Economic explanations for why banking crises occur focus on two influences: macroeconomic shocks that cause banks’ loans to become riskier, and banks’ tendency to fund themselves with large proportions of demandable debt or other very short-term debts. An adverse macroeconomic shock that raises the riskiness of bank debt leads risk-intolerant depositors to demand repayment. As banks reduce their lending to fund withdrawals, economic prospects of borrowers worsen and bank losses rise further, producing a substantial further increase in borrower and bank insolvency risk.
This simple explanation has a lot going for it; there has almost never been a major banking crisis that did not begin with a shock to the value of bank loans, and there is substantial evidence that banks have raised funds primarily in the form of risk-intolerant, short-term debt since commercial banking began about 2,600 years ago in ancient Athens. But there is a problem with this story: It does not do a good job of explaining why some places and some eras witness a large number of banking crises while others do not.
When you look at variation in the frequency and severity of banking crises worldwide over time, it’s striking how much the frequency has increased during the past four decades
Take an obvious point of comparison: the history of banking crises in Canada and the United States since 1837. Canada has had no banking crises during this 180-year period, while the United States has averaged about one every 12 years. Canada has a more volatile economy (hence bigger shocks), and has a higher average ratio of bank loans to gross domestic product (GDP). Banks in Canada, like those in the U.S., finance themselves primarily from short-term debt. A simple explanation about shocks and bank funding structure would suggest that Canada should have experienced more shocks than the U.S., but the opposite is the case.
A Global Pandemic
When you look at variation in the frequency and severity of banking crises worldwide over time, it’s striking how much the frequency has increased during the past four decades. Indeed, we are currently living thorough a global pandemic of banking crises unlike any the world has seen before. Compared to the period 1874 through 1913 – a time of significant macroeconomic volatility and heavy reliance on bank credit – the period of the same length from 1970 through 2009 saw 10 times as many banking crises (defined as episodes of significant bank failure and/or times of sudden system-wide withdrawal of deposits), and banking crises were five times as severe (measuring severity as the negative net worth of failed banks relative to GDP).
Could it be that variation in the way banks are regulated can explain differences across countries and over time in the degree of banking instability? That is certainly true. Regulation defines how banks are created, what they’re permitted to do, what they’re required to do, and how much protection they expect to receive from taxpayers, and these have important consequences for risk-taking. The consequences of regulatory differences across countries and over time are important drivers of the recent pandemic of banking instability. The two most important regulatory contributors to risk are the recent increase in the extent to which taxpayers around the world protect banks from loss, and the encouragement of banks’ involvement in real estate lending. These two policy choices could be called the two “800-pound gorillas in the room” of banking system risk.
Government structure matters in determining which political coalitions win in the struggle over control of the banking system.
In theory, more protection of banks could either increase or decrease systemic risk. On the one hand, protection insulates banks from withdrawal risk by supplying them with funds as needed, and this reduces the potential for bank illiquidity to magnify the shocks that produce banking crises. On the other hand, protecting banks may lead them to undertake more risk, thereby making the financial system more vulnerable to shocks. A large empirical literature has shown that the second influence outweighs the first: Overall, protection substantially raises financial system risk.
It’s accepted that the encouragement of real estate lending by banks unambiguously increases systemic risk. Real estate lending is not a natural niche for commercial banks. Real estate prices tend to be correlated with each other and move closely with the business cycle, making it hard to diversify a loan portfolio dominated by real estate loans. Furthermore, real estate assets have unique properties that make them illiquid (not easy to sell quickly for their fair value). Loans backed by real estate may, therefore, be particularly challenging to liquidate during a crisis as banks that face withdrawal pressures seek to reduce their lending. For these reasons, it isn’t natural for banks funded by short-term debts to specialize in real estate lending, and for the most part, historically, banking systems avoided doing so.
However, if a government wishes to subsidize real estate lending, it often finds that the most convenient way to do so is through banking regulation, despite the adverse consequences for systemic risk. The reason is simple: banks are among the most important entities that are chartered and controlled closely by governments.
These observations about the “two gorillas” of systemic risk raise deeper questions: Why are recent regulatory choices so different from those of the past? Why are so many societies now willing to tolerate banking regulations that predictably produce instability by protecting banks or encouraging banks to make real estate loans? The answers to those questions require political insight. My 2014 book with Stephen Haber, Fragile by Design: The Political Origins of Banking Crises and Scarce Credit, explores why and how political coalitions that produce banking regulations that cause instability have been formed in some times and places but not in others. In particular, we ask why coalitions in control of regulatory outcomes in the United States have been so much more willing to tolerate instability as part of their winning political bargains, and why the two gorillas of taxpayer protection of banks and subsidization of real estate lending risks have become more important recently than they were in the past.
When asking questions about the political influence on banking regulation, it is helpful to begin by recognizing that banking coalitions aren’t just about the interests of bankers. This makes the political economy of banking regulation a bit different from that of other industries. Political economy theories of regulation often focus on the “capture” of regulation by firms in the regulated industry. Because consumers of particular products tend not to organize as a group politically but producers within a particular industry can and do organize in pursuit of their collective interests, producers often are able to influence the regulation of their industry to favor their interests at the expense of consumers.
Not only has the subsidization of real estate lending risk helped to promote a wave of severe banking crises around the world, it has also reduced economic growth.
But this logic does not apply to banks. Banks’ customers – especially borrowers – often organize themselves politically. In the United States, the agrarian populist movement (which played a major role in financial regulation from the beginning of the republic until the latter half of the 20th century) largely reflected the interests of agricultural borrowers. Those borrowers successfully promoted particular bank regulatory policies – most importantly, “unit banking” laws that limited the establishment of branching banks – that they believed improved their access to local agriculture-related credit. Beginning in the 20th century, urban mortgage borrowers became an organized interest group (assisted by various real estate industry special interests) and used their political influence to push banking regulation (and other financial regulatory policies) to subsidize the risks of mortgage lending.
