Homeownership gaps among Indian reservations prove puzzling
An analysis of some persistent gaps in homeownership rates among American Indian reservations in the Ninth Federal Reserve District.
Published May 1, 2009 | May 2009 issue
The homeownership rate of American Indians is low relative to that of whites in the U.S. According to the U.S. Census Bureau's 2006 American Community Survey, 56 percent of American Indian households were homeowners, compared to 71 percent of white households. In rural areas, which encompass American Indian reservations, homeownership rates tend to be higher, but the disparity persists. In 2006, the homeownership rate of 62 percent for rural American Indian households was 13 percentage points less than the rate for rural white households.
The American Indian homeownership gap with respect to whites is thought to reflect a number of economic, institutional, and demographic differences. For example, a recent report from the Federal Reserve and the Brookings Institution notes that homeownership on Montana's Blackfeet reservation is constrained by low incomes, weak credit histories, a lack of mortgage financing, and the challenges of real estate lending on tribal trust lands.1/ This report also notes that the Blackfeet reservation's housing stock is "generally substandard," raising questions about whether a proper accounting for housing quality would show even wider gaps between American Indian and white homeownership. A better understanding of American Indian homeownership and the factors that influence it could lead to more effective policies to close the American Indian homeownership gap. (For more on this thought, see "A note about homeownership" below.)
This article takes a small step in that direction by documenting the fact that American Indian homeownership rates vary significantly among American Indian reservations, including those in the Ninth Federal Reserve District (District). Furthermore, large regional differences remain, even when standard measures of the two key factors stressed in the Blackfeet study—household income and housing quality—are accounted for. Although it is beyond the scope of this article, further thinking about why American Indian homeownership is high on some reservations and low on others has the potential to improve our understanding of the factors and policies that determine American Indian homeownership rates.2/
Gaps between the District and the nation
The large District reservations analyzed here have a lower rate of American Indian homeownership (53 percent overall) than their counterparts in the rest of the country (71 percent). (See the sidebar below for details on the reservations and data analyzed.) Figure 1 (see below) shows that income differences explain some, but not all, of this 18 percentage-point gap. For both areas (District and rest of U.S.) and three income categories, the rectangular boxes in Figure 1 show the typical range of homeownership rates among reservations.3/ For example, in the low-income category, typical homeownership rates for American Indian households on District reservations range from 32 to 46 percent, whereas the rest of the country shows a typical range of 50 to 66 percent. The horizontal line inside each box represents the median reservation, or the one that falls in the middle when all the area's reservations are ranked according to their homeownership rates. In the low-income category, the median District reservation has a homeownership rate of 40 percent, compared to a median of 58 percent among large reservations in the rest of the country. Thus, among low-income American Indian reservation households, the gap between the District and the rest of the country is the same as the overall gap of 18 percentage points.
Click on chart to view larger image.
The percentage of households owning homes rises with income, as shown by the increasingly high position of the boxes from left to right for each geography. In addition, the gap between District and other reservations narrows as incomes rise, as indicated by the greater overlap of the typical range of homeownership rates. Based on the median reservations, the gap narrows from 18 percentage points for low-income households to 15 points for medium-income households and just 8 points for high-income households. However, the gap never disappears. Thus, we see that although household income is an important influence on American Indian homeownership, it fails to fully explain the lower rate of American Indian homeownership on District reservations.
The thin lines extending above and below most of the boxes cover reservations that are outside the typical range for their area and income category, but not exceptionally so.4/ A few reservations have even more extreme homeownership rates, and these outliers are individually represented by having their names displayed at the heights corresponding to their homeownership rates. Generally, for each income category, homeownership rates that are atypically high by District standards would be considered merely typical, or just barely higher than typical, by national standards. District reservations whose rates are atypically low within a given income category have lower homeownership rates than all but a few national outliers. Thus, the American Indian homeownership gap between large District reservations and other large reservations is still apparent when each area's less typical reservations are considered. Understanding why American Indian homeownership on large District reservations is lower than in the rest of the country, even after income differences are taken into account, is both a challenge and a potential source of insight.
Gaps within the District
Our understanding is further challenged by the fact that American Indian homeownership rates differ significantly among the large reservations within the District. For the reservations analyzed here, American Indian homeownership rates, relative to income, tend to be lowest in North Dakota and South Dakota. This is shown in Figure 2 (see below), which takes the same form as Figure 1, but covers three areas within the District—North Dakota-South Dakota (ND-SD), Montana (MT), and Minnesota-Wisconsin (MN-WI). Within each area, the tendency for homeownership to rise with income is again evident. Within each income group, however, the typical ranges of American Indian homeownership rates are similar for MN-WI and MT reservations but lower for ND-SD reservations. The gap between the Dakotas and the other areas is especially distinct for the low- and high-income brackets.
