## Abstract

This paper identifies an exogenous variation in post–Civil War policy to examine the effect of land reform on racial inequality. The Cherokee Nation, located in what is now Oklahoma, permitted slavery and joined the Confederacy in 1861. During postwar negotiations, the Cherokee Nation agreed to provide free land for its former slaves. Using linked data that follow former slaves in the Cherokee Nation from 1880 to 1900, I find that racial inequality was lower in the Cherokee Nation in both 1880 and 1900. Land and the associated increase in incomes may have facilitated investment in both physical and human capital.

## I. Introduction

THERE are high levels of racial inequality in the United States. In 2013, the median black household earned 62% of the income and held 8% the assets of the median white household (DeNavas-Walt & Proctor, 2014; Kocchar & Fry, 2015), an economic inequality that dates back to the American Civil War. The previously enslaved had little property and incomes that were, on average, a quarter that of whites during the late 1860s (Higgs, 2008). Abolitionists and policymakers argued that providing freed slaves with land was critical for their economic advancement. However, no land redistribution bill became law. Southern freedmen's dreams of “forty acres and a mule” were, as Du Bois (1903) wrote in his critique of the Freedmen's Bureau, “destined in most cases to bitter disappointment.” References to the promised land and plow animal have remained in the American public discourse, reflecting a belief that the enduring inequality between blacks and whites could have been reduced if Reconstruction-era policymakers had granted each formerly enslaved family its own farm. In this paper, I consider the role of land redistribution in ameliorating racial inequality by exploiting a plausibly exogenous variation in postbellum land policies between the Cherokee Nation and the United States. The Cherokee Nation, located in what is now Oklahoma, permitted the enslavement of people of African descent and joined the Confederacy in 1861. During postwar negotiations, the United States insisted that the Cherokee Nation provide land for its former slaves.1

The role of land reform and redistribution in alleviating the poverty of historically disadvantaged groups has long been debated. While a large literature discusses the key role of property rights over land in economic development (Feder & World Bank, 1988; Besley, 1995; De Soto, 2000, 2001; Jacoby, Li, & Rozelle, 2002; Goldstein & Udry, 2008; Besley & Ghatak, 2010; and Galiani & Schargrodsky, 2010), empirical analyses of land redistribution have been rare (Bruce & Migot-Adholla, 1994). Scholars remain similarly conflicted on the potential role of land distribution in decreasing American racial inequality. An emerging “new orthodoxy” in the 1960s and 1970s held that Reconstruction was “fatally flawed” due to policymakers' failure to meaningfully redistribute Southern land from whites to the recently freed slaves (Belz, 1973, 107). More recent scholarship has questioned the conventional wisdom that land alone would have altered the economic conditions of former slaves, arguing that the Southern social and political environment would have counteracted any equalizing advantage that land offered (Higgs, 2008; Woodman, 1977, 2001; Engerman, 1980, 1982; Foner, 2007; Ransom, 2005).

An empirical analysis of the effects of free land on racial inequality has previously eluded researchers. Without a group of former slaves who were treated with access to free land, the construction of a counterfactual to analyze the economic impact of postbellum land distribution is difficult. The Cherokee Nation's land policy provides a fresh perspective. As a “domestic dependent nation,” the Cherokee Nation was in the unique position of being Southern with respect to economic organization while simultaneously adhering to a different legal regime.2 This allowed for the development of slavery in both the South and the Cherokee Nation before the Civil War but different policies toward the freed slaves after the war. To analyze the effects of land distribution on racial inequality, I digitized the 1860 Cherokee Nation Slave Schedules and a sample of the 1880 Cherokee Census. The Dawes Enrollment Cards, collected by the federal government between 1899 and 1908, facilitated construction of a linked sample that connects blacks in the Cherokee Nation from the 1880 Census to the 1900 U.S. Census.

This data set provides the first quantitative evidence on the economic status of former slaves in the Cherokee Nation. Despite wealth and human capital levels that were comparable to those of Southern blacks in 1860, former slaves in the Cherokee Nation prospered in the decades following the Civil War. Using several different metrics of asset ownership and agricultural output, I find they had higher absolute levels of wealth and income than Southern freedmen in 1880 and 1900. Racial inequality was strikingly lower in the Cherokee Nation than in the South. Table 1 summarizes these differences. The impact of land distribution went beyond a transfer of assets. Access to free land was associated with increased levels of investment in both physical and human capital. In 1880, Cherokee freedmen were more likely to plant fruit trees, an important form of immobile capital, than blacks in the South. Black Cherokee children had higher levels of school attendance and literacy in 1900 that were facilitated by higher parental income. Within the Cherokee Nation, the children of the richest black families in 1880 were more likely to become literate adults than the children of the poorest families. Overall, access to free land appears to have had a substantial positive effect on the economic conditions of former slaves in the Cherokee Nation. The impact was felt for several generations and suggests that land distribution policies may play an important role in alleviating racial economic inequality.

Table 1.
Black Populations in the Cherokee Nation and South, 1880 and 1900
Cherokee NationSouthDifferencee
A. 1880
Literate (%) 14 36 −21
[0.35] [0.48] [0.00]
Farm owners only
Acreage 41.80 35.00 6.80
[56.27] [31.09] [0.00]
Wealth in livestock (1880 $) 496.94 140.93 356.01 [706.47] [170.16] [0.00] B. 1900 Male household heads Own Home (%) 81 21 60 [0.39] [0.41] [0.00] Adults Literate (%) 55 37 18 [0.49] [0.48] [0.01] Children (10 to 18) Attended school (%) 41 35 [0.49] [0.48] [0.00] Months attended (>0) 5.43 3.95 1.48 [2.28] [1.99] [0.00] Literate (%) 59 47 12 [0.49] [0.50] [0.00] Cherokee NationSouthDifferencee A. 1880 Adults Literate (%) 14 36 −21 [0.35] [0.48] [0.00] Farm owners only Acreage 41.80 35.00 6.80 [56.27] [31.09] [0.00] Wealth in livestock (1880$) 496.94 140.93 356.01
[706.47] [170.16] [0.00]
B. 1900
Own Home (%) 81 21 60
[0.39] [0.41] [0.00]
Literate (%) 55 37 18
[0.49] [0.48] [0.01]
Children (10 to 18)
Attended school (%) 41 35
[0.49] [0.48] [0.00]
Months attended (>0) 5.43 3.95 1.48
[2.28] [1.99] [0.00]
Literate (%) 59 47 12
[0.49] [0.50] [0.00]

Standard errors are in brackets under sample means. $p$-values are in brackets under difference calculation.

Sources: Cherokee Nation calculations use the linked sample of Cherokee freedmen. Southern calculations for 1880 use the 1880 IPUMS and 1KF. 1900 U.S. calculations use the 1900 IPUMS with Indian Oversample. (IPUMS data from Ruggles et al., 2019.)

## II. Historical and Contemporary Context of Land Redistribution

### A. Land Redistribution in the Cherokee Nation and South

Shortly after the Civil War began in April 1861, abolitionists began calling for the federal government to confiscate Confederate land for the use of freed slaves. Southern land was seized under the Direct Tax Act (1862), the Confiscation Act (1862), and the Captured and Abandoned Property Acts (1863, 1864). In January 1865, General William Tecumseh Sherman issued Special Field Order Number 15, which authorized freed slaves to establish forty-acre farms on federally controlled land along the coasts of South Carolina, Georgia, and Florida. He later provisioned them with broken-down military mules (Hahn, 2008). As rumors of “forty acres and a mule” spread, former slaves began to migrate to the Sherman Reserve. The Civil War ended in April 1865. By June of that year, 40,000 freedmen believed they would receive official title to their nascent farms. However, President Andrew Johnson revoked Sherman's field order, and much of the land was quickly restored to its Confederate owners (Oubre, 2012).

Proponents of land distribution pressed their case and emphasized the freedmen's immediate need for land. General O. O. Howard, commissioner of the Freedmen's Bureau, wrote, “My task is a hard one, and I am convinced that something must be done to give these people and others the prospects of homesteads (Howard, 1908, 240).” Congressman Thaddeus Stevens introduced a bill in March 1867 to assign each liberated male slave forty acres and $100, arguing that homesteads were critical “not only [for] the happiness and respectability of the colored race, but their very existence” (Stevens [1867], 1998). The bill was defeated, despite the support of George Washington Julian, chair of the House Committee on Public Lands, and Charles Sumner, chair of the Senate Committee on Foreign Relations (Foner, 1980). Amid these debates over land redistribution, the federal government began postwar negotiations with the Cherokee Nation. In August 1861, the Cherokee Nation had declared its allegiance to the Confederacy: “The Cherokee people had its origins in the South; its institutions are similar to those of the Southern States.… [They] cannot but feel that their interests and destiny are inseparably connected with those of the South. The war now raging is a war of Northern cupidity and fanaticism against the institution of African servitude.”3 The alliance was tested by a series of military defeats, a Confederate request to borrow$50,000 in gold, and a conscription order for Cherokee men. Amid this turmoil, Union forces began a campaign into the Cherokee Nation and surrounded its capital. In August 1862, they arrested Principal Chief John Ross and transported him to Washington, DC. The Cherokee Nation split into two factions. One, led by Stand Watie, remained staunch supporters of the Confederacy. The other, directed by Ross in exile, allied itself with the Union. On June 23, 1865, Watie surrendered, the last Confederate general to do so. A federal commission was tasked with negotiating a treaty between the United States, the Union Cherokees, and the Confederate Cherokees. The U.S. government issued a statement demanding “80 acres for each person of color, formerly a slave of any Cherokee, who remain in said country.”4 Both Cherokee parties opposed this, and some refused to free their slaves. Cherokees called for the freedmen to be moved to a segregated area. Others believed that the United States should be responsible for removing all freedmen from the Cherokee Nation, “as it has freed them.”5

The United States did not support these proposals and was providing rations to former Cherokee slaves, many of whom had sought federal protection. Army officials arranged for freedmen to enter into wage and sharecropping contracts, while others began to farm abandoned lands. Cherokees continued to pressure U.S. officials to remove the freedmen. However, union officials believed they could become independent: “With lands set apart for the freedmen of the Indian nation, and the Freedmen located upon them, and a government, military or civil, organized and executed for their protection, they will, beyond doubt, soon become an industrious, intelligent and happy population.”6 The freedmen were authorized to continue occupying the abandoned lands (Littlefield, 1978). U.S. negotiators played the Cherokee factions against each other, and both sides eventually acceded to the federal government's wishes.7 Under the resulting 1866 treaty, the Cherokee's former slaves were granted “all the rights of native Cherokees” (Kappler, 1904, 944).8 With these rights and federal support, the Cherokee freedmen had three crucial advantages over their Southern counterparts. First, each freed Cherokee slave could claim land in the public domain.9 Second, because federal officials assisted the freedmen in becoming “reasonably well supplied with farming implements and seed,” freedmen who claimed land had initial working capital (Littlefield, 1978, 23). Third, the U.S. government enforced the Cherokee freedmen's property rights and ensured that they were not removed from claimed land (Littlefield, 1978). By the next season, many former slaves had established their own farms. Qualitative evidence suggests that the 1866 treaty had an impact on racial inequality in the Cherokee Nation. An 1892 editorial in the Afro-American Advocate noted that “the opportunities for our people in that country [the Cherokee Nation] far surpassed any of the kind possessed by our people in the U.S.”10

