Financial adviser users and financial literacy Bhanu Balasubramniana, Eric R. Briskera,* aCollege of Business Administration, The University of Akron, Akron, OH 44325, USA Abstract Using the 2012 National Financial Capability Study we determine what demographic character- istics are associated with individuals that use financial advisers and whether financial advisers have any impact on the financial literacy of their clients. We consider five types of financial advisers: Debt Counselors, Savings or Investment, Mortgage or Loan, Insurance, and Tax Planning. We find a significant increase in the use of financial advisers over the past decade. We also find that Savings or Investments advisers have the largest positive impact on the financial literacy of their clients, followed by Mortgage or Loan and Insurance advisers, even when controlling for financial education and potential endogeneity issues. © 2016 Academy of Financial Services. All rights reserved. JEL classification: D12; D14; D81 Keywords: Personal financial planning; Financial advisers; Financial planners; Financial literacy; Instrumental variable models; Propensity score matching 1. Introduction We can argue that financial clients follow the advice from their financial advisers without learning from their interactions or financial experience. On the other hand, it is possible that financial clients do learn from their financial experience and interactions with financial advisers. Thus, financial advisers may play a significant role in disseminating financial knowledge to their clients that positively impacts their financial literacy. In this article, using the survey data from the 2012 National Financial Capability Study (NFCS), conducted by the Financial Industry Regulatory Authority (FINRA) Investor Education Foundation, we first identify demographic characteristics that are associated with individuals that use financial * Corresponding author. Tel.: �1-850-766-8315; fax: �1-330-972-5970. E-mail address: ebrisker@uakron.edu (E. R. Brisker) Financial Services Review 25 (2016) 127–155 1057-0810/16/$ – see front matter © 2016 Academy of Financial Services. All rights reserved. advisers in the United States. We then determine if financial advisers have any impact on the financial literacy of their clients. Over the past few decades financial products have become more complex to the point that making financial decisions has become very challenging for most people. In addition, retirement planning decisions have become more difficult for many people because of a shift in employee retirement plans away from defined benefit plans and towards defined contri- bution plans where individuals must make their own investment allocation decisions. For example, in recent years there has been a large increase in the number of investment products available that are typically used for retirement savings such as thousands of new mutual funds, exchange traded funds, target date (life-cycle) funds, and international equity funds. There has also been an increase in the number of non-investment products such as mortgage loan products (i.e., adjustable rate mortgages, reverse mortgages, and interest-only mort- gages), insurance products (i.e., universal life insurance, annuities), and personal health insurance products. As these more complex financial products become available investors are more likely to make non-optimal financial decisions simply because they do not understand the financial products they are consuming or how to manage risks. As financial products have grown in complexity the importance of financial literacy in our society has also grown. Several studies have considered how financial literacy impact financial decision making and the well-being of individuals. For example, Campbell (2006), in a study of household finance, finds that poorer and less educated households having lower financial literacy levels are more likely to make investment mistakes when making financial decisions, and the presence of these types of households may inhibit complex financial innovation. Klapper, Lusardi, and Panos (2013) find that financial literacy is positively related to participating in financial markets and negatively related to the use of informal sources of borrowing, and that individuals with higher levels of financial literacy and unspent income are better able to deal with macroeconomic shocks. Other research has shown that financial literacy leads to higher portfolio returns (Calvet, Campbell, and Sodini, 2009), greater wealth and probability of investing in stocks (Van Rooij, Lusardi, and Alessie, 2011), better retirement planning (Lusardi and Mitchell, 2007a, 2007b), and lower cost of borrow- ing (Huston, 2012). On the other hand, lower financial literacy levels lead to higher mortgage costs (Moore, 2003), excessive financial burden because of debt (Gathergood, 2012), and other debt related problems (Lusardi and Tufano, 2009). In summary, all prior literature agree that improving financial literacy can lead to better outcomes for both individuals, households, and the entire economic system. To compensate for lower levels of financial literacy, individuals have the option to seek out expert financial advice from professional advisers who will assist them in making complex financial decisions. For example, Winchester and Huston (2014), using the Theory of Planned Behavior, find that low control beliefs are significantly associated with lower financial-goal progress, however, receiving expert financial advice can help reduce this negative effect and actually result in higher levels of goal progress compared with individ- uals having high control beliefs that receive no expert financial advice. However, we cannot assume that everyone using financial advisers have lower levels of financial literacy. It is certainly possible that many individuals that use financial advisers already have high levels 128 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 of education and financial literacy, but are not able to stay informed about new financial products or strategies on their own. Black, Ciccotello, and Skipper (2002) propose two personal financial advising delivery models that are based on portfolio theory. In the first model, known as the “specialist model,” the financial consumer, or household, works directly with multiple types of financial advisers such as a debt counselor, investment adviser, mortgage or loan counselor, insurance agent, and tax planner. In the second model, known as the “planner model,” the financial consumer, or household, works directly with a single financial adviser, and this financial adviser acts as the intermediary between the consumer and the relevant financial adviser specialists. The consumer’s choice between these two models depends on the amount and type of financial advice they are seeking. As both of these models use financial advisers, it is important to understand not only who uses financial advisers, but also what specific types of financial advisers they are using. In 2012, FINRA Investor Education Foundation conducted a National Financial Capabil- ity Study. This national survey collected data on demographic characteristics of each survey participant regarding gender, age, ethnicity, education level, marital status, income level, employment status, and the region they live in within the continental United States. The survey also asks if the individual has ever received a financial education from high school, college, an employer, or from the military, and if the respondent has used a specific type of financial adviser. The five specific types of financial advisers included were Debt Counsel- ing, Savings or Investments, Mortgage or Loan, Insurance, and Tax Planning. The survey also asks five questions designed by Annamaria Lusardi and Olivia S. Mitchell to assess the financial literacy of the survey respondents (see Table 1 and Appendix A for a list of these questions). In determining who uses financial advisers, we find that there has been a significant increase in the number of individuals who use financial advisers over the last 14 years. Elmerick, Montalto, and Fox (2002), using the 1998 Survey of Consumer Finances, found that only 21.20% of those surveyed claimed to have used a financial adviser. Using the more recent 2012 FINRA Survey, we find that over 53% of those surveyed claimed they used a financial adviser indicating the usage of financial advisers has more than doubled since the 1998 survey. Overall, we find that those who have received a financial education, are female, have higher education and income levels, and those that are self-employed are more likely to use a financial adviser. Those in younger age groups and non-married individuals are less likely to use financial advisers. When considering the specific financial adviser types separately, we find those having received a financial education and those self-employed are more likely to use all of the different adviser types. Females are more likely to use Insurance advisers, Black (non- Hispanic) individuals are more (less) likely to use Debt Counseling (Mortgage or Loan) advisers, and younger individuals are more likely to use Debt Counseling and Mortgage or Loan advisers. Individuals having higher education levels are more likely to use Savings or Investments and Mortgage or Loan advisers, those that are married are more likely to use Mortgage or Loan, Insurance, and Tax Planning advisers, and those with higher income levels are more likely to use all of the different types of advisers but Debt Counselors. 129B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Table 1 Financial literacy questions Panel A: Financial literacy questions Q1 Time value of money Suppose you had $100 in a savings account and the interest rate was 2% per year. After 5 years, how much do you think you would have in the account if you left the money to grow? Q2 Inflation Imaging that the interest rate on your savings account was 1% per year and inflation was 2% per year. After 1 year, how much would you be able to buy with the money in this account? Q3 Interest rates If interest rates rise, what will typically happen to bond prices? Q4 Interest rates A 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less. Q5 Diversification Buying a single company’s stock usually provides a safer return than a stock mutual fund. Panel B: Percentage Answering Financial Literacy Questions Correctly Q1 Q2 Q3 Q4 Q5 Overall 78.6% 66.1% 30.9% 79.4% 53.5% Financial Adviser User 82.1 71.4 36.3 84.9 61.7 Non-user 74.7 60.1 24.8 73.0 44.1 t stat 13.28*** 17.77*** 18.77*** 21.88*** 26.59*** Debt Counseling User 73.0 56.3 28.2 80.0 48.0 Non-user 79.2 67.1 31.2 79.3 54.0 t stat �5.91*** �9.18*** �2.82*** 0.74 �5.12*** Savings or Investments User 83.8 74.0 40.5 86.1 67.5 Non-user 76.2 62.5 26.5 76.2 47.0 t stat 13.61*** 17.64*** 20.36*** 18.24*** 29.78*** Mortgage or Loan User 82.6 71.0 36.5 88.6 63.2 Non-user 77.5 64.8 29.4 76.7 50.7 t stat 8.15*** 8.36*** 9.25*** 21.45*** 15.87*** Insurance User 81.6 70.5 35.4 85.1 60.7 Non-user 77.2 64.0 28.8 76.5 50.0 t stat 7.69*** 9.77*** 9.84*** 15.76*** 15.27*** Tax Planning User 82.8 71.7 41.5 85.9 66.2 Non-user 77.6 64.8 28.3 77.7 50.3 t stat 8.04*** 9.07*** 16.28*** 13.35*** 19.68*** Panel A identifies the five financial literacy questions asked in the survey. Panel B reports the percentage of individuals who answered each question correctly for financial adviser users vs. non-users. Results are reported for the full sample (Overall), for those financial adviser users who used at least one or more of the five types of financial advisers vs. non-users (Financial Adviser), and for each financial adviser type user vs. non-users, separately. All values are in percentage. t stats are reported indicating the significance in the difference between users vs. non-users. