Some Qualms About the Test of the Institutionalist Hypothesis of the Long‐run Behavior of Velocity

Abstract

This paper expresses econometric qualms about Bordo and Jonung's [1981] analysis of long-run velocity. They did not recognize that, for U S . and Canadian data, the log of velocity has a unit root. Hence, estimation of a log level regression may produce spurious regressions.When Bordo and Jonung's velocity equation is reestimated in rate of change form, permanent income is significant, contrary to their earlier conclusion. Moreover, using this approach gives a stronger result for one of the institutional variables in the velocity function, in the sense that the remaining variables become more significant. [1982] are the level of real per capita permanent income, the difference between the nominal yield on money and other financial assets, and the yield on physical assets. The effect of financial sophistication on demand for money has been recognized by Friedman and Schwartz [1982, sec. 6.31. However, they failed to capture the influence of financial sophistication on the demand for money through the use of proxy variables. Rather, they treated the effect of financial sophistication as an episode and allowed for its influence on velocity by adjusting the data. The task of empirically measuring financial sophistication was taken up by BJ [1981]. 537

Unit Root Testing: A Critique From Chaos Theory

“…Although hotly debated;' based upon the permanent incomellife cycle hypothesis and the rational expectations hypothesis, Hall (1978) has argued that personal consumption expenditures in the U.S. follow a random walk. Collectively, Mankiw and Shapiro (1985), Rose (1988), Raj and Siklos (1988), Cochrane (1988), and Tronzano (1987) have argued that various money, income, and price series, nominal and real interest rates, real exchange rates, and the log of velocity are also indistinguishable from random walks. This has led to a rising interest in random walks, and the tests by which they are recognized.…”

Some Qualms About the Test of the Institutionalist Hypothesis of the Long‐run Behavior of Velocity: Reply

“…Raj and Siklos [1988] reexamine our procedures and results in the light of some recent econometric advances. A log level form was used in order to study the trend behavior of velocity.…”