### B. Land and Economic Inequality

The absence of Southern land redistribution meant that former slaves in the South had initial wealth and income levels that were far lower than those of whites. Higgs (1982) found that black total property holdings were just 1/36 those of whites in 1880. This ratio improved slightly to 1/26 by 1890, and 1/23 by 1900 but was still only 1/16 in 1910.11 The income ratio was 1/4 in 1867 and 7/20 in 1900. Engerman (1982, 218) noted that Southern blacks' dearth of capital at emancipation “meant that, no matter how rapidly incomes grew, they would remain far behind those of whites for [a] very prolonged perio[d].” Property accounted for about one-third of national income in the decades following emancipation; closing the racial income gap without land redistribution would be difficult (Alston, 1990). DeCanio (1979) calibrated the impact of a land distribution policy on the time path of inequality. Allocating each freedmen household “forty acres and a mule” would have increased blacks' starting average wealth level to 60% that of whites and average income at emancipation to about half that of whites. Even in the absence of discrimination or market imperfections, a majority of the black-white income gap during the late nineteenth and early twentieth centuries was explained by initial capital inequality.

The majority of Southern black farmers became renters or sharecroppers: 80% in 1880 (Ransom & Sutch, 1973) and 75% in 1910 (Higgs, 2008). An economic system comprising primarily white landowners and black tenants can hinder black income growth. Because landlords were able to expropriate gains by not renewing a lease, tenancy arrangements lowered a farmer's incentive to make land-specific, productivity-enhancing investments. If landowners were unable to verify their tenants' effort and care over investments, then they too had less incentive to fund investments relative to a system of owner-operators. In the presence of such inefficiencies, incomes could be increased by transferring land from the owner to the tenant, reuniting investment and control over investment (Besley & Ghatak, 2010).

Black tenant farmers had limited options for purchasing land. Poorly functioning Southern credit markets shut many freedmen out from land ownership (Higgs, 2008; Mandle, 1978; James, 1981; Jaynes, 1986; Carlton & Coclanis, 1989; Ransom & Sutch, 2001). Land redistribution could have allowed blacks to own land without having to borrow or accumulate sufficient savings. McKenzie (2002) estimated that a sharecropper needed to save 10% of his income for eighteen consecutive years in order to purchase a farm. Landownership, as opposed to tenancy, may have also facilitated access to credit markets by providing collateral on loans, allowing freedmen farmers to finance risky or lumpy investments (Banerjee & Iyer, 2002).

Becker and Tomes's (1979, 1986) model of the relationship between parents' and children's incomes suggests that grants of land at emancipation could have increased the income and wealth of subsequent generations. Parents can improve their children's wealth by leaving bequests of land, property, or other assets. They can increase future earnings by investing in human capital. If land distribution increased former slaves' income levels, black parents could have overcome the obstacles their children faced in obtaining an education. The ability to self-finance education, particularly in the presence of credit constraints, can be a critically important mechanism for future income growth and economic mobility (Grawe & Mulligan, 2002). Margo (1987), Fishback and Baskin (1991), and Moehling (2004) all found that a family's economic status had a significant and positive impact on black children's educational outcomes in the South.

## III. Data

### A. The Linked Sample of Cherokee Freedmen

I constructed a linked census sample that contains detailed economic and demographic information on Cherokee freedmen households in 1880 and 1900. I relied on three data sources: the 1880 Cherokee Census, the 1900 U.S. Census, and the Dawes Enrollment Cards. The Cherokee Nation oversaw the collection of the 1880 Census. Detailed agricultural data, including acreage, crop yields, and livestock ownership, were recorded for each family, and demographic data, such as age, literacy, and occupation, were collected for each person. Both the agricultural and population data were listed on the same schedule. I collected and digitized a 60% sample of this census—all blacks in the Cherokee Nation and approximately 50% of the remaining population. It contains 11,899 people, 1,784 of whom are freedmen.

Between 1899 and 1907, the United States compiled a roll of all Cherokee citizens. The commission facilitating this process was charged with accounting for all people enumerated on the 1880 Census, including those who had died. Name, age, sex, and family members were recorded for each person on a Dawes Enrollment Card. The card's unique identification number was written next to the person's entry in the original 1880 Cherokee Census, providing a documented link between 1900 and 1880. The linking procedure had three basic steps. First, the 1880 Census provided the card number. Second, I used the card to discover the name and family members of the person in 1900. Third, I used this information to locate the individual in the 1900 U.S. Census via ancestry.com. I callected 1900 Census information for each linked person and his or her entire household. This census included a variety of economic and demographic information, including occupation, literacy, school attendance, and home ownership.12 I located 84% of people with Dawes information in the 1900 Census. Because both first and last names are known for both 1880 and 1900, women whose names changed with marriage were also linked. The linkage rate for men (84%) was comparable to that of women (82%). The data set includes 789 people from the 1880 Cherokee Census, 2,664 total individuals, and 470 households. Additional information is available in appendix I and Miller (2015).

### B. Other Census Data

To compare the Cherokee Nation to the Southern states in 1880, I use several preexisting cross-sectional samples. The 1% Public Use Microdata Sample of the 1880 U.S. Census (1880 IPUMS) contains information from the population schedules but has no numerical wealth or income data. The sample collected by Roger Ransom and Richard Sutch for their book, One Kind of Freedom (1KF, 2001), provides detailed information on farming households and matches farmers listed on the 1880 U.S. Agricultural Schedules to their respective entries in the 1880 Population Schedules.13 For analysis in 1900, I take advantage of the 1% Public Use Microdata Sample of the 1900 United States Census with American Indian Oversample (1900 IPUMS). This includes the standard 1-in-100 sample for the Southern states and a 1-in-5 sample of the American Indian Schedules, including the Cherokee Nation.

## IV. Identification Strategy

The variation in land policy between the Cherokee Nation and the Southern United States allows me to examine how the availability of free land influenced racial wealth and income inequality. For a given measure of mean wealth or income $Y¯$, I expect the racial gap to be smaller in the Cherokee Nation than in the South:14
$Y¯non-blackSouth-Y¯blackSouth>Y¯non-blackCherokeeNation-Y¯blackCherokeeNation.$
For 1880, I consider acres of land and value of livestock owned as measures of wealth. The total value of crops produced serves as a proxy for income. For 1900, I examine both home ownership and occupation. Theory suggests several mechanisms through which land distribution could have affected economic inequality. First, landowners had a greater incentive to invest in their land than tenants or landlords did. I use data on fruit trees as a measure of durable capital investment. Second, wealthier, landowning parents may have invested more heavily in their children's education than poorer families did. Literacy levels in 1880 and 1900 and school enrollment in 1900 provide information on human capital investment. I estimate
$ln(Yi)=βo+β1Blacki+β2CNi+β3(Blacki×CNi)+γXi+ɛi.$
(1)
$Yi$ denotes an economic outcome of interest for a given household. The vector $Xi$ includes demographic and land quality controls that I explore below. Black is an indicator variable equal to 1 if a household head's race is black and takes the value of 0 otherwise. $β1$ measures the location invariant effect of being black on $Y$. The CN variable is 1 if an individual lives in the Cherokee Nation and 0 if he lives in the southern United States. Its coefficient, $β2$, is an estimate of the effect of living in the Cherokee Nation relative to living in the South for a nonblack on outcome $Y$. $β3$, the coefficient on the interaction term, is the main coefficient of interest. Since the omitted category is a nonblack in the South,
$β3={E[Yi|black=0,CN=0,Xi]-E[Yi|black=1,CN=0,Xi]}-{E[Yi|black=0,CN=1,Xi]-E[Yi|black=1,CN=1,Xi]},$

and is a measure of the difference in the gaps, that is, the difference in racial inequality. A positive and significant estimate of $β3$ suggests that the racial gap was smaller in the Cherokee Nation than in the United States for the outcome of interest. For $β3$ to identify the effect of free land access on former slaves, the following conditions must hold.

### A. Differences in Land Policy Do Not Reflect Differences in Attitudes

Cherokee land policy reflected the Nation's lack of bargaining power during postwar negotiations with the United States. Relations between the Cherokees and their former slaves were strained after ratification of the treaty. In 1885, the U.S. Senate investigated the conditions of freedmen in the Cherokee Nation. A Missouri lawyer testified, “[The freedmen] do not enjoy the same rights and privileges that the balance of the nation do.… Their general treatment is very much like it has been of the colored people in the South in the past.… I should say that their treatment has been about like that of the colored people in Louisiana, Mississippi, and the other Southern States.”15 William Boudinot, the executive secretary of the Nation, stated unequivocally, “It is the policy of the nation that the two races should be separated.”16 Freedmen before the committee complained that they were denied voting rights, treated unfairly in jury trials, and had limited access to schooling. There were also complaints to the special committee and in separate petitions that Cherokees perpetrated violence against freedmen.

In 1906, another Senate committee visited the Cherokee Nation. Resistance to freedmen citizenship remained. One Cherokee woman declared, “We were forced to take them [the freedmen] in as citizens after the war of the rebellion … we do not think it was right to force all these negroes onto us as a punishment.” A lawyer from the Cherokee Nation observed, “There is a very strong prejudice between the Indian blood of the Cherokee Nation and the freedmen.” A judge succinctly summarized a popular opinion on the Cherokee freedmen's access to land: “There was just as much justice in it as there would have been if the Government had compelled the slave owners of Texas and Arkansas and Mississippi and Louisiana to have divided their lands with the negroes.”17

### B. Slaves and Slavery Were Similar in the Cherokee Nation and the South

Although no direct evidence on slaves' human capital or wealth exists, census records, slave laws, contemporary descriptions, and slave narratives indicate that Cherokee and Southern slaves lived in similar circumstances and entered freedom with similar types of human and physical capital. Based on such sources, some historians have concluded that “slavery among the Cherokees was little different than that in the white South” (Littlefield & Littlefield, 1976).

