Qualities and Effective-quantities of Slaves in New Orleans

A hedonic price index and a quality-adjusted quantity index are constructed in order to provide a clearer picture of how slave quality evolved. Controlling for slave quality, slave prices did not rise as much as previously thought. This is due to an increase in slave quality, especially during the earlier 1800s. Properly accounting for the effect of the 1808 prohibition of slave importation, there is evidence of a decline in slave quality. Finally, regressions reveal the puzzling fact that light skinned males sold at a premium, even though they would have been more immune to the local disease environment. INTRODUCTION Debate about the profitability of slavery and the necessity of the Civil War has raged since the 1860s. Contemporary debate is characterized by Genovese (1989 [1965]) who claimed that slavery was inefficient and doomed to failure, and Fogel and Engerman (1974), who emphasized the profitability of slavery. Properly answering this question centers on calculating the costs and benefits of slave capital, slave agricultural production, and slave society. In this paper I examine the first of these, focusing on the proper measurement of slave prices and quality. Slave prices depended upon the changing mix of slave qualities. To do investigate this, I employ a technique similar to that of Kotlikoff (1979), but extend his analysis to the issue of changing slave quality. Estimating real wealth in the antebellum South depends to a great extent on the value of slaves. Consequently, having accurate slave values is of particular interest. For example, if real slave prices had been rising due to an increase in quality then estimates of inflation are overstated, and real wealth understated. In this paper I use hedonic methods to provide a clearer picture of the movement of slave prices that is not due to any changes in slave quality. A byproduct of these procedures is that one can compute quality-adjusted quantities as well as prices. I show that slave prices did not increase as much as the unaltered data indicate. This implies that previous calculations of antebellum inflation and the net wealth of the South need revision. More importantly, I show that the quality of slaves---quality as measured by their purchasers---was decreasing. These factors, coupled with the decreasing prices of cotton and sugar, imply that the future of the slave economy was bleak. THE DATA My main data source is The New Orleans Slave Sample, 1804-1862, collected by Fogel and Engerman and used in their 1974 book Time on the Cross. Southwestern Economic Review 162 Each bill of sale for a slave in the New Orleans market---there were approximately 135,000 such sales---was notarized and kept in various notarial offices throughout the city, until they were gathered in the New Orleans Notarial Archives in 1867. Each bill of sale contained information about sellers, buyers, means of payment, and more importantly characteristics of the slaves. From these bills, one can observe the gender, skill level, age, occupation, and complexion of the slave, as well as whether they were sold as a part of a family unit. The Fogel and Engerman dataset is a sample of over 5,000 slaves from these bills. Following Kotlikoff (1979) I exclude those records that contain slaves sold in batches, because they list identical prices for each slave in the group. These prices are average prices and therefore do not indicate which slaves or which characteristics were more valued I have almost 3000 observations of slave prices (out of the original 5000) covering the years 1804-1862. This is an average of 51 slave sales per year, 13 per quarter, or 4 per month. I deflate all prices by an index representing the average price of commodities in the New Orleans area. Compiling and cross-checking data from 45 different periodicals, Cole (1938) produced a price index for New Orleans wholesale prices for the years 1800-1861. The range of goods for which the report prices is extensive; one hundred different commodities' prices were recorded, ranging from different grades of cotton and sugar, to molasses, lumber, pork, corn, furs, beef, cheese, gin, etc. Because of the ongoing War of 1812 (which lasted into 1815 in New Orleans), I do not have data on agricultural prices for the years 1812-14. For the purposes of my study, I use Cole's weighted all-commodity series as a measure of the general price level for the Louisiana region. AGGREGATION AND THE HETEROGENEITY PROBLEM Movements in slave prices can easily be described by taking averages in each period. These numbers are easy to interpret. For example, of the sample of 23 slaves sold in period 1, the average price was $373, 30% of the slaves were male, and the average age was 17 years. Dividing the average slave prices by the 1804 average price and multiplying by 100 yields the simple (naïve) index of relative slave prices. There are serious problems with such an aggregation procedure, because you run the risk of aggregating over heterogeneous product, so to speak. Each individual is unique, so each slave sale represents a change in quality. When slave prices and qualities are changing, how much of the increase in price is due to inflation, and how much to changes in the qualities of the slaves? This heterogeneity problem is addressed using hedonic regression techniques. A hedonic regression is fancy terminology for any regression of the price of a good onto its characteristics. The price of a slave depends on his characteristics. Throughout I will make use of regressions of the logarithm of slave prices on a set of slave characteristics and (sometimes) time dummy variables. In his groundbreaking 1979 study of slave prices, Kotlikoff used the set of slave characteristics described in Table 1 below, to which I add a series of geographic variables. Several of these variables deserve further elaboration. The variable MTHCRED indicates how many months of credit were offered to the buyer for purchasing a slave. There are several ways that interest can be reported. For example, when you buy a product, you can either agree to the purchase price of, say $1,000, plus ten percent interest payable at the end of the year. Or, you may report the Qualities and Effective-Quantities of Slaves in New Orleans 163 purchase price as $1,100 payable at the end of the year. Thus, increases in the length of credit should increase the reported price. The variable GUAR relates to Louisiana's “redhibition” laws. Under these laws, a buyer could sue for a refund the purchase price of the slave, if he could establish that, at the moment the slave was bought, the slave suffered from physical