Bayesian Investor Blog

I’ve been noticing more discussion recently about recession risks. Here are some disorganized thoughts.

Here’s an article suggesting that the 1949 recession is a good model for what we face.

The similarities that I consider important are:

• An unusual external force (war, pandemic) recently caused atypical disruptions.
• Monetary conditions became quite inflationary during the disruptions, with some return to stability as the disruptions eased.
• Interest rates during the resulting inflation were low enough that rising rates seemed more likely than falling rates.

That article evaluates monetary policy by looking at interest rates, and says the Fed did nothing.

Interest rates are an unreliable indicator of monetary policy, so I want to look further. There’s no TIPS spread or NGDP futures for that era, so I guess I’ll settle for money supply data.

That agrees with the interest rate evidence: after an unusually inflationary spurt from 1940 to 1945, there were 3 years of flat monetary base before the 1949 recession:

 

Compare that to the recent spurt in money supply lasting 18 months, and flattening for 6 months:

Today we have the TIPS spread, which provides clear evidence that the Fed isn’t planning to tighten monetary policy much.

The CPI peaked in the summer of 1948, followed by modest deflation. The S&P 500 peaked slightly earlier, dropped 20%, and recovered by late 1949.

Note that inflation statistics in the 1940s are a bit misleading due to wage and price controls. An ideal measure of inflation (that included black market prices and costs associated with shortages) would have shown more inflation during the war, and less during 1946-8. A weaker version is likely happening now, as various obstacles to raising prices and renegotiating wages are delaying the symptoms of last year’s inflationary monetary conditions.

Housing starts peaked in late 1947, a year before the start of the recession. It’s unclear whether housing has peaked this cycle – homebuilding companies are still reporting unusually large backlogs.

There have probably been some mild housing bubbles in Phoenix, and maybe Florida, as homes couldn’t be built fast enough to handle the people who fled from blue state cities due to COVID and/or COVID prevention. Those housing markets will likely return to normal this month, as buyers return to cities due to high gas prices and reduced COVID concerns.

There’s a large, unusual discrepancy between the recent report of a Q1 GDP decline, and the ISM purchasing managers report which says that business continued growing at an above-average rate. I’m guessing that the reports are reacting quite differently to the abnormal number of sick days in January. I consider the purchasing managers report to be more valuable for investment purposes. I think it’s very unlikely that a recession has started.

Possible differences from 1949:

In 1949 the Fed, influential economists, and the finance industry were committed to some sort of gold standard. That ensured that inflation would mostly stop. There may be a comparable commitment today to a 2% inflation rate, but the Fed’s willingness to correct mistakes seems less automatic than was the case under most versions of a gold standard. I’m fairly confident that inflation is moderating, but not confident about how much it will moderate.

There was a large influx of workers to the regular job market in 1945-6 (from a rapidly shrinking military). We have a much smaller influx of workers returning after being at home for the pandemic. (See also the great resignation.)

That’s likely to mean that wages will rise for a while, and still be low enough to leave some moderate labor shortages.

These differences clearly point toward a lower recession risk than was the case in 1948 (and a higher risk that inflation will cause the Fed to overreact sometime further in the future).

Fed funds futures suggest that interest rates will start declining in late 2023. The size of the predicted decline is much smaller than what I’d expect if markets thought they could predict when a recession would occur (although that mainly means it’s hard to predict the timing of a recession). Markets aren’t predicting lower interest rates than they were predicting a few months ago, so it’s pretty clear markets aren’t increasingly concerned about a 2023 recession.

I’ll guess there’s a 35% chance of a recession in 2023. [I increased that from 30% after seeing that Metaculus strongly disagreed with me.]

If we do get one, I expect it will be mild. My model of recession severity is influenced by how close we are to a situation where employers lay off workers rather than cut wages (or reduce wage increases that workers have come to expect as automatic adjustments for inflation). I.e. when nontrivial parts of the economy need to reduce wages (or expected wage increases) if they want to keep labor supply and demand in balance.

That situation is most likely when inflation has been low and declining. It’s much less likely after an unusual burst of inflation has caused nominal wages to rise, without creating expectations of steady wage increases. Our current situation is more like the latter than at any time since the 1940s. (Okay, the early 1970s were sort of close, so I should still worry a bit.)

Mortgage defaults are also relevant to recession severity, and the situation there is similar. The unexpected inflation has created housing prices which are high relative to mortgages, leaving a bit more slack in the system than usual.

P.S. Here are some of my thoughts about the stock market:

I predict that even if we do have a recession by 2023, the S&P 500 will not close below 3600.

Antti Ilmanen has a new book out. His prior book was thoughtful enough that I’ll want to read the new one this spring. (H/T Arnold Kling).

SPAC warrants look cheap. Most likely SPACs issued more warrants than long-term investors want to hold. There may also be some neglect of warrants because warrants were uncommon enough 5 years ago that few investors developed habits of noticing them. They might languish for another year or two, but are likely to do well in the next big bull market.

Interest rates have been rising a bit faster than I expected, and that seems to be causing a strong reduction in investor willingness to buy stocks in companies that won’t be profitable for years. That effect will take time to reverse.