The counted and the uncounted: the occupational structure of early American cities.

Historians have examined the occupational and social structures of American cities in the Revolutionary Era and the early National Period for several reasons. For example, Gary Nash argues the cities were the "crucibles" of the political consciousness which led to the American Revolution, and part of the source of that consciousness was growing urban economic inequality and class tensions.(1) Other scholars document these trends both during the Revolutionary Era and in the decades immediately after. They also argue that increasing economic inequality reflected emerging class differences and shaped urban political and social life.(2) Historians seeking a bottom-up view of urban life, as a counterbalance to the history of urban elites, depend on occupational, economic, and demographic statistics to describe the working population because of the paucity of literary evidence from ordinary people.(3) Finally, historians of working-class formation look at the occupational structures of early American cities to pinpoint the beginnings of the working class.(4)

Nearly all of these endeavors employ, at least in part, quantitative measures of occupational structure, income, or property holdings, almost always based on the three most accessible sources of such information: tax lists, city directories, and probate court inventories of wills. Nash, for example, uses tax lists and inventories of wills to calculate a variety of statistical measures of income distribution, property distribution, and occupational structure for Boston, New York, and Philadelphia between 1685 and 1775.(5) Alan Kulikoff offers a variety of qualitative and quantitative evidence for his conclusion of growing inequality in Boston between 1771 and 1790, but most of his statistics are based on calculations from tax lists.(6) Almost every one of the community studies of workers in the early 1800s presents tables of occupations based on city directories or tax lists.(7)

Reconstructing urban occupational and class structures from tax lists or city directories has been a useful starting point for understanding the social structures of early American cities. Occupational and economic reconstructions have contributed to investigations of such key questions as the social origins of the American Revolution(8) and the social tensions of the post-Revolutionary period, as well as to efforts to understand the beginnings of working-class formation. However, tax lists and city directories are a problematic source for such reconstructions because they exclude most of the population. Not only do they ignore people who were not considered part of the labor force, especially women and children, but they also exclude the majority of the labor force--in many cases, a large majority. The result is usually a sample population composed overwhelmingly of white males over twenty-one years old and, even among adult white males, skewed toward the prosperous and elite. Many of the most important works TABULAR DATA OMITTED which use these sources display great quantitative sophistication. Nonetheless, they rarely make systematic efforts to quantitatively estimate who has been left out. Since the omitted were usually the majority of the population, any conclusions about the shape of the labor force, the comparative size of various social groups or classes, or the extent or changing nature of economic inequality are potentially suspect without such estimates.

Table 1 demonstrates just how substantial these exclusions are. The calculations of occupational structures for the thirteen cities in Table 1 are based on data from various monographs and, for several cities, my own calculations from city directories. As the column labeled "N/Population" reveals, in no case do these occupational totals represent more than 20% of total population. The mean is just over 13%. Economic historian Stanley Lebergott's estimates of national labor force trends place the national labor force participation rate at 35.8% of the nation's population in 1800.(9) In most cities, given the influx of young adults looking for work, the labor force participation rate was probably higher than the national average. Thus, the occupational breakdowns for nearly all of the cities in Table 1 exclude the majority of the local labor force. For those cities in the top half of the table, more than two thirds of the labor forces are excluded. Only the figures for Germantown, and maybe Philadelphia, can represent even half of the local labor force.

Nearly everyone who has used early tax lists and city directories acknowledges that they exclude entire categories of people and undercount even those groups they seek to include. Nash cautions, in his occupational table, at the end of The Urban Crucible that "both inventories and tax lists, especially for Philadelphia, greatly understate the number of ordinary seamen...."(10) In a later essay about urban occupational structure, Nash and collaborators Billy G. Smith and Dirk Hoerder reiterate this point. They challenge the conclusions of scholars who impute a middle-class character to early American cities on the basis of the commercial and administrative functions of most cities. To bolster their case, they offer estimates of the number of sailors in Philadelphia, Boston and New York at various dates between 1756 and 1790 as well as minimum estimates of the number of laborers.(11) Yet, although the thrust of their argument is similar to the one presented here, they do not go far enough. They restrict their estimates to free adult males, emphasizing sailors as the most important group omitted from earlier occupational tabulations. Most other historians who present tables of urban occupational structure offer no corrective estimates of those left out of tax lists and city directories. I have found only two articles on urban occupational structure in this era which include calculation of occupations tabulated as a percentage of total population or a percentage of total labor force.(12)

In fact, unreported sailors are just the tip of the quantitative iceberg of the uncounted. In 1800, the Philadelphia city directory provided occupations for 7,542 individuals, while I estimate the probable labor force at about 16,000. Among the approximately 8,500 labor force participants excluded from the city directory tabulation, I estimate there were 1,878 uncounted sailors. Sailors were less than one fourth of those left out.

