|
Article Excerpt
Credit literacy depends, in part, on understanding credit reports and scores. The U.S. Government Accountability Office conducted a study in 2004 to assess consumers' knowledge of credit reports, credit scores, and the dispute resolution process. This study uses the Government Accountability Office data and estimates a series of ordinary least squares and quantile regressions to identify specific subgroups of the population that could benefit from more targeted consumer policies and financial education. The findings from this research have important implications for consumer educators, financial professionals, and policy makers, especially with respect to national strategies designed to improve consumers' financial well-being.
**********
The Fair and Accurate Credit Transactions Act (FACT Act) was enacted in 2003 and amended the Fair Credit Reporting Act, a federal law that regulated, in part, who was permitted to access consumers' credit report information and how that information could be used. The FACT Act addressed ongoing concerns about inaccuracies in credit reports by providing consumers with new tools to enhance the accuracy, security, and reliability of their financial information. For example, the act expanded access to credit information by entitling consumers to one free credit report annually from each of the three major credit reporting agencies (CRAs) (Experian, Equifax, and TransUnion). The FACT Act also included a provision, which mandated that the U.S. Government Accountability Office (U.S. GAO) evaluate consumers knowledge and experience with credit reporting.
As part of its response to the mandate, the GAO conducted a study in 2004 to assess consumers' knowledge of credit reports, credit scores, and the procedures needed to correct credit reporting errors (U.S. GAO 2005). This study extends the work of the GAO. Specifically, the GAO data are used to estimate ordinary least squares (OLS) and quantile regressions to determine how the impact of demographic factors varies across consumers with different levels of credit reporting knowledge. A priori one might think that consumers with widely varying credit knowledge scores would respond in very different ways to key determinants of financial behavior--differences that could be obscured by measuring only mean effects. Evidence from the data appears to support this hypothesis. Results from the standard OLS model identify, in general, demographic groups that are less credit knowledgeable than others. However, the findings from the quantile regressions identify more precisely specific demographic subgroups that are likely to benefit from more targeted financial education.
While knowledge-based surveys continue to be widely used, it is important to acknowledge that financial education research is moving toward using more behavior-type measures (Lyons 2005; Lyons, Chang, and Scherpf 2007; Lyons et al. 2006). However, documenting changes in actual financial behavior is often difficult. Follow-up interviews with participants are needed but rarely occur because of time and resource constraints within organizations (Lyons 2005; Lyons, Chang, and Scherpf 2007; Lyons et al. 2006). Also, it is not always practical or appropriate to collect behavior information. The appropriateness of particular indicators depends on program objectives, delivery methods, and target populations.
This study contributes to the literature by applying a fresh methodological approach to financial education research. Using quantile regression analysis, we show how valuable insight can be gained from a knowledge-based survey using data collected from a random sample of the U.S. population. Much of the existing research uses data collected from convenience samples such as participants of specific financial education programs. Thus, our study allows us to generalize our findings to the U.S. population as a whole. In particular, we demonstrate how quantile regression analysis can be a useful tool in identifying specific subgroups of U.S. consumers who could benefit from more targeted consumer policies and financial education.
LITERATURE REVIEW
A growing body of literature provides general insight into consumers' financial literacy levels and the factors that affect their financial knowledge and behavior. For an overview, see Bell and Lerman (2005), Braunstein and Welch (2002), Fox, Bartholomae, and Lee (2005), Hilgert, Hogarth, and Beverly (2003), Hogarth (2002), Hogarth, Beverly, and Hilgert (2003), Lyons (2005), Lyons et al. (2006), and National Endowment for Financial Education (2004). This research typically has concluded that providing financial information and education results in positive improvements in consumers' financial literacy levels.
Credit reports and credit scores affect many aspects of consumers' lives (U.S. GAO 2005). Both can influence lenders' decisions to grant credit and can affect a consumer's ability to get a job, rent an apartment, buy a car or a home, or purchase insurance. Higher credit scores can improve the likelihood that a loan will be granted and can lower the interest rates consumers receive on their loans. Given the growing importance that credit reports and credit scores play in the lives of consumers, it has become particularly important that researchers study credit literacy to better understand how consumer education policy can be used to help improve consumers' ability to manage their own credit and ultimately their long-term financial security. Yet, little research has been conducted in this area.
