Consumer Research Contributions to the Evaluation of Public Policy: the Case of Truth-In-Lending (Abstract)

George S. Day, Stanford University
William K. Brandt, Columbia University
[ to cite ]:
George S. Day and William K. Brandt (1974) ,"Consumer Research Contributions to the Evaluation of Public Policy: the Case of Truth-In-Lending (Abstract)", in NA - Advances in Consumer Research Volume 01, eds. Scott Ward and Peter Wright, Ann Abor, MI : Association for Consumer Research, Pages: 82-84.

Advances in Consumer Research Volume 1, 1974      Pages 82-84


George S. Day, Stanford University

William K. Brandt, Columbia University

[The authors are indebted to the National Commission on Consumer Finance for financial support of the study and to Murray Silverman and Terry Duetscher for their contributions. The complete study is reported in National Commission on Consumer Finance, Technical Study #1. Washington, D.C.: U.S. Government Printing Office, 1973.]

[George S. Day is Associate Professor of Marketing at Stanford University. William K. Brandt is Assistant Professor of Marketing at Columbia University.]

Two converging trends of potential interest to consumer researchers are explored in this paper: the methodology of evaluation research and the effectiveness of information disclosure as a means of assisting or protecting consumers. Research conducted by the authors for the National Commission on Consumer Finance illustrates the application of the methodology to evaluate the impact of Truth in Lending (TIL) legislation on consumer behavior. This paper demonstrates how concepts developed from consumer behavior theory were combined with the methodology of evaluation research to assess the result of this public policy.

All evaluations are designed to judge the results of programs or activities and to provide a basis for subsequent decisions about the program. The growing number of social programs and the corresponding concern over their costs and effectiveness have focused much attention on the formal evaluation of program results. The evaluator is faced with the same basic questions in virtually all cases: to what extent is the program achieving its goals? Although the basic evaluation question remains the same, circumstances beyond the evaluator's control frequently limit the choice of research designs. Several of these designs are listed below:

Most preferred

- Pure experimental design

- Quasi-experimental design

- Correlational design with statistical controls

- Qualitative audits by outside observers

- Narrative self-evaluation

Least Preferred

In this evaluation study of TIL [Truth in Lending was implemented in July 1969 and required uniform disclosure of annual percentage rate of interest and dollar finance charges for virtually all consumer credit purchases.] a correlational design with statistical controls was chosen to test the following hypotheses: (1) that TIL disclosure had no substantial effect on information search in comparison shopping for goods or credit, (2) that TIL disclosure had no substantial effect on purchase behavior to postpone or to use cash instead of credit, and (3) that the relative impact of TIL disclosure on low-income and minority consumer behavior was less than for other segments.

Three important models or concepts from consumer behavior theory were incorporated in the research design: the decision-process model, the hierarchy of effects model, and the concept of attitude structure. The Engel decision-process model [Engel, James; Kollatt, David; & Blackwell, Roger. Consumer Behavior. New York: Holt, Rinehart & Winston, 1968.] was further specified to include credit-related decisions which might occur in the purchase of major durable goods. Two central elements studied were the influence of credit availability on choice of retailer or credit source and the impact of information about credit costs provided by TIL.


The data were obtained from personal interviews with 793 California heads-of-household in October 1970 (fifteen months after TIL implementation) and reinterviews with the random subsample of 196 households in July 1971. The initial survey included a representative sample of 641 households from across the state and a separate probability sample of 152 black households. The reinterview subsample was drawn from the representative sample.

Both interviews gathered extensive data about the respondent's knowledge and attitudes towards consumer credit and credit sources, history of credit experience, and demographic and socio-economic characteristics. For ear or major household durable purchases during the previous year, an additional set of questions retraced the sequence of purchase decisions in considerable detail. For credit purchases additional questions focused on each important credit decision. Knowledge about credit costs was evaluated in two ways: (1) respondents who made a major credit purchase were asked the interest rate and dollar finance charges for the credit purchase, and (2) all respondents estimated the interest rate and the total cost (including finance charges) of a hypothetical credit purchase. Institutional knowledge or general consumer perceptions about major credit sources were measured in three ways: an assessment of relative interest rates of various credit sources, an evaluation of the perceived difficulty to borrow from various sources, and similarity judgments about each pair of sources.

For both actual and hypothetical purchases, respondents tended to underestimate interest rates and to overstate finance charges. For the hypothetical purchase 45 percent significantly underestimated the annual percentage rate while 66 percent overestimated the dollar finance charges. Despite the TIL disclosure information most consumers apparently did not understand the arithmetic conversion between annual percentage rate and dollar finance charges for installment credit.

The table below shows that while the net change in learning between the two interviews was 10 percent, the total amount of change was much greater with 15 percent forgetting and 25 percent learning something about interest rates in the interim.


Standardizing the row figures in this table shows that the 47 percent level of knowledge in July 1971 was an effective ceiling or equilibrium level. Without significant changes in the environment it is unlikely that consumer knowledge will rise much above this level.

Multiple measures designed to measure the relationship between TIL disclosure and consumer behavior showed: (1) that credit availability and better credit terms were not influential in consumers' choice of retailers, (2) that disclosure on credit contracts and monthly statements did not influence behavior, and (3) that accurate knowledge about credit costs was not strongly related to search for credit information and sources. The influence of TIL disclosure on the use of credit also revealed very minor effects. It should be noted when evaluating the disclosure-behavior relationship that fifteen months of TIL information may be too short a period to expect significant changes in consumer buying behavior.

An evaluation of possible constraints on the credit decision focused on the special problems of low-income and minority buyers. Three major constraints were identified: (1) that the credit decision is of little importance when compared with other decisions in the purchase process; (2) that two-thirds of all credit buyers rely on the retailer for credit (much higher for low-income and minority buyers) without giving thought to other lower cost sources; and (3) that although consumers understand the relative cost of credit sources, low-income and minority families perceive that credit is inaccessible to them from low cost sources.