Manpower in Economic Growth: The American Record since 1800

Classic Reviews in Economic History

Stanley Lebergott, Manpower in Economic Growth: The American Record since 1800. New York: McGraw-Hill, 1964. xii + 561 pp.

Review Essay by Robert A. Margo, Department of Economics, Boston University.

Manpower after Forty Years

During the first half of the twentieth century classical musicians routinely incorporated their personalities into their performances. One recognizes immediately Schnabel in Beethoven, Fisher in Bach, Cortot in Chopin, or Segovia in just about anything written for guitar. As the century progressed performance practice evolved to where the “text” — the music — became paramount. The ideal was to reveal the composer’s intent rather than putting one’s own stamp on the notes — the performer as conduit per se rather than co-composer.

Personal style played a major role in the early years of the cliometrics revolution. Hand a cliometrician an unpublished essay by Robert Fogel or Stanley Engerman, and I am quite sure she could identify the author after reading the first couple of paragraphs (if not the first couple of sentences). No one can possibly mistake a book by Doug North for a book by Peter Temin or an essay by Paul David for one by Lance Davis or Jeffrey Williamson. To some extent this is because personal style mattered at the time in economics generally — think Milton Friedman or Robert Solow. But mostly it mattered, I think, because these cliometricians were on a mission. Men and women on a mission put their personalities up front, because they are trying to shake up the status quo.

So it is with Stanley Lebergott. Indeed, of all the personalities who figured in the transformation of economic history from a sub-field of economics (I am tempted to write “intellectual backwater”) that eschewed advances in economic theory and econometrics to one that embraced them (I am tempted to write “for better and for worse”), Lebergott’s style was perhaps the most personal. In re-reading Lebergott’s most famous book — his Manpower in Economic Growth: The American Record since 1800 (1964) — one sees that style front and center on nearly every page, as well as the conflicting emotions as its author tried, not always successfully, to marry the anecdotal and archival snippets beloved by historians with the methods of economics. Manpower was (and is) substantively important for two reasons. First, prior to Manpower, the “economic history of labor” meant unions and labor legislation. By contrast, Lebergott made the labor market — the demand and supply of labor — his central focus and in doing so elevated markets and market forces to a central tendency in the writing of economic history. Second, Lebergott produced absolutely fundamental data — estimates of the labor force, industrial composition, unemployment, real wages, self-employment, and the like — that economic historians have relied on (or embellished) ever since.

These two accomplishments aside, I emphasize style not because, in Manpower‘s case, it is light years from the average article that I accept for publication in Explorations in Economic History. Economic history, like all economics, is vastly more technical than it was in the early 1960s. Burrowing into the style of Manpower reveals an author transfixed with what he perceived to be the grandness of the American experiment, the transformation of a second-rate colony into the greatest economy the world had yet seen. The core of Manpower would always be its 33 appendix tables and 252 (!) pages of accompanying explanatory text lovingly produced and so relentlessly documented as to drive any reader to distraction (or tears). So much the line in the sand, daring — indeed, taunting — the reader to do better. Lebergott knew that, in principle, one could do better, because he did not have ready access to all the relevant archival materials. I would conjecture, however, that he would always be surprised if anyone did, in fact, do better. Tom Weiss, himself one of the great compilers of American economic statistics, spent several years redoing Lebergott’s labor force estimates using census micro data rather than the published volumes that Lebergott relied on (Weiss 1986). In commenting on Weiss’s work, Lebergott (1986) characterized the differences between his original figures and the revisions as “very small beer” and then took Weiss to task for failing (in Lebergott’s) view to fully justify the revisions. “One awaits with interest,” he concluded, “further work by the National Bureau of Economic Research project of which this is a part.” When Georgia Villaflor and I (Margo and Villaflor 1987) produced a series of real wage estimates for the antebellum period drawing on archival sources that Lebergott did not use, I received a polite letter congratulating me but requesting more details and admonishing me to think harder about certain estimates that Lebergott felt did not mesh fully with his priors. There are thousands of numbers in those 33 appendix tables and one’s sense is that each number received the undivided attention of its creator for many, many, many hours.

