Prompted in part by the well-publicised work of Thomas Piketty (2014), economists have recently exhibited renewed interest in the issue of wealth and income inequality. In addition to Piketty’s work on the long-run tendencies toward increasing inequality, a number of other studies have explored the long-run persistence of social and economic status (Clark 2014, Long and Ferrie 2013). But less attention has been devoted to the way in which economic and political shocks may affect the persistence of social status.
In a recent paper, we examine the impact that the disruptions of the American Civil War and emancipation had on the wealth distribution by locating in the 1860 census those heads of household who were among the top 5% of wealth holders in 1870 (Dupont and Rosenbloom 2016). In doing so, we are able to determine what happened to these individuals over the course of this turbulent decade.
Our results suggest that while there was an entrenched southern planter elite that retained their economic status after the war, the turmoil of the 1860s opened greater opportunities for mobility in the South than was the case in the North.
As a result, there was much greater turnover among wealthy southerners than among comparably wealthy northerners.
For those interested in the impact of the Civil War and emancipation on the economic development of the postbellum South, these results require some revision of widely accepted views. It appears that the 1860s introduced considerably greater levels of turnover in the South than in other parts of the US. More broadly, these issues have salience for discussions of the dynamics of wealth and income inequality, and the role that economic and political shocks may play in the persistence of social status.
The economic impact of the Civil War
The southern slave economy permitted a small number of wealthy planters to accumulate extraordinary fortunes. The 1860 census data show that the median wealth of the richest 1% of Southerners was more than three times higher than for the richest 1% of Northerners. In a largely rural and agricultural economy, slavery eliminated the labour constraints that limited the size of northern farms and allowed for a much greater concentration of wealth (Wright 1970, 1978, Ransom 1989). Of course, there were different vehicles for wealth accumulation in the two regions – in the North, real estate accounted for two-thirds or more of property ownership among the top 45% of wealth holders, but personal property (which included slaves) made up close to three-fifths of total wealth held by the top 10% of Southerners.
The Civil War and emancipation destroyed an immense amount of Southern wealth. Given the prominent role of slaves among the property of the wealthiest southerners, it is reasonable to conjecture that the effects of emancipation may have been most pronounced at the top of the wealth distribution. While northern wealth holders above the 55th percentile experienced an approximately 50% increase in property holding over the 1860s, the value of property owned by southerners fell by nearly 75%. The drop was especially pronounced for personal property; those in the top 10% of the Southern wealth distribution experienced a 90% drop in the value of their personal property, while real property wealth was cut approximately in half. As a result, after the war the relative shares of real and personal property in the South converged toward those in the North, with real property making up 60-70% of wealth, at least among the wealthier household heads.
Creating a linked sample
Recent advances in the ability to search for individuals in the manuscript census reports of 1860 and 1870 make it possible to determine what happened to individual wealth holders during the 1860s. We have created a linked sample by starting with those at the top of the wealth distribution in each region (North and South) in 1870 and attempting to locate them in the 1860 census.
To construct our linked sample, we used the IPUMS to identify household heads with total 1870 property holdings that placed them among the top 5% in each region. Since it is still necessary to hand-collect the linked data, we chose the top 5% to provide a large enough sample of linked individuals for regional differences to be visible, while keeping the data collection effort manageable.
To increase the chances of locating individuals in our backward linkage approach, we restricted our analysis to household heads age 25 and older. To avoid the distorting effects of wealth transfers in old age, we imposed an upper limit of 75 years. This procedure generated a sample of 4,419 household heads (2,520 in the North and 1,899 in the South) to be linked backward to the 1860 census. We then searched for each of the individuals in our sample in the 1860 census using the Ancestry.com database based on first and last name, and year of birth calculated from reported age in 1870. For each linked individual we noted several aspects of link quality, including whether the place of birth was the same in both censuses, and whether we were able to identify other household members in both censuses.