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 130 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 With a large increase in the use of financial advisers in the United States raises the question if financial clients learn from their interactions with financial advisers. Bucher- Koenen and Lusardi (2011) declare that “those exposed to financially knowledgeable people become more financially knowledgeable themselves.” Kolb (1984), in the education litera- ture known as the “Experiential Learning Theory,” includes active experimentation, educa- tion, observation, and experience as essential components in the general model of learning. Bandura (1977) identifies observational learning as the process of social interaction impact- ing the knowledge of an individual. Therefore, individuals who use the services of a financial adviser are likely to learn from their interactions, discussions, experience, and financial outcomes. There is similar literature that shows that individual’s financial literacy improves when their parents have higher levels of education (Lusardi, Mitchell, and Curto, 2010), when they have financial experience (Johnson and Sherraden, 2007), and when they live in a zip code with higher average education levels (Lachance, 2014). Tang and Peter (2015) also show that financial education, financial experience, and parent’s financial experience exert positive influence on young adult’s financial knowledge. However, there are no studies that examine whether financial advisers have any impact on the financial literacy of their clients. We find that those who use financial advisers answer more of the financial literacy questions correctly compared to those who do not use financial advisers, even when controlling for demographic characteristics that are expected to impact financial literacy. We also control for previous financial education received that has already been shown to be positively related to financial literacy by Grimes, Rogers, and Smith (2010), and Walstad, Rebeck, and MacDonald (2010). When considering the different financial adviser types separately, we find that Savings or Investment advisers have the largest positive impact on financial literacy, followed by Mortgage or Loan and Insurance advisers, respectively. Similar to results reported by Robb, Babiarz, and Woodyard (2012) using the 2009 NFCS survey, we also find a negative relationship between using Debt Counseling1 advisers and financial literacy scores. It is possible that people having lower financial literacy get into debt and foreclosure problems and because they are poor learners, they do not learn from their experience or interaction with Debt Counselors. It is also possible that Debt Counseling services do not offer opportunities to expand their client’s financial literacy in the same manner that Savings and Investments advisers can since they advise in multiple areas pertaining to personal financial planning. We also find Tax Planning advisers do not have a significant impact on the financial literacy of their clients as they specialize in tax advising that is not necessarily related to financial literacy topics covered in the questions asked. Overall, our results indicate that financial advisers are serving more financial consumers in the United States in recent years, and a positive relationship exists between using certain types of financial advisers and higher financial literacy levels even when controlling for demographic characteristics, such as financial education and education levels, that are known to impact financial literacy. Robb, Babiarz, and Woodyard, (2012) and Allgood and Walstad (2016) find financial literacy influences the decision to use a financial adviser. Allgood and Walstad (2016) acknowledge that this relationship may reflect reverse causality, but they largely rule this possibility out without actually testing for reverse causality. In this study, we examine if 131B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 financial advisers actually do influence the financial literacy of their clients, and up to this point we have attempted to control for demographic characteristics that impact financial literacy, such as age, financial education, and education, and we still find a statistically significant positive relationship between financial literacy and using financial advisers. However, we have not directly dealt with the potential endogeneity issues that may exist between the use of financial advisers and financial literacy. We address this potential endogeneity issue by using instrumental variable (IV) Gener- alized Method of Moments (GMM) and IV Probit models where various instruments are used that are correlated with using a specific type of financial adviser but uncorrelated with financial literacy. To correct for the possibility of a confounding variables problem in our base-line model, we also conduct propensity score matching tests where we match financial adviser users with non-users having similar demographic characteristics and then determine how financial advisers impact financial literacy using this matched sample. We continue to find that Savings or Investments, Mortgage or Loan, and Insurance advisers have a positive influence on financial literacy scores, but Tax Planning advisers have no significant impact on financial literacy scores. However, we find that Debt Counseling advisers have a negative influence on financial literacy scores when using IV GMM estimation, but they have no significant impact on financial literacy scores when using IV Probit estimation. Overall, our results remain robust even when controlling for endogeneity and confounding variables issues indicating a significant positive relationship between financial literacy and using Savings or Investments, Mortgage or Loan, and Insurance advisers. We contribute to the individual/household financial literacy literature by providing evi- dence that specific types of financial advisers have a positive influence on financial literacy. This suggests clients learn from their financial experience and interactions with specific types of financial advisers rather than just take advice from them. To the best of our knowledge, this is the first article to consider the impact financial advisers have on the financial literacy of their clients. The remainder of the article is organized as follows. Section 2 provides details about the FINRA NFCS 2012 survey, and our multivariate analysis empirical results are reported in Section 3. Section 4 presents our IV models and propensity score matching results, and we conclude in Section 5. 2. Data and descriptive analysis 2.1. FINRA 2012 National financial capability study The FINRA Investor Education Foundation conducted the State-by-State Financial Ca- pability Survey that was developed in consultation with the U.S. Department of Treasury, the President’s Advisory Council on Financial Literacy, and multidisciplinary team including academics and practitioners. The data were collected through an online survey of 25,509 adults (18�) across the United States, with approximately 500 respondents per state plus the District of Columbia. The survey was self-administered by respondents on a website from 132 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 July to October of 2012 and provides an unprecedented set of information on financial behaviors across the United States. Survey respondents were asked to identify several demographic characteristics such as their gender, age, ethnicity, education level, marital status, income level, employment status, region of the United States they lived in, and if they have ever received a financial education in the past. They were also asked if they have used a Debt Counseling, Savings and Investments, Mortgage or Loan, Insurance, or Tax Planning adviser within the past five years. Available responses included YES, NO, Do Not Know, and Prefer not to say. After limiting our sample to respondents who answered either YES or NO for each of the financial adviser type use questions we are left with 22,218 out of the 25,509 survey respondents for our analysis. Out of these 22,218 survey respondents, 53.4% used at least one of the five types of financial advisers compared with 46.6% that had not. The survey respondents are also asked five basic finance questions designed by Annamaria Lusardi and Olivia S. Mitchell to assess their financial literacy. The five questions asked are designed to test the individuals understanding of basic concepts related to the time value of money, interest rates, inflation, and risk diversification. These questions are widely used as a measure of financial literacy in the literature. We will use the survey respondent’s answers to these five questions to measure their financial literacy level. 2.2. Descriptive analysis In Table 2 we report the percentage of each demographic characteristic group that used at least one type of financial adviser in the FA User column, and for each specific type of financial adviser in the Debt Counselor, Savings or Investments, Mortgage of Loan, Insur- ance, and Tax Planning columns. The demographic characteristics considered are Financial Education, Gender, Age Group, Ethnicity, Education, Marital Status, Income, Employment, and Region. These demographic characteristics have been shown to impact personal financial behavior and the probability of using a financial adviser when making financial decisions (Barber and Odean, 2001; Elmerick, Montalto, and Fox, 2002). We report t-stats indicating the significance in the difference between the two identified groups within each demographic characteristic. We find a large variation in the percentage of individuals using at least one type of financial adviser based on their Education and Income levels. There is a much larger percentage of individuals using financial advisers who are either a college graduate (62.9%) or have post graduate college (69.7%) compared with those who have either completed High School (44%) or less than High School (30.4%). Not surprisingly, we also find that that there is a monotonic increase in the percentage of financial adviser users across higher income levels. For instance, 75.1% of those earning more than $150,000 used a financial adviser compared with only 40.3% of those earning $15,000 to $24,999. When considering the specific finance adviser types, we find that a larger percentage of individuals falling in lower income categories and younger age groups reported using Debt Counselors. We find a larger percentage of individuals in lower age groups and higher income levels used Mortgage or Loan advisers. There are also a larger percentage of individuals falling in higher age groups, higher education levels, and higher income levels 133B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Table 2 Percent of individuals that use financial advisers FA User Debt Counseling Savings or Investments Mortgage or Loan Insurance Tax Planning Total % Using 53.4% 8.8% 31.6% 22.1% 32.9% 19.9% Financial Education Taken 65.6 11.5 42.4 29.5 42.9 27.9 Not taken 49.8 8.0 28.4 19.9 29.9 17.5 Taken - Not taken t test 20.70*** 7.23*** 18.15*** 13.58*** 16.70*** 15.01*** Gender Male 54.3 8.6 33.8 23.3 33.2 21.5 Female 52.7 8.9 29.9 21.2 32.7 18.6 Male - Female t test 2.29** �0.72 6.16*** 3.81*** 0.71 5.38*** Age group 18–24 43.7 9.0 26.1 17.2 26.5 16.3 25–34 52.6 13.9 29.3 28.5 34.9 20.4 35–44 52.4 11.5 26.1 27.0 35.1 18.2 45–54 50.0 8.1 27.1 20.9 32.7 17.8 55–64 55.4 7.1 34.9 19.8 33.3 20.1 65� 62.3 3.9 43.7 18.2 32.3 25.4 (18–24) - (65�) t test �13.90*** 7.28*** �14.10*** �0.99 �4.77*** �8.55*** Ethnicity White 54.2 7.5 32.6 22.4 32.6 20.2 Black 51.4 12.4 29.0 