Robert A. Margo, Wages and Labor Markets in the United States, 1820-1860. Chicago: University of Chicago Press, 2000. xii + 200 pp. $28.00 (cloth), ISBN: 0-226-50507-3.
Reviewed for EH.NET by Joshua L. Rosenbloom, Department of Economics, University of Kansas.
Published by EH.Net, July 2000.
The antebellum period in the United States has sometimes been described as a "statistical dark age." With the publication of this book, however, such a characterization seems less appropriate. Drawing on two significant bodies of archival data--the manuscripts of the Censuses of Social Statistics for 1850 and 1860, and payroll records of civilian employees at military installations throughout the country--Robert A. Margo (Vanderbilt University) constructs new series of nominal and real wage estimates for the years 1820 through 1860. Based on both the quantity of evidence Margo has assembled and the care with which it is used, his series are sure to become the standard sources for scholars interested in the economic history of this period.
The argument of this compact volume unfolds in a logical and easy to follow manner. In chapter 2 Margo begins by establishing the need for new antebellum wage series. Carefully reviewing the existing data, he demonstrates that despite the efforts of a number of scholars, the commonly accepted wage series are pieced together from a variety of different sources in a fashion that makes them potentially unreliable. Moreover, there are troubling discrepancies in the behavior of the existing series that cannot be resolved without resort to more data. Having established the need for a new data, he then offers a description of the data sources that he will use.
The first of Margo's sources, and the more novel, is the "Reports of Persons and Articles Hired" that were made by the quartermasters of military installations. These reports contain over 59,000 monthly wage observations for civilian workers employed at forts in all parts of the country between 1820 and 1860. It should be noted that this figure is likely to overstate the number of independent observations in the data set, however, because some individuals may have been employed for more than one month. Nonetheless these data are a substantial addition to our stock of knowledge about antebellum labor markets. One might, of course, be suspicious of how well wages paid by the military reflect conditions in the civilian economy, but Margo goes to some lengths to demonstrate that the wages recorded in the "Reports" accurately reflected labor market conditions in areas surrounding the forts. A second problem, about which Margo has less to say, is the uneven coverage of the data. This unevenness is especially pronounced for the frontier regions. In the Midwest, for example, of 9,525 observations, 7,394 are from Kansas, and another 1,180 are from Missouri. As one might expect from this geographic emphasis, most of the midwestern data (6,794 observations) come from the 1850s.
The second of Margo's sources, the Censuses of Social Statistics, are more familiar to economic historians, and the published returns from these censuses have been used before. But Margo's use of the manuscript schedules is novel, and allows him to exploit considerably more detail than previous scholars, though at the expense of narrowing the geographic coverage to a set of eight states for which the manuscripts have survived. In chapter 3 Margo describes procedures he employs to construct annual wage series, both nominal and real, for three occupation groups--common laborers, artisans, and white-collar workers--in four regions. For artisans and common laborers the census of social statistics data is used to establish benchmark wage levels in 1850 that are then extrapolated using estimates of year-to-year movements derived from the military payroll data. Because the census data do not cover white-collar workers, the level of wages for this group is fixed using the military payroll data alone. To convert nominal to real wages, Margo uses region-specific indexes of wholesale prices.
To estimate annual movements in wages from the military payroll data Margo employs a hedonic wage regression framework to aggregate information from different locations, occupations, and skill levels. That is, for each region and occupation group, Margo regresses wage observations on dummy variables for location, indicators of worker skill, specific occupations, season of the year, and individual years. Implicitly this approach assumes that within each region, after controlling for fixed effects of location and other characteristics, wage movements over time at one location will be the same as those at any other location. This seems at least plausible for more settled areas in the East, but I am less convinced of the stability of wage rates within more recently settled areas. For example, one must wonder if the relationship between wages in Kansas and Missouri and other parts of the Midwest was the same in the 1830s and 1840s as it was in the 1850s.
The next three chapters make use of the wage series constructed in chapter 3 to examine a variety of important questions relating to labor market performance and economic growth in more wide-ranging terms. Chapter 4 examines wage gaps between farm and non-farm employment. Antebellum economic growth was associated with the movement of labor out of agriculture and into non-agricultural employment, and recent research has raised the question of whether this movement occurred as quickly as would have been optimal, or if frictions prevented labor from moving to its most efficient uses. Using data from the census of social statistics, Margo shows that in 1850 wage gaps were small within counties, and, after adjusting for the cost of living, they were also small within states. Thus it appears that farm and non-farm labor markets were relatively tightly integrated.
In chapter 5, Margo turns to the issue of geographic integration, comparing real wages by occupation group across the Northeast, Midwest, South Atlantic, and South Central regions of the country. Consistent with the geographic redistribution of labor from east to west in this period, he finds that wages in Midwest and South Central regions were persistently higher than they were in the Northeast and South Atlantic. Although the differentials were initially quite large between the two northern regions--over seventy-five percent for artisans and thirty percent for common laborers--there was a clear tendency toward wage convergence over time. In contrast wage gaps were smaller in the South, but showed no clear trend. Of particular note, Margo finds evidence that the North-South wage gap that became prominent after the Civil War had its origins in the antebellum period.
Margo continues his investigation of market integration in chapter 6 through a case study of one important episode: the California Gold Rush. Using data on wages at California forts, he offers a new series of wage estimates for the period surrounding the discovery of gold, and California's entry into the Union. This series clearly captures the sharp spike in wages occasioned by a massive new resource discovery. But interestingly it also reveals that the Gold Rush had a lasting effect on the California economy, leaving wage levels in the state permanently higher than they had been before.
Margo outlines a new interpretation of antebellum labor markets in Chapter 7. First, the evidence he has assembled indicates that real wages grew at about one percent per year on average, a rate strikingly close to that of per capita GDP over the same period. This suggests, as Margo puts it "that economic growth did 'trickle down,' on average, to members of the antebellum working class" (p. 143). While this finding would seem to defuse the pessimists' case that the working class did not share in the benefit of antebellum growth, there were shorter periods during which real wages did decline--especially the late 1830s and the late 1840s and 1850s. Comparing the dynamics of wages and prices in these episodes it appears that much of the reason for this is that wages tended to adjust with a lag to price movements. As Margo notes, these episodes of declining wages might explain some of the decline in the height of cohorts born at these times. But since overall real wage levels were rising between the 1820s and 1860s, economic explanations of the antebellum decline in stature must rely upon channels of influence other than a decline in the standard of living. Second, as much of the evidence on spatial and sectoral wage differences indicates, antebellum labor markets appear to have performed quite well, responding to both sectoral and geographic shifts in demand.
Notwithstanding the limitations of the data, this is an important contribution to the statistical and quantitative underpinnings of our understanding of the economic history of the United States. Through extensive archival research and careful analysis Robert Margo has added significant detail and clarity to our understanding of the antebellum period.
Joshua L. Rosenbloom is author of "Strikebreaking and the Labor Market in the United States, 1881-1894," Journal of Economic History 58 (Mar. 1998), and "Was There a National Labor Market at the End of the Nineteenth Century? New Evidence on Earnings in Manufacturing," Journal of Economic History 56 (Sept. 1996). He is currently working on a project estimating the rate of economic growth in North America before 1800.
Copyright © 2000 by EH.NET. All rights reserved. This work may be copied for non-profit educational uses if proper credit is given to the author and the list. For other permission, please contact the EH.NET Administrator (administrator@eh.net; Telephone: 513-529-2850; Fax: 513-529-3308).
Posted: 21 July 2000