Through enumerator error, the 1860 U.S. Census Slave Schedules included slaves in the Cherokee Nation.18 Panel A of Table 2 contains summary statistics for the 1860 Cherokee data and the 5% 1860 Public Use Microdata Flat Sample of the Slave Population. The composition of the enslaved population is similar across both regions in mean age and percent female. The median slave holding, or number of people enslaved in a household, is the same. The Cherokee Nation's mean is lower due to the lack of the very large holdings that existed in the Southern states. The densities are similar, but with the higher end of the distribution truncated in the Cherokee Nation. Only 1.83% of Southern slave owners owned more than 56 slaves, the largest holding in the Cherokee Nation. Figure I in appendix II plots the kernel.

Table 2.
Comparison of Cherokee Nation and Southern United States
Cherokee NationSouth
A. Slave populations, 1859
Sampling rate 100
Population enslaved (%) 15 32.27
Female (%) 50.70 49.69
Mean age 19.15 20.13
Fugitives 14
Manumitted 17
Mean slaveholding 6.57 9.72
Median slaveholding
B. Agricultural information, 1859
Acreage (total) 102,500$a$ 56,832,153$b$
Acreage per capita 4.40 6.24
Total value of listed 2,884,350 1,081,679,455
Livestock (1859 dollars)$c$
Livestock per capita (1859 Dollars) 122.68 118.82
C. Freedmen populations, 1880
Female (%) 51.12 50.46
Mean age 19.65 20.92
Married (%)$d$ 64.73 70.81
Female-headed Households (%) 33.96; 26.13$e$ 18
Cherokee NationSouth
A. Slave populations, 1859
Sampling rate 100
Population enslaved (%) 15 32.27
Female (%) 50.70 49.69
Mean age 19.15 20.13
Fugitives 14
Manumitted 17
Mean slaveholding 6.57 9.72
Median slaveholding
B. Agricultural information, 1859
Acreage (total) 102,500$a$ 56,832,153$b$
Acreage per capita 4.40 6.24
Total value of listed 2,884,350 1,081,679,455
Livestock (1859 dollars)$c$
Livestock per capita (1859 Dollars) 122.68 118.82
C. Freedmen populations, 1880
Female (%) 51.12 50.46
Mean age 19.65 20.92
Married (%)$d$ 64.73 70.81
Female-headed Households (%) 33.96; 26.13$e$ 18

Sources: Cherokee Nation slave data are from the 100% sample of 1860 Slave Schedules for the Cherokee Nation. U.S. slave data are from the 1860 SlavePUMS. Cherokee Nation agricultural data are from an 1859 Commissioner of Indian Affairs Report. U.S. agricultural data are from 1860 U.S. Census of Agriculture. With the exception of the number of slaves, these data were estimates, not actual counts. The 1880 freedmen data are from 1880 Cherokee Census and 1880 IPUMS sample of the Southern states (IPUMS data from Ruggles et al., 2019). The South includes all states that joined the Confederacy.

$a$Acreage in cultivation. $b$Improved acres. $c$Livestock types include sheep, mules, horses, and cattle. $d$Percent married, spouse either absent or present, for people age 20 and over. $e$First result is for all families; second result is for all families except people listed as single.

Some Cherokees had plantations of 600 to 1,000 acres worked by enslaved labor (Miles, 2006). They grew crops for sale and profit. There was a distinction between field hands and house servants, with overseers directing the field hands and employing a gang labor system for larger slaveholdings.19 Slaves were hired out for manufacturing or other work when the fields were fallow or if the slave had a particularly valuable skill. This similarity in tasks suggests that Cherokee slaves did not acquire a different set of skills while in bondage. No comprehensive human capital or education measures for slaves were recorded during slavery. Table 3 reports literacy rates of Cherokee and Southern blacks by birth cohort in 1880 Census data. The literacy rates for cohorts reared under slavery were lower in the Cherokee Nation than in the South, suggesting that Cherokee slaves did not have higher levels of human capital.

Table 3.
Black Literacy Rates in the Cherokee Nation and South
Age RangeBirth CohortSouthCherokee Nation
11 to 15 1870 0.42 0.35
16 to 20 1865 0.38 0.22
21 to 25 1860 0.3 0.23
26 to 30 1855 0.23 0.14
31 to 35 1850 0.19 0.18
36 to 40 1845 0.16 0.11
41 to 45 1840 0.15 0.08
46 to 50 1835 0.15 0.07
51$+$ 1830 0.18 0.09
Age RangeBirth CohortSouthCherokee Nation
11 to 15 1870 0.42 0.35
16 to 20 1865 0.38 0.22
21 to 25 1860 0.3 0.23
26 to 30 1855 0.23 0.14
31 to 35 1850 0.19 0.18
36 to 40 1845 0.16 0.11
41 to 45 1840 0.15 0.08
46 to 50 1835 0.15 0.07
51$+$ 1830 0.18 0.09

Sources: Data are from 1880 Cherokee Census and 1880 IPUMS (Ruggles et al., 2019).

Billington (1982) compared the narratives of Indian-owned and white-owned slaves in the WPA Slave Narratives and Indian Pioneer Papers (Baker & Baker, 1996) by quantifying incidences of physical punishment, care, food availability, and attitudes toward former owners. He concluded that there were few differences between white and Indian slave owners. Naylor's (2008) analysis of slave narratives found commonalities in slave resistance in the Cherokee Nation and Southern states, including attempted escapes, theft, verbal or physical confrontation, and noncooperation. Slave owners reacted to such transgressions with denial of rewards, sale of the slave, and physical punishment. Use of the whip was practiced in the Cherokee Nation (Naylor, 2008). Physical punishment of slaves was a legally protected right of slave owners in the Cherokee Nation, and Cherokee slave laws greatly resembled those found in most Southern states. Cherokee slaves were forbidden from owning property and prohibited from learning to read or write, a law shared with about half of the Southern states. Teaching of any person of color—slave or nonslave—to read was illegal. There is no evidence of differential enforcement of these laws in the Cherokee Nation and the South, which is consistent with the claim the both Cherokee and Southern slaves had limited opportunities to obtain physical capital or formal schooling.

### C. Initial Conditions Were Comparable

Because Cherokee and Southern blacks had comparable initial levels of wealth and human capital at emancipation, any difference in the initial racial wealth gap would be the result of differences between nonblacks. Evidence indicates that nonblacks in both places shared similarities in institutions, economic activity, social organization, and wartime experiences. In addition, available data on farming demonstrate similarities between the two areas.

George Butler, a Southerner and the Cherokee's representative from the Superintendency of Indian Affairs, reported in 1859, “From their general mode of living, the Cherokees will favorably compare to their neighbors in any of the states.”20 The Cherokee Nation's government was explicitly modeled after that of the United States.21 The economy was, like the South's, primarily agricultural, and the main crops were corn, wheat, oats, and cotton. Cherokees built the same style houses and buildings as their neighbors in Arkansas and Kansas. The Cherokee Nation was home to incorporated towns, a weekly newspaper, and a Masonic lodge. Cherokees dressed in Western clothing, attended Christian churches, and bought patent medicines purported to be straight from New York (Littlefield, 1978). Cherokee citizens conducted business with stores and other establishments in the nearby Southern states, indicating integration into the Southern economy. The Cherokees also intermarried with the white population. According to an 1835 Census, approximately 8% of Cherokee households contained an intermarried white.22 Because of the Cherokees adoption of Western economic and social practices, they were deemed one of the “Five Civilized Tribes” during the nineteenth century (Frank, 2009).

Although data on agriculture are scarce in the years before the Civil War, panel B of table 2 contains the available information.23 The values of livestock owned per capita are quite close, but the acreage per capita calculations are not exactly comparable. Acres in cultivation were available for the Cherokee Nation. The Southern data reflect improved acres, or land “cleared and used for grazing, grass, or tillage, or which is now fallow, connected with or belonging to the farm” (U.S. Census, 1853). The Cherokee acreage was lower per capita. Part, if not all, of that difference can be attributed to the different acreage definitions used.

The Cherokees suffered ill effects of the Civil War similar to those in the Southern states. Seven officially recognized battles were fought in the Nation and guerrilla warfare was common (Perdue, 1979). Houses and barns were burned by the opposition of the time, and fields and fences fell into disrepair. Passing troops requisitioned or stole livestock and looted houses. Thornton (1992) estimates that the Cherokee population fell from 21,000 to 15,000 during the war.

### D. Comparable Policies That Otherwise Inhibited Black Economic Progress

Neorevisionist scholars have expressed skepticism about the efficacy of a postbellum land distribution program. As Higgs (2008, 80) summarized, “‘Forty acres and a mule’ was no panacea for the economic ills of the black population.” Conditions in the Southern United States could have inhibited any economic progress that land distribution spurred. Ransom (2005) identifies several potential impediments present in the postbellum South—including psychological trauma, disruptions in the cotton market, the withdrawal of federal troops, and credit scarcity—that also characterized the Cherokee Nation. Beyond the physical damage of the Civil War, “psychological wounds from the fighting” lingered in the Southern populace (Ransom, 2005, 367). The war was destructive for the Cherokee Nation, and its citizens continued to suffer from the after-effects. Writing in 1869, Brevet Major Jonathon Craig reported that “social demoralization consequent upon the late war, such as found to exist in the Southern United Sates, prevails to some extent” (Holland, 1956, 405). The Cherokee Nation grew cotton for sale both before and after the war, and both the Cherokee and Southern postwar economies experienced the same falling cotton prices. Federal troops remained in the South until 1876. By this time, the Federal military presence in the Cherokee Nation had already been greatly reduced. Union troops had occupied Fort Gibson in the Cherokee Nation in 1862. They remained until 1871, when most were removed except a small detachment that monitored the supply depot. The size of the detachment then fluctuated, based on federal needs, before being abandoned permanently in 1890 (Agnew, 2009). Ransom (2005) and Foner (2007) highlight the relative scarcity of Southern credit and the difficulties freedmen had in obtaining loans. Capital markets were no further developed in the Cherokee Nation. Because the National Banking Act did not apply to Indian Territory until 1890, the nearest national bank for much of this period was in Lawrence, Kansas. In the absence of a formal banking sector, merchants provided credit and became the primary source for loans within the Cherokee Nation (Hightower, 2013; Aldrich, 1979). With interest rates of 15% to 35%, and perhaps as high as 60%, Cherokee freedmen faced high costs in obtaining funds from the commercial sector (Hightower, 2013).

### E. Unaltered Group Composition

The composition of the Cherokee freedmen citizenry was not altered due to the availability of free land. While land access could have theoretically encouraged Southern slaves to move to the Cherokee Nation, restrictions of the treaty prevented such a land rush. Only former slaves of Cherokees who returned by January 1867 were eligible for citizenship. This provision was strictly enforced, and freedmen who were listed on the 1880 Cherokee Census rolls were those who could prove they were eligible for citizenship. If Southern blacks moved to the Cherokee Nation to gain access to free land, the movers would be those who had the lowest costs of moving and the most to gain from land—likely young men. Table 2 contains basic demographic characteristics of blacks in the Cherokee Nation and the South in both 1859 and 1880. A comparison finds no evidence of young black Southern men entering the Nation.

