M O R T G A G E LENDING D E T E R M I N A N T S IN A M A J O R M E T R O P O L I T A N A R E A

Harold Black and Robert L. Schweitzer

The purpose of this paper is twofold. First, the paper analyzes the deter- minants of mortgage lending terms at a commercial bank and at a mutual savings bank in a major metropolitan area. Secondly, the paper addresses the question whether the clientele of the institutions differ with respect to the mortgage applications. The data set consists of observations on mortgage lending terms and borrower and property characteristics for mortgage loan applications from each of the financial institutions. The survey was conducted from January 1978 to June 1978.

The coefficients from the estimated equations are tested for statistical significance. The estimated equations include the following variables. The loan terms are amount requested, percentage down payment, years to maturity, and interest rate. Although it is recognized that those terms may be interdependent, that particular problem is not considered here. Indeed, interdependence of loan terms exists given that the market for mortgages is not cleared solely by the movement of interest rates. That is due either to the inflexibility of mortgage rates because of political constraints such as usury ceilings or the willingness of the applicant and institution to trade off among the various terms. 1 In particular, a lower interest rate may be offered the applicant provided a higher down payment is forthcoming. Thus, to clear the market for mortgages, the rationing of credit is ac- complished through an interaction of the entire package of loan terms rather than simply via movements in the mortgage interest rate.

Moreover, in the case of mortgage applications, the determination of the willingness of the financial institution to grant the application will depend on the creditworthiness of the borrower. The functions used in this paper include as measures of creditworthiness ratio of monthly hous-

MORTGAGE LENDING DETERMINANTS 429

ing expense to gross monthly income, net worth, and the ratio of monthly debt payments to gross monthly income. 2

Also included in the equations are property characteristics of the resi- dence as well as neighborhood characteristics. These are age of the prop- erty, and percentage of households in the census tract in which the prop- erty is located that are headed by individuals who are black, female, or over 62 years of age. It is assumed here that certain properties are per- ceived, rightly or wrongly, as being in areas of high risk. Hence, in order for a loan to be noninsured and financed conventionally within a per- ceived high-risk area, the lender would seek to charge a higher interest rate or to make the terms of the loan more favorable (to the lender) in order to offset the additional perceived risk factor. If higher interest rates are prevented by ceilings, a higher percentage down payment may be required, or a shorter maturity. Such differences in terms might be applied to all applicants requesting loans in perceived high-risk areas, regardless of the personal characteristics of the applicants. 3

Finally, data were collected relating to the personal characteristics of the applicants. Those variables are race, age, sex, and marital status. Each is a prohibited lending factor in that financial institutions are pro- hibited from considering such factors in the loan decision. An additional variable, whether the applicant is a customer of the institution, is also included.

The above variables are employed in this study to determine whether the clientele of the commercial bank and the mutual savings bank are separable. Such a determination is important in these times of a general movement by financial institutions and by the Congress toward greater financial homogenization.

The question addressed below is whether commercial banks and mutual savings banks serve different customers. In order to address that point, multivariate probit analysis is employed. 4 The model specified includes all variables discussed above. Each variable is postulated as having a particular sign. The dependent variable is coded zero for a bank applicant and one for an applicant at the mutual savings bank. No applicant, in this instance, has made applications at both institutions. A positive coefficient implies that as the value of the independent variable increases, the value of the dependent variable approaches ! (the applicant applies at the mu- tual savings bank).

The assumption here is that the clientele are separable. It is anticipated that the applicants at the bank are more likely to apply for a lower amount and to have a higher down payment, a shorter loan maturity, lower

430 The Review of Black Political Economy

housing expense to gross income, higher net worth, and lower monthly debt to gross income ratio. The interest rate will be higher for mutual savings bank customers. It is further assumed that the bank customers will apply for younger property, be male, white, and married. Their proposed neighborhoods will have fewer Blacks, fewer female heads of house- holds, and younger households. Moreover, it is assumed that the appli- cants will most likely be customers of the banks than of the mutual savings banks. In essence, the implication is that bank customers will be those with the lower perceived risk (if the above can be assumed to be prima facie evidence of riskiness).

Two functions are estimated. They include the same variables, but the first contains all applications regardless of disposition while the second includes only applications that were accepted. Rejected applications could not be considered here due to the paucity of rejections. There are only eight bank rejections and 20 rejections at the mutual savings bank. Consequently, the first estimated equation explains who applies for loans at these institutions and the second equation explains the characteristics of the accepted applications.

Lastly, this paper has shown that the clientele of the two institutions with regard to mortgage applications are similar. The difference, how- ever, is on terms of loans rather than on personal characteristics or on economic bases. That is, no significant differences are found for race, marital status, and sex of the applicant. No significant differences are found for income and net worth. However, applicants at the bank have significantly less monthly debt payments, ceteris paribus. That may point to the assertion, often made by bankers, that they serve the mortgage demands of only their better customers and refer other less-preferred (higher-risk) customers to other institutions. However, such a conclu- siorv--that higher monthly debt provides an indication that the two popu- lations are different--is tentative given that the other economic variables are statistically insignificant.

NOTES

1. For example, see Ang and Willhour (I). Hester (4), in his work on loan offer functions, also discusses this point.

2. Net worth is defined here to be total assets less total debt. Total assets includes the market value of the current residence. Total debt is all debt outstanding including mortgage debt.

3. On this point see Black, Schweitzer and Mandell (2).

MORTGAGE LENDING DETERMINANTS 431

4. For a discussion of the technique see Boczar and Black, Schweitzer, and Mandell (3,2)

5. The age-of-applicant variable is not included in the final form due to lack of obser- vations.

LITERATURE CITED

1. Ang, James and Richard Willhour. "The Consumer Loan Supply Function of a Minority Bank: An Empirical Note." Journal of Money, Credit and Banking (May 1976).

2. Black, Harold; Robert L. Schweitzer; and Lewis Mandell. "Discrimination in Mortgage Lending." American Economics Review, (May 1978).

3. Boczar, Gregory E. "Competition Between Banks and Finance Companies: A Cross-Section Study of Personal Loan Debtors." Journal of Finance (March 1978).

4. Hester, Donald D. "An Empirical Examination of a Commercial Bank Loan Offer Function." Yale Economical Essays (Spring 1962).