Nominal Wage Rigidity and Industry Characteristics in the Downturns of 1893, 1929, and 1981 by Christopher Hanes. Published in volume 90, issue 5, pages 1432-1446 of American Economic Review, December 2000

Wage, price, and output indexes suitable for comparing cyclical movements across decades show a decrease in nominal wage flexibility (the change in wage inflation associated with output fluctuations) after the 1880s, following an increase in strike frequency linked to the spread of large-scale manufacturing. Cross-sectional data show that firms in industries experiencing more strikes in the 1880s were less likely to cut nominal wages in the depression of 1893. This and other evidence suggests that the nineteenth-century decrease in wage flexibility was caused by an increase in workers' bargaining power in the absence of binding wage contracts.

Wage, price, and output indexes suitable for comparing cyclical movements across decades show a decrease in nominal wage flexibility (the change in wage inflation associated with output fluctuations) after the 1880's, following an increase in strike frequency linked to the spread of large-scale manufacturing. Cross-sectional data show that firms in industries experiencing more strikes in the 1880's were less likely to cut nominal wages in the depression of 1893. This and other evidence suggests that the 19th-century decrease in wage flexibility was caused by an increase in workers' bargaining power in the absence of binding wage contracts.

Data from recent years indicate that employers are especially unlikely to cut the nominal wage rate paid for a job. Adjustments to real or relative wages that require absolute cuts in money wages are very rare, relative to the frequency one might expect to observe based on distributions of wage changes that do not require money wage cuts (Erica L. Groshen and Mark E. Schweitzer, 1995, 1997; David E. Lebow et al., 1999). A worker who remains with the same employer from one year to the next is accordingly unlikely to report an absolute reduction in the nominal wage he receives (David Card and Dean Hyslop, 1997; Shulamit Kahn, 1997; Joseph G. Altonji and Paul J. Devereux, 1999). The existence and causes of “downward nominal wage rigidity” have implications for macroeconomic outcomes and for practical issues in monetary policy. Presumably, efficient operation of labor markets requires many downward adjustments of relative wages, and occasional decreases in the overall wage level relative to product prices. If the average rate of wage inflation were close to zero, some of these relative or real wage adjustments would be blocked by a floor under current levels of nominal wage rates. George A. Akerlof et al. (1996) assert that downward nominal wage rigidity reflects fundamental preferences of workers, and argue that it implies a central bank should target a rate of price inflation greater than zero because “a target of zero inflation will impose permanent real costs on the economy” (p. 2). The notion that downward nominal wage rigidity is a fundamental constraint, so that inflation serves to “grease the wheels of the labor market,” is controversial even among those who accept the possibility of nominal rigidities in general. It has been argued that downward nominal wage rigidity, as distinct from such phenomena as “menu costs” [which would discourage adjustment of a wage in either direction (David Romer, 1993)], may be an artifact of an inflationary monetary regime, and would disappear in the absence of persistent inflation. “Nominal wage reductions would no longer be seen as unusual if the average nominal wage was not growing. Workers would not see them as unfair, and firms would not shy away from imposing them” (Robert J. Gordon, 1996, p. 62; see also N. Gregory Mankiw, 1996; William Poole, 1998; William B. English, 2000). To see whether downward nominal wage rigidity would exist in a noninflationary regime, it may be useful to examine data from times and places where there was no persistent wage inflation. For the United States, at least, that excludes data from years since the Second World War: throughout the postwar period, price inflation or real wage growth has been sufficient to keep average wage inflation above zero. Before the Second World War, on the other hand, long-run trend rates of price inflation were often close to zero or negative (Robert B. Barsky, 1987). If wage inflation was correspondingly low, U.S. historical data may offer tests of the proposition that downward nominal wage rigidity exists even in noninflationary

Most major American industrial business cycles from around 1880 to the First World War were caused by fluctuations in the size of the cotton harvest due to economically exogenous factors such as weather. Wheat and corn harvests did not affect industrial production; nor did the cotton harvest before the late 1870s. The unique effect of the cotton harvest in this period can be explained as an essentially monetary phenomenon, the result of interactions between harvests, international gold flows, and high-powered money demand under America's goldstandard regime of 1879–1914.