A potential problem with group composition could arise if there was a systematic difference between those who became citizens and those who did not gain postwar citizenship. There were 2,511 slaves in 1860 and an estimated freedmen population of 2,000 to 2,500 in 1866. Population growth was likely limited during the war, and these numbers suggest that there was only a small group of freedmen who were not in the Cherokee Nation at the war's close.

## V. Inequality in 1880

### A. Wealth Inequality in 1880

With their access to free land, Cherokee freedmen were more likely than Southern freedmen to own farms. Table 4 reports 1880 Census data on farm ownership rates. In the Cherokee Nation, 68% of all black male heads of households owned a farm, while 70% of all nonblack male heads did. The Southern data do not allow for a direct calculation of the ownership rate, but information from the IPUMS and 1KF samples can be used to make an estimate. The implied black farm ownership rate was 12%—less than a fifth that of the Cherokee Nation. The implied nonblack landownership rate was 52%. This suggests that access to free land was associated with a substantial closing of the racial gap in farm ownership.

Table 4.
Farm Ownership in the Cherokee Nation and the South, 1880
Farmers Who Owned Land (%)Male Household Heads Who Were Farmers (%)Implied Farm Ownership Rate (%)
BlackWhiteBlackWhiteBlackWhite
Cherokee 100 100 67.8 70.4 67.8 70.4
Nation
South 28.4 73.7 43.4 70.7 12.3 52.1
Farmers Who Owned Land (%)Male Household Heads Who Were Farmers (%)Implied Farm Ownership Rate (%)
BlackWhiteBlackWhiteBlackWhite
Cherokee 100 100 67.8 70.4 67.8 70.4
Nation
South 28.4 73.7 43.4 70.7 12.3 52.1

Source: Data are from 1880 Cherokee Census sample, 1880 IPUMS (Ruggles et al., 2019), and Ransom and Sutch (2001) sample. South includes all states that joined the Confederacy except Arkansas.

I use farm-level data for male heads of households from the 1880 Cherokee Census and the 1880 IKF sample to estimate equation (1).24 To ensure that the results are not driven by demographic or human capital differences, I control for the age, age-squared, and literacy of each farmer. I also include county-level soil fixed effects to minimize the impact that differences in land type have on the estimates.25 Robust standard errors are reported.

Farm ownership alone would have been insufficient to substantially narrow the racial income and wealth gaps. Free land could potentially induce marginal farmers with little practical experience or working capital to claim small plots with limited productive capacity. I estimate equation (1) where $Yi$ is acres in use.26 Results are in panel A of table 5. The estimates in column 1 include all farmers. Column 2 restricts the sample to farm owners. Appendix II provides two additional specifications, omitted here due to space considerations: I include a baseline case with no controls and restrict the sample to the Upper South—Cherokee Nation, Tennessee, Virginia, and North Carolina. By focusing on these states, I can exclude the possibility that the estimated differences in racial inequality are driven by regional differences in racial inequality between the Upper and Lower South. In all specifications, the estimated coefficient on the black variable confirms that blacks had farms that were substantially smaller. The interaction term, or the estimated difference in racial gaps, is statistically significant and large in magnitude, ranging from 0.31 to 0.639. Access to free land was associated with a shrinking of the racial gap in farm size by about 36 to 89 percentage points.

Table 5.
Inequality in 1880
A. ln(Acres in Use)B. ln(Value of Livestock Owned)C. ln(Crop Income)D. ln(Total Value)
12345678
Black −0.41*** −0.62*** −0.44*** −0.31*** −0.05 −0.09 −0.23*** −0.20***
[0.03] [0.06] [0.04] [0.06] [0.04] [0.06] [0.04] [0.05]
Cherokee Nation 0.01 0.07 1.35*** 1.22*** 1.14*** 1.14*** 1.26*** 1.20***
[0.10] [0.08] [0.07] [0.07] [0.05] [0.06] [0.06] [0.06]
Black $×$ Cherokee Nation 0.31*** 0.55*** 0.65*** 0.51*** 0.23** 0.31*** 0.46*** 0.46***
[0.09] [0.10] [0.08] [0.09] [0.08] [0.09] [0.06] [0.07]
Controls for farm size
Includes
All Farmers
Owners
Observations 7892 5202 7583 4928 7321 2,731 7823 5160
Adjusted $R2$ 0.13 0.14 0.44 0.45 0.51 0.57 0.55 0.58
A. ln(Acres in Use)B. ln(Value of Livestock Owned)C. ln(Crop Income)D. ln(Total Value)
12345678
Black −0.41*** −0.62*** −0.44*** −0.31*** −0.05 −0.09 −0.23*** −0.20***
[0.03] [0.06] [0.04] [0.06] [0.04] [0.06] [0.04] [0.05]
Cherokee Nation 0.01 0.07 1.35*** 1.22*** 1.14*** 1.14*** 1.26*** 1.20***
[0.10] [0.08] [0.07] [0.07] [0.05] [0.06] [0.06] [0.06]
Black $×$ Cherokee Nation 0.31*** 0.55*** 0.65*** 0.51*** 0.23** 0.31*** 0.46*** 0.46***
[0.09] [0.10] [0.08] [0.09] [0.08] [0.09] [0.06] [0.07]
Controls for farm size
Includes
All Farmers
Owners
Observations 7892 5202 7583 4928 7321 2,731 7823 5160
Adjusted $R2$ 0.13 0.14 0.44 0.45 0.51 0.57 0.55 0.58

All specifications include controls for soil type, age of farmer, and literacy of farmer. Robust standard errors are reported in brackets. Sampling weights are used. Sample includes male heads of household in the Cherokee Nation and 1KF samples who report positive values of the dependent variable in the sample counties. ***$p<0.001$, **$p<0.01$, and *$p<0.05$.

I next use the total value of livestock owned as a measure of wealth. Livestock can represent available capital, investment, or forgone consumption. Panel B of table 5 reports regression results for equation (1) where $Yi$ is the summed values of all horses, cattle, mules, sheep, and swine in 1880 dollars.27 Only farmers with positive livestock values are included. Because livestock requirements may be correlated with the size of a farm, I also control for acres in use. (The additional specifications are available in appendix II.) The estimated gap in the value of livestock is consistently and statistically significantly smaller in the Cherokee Nation.

Lower levels of livestock ownership by nonblacks in the Cherokee Nation do not drive the difference, as the estimated coefficient on the Cherokee dummy variable is positive. When southern renters and tenants are excluded, the difference in gaps remains large in magnitude.

### B. Income Inequality in 1880

Because revenues from the sale of crops were an important component of a farmer's annual income, I use the total market value of crop yields as a proxy for income. Although renters and sharecroppers paid either money or crops to their land's owner, I do not adjust their incomes to reflect this and they should be considered an upper bound. Panel C of table 5 reports estimates of equation (1) where $Yi$ is calculated as the total value of corn, cotton, oats, wheat, Irish potatoes, sweet potatoes, and fruit yields in 1880 dollars. Only farmers with positive values are included. The additional specifications are again in appendix II.

The results reported in table 5 include a covariate for the natural log of farm size. The estimated difference is now the racial income gap per acre, not per farm, and reflects differences in land productivity. In all specifications, the estimated coefficients on the interaction term are statistically significant and positive, indicating that the racial income gap was smaller in the Cherokee Nation than in the South. The estimated magnitude of the gap is 25 percentage points smaller per acre and demonstrates that the decrease in racial income inequality in the Cherokee Nation was not driven by changes in the relative size of black and nonblack farms. The estimated coefficient on Cherokee Nation is large at 1.14, suggesting that on a per acre basis, Cherokee farms produced more income than Southern farms. This is consistent with analyses that report farms in economies with high levels of land inequality are less productive than those with a more egalitarian distribution, particularly in the presence of credit constraints (Dorner, 1971, 1972; Eswaran & Kotwal, 1986; Lin, 1992). The estimated coefficient for the black variable approaches 0 and becomes statistically insignificant in this specification, which supports Wright's ([1986] 1997) finding that incomes of Southern black farm owners lagged behind those of their white counterparts due to their smaller, less fertile farms.

I construct a measure of the total liquidation value of farm products by summing the total values of livestock and crops. This can be considered as the cash a farmer could raise by selling all of his farm products. I estimate equation (1) where $Yi$ is the liquidation value of farm products. Panel D of table 5 reports these results. The estimated difference in gaps is positive and statistically significant for all specifications, including those reported in appendix II. Blacks in the Cherokee Nation have total liquidation values that are two to three times as high as blacks in the South. These results further support the hypothesis that access to free land could have substantially lessened racial income and wealth inequality following the Civil War.

These estimated differences in the racial wealth and income gaps may understate the true magnitude of the differences. By excluding nonfarmers, the estimates exclude a large number of poor blacks. An overwhelming majority of black nonfarmers worked as laborers: 85.1% of nonfarming black male heads of household in the South were laborers as were 82.9% in the Cherokee Nation. Because laborers typically earned lower incomes than farmers, including nonfarmers would likely decrease the average black income and wealth levels for both the South and the Cherokee Nation. The extent of the decrease would be proportionate to the share of laborers in the population. The high landownership rate of Cherokee blacks placed a limit on the number of laborers. The relatively larger number of Southern black laborers would likely cause the difference in the racial wealth and income gaps to increase.

## VI. Inequality in 1900

To examine racial inequality in 1900, I combine my sample of Cherokee freedmen households in 1900 with data on rural Southern households drawn from the 1900 IPUMS with Indian Oversample (Ruggles et al., 2019).28 For consistency, I consider the same Southern states included in the 1KF sample. The 1900 Census does not include any quantitative wealth or income data but did record information on home ownership. A household was recorded as owning its home if the dwelling's owner resided at the property. I also consider occupation and focus on farming to gauge the extent to which male household heads were able to leave behind wage labor for working their own farm. I restrict the sample to male heads of household and estimate a variant of equation (1):
$Pr(Yi)=Φ(βo+β1Blacki+β2CNi+β3(Blacki×CNi)+γXi+ɛi).$
(2)

$Yi$ is an indicator variable for home ownership or farming. $β3$, the difference in the gaps, remains the main coefficient of interest. The vector $Xi$ includes demographic controls for the head of household (age and literacy) and family size. Marginal effects are reported.