The results of our search are summarised in Table 1. In total, we were successful in linking 43% of household heads, or 1,918 individuals. Based on indicators of link quality, our confidence in the linkage process is quite high. In all but 51 cases, place of birth was identical in both censuses, and in 85% of cases we were able to find other household members (a spouse and/or child) that matched across the two censuses. Our linkage rate appears to be considerably higher than other studies that have sought to link 19th century census records. Ferrie (1996), for example, reports a success rate of just under 20% when linking forward from the 1850 to the 1860 census. The relatively high rate of success in linking in our sample likely reflects the greater stability of high wealth individuals.1 Backward linkage also eliminates the negative effects of mortality on linkage, which is especially important in light of evidence that about 8% of white males between ages 13 and 43 (in 1860) died in the war.2
Table 1. Characteristics of the linked sample
Patterns of wealth mobility in the 1860s
Transitions across the wealth distribution are summarised in Table 2. The rows of each table indicate the location in the 1860 wealth distribution, while columns correspond to 1870 wealth levels. To locate individuals in the 1860 wealth distribution we have used percentile cut-offs for total property ownership in each region calculated from the IPUMS 1% sample. In the lower panel of the table, we report the value of each cell as a percentage of the column total.
It is apparent that there was considerably more turnover among the ranks of top southern wealth holders than among northern wealth holders. While more than half of those in the top 5% of northern wealth holders had been in the same group in 1860, just one-third of top southern wealth holders in 1870 had enjoyed a similar status in 1860. Roughly the same proportion of the top 5% in each region was drawn from the next stratum of wealth holders in 1860 (90th to 95th percentile). On the other hand, our data suggest that the turmoil of the Civil War decade created much greater opportunities for those with moderate wealth in 1860 – between the 55th and 90th percentiles – to move up to the top of the wealth distribution. Nearly 40% of the wealthiest southerners in 1870 had been in this group in 1860, compared to less than one quarter of the richest northerners.
Table 2. 1860 wealth of 1870 top 5% of wealth holders, by region
The American Civil War and slave emancipation wiped out a vast amount of wealth, but they also impacted the distribution of wealth in important ways that are only revealed by examining the fates of individual households in the 1860s. By using the Ancestry.com database, we are able to improve on past efforts to link wealthy individuals across census years. Where previous studies have been confined to a few counties, we are able to study a random sample of wealthy individuals in all locations and link them across census years regardless of geographic mobility. The evidence confirms that wealth provided some insulation from the shocks of that decade, but we also find that it was a less effective insulator in the South than in the North.
Campbell, R B (1982) "Population persistence and social change in nineteenth century Texas: Harrison County, 1850-1880", The Journal of Southern History, 48: 185-204.
Clark, G (2014) The son also rises: Surnames and the history of social mobility, Princeton University Press.
Dupont, B and J Rosenbloom (2016) “The impact of the Civil War on southern wealth holders”, NBER, Working Paper.
Ferrie, J P (1996) "A new sample of Americans linked from the 1850 Public Use Micro Sample of the Federal Census of Population to the 1860 Federal Census Manuscript Schedules", Historical Methods, 34: 141-156.
Long, J and J P Ferrie (2013) “Intergenerational occupational mobility in Great Britain and the United States since 1850”, The American Economic Review, 103: 1108-37.
Piketty, T (2014) Capital in the twenty-first century, Harvard University Press.
Ransom, R L (1989) Conflict and compromise: The political economy of slavery, emancipation and the American Civil War, Cambridge University Press.
Schaefer, D (1985) "A statistical profile of frontier and new south migration, 1850-1860", Agricultural History, 59: 563-578.
Vinovskis, M A (1989) "Have social historians lost the Civil War? Some preliminary demographic speculations", The Journal of American History, 76: 34-58.
Wiener, J M (1976) "Planter persistence and social change", The Journal of Interdisciplinary History, 7: 235-260.
Wright, G (1970) "‘Economic democracy’ and the concentration of agricultural wealth in the Cotton South, 1850-1860", Agricultural History, 44: 63-93.
Wright, G (1978) The political economy of the Cotton South: Households, markets, and wealth in the nineteenth century, Norton.
 Schaefer (1985) used a backward linkage approach similar to ours and also found a high linkage rate of approximately 50% between the 1860 and 1850 censuses.
 The estimated death rate is from Vinovskis (1989: 38). Previous studies (Wiener 1976, Campbell 1982) have recognised death as a factor in non-persistence and typically look for potential heirs, but this approach is obviously limited in cases where there were no surviving heirs.