Russ Roberts is working on a paper suggesting that the fragility of our financial system could be the result of past bailouts, in which unsecured creditors and counterparties of financial institutions were always made whole. In some sense, the fact that the Fed fears contagion makes them turn such liabilities into ex post nearly riskless assets for investors.

This reminded me of the history of the economics of corporate capital structure. My version of that history follows.Take a company with a given set of investment projects (you can think of my favorite example, fruit trees). Should they be financed with equity or with debt?

Before economists got involved, the thinking was that shareholders could earn a higher return with more leverage (more debt). The thinking went something like this: Suppose that the fruit trees earn a 4 percent return per year, and debt costs 3 percent per year. If you go with 50 percent debt and 50 percent equity, then the equity holders earn (approximately) 5 percent per year. But if you go with 90 percent debt and 10 percent equity, then that leaves the equity holders with (approximately) 13 percent returns.*

This thinking was challenged by Merton Miller and Franco Modigliani. They said that the debt-equity mix should not matter! That is because investors can offset the leverage decisions of firms. If the firm is only 50 percent leveraged, and I want 90 percent leverage, I borrow most of the money to buy the stock. If the firm is 90 percent leveraged and I want 50 percent leverage, then I buy a combination of bonds and stock.

In fact, the easiest way to think of the Modigliani-Miller theorem is to assume that a single investor owns both the debt and the equity of the firm. If a firm issues $100 in stock and $100 in debt, and I own all of both, then why would I care if the firm changes to a capital structure of $20 in equity and $180 in debt? Either way, I still own the entire firm. As Miller liked to say, whether you cut a pizza into 6 slices or 8 slices, it’s still the same pizza.

Next, we introduce tax costs and bankruptcy costs. There is a tax advantage to using debt rather than equity, so if nothing else mattered, you go for almost entirely debt financing. However, there are bankruptcy costs–there is a loss of resources when a firm goes bankrupt. To reduce the probability of bankruptcy, you want to have some equity. The optimal capital structure is one where the marginal tax cost of equity is offset by the marginal benefit of the additional reduction in the probability of bankruptcy.

Next, we introduce principal-agent problems. Maybe as an equity-holder I am not sure that management will really pay out dividends–what if they go for salary and perks instead? So I would rather be a debt-holder.

Another agency problem is that management may be risk averse. They would rather have a sure salary than take a reasonable risk on behalf of shareholders. So you do what America’s financial wizards of the 1980’s did–you encourage leveraged buyouts, hostile takeovers, and other means to put pressure on management to maximize shareholder value.

Next, suppose we introduce the moral hazard issue. Suppose that the market gradually learns that the probability of a debtholder losing money in the case of the default of a large financial firm is really low, because the government almost always comes to the rescue.

What this amounts to is a very large subsidy to issuing debt. It should shift the balance in favor of high leverage. The capital structure at large financial firms should tend toward huge amounts of debt piled on relatively little equity.

Now, there is an offset. If you are a large financial firm with the ability to issue subsidized debt, you have a profit machine. Going bankrupt would mean that you lose your machine. So you do make some effort to remain solvent, in order to keep your machine. But basically, to get the most out of the machine, you go for the most leverage.

One doesn’t have to tell this capital structure story to make the point that the moral hazard caused by past bailouts was a cause of fragility. But I happen to like doing it this way.

In any case, Roberts points to some very disturbing implications.

*Suppose $100 investment in fruit trees, with $10 equity, $90 in debt. Annual profit is $4, of which $2.70 goes to debt holders, leaving $1.30 for dividends, or a 13 percent return.