Description
The Get Checking™ program is a “second chance” program that aims to provide financial education to consumers who were reported to ChexSystems by a previous financial institution for account abuse or mismanagement
Effectiveness of Financial Education on Financial Management Behavior and
Account Usage: Evidence from a ‘Second Chance’ Program
Rebecca Haynes-Bordas, D. E. Kiss and Tansel Yilmazer
Acknowledgements We thank Bharathi Nagarajan for her assistance with the collection of the
data and Patryk Babiarz for excellent research assistance. We also thank Angela C. Lyons and
Jeanne Hogarth and the seminar participants at the 2006 Federal Reserve Research Forum for
Closing the Wealth Gap: Building Assets among Low-Income Households.
_______________________________
R. Haynes-Bordas
Purdue Extension Marion County, 6640 Intech Blvd., Suite 120, Indianapolis, IN 46278-2012, USA
e-mail:[email protected]
D. E. Kiss
Department of Consumer Sciences and Retailing, Purdue University, 812 W. State Street, West Lafayette, IN
47907-1262, USA
e-mail: [email protected]
T. Yilmazer (Corresponding author)
Department of Consumer Sciences and Retailing, Purdue University, 812 W. State Street, West Lafayette, IN
47907-1262, USA
e-mail: [email protected]
Effectiveness of Financial Education on Financial Management Behavior and
Account Usage: Evidence from a ‘Second Chance’ Program
Abstract The Get Checking™ program is a “second chance” program that aims to provide
financial education to consumers who were reported to ChexSystems by a previous financial
institution for account abuse or mismanagement. Using data collected from Indiana
participants of the program, the first goal of this study is to investigate the success of the
program in impacting financial management behavior of the participants. The second goal is to
investigate the change of participants’ actual behavior in terms of account usage and asset-
building after the completion of the program. The findings show that the program was
successful in positively influencing the financial management behavior of Non-whites in terms
of recording transactions and communicating with financial institutions. Also, financial
management skills emphasized in the program, especially communicating with financial
institutions, have a significant positive effect on the actual behavior of the participants in terms
of obtaining a loan. Among the heterogeneous group of the unbanked, findings shed light on
the demographic groups, such as Non-whites and young adults, that could benefit the most
from this type of financial management education.
Keywords Account Usage, Asset-building, Financial Management Education, Get Checking™
Program, Unbanked
Effectiveness of Financial Education on Financial Management Behavior and
Account Usage: Evidence from a ‘Second Chance’ Program
The unbanked – those who rely on check cashing outlets and pawn shops for their
financial transactions – usually pay high costs for check cashing and electronic money
transfers (Barr 2004). They can also become victims of financial scams and predatory lending
for their short-term loan needs (Barr 2001; Seidman and Tescher 2003; Stegman and Faris
2003). In addition to addressing these safety and security issues, establishing banking
relationships is a key element in establishing financial stability and building financial assets
(Barr 2001; Barr and Sherraden 2005). Finally, unbanked households can improve their credit-
risk profiles and gain access to lower-cost sources of credit by joining the mainstream financial
system and improving relationships with financial institutions (Belsky and Calder 2004).
A number of organizations (National Endowment for Financial Education, Cooperative
Extension System, National Foundation for Consumer Credit, social service and community
based-organizations, and financial institutions) have developed programs to teach personal
financial education to low- and moderate-income households. In general, the objective of these
programs is to increase financial knowledge, positively impact motivation, and lead to changed
behavior. Vitt et. al (2000) was one of the first studies to catalog these various programs.
In addition, financial institutions and educators have specifically developed programs to
bring the unbanked into the mainstream financial services market. One example, Money Smart,
an adult financial education program developed by the Federal Deposit Insurance Corporation
(FDIC), aims to help low-and moderate-income individuals develop financial skills and
banking relationships.
1
Through Money Smart, the FDIC was able to link many banks and
1
For more information on this program, seehttp://www.fdic.gov/consumers/consumer/moneysmart/
community groups and establish 1,200 private and public partnerships (Financial Literacy and
Education Commission 2006).
Lyons, Chang, and Scherpf (2005) investigated if – and how – financial education
translated into behavior change for low-income populations. They concluded that participants’
prior level of financial experience may be more important than the amount of education
received and that the financial management education may have the greatest impact on
financial behaviors that can be readily altered in the short run. Finally, they challenged
researchers to “focus less on outcomes tied to individuals’ financial situations and more on
whether individuals are able to make sound financial decisions regardless of their financial
situation” (p. 41). In addition, Xiao et al. (2004) investigated the changes in the behavior of
participants of Money 2000
TM
.