But numbers do not a narrative make. Chapter One, “The Matrix,” has little in common with the archetypal introduction that gives the reader a roadmap and a flavor of the findings. It begins rather with an 1802 quote from “The Reverend Stanley Griswold” about the frontier that lay before the good minister. “This good land, which stretches around us to such a vast extent … large like the munificence of heaven … [s]uch a noble present never before was given to any people.” (Reviewer’s note: any people? Which people?) The first sentence goes on to describe an incongruous scene from Kentucky in 1832, “a petit bon homme” and his wife and their “little pile of trunks” sitting in a restaurant in the middle of (literally) nowhere. We then learn of a “great theme” of American history, that which motivated those who wrested the land from the “wilderness” — a belief in an open society, of which there were three elements. First, “hope” — an unabashed belief that things will always get better, and were better in America than in Europe. Second, “ignorance” — Americans were always willing to try something new, no matter how crazy. Third, America had a huge amount of space for people to spread out in. OK, the reader says, but where’s the economics? Ca. page 13 Lebergott emphasizes that the three elements made Americans unusually restless people, willing to move all the time. Ordinarily, Lebergott opines, it is the smaller (geographically-speaking) countries that have higher labor productivity because, ordinarily, people do not like to move. But Americans liked to move, he claims, and they did so on the slightest provocation. Excessive optimism, misinformation, and folly are core attributes of the American spirit and key factors in the American success story. In the end, the errors didn’t matter anyway (“small beer” indeed) because the land was so rich. More people moved to California in 1850 than could be rationally justified by the expected returns to gold mining but, as a result, California entered the aggregate production function sooner than otherwise. Labor mobility per se was a Good Thing, and American had it in abundance.

Chapter Two asks where all the workers would come from. Lebergott notes that certain labor supplies were highly predictable — slaves, for example. But once the slave trade was abolished the supply of slave labor grew at whatever the natural rate of increase. If the riches of America were to be tapped, free labor would have to be found — all the more difficult if the required number of workers to be assembled in any given spot was very large.

Another element of the Lebergott style is a dry wit, as evidenced in his exchange with Weiss. In a section on “[t]he Labor Force: Definition” we are told that ‘[t]he baby has contributed more to the gaiety of nations than have all the nightclub comics in history. We include the comic in the labor force … as we include [his] wages in the national income but set no value on the endearing talents provided by the baby.” In discussing the then-fashionable notion that the aggregate labor force participation rate (like other Great Ratios) was “invariant to economic conditions” Lebergott notes that small changes can nevertheless have great import. “The United States Calvary,” he observes, “was sent to the State of Utah because of the difference between 1.0 wives per husband and a slightly greater number.” The remainder of the chapter considers segments of the labor force whose labor was, indeed, “responsive to economic conditions” — European immigrants, internal migrants, (some) women and children as well as the impact of social and political factors on labor supply; it demonstrates the extraordinary flexibility of the American labor force and its responsiveness to incentives. While this conclusion would not surprise anyone today it was, I think, quite revolutionary at the time. It is as good an example of any I know of the power of historical thinking to debunk conventional wisdom derived from today’s numbers.

By now the reader is accustomed to Lebergott’s modus operandi — the opening paragraph that sometimes seems to be beside the point but really isn’t; quotations in the text from travelogues, diaries, plays, literature and what-not; obscure (to say the least) references in the footnotes; all interspersed with economic reasoning that has more than a tinge of what would be called today “behavioral” economics. In Chapter Three Lebergott talks about the “process” of labor mobility, which is really one extended probing into the relationship between mobility of various sorts and wage differentials. We get to see some univariate regression lines, superimposed in scatter-plots of decade-by-decade changes in the labor force at, say, the state level, against initial wage rates. Generally, labor flows were directed at states with higher initial wage rates, although Lebergott is quick to assert that “[m]igrants suboptimized” because the cross-state pattern was far less apparent at the level of regions. Next, Lebergott takes on the notion that economic development is an inexorable process of labor shifting out of agriculture. The American case, Lebergott claimed, challenges this notion. American workers shifted out of agriculture when the economic incentives were right; that is, when the value of the marginal product of labor was higher outside of agriculture.

The remainder of Chapter 3 is divided into two brief sections, both of which contain some of the most interesting writing in the book. In “Social Mobility and the Division of Labor,” Lebergott examines the relationship between occupational specialization and growth. In the nineteenth century most workers possessed a myriad of skills, farmers especially. They were jacks of all trades, masters of none. Lebergott speculates that this was a good thing because the master of none was more inclined to try something new, rather than assume he was, well, the master and therefore knew everything. If some fraction of novel techniques were successful, this could (under strong assumptions) lead to a higher rate of technical progress. “Origins of the Factory System” considers the problem posed earlier in the book of assembling large numbers of workers at a given location. Rather than pay higher wages, manufacturers turned to an under-utilized source of labor, women and children. Some years later, the ideas presented in this section would develop in full bloom in a celebrated article by Claudia Goldin and Kenneth Sokoloff (Goldin and Sokoloff 1982) on the role of female and child labor in early industrialization.

