In this post, I make some conjectures about U.S. economic growth over the next year or two.
Many people expect a depression, due to the current high unemployment numbers. But depressions aren’t caused by unemployment – that’s a symptom, with little predictive power.
The main cause of poor economic growth has been an inability to alter wages so that the supply of and demand for labor are in balance. That typically means deflation, or a large, unexpected decline in the inflation rate, combined with sticky wages.
So I’ll mostly focus on guessing whether we’ll have inflation or deflation.
Markets are predicting low inflation for the foreseeable future. My best guess is that prediction implies a growing economy, but it’s close enough to forecasting seriously suboptimal growth that I want to scrutinize it closely.
The market’s forecast might be somewhat distorted by people buying t-bonds for liquidity reasons, but it’s still likely to be the best forecast that most people can easily find.
Forces tending to create a prolonged depression
There was a massive deflationary shock in March, with many people trying to shift their wealth into safe and liquid assets (few other than dollars and dollar-like assets were safe enough in mid-March). That means that demand for dollar-denominated assets soared. In the absence of increased supply, that means the value of dollars increased, which means that dollar-denominated wages rose (in real terms) in ways that created obstacles to labor supply and demand coming back into balance.
Because we’re on a gold standard, the supply of money doesn’t expand to accommodate this demand. Wait a minute … some of you are likely objecting that we abandoned the gold standard. What I mean is that the Fed has inherited lots of habits from when it was firmly committed to the gold standard, so that it has somewhat anchored on a presumption that the money supply should increase at a slow, steady pace.
The Fed keeps adjusting its habits after each recession to be a bit closer to what was needed to stabilize the economy after the past few recessions, but it still focuses on policies that are somewhere in between a gold standard and a standard of targeting expected wage and/or price inflation.
So, the Fed will do more than it did in the previous recession to increase the money supply. But it seems likely to anchor on that previous performance, and increase the money supply by somewhat more than last time, even though markets are indicating that won’t be enough to hit the Fed’s target anytime in the foreseeable future. After all, if the Fed errs by creating too much inflation, the Fed will clearly get blamed, whereas if the Fed makes a slightly smaller error in the direction of deflation, plenty of voters/politicians will imagine it didn’t have the power to stop the recession.
Scott Sumner has suggested some other possible reasons for Fed under-reaction (using interest rates as a measure of monetary policy, fear of a large balance sheet). Those points all reinforce my main conclusion, while reminding me to not be too confident about the causes.
There’s no clean way to measure the how much more dollar-like assets investors wanted. I’ll venture a guess that it was something like 15-25% of the U.S. stock market value, so somewhere in the $3 to $5 trillion dollar range. The Fed added less than $2 trillion in March, and another trillion in April. Another measure is MZM, which says the Fed’s actions weren’t much offset by other factors that affect the money supply. So my guess is that the Fed has still under-reacted, but is maybe close to catching up.
Forces tending to promote a recovery that is fast compared to standard recessions
If demand for safe assets diminishes, that will have inflationary effects that offset the Fed’s caution.
It seems likely that the demand for safe, liquid assets peaked in March, and will diminish as we improve our ability to predict the effects of the virus on the economy.
I saw lots of large companies announce in late March that they’d borrowed everything they could on their existing line of credit-card-like borrowing facilities, often $400 million per company. These were not companies on the verge of collapse. These were companies with fairly average credit-worthiness who worried about a small chance of their credit drying up at a time when their need for cash was hardest to evaluate.
Businesses are seeing many fewer surprises this month, so their borrowing has slowed, and is likely quietly reversing (except in a few of the worst-hit industries, such as airlines and cruise ships, which just started to look safe enough that they can borrow large amounts). I expect continuing return to an economy with a normal level of predictability will continue driving lower demands for cash.
Many workers will somewhat predictably return to their old jobs, reducing their need for safe assets. And some of their demand for safe assets was due to uncertainty over how quickly their unemployment checks would start arriving.
So the overall demand for dollars and dollar-like assets will diminish somewhat steadily, at a faster rate than in normal recessions. In normal recessions, there are big sectors where the risk of being laid off / going bankrupt remains high for an extended time. This time, most of that risk was compressed into March, and the risks are declining, due to increased certainty that the pandemic will have a mediocre outcome that most of us will survive.
