The Real Problem With Inflation Expectations

Find and replace "inflation" with "NGDP growth"

Macroeconomists often talk about the necessity of “anchored inflation expectations”. The essential idea is that if the public starts expecting high inflation, their expectations will be self-fulfilling. This view has recently come under fire with the now-infamous Jeremy Rudd paper and a general skepticism around the emphasis on expectations. 

These criticisms generally center around problems with (1) measuring inflation expectations and (2) pointing out how the causal mechanisms by which expectations can influence actual inflation are very shaky. 

This post is going to be a strange exception. I call for economists to stop looking at inflation expectations, because they can be replaced by another, better variable. 

The Monetarist Causal Mechanism - Why Expectations Matter

First, let’s think about the general monetarist logic for why inflation expectations should matter for future inflation. Imagine that the “natural” real interest rate in the economy is 2%. Further assume the Fed currently has the nominal rate set at 4% and that the public’s inflation expectations are 3%.¹

This means that the public currently thinks that the real rate is 1% (3% inflation with a 4% nominal rate). That’s below the “natural” rate. The general way to interpret this is that borrowing will rise because it’s cheaper than in the “natural market-clearing scenario”. Since borrowing is the primary cause of money creation in the modern economy, this will lead to a higher money supply, and the higher money supply makes the inflation expectations become self-fulfilling. 

OK, that is a bit iffy. We have no idea how much the money supply will grow and how much inflation it will really cause. Instead, let me just give you a weaker claim: if you expect 20% inflation, you’ll borrow more money than if you expect 2% inflation. As a result, you’ll cause inflation to be higher. Makes sense right? So inflation expectations cause higher inflation (even if we’re unsure about the exact numbers). 

Unfortunately, that’s not completely right. 

Why We Borrow

Let’s think about why higher inflation expectations would cause people to borrow money. More specifically, what do we consider when borrowing? The standard answer would be that if we expect money to be worth less in the future, we will want more now so that we can spend it while we can get bang for our buck.

But…this logic forgets the fundamental problem with borrowing money: you have to pay it back. 

That is, in fact, the primary reason why increased inflation expectations would make us borrow more. We expect the money to be easier to pay back since we’ll have higher wages in the future. But what if we expected higher inflation without expecting higher wages?

This does happen by the way. We call them, you know, supply shocks. They increase the price level but reduce economic activity. As a result, if people expect a supply shock to cause further inflation, they will not expect their wages to significantly increase. Therefore, they will not be as capable of taking on more debt. 

We can even formalize this idea with a fairly simple stylized model. If you assume a consumer who receives income in two periods and has CRRA utility their borrowing will primarily depend on their future nominal income (which we can proxy with, you guessed it, Nominal GDP). In fact, if you assume logarithmic utility, then shocks to the price level don’t matter for borrowing at all if future income stays constant. And in the more general case, a supply shock that leaves nominal income unaffected could either increase or decrease borrowing based on the relative risk aversion parameter (this should be giving you flashbacks to how supply shocks can increase or decrease nominal GDP).²

So…what does this tell us about the real world?

Well, first of all, stop using inflation expectations. Though inflation expectations can affect borrowing, a much better proxy is expectations of nominal wages or nominal GDP. They tell us people’s beliefs about how capable they’ll be of paying off debt, and, therefore, how worthwhile it is to borrow more. Unanchored inflation expectations will not be the cause of inflationary spirals, that’ll be unanchored nominal GDP expectations. 

1. I’m mostly abstracting away from the measurement of expectations, but the TIPS spread would be my favored method.

2. This, by the way, does not imply that the canonical New Keynesian model is wrong about how inflation expectations affect inflation. In that model, a supply shock will decrease the natural rate, and, therefore, not have as large of an effect. But we do not observe the natural rate in the data, so the New Keynesian model often leads to bad intuition in the real world.