Thicker capital buffers do not prevent banking crises

I know I complained about the sorry state of academic research on banking in my previous post, but not all research makes me despair. In fact, I have long admired a number of ‘mainstream’ academic researchers, such as Borio, as well as Jordà, Schularick and Taylor. The latter’s research is top-notch and what they built what is surely one the best available historical databases of banking. Thanks to their data collection, they provide academics with resources that go beyond the narrow scope of US banking. Their dataset is available online.

Last September, they published a paper titled Macrofinancial History and the New Business Cycle Facts, which is quite interesting, although not as much as their ground-breaking previous papers. Nevertheless, it is based on excellent datamining and I strongly encourage you to take a look. One of the interesting charts they come up with is the following real house price index aggregated from data in 14 different countries. As we can see, real house prices have remained relatively stable (at least within a range highlighted the black lines I added below) until the 1970s. However they started booming from the 1980s, when Basel artificially lowered real estate lending capital requirements relative to that of other lending types.

But it is their most recent paper that particularly drew my attention. Published just a couple of weeks ago and highlighting Bagehot’s quote at the top of this post, Bank Capital Redux: Solvency, Liquidity, and Crisis argues that, contra the current regulatory logic, higher capital ratios do not prevent financial crises. In their words (my emphasis):

A high capital ratio is a direct measure of a well-funded loss-absorbing buffer. However, more bank capital could reflect more risk-taking on the asset side of the balance sheet. Indeed, we find in fact that there is no statistical evidence of a relationship between higher capital ratios and lower risk of systemic financial crisis. If anything, higher capital is associated with higher risk of financial crisis. Such a finding is consistent with a reverse causality mechanism: the more risks the banking sector takes, the more markets and regulators are going to demand banks to hold higher buffers.

As usual, their data collection is remarkable. This time, they collected Tier 1 capital-equivalent* numbers, as well as other balance sheet items, across 17 countries since the 19^th century. Here is the aggregate capital ratio over the period:

Unlike what most people – and economists – believe, they also demonstrate that capital ratios were on the rise in a number of countries in the years preceding the financial crisis:

Their finding is a blow to mainstream regulatory logic: capital ratios are useless at preventing crises and may well be a sign of higher risk-taking.

However, some of their findings do provide some justification for capital regulations. They find that

a more highly levered financial sector at the start of a financial-crisis recession is associated with slower subsequent output growth and a significantly weaker cyclical recovery. Depending on whether bank capital is above or below its historical average, the difference in social output costs are economically sizable.

While the fact that better capitalised banks are more able to lend during the recovery phase of a crisis sounds logical to me, I believe this result requires more in-depth analysis: it is likely that regulators in many countries forced banks to recapitalise after past crises or, as it was the case in the US in the post-WW2 era, that banks were also required to comply with a certain type of leverage ratio. This would have slowed their lending growth and impacted the recovery as they rebuilt their capital base to remain in compliance.

It may also be that, as they highlight in some of their previous research, banks suffered more from real estate lending, which was initially seen as safer and requiring thinner capital buffers, but which ended up damaging their capital position further and for longer periods of time once prices collapsed (relative to financial crises triggered by stock market crashes for instance). Whatever the underlying reason, this finding requires more scrutiny and granular analysis.

They also find

some evidence that higher levels and faster growth of the loan-to-deposit ratio are associated with a higher probability of crisis. The same applies to non-core liabilities: a greater reliance on wholesale funding is also a significant predictor of financial distress. That said, the predictive power of these two alternative funding measures relative to that of credit growth is relatively small.

See below the 17 countries aggregate loans/deposit ratio:

This is interesting, as we see that, unlike capital ratios, loans/deposit ratios were quite stable in recent decades relative to long-run average (in particular if we exclude the Great Depression period and its long recovery), around the 100% mark.

However, I will have to disagree with their finding that more ‘wholesale funding’ is a driver behind financial crises, even if there is some truth to it; although I happen to disagree strictly based on the evidence they provide. They base their reasoning on the wrong assumption that all non-deposit liabilities are necessarily other funding sources (see below the breakdown of liability types). This is incorrect: modern large universal banks have very large trading and derivative portfolios, which often account for 20% to 40% of the liability side of their balance sheet (although US banks under US GAAP accounting standards are allowed to net derivatives and therefore report much smaller amounts).

The key to figure out whether a bank is wholesale-funded is simply its loans/deposits ratio. A ratio above 100% indicates that a portion of loans has been funded using non-deposit liabilities. But as we’ve seen above, this ratio has never risen very high in the years preceding the financial crisis and used to be even higher in the 1870s.

Despite those minor disagreements and caveats, their research is of great quality and their dataset an invaluable tool for future analysis.

*Tier 1 capital is a regulatory capital measure introduced by the Basel rulebook