21.2 33.8 19.0 White - Black t test 3.59*** �10.41*** 5.15*** 1.96** �1.66* 2.05*** Education � High school 30.4 6.6 10.6 8.9 18.9 7.8 High school 44.0 7.2 23.1 16.0 25.2 13.3 GED 38.0 8.3 16.2 14.5 24.6 11.0 Some College 54.1 8.7 30.8 22.3 34.3 18.7 College graduate 62.9 11.6 40.6 28.5 39.5 25.3 Post-graduate 69.7 8.1 50.1 30.8 41.4 34.4 �HS–Post-grad t test �27.33*** �1.92* �32.70*** �19.71*** �16.76*** �24.07*** Marital status Married 59.6 8.6 36.2 26.5 36.7 24.0 Single 43.9 9.2 26.2 16.7 26.7 14.4 Separated 39.4 10.6 18.0 11.9 28.4 14.2 Divorced 47.3 8.7 23.9 16.8 29.4 13.9 Widowed 52.0 7.6 30.3 14.5 30.7 17.3 Married - Widowed t test 4.45*** 1.13 3.70*** 9.72*** 3.73*** 5.10*** Income �$15,000 27.4 6.2 11.5 7.9 16.5 6.4 $15K-$24,999 40.3 10.1 16.8 11.8 26.3 9.5 $25K-$34,999 45.7 9.5 22.4 16.7 28.6 13.1 $35K-$49,999 51.8 11.1 26.7 19.7 32.6 16.3 $50K-$74,999 59.2 9.1 35.1 25.6 36.8 21.3 $75K-$99,999 64.2 9.1 43.1 30.4 38.6 26.5 $100K-$149,000 70.3 7.0 50.0 32.8 42.0 32.8 � $150,000 75.1 6.8 57.6 36.8 45.3 43.4 �$15K - �$150K t test �34.09*** �0.87 �33.07*** �21.83*** �19.92*** �27.55*** Employment Self employed 60.1 9.2 36.3 25.7 41.3 28.8 Full-time 59.2 11.0 35.0 29.4 37.5 21.6 Part-time 51.4 9.6 31.8 20.2 32.3 20.0 Homemaker 45.4 8.0 21.3 19.9 28.7 15.9 Student 43.3 9.4 27.8 15.5 25.4 13.9 134 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 that use Savings and Investments advisers. In fact, there is a monotonic increase in the percentage of individuals using a Savings and Investments, Mortgage or Loan, Insurance, or Tax Planning adviser across both higher education and income levels. Individuals using Insurance or Tax Planning financial advisers are mixed across the other demographic groups, and Gender, Ethnicity, Marital Status, Employment, and Region reveal mixed results for each financial adviser type. In Table 1, Panel A, we identify the five financial literacy questions used in the survey, and in Table 1, Panel B, we report the percentage of individuals answering each of the five financial literacy questions correctly. The first row in Panel B, Overall, reports the percentage of survey respondents answering each question correctly for the entire sample, followed by user versus non-user comparisons for those using one or more of the different financial adviser types (Financial Advisor), and for each of the financial adviser types, separately. We also report t-stats indicating the significance in the difference between the percentages of financial adviser(s) users versus non-users who answered each question correctly. Overall, we find that 78.6% of individuals answered the question regarding the Time Value of Money correctly, but only 30.9% answered the question pertaining to the relation- ship between interest rates and bond prices correctly. Just over 66% answered the question pertaining to Inflation correctly, but nearly 80% answered the question regarding mortgage interest expenses correctly. Surprisingly, only 53.5% answered the question pertaining to diversification using mutual funds correctly. We find that a greater percentage of those using one or more of the five types of financial advisers, reported under Financial Adviser, answered all five questions cor- rectly compared with those not using any of the five types of advisers, and these differences are all highly significant at the 1% level. When considering the five types of financial advisers separately, we find that a higher percentage of individuals using Table 2 (Continued) FA User Debt Counseling Savings or Investments Mortgage or Loan Insurance Tax Planning Disabled 37.0 8.2 12.6 11.3 25.6 6.9 Unemployed 34.4 8.5 15.6 11.4 20.8 10.7 Retired 59.2 4.4 40.8 17.4 31.7 23.6 Self-employed - Retired t test 0.63 6.37*** -3.24*** 6.97*** 7.02*** 4.15*** Region Northeast 53.6 8.5 34.0 22.1 32.0 21.2 Midwest 53.6 7.7 31.6 21.5 33.8 19.8 South 50.4 9.2 29.5 19.9 31.0 18.7 West 57.1 9.4 32.8 25.5 35.2 20.7 Northeast - West t test �3.33*** -1.50 1.24 �3.79*** �3.28*** 0.62 The table shows the Percent of individuals surveyed that used financial advisers. The FA User column report results for people that used at least one or more of the five types of financial advisers. The specific financial adviser columns report the percentage of individuals using that specific type of adviser. All values are in percentages. t stats are reported indicating the significance in the difference between the two identified groups within each demographic characteristic. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 135B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Savings or Investments, Mortgage of Loan, Insurance, and Tax Planning advisers answered all five questions correctly compared to those not using those types of advisers (with the statistical significance of these differences being highly significant at the 1% level). However, we find that a lower percentage of those using Debt Counselors answered all of the questions correctly compared to those not using Debt Counselors except for the question regarding mortgage interest expenses (Q4). 3. Multivariate analysis 3.1. Financial adviser users To determine who is using financial advisers, we use multivariate OLS regressions to relate each demographics characteristic associated with individuals that used one or more of the five types of financial advisers while controlling for all other demographic characteristics. We also estimate a multinomial logit regression model to estimate the effect each demo- graphic characteristic has on the probability of using at least one of the five types of financial advisers while simultaneously controlling for the effects of all other demographic charac- teristics. The dependent variable in both models, Financial Adviser Users, is a dummy variable set equal to 1 if the survey respondent used at least one of the five types of financial advisers, or 0 otherwise. The demographic characteristics included are Financial Education, Female, Age Group, Black (Non-Hispanic), Education, Marital Status, Income, Employment, and Region. Financial Education is a dummy variable set equal to 1 if the respondent had any Financial Education training, or 0 otherwise. Female is a dummy variable set equal to 1 for females and 0 for males. Black (Non-Hispanic) is a dummy variable set equal to 1 if the person is Black (Non-Hispanic), or 0 otherwise. The categories within the Age Group, Education, Marital Status, Income, Employment, and Region demographic characteristic as identified in Table 3 are set equal to 1 if the respondent identifies themselves to be in that categorical group, or 0 otherwise. The first two columns in Table 3 report coefficient estimates and t-stats from OLS estimation. We find the OLS coefficient estimates for Financial Education and Female is positive and highly significant indicating people having received financial education in the past and females tend to use financial advisers. We also find there is a highly statistically significant monotonic increase in the use of financial advisers across higher Education and Income levels. Those in younger age groups and non-married people are less likely to use financial advisers. However, results are mixed for different Employment and Region cate- gories and not significant for Black (Non-Hispanic) indicating ethnicity does not appear related to using a financial adviser. The last three columns in Table 3 report Logit coefficient estimates, z-stats, and Odds Ratios from the Logit regression results. Positive (negative) coefficient estimates indicate that demographic characteristic is more (less) likely to use a financial adviser compared with the reference category identified in the table for each demographic characteristic. Consistent with the OLS results, we find that those having a Financial Education are 62% more likely, and Females 14% more likely, to use a financial adviser. We also find there is a monoton- 136 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Table 3 Financial adviser user OLS and Logit regressions Financial adviser users OLS coefficient t stat Logit coefficient z-stat Odds Ratio Constant 0.294*** 12.68 �0.921*** �8.47 0.398 Financial education 0.105*** 13.79 0.483*** 13.49 1.621 Female 0.030*** 4.52 0.135*** 4.44 1.144 Age group 25–34 �0.018 �1.25 �0.082 �1.28 0.921 35–44 �0.054*** �3.70 �0.243*** �3.71 0.784 45–54 �0.069*** �4.85 �0.309*** �4.83 0.734 55–64 �0.026* �1.77 �0.116* �1.74 0.891 65� 0.024 1.41 0.109 1.44 1.116 Black (Non-Hispanic) �0.008 �1.02 �0.035 �1.01 0.965 Education (reference category: Did not complete high school) High school 0.037*** 2.58 0.177*** 2.63 1.194 GED 0.005 0.26 0.035 0.43 1.036 Some college 0.100*** 7.21 0.443*** 6.76 1.558 College graduate 0.147*** 9.80 0.649*** 9.24 1.913 Post-graduate 0.159*** 9.77 0.717*** 9.25 2.049 Marital status (reference group: Married) Living-with-partner �0.022* �1.78 �0.099* �1.76 0.906 Single �0.043*** �5.34 �0.194*** �5.44 0.823 Income (reference category: Less than $15,000) $15,000 - $24,999 0.102*** 7.80 0.481*** 7.85 1.618 $25,000 - $34,999 0.134*** 9.73 0.610*** 9.71 1.840 $35,000 - $49,999 0.177*** 13.24 0.787*** 12.87 2.196 $50,000 - $74,999 0.230*** 17.35 1.004*** 16.45 2.729 $75,000 - $99,999 0.262*** 17.73 1.142*** 16.79 3.134 $100,000 - $149,999 0.305*** 20.04 1.349*** 18.76 3.853 $150,000 or more 0.339*** 20.01 1.532*** 18.26 4.628 Employment (reference category: Self employed) Employed full-time �0.038*** �3.05 �0.175*** �3.05 0.839 Employed part-time �0.039** �2.50 �0.176** �2.51 0.839 Homemaker �0.102*** �6.41 �0.453*** �6.32 0.636 Full-time student �0.097*** �4.41 �0.426*** �4.36 0.653 Disabled �0.066*** �3.63 �0.296*** �3.53 0.744 Unemployed �0.105*** �6.48 �0.481*** �6.38 0.618 Retired �0.036** �2.47 �0.169** �2.49 0.845 Region (reference category: Northeast) Midwest 0.021** 2.07 0.093** 2.07 1.098 South �0.002 �0.18 �0.007 �0.17 0.993 West 0.042*** 4.25 0.188*** 4.21 1.207 Observations 22,218 22,218 Adjusted R2 0.110 This table reports OLS and Logit regression results where the dependent variable, Financial Adviser Users, is set equal to 1 if the person used at least one of the five types of financial advisers, or 0 otherwise. The independent variables include Financial Education, Female, Age Group, Black (Non-Hispanic), Education, Marital Status, Income, Employment, and Region. Positive (negative) coefficient estimates indicate that demographic character- istic is more (less) likely to use a financial adviser compared to the indicated reference category. