# Why was Friedman so wrong about inflation?

In a famous 1977 lecture, Milton Friedman claims that inflation depends purely on the monetary policy and he proves it with a convincing chart (youtube video here)

I have reproduced his chart using FRED data (links below) and it works!

However -and this looks like a perfect joke about economists making predictions- from the year after his presentation the relationship did not hold anymore.

Any idea why the relationship got much fuzzier from 1979 on?

Data:

• Small movements are hard to see, so the initial "relationship" in this chart could be an optical illusion. Any chance there is a similar chart with the growth/increase rates of these time series? Aug 5 at 16:57
• @Giskard: sorry I did not zoom the initial part until 1978, but it looks exactly as Friendman's sldie (see the link in my question) Aug 5 at 17:00

Why was Friedman so wrong about inflation?

He actually was not wrong in a broad sense, only in more specific sense.

First, you are actually misrepresenting and slightly 'strawmaning' Friedman's ideas. Friedman never claimed that inflation depends solely on monetary policy.

Literally in the same video just few seconds later Friedman shows this graph:

And furthermore, Friedman literally in the video states [emphasis mine]:

Again you see the same story. Now interesting thing that you can see that the quantity of money for a while later in the seventies was running ahead of the price index, but now they are coming back together again. And that is behavior you will very often observe. The quantity of money per unit of output is the major factor that from the immediate sense determines the price index but it does not operate instantaneously.

So:

1. Friedman never claims that inflation is only determined by quantity of money rather that it is a major factor. Since he believed other factors (like velocity) are more or less constant in a long run he considered inflation to be always monetary phenomenon.
2. Friedman never claims that there is always tight relationship between money supply and price level, rather that in the long run they move together. On your graph you can still see they move together more or less in lockstep even though they no longer overlap.

To be more specific Friedman was a proponent of the Quantity Theory of Money, given by (see Mankiw Macroeconomics pp 87) :

$$MV=PY \implies P = \frac{MV}{Y}$$

Which tells us that the change in price level does not depend just on money supply $$M$$ but also real output $$Y$$ (that is why Friedman plots quantity of money per unit of output) and in addition on velocity of money $$V$$.

So the graph you construct in itself does not contradict Friedman thesis.

In addition, monetary scholarship confirms that inflation in the long run is caused by money supply, and thus is predominantly monetary phenomenon which support's the Friedman's main thesis (for example see Maune, Matanda, Mundonde, 2020; Barsky, 1987; Mankiw 2019; Romer: Advanced Macroeconomics 4th ed; Grauwe & Polan, 2005; Benati 2005; Frain 2004 and sources cited therein).

However, Friedman was wrong about inflation in more specific sense.

1. Friedman considered $$V$$ to be constant in the long-run, but recent evidence suggest that it is likely that there can be secular trends and long run changes in the velocity of money. As you can see prior to 90s the velocity was quite constant, but since 90s it significantly dropped, which can explain why the relationship is more fuzzy recently, the increase in money supply is being offset by drop in $$V$$.

1. Friedman was wrong in a sense that we now know that the relationship is not proportional.

Modern scholarship no longer relies on $$P=\frac{MV}{Y}$$ rather modern description of money market would be given by:

$$M/P = L(i,Y) \implies P = \frac{M}{L(i,Y)}$$

This is not a very large change since the signs of the relationship are still the same (i.e. $$P$$ still depends positively on $$M$$, negatively on $$Y$$ and positively on $$i$$ (since $$\partial L/\partial Y>0, \partial L/ \partial i<0$$), which affects $$V$$ (which can be also though to be function of $$i$$ in QTM). However, the relationship is no longer proportional and mechanism is a bit different.

Friedman, was likely also wrong about host of other details (science always marches forward), but his general thesis is still considered valid by majority of economists.

Friedman, "inflation is always and everywhere a monetary phenomenon", in the sense that price increases are produced by significant increases in the amount of money. Although this wording refers to the neutrality of money in terms of shaping real economic figures, it is worth noting that this neutrality does not apply in the short term when inflation is unexpected. An example, increases in the price level P leading to the same percentage increases in the nominal wage W do not change the actual wage W / P and therefore do not change employment and output. However, if price increases are unexpected and there is no corresponding increase in nominal wage, real wages, employment and output change in the short run, ie there is a real effect of unexpected inflation. Of course, the long-term adjustment of expectations to changes in the price level certainly changes the nominal wage so that the real wage eventually returns to the original level and, consequently, the neutrality of money is valid for a long time.

Basically inflation is always and everywhere a political phenomenon. The question is whether societies want low inflation. It is reasonable to doubt this today.