I see a few signs that workers are more willing to accept explicit wage cuts, so maybe the sticky wage effect is contributing less to unemployment than usual. But I don’t have good enough evidence to put much weight on this.
The Fed has been undershooting its inflation target for nearly 12 years. The Fed may be slow to change, but it will eventually get tired enough of erring in one direction to do something else. The resemblance between current conditions and war lends some plausibility to scenarios where the Fed manages to create more than the optimal amount of inflation within months.
Debt
People who forecast economic collapse often point to massive corporate debt, and claim that makes the economy fragile. In prior recessions, banks typically took most of the risks associated with loaning money to finance housing. They often took a year or two to decide what to do with questionable mortgages, and that uncertainty likely translated into extended periods of widespread economic uncertainty.
Here’s the story of how one such mortgage company, New Residential Investment Corp. (NYSE:NRZ), is coping: in the first quarter NRZ sold enough loans and loan-like securities to reduce its holdings from around $35 billion to around $16 billion. Most of those sales appear to have been done in mid-March. That got its equity up to over 25% of those loan values, versus barely 20% at the start of the quarter.
So NRZ experienced something like a super duper depression compressed into a month or so, and appears to now be in a less precarious position. [I own stock in NRZ.] Other mortgage companies that I’ve looked at seem to be handling the crisis almost as well. There’s a wide range of how much risk each one takes, but it looks like nearly all of them have their risks under control, and will recover in the absence of further large surprises.
Why is this different from prior mortgage loan crises? One big reason is that those mortgage companies borrow in ways that are subject to margin calls that work something like the margin calls on the stock market. Something apparently deters banks from using this kind of borrowing, but the current mortgage industry seems to have imitated the risk-management competence of stock brokers, likely at the cost of triggering a sharper stock market panic than would otherwise have happened (I’m guessing that some hedge funds were buying mortgages at fire-sale prices instead of buying stocks that were selling at fire-sale prices). I’m not sure this is an improvement over prior financial crises, but it has almost certainly shortened the duration of the crisis.
The mortgage industry stabilized at about the time when the Fed decided on March 23 to buy mortgage-backed securities if needed. Did the Fed announcement stop a cascade of margin calls? Or did the Fed manage to time its action just as the industry had sold enough to avoid further margin calls? I don’t see an easy way to distinguish those two possibilities.
In the last recession, mortgages became risky because housing prices were declining more than most people anticipated was possible. This year, housing prices were starting from levels that looked more sustainable, and what I read in relevant corporate announcements indicates they’re not changing much – both supply and demand seem to have been disrupted in comparable amounts.
Back in March, there was talk about massive nonpayment of rent. By now, enough of the newly unemployed have gotten unemployment compensation, and realized that delaying rents won’t help them much, for those fears to look unlikely. E.g. Preferred Apartment Communities (NYSE:APTS) said in its recent earnings report that its April housing and office rent collections were down only 2 to 3%, except for non-grocery retailers, which were down 32%.
Current unemployment is not much like typical unemployment. Some of it is due to to employment being outlawed (almost certainly temporary), some is due to various obstacles to consumer spending that look like they’re being relaxed, and some is due to the government actively rewarding workers for staying home. (Trump-the-campaigner has been has been eager to look like he wants to get people back to work. I can’t quite tell whether he’s campaigning against the Trump administration, whose policies are clearly designed to keep people away from work. Or at least they were 2 months ago).
While writing this, I came across this piece by an economist who has different, and often better, arguments for my general points. He’s likely a bit too optimistic about how quickly people will be willing to resume work and spending. I also have less confidence in the Fed than he does.
I expect a check-mark-shaped recovery. GDP seems to be recovering steadily, although at much less than an optimal rate. Even with an occasional setback due to additional, smaller waves of the pandemic, GDP seems likely to be higher in 2021 than in 2019.
P.S. Why are stock market bottoms looking more V-shaped recently? Is it due to a larger number of sophisticated traders who know that margin calls are making the market inefficient at the bottom?