This program is unique as it was developed based on the
Transtheoretical Model of Change, which is a framework that has been widely used to study
health-related behaviors, like quitting smoking, and has established that the efficacy of
treatment interventions increases with a person’s readiness for action. Their findings suggested
that the same principle can be applied to financial management behavior.
Although researchers have begun analyzing what is known about effective financial
education and the impact of financial education programs on consumer skills and behaviors,
there is still much to learn (Borden et al. 2008; Collins 2007; Fox et al. 2005; Fry et al. 2008);
Hogarth 2006; Lyons et al. 2006; Peng et al. 2007). The effectiveness of the programs in
improving unbanked consumers’ financial management skills, influencing their attitudes
towards financial institutions, and changing their actual behavior in terms of account usage and
asset-building, has not been well documented. In addition, little is known about which sub-
groups of the unbanked population benefit most from these educational programs.
The goal of this study is to document the role of financial management education in
impacting financial management behavior and affecting the asset-building behavior of
2
consumers who once had a checking account but had it closed by a financial institution. This
study uses data collected from Indiana participants of the Get Checking™ program. This
program is a “second chance” program that aims to provide financial education to consumers
who were reported to ChexSystems by a previous financial institution for account abuse or
mismanagement.
Get Checking™ was developed in 1998 in Milwaukee, Wisconsin when a University of
Wisconsin Extension, Cooperative Extension educator and staff from financial institutions and
non-profit educational organizations collaborated to create the program. As a “second chance”
program, it emphasizes financial education, restitution if money is owed to a previous financial
institution, and the opportunity to open a checking or savings account upon completion of the
program. In 2001, eFunds Corporation became a national partner. Across the country, financial
institution partners market the program to consumers and open deposit accounts for program
graduates. Educational partners teach the curriculum.
The content of the course includes an introduction to the Get Checking Program
TM
and a
discussion of the importance of choosing an account that is right for the participant. It includes
teaching participants how to manage a checking account. Finally, the basics of financial
planning, and the importance of account ownership and credit rating are emphasized.
Consumers register and pay a fee for the six-hour class. When they complete all class
requirements, including successfully passing a quiz, they earn a certificate. After paying
restitution to any financial institutions they owe money to, they present the certificate to a
participating financial institution and are able to open an account.
2
A follow-up survey was mailed to all central Indiana participants who completed the
program in 2003, 2004, and 2005. The survey included questions about the checking, savings,
and other types of accounts that the participants had opened since their completion of the
2
Seehttp://www.getchecking.org/ for more information about the program.
3
program. Using data from the follow-up survey, the first goal of this study was to analyze the
success of the program in impacting the financial management behavior of the participants. In
particular, this study investigated whether demographic characteristics of the participant have a
varying effect on recording transactions, reconciling bank statements, budgeting, and
communicating with financial institutions. The second goal of this study was to investigate the
actual behavior of participants in terms of account usage and asset-building since the
completion of the program. Specifically, this study investigated the effect of financial
management skills emphasized in the program, especially communicating with financial
institutions, on the likelihood of opening a savings, an asset, or a loan account.
Determining the effectiveness of financial education programs targeted towards low- and
moderate-income households is difficult for a number of reasons. First, the effects of the
program on the participant’s behavior may be very different at the end of the program
compared to six months after it ends. Therefore, the point in time when the impact of financial
management education is evaluated plays a role when investigating whether or not a program
improves financial literacy and changes behaviors related to financial management. In this
study, the interval between completing the program and responding to the survey was long
enough to provide a reasonable indication of actual behavior change by the respondents.
Second, selection issues create difficulty in measuring the effectiveness of the program.
If the participants who enroll in the program are individuals who are already self-motivated to
change their financial practices, any changes in the observed behavior could be a result of the
motivation and aptitude of the participant, not the impact of the financial management course
(Caskey 2006). Since the financial institution partners require all participants to provide the
Get Checking™ certificate of completion as a requirement to opening a checking or savings
account, in this study the selection issue plays a very small role.
4
In the next section, the literature describing what is known about the unbanked and how
they could benefit from participating in the financial services market is reviewed. Then,
characteristics of the unbanked who once had a checking account are analyzed. Because such
an analysis has not been previously published, it is included here in order to place the results
of this study in context. Following that, the data and methods of this study are described.
Finally, the results are discussed and limitations, implications, and conclusions are presented.
Literature Review
What Do We Know About the Unbanked?
Despite the challenges of being unbanked, according to the 2004 Survey of Consumer
Finances, the proportion of families who did not do business with a financial institution did not
change much from 2001 to 2004 (Bucks et al. 2006). In 2004, 8.7% of American families did
not do business with a financial institution and were considered unbanked. Moreover, 10.6%
of families did not have a checking account in 2004, a slight decrease from 12.7% in 2001
(Bucks et al. 2006).
