few other shortcomings in the DD exposition [more details and references can be found in White (1999, pp. 121-133)].

  • The DD bank does not have equity, all of its assets come from liabilities (deposits). The absence of equity (i.e. investors) gives no cushion to the bank to deal with losses.
  • The DD bank is particular in its functioning, it does neither make loans nor offer checking accounts. The DD bank accepts claims against the only consumption good in the economy. Clients deposit good x against a claim for a future amount of good x.
  • The DD bank is the only bank in the model. This results that the insolvency of the DD bank also becomes the insolvency of the whole system.

It is almost as if the DD bank was built to become prone to bank runs. But the DD banks is not a fair representation of how banks operate in the real world, let alone in a free banking market. Simplifying assumptions, to be useful, should fall inside the real world; if they fall outside then they are not providing an explanation of the market process but of some other situation.

But isn’t a clearinghouse, after all, a lender of last resort? Don’t banks lend funds each other? Yes, they do, and in some sense one bank may see other banks as their “lenders of last resort,” but they are not lenders of last resort in the sense that other banks will bailout insolvent banks. Bank A will be willing to lend funds (probably at a premium) to Bank B if the latter is illiquid but solvent at the moment. To pursue the strategy to lend funds to an insolvent bank will reverse the situation and make Bank A insolvent and Bank B solvent. In other words, Bank A will be giving away it’s solvency to Bank B. Now it’s Bank A the one that will be facing a potential bank run. The clearing house, then, does not only reduce transaction costs between banks, but also becomes a reliable source of information to the rest of the market of which banks are and are not solvent by looking to who banks lend money and at what premium. The expert point of view from a Bank may be more reliable and accurate than the hunch from their clients.

What that the empirical evidence has to say about the problem of bank runs?

George Selgin (1996, pp. 193-206) finds that the “conventional theory of banking crisis appears to fail the most elementary kind of empirical test. The theory implies not only that crisis are likely in any fractional-reserve banking system, but that they are especially likely in systems lacking any public lender of last resort or government deposit insurance. Readily historical evidence, however, contradicts both claims.” (p. 196)

Selgin presents several historical cases. Canada, for example, was as shocked as the United States by the collapse in prices and income after 1922. But Canada did not face any bank failure while the U.S. suffered one of its worst banking crisis. The Canadian system was less regulated and more free of government interference than the U.S. Another example is the relative performance of the English banks with respect to the Scottish banks. Again, England, with a more regulated system, did not perform as well as the Scottish free banking. Surely enough, there where problems in the cases of free banking as well, but there are differences in the relative performances. Selgin’s conclusion on why despite the historical evidence against banking instability the position that bank runs are a serious problem is of interest (p. 198):

Faced with this evidence of the conventional theory’s failure, one cannot help wondering how it managed to become so popular in the first place. Another look at the historical incidence of banking crises suggests an explanation. […] banking crises appear to have been a US specialty, with England earning second place in the banking crisis marathon. Most of our economic theories, including the conventional theory of banking crises, come from British and especially American economists, who know much more about the economic histories of their own countries than they know about experiences elsewhere. It is no wonder, therefore, that the received theory of banking crises appears, superficially at least, to fit the historical record of the United States and England, while bearing little connection to the experiences of many other nations. Even critics of the received theory have played into the hands of its proponents by relying solely on US experience to refute conventional assumptions, when evidence from other nations would make their task much easier. On the other hand, the few writers who have actually surveyed international experience tend to focus too much on comparing the United States with the “United Kingdom” (meaning England) in drawing general conclusions from their surveys. These writers are thus led to offer the presence of an “effective” or “dependable” public lender of last resort as the most important reason for the relative infrequency of panics in certain countries during certain periods, ignoring the more numerous cases […] in which panics were avoided despite the absence of a public lender of last resort.

A broad study of the history of banking is as important to find a way to get us closer to sound money as is to analyze the economy with models that actually resembles banks in the real world.

Nicolas Cachanosky is a doctoral student in economics at Suffolk University, as well as a previous Sound Money Essay Contest winner.

Image by Mantas Ruzveltas / FreeDigitalPhotos.net.

Nicolás Cachanosky

Nicolas Cachanosky

Nicolás Cachanosky is an Assistant Professor of Economics at Metropolitan State University of Denver. With research interests in monetary economics and macroeconomics, much of his recent work has focused on incorporating aspects of financial duration into traditional business cycle models. He has published articles in scholarly journals, including the Quarterly Review of Economics and Finance, Review of Financial Economics, and Journal of Institutional Economics. He is co-editor of the journal Libertas: Segunda Época. His popular works have appeared in La Nación (Argentina), Infobae (Argentina), and Altavoz (Peru).

Cachanosky earned his M.S. and Ph.D. in Economics at Suffolk University, his M.A. in Economics and Political Sciences at Escuela Superior de Economía y Administración de Empresas, and his Licentiate in Economics at Pontificia Universidad Católica Argentina.

Get notified of new articles from Nicolás Cachanosky and AIER.
AIER - American Institute for Economic Research

250 Division Street | PO Box 1000
Great Barrington, MA 01230-1000

Contact AIER
Telephone: 1-888-528-1216 | Fax: 1-413-528-0103

Press and other media outlets contact
[email protected]

Editorial Policy

This work is licensed under a 
Creative Commons Attribution 4.0 International License,
except where copyright is otherwise reserved.

© 2021 American Institute for Economic Research
Privacy Policy

AIER is a 501(c)(3) Nonprofit
registered in the US under EIN: 04-2121305