Inflation, defined as an increase in the general level of prices, is not directly observable more than the general price level itself (also called “overall price level”).  You cannot go to your local convenience store and order one unit of GDP, asking, “How much is it?” —that is, how much is the general price level for one unit of GDP, made of a fraction of an automobile, a few words of medical advice, a bit of an Amazon delivery service, a dozen bubble gums, etc.? Inflation is estimated by measuring the increase in the average price of some “representative” basket of goods and services. It is incorrect to take this estimate of inflation (the consumer price index or CPI, for example) and try to trace the causes of inflation back to observed individual prices, for the latter also incorporate relative price changes (among goods and services) not caused by inflation.

Consider the following standard pronouncement. The Wall Street Journal writes (“Record Diesel Prices Pressure European Drivers, U.S. Deliveries,” May 13, 2022):

Rising energy prices are a major factor contributing to the persistence of inflation

The following historical observation  can serve to falsify such a claim. On the chart below, let’s focus on the period from January 1999 to July 2008. Over this period the average price of energy for American consumers (calculated from a sub-index of the CPI) increased nearly non-stop and was 172% higher at the end of the period than at the beginning. But some other prices decreased. For example, the average price of consumer durable goods (including such things as cars and appliances) decreased by 12%; the cost of apparel, by 10%. While inflation was estimated to be 33%, some prices increased, other decreased.

And that is the point. Without inflation, relative prices would continue to change, some prices increasing, others decreasing. The measured total change of a given price is, by definition, the sum of its relative change (relatively to other goods, without inflation) and of the inflation rate:

total price change = relative price change + inflation.

One cannot say that the total price change “contributed to” inflation, because inflation has already been added to the relative price change to produce the total. If you obtain 3 by adding 1 to 2 (2+1), it does not make much sense to conclude that 3 “contributes” to 1 (except perhaps in a very formal algebraic way).

It is true that if a supply shock reduced  the production of oil and thus of all consumer goods that use this input, there would be a one-time increase in the general price level because the same quantity of money would be chasing fewer goods. But there is no reason why inflation would continue month after month, year after year, as we see for the CPI in our chart. A persisting inflation (what people usually mean by inflation) requires more new money chasing the existing (or a fortiori reduced) volume of goods.

This argument rests on standard microeconomic theory (which deals with relative prices and is also called “price theory” for that reason) and on a macroeconomic theory called monetarism.  But note that any counter-claim must also be based on some economic theory, or else it is just a vague ad hoc intuition. The media and alas even the financial press often rely on the latter.

Over the past few decades, monetarism of the Milton Friedman type has been abandoned or modified by economists. But this does not necessarily make it invalid, especially in its fundamental result that an increase in the money supply over and above the demand of the public for money will, after a lag, generate inflation. (I understand that co-blogger Scott Sumner defends a modified version of monetarism, which does not seem to contradict my claim above, at least in ordinary circumstances.)

An argument close to mine above is presented in a recent article by John Greenwood (Invesco, Ltd.) and Steve Hanke (John Hopkins University): “On Monetary Growth and Inflation in Leading Economies, 2021-2022: Relative Prices and the Overall Price Level,” Journal of Applied Corporate Finance 33:4 (2021). A paragraph from the abstract of the article provides a good summary:

The authors argue that this consensus [on current inflation being largely generated by “supply chain disruptions”] will prove to be wrong because of its failure to distinguish between relative price changes and changes in the overall price level. The movement of any single set of relative prices fails to convey information about the overall inflation rate. And as asserted by the Quantity Theory of Money, the overall inflation rate and price level are determined by changes in the money supply broadly measured. Changes in relative prices, on the other hand, result from changes in demand and supply in the real economy, making them independent of changes in the money supply. So, while a doubling of the money supply will result in a doubling of all nominal prices, relative prices remain unaffected by monetary policy.