Having an EDGE on the Bailout/Too Big to Fail Problem
I recently wrote an op ed about the “Estimated Debt Guarantee Expenses” or “EDGE” database that I created, which provides quarterly estimates since 1971 of what it might have cost to guarantee debt for all available publicly traded corporations over the next year. I’m not suggesting policy-makers should guarantee that debt, but having estimates of the potential dollar cost of bailouts handy might be of service to policymakers during periods of greater financial distress. In addition to corporation level analysis, the accompanying working paper summarizes my findings in the aggregate and across industries.
For example, the most severe periods of distress, ranked in order of the maximum quarterly aggregate estimate in Q1 2010 US dollars include: 1) Q1 2009 during the 2007-2009 financial crisis at $983 billion, 2) Q1 2020 at the beginning of the pandemic at almost $285 billion and 3) Q4 1998 after the Russian default at over $190 billion. Another key finding was that banking has the lowest industry average asset volatility but the highest industry average leverage, by which I mean reliance on debt instead of equity funding; that’s not conventional wisdom, but it’s what you might expect.
The database provides a potential bailout/Too-Big-to-Fail tracker in that it provides ex ante estimates of the costs of restoring solvency for all available publicly traded corporations, including banks and non-bank corporations. To understand what restoring solvency means intuitively, start with the figure below, which depicts a healthy bank to the left and non-bank corporation to the right:
The stylized balance sheets, as depicted in the figure, imply that the (market value of) assets equal (face value of) liabilities plus (market value of) equity. Now suppose, the corporation gets into trouble, such that the (market value of) assets lies below the (face value of) liabilities, as the equity investors gets wiped out. The EDGE database provides estimates of the cost of restoring solvency for such corporations, as depicted in the figure below:
For banks, you can compare the ex ante EDGE estimates with the ex post estimates that the Federal Deposit Insurance Corporation reports on its Bank Failures and Assistance Data webpage. For non-bank corporations, such data’s often not available, as distressed non-bank corporations tend to go through bankruptcy proceedings; however, if the government provides debt guarantees to non-banks, you could compare the costs with the ex ante EDGE estimates. With that intuition in mind, I’ll discuss the deeper motivation for the project that I did not touch on in the op ed.
Why Certain Corporations Get Bailed Out
In anticipation of my starting this project, during the pandemic in late 2020, I wrote a blogpost about why we see only certain corporations getting bailed out. The most common explanation holds that bailouts reflect problems arising from incentives, such as moral hazard. However, there may be other reasons, too, which I believe are important to keep in mind.
First, likely beneficiaries tend to be larger corporations that provide goods or services to the government. For example, during World War II, auto manufacturers produced weapons, and during the pandemic Ford made respirators. Airlines have long helped move troops during war and peace time. Large primary dealer banks provide valuable services as counterparties to the Federal Reserve and Treasury Department. When such corporations face distress, Congress may offer debt guarantees. I’m still investigating whether corporations in other industries have access to bailouts for similar reasons, and feel free to reach out to me or investigate for yourselves.
Second, while writing the working paper, a former student suggested I should also examine the employment angle, too, which turned out to be a great suggestion. Corporations with the highest estimated costs of guaranteeing debt during periods of greater financial distress on average also employed a much larger number of people than all other corporations. In ordinary periods, the average number of employees for corporations with the highest estimated guarantees and all others were fairly similar. Maybe this employment angle doesn’t apply to banks, but you can see this in other industries. For example, policymakers cited concerns about a sudden spike in unemployment as reasons for the government assistance for Chrysler when it faced distress both in 1979 and 2008.
What You Might Do About It
What could you do to limit any impact of debt guarantees on the government budget and taxpayer? In the 2020 blogpost and the working paper, I touch on how leverage restrictions – limits on how much of a corporation’s funding comes from debt – could be used to limit the size of, if not eliminate, bailouts. This idea came from John Cochrane’s suggestion to have airline leverage restrictions through capital requirements, as we have for banks. With less leverage, such corporations would be less likely to default, and the taxpayer would be less likely to be on the hook for direct losses, as well as indirect ones, if the corporation passes on any costs on customers. Holding all else constant, the less debt a corporation funds with, the farther away it will be from default (and as I also wrote in 2020, it is true that for some industries it makes more sense to fund with more debt than others, especially for tangible capital formation). To the extent that corporations that get bailouts tend to borrow lots to fund their operations, rather than because they simply have risky assets, leverage restrictions could limit bailouts.
Future Steps
The EDGE database has also inspired me to write about how you can create synthetic contingent liability. Currently, common equity shareholders, including executives who receive shares as part of deferred compensation face single liability in that they can lose their entire investment if a corporation goes bankrupt and that’s it. But bank regulators in the past have relied on contingent liability, such as double, triple or unlimited liability, where shareholders could lose not only their entire investment, but also be made to pay for at least some of the creditor losses. The extra penalties aimed to curb executive risk-taking. In my next post, I’ll explain how you might be able to do that synthetically, in the sense that you won’t have to change existing shareholder liability, using existing common equity shares and input from the EDGE database, and also how to do this for executive options compensation. Stay tuned.