At 89 pages, Chapter Four, “Some Consequences,” is the longest chapter in the book. The first few pages, highly influential, are given to the formation of a national labor market, revealed by changes over time in the coefficient of variation of wages across locations. We are then given an extended tour of the history of American real wages, back and forth between the relevant tables in the appendix, quotations from contemporaries and other anecdotal evidence. The “Determinants of Real Wage Trends” comes next. The first, productivity, is no surprise. The second, “Slavery,” isn’t really either, but here Lebergott’s contrarian instincts, I think, get the better of him. Lebergott would have the reader believe that, first, free and slave labor were close to perfect substitutes; and, second, slave rental rates contained a premium above what the slave would have commanded in a free labor market. Consequently, when slavery ended, wages fell and there was downward pressure on real wage growth for a time. No question that wages fell in the South after the Civil War but Lebergott’s analysis is incomplete at best. Slave labor was highly productive before the Civil War because of the gang system, and when the gang system ended, the demand for labor fell in the South. Because labor supplies were not perfectly elastic, wages fell too. “Immigration,” the third purported influence, had negative short run effects on wages but positive long run effects via productivity growth.

What follows next is a 25-page section that years later produced two high-profile controversies in macroeconomics. This is the (celebrated) section where Lebergott presents his long-term estimates of unemployment. In thinking today about his work, we would do well to remember that, at the time he prepared his estimates, the United States had only a relatively brief experience with the direct and regular measurement of unemployment, courtesy of the 1940 Census and the subsequent Current Population Survey (CPS). (By “direct” I mean answers to questions about a worker’s time allocation during a specific period of time — if you did not have a job during the survey week, were you looking for one?)

Like all the estimates in the book, Lebergott’s unemployment figures were the product of detailed, painstaking work that, inevitably, required strong assumptions. The fundamental problem was that, if one wanted annual estimates of unemployment, there was no way to obtain these directly from survey evidence prior to the CPS. For some benchmark dates one could produce tolerable direct estimates from the federal census, but the federal census was useless if one wanted to generate an estimate, say, for 1893 or, for that matter, 1933.

Lebergott’s solution was to rely on an identity. By definition, the labor force was the sum of employed and unemployed workers. One might not know the number of unemployed workers but perhaps one could extrapolate between benchmark dates the number of workers in the labor force and employment, one could estimate unemployment levels via subtraction.

The first high profile controversy involved Lebergott’s estimates for the 1930s, which included in the count of unemployed workers persons on work relief. After 1933 there were many such workers, and so, by historical standards, unemployment looks, of course, rather high. This generated a lot of theoretical work for macroeconomists who thought they had to explain how unemployment rates could remain above 10 percent while real wages were rising (after 1933).

Michael Darby (1976) suggested that this effort was misplaced because Lebergott “should” have included the persons on work relief in the count of employed workers. Darby showed that doing so made the recovery after 1933 look much more normal. I’ve written a few papers on this issue, and my view is somewhere in-between Darby and Lebergott (Margo 1991; Finegan and Margo 1994; see also Kesselman and Savin 1978). Ideally, in constructing labor force statistics we should be consistent over time, so if persons on work relief were “employed” in the 1930s we should consider adding, say, “workfare” recipients to the labor force (or, possibly, prisoners making license plates) today, but this ideal may not be achievable in practice. The real issue with New Deal work relief is not the resolution of a crusty debate between competing macroeconomic theories but whether the program affected individual behavior. Here I think the answer is a resounding yes — unemployed individuals in the 1930s did respond to incentives built into New Deal policies. Wives were far more likely to be “added workers” if their unemployed spouses had no work whatsoever, than if the spouse held a work relief job, so much so that, in the aggregate, the added work effect disappeared entirely in the late 1930s, because so many unemployed men were on work relief.

The second high-profile debate involved Christina Romer’s important work on the long-term properties of the American business cycle. Prior to her work it was (and in some quarters still is) a “stylized fact” that the business cycle today is less volatile than it was in the past. Lebergott’s original unemployment series combined with standard post-war series were often used to buttress claims that the macroeconomy become much more stable over time. Statistical measures of volatility estimated from the combined series clearly suggest this, whether volatility is measured by the average “distance” (in percentage points) between peaks and troughs or standard deviations.