Table 2 presents a more systematic effort to reconstruct Philadelphia's labor force in 1800. If we compare Table 2 with the city directory figures for Philadelphia in Table 1, we get a very different picture of the city. The Philadelphia in Table 1 is a very strange place, indeed. Merchants, professionals, shopkeepers, and other white collar occupations are 42% of the labor force. Add in the most prosperous master craftsmen, and the well-to-do and middle classes seem to constitute well over half. Philadelphia was among the most important seaports in the entire Atlantic economy, but not only are there few sailors, we also find no more than a handful of the other occupations necessary for the labor-intensive movement of goods by animal power and human muscle, longshoremen and wagon-drivers, for example. If we took the figures in Table 1 literally, it would seem that the work ethic was amazingly entrenched among the well-to-do in Philadelphia. The merchants, lawyers, and public officials--or their wives--must have washed their own clothes, cooked their own food, served their own meals, driven their own carriages, and tended their own gardens because the city directories include virtually no servants or service workers. A much more plausible assumption is that the half to three-quarters of the workforce left out of tax lists and city directories, in Philadelphia and most other cities, was mainly composed of working people including sailors, servants, carters, dockhands, and other menial occupations.

Table 2

Distribution of Philadelphia's Labor Force, 1800(a)

Demographic structure:

white males over 15                    10,516
white females over 15                   2,793
white children 10 to 15                   702
free black males over 13                1,036
free black females over 13                901
slaves                                     50

total                                  15,998

Occupational structure:

       merchants     shopkeepers   artisans and   sailors and   unskilled
          and         and white       craft          other      and manual
     professionals     collar       apprentices    transport     service

n       1670           1644            5706          2873         4105
%         10.4           10.3            35.7          18.0         25.7

a For sources and methods of estimate, see appendix.

Table 2, which presents estimates of the distribution of the entire Philadelphia labor force, suggests exactly that. The proportion of merchants, professionals, shopkeepers, and bookkeepers has shrunk by more than half. They are now outnumbered by their servants and day laborers. The Philadelphia of Table 2 was a much more plebian place with a minority of the prosperous and propertied surrounded by armies of servants and dressmakers, carters and dock hands, shoemakers' apprentices and journeymen bakers. It is much easier to understand why the urban propertied of early America harbored paranoid fears of their neighbors like those so cogently expressed by New York City merchant and Patriot leader Gouveneur Morris in 1774. Morris described an assemblage at a protest meeting as "poor reptiles," and he worried "with fear and trembling" that these people whom he had helped set in motion, "ere noon they will bite."(13)

Morris, or his Philadelphia counterparts, may indeed have been unrealistic if they saw the urban poor as a unified group likely to mobilize politically against the wealthy. The lives of a Black servant, an Irish dock hand, or a native seamstress were very different, and often they harbored as great a mistrust of each other as of the well-to-do. But the numbers in Table 2 suggest something of the basis of Morris' paranoia.

I have not yet made comparable occupational reconstructions of other cities, but there is every reason to expect they would resemble Philadelphia. New York and Boston were demographically similar to Philadelphia and were also thriving seaports. Many of the nation's other cities and towns had a somewhat different demographic structure, but, in most cases, a structure which would probably skew the occupational composition of the labor force even more toward the most menial and unrewarded forms of labor. In Southern cities such as Charleston, Norfolk, Savannah, and Richmond, slaves made up the majority of the workforce, and the white population included a substantial surplus of younger males, both characteristics suggesting even larger proportions of transport workers, servants, and laborers than in the Northern cities. In the medium-sized seaports of New England (e.g. Salem, Newburyport, New London) shipping and international trade dominated the local economies even more than in the big cities. Sailors and maritime-related occupations (e.g. dockhands, shipcarpenters) were probably proportionately even more important than in the larger seaports. Only in such interior market towns as Reading and Lancaster does it seem likely that merchants, professionals, storekeepers, and artisans made up a substantially larger proportion of the labor force than they did in Philadelphia.

