The author is Assistant Professor of Economics, University of Delaware, Newark, DE 19716.

Earlier drafts were presented at the Economic History Seminar, University of Pennsylvania, 1985; the Economics Seminar, University of Delaware, 1986; the Econometric Society Meetings, New Orleans, 1986; the Economics Seminar, University of North Carolina-Chapel Hill, 1987.

I wish to thank the participants of these seminars and Gary Becker, Stanley Engerman, Robert Fogel, David Galenson, Robert Gallman, Henry Gemery, Claudia Goldin, Saul Hoffman, Ken Koford, Ken Lewis, Robert Margo, Thomas Mroz, Richard Steckel, the editor, and the many anonymous referees for helpful comments on earlier drafts. I also wish to thank the City Archives of Philadelphia and the Pennsylvania Historical Society for providing microfilm of manuscript contract evidence.

¹ See Grubb, Farley, “The Incidence of Servitude in Trans-Atlantic Migration, 1771–1804,” Explorations in Economic History, 22 (07 1985), pp. 316–39.CrossRefGoogle ScholarGrubb, Farley; “Redemptioner Immigration to Pennsylvania: Evidence on Contract Choice and Profitability,” this JOURNAL, 46 (06 1986), pp. 407–18;Google ScholarHansen, Marcus Lee, The Atlantic Migration 1607–1860 (Cambridge, Mass., 1940), pp. 102–6;CrossRefGoogle ScholarHerrick, Cheesman A., White Servitude in Pennsylvania (Philadelphia, 1926), pp. 254–66.Google Scholar

² For recent studies of the market for indentured servants, see Galenson, David W., White Servitude in Colonial America (Cambridge, Mass., 1981);Google ScholarGrubb, Farley, “The Market for Indentured Immigrants: Evidence on the Efficiency of Forward-Labor Contracting in Philadelphia, 1745–1773,” this JOURNAL, 45 (12. 1985), pp. 855–68;Google ScholarGrubb, Farley, “Colonial Labor Markets and the Length of Indenture: Further Evidence,” Explorations in Economic History, 24 (01. 1987), pp. 101–6.CrossRefGoogle Scholar

³ For descriptions of the redemptioner system see Galenson, , White Servitude, pp. 13–15;Google ScholarGeiser, Karl F., Redemptioners and Indentured Servants in the Colony and Commonwealth of Pennsylvania (New Haven, 1901), pp. 5–76;Google ScholarGrubb, , “Redemptioner Immigration,” pp. 407–18;Google ScholarSmith, Abbot E., Colonists in Bondage (New York, 1947), pp. 3–42.Google Scholar The two most cited eyewitness accounts of the redemption auction are in Mittelberger, Gottlieb, Journey to Pennsylvania in the Year 1750 and Return to Germany in the Year 1754, Handlin, Oscar and Clive, John, trans. and eds. (Cambridge, Mass., 1960), pp. 16–18,Google Scholar and in letters written by Henry Melchior Muhlenberg reprinted in Diffenderifer, Frank R., “The German Immigration into Pennsylvania Through the Port of Philadelphia, and ‘The Redemptioners’,” Pennsylvania German Society, 10 (1899), pp. 189–93.Google Scholar

⁴ See the assessment in McCusker, John J. and Menard, Russell R., The Economy of British America, 1607–1789 (Chapel Hill, 1985), pp. 242–44.Google Scholar

⁵ Quoted in Durnbaugh, Donald F., “Two Early Letters from Germantown,” Pennsylvania Magazine of History and Biography, 84 (04. 1959), pp. 231–33.Google Scholar Sauer became a respected Pennsylvania-German printer. Previously, scholars have traced the redemption system only as far back as 1728, see Smith, A. E., Colonists in Bondage, p. 21;Google ScholarDiffenderifer, , “German Immigration into Pennsylvania,” pp. 172–73, 200–1. Advertisements for the sale of German immigrant servants in the American Weekly Mercury in Philadelphia indicate that the redemption system had not been adopted by German immigrants in 1722. Therefore, the redemption system appears to have begun sometime between 1722 and 1724.Google Scholar

⁶ See newspaper accounts reproduced in Diffenderifer, , “German Immigration into Pennsylvania,” pp. 200–1.Google Scholar

