Known Unknowns


Welcome to Known Unknowns, a newsletter that is an empiricist (thank you, David Hume), but also a rationalist (old-school kind, not the new Silicon Valley version).

How much is too much?

There is a strange economics argument brewing. The Biden administration has proposed a $1.9 trillion aid package, and even economists who are normally debt enthusiasts, like Olivier Blanchard and Larry Summers, are worried that it’s too big. Some relief is needed, but the important question is how much? Why should we spend more than what COVID cost the economy? Household balance sheets are in good shape, and once the economy opens, it’s poised to bounce back. So why give money to people who don’t need it?

Fans of the plan argue don't have answers other than, “Why not?” We need to do something, so why not make it as big as possible—and what’s the harm? After all, lots of people are suffering. This is the natural argument to make once you believe that there is no costs to running large debts. I’ve seen it argued in The Atlantic and The Economist that debt doesn’t matter, because we can monetize it and roll it over. And if debt doesn’t matter, why not send everyone a check?

Economists have spent years trying to figure out the magic number, i.e., the level of debt that tips an economy into a debt crisis or inflation spiral. And that number may not even exist. But just because there’s no magic number doesn’t mean that there’s no limit to debt.

It is important to understand the magic number doesn’t exist simply because it changes over time. The debt limit is structural. It depends on interest rates, on the state of the global economy, the age of your population, and economic growth. And here’s where it gets complicated. We may be able to manage lots of debt right now—maybe 150% of the GDP or more. But what if any of the factors I listed change over the next few years? Then the tolerable debt-to-GDP ratio changes, but you still have all that debt on your balance sheet.

So, will a $1.9 trillion relief package cause runaway inflation or a spike in interest rates? Probably not, at least not right away. But that doesn’t mean that it doesn’t pose costs and risks. I agree that we risk overheating the economy. But the costs of overheating aren’t just higher employment and inflation—we’re also introducing a distortion on prices and the price of risk, and that tends to have unintended consequences. Also, I’m not comfortable with the precedent that every time we have a shock, we send people checks (paid for by monetized debt), whether or not they suffered any actual economic setbacks. Relief for economic hardship is one thing, but “let’s send everyone money the Fed printed” just to make Americans happy is another.

Also, you can’t have it both ways. A structural change can mean that there’s less cost to carrying debt than there was in 1974, but it could also mean that there are fewer benefits, too, and I haven’t heard the benefits of the stimulus well-argued. Have you?

More on inflation

There is a generational divide on inflation. Young economists (anyone under 45) don’t seem to be worried. But economists over 45 are. I reckon that’s because anyone under 45 can’t remember a time when inflation wasn’t small and predictable. I’m old-school—not because I worry about inflation (anyway, I’ve never experienced it), but because I still believe in theory. When I was in grad school, we were all trained to be empiricists. Causal inference—and, later, RCTs—were the name of the game, and were how you found success as an economist. And I love data, because how else do we test our theories?

But data without theories is meaningless, and this is critical when we’re trying to make distinctions between what’s structural and what’s cyclical. We now have almost 40 years of data telling us that rates and inflation won’t go up. If you’re an empiricist, it seems safe to infer that we can assume the same going forward. But if you’re theorist, you might argue that we’re still in the middle of a debt cycle, and maybe we can’t conclude much about the next 50 years based just on the last 40 years of data.

My vintage of economists has only lived under one structural regime, but it could also change, and a theory helps you make this distinction.

Also, check out this picture I made.

Let’s say that we have a big boom later this year, on top of more stimulus. Throw in low interest rates, and that could mean high inflation. Well, in that case, the Fed will step up, right? I’m not so sure. This isn’t Volcker’s Fed. We’ve come to expect the Fed to keep employment up at all costs, so do you really think that they’ll slow a recovery to fight inflation—something many people can’t even conceive of? People still regard Janet Yellen’s 25 basis point increase during a boom as a moral failing.

Also, as my picture shows, the Fed is buying lots of our debt. If they start selling it, too, on top of the existing debt that we’re rolling over—well, things could get messy in the bond market.

It could be another structural change, and not just one that will lead to short-term inflation. Volcker convinced markets that the Fed took inflation seriously, and that belief contributed to low inflation rates over the past 40 years. But if the Fed doesn’t take strong action next time, that credibility will be gone, and who knows what the future will hold?

It will be interesting to see, and I think we’ll all learn a lot in the process.

While we’re throwing money around

You may have missed it, but Covid relief includes a bailout for multi-employer pension plans. Many of these plans, like the teamsters pension, were poorly managed and reform was definitely necessary. Now there are no costs to their fecklessness. The pensions are not only being bailed out for their current shortfall, they can reverse earlier benefit cuts--putting taxpayers on the hook for another 30 years. Does this have anything to do with the pandemic? Not at all. The unions have been angling for this for years.

But why not? If you believe that debt is meaningless, why not bail out everyone? Except this pension bailout it comes at the expense of a month of unemployment benefits.

Until next time, Pension Geeks!