Wages and Labor Markets in the United States, 1820-1860.

Whaples, Robert

By Robert A. Margo.

Chicago: University of Chicago Press, 2000. Pp. xii, 200. $28.00.

The American economy underwent a series of profound changes between 1820 and 1860--including the completion of the Erie Canal, the growth of commercial banking, industrialization in the Northeast, breakthroughs in the mechanization of agriculture, a 10-fold increase in cotton production, the rapid westward movement of population, the construction of over 30,000 miles of railroads, and (after 1845) the highest per capita rate of immigration in the country's history. Yet, our understanding of this period in which the United States experienced the onset of modern economic growth is surprisingly limited. Despite a consensus that per capita GDP rose substantially during this period, there has been considerable debate about whether the standard of living of the average worker rose. As Robert Margo (Vanderbilt) amply demonstrates in his thorough and painstaking review of previous studies, estimates of nominal and real wage trends and movements during this period are generally confined to the Northeast, rest on thin and perhaps unrepresentative bodies of evidence, and often involve questionable assumptions.

Margo scores an economic historian's hat trick: combining extensive data collection with careful, sensible, rigorous statistical analysis and an economically written, knowledgeable, wide-ranging interpretation of the findings. The core of his new evidence consists of newly collected payroll records containing nearly 60,000 monthly observations for civilian workers hired by the U.S. Army at forts and bases around the country between 1820 and 1860. These data are supplemented by evidence collected from the federal manuscript Censuses of Social Statistics of 1850 and 1860. Margo uses these data to construct annual nominal and real wage series for three groups of workers--artisans, common laborers and teamsters, and clerks--in four regions: the Northeast, Midwest, South Atlantic, and South Central. To do this, he uses hedonic regressions that control for fort location, workers' characteristics and occupations, and season before isolating the impact of year dummies on wage levels.

He concludes that in the aggregate, between 1820 and 1860, real wages grew at about 1% per year, approximately the same as the growth rate of real output per worker. Real wage growth was strongest for white-collar workers, but was fairly anemic for artisans in the Midwest and South Atlantic. In addition, Margo verifies that real wages followed an erratic course. Real wages generally grew sluggishly from the 1820s to the 1830s and fell between the 1840s and 1850s--as massive immigration glutted the labor market.

Next Margo examines the effectiveness of antebellum markets in allocating labor and dealing with economic change. He finds that the market was successful in shifting labor out of agriculture and into the burgeoning nonagricultural sector--wage gaps between the two sectors were fairly small. Consistent with the flow of labor in this period, Margo finds that wages in the West were persistently higher than in the East. Wages in the Midwest were initially considerably higher than in the Northeast but there was a marked convergence of northern wages over time. In the South wage gaps were much smaller and demonstrate little trend. In addition, Margo finds that the emergence of noticeably lower wages in the South Atlantic than in the Northeast predates the Civil War. The final empirical chapter examines the impact of the California Gold Rush on the western labor market. Because the short-run elasticity of labor supply was so small, the discovery of gold initially caused real wages to soar approximately 615% from la te 1847 and early 1848 to their peak in 1849. Real wages were soon eroded, but they remained permanently higher than they had been before the rush.

Margo's findings are important because they cover a wide geographical scope and a continuous period of four decades. The leading textbooks in the field will make major revisions on the basis of this volume. However, it is important not to push these data too far and pretend to a certainty about this period--especially when real wage trends in our own era are the subject of vigorous debate. Margo carefully and repeatedly points Out the fragility of his new series while defending their overall usefulness. He notes, for example, that the price series used to deflate nominal wages have biases that cannot be eliminated, that data are thin for certain region-occupation-year combinations, and that wages paid by the military may not perfectly reflect broader civilian pay levels.

Perhaps the greatest potential weakness of this study is the assumption that wages paid at forts that are often on the edge of a large census region are indicative of levels and trends for the broader region. This problem is probably least serious for the Northeast and worst for the Midwest and South Atlantic. Over three-fourths of the observations for Midwestern laborers and nearly two-thirds of the observations for Midwestern artisans come from Kansas. If the labor market near Kansas forts was out of step with the rest of the Midwestern labor market, the yearly dummy variables are likely to reflect the peculiarities of the Kansas market rather than the regional market, since the econometric specifications assume a time-invariant set of fort-specific effects. Although Margo is well-versed in economic history and labor history, he makes little reference to military history. To adequately understand this labor market data, we probably need to know what was happening during the period at places like Kansas' Fo rt Leavenworth--the "Queen of the Frontier Posts," a key staging ground during the Mexican War, through which thousands of wagons passed while traveling to the West, especially along the Santa Fe Trail. Could the isolation, dangers, and resultant compensating wage differentials paid to those working at such posts have ebbed and flowed at a pace that was dissimilar to the broader Midwestern labor market? I am somewhat wary.

Likewise, the South Atlantic data are dominated by observations from Florida--over 70% of South Atlantic laborer observations are from Florida, nearly 60% of artisans, and almost half of white-collar workers. Florida was surely on the periphery of this region, which extends up to Maryland. Moreover the bulk of the observations come from the period in which the Second Seminole War was raging--a war that left about 1400 American soldiers and 100 American civilians dead, mainly from disease. Although the timing of the war seems to have had an immense impact on wages in the South Atlantic, Margo does not mention its existence. Table 3A.3 shows that nominal wages in the South Atlantic soared between 1835 and 1836, as the war erupted, then stayed fairly high until the end of the war, which wound down around 1842. Likewise, the real wage data show the South Atlantic to be an outlier. In the years at the start of the war, real wages fell considerably in the other regions, but rose in the South Atlantic. Does a data set dominated by war-wracked Florida reflect broader trends in the South Atlantic during this period?

These concerns about matching military with economic history may not undermine Margo's broader conclusions, but they reemphasize potential problems with the data and the need to interpret them carefully.