⁷ See the discussion in Diffenderifer, , “German Immigration into Pennsylvania,” pp. 19194;Google ScholarGeiser, , Redemptioners and Indentured Servants, pp. 64–68.Google Scholar For an example of restrictive search clauses written into redemptioner transportation agreements see the passage contract for the ship Commerce reproduced in Strassburger, Ralph B., Pennsylvania German Pioneers (Norristown, 1934), vol. 3, pp. 131–34.Google Scholar

⁸ See the discussions in Eddis, William, Letters from America (Cambridge, Mass., 1969), pp. 39–40;CrossRefGoogle ScholarTrautmann, Frederic, “Pennsylvania Through a German's Eyes: The Travels of Ludwig Gall, 1819–1820,” Pennsylvania Magazine of History and Biography, 105 (01 1981), pp. 40–41. The speed of sale was derived from the evidence cited in Tables 1–3.Google Scholar

⁹ See also the discussions in Diffenderifer, , “German Immigration into Pennsylvania,” pp. 170–216, 238–55; Galenson, White Servitude, p. 15; Geiser, Redemptioners and Indentured Servants, pp. 56–70; Smith, A. E., Colonists in Bondage, pp. 20–22, 39–41.Google Scholar

¹⁰ Reprinted in Rupp, I. D., History and Topography of Northumberland, Huntington, Miffin, Centre, Union, Columbia, Juniata and Clinton Counties, Pa. (Lancaster, 1847), p. 57.Google Scholar For examples of similar denunciations in later years of the redemption trade see, Muhlenberg, Henry Melchior, The Journals of Henry Melchior Muhlenberg, Tappert, Theodore G. and Doberstein, John W., trans. (Philadelphia, 1942), vol. 2, p. 157;Google ScholarEddis, , Letters from America, pp. 38–40;Google ScholarFearon, Henry Bradshaw, Sketches of America (3rd edn., London, 1819), pp. 148–51.Google Scholar

¹¹ In addition, competition among shippers would cause passage debts to equal the cost of transportation, see Grubb, Farley, “The Market Structure of Shipping German Immigrants to Colonial America,” Pennsylvania Magazine of History and Biography, III (01 1987), pp. 27–48;Google ScholarGrubb, , “Redemptioner Immigration,” pp. 411–18.Google Scholar

¹² Derived from the sources cited in Tables 1–3 and The Pennsylvania Gazette 1728–1789, 25 vols. (reprinted, Philadelphia, 1968).Google Scholar

¹³ Mittelberger, , Journey to Pennsylvania, pp. 16–18.Google Scholar For the geographic distribution of servant purchasers in the Delaware Valley, see Grubb, Farley, “Immigrant Servant Labor: Their Occupational and Geographic Distribution in the Late Eighteenth-Century Mid-Atlantic Economy,” Social Science History, 9 (Summer 1985), pp. 249–75.Google Scholar

¹⁴ The 1785 to 1804 servant document was designed only to record German redemptioner contracts. The 1771 to 1773 servant document recorded indentured, redemptioner, and resident servants, see Grubb, Farley, “Servant Auction Records and Immigration into the Delaware Valley, 1745–1831: The Proportion of Females Among Immigrant Servants,” Proceedings of the American Philosophical Society, 132 (09 1988), forthcoming. Redemptioners were distinguished from other servants in the 1771 to 1773 records by their immigrant status and by the distinction of not being “assigned” to a master by a current owner. Unlike all other servants, redemptioners were not the property of anyone when they entered the servant auction, and they assigned themselves to their chosen master.Google Scholar

¹⁵ See Grubb, , “Colonial Labor Markets,” p. 102, for a discussion of the secular decline in contract lengths for indentured servants.Google Scholar

¹⁶ Within the 1787 to 1804 sample, controlling for age, the secular rate of decline in contract length was 3 percent for every five years, see Table 3. This estimate, however, is only marginally significant.