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 137B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 ically increasing likelihood of using a financial adviser across higher Education and Income levels, but Age Group is significantly related to a reduced probability of using a financial adviser for those in younger age groups. Compared with the univariate results presented in Table 2, this shows that it is income rather than age that increases the probability of an individual using a financial adviser. We also see that married people are more likely to use financial advisers than either single or “living-with-partner” couples, and Black (Non- Hispanic) does not impact the likelihood of using a financial adviser. When considering Employment, those that are self-employed are more likely to use financial advisers when compared with all other employment categories. Finally, we find that those from the Midwest (West) are 10% (21%) more likely to use a financial adviser compared with those from the South or Northeast. In Table 4 we estimate the same OLS and Logit models for each of the five financial adviser types separately. In this case, each financial adviser type to include Debt Counseling, Savings or Investments, Mortgage or Loan, Insurance, and Tax Planning is set equal to 1 if the person used that type of adviser, and zero otherwise. We use the same demographic characteristics and categories that were used in Table 3. In discussing the OLS estimation results (first two columns for each adviser type), we find that Financial Education is both positive and highly statistically significant for all financial adviser types indicating people having received financial education in the past are associated with using all of the five types of financial advisers. However, Females are only associated with using Insurance advisers with all other coefficient estimates for Female being not significant for other adviser types. We also find that Black (Non-Hispanic) is positively (negatively) related to using Debt Counseling (Mortgage or Loan) advisers, but not significantly related to using the other three financial adviser types. We find a monotonically decreasing relationship between Age Groups and Debt Counseling where lower (higher) Age Groups are positively (negatively) related to using Debt Counseling. We also find that the 65� age group is positively related to using Savings or Investment advisers, the 25 to 34 age group is positively related to using Mortgage or Loan advisers. For Education, people having a high school education or higher are associated with using Savings or Investment and Mortgage or Loan advisers, while only those having some college or higher are associated with using Insurance and Tax Planning advisers. For Marital Status, married couples are associated with using Mortgage or Loan, Insurance, and Tax Planning advisers. For Income, we find those in lower income levels are associated with using Debt Counseling while those in higher income levels are increas- ingly associated with using Savings or Investment, Mortgage or Loan, Insurance, and Tax Planning advisers. For Employment, all coefficient estimates are either significantly negative or not significant, indicating the reference category, Self Employed, is posi- tively related to using all of the financial adviser types compared with the other Employment groups. For Region, we find those in the West are associated with using Mortgage or Loan advisers, and those in the Midwest and West are associated with using Insurance advisers. The logit estimation results (last three columns for each adviser type) are consistent with the OLS results. Those receiving a Financial Education are from between 39% to 53% more likely to use all of the financial adviser types compared to those not receiving 138 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Table 4 Specific type of financial adviser user OLS and Logit regressions Debt Counseling Savings or Investments OLS coefficient t stat Logit coefficient z-stat Odds Ratio OLS coefficient t stat Logit coefficient z-stat Odds Ratio Constant 0.050*** 3.58 �2.901*** �15.27 0.0550 0.130*** 6.58 �2.181*** �16.15 0.113 Financial education 0.030*** 6.08 0.355*** 6.48 1.426 0.083*** 10.86 0.399*** 11.03 1.490 Female 0.004 0.92 0.034 0.67 1.034 0.003 0.47 0.019 0.59 1.019 Age group (reference category: 18–24) 25–34 0.038*** 4.11 0.367*** 3.70 1.444 �0.043*** �3.42 �0.279*** �3.82 0.757 35–44 0.016* 1.67 0.171 1.62 1.186 �0.106*** �8.36 �0.622*** �8.26 0.537 45–54 �0.014 �1.62 �0.166 �1.55 0.847 �0.088*** �7.09 �0.520*** �7.11 0.594 55–64 �0.019** �2.05 �0.225** �1.97 0.799 �0.030** �2.30 �0.210*** �2.82 0.811 65� �0.039*** �3.94 �0.675*** �4.52 0.509 0.028* 1.80 0.051 0.62 1.052 Black (Non-Hispanic) 0.030*** 5.94 0.325*** 6.04 1.384 �0.007 �1.04 �0.037 �0.96 0.964 Education (reference category: Did not complete high school) High school 0.005 0.68 0.080 0.64 1.083 0.038*** 3.65 0.451*** 4.79 1.570 GED 0.009 0.92 0.139 0.95 1.149 �0.000 �0.04 0.154 1.36 1.166 Some college 0.015* 1.95 0.218* 1.83 1.244 0.075*** 7.39 0.664*** 7.26 1.943 College graduate 0.032*** 3.59 0.397*** 3.17 1.488 0.142*** 12.10 0.979*** 10.37 2.663 Post-graduate 0.011 1.17 0.167 1.20 1.181 0.162*** 11.76 1.030*** 10.41 2.802 Marital status (reference category: Married) Living-with-partner �0.012 �1.53 �0.145 �1.56 0.865 0.004 0.39 0.012 0.19 1.012 Single �0.009* �1.78 �0.116* �1.92 0.890 0.001 0.09 0.010 0.26 1.010 Income (reference category: Less than $15,000) $15,000 - $24,999 0.042*** 5.42 0.577*** 5.39 1.780 0.039*** 3.94 0.366*** 4.45 1.443 $25,000 - $34,999 0.030*** 3.81 0.432*** 3.86 1.540 0.083*** 7.69 0.665*** 8.27 1.944 $35,000 - $49,999 0.045*** 5.64 0.588*** 5.41 1.801 0.114*** 10.56 0.842*** 10.93 2.320 $50,000 - $74,999 0.023*** 2.95 0.338*** 3.02 1.402 0.185*** 16.69 1.181*** 15.70 3.257 $75,000 - $99,999 0.018** 2.10 0.277** 2.24 1.319 0.252*** 19.31 1.473*** 18.30 4.361 $100,000 - $149,999 0.001 0.16 0.046 0.34 1.047 0.304*** 22.01 1.680*** 20.35 5.364 $150,000 or more �0.001 �0.05 0.016 0.11 1.016 0.364*** 22.28 1.923*** 21.13 6.840 Employment (reference category: Self employed) Employed full-time 0.007 0.84 0.064 0.70 1.066 �0.037*** �3.02 �0.183*** �3.06 0.832 Employed part-time 0.000 0.05 0.007 0.06 1.007 0.007 0.47 0.045 0.61 1.046 Homemaker �0.023** �2.40 �0.280** �2.31 0.756 �0.072*** �5.05 �0.417*** �5.25 0.659 Full-time student �0.012 �0.84 �0.121 �0.77 0.886 �0.033 �1.64 �0.143 �1.32 0.867 Disabled �0.000 �0.01 �0.001 �0.01 0.999 �0.077*** �5.15 �0.582*** �5.38 0.559 Unemployed �0.007 �0.73 �0.077 �0.61 0.926 �0.066*** �4.67 �0.436*** �4.89 0.647 Retired �0.018** �2.24 �0.321** �2.56 0.725 0.005 0.38 0.018 0.26 1.018 Region (reference category: Northeast) Midwest �0.011* �1.86 �0.142* �1.81 0.868 0.004 0.44 0.028 0.58 1.028 South 0.000 0.07 0.004 0.06 1.004 �0.010 �1.14 �0.056 �1.22 0.945 West �0.000 �0.01 0.005 0.07 1.005 0.004 0.47 0.028 0.58 1.028 Observations 22,218 22,218 22,218 22,218 Adjusted R2 0.023 0.127 Mortgage or Loan Insurance OLS coefficient t stat Logit coefficient z-stat Odds Ratio OLS coefficient t stat Logit coefficient z-stat Odds Ratio Constant 0.135*** 7.37 �2.215*** �14.87 0.109 0.207*** 9.53 �1.500*** �12.75 0.223 Financial education 0.060*** 8.40 0.336*** 8.69 1.399 0.097*** 12.26 0.431*** 12.52 1.539 Female 0.004 0.77 0.025 0.70 1.025 0.023*** 3.58 0.112*** 3.60 1.118 Age Group (reference category: 18–24) 25–34 0.032*** 2.76 0.117 1.51 1.125 0.011 0.87 0.044 0.66 1.045 35–44 �0.009 �0.80 �0.134* �1.68 0.874 �0.008 �0.57 �0.048 �0.70 0.953 45–54 �0.065*** �5.71 �0.458*** �5.73 0.633 �0.026** �1.97 �0.132* �1.94 0.876 55–64 �0.076*** �6.39 �0.533*** �6.36 0.587 �0.022 �1.58 �0.114 �1.62 0.892 65� �0.092*** �6.62 �0.635*** �6.58 0.530 �0.037** �2.27 �0.184** �2.28 0.832 Black (Non-Hispanic) �0.026*** �3.99 �0.163*** �3.95 0.850 0.011 1.50 0.053 1.49 1.054 Education (reference category: Did not complete high school) High school 0.030*** 3.18 0.353*** 3.40 1.423 0.006 0.51 0.079 1.02 1.082 GED 0.022* 1.83 0.282** 2.31 1.325 0.010 0.65 0.093 1.01 1.098 Some college 0.068*** 7.19 0.616*** 6.17 1.851 0.071*** 5.84 0.406*** 5.48 1.501 139B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Table 4 (Continued) Mortgage or Loan Insurance OLS coefficient t stat Logit coefficient z-stat Odds Ratio OLS coefficient t stat Logit coefficient z-stat Odds Ratio College graduate 0.086*** 7.89 0.700*** 6.78 2.015 0.088*** 6.49 0.472*** 6.05 1.603 Post-graduate 0.087*** 6.81 0.706*** 6.51 2.026 0.087*** 5.67 0.465*** 5.55 1.591 Marital status (reference category: Married) Living-with-partner �0.022** �2.04 �0.123* �1.84 0.885 �0.026** �2.13 �0.120** �2.03 0.887 Single �0.054*** �8.20 �0.357*** �8.02 0.700 �0.040*** �5.10 �0.188*** �5.01 0.829 Income (reference category: Less than $15,000) $15,000 - $24,999 0.023*** 2.76 0.336*** 3.50 1.400 0.082*** 7.07 0.517*** 7.32 1.677 $25,000 - $34,999 0.051*** 5.34 0.612*** 6.57 1.844 0.089*** 7.28 0.556*** 7.75 1.744 $35,000 - $49,999 0.068*** 7.09 0.732*** 8.14 2.080 0.118*** 9.78 0.695*** 10.02 2.004 $50,000 - $74,999 0.113*** 11.52 0.997*** 11.40 2.709 0.146*** 12.13 0.819*** 11.98 2.268 $75,000 - $99,999 0.146*** 12.40 1.151*** 12.42 3.163 0.150*** 10.89 0.834*** 11.20 2.303 $100,000 - $149,999 0.168*** 13.33 1.252*** 13.20 3.499 0.175*** 12.00 0.938*** 12.20 2.556 $150,000 or more 0.202*** 13.33 1.404*** 13.79 4.071 0.201*** 11.82 1.043*** 12.30 2.836 Employment (reference category: Self-employed) Employed full-time 0.005 0.40 �0.020 �0.32 0.980 �0.064*** �4.98 �0.291*** �5.23 0.747 Employed part-time �0.025* �1.85 �0.141* �1.75 0.868 �0.062*** �4.01 �0.271*** �3.88 0.763 Homemaker �0.052*** �3.76 �0.319*** �3.86 0.727 �0.111*** �7.09 �0.511*** �7.02 0.600 Full-time student �0.079*** �4.51 �0.491*** �4.01 0.612 �0.126*** �6.20 �0.581*** �5.65 0.559 Disabled �0.040*** �2.80 �0.328*** �2.88 0.720 �0.060*** �3.42 �0.261*** �2.97 0.770 Unemployed �0.064*** �4.87 �0.485*** �4.92 0.615 �0.118*** �7.63 �0.592*** �7.40 0.553 Retired �0.037*** �2.84 �0.254*** �3.21 0.776 �0.073*** �4.87 �0.331*** �4.91 0.718 Region (reference category: Northeast) Midwest 0.005 0.64 0.029 0.54 1.029 0.027*** 2.82 0.130*** 2.81 1.139 South �0.002 �0.22 �0.012 �0.24 0.988 0.003 0.31 0.013 0.29 1.013 West 0.037*** 4.35 0.223*** 4.32 1.250 0.028*** 2.93 0.133*** 2.92 1.143 Observations 22,218 22,218 22,218 22,218 Adjusted R2 0.074 0.051 Tax Planning OLS coefficient t stat Logit coefficient z-stat Odds Ratio Constant 0.166*** 9.28 �2.110*** �13.34 0.121 Financial education 0.063*** 9.11 0.379*** 9.46 1.461 Female �0.005 �0.83 �0.028 �0.75 0.972 Age group (reference category: 18–24) 25–34 �0.024** �2.15 �0.250*** �2.93 0.779 35–44 �0.073*** �6.44 �0.595*** �6.76 0.551 45–54 �0.074*** �6.64 �0.599*** �6.95 0.549 55–64 �0.067*** �5.76 �0.552*** �6.27 0.576 65� �0.036*** �2.61 �0.371*** �3.84 0.690 Black (Non-Hispanic) 0.005 0.80 0.036 0.83 1.037 Education (reference category: Did not complete high school) High school 0.007 0.83 0.193* 1.76 1.212 GED �0.002 �0.20 0.073 0.56 1.076 Some college 0.035*** 4.05 0.431*** 4.10 1.539 College graduate 0.069*** 6.79 0.637*** 5.87 1.891 Post-graduate 0.111*** 8.94 0.815*** 7.24 2.259 Marital status (reference category: Married) Living-with-partner �0.038*** �4.03 �0.300*** �3.89 0.740 Single �0.020*** �3.19 �0.139*** �3.02 0.870 Income (reference category: Less than $15,000) $15,000 - $24,999 0.026*** 3.31 0.382*** 3.63 1.465 $25,000 - $34,999 0.055*** 6.17 0.688*** 6.74 1.990 $35,000 - $49,999 0.080*** 8.84 0.898*** 9.21 2.455 140 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 a financial education. Females are 12% more likely to use Insurance advisers, but we find no other significant difference related to gender in using the other financial adviser types. Those that are Black (Non-Hispanic) are 38% more likely to use Debt Counseling advisers, but 15% less likely to use Mortgage or Loan advisers. We find a monotonically decreasing probability of using Debt Counseling and Mortgage or Loan advisers across higher Age Groups, but all other results related to Age Group are mixed. We find a monotonically increasing probability of using Savings or Investments and Mortgage or Loan advisers across higher Education levels, while those having some college or higher are far more likely to use Insurance and Tax Planning advisers. For Marital Status, we find that those that are married are 12% to 30% more likely to use Mortgage or Loan advisers, 11% to 17% more likely to use Insurance advisers, and 13% to 26% more likely to use Tax Planning advisers compared with those that are single or “living-with- partner.” We find a nearly monotonic decreasing probability of using Debt Counseling advisers across higher income levels, but a monotonically increasing probability of using Savings or Investments, Mortgage or Loan, Insurance, and Tax Planning advisers across higher income levels. For Employment, those that are Self-Employed are more likely to use any of the financial adviser types compared with the other Employment groups. For Table 4 (Continued) Tax Planning OLS coefficient t stat Logit coefficient z-stat Odds Ratio $50,000 - $74,999 0.120*** 12.90 1.164*** 12.30 3.202 $75,000 - $99,999 0.163*** 14.39 1.401*** 14.04 4.060 $100,000 - $149,999 0.214*** 17.35 1.650*** 16.26 5.205 $150,000 or more 0.307*** 20.13 2.041*** 19.10 7.698 Employment (reference category: Self-employed) Employed full-time �0.095*** �8.41 �0.583*** �9.15 0.558 Employed part-time �0.049*** �3.65 �0.264*** �3.25 0.768 Homemaker �0.084*** �6.28 �0.528*** �6.10 0.590 Full-time student �0.121*** �7.00 �0.798*** �5.99 0.450 Disabled �0.098*** �7.55 �0.871*** �6.59 0.419 Unemployed �0.084*** �6.51 �0.574*** �5.73 0.563 Retired �0.049*** �3.70 �0.277*** �3.67 0.758 Region (reference category: Northeast) Midwest 0.007 0.85 0.051 0.93 1.053 South �0.003 �0.39 �0.023 �0.44 0.977 West 0.004 0.47 0.031 0.58 1.032 Observations 22,218 22,218 Adjusted R2 0.088 This table reports OLS and Logit regression results where the dependent variable, Debt Counseling, Savings or Investments, Mortgage or Loan, Insurance, or Tax Planning is set equal to 1 if the person used that type of financial adviser, or 0 otherwise. The independent variables include Financial Education, Female, Age Group, Black (Non-Hispanic), Education, Marital Status, Income, Employment, and Region. Positive (negative) coeffi- cient estimates indicate that demographic characteristic is more (less) likely to use a financial adviser compared with the indicated reference category. