Most Inflation Stories Don't Add Up

Using price theory to sort through narratives that won't go away

You are reading Economic Forces, a free weekly newsletter on economics, especially price theory, without the politics. Economic Forces arrives weekly in the inboxes of over 15,000 subscribers. You can support our newsletter by sharing this free post or becoming a paid subscriber:

---

For a newsletter about price theory, we’ve written extensively about inflation. While this might seem odd—since price theory is all about relative prices and inflation is about the behavior of money prices—the tools of price theory are crucial to this debate.

The puzzle is simple but important: When prices rise rapidly across the economy, how do we determine what’s actually causing those increases? Everyone has their preferred explanation—supply chain disruptions, excessive demand, or maybe corporate greed. But rather than debating magnitudes, we should first ask: What’s the mechanism behind each claim? What would we expect to see in the data if that mechanism was actually driving inflation?

Price theory helps us in two ways. First, it provides testable implications about whether inflation is supply-driven or demand-driven. Second, price theory can help us to do is to distinguish between changes in relative prices and changes in money prices. This is clearly important because the emphasis on a particular price or subset of prices often confuses changes in relative prices with changes in inflation.

How Can We Determine Whether Inflation Was Caused by Supply or Demand?

Let’s start with the first point. How can price theory help us to distinguish between supply-driven and demand-driven inflation? (A subject Brian has written about previously.) Just draw supply and demand curves. If inflation is supply-driven, this would mean that the supply curve is shifting up and to the left. As a result, rising prices are associated with declines in output. On the other hand, if inflation is demand-driven, rising prices are associated with rising output.

Fortunately, at the macroeconomic level, there is an easy way to determine whether inflation is supply- or demand-driven: Nominal GDP.

The reason that nominal GDP is useful is that nominal GDP can be separated into two separate components, real GDP and the GDP deflator (the price level). As I discussed above, price and quantity move in opposite directions following a supply shock whereas they move in the same direction when there is an increase in demand. Suppose, for example, that nominal GDP is typically constant over time. A supply shock would result in a muted response of nominal GDP. In fact, nominal GDP might not change at all since the price level and output are moving in opposite directions. On the other hand, an increase in demand would cause both prices and output to rise, thus increasing nominal GDP.

Even so, there is an additional point about supply shocks. Typically, we think of supply disruptions having an effect on the relative prices of goods. Certain goods are subject to the supply disruption. This causes these prices to rise relative to others. There might also be secondary effects on related prices. Since we measure the price level using a weighted average of prices, these changes can cause the price level to increase. However, inflation refers to a sustained increase in the price level over time. A supply shock increases the price level, not the inflation rate.

Of course, one might argue that the pandemic created a situation in which most goods were subject to some type of supply disruption. However, the same logic applies. All else equal, one should expect the supply chain disruption to increase the price level. But this is another reason to focus on nominal GDP rather than solely on the price level or the inflation rate.

It is a bit more difficult to disentangle these effects in a growing economy with positive rates of inflation. Nonetheless, one can think of the distinction between supply-side and demand-side factors for changes in NGDP growth as follows. In a growing economy, a supply-driven inflation would be expected to result in slower — or even negative — real GDP growth and rising inflation. Since these two components should be changing in opposite directions, one would expect a supply-driven inflation to result in little change in the growth rate of nominal GDP growth — after all, real GDP growth declines and inflation rises. One thing that we know for certain is that it is not going to cause a surge in nominal GDP growth since the two components are moving in opposite directions.

On the other hand, a demand-driven inflation would be expected to increase both real GDP growth and inflation. This means that we should observe an acceleration in nominal GDP growth under demand-driven inflation.

If one looks at nominal GDP growth in recent years, it certainly seems like inflation is primarily driven by demand. Below is a graph of nominal GDP growth from 1987 to the present. As shown, nominal GDP growth grew at an average annual rate of around 5% over this period. At the start of the pandemic, nominal GDP declined before the growth rate accelerated quite rapidly. The growth rate has since declined to its average over this period.