More careful use of occupational and social statistics gleaned from tax records, city directories, and other similar sources is important not only because of the intrinsic value of historical accuracy, but also because these statistics have been one of the underpinnings of much of the historiography of early American history, from the colonial era through the early national period. At a minimum, generally-accepted conclusions about the extent of urban economic inequality in the Revolutionary Era, the degree of class polarization, and the extent of change from earlier decades cannot be given full credence without more thorough recalculations which include the uncounted. Since the uncounted were disproportionately, female, non-white, young and poor, my suspicion is that such recalculation would reveal even greater inequality than has been heretofore supposed.

The question of whether the inequality of the Revolutionary Era represented a major change from colonial patterns is also problematic. A generation of historians who came of age during the 1960s counterculture tend to romanticize and idealize the Gemeinschaft of colonial society. Since slavery and indentured servitude were even more important in the colonial era than later on, the proportion of uncounted in earlier records may be even higher than in the late 1700s or early 1800s. Trends in inequality cannot be understood without imbedding their examination in a larger discussion of demographic changes and changes in labor systems.(14)

My own particular interest in these early American urban occupational statistics is the role they have played in the historiography of the American working-class. I began to investigate these statistics as part of a larger work on American working-class formation. Quickly, I became disatisfied in how I saw them being used. The model worker for much of that historiography has been the white male urban artisan undergoing deskilling as a result of division of labor, and, later, technological change.(15) Such artisans were among the most vocal critics of capitalist development. Their organizational efforts, as well as their capacity to accumulate modest amounts of property, left a paper trail further increasing their historical visibility. Not coincidentally, they are the only group of urban workers with a substantial representation in tax records and city directories.

But urban artisans were only one of several cohorts of working people whose lives were being fundamentally altered by commercialization and industrialization. Servants, transport workers, laborers, rural outworkers and farm workers all experienced profound change. Subsuming working-class formation under the stow of urban artisans misses the majority of workers, and avoids confrontation with the difficult problem of how changes in the class system intersected with changes in the gender system and the racial order. Part of creating a more inclusive and more gender and racially-conscious analysis of working-class formation must include greater attention to whom quantitative social historians have counted and whom they have not.

Appendix: Method of Estimate of Philadelphia's Labor Force, 1800

I estimated the size of Philadelphia's labor force in 1800 by multiplying the known populations (from the census) of white males over 15, white females over 15, white children 10 to 15, free blacks, and slaves(16) by labor force participation rates for each group. For white males over 15, I used a labor force participation rate of .885, the national rate in the 1850 census,(17) the first census to report detailed occupational data. This rate did not vary significantly between 1850 and 1900 (and not in a consistent direction) so it seems unlikely that it would have been dramatically different in 1800. For white females over 15, I used a rate of .25. Again this ratio is consistent with trends for urban areas reported by later censuses beginning in 1870, the first census to report such data.(18) For children aged 10 to 15, I used a rate of .15. The national rate in 1870 was .13, in 1880 .16.(19) For free blacks, I assumed that the labor force participation rate for adult males was the same as for white males, but the labor force participation for adult females was double the rate for white females,(20) the same ratio of black females to white females which prevailed in Philadelphia in 1890, the first census to provide this information.(21) I assumed that 50 of the 55 slaves reported in Philadelphia by the census were in the labor force.(22)

The result is a labor force of 15,998. Tabulating the listings in the Philadelphia directory for 1800 yielded occupations for 7,542 individuals leaving an additional 8,456 for whom occupations are unknown. Most of them can be plausibly accounted for by estimating the number in five large occupational groups almost completely ignored by the city directory: apprentices, sailors, carters, domestic servants, and laborers.

