Profitwise News and Views

September 2003 Issue

Seeds of Growth Conference Summary

Seeds of Growth Sustainable Community Development: What Works, What Doesn't, and Why

Session One - Counseling and Intervention
The four papers presented in the first session focused on counseling and financial education programs: pre-purchase counseling, credit counseling, retirement planning seminars, and the relationship of financial knowledge to financial behaviors. Each of the papers finds that increasing financial knowledge - whether through counseling, teaching, or experience - improves both financial behaviors and financial performance.

The first paper by Abdighani Hirad and Peter Zorn, Freddie Mac's vice president of Housing Economics, A Little Knowledge is a Good Thing: Empirical Evidence of the Effectiveness of Pre-Purchase Homeownership Counseling, examines data gathered on 40,000 mortgages originated between 1993 and 1998 under Freddie Mac's Affordable Gold program. Recipients of mortgages made under the Affordable Gold program are required to receive counseling, although it is up to the lender to determine the type and source of counseling borrowers receive. When an Affordable Gold loan is submitted to Freddie Mac, the lender must record the organization that provided the counseling (lender, nonprofit, government agency, or "other") and the type of counseling that was provided (classroom, home study, individual counseling, or "other").

The authors present empirical evidence that pre-purchase counseling reduces the likelihood of 90-day delinquency. The reduction in that default risk can be confidently attributed to the counseling. Although there is no significant variation in the rate of risk reduction associated with specific counseling providers,¹ there is a wide variation in the rate of risk reduction among different types of counseling.

On average, borrowers getting counseling are 19 percent less likely to go into 90-day delinquency than borrowers not receiving counseling, regardless of counseling type or counseling provider, Hirad and Zorn find. Individual counseling is the most effective and provides an estimated 34 percent reduction in 90-day delinquency rates. Classroom and home study provide 26 and 21 percent risk mitigation, respectively. Telephone counseling provides no statistically significant reduction in delinquency rates (Figure 1).

Zorn, who presented the paper, noted that the study ignores the possible effects of timing of delinquency. It may be that counseling simply delays delinquency.

The authors conclude that counseling can play a crucial role in expanding affordable homeownership opportunities. They raise the possibility that counseling might have a larger impact if it were limited to individual- or classroom-type counseling, and they suggest that more recent data may show an even bigger impact of counseling due to the increasing sophistication of the counseling industry.

In The Impact of Credit Counseling on Subsequent Borrower Credit Usage and Payment Behavior, authors Gregory Elliehausen, E. Christopher Lundquist, and Michael Staten, director of the Credit Research Center at Georgetown University, find that credit counseling improves clients' credit profile and financial behavior.

Michael Staten's paper begins with an overview of the credit counseling industry. The number of people seeking credit counseling grew from 500,000 in 1991 to 2.5 million in 2001. Some reasons offered for such rapid growth are increased bankruptcy filings (875,000 declared bankruptcies in 1991; 1.5 million in 2001), and the now widely accepted use of counseling as part of the prescription for both bankruptcy reform and anti-predatory lending legislative initiatives.

At the same time, market dynamics impact the counseling industry. Fees from debt management plans (DMP),² which until recently produced enough revenue to subsidize non-DMP counseling, are declining while new market entrants are also driving down the cost of counseling.

The study tracked 14,000 clients who received counseling between April and August of 1997 at one of five National Foundation for Credit Counseling (NFCC) member organizations.³ A random sample of 98,000 borrowers with credit reports in both 1997 and 2000 provided a comparison group of similar borrowers who received no known counseling.⁴ The paper documents the short-term and long-term impacts of credit counseling on clients' borrowing and payment behavior. The counseling was provided by nonprofits for people who did not adopt a debt management plan.

Each member of the counseled group had at least one 60- to 90-minute individual session with a certified counselor. Each session included a discussion of the family budget, financial goals, and spending patterns. Each participant left with a customized, written action plan. Follow-up sessions were held if needed.

Two credit bureau reports, one in June of 1997 and the other in June of 2000, were obtained for each borrower in the counseled group⁵ and a random sample of non-counseled borrowers with similar characteristics from the same geographic area. The authors compared the experiences of counseled and non-counseled borrowers to borrowers having similar risk profiles at the beginning of the period.

