The Relationship Between Money and Prices: Some Historical Evidence Reconsidered

This article describes a debate about the validity of the quantity theory of money and offers further evidence against it. The evidence is primarily from the North American colonies of Virginia, New York, and Pennsylvania and regards the issue of measuring the money supply. Studies have shown that changes in colonial money and inflation are inconsistent with the quantity theory. Some have argued that those studies measure money wrong: specie belongs in the measure because the colonies were on a fixed exchange rate system with Britain; changes in colonial paper money were offset by specie flows. When specie is counted, the quantity theory stands. This study responds with evidence that the critics are wrong: the colonies had no such fixed exchange rate regime, and movements in the stock of colonial paper currency cannot have been offset by specie flows. The views expressed herein are those of the author and not necessarily those of the Federal Reserve Bank of Minneapolis or the Federal Reserve System. [James] Madison entertained an intelligent view of the causes affecting the value of paper money. “It depends on the credit of the State issuing it, and on the time of its redemption; and is no otherwise affected by the quantity than as the quantity may be supposed to endanger or postpone the redemption.” —Albert Bolles, 1884 Central to most thinking about monetary theory and monetary policy is a version of the quantity theory of money. According to Lucas (1980, p. 1005), “two central implications of the quantity theory of money . . . [are] that a given change in the rate of change in the quantity of money induces (i) an equal change in the rate of price inflation; and (ii) an equal change in nominal rates of interest.” Lucas goes on to state (p. 1005) that these propositions “possess a combination of theoretical coherence and empirical verification shared by no other propositions in monetary economics.” While Lucas does not state what this empirical verification consists of, it seems safe to assume that it includes the findings of Friedman and Schwartz (1963, p. 676) that, since the Civil War, “changes in the behavior of the money stock have been closely associated with changes in economic activity, money income, and prices . . . . The interrelation between monetary and economic change has been highly stable.” It also likely includes the claim of Friedman (1960, p. 2) that, since World War II, “no country succeeded in stemming inflation without adopting measures directed at restraining the growth of the stock of money,” as well as the conclusion of Schwartz (1973, p. 264) that, at least since the time of Alexander the Great, “long-run price changes consistently parallel the monetary changes, with one exception for England in the sixteenth century.” These conclusions and Lucas’ propositions have been so firmly held by economists that they are often built into (rather than derived from) economic models. They also influence everyday thinking about the role of the Federal Reserve System, in that the central bank is charged (under this view) with preventing secular inflation, increases in interest rates, and so on. However, despite Lucas’ assertions about theoretical coherence and empirical verification, the quantity theory propositions described above have come under sharp theoretical and empirical scrutiny. On theoretical grounds, the asserted effects of monetary changes on prices and inflation have been challenged by Wallace (1981) and by Sargent and me (1986, 1987). In particular, we have produced economic models in which the consequences of monetary changes, even for nominal magnitudes, depend crucially on how such changes are accomplished. Loosely speaking, our work directs economic observers to examine the consolidated balance sheet of a nation’s treasury and central bank. Monetary changes that affect total liabilities on this consolidated balance sheet (without compensating changes in assets) will have the effects Lucas predicts. However, monetary changes that do not result in changes in this consolidated balance sheet can actually be irrelevant for prices and interest rates. To illustrate this point, Sargent and I (1987) provide an example of a once-andfor-all change in the money stock that produces no changes in prices or interest rates. These Wallace/Sargent-Smith results have some quite dramatic implications for the conduct of monetary policy. One is that open market operations accomplished with fiscal policy held constant (that occur with the consolidated balance sheet of the treasury and the central bank unaltered) have no effect on prices. Another implication is that government attempts tomanage foreignexchange rates can be effective only if accompanied by fiscal actions that have redistributiveconsequences. (SeeSargentandSmith1986.) Of course, if these theoretical results lack empirical verification, as Lucas implicitly suggests, then the results are rightly not of great interest to economic policymakers or monetary economists. However, at least on the surface, there appears to exist quite strong empirical support for them. For instance, Sargent (1982), Bomberger and Makinen (1983), Makinen (1984), Smith (1984; 1985a,b), Wicker (1985), White (1986), and Imrohoroglu (1987) provide evidence of a number of episodes in which very large monetary changes occur (in some cases, over quite long periods) and in which price levels and currency values are extremely stable. In most of these cases, it is fairly apparent that the monetary changes were accomplished without significant effects on the consolidated balance sheet of the relevant treasury and central bank. These episodes thus provide a wide range of empirical support for the Wallace/Sargent-Smith view and against the Lucas version of the quantity theory. That more such evidence will appear seems likely as well, since Redish (1985) suggests the existence of similar evidence for periods in early Canadian history, for instance. Given the cumulation of this kind of evidence and its important implications for monetary economics, it seems appropriate to briefly review the findings of some of this literature, as well as some reactions to these findings. Sargent (1982) has examined the experiences of four European economies as they emerged from hyperinflations after World War I. One of his findings is that each of these economies experienced extremely rapid growth in its money supply for some time after the price level had been stabilized. Post-hyperinflation Germany, for instance, saw its money supply increase by a factor of nearly four in the year following price stabilization. Sargent argues that these monetary changes were accomplished without altering the net balance sheet positions of the relevant treasury and central bank. Thus, these episodes support the propositions derived by Wallace, Sargent, and me. Subsequently, Bomberger and Makinen (1983) and Makinen (1984) have accumulated similar evidence based on the experiences of other countries emerging from hyperinflations. The evidence presented by Sargent (1982) is not universally regarded as being inconsistent with the quantity theory, however. Under one interpretation, the hyperinflations essentially destroyed the monetary systems of these economies, which were then simply remonetizing after the stabilizations.Another interpretation is that the reforms that accompanied price stabilization required some adjustment in the expectations of agents: Changes in expectations over time increased the demand for money, preventing increases in the money supply from producing inflation. Thus, further presentation of evidence is called for. I have presented an array of evidence consistent with Sargent’s (Smith 1984 and 1985a,b). Moreover, much of this evidence is not readily explained by appealing to changes in monetary systems or expectations. In particular, many researchers have observed that, in the British North American colonies, there were several episodes in which the money supply apparently changed dramatically over long periods. These changes were quite often not accompanied by any price level movements. For instance, in 1760–70, the colony of New York reduced its per capita currency supply 86 percent, but available evidence indicates that the price level fell only 3 percent over the same period. This kind of experience was repeated in different colonies and different time periods. In addition, these monetary changes were accomplished with only minor changes in the (consolidated) government balance sheet. Hence, these observations are consistent with the Wallace/Sargent-Smith propositions and inconsistent with the quantity theory. Moreover, since no regime changes (or monetary reforms) had occurred, the counterarguments that are available against Sargent’s interpretation of events are not available in the colonial context. Limitations in the kind of data that are available for the colonial period have, however, led to some questioning of this interpretation of events. Specifically, the only data that are available on colonial money supplies are measures of the amount of paper money issued by each colony. I have related this money supply measure to movements in colonial prices and exchange rates, finding that in many cases large money supply movements produced no changes in price levels or currency values. Still, in addition to their own paper currencies, the colonies had stocks of specie (coins) that circulated within them. Since no data on colonial specie stocks exist, any money supply measures necessarily omit this component of the money stock. In fact, I have discussed this omission (in Smith 1984 and 1985a,b) and presented some arguments about why the inability to measure the quantity of specie is unlikely to be of concern in interpreting the colonial evidence. These arguments center on indications that the specie stock was generally a fairly small component of the colonial money supply. Subsequent work by