Click on chart to view larger image.
For each income group, reservation homeownership rates also vary widely within each of the three District areas. For example, in the low-income bracket, American Indian homeownership rates for large reservations in the Dakotas cluster around a median of 31 percent, but there are two outliers—Pine Ridge and Turtle Mountain—with substantially higher rates. Turtle Mountain's rate is well above average for low-income American Indian households in either MT or MN-WI (although it is still well below the 73 percent rate for Montana's Crow reservation). In the medium-income category, Turtle Mountain's American Indian homeownership rate of 78 percent exceeds all District reservations analyzed, and its 75 percent rate for high-income households also ranks high. The high rate of American Indian homeownership across income groups on Turtle Mountain raises questions about why this reservation differs from its peers in ND-SD.
Can housing quality explain the gaps?
As noted above, the quality of housing on American Indian reservations is sometimes low. This raises the possibility that focusing solely on reservation homeownership rates might be misleading. For example, suppose a reservation has a high homeownership rate but also has exceptionally low-quality or crowded houses. In that case, we might view its apparent homeownership success as less significant than if there were no offsetting quality gap. In this section, we assess the possibility and find only limited evidence of it.
Good data on all aspects of reservation housing quality are not readily available, but Census 2000 provides standard (yet imperfect) measures of some dimensions of quality. There are no area-wide housing quality differences in these data that are big enough to account for the regional differences in homeownership rates shown in Figures 1 and 2. However, when we look at individual reservations within the District, we find a few cases where the incidence of mobile home ownership, substandard kitchen and plumbing facilities, or older or crowded homes is unusually high on reservations with relatively high American Indian homeownership rates.
Although most mobile homes manufactured today meet quality standards on par with those of site-built homes, many older mobile homes are less durable. For this reason, census data on mobile homes—and especially mobile homes subjected to the harsh climates typical of District reservations—are considered a potential indicator of low-quality housing. Nonetheless, mobile home ownership explains only a few of the disparities in Figures 1 and 2. Both nationally and within the District, 23 percent of homes owned by American Indian households on reservations are manufactured units, making mobile homes a nonfactor with respect to how the District's American Indian homeownership rate compares to the national rate. Within the District, American Indian ownership of mobile homes is highest on ND-SD reservations, where overall American Indian homeownership rates are low, so it doesn't explain away the higher American Indian homeownership rates on MT or MN-WI reservations. However, it may partly explain why American Indian homeownership is higher on the Pine Ridge and Turtle Mountain reservations than on the reservations in the Dakotas, as about 40 percent of homeowners on these two reservations own mobile homes—the highest rate within the District. Turtle Mountain's American Indian homeownership rate is still fairly high, even considering the higher usage of mobile homes there.
The remaining measures of housing quality show only a few notable differences across reservations. The age of a home is considered a quality indicator because structures deteriorate over time. Among American Indian homeowners on reservations, the distribution of homes by decade built is similar in the District and the nation. The same is generally true when comparing reservations within the District, except that three Montana reservations (Rocky Boy's, Fort Belknap, and Northern Cheyenne) have an unusually high percentage of homes built in the 1970s and an unusually low percentage built in the 1990s. Plumbing and kitchen facilities are present in 94 to 100 percent of homes on most reservations in the District, slightly above average for reservations nationally. The one exception is the Pine Ridge reservation in South Dakota, where a lower percentage of homes with plumbing (78 percent) or kitchen facilities (84 percent) partly offsets its relatively high (for the Dakotas) homeownership rate among low-income households.
Crowding, or the number of persons per room, is generally higher in the American Indian community than in the U.S. as a whole. To the extent that a reservation's houses are unusually small, or to the extent that it is more common for multiple families on a reservation to crowd together in a single home, crowding could be a form of low housing quality that might lead us to discount high homeownership rates on some reservations. However, among American Indian households on large District reservations, the number of persons per room is comparable to that of reservations nationally. Differences are generally small within the District, too, except that on three District reservations, the percentage of owner-occupied homes with more than one person per room is well above the national reservation average of 18 percent. These three reservations are Pine Ridge (32 percent), Rocky Boy's (29 percent), and Crow (26 percent). Further study would be needed to determine whether crowding might offset the relatively high rate of homeownership among low-income American Indians on the Pine Ridge and Crow reservations.