The study found that borrowers who received financial counseling in 1997 generally improved their overall credit profile - Empirica risk score⁶ - over the next three years relative to non-counseled borrowers who had similar credit profiles in 1997. These findings stand, according to the authors, when correcting for self-selection bias and holding constant other factors (job loss, divorce, etc.) that may influence the performance of both groups over time.⁷

The impact was the greatest for borrowers who were least credit-worthy at the start of the observation period. For counseled borrowers in the lowest-scoring 10 percent of initial risk scores, the average score improvement over the subsequent three years for counseled borrowers equated to a 38 percent reduction in the predicted likelihood of a bankruptcy or charge-off. Improvement in the risk scores attributable to counseling diminished for borrowers in higher initial score ranges.

Notably, the researchers observed that counseled borrowers experienced improvements across a range of credit behaviors relative to similar non-counseled buyers even when their risk scores didn't improve relative to the comparison group. The authors note improvements among the counseled group in several categories of credit utilization, including:

Number of accounts with positive balances (fewer)
Total debt
Non-mortgage debt
Credit card debt
Percentage utilization of bank card lines
Number of 30+ and 60+ day delinquencies.

Staten cautioned that the behavioral findings may apply only to one-on-one counseling using the specific counseling protocol used by NFCC member agencies.

In Financial Education and Retirement Savings, Robert Clark, Ann McDermed, Kshama Sawant, and Madeleine d'Ambrosio explore the impact of financial education seminars on individuals" financial goals and behaviors. The rise of defined benefit plans and the potentially lower benefits from Social Security mean that individuals will have increasing responsibility for their own retirement savings.

The analysis is based on a standard economic lifecycle model of decision making: individuals start with retirement goals (when to retire, amount needed to retire) and make decisions based on the present level of wealth and savings rate needed to reach the goal. Individuals pay for retirement through reduced consumption in the course of their working life.

The standard model assumes that individuals have an estimate of their expected lifetime earnings and both a grasp of basic financial mathematics and the risk-return trade-off among investment choices. The authors question the validity of these assumptions across the broad population.

The authors surveyed 633 participants in financial education seminars conducted by the Consulting Services division of Teachers Insurance and Annuity Association - College Retirement Equities Fund (TIAA-CREF) across the country. The TIAA-CREF consultants asked participants to complete a pre-seminar survey dealing with retirement age and income goals and the likelihood (in their view) of achieving those goals. Participants then took part in a seminar designed to provide financial information that would assist individuals in the retirement planning process. Following the seminar and before the participants left the room, another survey asking them to indicate changes in retirement goals or changes in anticipated financial behaviors based on information obtained in the seminar was completed. The findings include:

Twelve percent of participants changed their retirement age goals.
Those with relatively low expected retirement ages were more likely to raise their retirement age.
Twenty-eight percent changed their retirement income goal.
Thirty-seven percent of those with desired income replacement rates of 65 percent or less raised their target.
Forty-one percent planned to establish a supplemental pension plan.
Forty-three percent of those who already had a supplemental retirement plan planned to increase contributions to supplemental plans.
One-third of participants intend to alter their investment allocation between equities and bonds.

"Financial education does, in fact, cause individuals to rethink their retirement goals and to change or to plan to alter their savings and investment behavior. Most people making changes in their retirement goals tend to raise their retirement age, increase their desired income level, and increase their savings rates. Women are more likely than men to make changes across all of those dimensions," stated Clark. A follow-up paper focusing on gender differences in this research will be available on the TIAA-CREF website in the near future.

The analysis indicates that new information provided in the seminar caused participants to alter their retirement age and income goals and behavior, making them more consistent with economic expectations. The authors call for expanded financial education programs based on their analysis and the trend toward more self-directed retirement plans.

Financial education tends to alter investment behavior and increase retirement savings. These changes should enhance retirement income security and potentially influence public policy regarding social security and employee pension plans.

The final paper of the session, Patterns of Financial Behaviors: Implications for Community Educators and Policy Makers, by Jeanne Hogarth, Marianne Hilgert, and Sondra Beverly, was "motivated by our desire to help community educators "triage" clients," according to Hogarth, a Consumer and Community Affairs program director for the Federal Reserve Board.

"You've heard that credit counseling is effective, but we all know that it's very costly to do one-on-one counseling. Seminars are effective and less costly, but some people don't do as well in the group setting. Some people may benefit from simply giving them a brochure. So the question is: how do we cut those slices - how do we triage clients to make the most effective and optimal use of our resources?" Hogarth stated.

The paper delivers a thorough overview of the literature on financial behavior, financial education, and financial knowledge of individuals and households, summarizing evidence on financial knowledge, the relationship between knowledge and behavior, and the effects of financial education on behavior. Unlike previous research, this study directly tests the assumption that education increases knowledge, which in turn affects behavior.