For apprentices, I began with an 1820 estimate of 8,000 apprentices in New York City when New York's population was 123,706.(23) Assuming the same ratio of apprentices to total population in Philadelphia in 1800 would yield a total of 2,666. Assuming that twenty years earlier the crafts were slightly less commercialized, I deducted 10 percent, leaving 2,400. Of these I added 2,000 to the artisans category and 400 to unskilled and manual service (since young women were apprenticed as domestics).(24)

For sailors, I estimated a total of 2,096 based on 262 ship captains listed in the city directory and a mean crew size of eight plus the captain (included with non-manual).(25) The city directory lists only 191 mariners and 27 pilots. I added the difference (1,878) to the transportation category.

The Philadelphia directory listed only 110 carters, an impossible figure for a bustling seaport in an era when almost all goods had to be moved to and from the docks by wagon. We can make a more plausible estimate by looking at data on New York's carters. New York City licensed its carters in this period, and the New York City directory listed 1,200 carters in 1806.(26) Philadelphia had a more populous hinterland close to the city than New York, so it seems likely that it had at least as large a proportion of carters in its population as New York. If carters made up the same proportion of Philadelphia's population as they did in New York, there should have been 618 of them, so I added 508 carters to the total for transportation.

The directory lists no servants and only a handful of individuals in other service occupations. An 1830 statement by the Philadelphia Society for the Encourgement of Faithful Domestics claimed "at least the same number of domestics as houses in the city."(27) That estimate seems high. The 1880 census (the first believed to give relatively accurate data for domestic servants) reports ratios of servants to families of less than 1-to-4 in most northern cities.(28) Servants made up 3.6 percent of the local population in Philadelphia in 1870 and 3.9 percent in 1880.(29) With approximately 7,200 households in 1800, using the 1830 estimate would mean 7,200 servants, while using the 1880 ratio of servants to population yields 1,607 servants. Presumably, before the impact of mass immigration and industrialization, the proportion of servants in Philadelphia would have been higher than in 1880. So an estimate of 2,000, only slightly above the 1880 ratio, seems like a conservative guess.(30)

I assumed unskilled laborers made up at least the same proportion of total population as in 1870 (3.1 percent).(31) While the city was less industrial in 1800, critical activities such as unloading ships and construction were even more labor intensive in 1800 than later. To equal 3.1 percent of the population, laborers would have numbered 1,277 in 1800, so I added 1,075 laborers to the 202 already included in the directory.

With these additions for apprentices, sailors, carters, servants, and laborers, there were 595 individuals still unaccounted for. It seems unlikely that the city directory would have missed many merchants, doctors, shopkeepers, or other business owners or professionals. These still unaccounted-for individuals were probably mainly in manual occupations. However, I made the more conservative assumption of distributing them in the five occupational categories of Table 2 in the same proportion as the rest of the population.

A margin of error is obviously possible with these methods. In order to insure that such potential errors did not bias the results in the direction of overstating the number of unskilled, I made conservative estimates of the number of servants and the distribution of the unallocated remainder. The figures presented here thus still probably slightly understate the proportion of unskilled labor in the Philadelphia labor force.

History Department Pittsburgh, PA 15260

ENDNOTES

1. Gary B. Nash, The Urban Crucible: Social Change, Political Consciousness, and the Origins of the American Revolution (Cambridge, MA, 1979).

2. Among the most widely cited is, Allan Kulikoff, "The Progress of Inequality in Revolutionary Boston," William and Mary Quarterly 27 (July 1971): 375-412; Jackson T. Main, The Social Structure of Revolutionary America (Princeton, 1965) makes extensive use of tax records, but places less emphasis on class tensions.

3. Perhaps the most noteworthy example of this genre is Billy G. Smith, The "Lower Sort" Philadelphia's Laboring People, 1750-1800 (Ithaca, 1990).

4. Charles G. Steffen, The Mechanics of Balitimore: Workers and Politics in the Age of Revolution, 1763-1812 (Urbana, 1984); Howard B. Rock, Artisans of the New Republic: The Tradesmen of New York City in the Age of Jefferson (New York, 1979).