### A. Home Ownership in 1900

I estimate equation (2) where the dependent variable is 1 if a household's dwelling is owner occupied. Results are in panel A of table 6. The baseline specification is in column (1). In column 2, I add demographic controls. Column 3 also includes a dummy variable equal to 1 if the head of household reports an occupation of farmer. There is a clear pattern: blacks in general are less likely to own their home than nonblacks. This is true even after controlling for differences in human capital attainment and occupation, a finding generally present in twentieth-century data (Collins & Margo, 2001). Racial inequality remained lower in the Cherokee Nation. In the baseline case, the gap was around 25 percentage points smaller. Even after controlling for literacy and farming, the gap was still 12 percentage points.

Table 6.
Inequality in 1900
Panel A: Home Owner $=$ 1Panel B: Farmer $=$ 1Panel C: Home Owner $=$ 1
12345678
Black −0.33*** −0.26*** −0.25*** −0.11*** −0.10*** −0.36*** −0.31*** −0.31***
[0.01] [0.01] [0.01] [0.01] [0.01] [0.01] [0.01] [0.01]
Cherokee Nation 0.38*** 0.44*** 0.44*** 0.09*** 0.10*** 0.35*** 0.40*** 0.40***
[0.02] [0.02] [0.02] [0.02] [0.02] [0.02] [0.02] [0.02]
Black $×$ Cherokee Nation 0.24*** 0.14*** 0.13** 0.13*** 0.11*** 0.31*** 0.25*** 0.25***
[0.04] [0.04] [0.04] [0.03] [0.03] [0.04] [0.04] [0.04]
Controls for
Age
Literacy
Family Size
Farmer
Sample Includes
All households
Owners only
Farmers only
Observations 24,686 24,662 24,662 24,738 24,712 15,684 15,669 15,669
Pseudo-$R2$ 0.08 0.16 0.17 0.01 0.05 0.09 0.17 0.17
Panel A: Home Owner $=$ 1Panel B: Farmer $=$ 1Panel C: Home Owner $=$ 1
12345678
Black −0.33*** −0.26*** −0.25*** −0.11*** −0.10*** −0.36*** −0.31*** −0.31***
[0.01] [0.01] [0.01] [0.01] [0.01] [0.01] [0.01] [0.01]
Cherokee Nation 0.38*** 0.44*** 0.44*** 0.09*** 0.10*** 0.35*** 0.40*** 0.40***
[0.02] [0.02] [0.02] [0.02] [0.02] [0.02] [0.02] [0.02]
Black $×$ Cherokee Nation 0.24*** 0.14*** 0.13** 0.13*** 0.11*** 0.31*** 0.25*** 0.25***
[0.04] [0.04] [0.04] [0.03] [0.03] [0.04] [0.04] [0.04]
Controls for
Age
Literacy
Family Size
Farmer
Sample Includes
All households
Owners only
Farmers only
Observations 24,686 24,662 24,662 24,738 24,712 15,684 15,669 15,669
Pseudo-$R2$ 0.08 0.16 0.17 0.01 0.05 0.09 0.17 0.17

Marginal effects are reported. Robust standard errors are reported in brackets. Sampling weights are used. Sample includes male household heads in the Cherokee Nation and South. ***$p<0.001$, **$p<0.01$, and *$p<0.05$.

### B. Occupation in 1900

Within the rural South, upward occupational mobility frequently involved progressing up the “tenure ladder” from wage worker to sharecropper to tenant. The highest “rung” was the independent farmer. Because income typically increased as one progressed up the ladder, racial income inequality would be associated with fewer black farmers relative to white farmers (Alston & Ferrie, 2005). Panel B of table 6 reports estimates when a farmer dummy is the dependent variable. The racial gap in farming occupation is statistically significantly smaller in the Cherokee Nation, and blacks in the Cherokee Nation are absolutely more likely to be farmers than blacks in the South. The census definition of “farmer” was broad and included landowners, sharecroppers, and tenants. There was, however, a large practical distinction between farmers who worked others' land and farmers who worked their own land. While a direct measure of farm tenancy does not appear in the population schedules of the 1900 U.S. Census, home ownership is highly correlated with landownership. In panel C, I restrict the sample to farmers to examine ownership. The difference in the racial ownership gap is large and statistically significant, ranging from 25 to 31 percentage points, suggesting that racial inequality in farm ownership continues to be smaller in the Cherokee Nation.29

## VII. Potential Mechanisms

### A. Investment in Fruit Trees

Due to the risk of expropriation, landowners have a greater incentive to invest in productivity-enhancing, location-specific improvement than tenants do (Alchian & Demsetz, 1973; Demsetz, 1967). Such investments are one mechanism by which landowners can increase income, both overall and on a per acre basis. Land use patterns suggest that Cherokee freedmen were more likely than Southern freedmen to undertake a certain type of investment: the planting of orchards. Orchards are a long-term investment; they require a large initial outlay, and the land produces little to no income during the initial gestation period (Rhode, 1995). Once the trees start producing fruit, they have a valuable productive life. At median 1880 yields and 1880 prices, a farmer could earn $58 per acre planted in apple trees and$7 per acre planted in corn. An assurance of stable property rights is critical when making such a long-term investment.

The Cherokee freedmen were much more likely to invest in orchards than were Southern freedmen: 60 percent of Cherokee freedmen farmers had planted fruit trees. In the South, slightly over 5% of black owners and 2% of black farmers had. Table 7 reports the results of regressing equation (2) where the dependent variable takes a value of 1 if any acreage is planted in fruit trees. The estimated coefficient for the black indicator variable is negative and significant in all regressions. The interaction term is positive in all regressions, and the racial gap in fruit tree acreage was 32 to 54 percentage points smaller in the Cherokee Nation than in the South. These results are robust to the inclusion of soil fixed effects, suggesting that differences in growing conditions are not driving the results.

Table 7.
Investment in Orchards, 1880
All FarmersOwners
1234
Black −0.17*** −0.10*** −0.14*** −0.11***
[0.01] [0.01] [0.02] [0.03]
Cherokee Nation 0.38*** 0.37*** 0.21*** 0.22***
[0.02] [0.04] [0.05] [0.05]
Black $×$ Cherokee Nation 0.44*** 0.28*** 0.34*** 0.28***
[0.05] [0.06] [0.07] [0.07]
Control for farm size
Observations 7,961 7,892 5,202 5,202
Pseudo-$R2$ 0.08 0.20 0.14 0.16
All FarmersOwners
1234
Black −0.17*** −0.10*** −0.14*** −0.11***
[0.01] [0.01] [0.02] [0.03]
Cherokee Nation 0.38*** 0.37*** 0.21*** 0.22***
[0.02] [0.04] [0.05] [0.05]
Black $×$ Cherokee Nation 0.44*** 0.28*** 0.34*** 0.28***
[0.05] [0.06] [0.07] [0.07]
Control for farm size
Observations 7,961 7,892 5,202 5,202
Pseudo-$R2$ 0.08 0.20 0.14 0.16

Dependent variable is 1 if any acres planted in orchards. Controls for soil, literacy, and age included. Marginal effects are reported. Robust standard errors are reported in brackets. Sampling weights are used. Sample includes male heads of household in the Cherokee Nation and 1KF samples that report positive acres in use in the sample counties. ***$p<0.001$, **$p<0.01$, and *$p<0.05$.

The increased propensity to plant orchards in the Cherokee Nation overall suggests a potential explanation for the smaller average acreage of nonblacks in the Cherokee Nation. Fruit cultivation tends to take place as small-scale, intensive farming, while crops such as corn or wheat tend to be grown on a larger, more extensive scale (Rhode, 1995). The Cherokee Nation's smaller average farm size may be a reflection of its greater number of fruit trees.

### B. Education and Literacy

Freed slaves had lower average levels of human capital than whites, as measured by literacy rates and schooling completed. In 1870, 81% of southern blacks were illiterate, while only 8.5% of whites were (Collins & Margo, 2003). Schools established by both the Freedmen's Bureau and concerned African Americans helped to narrow both the black-white literacy gap and the gap between free blacks and previously enslaved blacks in the Southern states (Sacerdote, 2005). However, the impact of these schools was limited due to access and quality issues (Margo, 1990). High levels of land inequality may have inhibited public investment in education within the South (Galor, Moav, & Vollrath, 2009). Blacks in the Cherokee Nation faced considerable hurdles in obtaining an education. Cherokee schools were racially segregated. Because the Freedmen's Bureau had no jurisdiction in the Cherokee Nation, it opened no schools. The Cherokee government provided only teachers for schools, with the local black community responsible for all other expenses. A minimum enrollment policy restricted public school access in areas with smaller populations. Cherokee freedmen testifying before a Senate committee in 1885 spoke of difficulties in attending public schools. Some local communities opened tuition schools. Other families sent their children to Kansas.30

In panel A of table 8, I estimate equation (2) for school attendance for 1900 and restrict the sample to children ages 10 to 18. I include demographic controls (age, sex, and family size) in column 1. Economic controls are added in column 2 (literacy of the household head, homeownership, and the household head's occupation as farmer). The coefficient on the interaction terms is positive and statistically significant. While part of this difference is due to overall lower school attendance in the Cherokee Nation, Cherokee freedmen children have absolutely higher attendance than Southern black children. Once economic factors are included, the size of the gap falls from 16 to 8 percentage points and becomes only marginally statistically significant, suggesting that the Cherokee freedmen children's higher levels of human capital accumulation may be due to their parents' higher levels of human and physical capital.

Table 8.
1900 Human Capital Attainment
A. School AttendanceB. Months of SchoolC. Literacy
12345678
Black −0.25*** −0.14*** −0.80*** −0.39*** −0.35*** −0.19*** −0.43*** −0.48***
[0.01] [0.01] [0.04] [0.04] [0.01] [0.01] [0.00] [0.00]
Cherokee Nation −0.10*** −0.13*** 0.76*** 0.73*** −0.17*** −0.17*** −0.18*** −0.20***
[0.02] [0.02] [0.10] [0.10] [0.03] [0.02] [0.01] [0.01]
Black $×$ Cherokee Nation 0.16*** 0.08* 0.71*** 0.48** 0.19*** 0.12*** 0.26*** 0.23***
[0.03] [0.03] [0.18] [0.17] [0.02] [0.02] [0.01] [0.01]
Controls
Demographic
Economic
Observations 31,471 31,471 15,519 15,519 31,471 31,471 73,811 69,109
Pseudo-$R2$ 0.12 0.14 0.03 0.15 0.13 0.201 0.14 0.21
A. School AttendanceB. Months of SchoolC. Literacy
12345678
Black −0.25*** −0.14*** −0.80*** −0.39*** −0.35*** −0.19*** −0.43*** −0.48***
[0.01] [0.01] [0.04] [0.04] [0.01] [0.01] [0.00] [0.00]
Cherokee Nation −0.10*** −0.13*** 0.76*** 0.73*** −0.17*** −0.17*** −0.18*** −0.20***
[0.02] [0.02] [0.10] [0.10] [0.03] [0.02] [0.01] [0.01]
Black $×$ Cherokee Nation 0.16*** 0.08* 0.71*** 0.48** 0.19*** 0.12*** 0.26*** 0.23***
[0.03] [0.03] [0.18] [0.17] [0.02] [0.02] [0.01] [0.01]
Controls
Demographic
Economic
Observations 31,471 31,471 15,519 15,519 31,471 31,471 73,811 69,109
Pseudo-$R2$ 0.12 0.14 0.03 0.15 0.13 0.201 0.14 0.21

Marginal effects reported for panels A and C. Dependent variable is 1 if a child attended school or is literate. Demographic controls included age, sex, and family size. Family size excluded for adults. Economic controls are literacy of the household head, home ownership, and the household head's occupation as farmer. Robust standard errors are reported in brackets. Sampling weights are used. Sample includes Cherokee Nation and South in 1900. ***$p<0.001$, **$p<0.01$, and *$p<0.05$.