The study uses data gathered in November and December 2001 through the University of Michigan Survey of Consumers. The consumers in the survey were asked what financial products they own. They were asked about their financial management practices - bill paying, savings, and investing. They were also asked questions designed to determine their financial knowledge, how they learned about financial management, and how they would prefer to learn in the future.

Using the responses to the financial management and financial product ownership questions, the authors sorted the respondents into "high-middle-low" performance categories in each of three financial behavior categories: cash flow management, savings, and investment. The results of that sorting are shown in Table 1.

Among the findings: about two-thirds of respondents were adept at cash-flow management (88 percent pay their bills on time), about one-third had good savings behaviors, and about one-fifth had good investment behaviors (45 percent said they had a 401(k)). At the low end of the behavior measures, about 2 out of 5 respondents don't invest or save for retirement or long-term goals.

Hogarth highlighted three specific categories:

Superstars - the 10 percent of respondents who are "high" in every category.
Take the Next Step - the biggest group at 34 percent who are reasonably adept at cash flow management, but need help in organizing a savings and investment plan.
Need All the Basics - the 11 percent of respondents who are "low" in everything and who need basic, remedial work.

The study also includes a multivariate analysis to determine what characteristics were shared among those ranked "high" in the financial behavior categories. Variables in the analysis included: financial knowledge, financial learning experiences, measures of economic stability, motivation, and a cross-section of socioeconomic and demographic characteristics.

The analysis revealed that only two variables were consistently significant across the cash management, savings, and investing behaviors: financial knowledge and learning method. The knowledge effect was the largest for cash-flow measures.

Hogarth noted the correlation between certain demographic characteristics and the need for basic financial education. Young, single female heads-of-household who are either Black or Hispanic tend to need basic financial education, are less likely to be homeowners, have credit cards, set goals, or read about money management.

Respondents who said they learned from personal experience and family and friends were more likely to have high scores on cash-flow management, savings, and investment, although the largest effect for these learning experience variables occurred with savings. The authors asked about learning preferences. Households with more financial education needs preferred learning by video. They are less likely to prefer Internet [delivery] and seminars. Households with less financial education needs prefer learning on their own. They like media, brochures, and videos. They are less likely to attend seminars.

The authors conclude that first, financial knowledge and financial learning experiences are significant. One probable way to increase financial knowledge is to gain financial education, although the authors acknowledge that causality may flow either way.

Second, the study concludes that the necessary scope and methods for delivering financial education vary with the recipient. There are various topics for financial education and various learning needs within each topic. Resources and delivery techniques need to be matched with learners. There is no "box of curricula that you can just pick up and use."

Concluding her remarks, Hogarth emphasized the importance of placing financial education in the highest context: "I think it is worth reiterating that financially secure families are, in fact, an appropriate goal for policymakers, community educators, practitioners, and researchers. Besides contributing to [an] effective, efficient market place...stable, secure families contribute to a stable, secure community, state, and nation."

Notes
¹ "The differentials we see in the marginal effects come from the mix of counseling administered by each provider, not from statistically significant differences in administering any given type of counseling" (A. Hirad and P. Zorn, 2003, "A Little Knowledge is a Good Thing: Empirical Evidence of the Effectiveness of Pre-Purchase Homeownership Counseling" page 15).

²A Debt Management Plan is an individualized financial consolidation program through sponsoring agencies (either for-profit or nonprofit) established to pay down debt and repair credit histories.

³The member organizations were Consumer Credit Counseling Services offices in the cities of Atlanta, San Francisco, Phoenix, Dallas, and Farmington Hills, Michigan.

⁴Similar borrowers were identified as those borrowers who had credit reports in both 1997 and 2000, resided in the same three-digit zip codes as the counseled group, whose names did not appear on the list of clients counseled by the participating counseling agencies, and whose Empirica score (see endnote v) value fell within the same range as observed in the group of counseling recipients.

⁵Each client signed a release consenting to use of his or her depersonalized data for research purposes. The credit bureau matched client names to credit reports and stripped all personally identifiable information prior to releasing data to the research team.

⁶The Empirica risk score is a proprietary score used by Trans Union as an overall measure of a borrower's predicted likelihood of delinquency or default (similar to the Fair Issac and Company scores). A lower Empirica score corresponds to a higher predicted frequency of charge off/repossession/bankruptcy; a higher Empirica score corresponds to a lower predicted frequency of charge-off/repossession/bankruptcy over the following 24 months.

⁷Borrowers who choose to seek counseling help may be more motivated to take corrective steps to improve their credit profile.