The Quality of Bank Lending to Businesses

I've written a fair bit in recent weeks about the US banking sector's exposure to losses stemming from bad mortgages. My conclusion: banks are in pretty good shape to handle a good number of mortgage defaults, but if the number of defaults rises significantly beyond current levels, or if the economy starts sliding toward recession in the next year or two, then the damage to the financial health of banks could be very serious.

A related but harder-to-answer question is to what degree banks' loans to businesses could also be susceptible to rising default rates. Calculated Risk took a look at this issue a week or two ago, specifically focusing on bank lending to businesses for real estate development, and today David Weidner of Marketwatch addresses the question of how safe business loans are more generally:

A mortgage crisis of their own
Commentary: Will companies go way of subprime borrowers?

NEW YORK (MarketWatch) -- The subprime credit crisis has been in swing for a few weeks now. As part of the cycle, we've all heard the stories about people drowning in debt. Maybe you know the fellow with the adjustable-rate mortgage who's slapped toward bankruptcy with every reset.

...Unfortunately, Joe America's not alone in experimenting with the dangerous combination of cheap money and supersized amounts. Corporate America may be next.

Much like the consumer-credit cycle, low-interest loans and easy-to-come-by debt fueled economic expansion and a mergers and acquisitions boom in the U.S. market during the last few years. Companies have increasingly been leveraging their balance sheets, putting them at risk of default in the next 12 to 18 months, according to the analysts at Standard & Poor's.

A corporate credit crunch isn't directly linked to the mortgage market, but in the credit cycle the markets tend to move in unison or with one lagging the other. A mortgage-market meltdown would also make it tougher on Main Street and rash of defaults could cause a chain reaction: making it tougher for private equity shops to borrow and buy, slowing mergers and IPOs and shutting down much of Wall Street.

To put a little meat (i.e. data) on these bones, I took a look at some of the statistics that the Fed collects on bank lending to businesses. Four times each year the Fed conducts a survey of bank lending in which they ask banks to categorize the new loans that they've made in terms of riskiness, and to report what fraction of their new loans were backed by collateral (among other things).

The following chart shows the pattern in these measures over the past decade. The light green line shows the percent of new bank loans that were classified as being of "minimal" or "low" riskiness (as opposed to loans that have "moderate", "acceptable", or higher levels of risk). The blue line shows the percent of the least risky loans that were backed up by collateral. The red line shows the percent of higher-risk loans (i.e. loans of "moderate" riskiness, which turns out to be the largest single category of new loans made by banks) that were backed up by collateral.


Note: series show 4-quarter moving averages. Source: Federal Reserve Board Survey of Terms of Business Lending.

By these measures, the quality of bank lending to businesses is not wholly reassuring. Of particular concern is the decrease since 2004 in the collateral that banks are securing for the lending that they do to their riskier clients. In fact, for both low- and high-quality loans, banks are at or near record-low rates of collateralization for the loans they are making.

Now, to some degree this may simply be a function of the types of loans that banks are making, and not an indication that they've been making riskier loans. Loans for businesses to develop real estate are easier for businesses to collateralize than loans for businesses to secure working capital, for example (because the business can promise the real estate itself as collateral). So perhaps banks have simply been making more of the latter type of loan in recent years.

But that explanation seems at odds with the fact that the series began declining over two years ago, when the real estate boom was still in full swing. It's also worth noting that in 2006 real estate development was still considerably above 1999-2000 levels - yet rates of collateralization on bank loans were significantly lower.

Furthermore, the fact that banks have been unable to categorize more of their business loans as being relatively low-risk, despite record corporate profits, suggests the possibility that banks have indeed been increasing the riskiness of the portfolio of loans that they hold, especially since 2004. By comparison, at this phase in the last business cycle (during the expansion of the 1990s) banks were able to classify roughly one-third of their loans as minimal or low-risk loans. For the past few years, on the other hand, the ratio has only been in the range of 20-25%.

Finally, this suggestive evidence that the quality of bank lending to businesses has gone down in recent years is consistent with economic theory. There are good theoretical reasons to think that banks make more risky loans when the economy is doing well. See for example this Richmond Fed paper by economist John Weinbert, "Cycles in Lending Standards?"

My conclusion is that banks' loan portfolios have indeed declined in quality during this business cycle, just as they have in all previous business cycles. So while banks seem to have strong reserves at first glance, they are also exposed to a lot of risk, both from the housing market downturn and from their business lending. Let's hope that any coming economic slowdown remains very mild, because if not, it's easy to see how banks could end up suffering a lot more than most people are now predicting.