I consider months of school attended by those children who attended any school in panel B and literacy in panel C. Black children overall attended school for fewer months and had lower levels of literacy. However, those in the Cherokee Nation outperformed their Southern counterparts. Racial educational inequality was smaller in the Cherokee Nation. When economic controls were included, the difference between the Cherokee Nation and the South shrank to .48. This again suggests that the Cherokee freedmen's increased schooling was due in part to higher parental income and wealth levels. In the final two columns, I consider the literacy of adults. In 1880, the Cherokee freedmen had no clear human capital advantage over Southern freedmen. Literacy rates were lower for each birth cohort, and the black adult literacy rate was 21 percentage points lower than the Southern rate. (See tables 1 and 3.) By 1900, this pattern reversed. In both the basic specification and after demographic controls are included, the adult racial literacy gap was 23 to 26 percentage points smaller in the Cherokee Nation. This was partly due to a lower nonblack adult literacy rate in the Nation.

Using only the linked Cherokee Nation data, I directly test the relationship between parental income levels and investment in children by estimating
$Pr(Literatei)=Φ(βo+β1ParentLiti+β2Q1i+β3Q2i+β3Q3i+γXi+ɛi)$
(3)

where the dependent variable is an indicator equaling 1 if a person was literate in 1900. ParentLit is 1 if either parent was literate. Q1 through Q3 are indicator variables for 1880 household income quartiles (as calculated from the sum of farm products). The omitted quartile, Q4, corresponds to the highest income level. The vector $Xi$ contains demographic controls. Marginal effects are reported in table 9. I first restrict analysis to adults in 1900 who were children (i.e., less that 18 years of age) in 1880. All quartile coefficients are negative in sign and increasing in magnitude as income falls, although the estimated coefficient is significant only for the lowest quartile. The estimated coefficients on the lowest two quartiles are significant when demographic controls are included. The children of poorer families were less likely to become literate adults, even after controlling for parents' literacy. The Cherokee freedmen's overall higher adult literacy rates in 1900 may be partly explained by higher parental income in 1880.

Table 9.
Income and Intergenerational Human Capital Attainment
Literate in 1900
123456
Parent literate 0.13*** 0.13*** 0.11** 0.12*** 0.15*** 0.13***
[0.04] [0.05] [0.05] [0.05] [0.05] [0.05]
Q1  −0.22* −0.28**  −0.12** −0.09*
[0.13] [0.14]  [0.05] [0.05]
Q2  −0.09 −0.12*  −0.1 −0.06
[0.07] [0.07]  [0.07] [0.07]
Q3  −0.05 −0.05  −0.03
[0.05] [0.05]  [0.06] [0.06]
Controls
Demographics
Observations 406 406 399 496 496 496
Pseudo-$R2$ 0.02 0.03 0.07 0.01 0.02 0.08
Literate in 1900
123456
Parent literate 0.13*** 0.13*** 0.11** 0.12*** 0.15*** 0.13***
[0.04] [0.05] [0.05] [0.05] [0.05] [0.05]
Q1  −0.22* −0.28**  −0.12** −0.09*
[0.13] [0.14]  [0.05] [0.05]
Q2  −0.09 −0.12*  −0.1 −0.06
[0.07] [0.07]  [0.07] [0.07]
Q3  −0.05 −0.05  −0.03
[0.05] [0.05]  [0.06] [0.06]
Controls
Demographics
Observations 406 406 399 496 496 496
Pseudo-$R2$ 0.02 0.03 0.07 0.01 0.02 0.08

Dependent variable is 1 if a person was literate. Marginal effects are reported. Robust standard errors are reported in brackets. Sample includes blacks in the Cherokee Nation in 1900. ***$p<0.01$, **$p<0.05$, and *$p<0.1$.

I estimated equation (3) for children ages 10 through 18 in 1900. ParentLit is 1 if either parent was literate in 1900. Q1 through Q4 are now indicator variables for the head of household's 1880 income quartile. The results are similar to those for adults, although estimated coefficients are generally smaller in absolute value than in the adult regression. Children in 1900 were less likely to be literate if their grandparents were located lower in the income distribution. After controlling for demographic characteristics, a child was 9% less likely to be literate in 1900 if his or her grandparents' household was in the lowest income quartile, even after controlling for parental literacy.

## VIII. Conclusion

High levels of racial inequality have been a hallmark of the American economy since emancipation. Scholars have long debated the extent to which this inequality could have been reduced or eliminated through policies of land reform or redistribution. In this paper, I developed an empirical strategy to exploit a plausibly exogenous policy variation in order to identify the impact of free land on the economic outcomes of former slaves. Freed slaves in the Cherokee Nation could claim land, while Southern freedmen's hopes for “forty acres and a mule” were never realized. My results suggest that access to free land increased the incomes and wealth of black families during the decades following emancipation and dramatically lowered levels of racial inequality.

I also considered how access to land and the subsequent increase in incomes affected the choices available to former slaves in the Cherokee Nation. I found that blacks in the Cherokee Nation had increased investment in both physical and human capital, they were more likely to plant fruit trees, and they had higher literacy and school attendance rates in 1900. Taking advantage of the unique linked nature on my data set, I find that the gains in human capital achievement were related to 1880 income and wealth levels. Both the children and grandchildren of the poorest families in 1880 were less likely to be literate in 1900.

Taken together, these findings suggest that land reform and distribution may have an important role in alleviating inequality, particularly in primarily agricultural economies like that of the postbellum South. The failure of the federal government to implement a land distribution policy was lamented by many freed slaves, who believed that landownership was key for their future success. When General Sherman and Secretary of War Edward Stanton met with several black religious leaders in 1865, the formerly enslaved asserted, “The way we can best take care of ourselves is to have land, and turn it and till it by our own labor” (Berlin et al., 1993, 175). The economic success of the Cherokee freedmen suggests that they were correct.

## Notes

1

The phrase “forty acres and a mule” has appeared in sources as diverse as the writings of W. E. B. Du Bois; the film Gone with the Wind; and the rapper Tupac Shakur's 1996 song “Letter to the President,” which inquired, “What happened to our 40 acres and a mule, fool?” (Shakur, 1999). John Conyers Jr. (D–Michigan) first introduced HR-40, “The Commission to Study Reparation Proposals for African-Americans Act,” in 1989. The bill's number, 40, was chosen, “as a symbol of the forty acres and a mule that the United States initially promised freed slaves.”

2

John Marshall declared the Cherokee Nation a “denominated domestic dependent nation” in Cherokee Nation v. Georgia, 30 U.S. 1 (1831). The practical implication was that the Cherokee Nation's government could enact and enforce its own laws and policies. However, these laws could be overridden if the U.S. Congress explicitly passed contradicting legislation.

3

“Declaration by the people of the Cherokee Nation of the causes which have impelled them to unite their fortunes with those of the Confederate States of America,” approved October 28, 1861, reprinted in The War of the Rebellion: A Compilation of the Official Records of the Union and Confederate Armies, series I, vol. 13, 503–505 (1902).

4

Statement of the Commissioner to the Northern Cherokees as to what Points the Government Would Insist upon, reprinted in Thoburn (1924).

5

Letter from John B. Sanborn to James Harlan, January 5, 1866, quoted in Bogy (1866), 285.

6

Ibid.

7

As in the South, former Confederates were given amnesty, had their confiscated property returned, and received no compensation for their emancipated slaves.

8

Beyond land, the exact rights conferred on the freedmen became a point of contention. As part of the treaty, the Cherokee Nation sold land to the United States. The payment was to be distributed among Cherokee citizens in 1883, but the Cherokee National Council decreed freedmen were ineligible for payments. A settlement reached in 1896 provided eligible freedmen with up to \$295.35 (Smith & Wolfe, 1896). No payments were received before the 1880 Census. Resistance to Cherokee freedmen citizenship continued, and their citizenship was cancelled in 1983. In August 2017, a federal judge ruled that the freedmen were entitled to citizenship.

9

Once land was claimed, ownership rights were similar to those of typical fee-simple ownership. As long as the land was not abandoned, its owner held heritable usufructuary rights. The land and its improvements could be sold, used as collateral, bequeathed in wills, or improved on. However, only Cherokee citizens were able to hold these rights. See Bloom (2002).

10

Afro-American Advocate, February 19, 1892, quoted in Littlefield (1978, 69).

11

Higgs (2008) uses Georgia property tax returns. Margo (1984) has wealth information for Arkansas, Louisiana, North Carolina, and Virginia. He finds a similar temporal trend to Higgs: blacks accumulated property at a faster rate than whites, but the black-white wealth gap remained large. However, Margo's evidence suggests that black wealth grew more slowly.

12

The 1900 agricultural schedules were destroyed for most of the United States, and farm-level agricultural data are not available for 1900.

13

See “Appendix G: Data Appendix” of 1KF for additional information. Not all Southern counties were included in the sample, and I use the provided sampling weights during estimation. The sample contains farms from all Confederate states, excluding Arkansas. The Arkansas agricultural schedules were of poor quality. See Ransom and Sutch (2001, 413).

14

“Black” refers to people of African descent who could have been enslaved. I generally use black, former slave, and freedman interchangeably. “Nonblack” refers to people of all other races. In the South, nonblacks were primarily white. In the Cherokee Nation, nonblacks were primarily “adopted whites” (white people who had gained Cherokee citizenship through marriage) or “native Cherokees” (people not of color who were born into Cherokee citizenship). The child of an “adopted white” and a “native Cherokee” was a “native Cherokee.”

15

United States (1886, 3).

16

Ibid., 76.

17

All quotes in this paragraph were taken from Carter (1999, 106).

