Section Three: Analysis of Conventional Mortgage Lending within the Urban American Indian Community

Theme One: Analysis of HMDA

Original research conducted by the Roy Wilkins Center for Human Relations and Social Justice attempted to determine whether a gap between the rejection rates for mortgage applications for whites and non-whites existed in the Upper Midwest (i.e. Iowa, Minnesota, Nebraska, North Dakota, South Dakota, and Wisconsin). This analysis found that a gap or disparity did exist between rejection rates for whites and non-whites and more specifically, between whites and American Indians.

The rejection rate for mortgage loan applications for American Indians in the Upper Midwest was 11.8 percent in 1992. This means that of the total applications for mortgage loans reported under HMDA for American Indians, 11.8 percent were rejected. For whites during that same time, only 7.5 percent of the loan applications were rejected. (Myers, et al., 1993)

The analysis went further to determine whether this gap between rejection rates could be explained by characteristics of the loan, applicant, or neighborhood. To do this, the 1992 HMDA data was merged with the 1990 Census data in order to describe characteristics of the neighborhood from which the loan applications were generated. (1992 data can be applicable to the current mortage lending situation. Interest rates had fallen by this time and are similar to current rates. The demand for mortgage lending in the conventional market should be very similar between 1995 and 1992.)

1992 Mortgage Loan Rejection Rates by Race of Applicant, Upper Midwest
Type of Loan	White	American Indian
Conventional Loans	7.00%	11.80%
Federal Housing Administration (FHA) Insured Loans	11.20%	10.40%
Veterans Affairs Loans (VA)	9.40%	13.80%
Farmers Home Administration (FmHA)	33.30%	*100.00%
Purpose of Loan	White	American Indian
Home Purchase	10.10%	16.40%
Home Improvement	10.80%	14.10%
Refinancing	4.70%	7.20%

Source: HMDA, 1992
* N=1

The following information describes the number and characteristics of the loans applied for by American Indians reported during 1992. In addition, the reason for denial (or rejection) is also noted. These reasons were provided by the lenders. The total number of applications for American Indians in this Upper Midwest dataset is 1,468.

Type of Loan	Percent
Conventional Loans	84.20%
Federal Housing Administration (FHA) Insured Loans	11.80%
Veterans Affairs Loans (VA)	4.00%
Farmers Home Administration (FmHA)	0.00%
Purpose of Loan	Percent
Home Purchase	35.20%
Home Improvement	19.60%
Refinancing	45.10%
Multi-family Dwelling Purchase	0.10%
Reasons for Denial	Percent
Debt to Income Ratio (too high)	19.30%
Employment History Insufficient	5.20%
Credit History	35.40%
Collateral Insufficient	9.40%
Insufficient Cash	4.20%
Unverifiable Information	1.60%
Credit Application Incomplete	2.10%
Mortgage Insurance Denied	0.50%
Other	6.80%

Data: HMDA, 1992

For American Indians in Minnesota, the HMDA dataset had 523 loan applications in 1992. The mean (or average) income for American Indian applicants was $52,470. The average amount for loans requested by Indian applicants was $65,580. Similar to the Upper Midwest dataset, American Indian applications in Minnesota had a rejection rate of 11.3 percent.

One aspect of this analysis was to examine where mortgage loan applications were originated, rejected, and "accepted." ("Accepted" means "not rejected." There are instances where loan applications may not have been rejected, but were not accepted either. For example, loan applications may have been withdrawn during the process by the applicant.) To do this, we created eight categories to describe particular areas and neighborhoods by income. These categories are broken down into percentages of the state median income for a family of four ($30,909) (From the City-County Databook, 1994. State medial household income in 1989):

less than 40%..less than $15,960
40% to 60%..$15,960 - 23,940
60% to 80%..23,940 - 31,920
80% to 100%..$31,920 - 39,900
100% to 120%..$39,900 - 47,880
120% to 140%..$47,880 - 55,860
140% to 160%..$55,860 - 63,840
greater than 160%..greater than $63,840

Categories for describing neighborhoods which are predominately white, integrated by race and ethnicity, and predominately people of color were also constructed. The following six categories describe the percentage of nonwhites living in a census tract as defined by the U.S. Census Bureau:

less than 10 percent nonwhite population
10 percent to 20 percent nonwhite population
20 percent to 30 percent nonwhite population
30 percent to 40 percent nonwhite population
40 percent to 50 percent nonwhite population
greater than 50 percent nonwhite population

The following breakdown of the total 523 American Indian loan applications for 1992 in Minnesota by neighborhood characteristics are:

