Appendix II Demand for Labor, 1662-1679

While historians have not ignored the intensification of the production process in explaining the tremendous increase in tobacco production across the seventeenth century, they have noted that overall growth in the Chesapeake was basically driven by putting an increased number of hands to work on more land. ¹ The analysis of tobacco productivity suggests that significant improvements could be obtained through acquisition of servants. Although planters might have been constrained in short-run plans for intensification decisions, perhaps in the intermediate- or long-run they responded differently in their plans for extensification. (For more on this contrast between intensification and extensification, see Appendix VI.)

In order to analyze the demand for labor we need to have fairly good data on the quantity and quality of labor supply and the price paid for labor. Before the 1690s Virginia planters met their labor needs primarily through the market for white (primarily English) servants, but thereafter they increasingly turned to the market for African slaves. ² The lack of good data on the actual annual supply of slaves until well into the eighteenth century, vitiates a rigorous analysis of the demand for slave or combined servant/slave labor, or the transition from servant to slave labor, in the seventeenth century. In addition, we only have good data on the actual supply of white servants for the years 1662-1679 (discussed below) and thus our analysis of the demand for labor is restricted to these years. One positive note, however, is that since there was no strictly simultaneous trans-Atlantic labor market because of the time constraints on communication and transportation‑-or, as economists would say, the supply and demand equations were predetermined‑-we need not consider the many factors which affected labor supply in England in order to understand planter demand. ³

On the price paid for labor, the best proxy we have is Russell R. Menard’s annual series on the value of a male servant with four or more years left to serve, which he finds most reliable for the years 1662-1709. ⁴ In employing the series we presume a fairly well-developed market for servants across the Chesapeake. Although this subject deserves far more attention, those who have studied the evidence agree with Richard B. Morris that “in the tobacco provinces the buying and selling of servants and the hiring of them out on wages was as common as the marketing of the sotweed.” ⁵ However, if the physical transfer of servants from one master to another has been exaggerated, competition for freshly imported servants would undoubtedly have helped establish a current price for labor.

Menard’s servant value series does have several potential problems. In the years prior to 1681, servant values were usually given in probate and other records in pounds of tobacco and to maintain consistency Menard converted all extraneous prices in currency into tobacco equivalents by dividing by the current tobacco price. Because the value of the servant in pounds of tobacco may be highly correlated with current tobacco prices, we will need to isolate the impact of current tobacco prices as an independent variable in order to isolate the significance of our other price models. ⁶ (See Figure 4.) Also Menard groups his servant values in three- and four-year averages (perhaps because there were not enough observations in any one year to justify annual values) and this may tend to introduce a degree of negative autocorrelation. (On autocorrelation, see Appendix IV.)

Figure 4

Although we still have much to learn about the nature of the market in bond servants in the Chesapeake, we do know quite a bit about the supply of “indentured” servants, those servants who signed a contract or “indenture” specifying terms of service before leaving England. Historians have reached a general consensus that these indentured servants were for the most part Englishmen quite representative of the lower-middle and upper-lower classes of English society seeking to better their condition. ⁷ The work of David Galenson shows furthermore that in England and the colonies the market in indentures functioned to a degree, however imperfect, like a modern labor market. ⁸

Unfortunately, social and economic historians, in devoting so much attention to these indentured servants, have neglected those servants who came to the Chesapeake without indentures and served not by any written contract but by the “custom of the country,” specified after the mid-seventeenth century by colonial statute. Although scholars of early America have yet to provide any good estimates of the exact ratio of unindentured to indentured servants, unindentured servants certainly were a highly significant part and perhaps even a majority of the bond labor force. ⁹ We do know that these unindentured servants were substantially younger, more likely of non-English origin, less skilled, and in general of lower labor quality than indentured servants and included among their number throughout the seventeenth century many “spirited” or kidnapped children. ¹⁰ Some Chesapeake historians have noted evidence of a declining quality among the unindentured servants over the second half of the seventeenth century especially after 1680‑-more boys, women, Irishmen, and felons. ¹¹ A greater supply of unindentured servants might have driven down the price of labor simply by both flooding the market and providing a lower quality good. However, changes in quality might have had minimal impact on the value of labor if, as Galenson notes, the seventeenth-century servant market, whether by indenture or custom, compensated for lower quality with increased years of service in order to equalize prices across servants. ¹²

On the actual supply of servants, we have very poor data on total numbers of white bond servants for any particular year let alone enough to create an annual time series. We are better able to estimate the total numbers of unindentured servants because, beginning in 1660 in Virginia and 1662 in Maryland, masters began to bring their unindentured servants into court to have their ages adjudged. It is not quite clear how or why this practice started in Virginia because there were no statutes requiring registration until 1662. Certainly, if there were increasing controversies over terms of service, masters had a definite incentive to have their servants’ ages adjudged when they were as young-looking as possible because, by the custom of the country, younger servants served longer terms (a precedent codified in Maryland as early as 1639 and Virginia as early as 1643). ¹³ In 1658 the Virginia Assembly mandated that the county courts be the judges of servant ages, but surely the courts had always acted as such judges. On the other hand, although regular registration in Maryland did not start until 1662, the Maryland Assembly had passed a statute on registration as early as 1654, requiring that every master should bring all his servants into court to have their ages adjudged if unindentured or to have their indentures recorded. But neither requirement seems to have been fulfilled in Maryland until a penalty was imposed in 1661. ¹⁴

