Relative price determination in the medium run: the influence of wages, productivity, and international prices.

• Author: Strauss, Jack

1. Introduction

   Balassa (1964) and Samuelson (1964) (B-S) provide the benchmark model for the determination of long-run real exchange rates and the relative price of nontradeables. Larger productivity differentials in the tradeable goods sector relative to the nontradeable sector lead to increases in the relative price of nontradeables, unit labor costs, and real exchange rates. The B-S productivity differential model relies on three critical assumptions in the long run: labor mobility, competitive labor markets, and purchasing power parity (PPP). This work tests these assumptions for the short and medium run for France, Germany, Japan, the U.K., and the U.S., and explains the implications for relative price determination by traded and nontraded sectors.

   The data reveal substantial violations in the assumptions of the productivity differential model. Significant real wage and unit labor cost (real wages minus productivity) growth differentials exist between industries as well as between traded and nontraded sectors, and these differentials only gradually diminish over time. Results strongly reject real wage equalization and support the presence of industry- and sectoral-specific human capital or compensating wage differentials. Further, the data reject the one-for-one relationship between productivity and real wage growth rates. The paper finds that innovations in domestic productivity and real wages affect domestic relative prices more than world prices in both the traded and nontraded sectors, which indicates strong deviations in PPP particularly for Germany, Japan, and the U.K.

   The assumptions of competitive labor markets, labor mobility, and PPP implicit in the productivity differential model imply the following asymmetries between the sectors. First, in the traded sector, competitive labor markets cause real wage and productivity growth rates to equalize, producing constant unit labor costs; whereas, in the nontraded sector, wage equalization across sectors and relatively slower productivity growth causes both rising unit labor costs and higher relative prices of nontradeables. Second, in the traded sector, competitive markets for tradable goods cause productivity and real wage innovations to possess no effect on price formation; whereas, in the nontraded sector, traded real wages and nontraded productivity influence unit labor costs and relative price formation. Thus, world prices affect tradable prices, and nontraded/traded productivity differentials affect the relative price of nontradeables. Real wage shocks or real wage differentials across sectors do not affect the relative price of nontradeables.

   The data do not support these asymmetries. Substantial interindustry wage and productivity differentials exist in both traded and nontraded sectors, indicating significant and large movements in unit labor costs. Moreover, innovations in traded productivity cause significant, but not one-for-one, wage movements in both sectors; results are similar in the nontraded sector. This demonstrates that productivity increases are not completely arbitraged by workers entering a sector; an outcome that is consistent with incomplete labor mobility due to industry/sectoral-specific human capital.

   Results further do not support the asymmetry of relative price formation. In both sectors, wages and productivity innovations significantly explain relative price determination positive wage (productivity) shocks represent adverse (favorable) supply shocks to a sector and imply rising (falling) relative prices in that sector. This relationship increases over time in both sectors; whereas, the productivity differential model implies increasing (decreasing) influence of world prices over time in the traded (nontraded) sector. In the medium run, domestic wage and productivity innovations can explain 21-64% of relative price movements. World prices play only a minor role in both traded and nontraded relative price formation in most economies. Thus, results do not endorse the asymmetric traded/nontraded relationship between productivity, wages, and relative prices: relative price formation does not vary substantially between sectors and is significantly explained by wages, productivity, and sometimes world prices.

   The assumptions of PPP, competitive labor markets, and labor mobility may be reasonable in the very long run. However, in the short to medium run (defined as 1-4 years), industry- or sector-specific factors such as differing human capital across industries or sectors implies that impediments to labor mobility and competitive labor markets may exist. Further, the prevalence of pricing-to-market strategies, hysteresis, imperfect information, and inertia in consumer habits suggests the failure of equalization for tradeable goods in the short to medium run.

   The remainder of the paper is as follows. Section 2 contains a brief review of the literature. Sections 3 and 4 present the model and data. Section 5 discusses results. Section 6 concludes with a summary of the evidence.

2. Review of the Literature

   A large amount of literature exists supposing wage differentials across industries and/or sectors (Dickens and Katz 1987; Katz and Summers 1989; Helwege 1992; Fallick 1993). This literature explains wage differentials using efficiency wage models (based on morale, turnover, or shirking), search theory, unionization, and segmented or sectoral labor markets. For instance, Gibbons and Katz (1992, p. 515) observe: "Several recent studies have shown that there are large and persistent wage differentials among industries, even after controlling for a wide variety of worker and job characteristics." They conclude that wage differentials are due to unmeasured ability. Katz and Summers (1989) present numerous empirical results concerning the presence and pattern of large interindustry wage differentials and find that workers in high-wage industries earn substantial rents, which impedes labor mobility.

   Helwege (1992) finds significant evidence of industry-specific human capital and demonstrates that random wage shocks create industry wage variation that persists for substantial periods of time. Subsequent work by Fallick (1993) presents additional evidence in support of industry-specific factors impeding labor mobility. Neal (1995, p. 653) shows that "workers apparently receive compensation for some skills that are neither completely general nor firm-specific but rather specific to their industry or line of work." When workers transfer industries, Neal finds they are unable to capitalize on their experience and tenure because the acquired skills and seniority are less useful in other industries; thus, "industry specific skill may constitute an important component of the typical workers human capital" (p. 653).

   A plethora of literature also exists that examines the merits of PPP (for a literature survey, see Breuer 1994). Much of this work shows that PPP fails, particularly in the short to medium run (Mark 1990; Engel 1993). Recent literature shows that PPP for tradables does not hold on a continuous basis but does so in the long run (Lothian and Taylor 1996); and moreover this suggests that we need an alternative model for the short/medium run. Panel estimates show that deviations possess a half-life of 3-5 years (Frankel and Rose 1996; Wu 1996; Wei and Parsley 1995); therefore, using a medium run or distributed lag specification, the influence of world prices on traded prices may be discernible.

   Several predictions of the productivity differential model are explored in the literature. Hsieh (1982) examines the effects of productivity shocks on real exchange rates. DeGregorio, Giovannini, and Wolf (1994) show that productivity differentials and shifts in demand determine innovations in the relative price of nontradeables; their work principally examines the determinants of traded and nontraded goods inflation through a series of correlations, particularly comparing Core European Monetary System to noncore economies. DeGregorio and Wolf (1994) test the influence of terms of trade, government spending shocks, and changes in traded/nontraded preferences on the real exchange rate. Using a neoclassical general equilibrium model, Asea and Mendoza (1994) examine the influence of productivity shocks on the relative price of nontradeables; their work concerns "the cross-sectional implications of the Balassa-Samuelson model rather than its time series implications." Using annual data, they pool the data and use the Hodrick-Prescott filter to ascertain the long-run properties of the data.

   Recently, Strauss (1996) using...