The market experienced a complete change of landscape in March from a very bullish outlook of no recession to a severe crisis within two to three weeks. While a particular bank run is often unpredictable, the backdrop that causes it is not. Two weeks ago, I argued that this time was not a specific bank problem but a systematic one. Now the stories have become clear that the few failed banks were quite sound in asset holdings (such as sovereign bonds) and liquidity. By contrast, the earnings potential of these failed banks worsened significantly.
While banks always do KYC—know your clients, nowadays people are talking about KYB—know your banks. Two key market indicators tell of a bank’s potential danger: Its stock price and CDS rate (credit default swaps rate). The latter no doubt reflects the likelihood of bankruptcy, but it shoots up very late yet sharply until the problem is exposed. For the stock price, one has to pay attention to the longer-run trend rather than short-term movements. Stock price, in principle, reflects earnings potential, which is convenient for those who do not read accounting reports.
A more straightforward way to judge the potential systematic risks among countries is by examining the macro version of earnings potential. That is the so-called famous “g–r” principle. While bank earnings are revenues minus costs, a country’s corresponding inflow and outflow are naturally GDP growth and interest rate with both in real terms. This does not tell which banks will be in trouble but which country. Previous analyses tended to take this as an indicator of whether a nation will fall into a debt crisis.
The accompanying chart with the latest data shows U.S. and Switzerland have the lowest “g–r,” and it is precisely these two places having bank failures. A glimpse of the curves suggests a potential danger could be when “g–r” gets below about two percent. This is necessary but not sufficient, meaning that a banking crisis happens when below about two percent is satisfied, but the converse is not true.
Perhaps another cause of a crisis is when “g–r” drops too quickly; that is, the slope may matter. If this were the case, then the crisis might not be confined to U.S. and Switzerland but also Eurozone and UK, which have much higher (extra four percent) “g–r.” Such a hypothesis cannot be concluded from the existing dataset as (fortunately) clusters of banking crises are rare in these four places. However, it is not hard to imagine “g” is much more volatile than “r” and is highly commoving across countries. The rest will follow whenever one or two of them enter into a recession.
That says, the current relatively high “g–r” in Eurozone and UK is not guaranteed at all, where the situation can suddenly change. The logic here seems to flow more from economy to crisis than the other way around. This is, in fact, the case, as a crisis is often a canary.