No one would deny that supply-side factors played a role. We know that supply chain disruptions took place. However, if you take the basic price-theoretic insight seriously, it seems quite clear that inflation was primarily demand-driven. The period of high inflation was associated with an acceleration in the growth of nominal GDP, precisely what a price theorist would predict from demand-driven inflation.

Long-time readers will recall that I argued that inflation was entirely predictable using a simple forecasting model that included only two variables: inflation and the growth rate of the money supply (as measured by Divisia M2). This forecasting model suggested that, using only data available as of the end of 2020 that we should expect rising rates of inflation throughout 2021. In addition, the model predicted, using only data through 2021 that the U.S. would experience persistently high rates of inflation throughout 2022. That model’s predictions turned out to be quite accurate, despite its simplicity.

What About Markups or “Greed”?

As one would expect, price theory offers additional lessons about how to evaluate supply-based arguments. Some argued that rising markups were evidence of supply-driven inflation. Some version of this argument was that when costs rise, firms use these rising costs as justification not only to raise their prices, but to increase their markups. However, as Brian pointed out, this doesn’t make sense. Firms with market power don’t need the cover of rising costs to increase their prices.

In a previous post, I argued that the standard price-setting model suggests that rising markups are actually incompatible with supply-driven inflation. One can think of supply shocks as increasing marginal cost. A rising marginal cost of production will indeed result in higher prices being charged by firms. However, the gap between price and marginal cost typically declines as marginal costs rise. Thus, one should expect to see declining mark-ups in a supply-driven inflation. If mark-ups are rising, this can only be explained by increases in demand.

Whether applying price theory as a tool by which to interpret the data or by using standard price theoretic models to assess the claims of those who argue that the inflation was supply-driven, it seems quite clear to us that the inflation was primarily demand-driven. This is not to say that supply-side factors didn’t matter. Certainly, supply-side factors likely made things worse. However, demand-side factors seem to be dominant and the supply-side arguments don’t seem to hold up well in the face of scrutiny.

Muddling Through the Data Without Price Theory

Given all of this, Brian and I were surprised to see a recent op-ed by Peter Orzag in The Washington Post arguing that the inflation was supply-driven after all. We were equally surprised at the weak evidence provided to resurrect this argument.

The main direct evidence that Orzag claims to provide is data on delivery times. He shows that delivery times increased by several standard deviations above the mean. However, his explanation for why this is evidence of supply-driven inflation isn’t the slam dunk that he thinks it is. For example, he writes:

Why did these effects play out over such a long time? At the start of the pandemic, Americans shifted their spending from services (like travel, eating out and going to the movies) to goods (like computer hardware and exercise equipment) — just as a snarled supply chain caused those goods to be in short supply. This caused prices to spike.

Supply chain disruptions, as traditionally measured by delivery times, largely righted themselves by 2022, so on the surface it might seem like these effects would have then, too. But in response to the severity of the shock, which has no modern comparison, inventory managers were cautious through at least this summer.

As a global CEO recently said, companies did not want to get caught not having what consumers wanted in stock again. Increasing prices to dampen demand and extend inventory life meant that the prices charged to consumers rose further above the cost to the seller. This meant profit margins went up, not because of “greedflation” but instead because of a rational response to businesses managing their inventories.

The way that this is written is to suggest that we know the inflation was supply-driven because delivery times increased as firms struggled to maintain inventories. The quote from the CEO is designed to lend support for this argument. But what this example fails to do is address what would happen to deliveries and inventories following a massive increase in demand. Similarly, as I have written about before, asking suppliers why prices are rising is fraught with error.

In that previous newsletter, I wrote the following:

In University Economics and its subsequent iterations, Alchian and Allen give the example of a butcher. Since demand fluctuates from day-to-day, the butcher will tend to have an inventory of beef. This inventory acts as a buffer against these day-to-day fluctuations in the demand for beef. On days when demand is high, the inventory will tend to be drawn down. On days when demand is low, there will be more inventory left over. However, suppose that there is an increase in the overall demand for beef (i.e., people are buying more beef, on average, every day). In this case, the butcher will see his inventory drawn down by more than anticipated. The butcher might even run out of beef. As these inventories are drawn down, the butcher will place a bigger order with the meat packer-supplier. The meat packer-suppliers will then see a draw down in inventories as well and will then want to purchase more cattle.