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 141B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Region, those in the West are 25% more likely to use Mortgage or Loan advisers, and those in the Midwest and West are 14% more likely to use Insurance advisers. 3.2. Financial literacy We will now determine if financial advisers have an impact on the financial literacy of individuals that use them versus those that do not. We first estimate a linear OLS regression model to determine if using at least one of the five types of financial advisers impacts the number of correct answers giving for the five financial literacy questions asked in the survey. We then estimate an Ordered Logit regression model to determine if using at least one of the five types of financial advisers impacts the probability of answering more of the financial literacy questions correctly. The dependent variable in both models, Financial Literacy Score, is set equal to the number of questions the survey respondent answered correctly ranging from 0 to 5. The independent variable of interest, FA User, is a dummy variable set equal to one if the survey respondent used at least one of the five types of financial advisers in the past five years, or 0 otherwise. We then re-estimate both models where we replace the FA User independent dummy variable with separate dummy variables for each financial adviser type to include Debt Counseling, Savings or Investments, Mortgage or Loan, Insurance, and Tax Planning that are set equal to 1 if the survey respondent used that type of financial adviser, or 0 otherwise. It is important for us to include the demographic characteristics Financial Education, Female, Age Group, Black (Non-Hispanic), Education, Marital Status, Income, Employment, and Region as control variables in our models since many of them are expected to influence financial literacy. The first two columns in the left panel of Table 5 report the coefficient and t-stat results from the OLS model specification using FA User as the main independent variable of interest. We first notice that Financial Education is positively related to higher Financial Literacy Scores as expected, but Female and Black (Non-Hispanic) are significantly related with lower Financial Literacy Scores. We also find a monotonically increasing positive relationship with higher Financial Literacy Scores across higher Age Groups, Education levels, and Income levels. Both Married and Self-Employed people scored higher Financial Literacy Scores compared with the other categories under Marital Status and Employment, and those in the Northeast scored lower Financial Literacy Scores compared with the other categories under Region. Our main variable of interest, FA User, is positive, and highly statistically significant, indicating that those using at least one of the five types of financial advisers answered more of the financial literacy questions correctly compared with those that did not use any financial advisers. This is our first evidence that financial advisers positively influence the financial literacy of their clients, even after controlling for other demographic characteristics that are also positively related to higher financial literacy scores. The last three columns of the left Panel in Table 5 report the Ordered Logit regression model results. In this model specification, a positive (negative) coefficient estimate indicates that characteristic increases (decreases) the probability of answering more of the financial literacy questions correctly. Consistent with the OLS results, we find that Financial Educa- tion increases the probability of answering more financial literacy questions correctly by 142 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Table 5 Financial Literacy OLS and Ordered Logit regressions Financial Literacy Score: 0 to 5 Financial Literacy Score: 0 to 5 OLS coefficient t stat O-Logit coefficient z-stat Odds Ratio OLS coefficient t stat O-Logit coefficient z-stat Odds Ratio Constant 1.639*** 27.63 3.986*** 42.39 53.833*** 1.667*** 28.16 3.952*** 42.08 52.044*** FA User 0.202*** 11.88 0.289*** 11.26 1.335*** Debt Counseling �0.277*** �9.38 �0.432*** �9.67 0.649*** Savings or Investments 0.154*** 7.30 0.247*** 7.73 1.280*** Mortgage or Loan 0.109*** 5.05 0.150*** 4.59 1.161*** Insurance 0.063*** 3.12 0.075** 2.50 1.078** Tax planning �0.016 �0.68 �0.023 �0.64 0.978 Financial education 0.265*** 13.47 0.414*** 13.76 1.512*** 0.271*** 13.75 0.425*** 14.09 1.529*** Female �0.455*** �26.83 �0.728*** �27.89 0.483*** �0.451*** �26.59 �0.721*** �27.62 0.486*** Age group (reference category: 18–24) 25–34 0.189*** 5.31 0.244*** 4.56 1.276*** 0.198*** 5.56 0.258*** 4.83 1.294*** 35–44 0.445*** 12.21 0.623*** 11.32 1.865*** 0.455*** 12.48 0.642*** 11.65 1.901*** 45–54 0.637*** 17.87 0.897*** 16.63 2.452*** 0.640*** 17.95 0.906*** 16.79 2.475*** 55–64 0.751*** 20.24 1.079*** 19.18 2.943*** 0.754*** 20.31 1.085*** 19.25 2.958*** 65� 0.881*** 20.67 1.289*** 19.86 3.630*** 0.882*** 20.70 1.291*** 19.86 3.637*** Black (Non-Hispanic) �0.294*** �15.16 �0.440*** �14.99 0.644*** �0.284*** �14.63 �0.427*** �14.53 0.652*** Education (reference category: Did not complete high school) High school 0.347*** 9.45 0.466*** 8.45 1.594*** 0.346*** 9.44 0.469*** 8.49 1.598*** GED 0.265*** 5.90 0.380*** 5.64 1.462*** 0.265*** 5.91 0.386*** 5.73 1.471*** Some college 0.714*** 20.00 0.990*** 18.35 2.691*** 0.716*** 20.06 0.995*** 18.43 2.705*** College graduate 0.963*** 25.05 1.403*** 24.00 4.067*** 0.966*** 25.14 1.410*** 24.09 4.098*** Post-graduate 1.111*** 26.32 1.700*** 26.29 5.474*** 1.108*** 26.26 1.696*** 26.18 5.450*** Marital status (reference category: Married) Living-with-partner �0.072** �2.26 �0.113** �2.40 0.893** �0.077** �2.42 �0.122*** �2.60 0.885*** Single �0.016 �0.78 �0.005 �0.15 0.995 �0.019 �0.94 �0.011 �0.34 0.989 Income (reference category: Less than $15,000) $15,000 - $24,999 0.135*** 4.00 0.187*** 3.72 1.206*** 0.154*** 4.58 0.214*** 4.26 1.239*** $25,000 - $34,999 0.231*** 6.62 0.342*** 6.49 1.407*** 0.243*** 6.98 0.359*** 6.81 1.431*** $35,000 - $49,999 0.368*** 10.81 0.513*** 10.01 1.670*** 0.385*** 11.33 0.537*** 10.48 1.711*** $50,000 - $74,999 0.531*** 15.60 0.742*** 14.43 2.100*** 0.536*** 15.76 0.748*** 14.55 2.114*** $75,000 - $99,999 0.644*** 16.96 0.941*** 16.25 2.561*** 0.640*** 16.84 0.933*** 16.06 2.541*** $100,000 - $149,999 0.755*** 19.03 1.146*** 18.88 3.144*** 0.744*** 18.69 1.126*** 18.46 3.083*** $150,000 or more 0.835*** 18.55 1.372*** 19.52 3.942*** 0.818*** 18.02 1.341*** 18.93 3.824*** Employment (reference category: Self-employed) Employed full-time �0.070** �2.20 �0.127*** �2.61 0.881*** �0.069** �2.15 �0.125** �2.57 0.882** Employed part-time �0.123*** �3.13 �0.204*** �3.41 0.816*** �0.126*** �3.21 �0.209*** �3.50 0.812*** Homemaker �0.165*** �4.10 �0.235*** �3.87 0.790*** �0.170*** �4.21 �0.244*** �4.01 0.783*** Full-time student 0.157*** 2.91 0.197** 2.40 1.218** 0.154*** 2.85 0.187** 2.29 1.206** Disabled �0.219*** �4.68 �0.344*** �4.88 0.709*** �0.214*** �4.58 �0.337*** �4.77 0.714*** Unemployed �0.104** �2.48 �0.177*** �2.81 0.838*** �0.104** �2.49 �0.179*** �2.84 0.836*** Retired �0.095** �2.53 �0.146** �2.53 0.864** �0.101*** �2.67 �0.154*** �2.67 0.857*** Region (reference category: Northeast) Midwest 0.064** 2.54 0.098** 2.54 1.103** 0.063** 2.48 0.094** 2.44 1.099** South 0.041* 1.72 0.045 1.22 1.046 0.042* 1.77 0.046 1.26 1.047 West 0.167*** 6.66 0.240*** 6.29 1.272*** 0.169*** 6.75 0.243*** 6.37 1.276*** Observations 22,218 22,218 22,218 22,218 Adjusted R2 0.296 0.297 This table reports OLS and Ordered Logit regression results where the dependent variable, Financial Literacy Score, is set equal to the number of financial literacy questions answered correctly, ranging from 0 to 5. The independent variables include Financial Education, Female, Age Group, Black (Non-Hispanic), Education, Marital Status, Income, Employment, and Region. The independent variable of interest in the left panel, FA User, is set equal to 1 if the person used one or more of the five types of financial advisers, or 0 otherwise. The independent variable of interest in the right panel, the five specific financial adviser types, are set equal to 1 if the person used that financial adviser type, or 0 otherwise. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 143B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 51%. We also find that Female is associated with a 52% decline in the probability of answering more financial literacy questions correctly which is consistent with Lusardi and Mitchell (2008) that find females have lower levels of financial literacy. Black (Non- Hispanic) is associated with a 36% decline in the probability of answering more financial literacy questions correctly. We find a monotonically increasing probability of higher Financial Literacy Scores across higher Age Groups, Education levels, and Income levels. For Marital Status and Employment, those “living-with-partner” are 11% less likely to answer more financial literacy questions correctly compared with those that are married, and self-employed individuals are more likely to answer more questions correctly. For Region, those living in the Midwest and West are 10% and 27% more likely to answer more financial literacy questions correctly, respectively. Our main variable of interest, FA User, indicates that those using financial advisers are 33.5% more likely to answer more financial literacy questions correctly compared with those that did not use financial advisers. This estimate is highly statistically significant, and reinforces the OLS results that financial advisers positively influence the financial literacy of their clients, even when controlling for other demographic characteristics that are also positively related to higher financial literacy scores. In the right panel of Table 5 we re-estimate both the OLS and Ordered Logit models where we replace the FA User independent dummy variable with dummy variables for each financial adviser type, separately, to determine the impact each adviser type has on financial literacy. The results for the demographic characteristic control variables are similar to those reported for the original model specification above. In the first two columns using the OLS specification, we