Leverage Cycles: More Schizophrenia
Click here for a printer friendly version of this article.
One of the triggers for last year’s financial crisis was the unacceptably high leverage ratios of financial institutions. Leverage measures the relationship between a bank’s total assets to the capital contributed by its owners. A higher leverage ratio can increase a bank’s profit margins. It also reduces the bank’s margin for error; a 30-to-1 leverage ratio means that a 4% decline in the market value of a bank’s assets would leave it technically insolvent. Leverage ratios also measure the amount of new loans a bank can extend for every $1 of internally generated capital. As a result, reducing leverage at financial institutions is tantamount to reducing the availability of credit for U.S. households.
In 1975, household debt (principally mortgage and consumer indebtedness) was equal to only 44.8% of GDP. By 2007, household debt had reached $13.7 trillion, or 97% of GDP. This extraordinary growth in household borrowing was facilitated by a financial system whose indebtedness grew even more rapidly: between 1975 and the 2008 financial crisis, financial sector debt grew from 15.9% of GDP to 118.3% of GDP, or $17 trillion.
In recent years, a strong positive correlation developed between leverage and the balance sheet size of key financial institutions, most notably investment banks. This means that as the value of investment banks’ assets increased, so too did the percentage of those assets that they borrowed to finance. This is counterintuitive and at odds with what is normally experienced among households. Consider a family that borrows $200,000 to buy a $250,000 house. When the value of the house increases by 10% to $275,000, the family’s home equity grows from $50,000 to $75,000 and its financial leverage (debt-to-equity ratio) declines by one third. Leverage and asset values are inversely, or negatively, related for any household or passive investor.
The opposite was true for investment banks. They responded to every one dollar increase in the value of their assets with more than one dollar of additional borrowing. Between 2003 and 2007, Lehman Brothers’ on-balance sheet assets grew from $312 billion to $691 billion (by 121%) while its gross leverage ratio (total assets divided by shareholder’s equity) increased from 23.69-to-1 to 30.72-to-1. This suggests that each dollar of new equity capital financed $40 of incremental balance sheet growth during this period. The direction and magnitude of the relationship was roughly the same for Bear Stearns, Merrill Lynch, Morgan Stanley, and Goldman Sachs from 1998 to 2007. The leverage of the housing government-sponsored enterprises (GSEs) was over 250-to-1 during this period, when measured as the ratio of their total book of business (on-balance sheet assets plus guaranteed securities) to common shareholders’ equity.
In practice, large financial institutions make decisions about balance sheet size and leverage by calculating the value-at-risk (VaR) of their trading portfolio pursuant to the guidance of Basel II. VaR estimates the losses that a bank could expect in a “worst case” scenario. The VaR helps to inform banks of the minimum amount of equity capital necessary to maintain a sufficiently low probability of insolvency. For example, if a bank’s VaR estimate suggests that it has a 1% probability of suffering a loss as great as $10 billion over the course of one month, holding $30 billion in equity capital should, theoretically, provide a cushion large enough for the bank to remain solvent for all but the rarest events.
The problem, as numerous banks discovered in 2008, was the difficulty in determining the worst case scenario. The motivation for Basel was that banks held the same minimum capital ratio irrespective of the riskiness of their assets. A bank that held nothing but U.S. Treasury notes and loans to investment grade corporations would be required to hold the same minimum amount of equity capital as a bank that lent solely to subprime mortgage borrowers. While it may be easy qualitatively to recognize that the subprime bank is the riskier of the two, it is next to impossible to determine quantitatively the difference between the two banks’ maximum probable losses. The Basel I process settled on risk weightings for assets based on credit rating but maintained a minimum capital standard as a buffer. Basel II moved further, allowing banks to set their own capital levels based on the worst case scenarios generated by their own internal models. The predictable result was more leverage in the banking system.
The Obama Administration now wishes to roll the clock back and increase the amount of capital banks must hold. But the Administration has shown no tolerance whatsoever for the contraction in household credit that would inevitably result from the corresponding decline in financial sector leverage. Until the intellectual and political instincts of the Administration are better aligned, expect a continuation of the status quo.