Saturday, 7 April 2012

Alas Scott and Joan

For some time now Scott Sumner has been giving the late, great Joan Robinson a hard time about this pronouncement, which appears in her 1938 review of Bresciani-Turroni’s study of the German hyperinflation:
An increase in the quantity of money no doubt has a tendency to raise prices, for it leads to a reduction in the rate of interest, which stimulates investment and discourages saving, and so leads to an increase in activity. But there is no evidence whatever that events in Germany followed this sequence.
Let’s break that down into bite-sized morsels, to see what the trouble appears to be:

An increase in the quantity of money has a tendency to raise prices. Very few economists would argue with that.

An increase in the quantity of money leads to a reduction in the rate of interest. This is the “liquidity effect” and there is ample evidence for it.

A reduction in the rate of interest stimulates investment and discourages saving. Maybe that could do with a ceteris paribus qualification or something, but really that seems like nit-picking. We’re talking about a hyperinflation, so it’s not as if there is any chance of a lower nominal interest rate being neutralised by lower inflation.

A stimulus to investment which also discourages saving leads to an increase in activity. Since the economy under discussion was operating at something like full capacity, it might have been better to stress that much of the “activity” consisted of frantic shopping, but again that’s surely obvious in context.

There is no evidence whatever that events in Germany followed the above sequence. Her paper discusses the sequence of events: “the violent inflation which set in in the second half of 1921 was inaugurated by a sudden fall in the mark exchange (May, 1921, 15 marks = 1 gold mark; November, 1921, 63 marks = 1 gold mark).” Whether this was caused by reparations payments or some other shock doesn’t much matter. The important point is that the inflation was transmitted through the prices of tradable goods, while other prices and wages lagged. “Moreover, the geographical diffusion of prices supports the argument, the movement spreading from the great ports and commercial centres to the interior of the country (p. 135).” Maybe modern scholars disagree with her about this but the few sources I’ve looked at seem to support her account.

In short, Joan Robinson’s pronouncement seems fairly innocuous to me. What’s bothering Scott Sumner? Let’s sample his comments:

1. "So easy money couldn’t possibly have caused the German hyperinflation because German interest rates were not very low. And everyone knows that easy money is associated with low interest rates. I won’t insult the intelligence of my readers by explaining what is wrong with her reasoning."

2. "So money couldn’t have been easy when German prices were soaring, because nominal interest rates weren’t that low."

3."...people like to mock how Joan Robinson said (in 1938) that easy money couldn’t have caused the German hyperinflation, after all, interest rates weren’t low."

Notice that in all these cases Scott Sumner attributes to Joan Robinson the view that interest rates were not low, or not very low, or all that low. Clearly she doesn’t express that view in the two sentences he quoted. Does she do so elsewhere in the paper? She does discuss interest rate policy. She says that the Reichsbank might have curbed the inflation at an early stage if it had followed a stricter policy:
If the quantity of money had not expanded it may be supposed that the rate of interest would have been driven up, investment impeded, and saving encouraged, so that unemployment would have appeared again and the rise in money wages would have been brought to an end. But, in fact, the budget deficit, the policy of the Reichsbank in ‘meeting the needs of trade’ and the various official and unofficial supplementary currencies which were improvised, combined to meet the demand for money, the short rate of interest did not begin to rise appreciably till July, 1922, and no obstacle was put in the way of the inflation. It is true that in 1923 the short rate of interest rose to such heights that loans were taken at 20 per cent per diem (though the maximum reached by the Reichsbank discount rate was only 90 per cent per annum). But by this time the expectation of a continued rise in prices was so strong that it was impossible for high interest rates to discourage entrepreneurs from investment or to restore the motive for saving to the ordinary public. Thus the fact that high rates did not stop the inflation in 1923 cannot prove that they would have been equally powerless in 1921. The champions of the quantity theory, therefore, may reasonably contend that it was the increase in the quantity of money which permitted the inflation to take place. [Emphasis mine]
There you have it.