find that only Savings and Investments, Mortgage or Loan, and Insurance advisers are positively related to higher Financial Literacy Scores, but Tax Planning advisers do not significantly impact Financial Literacy Scores. The Ordered Logit model reveals that Savings and Investments increase the probability of obtaining a higher Financial Literacy Score by 28%, while Mortgage or Loan and Insurance advisers increases this probability by 16% and 8%, respectively, and Tax Planning advisers have no significant impact on financial literacy. We find that Financial Literacy Scores are negatively related to Debt Counseling advisers, where Debt Counselors appear to negatively influence the financial literacy of their clients and reduce the probability of a higher Financial Literacy Score by 65%. Robb, Babiarz, and Woodyard (2012), using the 2009 NFCS survey, also find a negative relationship between using Debt Counselors and financial literacy scores because those individuals appear to make financial mistakes in the first place and do not learn from their experiences. Gathergood (2012) suggests debt counseling services involves offering solutions to existing debt prob- lems, such as debt-repayment plans, usually to clients that lack self-control, have lower levels of financial literacy, and are considered poor learners. It appears that the Debt Counselor results are consistent with this characterization of individuals that use Debt Counselors. This is also consistent with our univariate results reported in Table 1, Panel B, where we find that Debt Counselor users perform very well on the financial literacy question related to the total interest paid over a 15-year mortgage versus a 30-year mortgage, but underperform on the other financial literacy questions. This suggests that Debt Counselors do not influence the financial literacy of their clients in the same manner that other adviser types 144 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 do because they focus on topics specifically related to interest rates and interest payments related to debt products. In summary, using both OLS and Ordered Logit models, we find that financial advisers, in general, have a positive influence on the financial literacy of the individuals that use them suggesting they play a role in helping disseminate financial knowledge to their clients rather than just offering them advice about what they should do in making financial decisions. When considering the five different types of financial advisers separately, we find that it is Savings or Investment advisers that have the largest impact on financial literacy, followed by Mortgage or Loan and Insurance advisers, respectively. This is understandable since the five financial literacy questions asked pertain to the time value of money, interest rates, inflation, and risk diversification which are topics that a Savings or Investment adviser should cover with their client. Tax Planning advisers, which we find have no influence on Financial Literacy Scores, and Debt Counselors, which have a negative influence on financial literacy scores, would be less likely to cover these types of topics with their clients. 4. Endogeneity and confounding variables There is the possibility that our model suffers from endogeneity issues in the choice to use a financial adviser. It is possible that those having higher levels of financial literacy are more likely to use financial advisers. Even though we control for demographic characteristics that we find to be directly related to financial literacy in all of our regression specifications, such as Financial Education and Education, this may not be enough to control for this endoge- neity bias. Lacking longitudinal data that would directly address this problem, we must try to control for endogeneity using econometric techniques given our data limitations. We address these concerns in this section using an Instrumental Variables (IV) model and Propensity Score Matching. 4.1. Instrumental variables model To address the potential endogeneity issues between financial literacy and the choice to use a financial adviser, we implement a two-stage IV estimation model. We estimate IV models for each of the financial adviser types to include Debt Counseling, Savings or Investments, Mortgage or Loan, Insurance, and Tax Planning using both the GMM estimator and IV Probit estimation.2 In the first stage, we use responses to available survey questions that are expected to predict the choice to use a specific type of financial adviser as instruments, but these instruments must be uncorrelated with the second stage dependent variable that is the Financial Literacy Score. The same demographic control variables from Table 5 are also used in both the first- and second-stage model specifi- cations. Table 6, Panel A, identify the instruments (survey questions) used for each of the five financial adviser types. We also report the first-stage coefficient estimates for the instruments used from the GMM estimator and IV Probit estimation, separately.3 All instruments used are positively related to using each specific type of financial adviser, and they are all highly 145B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Table 6 Instrumental variable regressions Panel A: First stage instruments for each financial adviser type Instruments used Instrument survey question: Instrumented: Debt Counseling Home foreclosure Have you been involved in a foreclosure process on your home? Declared bankruptcy Have you declared bankruptcy in the last two years? Instrumented: Savings or Investments Employer retirement plan Do you have any retirement plan through a current or previous employer? Other retirement plan Do you have any other retirement accounts not through an employer? Instrumented: Mortgage or Loan Own your home Do you currently own your home? Home foreclosure Have you been involved in a foreclosure process on your home? Instrumented: Insurance Health insurance Are you covered by health insurance? Life insurance Do you have a life insurance policy? Instrumented: Tax Planning Earn salary or wages Did you receive any income in the form of Salary or Wages? Social Security benefits Did you receive any income from Social Security retirement benefits? First stage instruments coefficient estimates IV GMM t stat IV Probit z-stat Debt Counseling: Home foreclosure 0.166*** 9.29 0.166*** 15.38 Declared bankruptcy 0.381*** 18.87 0.381*** 34.28 Savings or Investments: Employer retirement plan 0.082*** 11.33 0.082*** 11.41 Other retirement plan 0.285*** 34.95 0.285*** 39.75 Mortgage or Loan: Own your home 0.119*** 18.81 0.119*** 17.97 Home foreclosure 0.161*** 9.25 0.161*** 10.65 Insurance: Health insurance 0.026*** 2.98 0.026*** 2.82 Life insurance 0.093*** 12.51 0.093*** 12.55 Tax Planning: Earns or wages 0.035*** 5.28 0.035*** 5.18 Social Security benefits 0.099*** 10.13 0.099*** 11.12 Panel B: Second stage dependent variable: Financial Literacy Score (0 to 5) Financial Literacy Score: 0 to 5 Financial Literacy Score: 0 to 5 IV GMM t stat IV Probit z-stat IV GMM t stat IV Probit z-stat Constant 1.756*** 27.24 0.677*** 7.02 1.587*** 23.04 0.531*** 4.99 Debt Counseling �0.870*** �7.57 �0.071 �0.39 Savings or Investments 1.174*** 16.10 1.248*** 7.69 Financial Education 0.313*** 16.15 0.366*** 7.86 0.189*** 8.67 0.303*** 5.81 Female �0.445*** �25.72 �0.212*** �6.40 �0.456*** �24.67 �0.238*** �6.66 Age group (reference category: 18–24) 25–34 0.221*** 5.71 0.203*** 3.76 0.195*** 4.58 0.173*** 2.89 35–44 0.446*** 11.20 0.235*** 4.17 0.508*** 11.56 0.308*** 4.75 45–54 0.603*** 15.64 0.394*** 7.00 0.656*** 15.44 0.466*** 7.26 55–64 0.725*** 18.25 0.505*** 8.16 0.714*** 16.48 0.490*** 7.20 65� 0.850*** 19.10 0.590*** 7.51 0.785*** 16.36 0.523*** 6.17 Black (Non-Hispanic) �0.269*** �13.02 �0.120*** �3.40 �0.278*** �12.69 �0.120*** �3.18 146 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Table 6 (Continued) Financial Literacy Score: 0 to 5 Financial Literacy Score: 0 to 5 IV GMM t stat IV Probit z-stat IV GMM t stat IV Probit z-stat Education (reference category: Did not complete high school) High school 0.354*** 8.81 0.318*** 6.32 0.338*** 8.15 0.315*** 5.74 GED 0.273*** 5.57 0.227*** 3.64 0.289*** 5.72 0.283*** 4.16 Some college 0.749*** 19.25 0.603*** 11.72 0.670*** 16.53 0.540*** 9.59 College graduate 1.019*** 24.57 0.722*** 11.68 0.850*** 19.22 0.590*** 8.52 Post-graduate 1.151*** 26.39 0.687*** 8.86 0.976*** 20.47 0.555*** 6.34 Marital status (reference category: Married) Living-with-partner �0.092*** �2.91 �0.032 �0.58 �0.084** �2.43 �0.036 �0.60 Single �0.039* �1.82 �0.119*** �3.20 �0.031 �1.39 �0.126*** �3.15 Income (reference category: Less than $15,000) $15,000 - $24,999 0.185*** 5.04 0.153*** 3.02 0.120*** 3.16 0.118** 2.18 $25,000 - $34,999 0.277*** 7.16 0.118** 2.22 0.162*** 4.00 0.034 0.58 $35,000 - $49,999 0.435*** 11.73 0.262*** 4.73 0.283*** 7.21 0.129** 2.10 $50,000 - $74,999 0.594*** 16.35 0.450*** 7.59 0.372*** 9.10 0.251*** 3.59 $75,000 - $99,999 0.705*** 17.59 0.362*** 5.27 0.420*** 8.96 0.107 1.27 $100,000 - $149,999 0.810*** 19.98 0.579*** 6.97 0.475*** 9.55 0.245** 2.40 $150,000 or more 0.899*** 19.86 0.580*** 5.60 0.506*** 8.97 0.197 1.56 Employment (reference category: Self-employed) Employed full-time �0.068** �2.12 0.029 0.45 �0.033 �0.96 0.081 1.16 Employed part-time �0.129*** �3.18 �0.013 �0.18 �0.132*** �3.06 �0.008 �0.10 Homemaker �0.205*** �4.89 �0.043 �0.58 �0.074* �1.66 0.076 0.95 Full-time student 0.135** 2.34 0.178* 1.88 0.149** 2.39 0.181* 1.76 Disabled �0.232*** �4.79 �0.074 �0.91 �0.140*** �2.76 0.005 0.06 Unemployed �0.137*** �3.17 �0.020 �0.27 �0.027 �0.58 0.083 1.04 Retired �0.119*** �3.21 �0.092 �1.15 �0.099** �2.52 �0.110 �1.29 Region (reference category: Northeast) Midwest 0.054** 2.07 0.050 1.06 0.067** 2.39 0.089* 1.76 South 0.039 1.59 0.085* 1.91 0.049* 1.85 0.110** 2.31 West 0.172*** 6.71 0.144*** 2.99 0.167*** 6.08 0.167*** 3.24 Observations 22,069 22,069 21,095 21,095 Adjusted R2 0.275 0.199 Financial Literacy Score: 0 to 5 Financial Literacy Score: 0 to 5 IV GMM t stat IV Probit z-stat IV GMM t stat IV Probit z-stat Constant 1.669*** 24.94 0.552*** 5.31 1.657*** 21.98 0.405*** 3.24 Mortgage or Loan 0.410*** 2.61 0.938*** 3.59 Insurance 0.306 1.58 1.383*** 3.85 Financial Education 0.260*** 12.28 0.293*** 5.99 0.249*** 9.27 0.206*** 3.45 Female �0.456*** �26.50 �0.225*** �6.71 �0.459*** �25.47 �0.249*** �6.94 Age group (reference category: 18–24) 25–34 0.183*** 4.72 0.172*** 3.13 0.172*** 4.33 0.183*** 3.18 35–44 0.438*** 11.06 0.252*** 4.40 0.428*** 10.52 0.249*** 4.13 45–54 0.641*** 16.26 0.460*** 7.70 0.616*** 15.42 0.441*** 7.21 55–64 0.768*** 18.72 0.581*** 8.81 0.743*** 18.10 0.545*** 8.20 65� 0.911*** 19.75 0.676*** 8.19 0.884*** 19.25 0.653*** 7.78 Black (Non-Hispanic) �0.277*** �13.37 �0.088** �2.44 �0.298*** �14.49 �0.132*** �3.57 Education (reference category: Did not complete high school) High school 0.339*** 8.56 0.297*** 5.71 0.356*** 8.91 0.304*** 5.62 GED 0.250*** 5.17 0.204*** 3.20 0.262*** 5.36 0.209*** 3.12 Some college 0.708*** 17.92 0.535*** 9.74 0.717*** 17.49 0.502*** 8.26 College graduate 0.959*** 22.47 0.641*** 9.69 0.972*** 21.93 0.619*** 8.50 Post-graduate 1.119*** 24.89 0.605*** 7.42 1.121*** 23.96 0.574*** 6.63 Marital status (reference category: Married) Living-with-partner �0.075** �2.38 0.001 0.01 �0.068** �2.15 0.006 0.10 Single �0.012 �0.55 �0.063 �1.58 �0.012 �0.55 �0.053 �1.30 Income (reference category: Less than $15,000) $15,000 - $24,999 0.140*** 3.89 0.134*** 2.63 0.131*** 3.28 0.054 0.89 $25,000 - $34,999 0.226*** 5.82 0.071 1.28 0.237*** 5.56 0.014 0.22 $35,000 - $49,999 0.369*** 9.81 0.204*** 3.50 0.379*** 8.62 0.115 1.58 $50,000 - $74,999 0.524*** 13.09 0.346*** 5.24 0.536*** 11.27 0.253*** 3.06 