Further research warranted
As we have seen, American Indian homeownership rates on large District reservations tend to be lower than on large reservations in the rest of the country, even after taking account of differences in income. Within the District, homeownership rates by income group are lowest in North Dakota and South Dakota. Housing quality differences between reservations do not seem to explain away these regional differences.
However, homeownership rates for American Indians on reservations vary widely, even within a specific region and among households with similar incomes. Homeownership rates are affected by many nonincome factors, including personal factors like net worth, access to credit, age, and marital status, as well as environmental factors such as the prevalence of trust land and the relative costs of renting vs. owning. In preliminary analyses, some of these factors were found to be statistically associated with the differences in American Indian homeownership rates shown in Figure 2. Still, significant differences remain that have yet to be explained. The large variations across reservations with respect to American Indian homeownership rates are a challenge to social scientists and policymakers and appear to be an appropriate topic for further discussion and research.
Federico Burlon, a student at Macalester College, completed the research for this article while serving as a Community Affairs intern at the Federal Reserve Bank of Minneapolis in 2008. He is grateful for the guidance of Professor Raymond Robertson at Macalester, who served as his faculty supervisor during the internship.
A note about homeownership
We do not mean to suggest in this article that homeownership is the best housing arrangement for all families. (For a recent critique of policies that promote homeownership, see Stephen Slivinski's article "House Bias: The Economic Consequences of Subsidizing Homeownership" in the Fall 2008 edition of the Federal Reserve Bank of Richmond's Region Focus magazine, available at www.richmondfed.org.) However, large ethnic or racial gaps in homeownership rates have been of concern to policymakers because they may reflect broader economic inequalities or discrimination. Better understanding of these gaps and their causes can help policymakers decide what kinds of programs, if any, may be appropriate to narrow them.
"American Indian" vs. "Native American"
Views within the indigenous community differ regarding whether "American Indian" or "Native American" is the preferred term. We use "American Indian" in this article, for consistency with the terminology used by our main data source, the U.S. Census Bureau.
A framework for comparing reservation homeownership rates
A note from the authors
Our article uses Census 2000 data to analyze American Indian homeownership rates on American Indian reservations while controlling for household income. Although these data are nearly ten years old and do not always align with other federal and tribal figures on homeownership, their level of detail and nationwide scope are unavailable elsewhere and provide a useful starting point for comparisons. We identify American Indian households as those whose heads reported their race as American Indian only, but similar results hold with a broader definition.
To control for household income, we report homeownership rates by three income categories. To obtain enough observations in each category, we define low-income American Indian households as those with incomes less than $20,000 (in 1999 dollars). Medium-income American Indian households are defined as those with incomes between $20,000 and $49,999. (By comparison, in 1999, the median household income for the whole U.S. was $42,000 and 22 percent of households had incomes less than $20,000.) The remaining American Indian households are deemed high-income. (We have obtained results similar to those reported here using more detailed categories.) On reservations with small populations, the number of households in the medium- and high-income categories is often too small to compute a meaningful homeownership rate. Therefore, reservations with fewer than 500 households are omitted, which cuts the number of District reservations analyzed from 41 to 24 and eliminates all reservations located in the Upper Peninsula of Michigan.Because the pattern of reservation homeownership was similar in North Dakota and South Dakota, and because the large Standing Rock reservation straddles their border, the reservations in these two states are grouped together for some purposes. Similar considerations led us to also group the reservations in Minnesota and Wisconsin. Thus, we analyze homeownership rates on large District reservations in five states, divided for some purposes into the following three groups: 1) North and South Dakota, 2) Montana, and 3) Minnesota and Wisconsin. We also compare these large District reservations to other large reservations (those with 500 or more American Indian households) nationally.
2/ To learn how a similar approach was used to study the factors affecting homeownership in the Hmong-American community, see Accounting for Regional Migration Patterns and Homeownership Disparities in the Hmong-American Refugee Community, 1980–2000, by Richard M. Todd and Michael Grover, Federal Reserve Bank of Minneapolis, October 2008.
3/ Specifically, for each income category and geography, 25 percent of the reservations analyzed have homeownership rates lower than the bottom end of the box, and 25 percent have homeownership rates higher than the top end of the box. The length of the box, known as the interquartile range (IQR), thus shows the range of typical homeownership rates within each category.
4/ These lines, known as "whiskers," extend as far as one and a half times the IQR. The whisker will be shorter than that if the last reservation above or below the box is less than one and a half IQRs from the box. In that case, the whisker stops at the reservation that is most distant from the box.