18

As “Indians not taxed,” Cherokees and their slaves were excluded from census returns. An Arkansas census enumerator mistakenly collected slave data for Cherokee Nation (Doran, 1978).

19

The WPA Slave Narratives of Lucinda Vann and Henry Henderson, former Cherokee slaves, provide details of the organization and treatment of slave labor on large Cherokee plantations (Baker & Baker, 1996).

20

United States (1860, 19).

21

The Cherokee Nation had a written constitution, a bicameral legislature, a chief executive with the title of principal chief, a cabinet, and a supreme court. The government also ran a system of public schools. See Abbott (1987).

22

Author calculations from the 1835 Cherokee Census. I thank David M. Wishart for digitizing and sharing these census data.

23

Although a time trend would be ideal, I have located comparable data only for 1859.

24

By limiting the analysis to farmers, I can calculate monetary values for income and wealth. The trade-off for these estimates is the exclusion of nonfarm households. The focus on farms is less restrictive in 1880 than it would be today. According to the 1880 IPUMS (Ruggles et al., 2019), 91.3% of all Southerners lived in rural areas, and 92.7% of black Southerners lived in rural areas, with 70.02% of black male heads of households either farmers or agricultural laborers. An additional 26.6% of black male household heads were classified solely as laborers and likely involved in agriculture. Rural Southern whites were similarly concentrated within agriculture, with 76.6% being employed as farmers or in other agricultural occupations. Southern counties that grow primarily rice, sugar, or tobacco are excluded from the analysis. These cash crops use different farming technologies and have larger minimum efficient scales than the more common Southern crops, such as corn and cotton.

25

Soil types are taken from the soil map included in Tenth Census of the United States, vol. 5: Report on Cotton Production in the United States (Hilgard, 1884).

26

Acres in use are reported in both the Cherokee and the U.S. Census data. Other reported land size measures differ. Acres in use include all tilled fields, acres not planted due to crop rotation, acres in pasturage, acres in orchards, and acres with structures.

27

All crop and livestock price data are from U.S. Department of Agriculture Statistical Bulletin No. 16, Prices of Farm Products Received by Producers (Bureau of Agricultural Economics, 1927), with the exception of fruit prices. The USDA bulletin did not report fruit prices until the 1900s. I obtained fruit prices from an article in the Cherokee Advocate, March 16, 1881. Although the Cherokee Census included the acreage devoted to orchards, it did not record yields. Neglecting to include fruit yields would substantially underestimate Cherokee farm incomes. The 1KF sample does contain information that can be used to calculate median fruit yields per acre for Southern farms. I use these to impute the value of fruit production for each Cherokee farm and add it to the value of other crops for Cherokee farms. I use actual yields to calculate Southern crop income. As a robustness check, I repeat the income regressions using both Southern mean fruit yield and excluding fruit entirely. These estimates are reported in appendix II, and the difference in the income gaps is always statistically significantly smaller in the Cherokee Nation.

28

In anticipation of Oklahoma statehood, whites flocked to Indian Territory. I exclude noncitizen residents of the Cherokee Nation. Because many of these whites rented property in anticipation of later claiming land, the estimate of $β3$ increases if they are included.

29

The Cherokee Nation bordered the Confederate state of Arkansas. While the adjacent area similar had climate and geography, former slaves lacked comparable access to free land. In appendix II, I use 1880 and 1900 Census data to examine the racial gap in farm occupation and homeownership. Racial inequality is lower in the Cherokee Nation than in Arkansas.

30

United States (1986, pp. 5–6, 9–10, and 21–22).