Romer (1986) argued that, to a large degree, this apparent decline in volatility was a figment of the way the original data were constructed. In particular, in constructing his annual series, Lebergott assumed (among other things) that deviations in employment followed one-for-one deviations in output. Romer invoked Okun’s law, arguing that the true relationship was more like 1:3. Constructing post-war series by replicating (as close as possible) Lebergott’s procedures produced a new series that was not less volatile than the pre-war series, thereby contradicting the stylized fact that the macroeconomy became more stable over time. This was, needless to say, a controversial conclusion, with many subsequently weighing in. Now that the dust is settled, my own view — a view I think that many share, although I could be wrong — is that there is definitely something to Romer’s argument; at the very least, she demonstrated (as she claimed in her original article) that before one draws conclusions from historical time series, one should be very familiar with how the series are constructed. Chapter Four ends with another of Lebergott’s meditations on the alleged constancy of aggregate parameters — in this case, factor shares.

Chapter Five (“Some Inferences”) concludes the narrative portion of the book. It repeats the book’s earlier mantra that “Yankee ingenuity” and initiative, especially that embodied in immigrants, were central to American success as opposed, say, to “factor endowments.” It ruminates on how highly mobile labor influenced the choice of technique, in ways familiar to the first generation of cliometricians, especially those who found H.J. Habakkuk a source of (repeated) inspiration. It notes how “thickening markets” made finding continuous work easier over time, reducing the wage premium associated with unemployment risk. Today’s economic historians, infatuated with “institutions” v. “geography” would probably disagree with the emphases in the chapter but I think there is much to admire in Lebergott’s “inferences.”

Some economic historians make their mark as much through their graduate students as their writings. Lebergott spent his academic career in a liberal arts college and did not, therefore, directly produce graduate students like a William Parker, Robert Fogel or (more recently) Joel Mokyr. In certain ways he was an outsider to economic history, an economist with a vast and deep appreciation for history in all of its flavors, who saw the past for what it can say about the present, not as an end in itself like a more “traditional” historian would. Compared with other classic works of cliometrics such as Fogel’s Railroads and American Economic Growth or North and Thomas’s The Rise of the Western World, Manpower‘s quirkiness can be a frustrating, more suitable for dabbling than a sustained read. By today’s standards the book falls short in its treatment of racial and ethnic differences (gender is more balanced) although this would hardly distinguish it from most other work in economics and economic history at the time. Yet Lebergott’s influence on economic history has been profound. There are few activities that economic historians can engage in of greater consequence than reconstructing the hard numbers. In this line of work Lebergott had few peers. Manpower put the labor force — people — at the center of economic history, not the bloodless “agents” of economic models but real people. As if to underscore this, the style asserts, like a triple fff in music: a real person not a (bloodless) “social scientist” wrote this book, one in deep and abiding awe of the economic accomplishment of his forbearers.

References:

Darby, Michael. 1976. “Three and a Half Million US Employees Have Been Mislaid: Or, An Explanation of Unemployment, 1934-1941,” Journal of Political Economy 84 (February): 1-16.

Finegan, T. Aldrich and Robert A. Margo. 1994. “Work Relief and the Labor Force Participation of Married Women in 1940,” Journal of Economic History 54 (March): 64-84.

Goldin, Claudia and Kenneth Sokoloff. 1982. “Women, Children, and Industrialization in the Early Republic: Evidence from the Manufacturing Censuses,” Journal of Economic History 42 (December): 741-774.

Kesselman, Jonathan R. and N. E. Savin. 1978. “Three and a Half Million Workers Were Never Lost,” Economic Inquiry 16 (April): 186-191.

Lebergott, Stanley. 1964. Manpower in Economic Growth: The American Record since 1800. New York: McGraw-Hill.

Lebergott, Stanley. 1986. “Comment,” in Stanley Engerman and Robert Gallman, eds., Long Term Factors in American Economic Growth, pp. 671-673. Chicago: University of Chicago Press.

Margo, Robert A. 1991 “The Microeconomics of Depression Unemployment,” Journal of Economic History 51 (June): 333-341.

Margo, Robert A. and Georgia Villaflor. 1987. “The Growth of Wages in Antebellum America: New Evidence,” Journal of Economic History 47 (December): 873-895.

Romer, Christina. 1986. “Spurious Volatility in Historical Unemployment Data,” Journal of Political Economy 94 (February): 1-37.

Weiss, Thomas. 1986. “Revised Estimates of the United States Workforce, 1880-1860,” in Stanley Engerman and Robert Gallman, eds., Long Term Factors in American Economic Growth, pp.641-671. Chicago: University of Chicago Press.

Robert A. Margo is Professor of Economics and African-American Studies, Boston University, and Research Associate, National Bureau of Economic Research. He is also the editor of Explorations in Economic History.