147B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Table 6 (Continued) Financial Literacy Score: 0 to 5 Financial Literacy Score: 0 to 5 IV GMM t stat IV Probit z-stat IV GMM t stat IV Probit z-stat $75,000 - $99,999 0.626*** 13.58 0.235*** 2.97 0.659*** 12.89 0.192** 2.09 $100,000 - $149,999 0.739*** 15.28 0.427*** 4.51 0.767*** 14.04 0.356*** 3.27 $150,000 or more 0.827*** 14.97 0.401*** 3.41 0.842*** 13.57 0.268** 2.09 Employment (reference category: Self Employed) Employed full-time �0.081** �2.53 0.018 0.28 �0.067** �1.97 0.077 1.07 Employed part-time �0.120*** �2.99 0.012 0.16 �0.124*** �2.98 0.028 0.35 Homemaker �0.168*** �3.99 �0.000 �0.00 �0.152*** �3.26 0.088 1.01 Full-time student 0.168*** 2.86 0.224** 2.30 0.160** 2.47 0.265** 2.41 Disabled �0.218*** �4.51 �0.042 �0.50 �0.211*** �4.27 �0.007 �0.08 Unemployed �0.113*** �2.60 0.032 0.42 �0.093* �1.92 0.134 1.52 Retired �0.087** �2.34 �0.062 �0.77 �0.089** �2.27 �0.034 �0.40 Region (reference category: Northeast) Midwest 0.063** 2.44 0.051 1.06 0.056** 2.10 0.026 0.51 South 0.038 1.56 0.081* 1.80 0.035 1.44 0.083* 1.80 West 0.155*** 5.92 0.109** 2.19 0.162*** 6.17 0.111** 2.17 Observations 22,020 22,020 21,645 21,645 Adjusted R2 0.286 0.288 Financial Literacy Score: 0 to 5 IV GMM t stat IV Probit z-stat Constant 1.753*** 22.38 0.595*** 4.75 Tax Planning �0.162 �0.59 0.591 1.24 Financial Education 0.292*** 11.44 0.319*** 5.63 Female �0.455*** �26.52 �0.219*** �6.46 Age group (reference category: 18–24) 25–34 0.166*** 4.24 0.194*** 3.45 35–44 0.405*** 9.23 0.250*** 3.75 45–54 0.583*** 13.51 0.429*** 6.43 55–64 0.704*** 16.22 0.524*** 7.48 65� 0.842*** 18.68 0.609*** 7.44 Black (Non-Hispanic) �0.294*** �14.55 �0.122*** �3.42 Education (reference category: Did not complete High School) High school 0.364*** 9.14 0.321*** 6.23 GED 0.270*** 5.54 0.240*** 3.76 Some college 0.742*** 18.74 0.576*** 10.52 College graduate 1.012*** 22.58 0.704*** 9.88 Post-graduate 1.174*** 22.39 0.657*** 6.85 Marital status (reference category: Married) Living-with-partner �0.087*** �2.66 0.004 0.06 Single �0.027 �1.26 �0.098** �2.52 Income (reference category: Less than $15,000) $15,000 - $24,999 0.161*** 4.37 0.132** 2.52 $25,000 - $34,999 0.285*** 6.97 0.092 1.54 $35,000 - $49,999 0.437*** 10.29 0.215*** 3.16 $50,000 - $74,999 0.611*** 12.60 0.377*** 4.53 $75,000 - $99,999 0.751*** 12.52 0.279*** 2.66 $100,000 - $149,999 0.868*** 12.09 0.438*** 3.27 $150,000 or more 0.973*** 10.16 0.394** 2.16 Employment (reference category: Self-employed) Employed full-time �0.098** �2.36 0.065 0.81 Employed part-time �0.146*** �3.45 �0.002 �0.03 Homemaker �0.199*** �4.18 �0.005 �0.06 148 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 statistically significant at the 1% level. This indicates they are legitimate instruments to use in predicting the probability of using each type of financial adviser. Table 6, Panel B, report results for the second-stage estimates for each model where Financial Literacy Score is the dependent variable, and the predicted financial adviser type user from the first stage model and the demographic characteristics are used as independent variables. Using the GMM estimator we find that Debt Counseling remains negatively related to Financial Literacy Scores, however, the IV Probit estimates indicate that Debt Counseling is not significantly related to Financial Literacy Scores. Savings or Investments and Mortgage or Loan remains positively related to higher Financial Literacy Scores using both the GMM estimator and IV Probit estimates. Insurance advisers remain positively related with higher Financial Literacy Scores only when using the IV Probit estimates, and not significant when using the GMM estimator. Tax Planning remains not signifi- cantly related to Financial Literacy Scores using both the GMM estimator and IV Probit estimates. Overall, our results in Table 6 are similar to those reported in Table 5 with the exception that Debt Counseling is not significant when using the IV Probit estimate, and Insurance is not significant when using the IV GMM estimator. Savings or Investments and Mortgage or Loan advisers appear to have the strongest impact on the financial literacy of their clients, even when controlling for demographic characteristics that are shown to impact financial Table 6 (Continued) Financial Literacy Score: 0 to 5 IV GMM t stat IV Probit z-stat Full-time student 0.116* 1.76 0.209* 1.86 Disabled �0.236*** �4.24 �0.015 �0.15 Unemployed �0.131*** �2.69 0.025 0.29 Retired �0.107*** �2.70 �0.089 �1.06 Region (reference category: Northeast) Midwest 0.060** 2.33 0.057 1.19 South 0.036 1.49 0.089** 1.98 West 0.173*** 6.77 0.150*** 3.06 Observations 21,911 21,911 Adjusted R2 0.287 This table reports Instrumental variable (IV) model results where the first stage dependent variable, Debt Counseling, Savings or Investments, Mortgage or Loan, Insurance, or Tax Planning is set equal to 1 if the person used that type of adviser, or 0 otherwise. The independent variables used in the first stage model include instruments for each financial adviser type and the demographic control variables Financial Education, Female, Age Group, Black (Non-Hispanic), Education, Marital Status, Income, Employment, and Region. Panel A report the instruments used for each financial adviser type and IV GMM and IV Probit coefficient estimates and t stats (z-stats) for the instruments from the first stage model results, respectively. The second stage model dependent variable is the Financial Literacy Score that is set equal to the number of financial literacy questions answered correctly, ranging from 0 to 5. The independent variables used in the second stage model include the predicted financial adviser user variable from the first stage model and the same demographic control variables used in the first stage model. Panel B report the IV GMM and IV Probit coefficient estimates and t stats (z-stats) for the second stage model results for each financial adviser type separately. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 149B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 literacy and when controlling for potential endogeneity issues between financial literacy and the choice to use a financial adviser. 4.2. Propensity score matching In an attempt to reduce any confounding variable biases in our results we implement a Propensity Score Matching model to estimate the impact financial advisers have on the financial literacy of their clients. We first estimate the same logit regression used in Table 3 where FA User, the dependent variable, is regressed on to the different demographic characteristic control variables to include Financial Education, Female, Age Group, Black (Non-Hispanic), Education, Marital Status, Income, Employment, and Region. Using the propensity scores from this logit estimation we perform a nearest neighbor match that ensures that a financial adviser user is paired with a non-user with statistically the same demographic characteristics. We conduct the matching with replacement to achieve the best match for each financial adviser user with a non-user. We separately conduct the matching without replacement to ensure we have a 1:1 match for each financial adviser user with a non-user that maximizes our matched sample size. We then use these matched samples to estimate an OLS univariate regression model where the dependent variable is Financial Literacy Score and the independent variable is FA User. We repeat this process using the logit model from Table 4 for each of the financial adviser types separately. Table 7, Panel A, report results using propensity score matching with replacement. Consistent with earlier results, we find there is a positive and statistically significant relationship between FA User and Financial Literacy Scores. When considering the financial adviser types separately, we find that Debt Counseling is negatively related to Financial Literacy Score, while Savings or Investments, Mortgage or Loan, Insurance, and Tax Planning are all positively related to Financial Literacy Scores. Table 7, Panel B, report results using propensity score matching without replacement. We again find that there is a positive and statistically significant relationship between FA User and Financial Literacy Scores. When considering the financial adviser types separately, we continue to find that Debt Counseling is negatively related to Financial Literacy Score, while Savings or Investments, Mortgage or Loan, and Insurance are all positively related to Financial Literacy Scores. However, we find that Insurance is not significantly related to financial literacy when using matching without replacement. These results continue to indicate that Savings or Investments advisers have the strongest influence on the financial literacy of their clients, followed by Mortgage or Loan and Insurance advisers. Debt Counselors is consistently shown to be negatively related to financial literacy, and Tax Planning advisers are only marginally, or not at all, related to the financial literacy of their clients. 5. Conclusions Using the 2012 FINRA Investor Education Foundation National Financial Capability Study, we determine the demographic characteristics associated with people that use finan- 150 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 cial advisers, and the impact financial advisers have on the financial literacy of their clients. We consider five different types of financial advisers: Debt Counselors, Savings or Invest- ments, Mortgage or Loan, Insurance, and Tax Planning. We first find that there has been a significant increase over the last decade in the number of individuals who use financial advisers. Elmerick, Montalto, and Fox (2002), using the 1998 Survey of Consumer Finances that is a similar survey to the 2012 FINRA survey we are using, reported that only 21% of those surveyed had used a financial adviser. Using the more recent 2012 FINRA survey, we find that over 53% of those surveyed reported using a financial adviser indicating the usage of financial advisers has more than doubled since the 1998 survey was conducted. When considering the demographic characteristics associated with those using a financial adviser, we find that females are 13% more likely, and those that have received a financial Table 7 Propensity score matching OLS Univariate Regression: Financial Literacy Score (0 to 5) � Adviser user dummy FA User Debt Counseling Savings or Investments Mortgage or Loan Insurance Tax Planning Panel A: Propensity score matching with replacement Number of users 11,872 1,949 7,025 4,914 7,312 4,422 Matches 4,494 1,660 3,692 3,151 4,142 2,872 Intercept 2.95*** 3.06*** 3.20*** 3.21*** 3.14*** 3.31*** (145.09) (91.52) (146.47) (134.12) (148.88) (132.08) Coefficient 0.41*** �0.20*** 0.32*** 0.21*** 0.19*** 0.17*** (17.14) (�4.47) (11.79) (6.95) (7.27) (5.37) Panel B: Propensity score matching without replacement Number of users 10,346 1,949 7,025 4,914 7,312 4,422 Matches 10,346 1,949 7,025 4,914 7,312 4,422 Intercept 2.77*** 3.06*** 3.31*** 3.29*** 3.25*** 3.45*** (201.72) (99.02) (209.72) (171.76) (204.81) (171.76) Coefficient 0.60*** �0.20*** 0.21*** 0.13*** 0.08*** 0.04 (30.69) (�4.65) (9.28) (4.68) (3.75) (1.25) This table reports Propensity score matching results where we first estimate the Logit regression in Table 3 to match each Financial Adviser user with a non-user having statistically the same Financial Education, Gender, Age Group, Ethnicity, Education, Marital Status, Income, Employment, and Region. We then estimate a Univariate regression using this matched sample where the dependent variable is the Financial Literacy Score that is set equal to the number of financial literacy questions answered correctly from 0 to 5, and the independent variable is the FA User dummy that is set equal to 1 if the person used one or more of the five types of financial advisers, or 0 otherwise. We repeat this process for each financial adviser type separately to include Debt Counseling, Savings or Investments, Mortgage or Loan, Insurance, and Tax Planning using the logit regressions from Table 4 to determine the propensity score matches. The dummy variables Debt Counseling, Savings or Investments, Mortgage or Loan, Insurance, and Tax Planning are set equal to 1 if the person used that type of financial adviser, or 0 otherwise. Panel A report results where propensity score matches were created with replacement in order to determine the best match for each observation. Panel B report results using matching without replacement in order to increase the sample size used in the Univariate regressions. