Your Notebook is now a Draft.

Once it's ready, please submit it for review by our team of Community Moderators. Thank you!


This content now needs to be approved by community moderators.


This essay was submitted and is waiting for review.


Two Theories of Inflation

by Arnold_Kling {{qctrl.question.publish_time | dateStr}} Edited on {{qctrl.question.edited_time | dateStr}} {{"estimatedReadingTime" | translate:({minutes: qctrl.question.estimateReadingTime()})}}
  • Facebook
  • Twitter

  • From March of 2001 until March of 2021, prices in the United States as measured by the Consumer Price Index, rose by an average of 2.1 percent per year. Over the following four months the CPI increased at an annual rate of 8.7 percent*.

    One theory of inflation is that the future will look mostly like the twenty years prior to March. Call this the Mean Reversion Theory, or MRT. Many economists now seem to believe in the MRT. In July, Reuters polled more than 200 economists, and over 70 percent of them said that recent high inflation is transitory.

    Financial markets too reflect a belief in MRT. On August 4th the market forecast known as the ten-year breakeven inflation rate (T10YIE) was 2.35 percent. This is the inflation rate needed to make a 10-year inflation-indexed U.S. Government bond yield the same amount as a 10-year bond not indexed to inflation. If inflation is higher than 2.35 percent over the next ten years, one would be better off holding inflation-indexed bonds. If inflation is lower than 2.35 percent, one would be better off holding regular bonds. I'll note here that my largest portfolio holding is inflation-indexed bonds.

    What's an alternative theory to MRT? To illustrate, let's start with Milton Friedman's famous words, "inflation is always and everywhere a monetary phenomenon." It's when people pay more money for the same goods, and they do this when they have more money to spend.

    Suppose the economy consisted of one good, 'corn,' and we produced 100 bushels per year. If we spent $100 per year on corn, then the price of a bushel would be $1. If we spent $500 per year on corn, then the price per bushel would be $5.

    What then determines how much we spend on a good? Although Friedman and other monetarists use a narrower definition of money, I prefer to include all forms of paper wealth created by the government. This means I treat government spending financed by bonds as approximately equivalent to government spending financed by printing money. We can call this modified form of monetarism the Government Debt Theory of inflation (GDT).

    The real world gets complicated

    In the real world, production is not fixed. There are unemployed resources---notably, labor. If there are enough unemployed resources, the thinking goes, and we increase spending from $100 to $500 per year on corn, this additional spending will result in fewer labor resources going unemployed. This additional labor will result in 500 bushels being produced rather than 100, and the price will remain $1 per.

    I have noted that Modern Monetary Theory "seems to believe that in a fiat money regime inflation only appears when the economy is at full capacity, with no unemployment. But we know from the experience of the 1970s that this is not the case." Today, with car makers facing a chip shortage, we cannot simply throw unemployed workers at the problem to increase production and thereby keep prices from rising.

    Measuring inflation

    How do we measure inflation in the first place? In the real world, we produce millions of goods and services, not just corn. To arrive at an overall measure of "the" level of prices, we select a weighted average of some of the prices that we observe. For example, The Consumer Price Index uses weights based on the purchases of a supposedly typical urban consumer.

    To extract the underlying trend in inflation from short-term fluctuations in the CPI, economists use a variety of different weighted averages. One approach is to weight the food and energy components at zero on the grounds that volatility in these components is often misleading. This provides what is known as the "core CPI." And notably, the rise in core CPI has been essentially the same as that in the regular CPI.

    Still, other inflation indicators can tell different stories. In particular, Tyler Cowen points out that median inflation in recent months is barely more than 2 percent. Mathematically, recent inflation is highly skewed, with most prices going up very little but a few prices soaring by a lot. Now, this observation might justify Mean Reversion Theory on the grounds that getting back to low inflation requires only a few markets to settle down.

    The Government Debt Theory of inflation

    If Sally Spender borrows $1000 from Larry Lender, then neither one is wealthier. Larry is minus $1000 in cash but plus $1000 in the form of an IOU from Sally. Sally is plus $1000 in cash but minus $1000 in the IOU she gave to Larry.

    But suppose that the government writes a $1000 stimulus check to Sally and borrows the money from Larry. Sally has $1000 and does not owe anybody anything, so she is wealthier by $1000. Larry is minus $1000 in cash but has a $1000 government bond, so his wealth is unchanged. So overall, there is $1000 more in paper wealth in the economy. This is the notion that underlies Government Debt Theory (GDT).

    I should point out that this interpretation of stimulus is not without controversy. Harvard economist Robert Barro famously asked "Are Government Bonds Net Wealth?". He pointed out that if people anticipate future tax increases to pay off government debt, then these expected tax liabilities will cancel out the increase in paper wealth. If Barro is correct and this sort of equivalence holds, then GDT is false. But, I do not think that Barro is correct.

    Instead, I take the view that paper wealth created by the government will eventually lead to higher spending. Pandemic-era deficits are adding $6 trillion of this form of paper wealth to an economy with total spending (nominal GDP) of $20 trillion. The Federal Reserve has purchased many of the bonds, but to do so it has provided assets to the private sector. To me, it does not matter how much of the paper wealth was "monetized." Government debt is private sector wealth, according to GDT. So far, most of this wealth has been saved. But as households and businesses start to spend, inflation is bound to pick up, in my view.

    Testing the GDT

    The next eighteen months will provide multiple tests of the GDT. I believe there is a 70 percent chance that the rate of inflation, as measured by the CPI, will be more than 3 percent between December 2021 and December 2022.

    If inflation does indeed fall back to below 3 percent and stay there, however, then this will provide evidence against GDT, and I will have to give some other theory more weight.

    What about the financial markets, with their expectations of low inflation? What will happen to the so-called real interest rate (the interest rate minus the rate of inflation) if inflation remains high? For example, the interest rate on three-month prime commercial paper was roughly 0.25 percent in March. In the following three months CPI inflation was at an annual rate of 9.7 percent for a real interest rate of 0.25 minus 9.7, or negative 9.45 percent! That's a sweet deal for corporate borrowers and a raw deal for savers.

    I believe it is a safe prediction that in 2022 we will not see any three-month period with such a low real interest rate. Following from that, I predict with 60% confidence that if inflation does not die down, then for any months in 2022 where it remains above 3 percent, the interest rate on three-month commercial paper will be no lower than 3 percentage points below the rate of inflation. So, if inflation is, say, 6 percent, then the commercial paper rate will be at least 3 percent---and quite possibly higher than that.

    And how about the interest rate on government debt? If savers can get 3 percent on prime commercial paper, then they are unlikely to accept much less than that on three-month Treasury bills. But if the interest rate on T-bills hits 3 percent, that might be enough to precipitate a government debt crisis. I would define a debt crisis for the U.S. Government as one which forces Congress to pass at least $500 billion in spending cuts with a Democratic President signing the legislation. The probability of such a crisis emerging in the next two years is low, but it is not zero. Let's say 2 percent.

    Do you disagree that inflation is likely to remain elevated over the next eighteen months, as GDT leads me to predict? If so, what theory do you prefer? The theory that high inflation is transitory? A theory that the Fed will step in and keep inflation at low levels? The theory that I here associate with Modern Monetary Theory, that inflation will only rise when there is hardly any unemployment? A theory that we will have another recession in the months ahead, due to COVID or some other factor?

    * To obtain this, divide the July CPI of 272.265 by the March CPI of 264.793. Take the result to the third power to annualize it, and then subtract one.

    Economy & Business
    Submit Essay

    Once you submit your essay, you can no longer edit it.