In 2004, the U.S. GAO conducted a literature search to identify leading research related to consumers' knowledge of credit reporting issues and the extent to which consumers reviewed their credit reports and credit scores (U.S. GAO 2005). An attempt was made to evaluate the quality of this research, the data used, and the importance of the findings. However, limited documentation on survey design, methodology, and sampling techniques made it difficult to conduct a thorough review (see U.S. GAO [2005], appendix III). According to the GAO report, few studies attempted to investigate consumers' knowledge of credit reporting issues. Of those studies that existed, most were primarily descriptive in nature and few used data collected from random samples of known populations.
The GAO cited a few studies for which there was some indication that a random sample of a known population had been used. The following are highlights from these studies that can serve as background information and later comparisons for the findings presented in this article. For example, research conducted by American Association of Retired Persons (2003), Consumer Federation of America (2003), Hilgert, Hogarth, and Beverly (2003), and Consumer Federation of America and Providian (2004) investigated consumers' general knowledge about credit reports and scores including how information is collected, how credit reports and scores can be obtained, the factors that impact credit histories, and consumers' experiences in reviewing their own credit reports and scores. The findings from these studies tended to show that many consumers lack basic knowledge about credit reports and scores. One study by Consumer Federation of America and Providian (2004), which focused primarily on consumers' knowledge about credit scores, found that most consumers do not understand their scores, even when they think their knowledge of credit is good. Specifically, those who thought their knowledge of credit scores was "good" frequently answered credit knowledge questions incorrectly. Not surprisingly, it also was reported that those who considered their knowledge of credit scores to be "poor" were more likely to know less about credit scores.
Additional research has investigated the credit reporting process including the accuracy of credit reports and the dispute process used to correct errors (i.e., Avery et al. [2003] and Avery, Calem, and Canner [2004]). However, these studies do not adequately capture the extent to which consumers understand the dispute resolution process and their consumer rights with respect to credit reporting.
So what have we learned? Previous research provides basic insight into consumers' knowledge of credit reports and scores and the credit reporting process, but these are primarily descriptive in nature and are limited by the design, methodologies, and sampling techniques. Few have used random samples of a known population. Also, few have used rigorous empirical methods to gain a more thorough understanding of consumers' credit literacy. In fact, to our knowledge, we are not aware of any other studies that have used quantile regression analysis to investigate consumers' financial knowledge, which allows us to conduct more targeted analysis and elucidate those factors that have a notable effect on particular segments of the population.
Koenker and Basset (1978) were the first to introduce quantile regression as an alternative to OLS. The primary advantages of the quantile approach are its robustness to violations of the assumption of a normal, or "Gaussian," distribution of errors, to outlier observations, and to monotonic transformations of the dependent variable. Moreover, quantile regression provides a more complete characterization of the impact of independent variables than OLS.
Researchers have applied quantile methods to an increasingly eclectic array of topics (for an overview, see Koenker and Hallock [2001]). Some researchers have used quantile regression to investigate the effect of education on the distribution of earnings (Bedard 2003; Buchinsky 1994, 1998a, 1998b). The quantile studies that most closely relate to ours examine academic test scores.
For example, Eide and Showalter (1998) used quantile regression to examine the relationship between school quality and student performance. The dependent variable in their model was the change in the score on a standardized mathematics test between the student's sophomore and senior year. They measured the effect of the key independent variables on changes in test scores at the 5th, 25th, 50th, and 95th percentiles, in addition to measuring the mean effect of these variables using OLS. The OLS regression yielded insignificant results for the key policy variables, such as pupil-teacher ratio, length of the school year, district level per pupil expenditures, and educational attainment of the instructors. The quantile regressions, however, revealed positive effects for some measures of school performance. For instance, Eide and Showalter found that lengthening the school year had positive and significant effects on students, who had test score gains at the 50th, 75th, and 95th percentiles. Also, per pupil district-level expenditures had a significantly positive effect on test scores gains for students at the fifth percentile. By supplementing...
|