5. Nash, The Urban Crucible, esp. 387-91, 395-401.

6. Kulikoff, "The Progress of Inequality," esp. 377, 380-87.

7. Steffen, The Mechanics of Balitimore, 13-17; Paul G. Faler, Mechanics and Manufacturers in the Early Industrial Revolution, Lynn, Massachusetts, 1780-1860 (Albany, 1981), 149; Cynthia J. Shelton, The Mills of Manayunk: Industrialization and Social Conflict in the Philadelphia Region, 1787-1837 (Baltimore, 1986), 83; Rock, Artisans of the New Republic, 13, 15-16n13.

8. J. Franklin Jameson, The American Revolution Considered As a Social Movement (Princeton, 1926).

9. Stanley Lebergott, Manpower in Economic Growth: The American Record Since 1800 (New York, 1964), p. 510. Many critics, especially feminist scholars, argue that the percentage should be much higher. Methods of estimating early labor force composition conventionally employed by economic historians exclude most women except slaves. For an insightful discussion of this gender bias in conventional labor market categories and a creative attempt to analyze the economic significance of unpaid female labor, see Jeanne Boydston, "To Earn Her Daily Bread: Housework and Antebellum Working-Class Subsistence," Radical History Review 35 pp. 7-25 and "Home and Work: The Industrialization of Housework in the Northeastern United States from the Colonial Period to the Civil War," Ph.D. diss. (Yale, 1984). See also Robert W. Smuts, "The Female Labor Force: A Case Study in the Interpretation of Historical Statistics," Journal of the American Statistical Association 55 (March 1960): 71-79.

10. Nash, The Urban Crucible, 391.

11. Gary Nash, Billy G. Smith, and Dirk Hoerder, "Laboring Americans and the American Revolution," Labor History 24 (Summer 1983): 414-439. This essay is a response to Hermann Wellenreuther, "Labor in the Era of the American Revolution: A Discussion of Recent Concepts and Theories," Labor History 22 (Fall 1981): 573-600. Wellenreuther subsequently discounts the importance of the undercounting emphasized by Nash, Smith, and Hoerder. Wellenreuther, "Rejoinder," Labor History 24 (Summer 1983): 440-454.

12. Lee Soltow, "Socioeconomic Classes in South Carolina and Massachusetts in the 1790s and the Observations of John Drayton," South Carolina Historical Magazine 81 (October 1980): 283-305; Sharon V. Salinger, "Colonial Labor in Transition: The Decline of Indentured Servitude in Late Eighteenth Century Philadelphia," Labor History 22 (Spring 1981): 165-191, esp. 175, 182. Using a demographic model, Salinger estimates the size of the Philadelphia labor force in 1800 at 14,336. She adds 13% of the population to the taxables to represent those under 21 who were employed, and assumes that 25% of the adult females were in the labor force. However, the labor force calculations are not a central part of her argument. They are central to Soltow's essay, although his emphasis is different from the one presented here. Soltow argues that the number of unskilled is much larger than those included in tax lists and city directories, but that this is not evidence of class division because they are mostly young. An anonymous reader to an earlier draft of this essay complained that the calculation of occupations tabulated as percentage of population "is something that on the face of it is obvious." Simple? Yes. Obvious? The failure of so many otherwise quantitatively-sophisticated scholars to include such calculations suggests that it is not obvious.

13. American Social History Project, Who Built America? Working People and the Nation's Economy, Politics, Culture, and Society, Volume 1 (New York, 1989), p. 145.

14. Salinger, "Colonial Labor in Transition," does this. Her approach needs to be more widely adopted.

15. Among the most prominent are Sean Wilentz, Chants Democratic: New York City and the Rise of the American Working Class, 1788-1850 (New York, 1984); Alan Dawley, Class and Community: the Industrial Revolution in Lynn (Cambridge, MA, 1976); Steven J. Ross, Workers on the Edge: Work, Leisure, and Politics in Industrializing Cincinnati, 1788-1890 (New York, 1985); Bruce Laurie, Artisans into Workers: Labor in Nineteenth-Century America (New York, 1989); Steffen, The Mechanics of Balitimore; Faler, Mechanics and Manufacturers.

16. Return of the Whole Number of Persons Within the Several Districts of the United States (Washington, D.C., 1801), p. 2A.

