, Income and Substitution: Evidence from Three Countries

Dallas S. Batten and IL W. Hafer /1 /‘ll, NUMBER of studies have demonstrated a r’ela- and economic activity (income I in the United States, tively close empirical relationship between changes in Germany and Japan within the framework of a simple atransaction-based measure of money and changes in reduced-form model. l’hese countries were chosen nominal income. This relationship, found for a variety primarily because they are the most influential in of economies, suggests that monetary policymakers international financial markets; indeed, some have can directly influence the path of nominal income suggested that their actions should 4 over time by changing the growth of the domestic be coordinated. A reduced-form model is used, be- money stock.’ cause it is a convenient for’m for testing the relative impact of specific variables on economic activity. A It has been argued recently, however, that the rela- sufficient condition for currency substitution to be tionship between domestic money growth and eco- important is that some measure offoreign influence — nomic activity may be affected by external factors! If money growth or movements — have a domestic residents consider domestic and foreign significant impact on domestic income after- account- or other financial assets) as relatively close ing for the impact of domestic money growth! substitutes, for example, then changes in relative pref- erences for domestic and foreign assets will motivate them to reallocate their portfolios. This portfolio ad- justment will affect the domestic demand for all as- sets, including domestic money.’ This hypothesis, known as currency substitution, suggests that, if the In a world of freely floating exchange rates, domes- demand for domestic money is dependent inter a/ia tic monetary authorities theoretically are insulated on external factors, domestic money growth may not from monetary shocks from abroad. Because mone- affect domestic economic activity to the degree antici- tary authorities have no obligation to maintain their pated by policvmakers. currencies’ foreign exchange value, an expansion or contr’action of one country’s money supply does not This article tests whether currency substitution has necessitate automatic policy reactions by other na- affected the relationship between domestic money tions. Instead, exchange rates fluctuate in response to relative movements in money supplies. Thus, mone- tary actions in one country do not necessarily impinge Dallas S. Batten and R, W. Hafer areresearch officers at the Bank of St. Louis. Paul G. Christopher provided research on the policy actions of another; each country is able ass/stance. The authors would like to thank Jeff Bergstrand for his to pursue its own domestic policy program. comments. ‘See, for example, Batten and Hater (1983), ‘This policy coordination scheme is attributed to McKinnon (1984). ‘SeeMiles (1978), Brittamn (1981) and McKmnnon (1982). ‘This condition is only sufficient in that foreign influences may affect ‘Within this context, one of the initial approaches to the investigation domestic activity within the structural econometric specification, but of currency substitution was within a money demand framework. may not be identifiable in the reduced-form specification that we See Batten and Hater(1984a) andthe referencescontained therein. have employed. FEDERAL RESERVE SANK OF St LOUIS MAY 1985

Some analysts argue that this insular’ proper-tv of cr-eases the power of the statistical tests and, therefore, floating exchar’rge rates br-eaks dowry when there is the confidence one can place in the results. currency substitution, If dr~mestic