![]()
CHAPTER 1
Introduction
1.1Purpose of the study
The purpose of this study is to present and empirically test a coherent framework explaining the secular behavior of the income velocity of money. Annual time series of velocity for a number of countries over the last century display a U-shaped pattern. In addition, in the post— World War II period, velocity of a large number of countries, classified by level of economic development, exhibits a similar U-shaped pattern.
Existing theories can explain each section of the velocity curve — the falling, flat, and rising parts — but the overall pattern is not consistent with any one prevailing theory. In this book we offer a com-prehensive explanation for velocity's secular behavior. The explanation is to a large degree an extension and development of the approach of the Swedish economist Knut Wicksell, who stressed the role of substitution between monetary assets. This approach, which emphasizes institutional variables, is then incorporated into the arguments of the traditional long-run money demand (or velocity) function, namely a scale variable and rates of return on assets alternative to money.
Four types of empirical evidence provide strong confirmation of our approach: econometric studies of the long-run velocity function for five countries — the United States, Canada, the United Kingdom, Sweden, and Norway; a cross-section study of about eighty countries in the postwar period; a case study of the Swedish monetization process in the fifty years preceding World War I; and an examination of the time series properties of velocity.
In contrast to standard macrostudies of the demand for money, or velocity, we focus on institutional factors. For two reasons, these factors have rarely been incorporated in studies of the demand for money. First, since most studies deal with shorter time periods, they assume (implicitly or explicitly) that institutional factors are constant. Second, empirical measures of institutional developments are difficult to construct.
This study draws on both monetary theory and economic history. Although our approach is a macro one, the empirical regularities we highlight lend themselves to theorizing on the microfoundations of the exchange mechanism and of the monetary system.
1.2Plan of the study
Chapter 2 presents evidence of a U-shaped pattern of velocity for a number of industrial economies in the past 100 years and reviews existing theories of the long-run behavior of velocity. No one theory is able to explain the secular pattern described. Chapter 3 presents a coherent explanation of the U-shaped pattern based on the concept of substitution between different monetary assets induced by institutional forces. This explanation is inspired by the work of Knut Wicksell.
Chapter 4 contains a series of econometric tests of the hypothesis presented in Chapter 3. The secular fall in velocity is explained by monetization and the growth of banking, the secular rise by financial sophistication and improved economic stability. Evidence for each of five countries treated separately and then pooled and treated as one entity is examined here. Chapter 5 studies in greater detail than Chapter 4 the Swedish monetization process and its effects on velocity over the period 1871 — 1913. Evidence on the monetization process is derived from the growth of commercial banking, changes in labor market contracts, and developments in commodity markets.
Chapter 6 continues the investigation of the hypothesis in Chapter 3 using IFS data for about eighty countries for the period since the 1950s. According to our explanation, for low-income countries velocity should fall, for high-income countries it should rise, and for middle-income countries it should be roughly constant. With two major exceptions, this pattern is confirmed. First, velocity is falling in Italy, Germany, and Japan but rises in all other high-income countries. Second, velocity rises in high-inflation countries such as Brazil, Israel, and Chile when, according to our approach, velocity for these upper middle-income countries is expected to be roughly constant.
Chapter 7 explores the time series properties of velocity, demonstrating that velocity over the past century for the five countries examined in Chapter 4 displays a random walk without drift. A random walk implies that changes in velocity cannot be predicted from past velocity changes, but it does not imply that past changes in other variables deemed important determinants of velocity cannot be used in prediction. Our finding that previous changes in the institutional and traditional determinants of velocity isolated in Chapter 4 can be used to predict future changes in velocity reconciles the stochastic behavior of velocity with our approach. Finally, Chapter 8 summarizes the results and discusses their implications.
![]()
CHAPTER 2
Evidence and theories of the long-run behavior of velocity
2.1The secular picture
The long-run behavior of the income velocity of money (V2) from the 1870s to the 1970s in the United States, Canada, the United Kingdom, Sweden, and Norway is displayed in Charts 2.1 through 2.5. The velocity curves are calculated using a broad definition of money, M2, where M2 is defined as the sum of currency, demand, and time deposits. M2 is the only monetary aggregate available over the entire data period and for the five countries covered by the charts. National income is measured by NNP at market prices for the United States and the United Kingdom, by GNP at market prices for Canada, and by GDP at market prices for Sweden and Norway. (For sources of the data used see Appendix 1A.)