## REFERENCES

REFERENCES
Abbot
,
Devon
, “
‘Commendable Progress’: Acculturation at the Cherokee Female Seminary,
American Indian Quarterly
11
(
1987
),
187
201
.
Agnew
,
, “
Fort Gibson (fort)
,” in
The Encyclopedia of Oklahoma History and Culture
(
2009
), http://www.okhistory.org/publications/enc/entry.php?entryFO033.
Alchian
,
Armen A.
, and
Harold
Demsetz
, “
,”
Journal of Economic History
33
:
1
(
1973
),
16
27
.
Aldrich
,
Duncan M.
, “
No General Stores, Retail Merchants, and Assimilation: Retail Trade in the Cherokee Nation, 1838–1890,
Chronicles of Oklahoma
57
(
1979
),
119
136
.
Alston
,
Lee J.
, “Issues in Postbellum Southern Agriculture” (pp.
207
228
), in
Louis
Ferleger
, ed.,
Agriculture and National Development: Views on the Nineteenth Century
(
Ames
:
Iowa State University Press
,
1990
).
Alston
,
Lee J.
, and
Joseph P.
Ferrie
, “
Time on the Ladder: Career Mobility in Agriculture, 1890–1938,
Journal of Economic History
65
(
2005
),
1058
1081
.
Baker
,
T. Lindsey
, and
Julie P.
Baker
, eds.
WPA Oklahoma Slave Narratives
(
Norman
:
University of Oklahoma Press
1996
).
Banerjee
,
A.
, and
Lakshmi
Iyer
, “
History, Institutions, and Economic Performance?: The Legacy of Colonial Land Tenure Systems in India,
American Economic Review
95
(
2002
),
1190
1213
.
Becker
,
Gary S.
, and
Nigel
Tomes
, “
An Equilibrium Theory of the Distribution of Income and Intergenerational Mobility
,”
Journal of Political Economy
87
(
1979
),
115
1189
.
Becker
,
Gary S.
, and
Nigel
Tomes
Human Capital and the Rise and Fall of Families,
Journal of Labor Economics
4
(
1986
),
S1
S39
.
Belz
,
Herman
, “
The New Orthodoxy in Reconstruction Historiography
,”
Reviews in American Histor
1:1
(
1973
),
1
47
.
Berlin
,
Ira
,
Joseph P.
Reidy
,
Steven F.
Miller
, and
Leslie S.
Rowland
, eds.,
Freedom: A Documentary History of Emancipation, 1861–1867
, series 1, vol. 2:
The Wartime Genesis of Free Labor: The Upper South
(
Cambridge
:
Cambridge University Press
,
1993
).
Besley
,
Timothy
, “
Property Rights and Investment Incentives: Theory and Evidence from Ghana,
Journal of Political Economy
103
(
1995
),
903
937
.
Besley
,
Timothy
, and
Maitreesh
Ghatak
, “
Property Rights and Economic Development
” (5:4525–4595), in
Dani
Rodrik
and
Mark
Rosenzweig
, eds.,
Handbook of Development Economics
(
Amsterdam
:
Elsevier
,
2010
), doi:10.1016/B978-0-444-52944-2.00006-9.
Billington
,
Monroe
, “
Black Slavery in Indian Territory: The Ex-Slave Narratives
,”
Chronicles of Oklahoma
60
(
1982
):
56
65
.
Bloom
,
Khaled J.
, “
An American Tragedy of the Commons: Land and Labor in the Cherokee Nation, 1870–1900,
Agricultural History
76
(
2002
),
497
523
.
Bogy
,
Hon. H. L.
,
Report of the Commissioner of Indian Affairs for the Year 1866
(
Washington, DC
:
Government Printing Office
,
1866
).
Bruce
,
John W.
, and
Shem E.
,
Searching for Land Tenure Security in Africa
(
Dubuque, IA
:
Kendall Hunt
,
1994
).
Bureau of Agricultural Economics
,
US Dept of Agriculture Statistical Bulletin 16
(
Washington
:
Government Printing Office
,
1927
).
Carlton
,
David L.
, and
Peter A.
Coclanis
, “
Capital Mobilization and Southern Industry, 1880–1905: The Case of the Carolina Piedmont
Journal of Economic History
49:1
(
1989
)
73
94
.
Carter
,
Kent
,
The Dawes Commission: And the Allotment of the Five Civilized Tribes, 1893–1924
(
Orem, UT
:
Ancestry Publishing
,
1999
).
Collins
,
William J.
, and
Robert A.
Margo
, “
Race and Home Ownership: A Century-Long View
,”
Explorations in Economic History
38
:
1
(
2001
),
68
92
.
Collins
,
William J.
, and
Robert A.
Margo
Historical Perspectives on Racial Differences in Schooling in the United States
,” NBER working paper
9770
(
2003
).
DeCanio
,
Stephen
, “
Accumulation and Discrimination in the Postbellum South
,”
Explorations in Economic History
16
(
1979
),
182
206
.
Demsetz
,
Harold
, “
Toward a Theory of Property Rights,
American Economic Review
57
(
1967
),
347
359
.
DeNavas-Walt
,
Carmen
, and
Proctor
,
U.S. Census Bureau, Current Population Reports, P60-249, Income and Poverty in the United States: 2013
(
Washington, DC
:
U.S. Government Printing Office
,
2014
).
De Soto
,
Hernando
,
The Mystery of Capital: Why Capitalism Triumphs in the West and Fails Everywhere Else
(
New York
:
Basic Books
,
2000
).
De Soto
,
Hernando
The Mystery of Capital. Finance and Development
(
Washington, DC
:
International Monetary Fund
,
2001
), http://www.imf.org/external/pubs/ft/fandd/2001/03/desoto.htm.
Doran
,
Michael F.
, “
Negro Slaves of the Five Civilized Tribes,
Annals of the Association of American Geographers
68
(
1978
),
335
350
.
Dorner
,
Peter
,
Land Reform in Latin America: Issues and Cases
(
:
Land Economics for the Land Tenure Center, University of Wisconsin–Madison
,
1971
).
Dorner
,
Peter
Land Reform and Economic Development
(
New York
:
Penguin Books
,
1972
).
Du
Bois
, and
William Edward
Burghardt
,
The Souls of Black Folk: Essays and Sketches
(
Chicago
:
A. C. McClurg
,
1903
).
Engerman
,
Stanley L.
, “
Review of The Roots of Black Poverty: The Southern Plantation Economy after the Civil War by Jay R. Mandle, The Wheel of Servitude: Black Forced Labor after Slavery by Daniel A. Novak, Forty Acres and a Mule: The Freedmen's Bureau and Black Land Ownership by Claude F. Oubre, and Social Origins of the New South: Alabama, 1860–1885 by Jonathan M. Wiener,
Journal of Social History
13
(
1980
),
490
497
.
Engerman
,
Stanley L.
Economic Adjustments to Emancipation in the United States and British West Indies
Journal of Interdisciplinary History
13
(
1982
),
191
220
.
Eswaran
,
Mukesh
, and
Ashok
Kotwal
, “
Economic Journal
96
(
1986
),
482
498
.
Feder
,
Gershon
, and
World Bank
,
Land Policies and Farm Productivity in Thailand
(
Baltimore
:
Johns Hopkins University Press
,
1988
).
Fishback
,
Price V.
, and
John H.
, “
Narrowing the Black-White Gap in Child Literacy in 1910: The Roles of School Inputs and Family Inputs
,” this review
73
(
1991
),
725
728
.
Foner
,
Eric
,
Politics and Ideology in the Age of the Civil War
(
New York
:
Oxford University Press
,
1980
).
Foner
,
Eric
Nothing But Freedom: Emancipation and Its Legacy
(
Baton Rouge
:
Louisiana State University Press
,
2007
).
Frank
,
Andrew K.
, “
Five Civilized Tribes
,” in
The Encyclopedia of Oklahoma History and Culture
(
Oklahoma City
:
Oklahoma Historical Society
,
2009
).
Galiani
,
Sebastian
, and
Ernesto
Schargrodsky
, “
Property Rights for the Poor: Effects of Land Titling,
Journal of Public Economics
94
(
2010
),
700
729
.
Galor
,
Oded
,
Omer
Moav
, and
Dietrich
Vollrath
, “
Inequality in Landownership, the Emergence of Human-Capital Promoting Institutions, and the Great Divergence
,”
Review of Economic Studies
76
:
2
(
2009
),
143
179
.
Goldstein
,
Markus
, and
Udry
,
Christopher
, “
The Profits of Power: Land Rights and Agricultural Investment in Ghana
,”
Journal of Political Economy
116
(
2008
),
981
1022
.
Grawe
,
Nathan, D.
, and
Casey B.
Mulligan
, “
Economic Interpretations of Intergenerational Correlations,
Journal of Economic Perspectives
16
(
2002
),
45
58
.
Hahn
,
Steven
,
Land and Labor, 1865
, vol.
1
(
Chapel Hill
:
University of North Carolina Press
,
2008
).
Higgs
,
Robert
, “
Accumulation of Property by Southern Blacks before World War I
,”
The American Economic Review
72
:
4
(
1982
),
725
737
.
Higgs
,
Robert
Competition and Coercion: Blacks in the American Economy, 1865–1914
(
Cambridge
:
Cambridge University Press
,
2008
).
Hightower
,
Michael J.
,
Banking in Oklahoma before Statehood
(
Norman
:
University of Oklahoma Press
,
2013
).
Hilgard
,
Eugene W.
,
Tenth Census of the United States, Vol. 5: Report on Cotton Production in the United States
(
Washington
:
Government Printing Office
,
1884
).
Holland
,
Cullen Joe
, “
The Cherokee Indian Newspapers, 1828–1906: The Tribal Voice of a People in Transition
,” PhD diss.,
University of Minnesota
,
1956
.
Howard
,
Oliver Otis
,
Autobiography of Oliver Otis Howard, Major General United States Army
(
Charlotte, NC
:
Baker & Taylor
,
1908
).
Jacoby
,
H. G.
,
Guo
Li
, and
Scott
Rozelle
, “
Hazards of Expropriation: Tenure Insecurity and Investment in Rural China,
American Economic Review
92
(
2002
),
1420
1447
.
James
,
John A.
, “
Financial Underdevelopment in the Postbellum South
Journal of Interdisciplinary History
11
(
1981
),
443
454
.
Jaynes
,
Gerald David
,
Branches without Roots: Genesis of the Black Working Class in the American South, 1862–1882
(
New York
:
Oxford University Press
,
1986
).
Kappler
,
Charles J.
, “
Indian Affairs: Laws and Treaties
,” in
Treaties
, vol.
2
(
Washington, DC
:
Government Printing Office
,
1904
).
Kocchar
,
Rakesh
, and
Richard
Fry
, “
Wealth Inequality Has Widened along Racial, Ethnic Lines since End of Great Recession
” (
Washington, DC
:
Pew Research Center
,
2015
).
Lin
,
J. Y.
, “
Rural Reforms and Agricultural Growth in China,
American Economic Review
82
(
1992
),
34
51
.
Littlefield
,
Daniel F.
,
The Cherokee Freedmen
(
Westport, CT
:
Greenwood Press
,
1978
).
Littlefield
,
Daniel F., Jr.
, and
Mary Ann
Littlefield
, “
The Beams Family: Free Blacks in Indian Territory
,”
Journal of Negro History
61
(
1976
),
16
35
.
Mandle
,
Jay R.
, “
Review of Essays on Caribbean Integration and Development by William G. Demas,
Journal of Economic History
38
(
1978
),
531
532
.
Margo
,
Robert A.
, “
Accumulation of Property by Southern Blacks before World War I: Comment and Further Evidence
,”
American Economic Review
74
(
1984
),
768
776
.
Margo
,
Robert A.
Accounting for Racial Differences in School Attendance in the American South, 1900: The Role of Separate-But-Equal
,” this review
69
(
1987
),
661
666
.
Margo
,
Robert A.
Race and Schooling in the South, 188–1950: An Economic History
(
Chicago
:
University of Chicago Press
,
1990
).
McKenzie
,
Robert Tracy
,
One South or Many? Plantation Belt and Upcountry in Civil War–Era Tennessee
(
Cambridge
:
Cambridge University Press
,
2002
).
Menard
,
Russell Trent Alexander
,
Jason
Digman
, and
J.
David Hacker
,
Public Use Microdata Samples of the Slave Population of 1850–1860
(
Minneapolis
:
University of Minnesota, Minnesota Population Center
,
2004
)
Miles
,
Tiya
,
Ties That Bind: The Story of An Afro-Cherokee Family in Slavery and Freedom
(
Berkeley
:
University of California Press
,
2006
).
Miller
,
Melinda
, “
Dawes Cards and Indian Census Data. Historical,
Methods: A Journal of Quantitative and Interdisciplinary History
48
(
2015
),
214
229
.
Moehling
,
Carolyn M.
, “
Family Structure, School Attendance, and Child Labor in the American South in 1900 and 1910
,”
Explorations in Economic History
41
:
1
(
2004
),
73
100
.
Naylor
,
Celia E.
,
African Cherokees in Indian Territory
(
Chapel Hill
:
University of North Carolina Press
,
2008
).
Oubre
,
Claude F.
,
Forty Acres and a Mule: The Freedmen's Bureau and Black Land Ownership
(
Baton Rouge
:
Louisiana State University Press
,
2012
).
Perdue
,
Theda
,
Slavery and the Evolution of Cherokee Society
(
Knoxville
:
University of Tennessee Press
,
1979
).
Ransom
,
Roger L.
, “
Reconstructing Reconstruction: Options and Limitations to Federal Policies on Land Distribution in 1866–67
,”
Civil War History
51
(
2005
),
364
377
.
Ransom
,
Roger L.
, and
Richard
Sutch
, “
The Ex-Slave in the Post-Bellum South: A Study of the Economic Impact of Racism in a Market Environment
,”
Journal of Economic History
33
:
1
(
1973
),
131
148
.
Ransom
,
Roger L.
, and
Richard
Sutch
,
One Kind of Freedom: The Economic Consequences of Emancipation
(
Cambridge University Press
,
2001
).
Rhode
,
Paul W.
, “
Learning, Capital Accumulation, and the Transformation of California Agriculture,
Journal of Economic History
55
(
1995
),
773
800
.
Ruggles
,
Steve
,
Sarah
Flood
,
Ronald
Goeken
,
Josiah
Grover
,
Erin
Meyer
,
Jose
Pacas
, and
Matthew
Sobek
,
IPUMS USA
, version 9.0 [dataset] (
Minneapolis, MN
:
IPUMS
,
2019
), https://doi.org/10.18128/D010.V9.0
Sacerdote
,
Bruce
, “
Slavery and the Intergenerational Transmission of Human Capital
,” this review
87
(
2005
),
217
234
.
Shakur
,
Tupac
, “
Letter to the President
,”
Still I Rise
, Interscope Records (
1999
).
Smith
,
Samuel
, and
Richard M.
Wolfe
, “
Acts of the Cherokee Council to Equalize Payments from Cherokee Outlet Funds
,” in
Annual Report of the Commissioner of Indian Affairs
(
Washington, DC
,
1896
).
Stevens
,
, “
Speech of the Hon. Thaddeus Stevens of Pennsylvania, Delivered to the House of Representatives, March 19, 1867, on the Bill (H.R. No. 20) Relative to Damages to Loyal Men, and for Other Purposes
,” in
Beverly Wilson
Palmer
and
Holly Byers
Ochoa
, eds.,
The Selected Papers of Thaddeus Stevens
, vol.
2
: April 1865–August 1868 (
Pittsburgh
:
University of Pittsburgh Press
,
1998
).
Thoburn
,
Joseph B.
, “
The Cherokee Question,
Chronicles of Oklahoma
2
(
1924
),
141
242
.
Thornton
,
Russell
,
The Cherokees: A Population History
(
Lincoln
:
,
1992
).
United States
,
Report of the Commissioner of Indian Affairs, Accompanying the Annual Report of the Secretary of the Interior for the Year 1859
(
Washington, DC
,
1860
).
United States
Testimony Taken by the Committee on Indian Affairs of the United States Senate in Relation to the Conditions of the Indian Tribes in the Indian Territory, and upon Other Reservations under Resolutions of the Senate of June 11 and December 2, 1884 and February 23, 1885
(
Washington, DC
:
Government Printing Office
,
1886
).
Woodman
,
H. D.
, “
Sequel to Slavery: The New History Views the Postbellum South,
Journal of Southern History
43
(
1977
),
523
554
.
Woodman
,
H. D.
One Kind of Freedom after 20 Years
,”
Explorations in Economic History
38
:
1
(
2001
),
48
57
.
Wright
,
Gavin
,
Old South, New South: Revolutions in the Southern Economy since the Civil War
(
Baton Rouge
:
Louisiana State University Press
, [1986]
1997
).

## Author notes

I appreciate helpful suggestions from Brigitte Madrian (the editor), anonymous referees, and my dissertation committee (Warren Whatley, Ben Chabot, Martha Bailey, Maggie Levenstein, Tiya Miles, and the late Gary Saxonhouse). I thank Stanley Engerman, Price Fishback, Timothy Guinnane, Joanna Lahey, Naomi Lamoreaux, Sara LaLumia, Bob Margo, and Gavin Wright for their useful suggestions and conversations about this project. I benefited from the comments of seminar participants at the University of Michigan, NBER Development of the American Economy Summer Institute, Economic History Association Annual Meetings, and the Social Science History Association Meetings. Roger Ransom and Richard Sutch kindly shared their data and answered any questions I had. I thank the National Science Foundation (award SES-0619588), the Sasakawa Young Leaders Fellowship Fund, and the Economic History Association for their generous support.

A supplemental appendix is available online at http://www.mitpressjournals.org/doi/suppl/10.1162/rest_a_00842.