Since, in short-run, the supply of cattle is fixed, this greater competition for a fixed supply of cattle will tend to lead to higher prices for cattle. The meat packer-suppliers will then raise their prices and tell the butcher that they have to raise their prices because the price of cattle has gone up. The butcher will then raise his prices and tell the customer that he is raising his prices because his costs have gone up.

After presenting this example, Alchian and Allen then ask the reader to think about who is responsible for the increase in the price. The answer is clearly the consumers whose demand has risen. They are the ones who set in motion the process by which prices increased. Yet, if we were to ask meat packer-suppliers why prices have gone up, they will blame the cost of cattle. If we ask the butcher why prices have gone up, he will blame the rising prices of the meat packer-suppliers.

This sounds exactly like the narrative that Orzag constructs, complete with reference to a CEO’s opinion. However, in the example above, the cause is demand and not supply.

The main point is that it is futile to ask sellers why prices are rising. Increases in demand and insufficient supply will actually look the same to the seller. All the seller sees is that it is harder to replenish inventories and that his costs are rising. Those observations don’t tell us anything about the source of the problem.

The beginning of Orzag’s statement also relies on a bit of a rhetorical trick. He starts by saying that spending on services declined and were replaced by spending on goods. The way that the sentence is written seems to suggest to the reader that this was a mere substitution of one type of consumption for the other. However, to the extent to which such substitution took place, one might care about the relative magnitudes. Was the increase in the demand for goods greater than, less than, or equal to the decline in the demand for services? Lucky for us, there is a simple measure of aggregate spending in the economy, nominal GDP, which I have already shown experienced accelerated growth during this period. Total spending was going up. Once you recognize that, Orzag’s argument becomes almost indistinguishable from the Alchian and Allen example.

Orzag then goes on to cite some papers that seem to reject demand-driven explanations of inflation. In particular, he cites studies that suggest that inflation didn’t seem to vary across countries or states in conjunction with the size of their fiscal stimulus.

But we here at Economic Forces are not alone in our assessment of the importance of demand-side factors. In fact, we can also cite papers that find precisely the opposite of Orzag’s preferred papers.

For example, Robert Barro and Francesco Bianchi focus on fiscal policy (another demand-side factor) across OECD countries and find that their fiscal policy variable has a positive and statistically significant effect on the rate of inflation during this period.

The interaction between monetary and fiscal policy is also important. Another paper, by Christina and David Romer, argues that the policymakers at the Federal Reserve had begun to emphasize the need for robust labor markets to promote inclusion. As a result, these policymakers were unwilling to respond adequately to rising inflation without seeing evidence that the economy had reached full employment.

Regardless, I don’t want to make this post about who has the better studies. The point of Economic Forces is that price theory is your guide. Data do not speak. Price theory is the foundation upon which one can interpret the data.

Over the past few years, Brian and I have consistently outlined precisely what price theory has to say about investigating the recent inflation. We have not only explained the predictions that price theory makes, but also how one would go about assessing whether those predictions are evident in the data, as well as what we would expect to see if we were wrong.

Price theory rejects the idea that we can ask people why prices are rising. Price theory rejects the idea that supply-driven inflation is associated with higher rates of nominal GDP growth. Price theory rejects the idea that rising mark-ups are driven by rising costs. Price theory rejects the idea that the inability to keep goods on the shelf necessarily represents a supply-side problem; it could just as easily be driven by demand.

These price theoretic insights are important in assessing the evidence. “Looking at the data” simply will not do. One needs the foundation in price theory to assess competing claims as to how the data were generated. On the topic of demand-driven inflation, the dog did in fact bark.