1992 American Indian Loan applications in Minnesota by Census Tracts with the following average annual income compositions
Census tracks by average annual household income	Percent of all loan applications
less than $15,960	1.00%
$15,960 to $23,940	3.30%
$23,940 to $31,920	5.70%
$31,920 to $39,900	17.60%
$39,900 to $47,880	22.90%
$47,880 to $55,860	18.50%
$55,860 to $63,840	17.40%
greater than $63,840	13.60%

Data: HMDA, 1992

1992 American Indian Loan Applications in Minnesota by Census Tracts with the following racial and ethnic (nonwhite) composition
Census tracks by percent of non-white population	Percent of all loan applications
less than 10% nonwhite	88.30%
10% to 20% nonwhite	7.60%
20% to 30% nonwhite	0.80%
30% to 40% nonwhite	0.20%
40% to 50% nonwhite	*0.00%
More than 50% nonwhite	3.10%

Data: HMDA, 1992; * no applications were made (n = 0)

An interesting note is the number of loan applications generated in low and moderate income census tracts and in tracts which are predominately nonwhite. There were five (5) loan applications made in the lowest income census tracts. Four of those loans fell into census tracts with greater than 50 percent people of color. An additional seven (7) loan applications came from census tracts with median incomes between $15,960 and $23,940 and which are greater than 50 percent nonwhite.

When examining the applications for these tracts in more detail, we found seven applications by American Indians were generated in the census tracts which make up the Phillips neighborhood in Minneapolis. All seven applications were accepted. (The median income for these loan applications was $14,700. Five of the loans were for home purchase and two were for refinancing.)

We can also determine the number of loan applications rejected by "neighborhoods." For example, as shown in the first line, those applications generated in the lowest income census tracts had a rejection rate of 40 percent. The rejection rates for all American Indian loan applications by neighborhood characteristics are as follows.

American Indian Loan Applications
Census tracks by percent of non-white population	Percent of loans rejected
less than 10% nonwhite	10.20%
10% to 20% nonwhite	15.00%
20% to 30% nonwhite	0.00%
30% to 40% nonwhite	0.00%
40% to 50% nonwhite	*0.00%
More than 50% nonwhite	**37.50%

Data: HMDA, 1992
* n = 5, ** n = 30

American Indian loan Applications
Census tracks by average annual household income	Percent of loans rejected
less than $15,960	*40.00%
$15,960 to $23,940	5.90%
$23,940 to $31,920	**26.70%
$31,920 to $39,900	12.00%
$39,900 to $47,880	11.70%
$47,880 to $55,860	12.40%
$55,860 to $63,840	8.80%
greater than $63,840	4.20%

Data: HMDA, 1992
* no applications were made (n = 0), ** n =16

The HMDA data also allows us to examine the actual location where loan applications were made. By determining the census tract numbers for the central city limits of Minneapolis and St. Paul, we found that 23.7 percent of all loan applications were made within the city of Minneapolis and 13.5 percent were made within the city of St. Paul. These figures do not include the percentage of loans granted or rejected. Because HMDA data is collected for Metropolitan Statistical Areas (MSA's), we can assume that a substantial portion of the remaining loan applications were generated within the seven county metropolitan area. Further examination of the census tracts will determine the exact percentage.

Similar to the loan applications for American Indians in the entire Upper Midwest, the reasons for the denial of loan applications within the state of Minnesota follow the same pattern. See the chart on the opposite page.

One special note should be made: A substantial portion of the loan applications rejected did not have a reason reported by the lending institutions. Roughly 32 percent of the 59 loan applications made by American Indians did not have a reason for denial reported. For home purchase loans only, there were 26 loan applications rejected. The primary reasons for the rejections of home purchase loans as compared to all of the loan applications were for debt-to-income ratio and bad credit; approximately 19.2 percent of home purchase loans were rejected for high debt-to-income ratios and 30.8 percent were rejected for bad credit, respectively.

In addition to the reasons provided by the banks and the cultural implications and economic barriers described in the earlier portion of this report, are there other explanations for the gap between rejection rates for American Indians and whites? Is there something about the mortgage lending process that could impact the outcome of the lending decision?

1992 Loan Applications for American Indians in Minnesota: All types of loans
Reasons for Denial	Number of loans rejected	Percent of all loans rejected
Debt to Income Ratio (too high)	13	22.00%
Employment History Insufficient	0	0.00%
Credit History	18	30.50%
Collateral Insufficient	4	6.80%
Insufficient Cash	2	3.40%
Unverifiable Information	0	0.00%
Credit Application Incomplete	1	1.70%
Mortgage Insurance Denied	0	0.00%
Other	2	3.40%
No reason given	19	32.00%

Data: HMDA, 1992. n=59

Next Section

The Well-Being of American Indian Children in Minnesota: Economic Conditions, 1994