Regardless of origins of registration, during the period 1662-79, the general practice appears to be that masters who brought or imported a servant into the county‑-and who expected the servant to serve more than the minimum number of years specified by the custom of the country for an adult servant‑-were required to bring the servant into the county court to have the servant’s age adjudged. During this period the time limit for registration was either expressed in terms of months or court sessions, varying from three to four months or two to three court sessions. The penalty likewise varied: loss of one year’s time, one thousand pounds of tobacco, or reduction of service to the minimum established by custom for adult servants. Since by statute additional years of servitude could be acquired for any unindentured servants below the age of 22 in Maryland and even among indentured servants this age group comprised two-thirds to three-quarters of the flow, we may expect that registration captured a large percentage of the total unindentured flow. In Virginia, on the other hand, where additional years were only acquired for those below the age of 16 until 1666 and 19 thereafter, a far smaller proportion of the unindentured flow was likely registered. ¹⁵ We are able to determine the number of servants registered annually for several years for counties (Charles, Talbot, Sumter, and Prince George in Maryland and Lancaster, Northumberland, York, and Middlesex in Virginia) comprising about a fifth of the total Chesapeake tithables for the late seventeenth century. ¹⁶ Presuming these counties, when aggregated, to be representative of the entire Chesapeake and the proportion of the flow captured to remain fairly constant, we have a fairly good proxy for unindentured servant flow.

For the higher quality, indentured servant flow, we are in a far worse position for estimating total numbers. The only good time-series data available on indentured servants for the seventeenth century are the Bristol registers for the years 1654-86, with data from years after 1679 either missing or suspect. ¹⁷ Since Bristol emigration made up only a part (perhaps 10-25%) of the total flow of indentured servants to the Chesapeake and was far outweighed in importance by Middlesex and London, and as some evidence suggests that outport merchants followed different market strategies than London merchants, we cannot fairly assume that Bristol indentured servant flows were representative of the overall indentured flows. ¹⁸ Furthermore, since the Bristol registration system was not designed simply to record indentures but to enforce the registration of all emigrants leaving Bristol for servitude in the colonies, we might presume that the Bristol registers capture not just the higher quality indentured flow but combined high and low quality labor supply. However, although Bristol registration requirements may indeed have led to the curtailing of spiriting and perhaps the indenturing of servants who in the past may have been shipped out without indentures, the law did not totally eliminate the unindentured flow from Bristol and we may assume that the Bristol servant flow fell somewhere between the select nature of a mere register of indentures and the comprehensive nature of a register of all emigrants. ¹⁹ Regardless of what the Bristol registers represent, they are the only source capable of capturing relative changes in either the indentured or total servant flow in the seventeenth century and remain an integral part of any analysis of the demand for labor in the seventeenth-century Chesapeake.

In addition to these measures of servant importation, we can also include changes in the tithable population as a measure of local labor supply although such changes would reflect the effect of young males coming of age as well as free immigration and emigration.

Other factors impacted planter demand for bond servants. For unindentured servants, demand was a function of the customs specified by statute, including years of service and the magnitude of freedom dues. For indentured servants, demand was affected by the particulars of the indenture including years of service and special privileges as well as customary privileges like freedom dues. With only ambiguous evidence on such changes apart from customary length of service, we will focus on changes on the impact of length of service on the value of labor and simply presume the other factors are constant over the period under consideration. ²⁰

Each additional year of service increased the value of a servant although the greatest impact of additional years occurred up to four years’ service, since the benefits of longer terms were for the most part balanced by the lower quality of labor of younger servants. By including only servants with four or more years left to serve, Menard’s servant value series tends to eliminate much, though not all, of the distortion due to changes in years of service. ²¹ Assuming no sampling bias (in other words, the average value reflects a servant with the average number of years left to serve in any particular year), we still have to consider statutory changes in years of service which the planters themselves recognized increased the value of servants. ²² In the period 1660-6, both Maryland and Virginia significantly increased the customary years of service for younger unindentured servants. The exact impact of these statutory changes on the average length of servitude prescribed for an unindentured servant is difficult to determine since it depends on the age distribution among unindentured servants. For the period 1658-79, Lorena S. Walsh shows a fairly constant age distribution among male servants whose ages were identified in Charles County, Maryland. ²³ Assuming a similar age distribution (10-14 35%, 15-19 45%, 20-24 20%) across Maryland and Virginia we can estimate the aggregate impact of statutory changes in customary years of service. In Maryland, the average term for an unindentured servant would have increased by one year from 6.55 to 7.55 years in 1666, while in Virginia the term would have increased from 6.54 to 7.86 years in 1662 and to 8.55 years in 1666. Further assuming three-quarters of the unindentured servants went to Virginia during these years (based on the proportion of tithables between Maryland and Virginia), we can approximate the impact of changing custom by a step function which increases from 0 to 1 in 1662 and to 1.75 in 1666.