¹⁷ See the discussions in Diffenderffer, , “German Immigration into Pennsylvania,” pp. 141–315; Galensonp, White Servitude, p. 13; Mittelberger, Journey to Pennsylvania, pp. 16–18; Smith, A. E., Colonists in Bondage, pp. 20–22.Google Scholar

¹⁸ Accumulation of wealth may not have been significantly reduced by entering America as a servant. Average personal savings rates were around 15 percent in the early ninetenth century. The savings rates for the poorest segment of society was probably lower. See Adams, Donald R. Jr, “Earnings and Savings in the Early 19th Century,” Explorations in Economic History, 17 (04 1980), pp. 118–34;CrossRefGoogle ScholarAdams, Donald R. Jr, “The Standard of Living during American Industrialization: Evidence from the Brandywine Region, 1800–1860,” this JOURNAL, 42 (12 1982), pp. 914–15.Google Scholar

¹⁹ The yearly income of resident workers was estimated by multiplying the average daily wage rate by an estimated number of workdays per year. The average number of days a fully employed laborer could expect to work per year is not known. Smith, B. G., “Material Lives,” p. 188, suggests that “five of every six work days or ten of every twelve months” would be a reasonable estimate. The estimate of resident artisan incomes by B. G. Smith and Adams were substantially different. Because these different estimates could not be reconciled, both were reported in Table 1. Only resident wage rates which excluded room and board were considered.Google Scholar

²⁰ Between 1793 and 1807, a few ship manifests recorded the occupations of German immigrants arriving in Philadelphia, see Strassburger, , German Pioneers, vol. 3. Among 579 immigrants with recorded occupations, 58 percent were artisans or professionals, 26 percent were farmers, and 11 percent reported no trade. Because only the poorest 45 percent of the immigrants entered servitude, farmers and those without a trade may have predominated among redemptioners. The contract lengths and prices of 66 immigrants, who reported artisan trades and were traced in the servant records, were significantly lower than for the average servant reported in Table 1 (t–statistic of 1.66 on the difference in average contract length, and t–statistic of 2.63 on the difference in average contract price). The comparison indicates that the average redemptioner was probably not a skilled artisan. Differences in productivity and work amenities between immigrant servants and free resident laborers that could not be directly measured may explain any residual income differences. For example, free labor may have received a premium for uncertainty in the duration and variance of employment. By contrast, servants may have received a premium for surrendering to the employer the right to inflict moderate corporal punishment and the right to make labor-leisure decisions at the margin. The net effect on relative compensation of these differing work amenities is uncertain. In addition, language and custom differences may have reduced immigrant productivity and thus lowered redemptioner relative compensation per reported skill level.Google Scholar

²¹ Quoted in Strassburger, , German Pioneers, vol. 1, p. xxxvii; Diffenderffer, “German Immigration into Pennsylvania,” p. 192. Mittelberger, Journey to Pennsylvania, pp. 16–18, gives a similar description of the redemption auction in 1750.Google Scholar

²² See Grubb, , “The Market for Indentured Immigrants,” pp. 855–68.Google Scholar

²³ For a discussion of hedonic indices see, Rosen, Sherwin, “Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition,” Journal of Political Economy, 82 (01 1974), pp. 34–55;CrossRefGoogle ScholarGriliches, Zvi, ed., Price Indexes and Quality Change (Cambridge, Mass., 1971).CrossRefGoogle Scholar

²⁴ The evidence was constructed by matching the names of redemptioners in the servant sale records to the names of immigrants, ships, and captains in ship passenger lists and custom house records. Merging these data sources was necessary to separate immigrant servants from resident servants and to expand the information about each servant, such as determining the servant's age, family structure, port of origin, literacy, and the time taken to form a contract after arrival. The matching procedure differed for the British and German samples and for the two sample periods, 1771 to 1773 and 1787 to 1804. Therefore, the model was estimated separately for the two time periods and the two ethnic groups.

²⁵ The sample of German redemptioner servants from 1771 to 1773 used in Table 2 was constructed in the same manner as the evidence used in Heavner, Robert O., “Indentured Servitude: The Philadelphia Market, 1771–1773,” this JOURNAL, 38 (09 1978), pp. 701–13;Google ScholarHeavner, Robert O., Economic Aspects of Indentured Servitude in Colonial Pennsylvania (New York, 1978), pp. 48–101.Google Scholar Although constructed in the same fashion, the sample in Table 2 is much larger, includes more independent variables, and differs in the number of contracts with education and training than the sample reported by Heavner. Heavner mistakenly identified the contracts in the sample as indentured servant rather than redemptioner. This mistake led Heavner to apply the wrong model to the evidence. He used contract price as the dependent variable and all other information as independent variables. Heavner interpreted this regression as measuring differences in servant human capital through systematic changes in contract prices. This error explains numerous puzling results, such as insignificant coefficients on literacy, skill, and training, but a large price reduction for education, and lower compensation for Germans migrating from London compared with Rotterdam. Correctly interpreted, Heavner's regression explains the variance in transportation debts rather than servant productivity. There is little reason to expect passage debts among adults to vary with literacy, skill, or training. The 43 percent lower passage debt for redemptioners asking for education may simply reflect the half-fare charged children under age 16. Finally,₁ lower passage debts for Germans sailing from London would be expected because passage fares from London were lower than from Rotterdam.

