The Fed’s 1932 Bond-Buying Experiment 2: Automatic Adjustments Via Gold Conversion

We’ve been talking about the Federal Reserve’s 1932 bond-buying experiment.

May 21, 2017: The Fed’s 1932 Bond-Buying Experiment

Today, we will fill out the story with a little look at what Milton Friedman had to say about it in A Monetary History of the United States, 1867-1960 (1963), and also Allan Meltzer, in A History of the Federal Reserve, Volume I (2003). Here’s Friedman:

[President of the New York Federal Reserve George L.] Harrison told the executive committee of his directors on April 4 [1932] that apparently “the only way to forestall some sort of radical financial legislation by Congress, is to go further and faster with our own [bond-buying] program.” … Ogden L. Mills, since February 13, 1932, Secretary of the Treasury, who had all along been in favor of more extensive action, stated: “For a great central banking system to stand by with a 70% gold reserve without taking active steps in such a situation was almost inconeivable and almost unforgivable. The resources of the System should be put to work on a scale commensurate with the existing emergency.” … After the initial program was voted on April 12, the System bought $100 million of government securities per week for five weeks. At the May 17 meeting, the Conference again voted another $500 million open market purchase … In June, [Harrison] … suggested … that the purchases each week be geared to the maintenance of member bank excess reserves at a figure somewhere between $250 and $300 million …

By the end of June, … total purchases of $1 billion had offset a loss of $500 million in gold and a reduction of $400 million in discounts and bills bought, leaving a net increase of $100 million in Federal Reserve credit outstanding. To Owen D. Young, this meant that “most of our efforts had, in reality, served to check a contraction of credit rather than to stimulate an expansion of credit.” (pp. 384-386)

This is not a very detailed description, but a few items can be discerned. One is that the idea that the gold standard system was a self-regulating system, in which base money was essentially generated in an automatic fashion dependent upon the value relationship of the dollar vs. gold, does not seem to have been at the forefront of their minds. Even while the gold standard system was actually displaying its self-regulating character in the clearest possible terms, for all to see, they didn’t see it. Their bond purchases were cancelled out because they had to be cancelled out, by one mechanism or another, intentionally or not, to maintain the value of the dollar at its gold parity. In the end, gold conversion served as the ultimate backup for this process. There is no evidence of intentional “sterilization” here — the offsetting of intentional asset purchases by intentional asset sales elsewhere. They were dismayed to find that they did not produce the increase in bank reserves (of $250m-$300m as was their goal) that they intended. Actually, I think that at least some of the people involved understood this well enough, which is why they were opposed to the program at the beginning, and more opposed at the end, when it was proven that it could accomplish nothing. We are dealing with the limitations of the historian here: if Friedman was incapable of understanding this concept (he was incapable), then we can’t expect him to be able to coherently relate the arguments of those who could understand it.

Meltzer did not have very much to add to this narrative. He did have this interpretation, however:

Reductions in discounts and advances were the other main offset to purchases. These also remained far below the volume of open market purchases. During the peak purchase period, from March to July, member bank discounts declined $133 million, 14 percent of open market purchases. Together gold and discounts offset 64 percent of purchases. With discounts reduced and many foreign balances withdrawn, the offset would have fallen had purchases continued. Economic recovery would have reversed the gold flow and the reduction in member bank borrowing. (p. 372)

This is the typical Monetarist “more money makes everything better” argument. More bond purchases would have simply led to more gold outflows, as the automatic elements of the system (particularly convertibility) offset any attempts at discretionary tomfoolery. In practice, discount lending worked in a similar way — banks with reserve inflows from the bond purchases, creating more reserves than they wished to hold, would naturally pay off their borrowings from the Fed. What actually happened, in the second half of 1932, is that further reductions in discount lending were matched by a steady increase in base money demand, which was accommodated by large gold inflows. From a low of $3,654 million of gold in July 1932, gold reserves rose to $4,260 million in January 1933. The automatic elements of the gold standard system were perfectly capable of providing as much base money as the market demanded, in this case adding about $600 million to offset other factors.

We can guess pretty easily what would have been the result, in the first instance, if the Fed had purchased an additional $1,000 million of government bonds by January 1933. The combination of gold outflows and further reductions in discounting (there was still $255 million in January 1933) would have been about $400 million. But this is a bit glib: what probably would have happened is, people would have watched this spectacle, and by January 1933, they might have figured that the Fed had become the money-printing idiot that Meltzer wanted them to be. Then, they would start to panic.