The five charts show that velocity has exhibited a secular U-shaped pattern over the past century in the five countries, most prominently in Sweden, Norway, Canada, and the United States. The dating of the turnaround differs, however, across countries. There are also considerable cyclical fluctuations in the velocity curves; specifically, the depressions of the 1920s and 1930s are commonly reflected in substantial declines in velocity.
Velocity was falling in the United States prior to the mid-1940s, when the turnaround occurred (Chart 2.1). It has displayed an upward trend since then. The depression in the 1930s is associated with a sharp fall in velocity. The Canadian curve (Chart 2.2) has great similarities with the American one, with a turnaround in the 1940s after a sharp cyclical downturn around 1930. For the United Kingdom, velocity falls from around 1910 onward, with a turnaround occurring in the mid-1940s (Chart 2.3). Prior to World War I, however, there is a weak upward trend.1 The U-shaped pattern is clearest for Sweden (Chart 2.4); the bottom was reached in 1922, which was a year of deep depression. Norway (Chart 2.5) and Sweden exhibit similar patterns until 1939, when World War II interrupts the data series for the former, allowing the conjecture that had Norway not been involved in the war, its velocity would have continued to behave in the Swedish mode.
Chart 2.1.Income velocity of money (V2) in the United States, 1870–1975. (Source: Appendix 1A.)
Chart 2.2.Income velocity of money (V2) in Canada, 1870–1975. (Source: Appendix 1A.)
Chart 2.3.Income velocity of money (V2) in the United Kingdom, 1870—1975. (Source: Appendix 1A.)
Chart 2.4.Income velocity of money (V2) in Sweden, 1871–1971. (Source: Appendix 1A.)
Chart 2.5.Income velocity of money (V2) in Norway, 1870–1974. (Source: Appendix 1A.)
To sum up, for these five countries, for which good data covering long periods are available, velocity follows a U-shaped curve.
Lack of data generally prevents the study of the secular experience of most other countries. However, acceptable data on velocity exist for a few countries other than the five just discussed. Charts 2.6 through 2.13 cover the experience of Australia, Denmark, Finland, France, Germany, Holland, Italy, and Japan. With the exception of France, an M2 definition of the money stock has been used to maintain comparability with the first five charts. (For the data used see Appendix 1A.)
The Danish velocity curve (Chart 2.6) displays a pattern similar to the Swedish and Norwegian curves. The fall in velocity is pronounced prior to World War I, but the postwar rise is limited. The Finnish data suggest a fall in velocity prior to the liberation of Finland from Russia in 1919, a rapid rise in the early 1920s followed by a fall, and a secular rise after the mid-1940s interrupted by a fall in the 1950s (Chart 2.7). The income velocity of money in Germany falls prior to 1914 (Chart 2.8). The two world wars are excluded due to lack of reliable data. After World War II, velocity shows a downward trend. In Holland, income velocity exhibits a pronounced secular decline until World War I, a postwar rise followed by a decline until World War II, and a strong upward trend after World War II (Chart 2.9). Sections of this pattern conform closely to sections for other countries.
Chart 2.6.Income velocity of money (V2) in Denmark, 1870–1974. (Source: Appendix 1A.)
Italy shows a continuous fall in velocity for the entire period 1870–1975 interrupted by a sharp rise after World War II (Chart 2.10). The velocity curve of Japan (Chart 2.11) displays some similarity to those of Germany and Italy. It falls secularly from the 1890s to the 1980s interrupted by a shift upward of the curve following World War II. (For further discussion of the post–World War II pattern of velocity of these three countries, see Chapter 6.) The income velocity of money in Australia (Chart 2.12) suggests a U shape with a triple bottom: a fall until the mid-1890s, a rise until the 1920s, then a fall to 1932, a rise until the war, then a wartime fall, and a secular upward trend thereafter. The income velocity of France during the period 1870—1969, based on an M1 definition (since consistent M2 series were not available for most of the period), displays a falling trend prior to World War I, followed by a brief rise in the early 1920s and then a decline in the rest of the ...
