Academic journal article Journal of Money, Credit & Banking

The Exchange Rate and the U.S. Wage Process: An Intensive Empirical Investigation

Academic journal article Journal of Money, Credit & Banking

The Exchange Rate and the U.S. Wage Process: An Intensive Empirical Investigation

Article excerpt

The significant slowdown in wage growth and the number and frequency of concessions that characterized the U.S. labor market in the 1980s have generated a large literature attempting to explain their causes and their persistence even in the face of low and declining unemployment. One of the explanations that has received particular attention is based on international competiveness forces. According to Dornbusch (1987, p. 14), "wages respond to the competitive pressure of an appreciation or depreciation in affected industries" directly, in addition to the indirect response caused by changes in the cost of living.

Dornbusch and Fischer (1986) provided evidence that wage growth in U.S. manufacturing was significantly influenced by the exchange rate in the direction suggested by their hypothesis. Additional research on the issue, however, (for example, Abowd and Freeman 1987, Blanchard 1987, Vroman and Abowd 1988, Bell, Hall, and Hayes 1988) has not led to conclusive results.(1) An alternative interpretation of the wage disinflation is that it was due to a change in "the wage norm." According to Mitchell (1985, p. 584), "as the numbers of industries affected [by concessions] grew, the conventional explanations - deregulation, foreign competition, general economic distress - became harder to credit, and the alternative hypothesis of a demonstration effect became more difficult to reject." The evidence for a shift in the wage norm is also mixed.

This paper contributes to this literature through an intensive investigation of several empirical issues. First, it attempts to distinguish between a slowdown in wage growth caused by a shift in the "wage norm" and that caused by foreign competition. Second, it modifies the sample used by Dornbusch and Fischer and others in two respects. It first excludes the fixed exchange rate period because, as Obstfeld (1985, p. 426) has observed, the process by which wages are set may depend on the exchange rate regime. A Chow test indeed indicates that the fixed exchange rate period should be excluded. Further, the main issue investigated here is not as meaningful in the fixed exchange rate period during which changes in the nominal or real value of the U.S. dollar were relatively small compared to those experienced in the post-1972 period. It then extends this sample to 1987:3, a period during which the dollar reversed all of its previous rise, thus allowing for a more precise evaluation of the exchange rate's role in wage formation. Third, it tests for "competitiveness effects" not only in the manufacturing sector but also in sectors such as services and construction, which have traditionally been thought of as immune to foreign competition, as well as on the economy-wide average private hourly earnings. Further, this paper discriminates between two possible interpretations of a significant exchange rate effect. According to the asset view of exchange rates, the exchange rate may be capturing forward-looking behavior rather than competitiveness effects as implied by the Phillips-curve specification. None of the previous papers has addressed this issue. I introduce another forward-looking variable to discriminate between the two. Fourth, it provides some evidence concerning the stability of the estimated exchange rate coefficients during the dollar's two recent major swings.

The analysis is carried out in a Phillips curve specification. I employ this specification for two reasons: First, the result of this paper can be directly compared to the previous literature that exclusively uses the Phillips curve approach. Attempts to expand the scope of the paper by adopting an error-correction model were not fruitful because the four variables included in the model do not seem to possess a common cointegrating vector. The failure to find cointegration is not necessarily damaging to the relationships uncovered below. Second, a more general VAR system exploring the main issue of this paper is employed in Cha and Himarios (1990). …

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.