17. J.D.B. DeBow, Compendium of the Seventh Census (Washington, D.C., 1854), pp. 55, 69, 128. DeBow's figures do not distinguish between white males and free blacks, but the number of free blacks was quite small, and there is no reason to suggest a different labor force participation rate for them.

18. Based on calculations from the 1870 census, the first census to include data useable for this purpose, the labor force participation rate for females 16 and over in Philadelphia was .218, but the 1870 census is widely believed to have undercounted female occupations. Ninth Census, Volume 1: The Statistics of the Population of the United States (Washington, D.C., 1872), pp. 654, 768-69, 794. The .25 female Philadelphia labor force participation rate is also consistent with national estimates for 1800 I am developing in my work-in progress: Richard Oestreicher, "A Social History of American Workers, 1800-1900," Chapter 1 and Appendix A (in author's possession).

19. Stanley Lebergott, Manpower in Economic Growth, p. 53.

20. For the age and gender structure of the free black population of Philadelphia, I used 1820 data, since the 1800 census did not report it. The 1820 census breaks down ages of free blacks into under 14, and 14 to 25, while the breakdown for whites was 10 to 15 and 16 to 25. Rather than attempt to correct for this discrepancy, I treated free black 14 and 15 year olds as adults, and omitted the 10 to 13 year olds. Presumably the higher labor force participation rates for the 14 and 15 year olds (.885 and .5 versus the rate of .15 used for white 14 and 15 year olds) offsets the omission of free black 10 to 13 year olds. In any event, the total free black population was only 10.2 percent of Philadelphia's 1800 population, and the 10 to 15 year age bracket was less than 10 percent of the free black population, so the discrepancies in reporting ages for free blacks affect less than 1 percent of Philadelphia's population. Census for 1820 (Washington, D.C., 1821), p. 17.

21. Report on the Population of the United States at the Eleventh Census: 1890, Part II (Washington, D.C., 1897), pp. 127, 710. Among white females 22.0 percent of the total female population was reported in the labor force, among black females 43.2 percent.

22. All of these slaves were adults, many probably aged. Pennsylvania had passed a gradual abolition law in 1780 under which the children of slaves were free.

23. Rock, Artisans of the New Republic, p. 314. Using Rock's figure, the ratio of apprentices to population in New York City in 1820 is 6.5 percent. Using Charles Steffen's tabulations of the annual number of apprenticeships in Baltimore in the 1790s, and assuming a mean length of seven years, the ratio of apprentices to population in Baltimore in the mid 1790s was 4.9 percent. Charles G. Steffen, The Mechanics of Baltimore, pp. 14, 32.

24. In Philadelphia between 1771 and 1773, 18.1 percent of apprenticeships were for girls apprenticed to "housewifery." Ian M. G. Quimby, Apprenticeship in Colonial Philadelphia (New York, 1985).

25. Stanley Lebergott estimates an average of one crew member per twenty tons on American ships in the early 1800s. See, "Labor Force and Employment in 1800," in Volume 30 of Studies in Income and Wealth: Output, Employment, and Productivity in the United States After 1800 (New York, 1966), pp. 166-67. Billy G. Smith reports an average of 178 tons for ships constructed in Philadelphia in 1796. Smith, "Struggles of the 'Lower Sort,'" p. 225. Applying Lebergott's ratio to this ship tonnage yields a crew size of nine.

26. Rock, Artisans of the New Republic, pp. 29, 211-12.

27. "Address to the Public of the Society for the Encouragement of Faithful Domestics" (Philadelphia, July 20, 1830), p. 1 as quoted by Lebergott, "Labor Force and Employment in 1800," p. 203.

28. David M. Katzman, Seven Days a Week: Women and Domestic Service in Industrializing America (Urbana, 1981), p. 286.

29. The Statistics of the Population of the United States, Volume 1 (Washington, D.C., 1872), pp. 380, 794; Statistics of the Population of the United States at the Tenth Census, Volume 13 (Washington, D.C., 1883), pp. 424, 894.

30. Since 400 apprentices were also allocated to domestic service, the total for the category becomes 2,400, but that is still only one-third the one-servant-per-household estimate of the 1830 society report.

31. The Statistics of the Population of the United States, Volume 1 (Washington, D.C., 1872), pp. 380, 794.