Academic journal article Quarterly Journal of Business and Economics

Monetary Models of the Canadian-U.S. Exchange Rate: A Reexamination of Empirical Evidence, 1971-1986

Academic journal article Quarterly Journal of Business and Economics

Monetary Models of the Canadian-U.S. Exchange Rate: A Reexamination of Empirical Evidence, 1971-1986

Article excerpt

INTRODUCTION

Boothe (1983) argues that monetary models were appropriate for describing the Canadian foreign exchange market for the 1971-1978 period and that such models could have been used to speculate profitably on the Canadian-U.S. exchange rate over the 1974-1978 period. These results diverge sharply from other research (Meese and Rogoff, 1983; Backus, 1984) which finds little statistical evidence to support such results and determines that the Canadian-U.S. exchange rate is approximated more closely by a random walk model.(1) The finding of either a random walk or a monetary model representation of the exchange rate does not provide a sufficient test of the efficient market hypothesis (Bilson, 1981; Longworth, Boothe and Clinton, 1983; Levich, 1985). Nevertheless, it is surprising that such opposing conclusions about the Canadian-U.S. foreign exchange rate have not been reconciled. Whether the Canadian-U.S. exchange rate should be modeled as white noise or some other process is of interest equally to economists and to hommes d'affaires.

The purpose of this paper is to reexamine the data available for the 1970s and 1980s and to test whether there is support for a monetary model representation of the Canadian-U.S. exchange rate. The objectives are twofold: first, we estimate monetary models using quarterly data for the period 1971-1978, consistent with the work of Boothe (1983), and examine the statistical evidence to support the conclusion that variations in the Canadian-U.S. exchange rate are explained by these models. Second, the monetary models are re-estimated using alternative definitions of some variables, monthly observations, and alternative periods. We find that the estimated models are not robust to the data and that minor variations in sample definition cause significant changes in the estimated parameters of the models. Further, an examination of the time-series structure of the monetary variables indicates a nonstationarity condition which questions the validity of a monetary model representation of the Canadian-U.S. exchange rate.

THE MODELS

Monetary models of exchange rate determination start from an assumption of perfect capital mobility. Differences among models are based on choices regarding the applicability of three theoretical principles: purchasing power parity (PPP) (Dornbusch, 1976; Frenkel, 1980; Levich, 1985); real or nominal interest rate parity (Clendenning, 1970; Aliber, 1973; Frenkel and Levich, 1977); and the Fisher equation (Krueger, 1983).

The purchasing power parity doctrine and the interest rate parity theorem are used in model development to define equilibrium conditions. The former maintains the existence of an equilibrium relationship between the purchasing power of currencies over time. The latter maintains an equilibrium between the difference in the spot and equilibrium exchange rates(2) and the interest rate differential (nominal or real depending on inflationary expectations). With inflationary expectations, the Fisher equation must be used to set the real rate of interest equal to the nominal rate, less the expected rate of inflation. If real rates of interest are equalized between two countries through capital mobility, then nominal interest rate differentials must reflect differences in expected inflation rates. We will examine the most common monetary models.

In presenting the models, the following symbols are used: m is nominal money demand, p is the price level, y is real national income, pfx is the price of one U.S. dollar in Canadian dollars, r is the nominal interest rate, [Pi] is the expected inflation rate, [[Beta].sub.ij] represents regression coefficients, and [[Epsilon].sub.i] is a random error term. We define U.S. variables using superscript * and Canadian variables with no superscript. Lower case letters indicate a logarithmic transformation.

The purchasing power parity doctrine holds that the exchange rate (pfx) and the national price level will adjust to maintain a given currency's purchasing power across countries (Levich, 1985). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.