Fun With The Fed's "Stress" Test

Authored by Nicholas Colas via,

Like many of you, I spent part of the late afternoon reading through the results of the Federal Reserve’s annual stress test of the US banking system. At first blush, everything looks good. Everyone passed. Now investors can look forward to next week’s Comprehensive Capital Analysis and Review, when we hear what sorts of buybacks and dividends these institutions can pay in the year ahead.

It’s difficult not to gasp a little when you look at the “Severely adverse” scenario the Fed used in this year’s review:

  • A 7 quarter recession that leaves GDP 7.5% lower than prior highs

  • Unemployment rises to 10% over the same period

  • Inflation drops to 1% and short term Treasuries yield close to zero

  • At the same time, investors shun long term Treasuries and yields there remain unchanged

  • Investment grade and mortgage yield spreads blow out to 5.75% and 3.5% respectively

  • US stocks drop by 65% in early 2019, and the VIX goes over 60

  • House prices drop 30% and commercial real estate by 40%

I don’t know about you, but to me that looks like the recipe for revolution more than a regulatory what-if scenario, but let’s go with it.

Two points:

#1) If the US banking system is really as robust as the Fed’s analysis indicates, doesn’t that merit a higher multiple on domestic stocks than historical norms? A bullet proof banking system should be worth a systematic premium over a shaky one, after all. America’s banks should be able to provide capital in even a severe downturn, make orderly markets if they have capital markets desks, and generally act as they would at any other point in the economic cycle. That would be a welcome change.

Now, there is a headwind to this positive case: the Financial sector is 14% of the S&P 500, and the regulatory process that makes them systematically safe also reduces their structural return on capital. So we’re unlikely to see much P/E expansion in the group, but every other sector - consumer or industrial - should get an uptick. A robust financial system should make the next recession easier than most previous ones by limiting any shock to the supply of credit, after all. Trough earnings will be higher than expected, and valuations should expand.

This is ultimately a cyclical argument, which means we’ll have to see it work in the next downturn before investors are willing to pay higher multiples. But for those investors who actively consider 5-10 year equity investment horizons, this is a bullish case we’ve not read anywhere else.

#2) If the Fed’s stress tests have a fatal flaw, it is that they assume a “V” bottom from the near-death levels they outline in their requirements. This doesn’t get much attention in the Fed’s document (link at the end of this note) but as near as we can tell, this is what they assume in the “Severely adverse” scenario for a recovery:

  • Unemployment improves in the 6 quarters after the peak by almost 2 points

  • GDP growth turns positive 4 quarters after its trough

  • US equity markets recover almost all their losses in the 2 years after the bottom

  • Corporate bond spreads recover to near 2014 levels 7 quarters after they peak

  • The VIX breaks below its long run average (20) just 2-3 quarters after the Dow trades at 10,000

Some of those – GDP and unemployment – we can see; the Dow’s recovery, bond spreads tightening and the VIX seem fanciful at best. We understand the Fed’s intent is to model another 2007-2008 scenario, but perhaps the next downturn will be different. For example:

  • The next recession, even a garden-variety contraction, will come just as artificial intelligence, robotics, drones, and other technologies hit their strides. All those advances have been invisible in the unemployment data because the US economy is growing.

    Take it from an old cyclicals analyst: companies do their big restructurings in economic downturns, not expansions. The next period of contraction is when capital will substitute for labor at an accelerating rate.
  • At the start of 2007, US public debt-to-GDP was 63%; it is now 106% (yes, we include Social Security). There will potentially be less room for fiscal stimulus in the next downturn, especially since every other developed economy has the same problem (which may be why the Fed’s assumptions include no change to long term yields).

The upshot here is that the next recession – especially a severe one – may not be anywhere near as easy on the way out as the Fed’s scenarios portray. 

So yes, the system is safe at the bottom of a 2007-style crack. But what if the recovery is much slower?

Since the Fed assumes a “V”, what happens if it is an “L”. We don’t know; they don’t include that possibility.

And what does it mean that the most systemically important banks in the world are in a bear market?