Bank Lending, Capital, Booms, and Busts
The volume of bank loans outstanding grows and shrinks with the business cycle. Growth slows just before a recession, and total volume shrinks after the recovery begins, typically bottoming out a few months later. In economist’s jargon, the amount of bank loans outstanding is a lagging indicator. Recent weekly data indicate that loan growth has already reached negative territory, meaning that total lending is now contracting. Over the past 30 years, this occurrence has indicated that the turn-around point in the business cycle has already been reached, but as they say in forecasting, past results are no guarantee of future success. While the procyclical pattern is evident with a longer range of annual data, a look at the 1930s shows that loans can contract for years before the bottom arrives.
Certainly many factors contribute to the cyclical pattern of loan growth. Both supply and demand contribute: investments look riskier to banks in a recession, and they tighten standards. Firms see fewer prospects for growth and they borrow less. (For more on this, see this Economic Trends article.) The current crisis has brought a lot of attention to the sometimes obscure role that bank capital plays in lending levels. One concern is that bank capital, which is intended to serve as a buffer against losses, tends instead to accentuate booms and busts. The theory is that capital requirements allow banks to increase their leverage in good times because loans look safe and risk measures decrease. But when times get worse and risk measures increase, capital requirements increase and make loans more expensive. So rather than lean against the wind, policy runs with the prevailings.
Bank capital, though, is a complex subject, and there are a variety of capital measures and related ratios, all measuring slightly different things. The simplest is common equity to total assets. Also popular is leverage, which is just the inverse ratio, that is, total assets to common equity. A somewhat broader definition of capital adds in some forms of preferred stock, resulting in Tier 1 capital. Adding in other liabilities, such as subordinated debt and the loan-loss reserve, defines Tier 2 capital. Assets might be simply summed up, or they might be weighted by a risk factor (gold bullion gets a zero risk factor, most commercial and consumer loans get a 100 percent risk weight).
For the period over which we have good data (slightly more than a decade), the ratio of Tier 1 capital to assets does not have a strong cyclical component, though it drops before the current recession and rises later on.
The Tier 1 risk-based capital ratio shows stronger cyclicality, repeating that pattern in both recessions.
Looking at leverage, for which we have a longer data series, however, this pattern has been hard to detect. Any cyclical changes are dominated by longer-run shifts. This does not necessarily mean that capital is unimportant for explaining bank lending behavior, just that the effects may be more subtle.
Over the very long run, bank capital has been decreasing, with major drops following the creation of national banks and the introduction of deposit insurance. Movement since the 1950s has been smaller, with perhaps a slight upward trend in the 1980s.