Journal of International
Money and Finance (1986),
5, 157-l
Comments
on Garber
STEPHEN
Department
of Economics,
59
MCCAFFERTY
Ohio State University,
Columbus OH 43210,
USA
Peter Garber’s paper has carefully documented flows of US paper currency into and out of the United States during the interwar years. As a result, his paper proposes alternative for currency held monetary
series to those provided by Friedman and Schwartz (1963) by the domestic public and the domestically held US
base for the period
Schwartz’s
attempt
to correct
in question. their
W’hile it is true that Friedman
series for international
and coin, they attempt no such adjustment for paper the Garber paper represents a very exhaustive review
movements
and
of gold
currency flows. As such of the archival evidence
and some very sensible speculation to fill in for what cannot be documented. When I first saw the title and introduction to this paper, I expected that Garber part
might
forcefully
the
conventional
of
discussion
avoids
be advancing
an alternative
wisdom
such speculation,
about although
the
way of interpreting
Great
some
at least
Depression.
Garber’s
of his findings
are subject
to that type of interpretation. Garber cites monthly Federal Reserve data showing significant shipments of US paper currency to Europe in September and October 1930. While such movements would not affect the worldwide US monetary base or measurements of the domestically held base, such shipments would lower the monetary base actually held in the United States. This means that the time series for the American
held monetary
base may not have been as
smooth as we have thought and that these currency shipments may therefore have played a role in the banking crises and related monetary factors in 1930 and 1931.
This interpretation
not offsetting these Garber also notes
might
lead one to fault the Federal
paper currency movements. that huge net receipts of paper
currency
Reserve
System
for
were reported
in
1932 and 1933. Such receipts would suggest that Friedman and Schwartz’s estimates of the American held monetary base for this period, which show a decline, may also be misleading. If we factor in these previously undetected currency relatively
inflows, constant
the American held monetary base may actually have been in the later period. If this were the case, much of the criticism
of the US Federal Reserve System for allowing the monetary base to fall in 1933 might be unfounded. While Garber considers a number of differing sets of assumptions to translate his raw data on paper money flows into alternative measures of the American held monetary base most do not differ significantly enough from those of Friedman and Schwartz to provide much support for conjectures such as those advanced above. In this spirit Garber only tentatively suggests a possibly
significant
rise in US paper
0261-5606/86/Ol/Sl57-03803.00
0
dollar
holdings
in Europe
1986 Butterworth& Co (Publishers)Ltd
for the year
S158
Comments
on Carher
1932, which would be the counterpart of an undetected drop in the American held monetary base for that year. Almost all of Garber’s data and his resulting analysis are based on Federal Reserve data on paper currency shipments and receipts from Europe reported by selected New York banks from 1923-1964. These banks were thought to represent the major shippers of such currency and most of the variability of these currency flows were thought to be European in nature. Such net flow data would need to be combined with either an initial or terminal European stock to come up with a continuous series on the European held stock of US paper currency. Garber first assumes that the European stock, which is believed to have grown substantially prior to 1923 as the result of European hyperinflations, gradually declined to approximately zero by 1938. If the Federal Reserve data then exactly measure the net flow of currency from Europe, his first set of reconstructed figures for the domestically held US currency and monetary base will be accurate. However, it should be remembered that such figures rely on a correct determination of the final stock and an exact measurement of the net flows. If both gross flows that are used to measure the net flows are each only an equal fraction of the correct totals there will be a cumulative error in the reconstructed figures. The same sort of cumulative error will occur if either shipments or receipts are systematically underreported. However, the general time series pattern of such a series will be reflective of the true series as long as the data are qualitatively correct. The series that Garber reconstructs in this manner are not very different from those reported by Friedman and Schwartz. Garber’s other series are constructed based on a number of alternative hypotheses which are, as Garber admits, designed to be more speculative in nature. Garber assumes that the net receipts of currency are proportional to the outstanding stock and that the US currency-deposit