²⁶ Indentured servant contracts were relatively uniform. Contract lengths were typically fixed in whole years with almost no variance in contract lengths among adults. Few indentured servants migrated in family units, and few indentured contracts contained stipulations other than customary freedom dues.

²⁷ The surviving German passenger ship lists from 1771 to 1773 only consist of loyalty oath signatures. Loyalty oaths were only required of German immigrant males above age 16, see Strassburger, , German Pioneers, vols. 1–2. Matching the names of German redemptioners with the names on the loyalty oaths produced a German servant sample of adult men, exclusively. This sample was used in Table 2.Google Scholar

²⁸ Longer contracts associated with education and training may have been caused either by the lower productivity of these servants or by the cost of acquiring these marketable skills. The cost of education and training to the master was both out-of-pocket expenses and the lost work time both of the servant and master. Education was stipulated in contracts either as a stated number of months of schooling or as being taught to read and write.

²⁹ For example see, Galenson, , White Servitude, pp. 97–113.Google Scholar

³⁰ Mittelberger, , Journey to Pennsylvania, p. 18, stated, “The very young, between the ages of 10 and 15, have to serve until they are 21.” He also stated that children under age 5 “must stay in service until they are 21 years of age.” Table 3 indicates that Mittelberger's claim was more of a rule of thumb for providing information to prospective emigrants than a mandatory or customary rule for assigning contract lengths to children.Google Scholar

³¹ For example see, Galenson, , White Servitude, pp. 105–7;Google ScholarGalenson, David, Traders, Planters, and Slaves (Cambridge, Mass., 1986), pp. 53–114;CrossRefGoogle ScholarFogel, Robert W. and Engerman, Stanley L., Time on the Cross (Boston, 1974), pp. 75–77.Google Scholar

³² For a discussion of forecast error in the indentured servant market see, Grubb, , “The Market for Indentured Servants,” pp. 866–67; Grubb, “Colonial Labor Markets,” pp. 102–3.Google Scholar

³³ See, for example, Galenson, , Traders, Planters, and Slaves, pp. 71–92.Google Scholar

³⁴ An alternative explanation would be that the sick postponed their sale, hoping to recover sufficiently from the arduous voyage to raise their labor value and thus shorten their contracts. This behavior would cause shorter contracts to occur early in the auction because healthy servants who could bargain for shorter contracts would choose to sell themselves first. See the description in Mittelberger, , Journey to Pennsylvania, p. 16.Google Scholar Sickness, however, was not very important. Debarkation morbidity was under 3.5 percent among German passengers, see Grubb, , “Morbidity and Mortality,” pp. 573–77.Google Scholar

³⁵ Purchasers who were listed as merchants were the exception to the general occupation-contract length pattern. The merchant category was relatively broad and the kind of work expected of the servant is unclear. Merchants may have been wealthy enough to purchase household servants to do domestic chores. Thus merchants may have purchased younger, inexperienced servants who could not negotiate shorter contracts because of their lower productivity.

³⁶ The productivity difference could be considerable. For example, Ulster Irish were able to negotiate 23 percent shorter contracts than Irish redemptioners from Dublin, and German redemptioners departing from London were able to negotiate 12 percent shorter contracts than Germans leaving through Rotterdam. All British ports were significantly different from one another, except for London versus South Irish ports. The partial F-test was 2.49 for Ulster versus South Ireland (significant at the 0.11 level), 4.08 for Ulster versus London (significant at the 0.04 level), and 0.66 for South Ireland versus London (insignificant). Among indentured servants sold in Philadelphia, Ulster servants were also found to be the most productive, see Grubb, , “Colonial Labor Markets,” p. 105.Google Scholar