The U.S. Dollar During WWI and the
Recession of 1920
March 25, 2012
This week we will look at a funny and little-known episode in the
history of the U.S. dollar and Federal Reserve, notably the wartime
years of WWI and the Recession of 1920 that immediately followed.
The Federal Reserve was, of course, signed into law in 1913, just in
time for WWI. There was something of a liquidity crisis upon the
outbreak of war, but that was handled by the existing "bank
clearinghouse system" which was legitimized by the Aldrich-Vreeland
Act of 1908. The Federal Reserve system had not yet been set up, and
was not operational.
The U.S. entered the war in April 1917. This coincided with the time
when the Federal Reserve banks became operational. Virtually from
day one, the Fed was pressured into "keeping interest rates low" for
the flood of government borrowing that was happening to finance the
war. Here's how Richard Timberlake describes it, in his wonderful
and essential book Monetary
Policy in the United States: an Intellectual and Political History:
The Federal Reserve banks were just
about operational when the United States became involved in World
War I. Almost immediately, the U.S. Treasury Department asserted
its dominance. The discount policy of the Board was to maintain
rates "in harmony with the low interest rates borne by the
Government loans," stated the Board's annual report for 1917. Even
more pointed was a section in the annual report for 1918, ...
which began: "The discount policy of the Board has necessarily
been coordinated . . . the Treasury requirements and policies,
which in turn have been governed by demands made upon the
Treasury for war purposes." Again, in 1918 the annual report
admitted that discount rates were based upon rates currently borne
by government securities, and "must for the time being continue to
be fixed with regard to Treasury requirements." The Board
"recognized its duty to cooperate unreservedly with the Government
to provide funds needed for war."
The report for 1919 excused the Fed's subservience to the Treasury
by asserting that no higher rate structure would have been
sufficient to persuade the securities market to absorb the
Treasury issues: "It was necessary to cooperate with the Treasury
in every way," the report claimed, "to facilitate first the sale
of Government securities and then their absorbtion by investors."
As you might imagine, this involved printing money. Here's what
happened to the monetary base in the U.S. during that time. This
data is from Milton Friedman's Monetary
History of the United States, 1867-1960:
Actually, something funny starts to happen right around the
beginning of 1915. With a little more sleuthing, we might figure out
what it is. For today, we'll just point it out as something to be
filled in later.
Here's what it looks like when we continue the series into 1940.
This data is from the St. Louis Fed:
There are a couple interesting things on this chart. For one, hardly
anything at all happens during the 1920s, despite the booming
economy. As the Great Depression begins at the end of 1929, the
monetary base begins to expand by quite a bit. This is what you'd
expect to see, as banks would be under pressure of withdrawals and
holding paper banknotes was probably becoming more popular. The
"money supply shrank" stories you've heard are fiction -- they
represent M2, which is mostly bank deposits, not base money.
Getting back to our original story ...
During the wartime years, there is an embargo on gold in the U.S.,
so nobody can buy, sell or export gold. The Bank of England had
effectively suspended redeemability of its banknotes into gold in
August 1914, rendering the British pound a floating currency. The
pound was also printed to finance the British government's war
effort. At the beginning of the war in 1914, the estimated pound
base money supply was around £200 million, according to the
book Sterling: the History of a
Currency. By mid-1919, the pound base money supply had
expanded to around £550 million.
Here is a graph of the gold bullion reserves of U.S. banks:
We can see bullion reserves rise in the beginning. I interpret this
as a rush of European interest in holding dollars and dollar assets,
as their governments were all at war with each other. Then, there is
an odd flatline around mid-1917-mid 1919, and a brief decline. The
flatline is probably the gold embargo period.
Here's a graph of U.S. gold imports and exports. Negative denotes an
We see a brief outflow at the outbreak of war in Europe, then an
inflow, which once again suggests Europeans rushing to hold dollars.
This switches to an outflow in mid-1917 and goes negative. Probably,
redemptions of banknotes for gold are happening here. From mid-1918
to mid-1919, there's a flatline, which is probably the embargo on
exports of gold. When the embargo is lifted after the end of the
war, there's a big gold outflow. The Fed and banks react by
shrinking the monetary base (above), which turns the outflow to an
inflow. The source for all these is the Fed's Banking and Monetary Statistics,
available at the St. Louis Fed FRASER archive. (I'm making the
sources clear so you are able to make graphs like these yourself if
So, basically what happened is the Fed was pressured by the Treasury
into facilitating the funding of the war effort, which involved
printing too much money. This led to gold redemptions, which were
stopped by government edict. After the war, there was a big rush of
gold redemptions, signalling that the dollar was much too weak
compared to its gold parity. The Fed and the banks (the Fed was not
yet the monopoly issuer of currency then) reacted by shrinking the
The combination of the demobilization of soldiers, the decrease in
war-related demand for goods, and the monetary deflation resulted in
the recession of 1920, in which prices fell considerably. This graph
is titled "Consumer Price Index," but I think it is really the
Bureau of Labor Statistics Wholesale Price Index, which is more of a
broad commodity index, and not really comparable to the CPI which
wasn't compiled until 1940. However, this index still stabilizes at
a much higher level than before the war began. (The BLS WPI began to
be compiled in 1919, but data was gathered back to 1913, which is
exactly when this time series starts.)
There are a few more things to say about this time period, but I
think I will save that for later.