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively, with t stats reported in parenthesis. 151B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 education in the past nearly 50% more likely, to use financial advisers. Those that are married, or self-employed, are also more likely to use financial advisers. We find that there is a monotonically increasing probability of using financial advisers across higher education and income levels, but younger age groups are less likely to use financial advisers. Surpris- ingly, ethnicity, after controlling for all other demographic characteristics, does not appear to significantly impact the likelihood of using a financial adviser. Finally, we find that respondents from the Midwest and West are more likely to use a financial adviser than respondents from the South or Northwest. When considering the demographic characteristics of those that used a specific type of financial adviser we find that those receiving a financial education in the past are more likely to use all of the five financial adviser types compared with those that did not. Females are more likely to use Insurance advisers compared with males, but not more likely to use any of the other four adviser types. Black (Non-Hispanic) people are more likely to use Debt Counselors but less likely to use Mortgage or Loan advisers. We find that older people are more likely to use Savings and Investments advisers but younger people are more likely to use Debt Counselors and Mortgage or Loan advisers. Those having a High School education or higher are more likely to use Savings or Investment advisers, but only those having some college or higher are more likely to use Insurance or Tax Planning advisers. People with higher income levels are less likely to use Debt Counselors, but more likely to use all of the other adviser types. Those with lower income levels are more likely to use Debt Counselors. Married couples are more likely to use Mortgage or Loan, Insurance, and Tax Planning advisers, and those that are self-employed are more likely to use all of the five financial adviser types. Finally, we find that those living in the West are more likely to use Mortgage or Loan advisers, and those in the Midwest and West are more likely to use Insurance advisers. Finally, after controlling for demographic characteristics that are shown to impact finan- cial literacy, we continue to find a positive and statistically significant relationship between financial literacy scores and the use of financial advisers suggesting they have a positive influence on the financial literacy of their clients. When considering the different financial adviser types separately we find that Savings or Investments, Mortgage Loan, and Insurance advisers have the highest positive influence on the financial literacy of their clients. However, Debt Counselors appear to have a negative influence on the financial literacy of their clients which is consistent with Robb, Babiarz, and Woodyard (2012) suggesting those individuals appear to make financial mistakes in the first place and do not appear to learn from their experiences. We find that Tax Planning advisers do not significantly impact the financial literacy scores of their clients. The relationship between the choice to use a financial adviser and financial literacy may suffer from endogeneity issues and confounding variable biases in the models we use, so we attempt to correct these biases using Instrument Variables and Propensity Score Matching methodology. Our results remain robust indicating that Savings or Investment, Mortgage or Loan, and Insurance advisers have a positive influence on the financial literacy of their clients. Even though it is difficult to measure the impact financial advisers have on the financial literacy of their clients with precision, we have established that a positive relationship exists between financial advisers and financial literacy after controlling for demographic charac- teristics that are shown to impact financial literacy. This relationship also remains robust 152 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 when controlling for the possibility of endogeneity and confounding variable problems that may exist in a model relating financial adviser use with financial literacy scores. The positive relationship we find that exists between financial adviser use and financial literacy scores suggest that clients appear to learn from their interactions with financial advisers. These results show that financial advisers do not just offer financial advice, but also play a significant role in dissemi- nating financial knowledge to their clients that has a positive impact on financial literacy. Future research using longitudinal data that measures financial literacy before and after using a financial adviser, or that is able to track the financial adviser-client relationship through time, would be able to measure the impact financial advisers have on the financial literacy of their clients with more precision. Notes 1 Debt counseling refers to the services provided to individuals who have problems repaying their debt on time. For further details, please see: http://www.consumer finance.gov/askcfpb/1449/whats-difference-between-credit-counselor-and-debt- settlement-company.html. 2 We also estimated the IV model using two-stage Least Squares (2SLS) with results nearly identical to the reported GMM results. 3 Coefficient estimates for the demographic control variables are nearly identical to what is reported in Table 4. Acknowledgment Research Funded by The University of Akron-Finance Advisory Board 2014 Economic Summit Awards Program Grant. Appendix A: Financial literacy questions Below are the five financial literacy questions asked in the survey that is used to measure the financial literacy of the survey respondent: 1. Suppose you had $100 in a savings account and the interest rate was 2% per year. After 5 years, how much do you think you would have in the account if you left the money to grow? a. More than $102 b. Exactly $102 c. Less than $102 d. Don’t know e. Refused 2. Imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. After 1 year, how much would you be able to buy with the money in this account? 153B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 a. More than today b. Exactly the same c. Less than today d. Don’t know e. Refused 3. If interest rates rise, what will typically happen to bond prices? a. They will rise b. They will fall c. They will stay the same d. There is no relationship between bond prices and the interest e. Don’t know f. Refused 4. A 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less. a. True b. False c. Don’t know d. Refused 5. Buying a single company’s stock usually provides a safer return than a stock mutual fund. a. True b. False c. Don’t know d. Refused References Allgood, S., & Walstad, W. B. (2016). The effects of perceived and actual financial literacy on financial behaviors. Economic Inquiry, 54, 675–697. Bandura, A. (1977). Social Learning Theory. Englewood Cliffs, NJ: Prentice-Hall. Barber, B. M., & Odean, T. (2001). Boys will be boys: Gender, overconfidence, and common stock investment. The Quarterly Journal of Economics, 116, 261–292. Black, K. Jr., Ciccotello, C. S., & Skipper, H. D., Jr. (2002). Issues in comprehensive financial planning. Financial Services Review, 11, 1–9. Bucher-Koenen, T., & Lusardi A. (2011). Financial literacy and retirement planning in Germany. Journal of Pension Economics and Finance, 10, 565–584. Calvet, L. E., Campbell, J. Y., & Sodini, P. (2009). Measuring the financial sophistication of households. American Economic Review, 99, 393–398. Campbell, J. Y. (2006). Household finance. Journal of Finance, 59, 1553–1604. Elmerick, S. A., Montalto, C. P., & Fox, J. J. (2002). Use of financial planners by U. S. households. Financial Services Review, 11, 217–231. 154 B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155 Gathergood, J. (2012). Self-control, financial literacy, and over-indebtedness. Journal of Economic Psychology, 33, 590–602. Grimes, P. W., Rogers, K. E., & Smith, R. C. (2010). High school economic education and access to financial services. Journal of Consumer Affairs, 44, 317–335. Huston, S. J. (2012). Financial literacy and the cost of borrowing. International Journal of Consumer Studies, 36, 566–572. Johnson, E., & Sherraden, M. (2007). From financial literacy to financial capability among youth. Journal of Social Welfare, 34, 119–146. Klapper, L., Lusardi, A., & Panos, G. A. (2013). Financial literacy and its consequences: Evidence from Russia during financial crisis. Journal of Banking & Finance, 37, 3904–3923. Kolb, D. A. (1984). Experiential Learning: Experience as the Source of Learning and Development. Englewood Cliffs, NJ: Prentice-Hall, Inc. Lachance, M. (2014). Financial literacy and neighborhood effects. Journal of Consumer Affairs, 48, 251–273. Lusardi, A., & Mitchell, O. S. (2007a). Baby boomers retirement security: The roles of planning, financial literacy, and housing wealth. Journal of Monetary Economics, 54, 205–224. Lusardi, A., & Mitchell, O. S. (2007b). Financial literacy and retirement preparedness: Evidence and implication for financial education. Business Economics, 42, 35–44. Lusardi, A., & Mitchell, O. S. (2008). Planning and financial literacy: How do women fare? American Economic Review: Papers and Proceedings, 98, 413–417. Lusardi, A., Mitchell, O. S., & Curto, V. (2010). Financial literacy among the young. Journal of Consumer Affairs, 44, 358–380. Lusardi, A., & Tufano, P. (2009). Debt literacy, financial experiences, and overindebtedness. NBER working paper 14808, National Bureau of Economic Research, Inc. Moore, D. (2003). Survey of financial literacy in Washington State: knowledge, behavior, attitudes and experi- ences. Technical report 03–39. Social and Economic Sciences Research Center, Washington State University. Robb, C. A., Babiarz, P., & Woodyard, A. (2012). The demand for financial professions’ advice: The role of financial knowledge, satisfaction, and confidence, Financial Services Review, 21, 291–305. Tang, N., & Peter, P. C. (2015). Financial knowledge acquisition among the young: The role of financial education, financial experience, and parents’ financial experience. Financial Services Review, 24, 119–137. Van Rooij, M., Lusardi, A., & Alessie, R. (2011). Financial literacy and stock market participation. Journal of Financial Economics, 101, 449–472. Walstad, W. B., Rebeck, K., & MacDonald, R. A. (2010). The effects of financial education on the financial knowledge of high school students. Journal of Consumer Affairs, 44, 336–357. Winchester, D. D., & Huston, S. J. (2014). Does a relationship with financial service professional overcome a client’s sense of not being in control of achieving their goals? Financial Services Review, 23, 1–23. 155B. Balasubramnian, E.R. Brisker / Financial Services Review 25 (2016) 127–155