Some historians have suggested that demand for labor might have risen in the second half of the seventeenth century as a result of fewer losses from mortality due to seasoning. However, there is little evidence to suggest when changes, if any, occurred and this easy generalization based on anecdotal evidence has been questioned by Darrett B. Rutman and Anita H. Rutman. ²⁴

Test Results

The best-fit equation for servant prices for the years 1662-1679 is presented in Table 3. Figure 5 graphically reveals the closeness of the match between predicted and actual measures of servant prices.

Table 3

Figure 5

Current tobacco price (PRICE) was negatively correlated with the value of servants suggesting that, for every tenth of a penny less that merchants offered for tobacco, ceteris paribus, planters paid about 37 pounds of tobacco more for a servant. This negative correlation reflects the tendency for the real value of imported commodities to maintain their value to some degree in the face of the inflationary/deflationary pressures of rising/falling tobacco prices, although the negative correlation may also capture to some degree the impact of current prices on demand.

With any inflationary/deflationary effect held constant, the negative coefficients on the average and minimum price models (the same models found to best explain tobacco productivity) reveal the basic similarity between tobacco production and labor-demand decisions. They show that rising average and minimum tobacco prices led to lower servant values and thus decreased demand for labor and that falling average and minimum tobacco prices led to higher servant values and thus increased demand for labor. Ceteris paribus, a fall in the lagged four-year running average price of tobacco (PL4) of a tenth of a penny led to an increase in servant value of 78 pounds of tobacco and a fall in the eight-year minimum price (PMIN8) of the same amount led to an increase of 98 pounds of tobacco. Thus‑-taking the impact of current, average, and minimum prices together‑-as a result of the fall in tobacco prices, holding all other factors constant, masters in 1676, when prices were relatively low, would have paid 1194 pounds of tobacco more for a servant than they did in 1662 when prices were relatively high. ²⁵ On the other hand, rising prices after Bacon’s Rebellion would have resulted in masters in 1679 paying 68 pounds of tobacco less per servant than they had in 1676. ²⁶

However, before jumping to the conclusion that tobacco productivity and demand for labor completely paralleled each other, one should note the reversal of the sign on the check variable (CHECK) between Tables I and III. The check variable was designed to measure whether those years in which tobacco prices fell from the previous year had any different impact from years in which prices rose or stayed the same, regardless of the amount or rate of fall. In contrast to the stimulating effect on tobacco productivity suggested by the positive coefficient on CHECK in Table I, the negative coefficient in Table III suggests that such a fall in prices reduced the demand for labor. Thus, a short-run fall in tobacco prices had a statistically different impact than a short-run rise in tobacco prices, reducing the value of a servant and thus the demand for labor while simultaneously stimulating tobacco productivity. It is unclear whether this reflects a constraint on an expansionary impulse due to income limitations in times of falling tobacco prices, a retreat into greater dependency on family labor supply while simultaneously intensifying household production, or something else altogether, but it suggests that extensification and intensification were not exactly parallel responses of planters.

The negative coefficient on unindentured servants (UNIN2) confirms expectations based on quantity and/or quality effects, but the positive coefficient on Bristol indentured servants (BRIS1) suggests that the quality effect overrode the quantity effect. However, the positive coefficient is also consistent with the belief that Bristol merchants tended to supply servants to the Chesapeake at times when, due to falling tobacco prices, planter demand for labor was high.

The geographic (SUM3) and demographic (DPOP1, POP1) expansion of the Chesapeake all had a stimulating effect on the value of a servant, but these were balanced by a strong negative time trend (YEAR). The negative time trend also tends to challenge arguments which suggest that a decline in seasoning losses in the second half of the seventeenth century led to an increased demand for servants.

Interestingly, the step variable designed to captured the effect of statutory changes in length of servitude (STEP) had an unexpected negative coefficient suggesting that statutory efforts to increase the time of servitude for unindentured servants actually led to a decrease in the value of servants of 434 pounds of tobacco after 1666. Why longer terms should have decreased the value of a servant is not readily apparent. Perhaps STEP simply captures some nonlinear element in the negative time trend unrelated to changes in custom.

These findings overall show that planters responded to falling and rising average tobacco prices and minimum tobacco prices much as they did for tobacco productivity–increasing demand for labor when tobacco prices fell and decreasing demand for labor when tobacco prices rose. However, the reversal of the sign on CHECK suggests that for planters intensification was a far more clear-cut response to falling tobacco prices than extensification. Recognizing that the purchase of additional servants entailed greater indebtedness/dissaving and risk of catastrophic failure not associated with simply intensifying production with present labor supply, one might readily comprehend why planters opted for intensification. Although the ability to pay might have been offset by tightening the belt in other areas‑-including increased frugality and/or self-sufficiency and selling off or delaying alternative investments‑-seemingly such options were not as desirable as simply increasing overall effort.