Why didn’t Canadian borrowers succeed in forming winning political coalitions to control the structure of Canada’s banking system, as their neighbors to the south had done? It was not for lack of trying. For example, in the 19th century, agrarian interests tried to abolish the branch banking system of Canada and replace it with a U.S.-style unit banking system. They also attempted to pass financial assistance packages for rural borrowers similar to those in the U.S., but those efforts failed.
As Stephen Haber and I explain in chapter 9 of Fragile by Design, the structure of the Canadian government (the centralization of regulatory rule-making and the powerful role of its unelected Senate) made it impossible for rural borrowers to succeed. Government structure matters in determining which political coalitions win in the struggle over control of the banking system.
A Lobbying Connection
Interestingly, U.S. history shows a connection between regulatory lobbying to protect banks and the subsidization of particular classes of loans. Rural borrowers were a crucial force pushing for deposit insurance to be enacted in individual states, and later at the national level in 1933, as a means of expanding their access to credit. Protection of bank deposits created value for protected banks, which regulation forced them to share with bank borrowers in the form of more favorable credit terms for rural credit.
The same sorts of rent-sharing arrangements from protection occurred in U.S. mortgage market and banking regulation during the 1990s and 2000s, and such protection was a major means of government subsidization of mortgage lending risk. Deposit insurance, implicit protection of other liabilities of too-big-to-fail banks, and implicit guarantees of the debts of the government-sponsored enterprises (GSEs) Fannie Mae and Freddie Mac created protection rents that banks and GSEs were forced to share in the form of mortgage risk subsidization (as described in detail in chapters 7 and 8 of Fragile by Design).
Lax prudential regulation of banks and GSEs was part of the creation of rent from protection. Lax regulation permitted increased risk of default, which increased the size of the transfer from the government to the coalition receiving it.
Recent U.S. banking history, therefore, has been a prominent example of the marriage between the two gorillas of systemic risk: real estate lending risk subsidization and government protection of banks. Unfortunately, this unholy union has become a global pattern over the past four decades. Indeed, the mutually reinforcing systemic risks produced by marriages of these two gorillas around the world can explain the current pandemic of crises that have been so destructive to banking systems.
Deposit insurance by governments and other ad hoc protections of bank liabilities by governments were relatively rare prior to 1980. As Figure 1 shows, however, protection of bank debts expanded dramatically in recent years. A 2008 study by Demirgüç-Kunt, Kane, and Laeven shows that both domestic and international political pressures played a role in explaining the recent wave of expanded protection of banks. Internationally, the International Monetary Fund, the World Bank, and the European Union all encouraged the expansion of safety nets for banks. In part, that advocacy likely reflected the desire to solve short-term banking system problems by using insurance to give banks breathing room. However, the long-term consequences for increasing system risk have been extremely debilitating.
A 2015 study by Jordà, Schularick, and Taylor shows that the proportion of real estate lending in bank portfolios in the countries in their sample has grown dramatically since 1980, from about 40% of loans to nearly 60%. In a new study, Sophia Chen and I analyze the extent to which the expansion of government protection of banks and the creation of protection rents for banks that this implies help to explain the increase in bank involvement in real estate lending. (Calomiris, Charles W., and Sophia Chen. “How Deposit Insurance Promotes Real Estate Lending. Columbia Business School Working Paper, June 2017) We find strong evidence that exogenous increases in bank protection will lead to resulting increases in mortgage lending.
To identify causality properly, it is necessary to identify exogenous changes in protection – that is, changes that are not themselves a consequence of local banks’ condition or related political responses to their condition. We use Demirgüç-Kunt, Kane, and Laeven evidence that international political pressures – which cannot plausibly be explained by the developments in the countries adopting or expanding protection – explain much of the adoption and expansion of protection. Formally speaking, our econometric analysis occurs in two stages: First, we show that external political factors explain the expansion of protection; second, we show that changes in protection that can be traced to those factors also result in significant expansion of bank mortgage lending. As in the case of the U.S. as described above, protection of banks creates rents that political coalitions direct toward risky real estate lending.
Not only has the subsidization of real estate lending risk helped to promote a wave of severe banking crises around the world, it has also reduced economic growth. A 2015 study by Cournède and Denk shows that the growth in bank consumer lending (which is mainly mortgage lending) has been associated with an economic growth-reducing expansion of bank lending.
Together, these various studies show that the current pandemic of banking crises, with their unprecedented costs to taxpayers and adverse consequences for economic stability and growth, should be viewed from a broad political economic, rather than a narrow economic, perspective. Banking crises are not an inevitable consequence of business cycles or the structure of banks. Countries that make political choices that eschew government protection of banks and government subsidization of real estate lending risk can avoid the costly systemic risk increases that plague the world’s banking system today. Of course, that doesn’t mean that it’s easy to reduce systemic risk that arises from the two gorillas; it is easier to identify politically driven problems that to organize political coalitions to correct them.
About The Author:
Charles W. Calomiris is Henry Kaufman Professor of Financial Institutions at Columbia Business School, Director of the Business School’s Program for Financial Studies and its Initiative on Finance and Growth in Emerging Markets, and a professor at Columbia’s School of International and Public Affairs. His research spans the areas of banking, corporate finance, financial history, and monetary economics.
He is a Distinguished Visiting Fellow at the Hoover Institution, a Fellow at the Manhattan Institute, a member of the Shadow Open Market Committee and the Financial Economists Roundtable, and a Research Associate of the National Bureau of Economic Research. He received a B.A. in economics from Yale University, magna cum laude, and a Ph.D. in economics from Stanford University.