ratio remains constant. Garber’s first assumption is consistent with a fixed velocity of circulation of US currency in Europe combined with a fixed proportion of European US currency spending flows being repatriated to the United States. It is this set of assumptions which leads Garber to entertain the possibility of a dramatic rise in American held currency in Europe in 1932. The time series properties of these final series rely in a crucial way on Garber’s assumptions of a fixed velocity of circulation of US currency in Europe and that the net flow of dollars back to the United States is proportional to the rate at which they are spent. This means that in years in which the flow of dollars back to the United States rises dramatically, such as 1932, his estimate will reflect an increase in the stock of US dollars held in Europe. This can only occur if there is simultaneously a large unreported flow of dollars into Europe which would be inconsistent with Garber’s assumption that the Federal Reserve data is in fact accurate. It is this particular implication of Garber’s assumptions which cause his final reconstructed data series to conflict with a more intuitive discussion of the implications of the data for the years 1932 and 1933. Casual inspection of the data suggests possible problems in the years 19321934 in any event. In these years currency shipments to Europe fell to almost exactly zero. However, as Garber points out, these were years of strict capital controls in Europe. A ban on US shipments of currency also began in March
s159
STEPHEN MCCAFFERTY
1933.
Therefore
even if the enforcement
of such measures
we would certainly expect such shipments to no longer banking system. Therefore any reconstruction based data may be somewhat suspect. While I am always fascinated by the possibility of the data and historical of the likelihood
record,
was not very strict,
be reported through the on the Federal Reserve
of alternative
interpretations
in this case I have to be somewhat
of a significant
impact
on the American
suspicious
held monetary
due to foreign paper currency substitution. The currency rather imprecisely defines the term to include everything
base
substitution literature from a borrowed unit
of account
to inclusion of foreign interest rates in domestic demand for money the present paper very narrowly defines currency functions. However, substitution in the form of a foreign demand for components of the US monetary base. I would argue that such a phenomenon is unlikely to be quantitatively important and therefore one is likely, as Garber generally finds, to be hard pressed to find examples when currency substitution considerations lead to dramatic differences in ourinterpretation of major historical events such as the Great Depression. Unit-of-account currency inflation, holdings.
substitution,
which
occurs
during
periods
of high
is only an aid in arithmetic and involves no actual foreign currency Store-of-value currency substitution would denote a demand for
protection against political and exchange risk. However, such a demand would, except in the case of very effective exchange controls, normally take the form of a demand finally
for
foreign
left with
interest
bearing
medium-of-exchange
substitute
foreign
episodes
did occur
monies
to avoid
in Europe
assets,
not
currency
during
the shoe
foreign
currency.
substitution. leather
the period
Here
costs
Garber
We
of inflation. studied
are
individuals Such
and the US
dollar was often the currency of choice. However, while the demand for US dollars as a means of payment may have risen dramatically, this does not automatically translate into a significant increase in the demand for US currency, particularly paper currency. As long as the
foreign
crisis
of confidence
simply
reflects
a lack
of confidence
in the
domestic monetary authority’s ability to provide stable domestic purchasing power, it is likely that the domestic banking system would be able to provide a means
of foreign-currency-denominated
rather
on foreign-denominated
deposits.
payments
based
not on currency
It is only in cases in which
but
banking
regulations are unusually successful or in which a crisis of confidence encompasses the domestic banking system as well as the stability of the domestic price level that we should expect the phenomenon of currency substitution to be reflected in a significant demand for foreign base money. Therefore
we would
expect
that only in the most unusual
of circumstances
like
the one that Garber considers in his most speculative view of the data that currency substitution should play a very important role in our interpretation of economic events. Garber’s paper provides an important set of possible improvements to the Friedman and Schwartz data. He has most carefully and exhaustively uncovered evidence on previously ignored international movements of US paper currency and provided interesting sources of additional speculation about the monetary factors surrounding the Great Depression.