The unbanked community is not a homogenous group in terms of their financial and
demographic characteristics. Data from local and national surveys indicate that households
without banking relationships tend to have lower income and net worth; are younger,
unemployed, and less educated; are headed by Blacks, Hispanics, and single females; and rent,
rather than own, their residence (Berry 2004; Dunham 2001; Hogarth et al. 2003; Hogarth et
al. 2005; Paulson and Rhine 2008).
There are a variety of reasons cited by the unbanked for not having a checking account.
The Survey of Consumer Finances (SCF) asks respondents that do not have a checking account
to give a reason for not having an account. Commonly reported reasons include not writing
enough checks to make account ownership worthwhile, not having enough money, and not
liking to deal with banks (Bucks et al. 2006). Recent studies have discussed reasons why low-
5
income consumers choose to remain unbanked and use alternative financial service providers
(Belsky and Calder 2004; Berry 2004; Dunham 2001). A number of different reasons are
mentioned in these studies. First, consumers may prefer to use alternative providers for
convenience because these providers are more likely than banks to be open outside traditional
business hours and have staff that speak their language. Second, some consumers mistrust
banks due to past negative experience in the United States or another country. Third, low-
income consumers may not have a bank account because they do not perceive that they have
surplus cash to save. Fourth, poor credit scores prevent some households from having a
checking or saving account.
How Could the Unbanked Benefit from Participating in the Financial Services Market?
Consumers who use financial service centers to cash checks usually pay high fees for
these services. To cash a check, a consumer usually pays between 2% and 3.5% of the face
value of the check (Financial Service Centers of America n.d.).
3
These financial service
centers also offer services including money orders, money wire transfers, automatic teller
machine access, government benefit and payroll payments, payday loans, and electronic tax
preparation. According to the Survey of Non-Bank Institutions conducted in four markets
(Atlanta, Boston, San Antonio and San Diego) in December, 1999 and January, 2000, the
average fee for check cashing was between $4.10 and $6.68, and the average fee for money
orders was between $0.40 and $0.61 (Bachelder and Ditzion 2000). In 28 states, check cashing
fees are regulated. In Indiana, for example, a financial service center cannot charge check
cashing fees in excess of $5.00 or 10% of the face amount of the check (Indiana Department
of Financial Institutions n.d.). In a survey of low-income neighborhoods in New York City and
Los Angeles, those who used check cashing outlets to cash their checks paid $3.38 on averag
e
3
The industry reports that in approximately 11,000 financial service centers, 180 million checks totaling $55
billion are processed annually.
6
per ch
s
wrence (2001) reported that approximately 35% of the customers
surve
k
an
ry lending include targeting vulnerable populations (less educated and
unban
to
eck cashed and those who purchased money orders paid $1 for a money order (Dunham
2001).
Those that have a poor credit history or lack credit cards may use payday lenders for their
short terms credit needs. Payday loans carry high implicit annual interest rates that result from
an approximately $15 fee for each $100 borrowed. For an average size loan of $300, the fee for
a two-week loan translates into an APR of 380% (Stegman and Faris 2003). Payday lenders
extended 26 million to 47 million loans in 2000, totaling over $8 billion to 14 billion (Stegman
and Faris 2003).
Due to the high cost and the short term, many borrowers cannot repay their
original loan by their next payday and renew the loan by paying another fee. These borrower
take out payday loans repeatedly throughout the year and get caught in a “debt trap.” Using
data from a national survey conducted by the Community Financial Services Association of
America, Elliehaussen and La
yed renewed their payday loan one to four times; 20% of customers renewed their loans
nine or more times in a year.
Subprime lending businesses target those who cannot secure credit at prime rates. There
has been an increase in the size of the industry as home purchase and home refinance loans by
subprime lenders increased by 760% and 890%, respectively, from 1993 to 1999 (Immergluc
and Wiles 1999). Subprime firms typically charge borrowers higher fees and interest rates th
“prime” lenders and demonstrate predatory lending behavior (Immergluck and Wiles 1999).
Examples of predato
ked), packing loans with unnecessary fees, and concealing the true cost of financing with
balloon payments.
Increasing asset accumulation among low- and moderate-income households is critical
lowering the dependence on high-cost short-term credit, decreasing the risk of financial stress
due to income loss or unexpected expenses, and improving the prospects for asset-building
7
through homeownership. Without a checking or savings account, however, low- and moderate
income households face barriers to asset accumulation (Barr 2001; Barr and Sherraden 2005).
Dunham (2001) showed that across different income groups, individuals with bank acco
were more likely to save regularly than unbanked individuals. Gale and Carney (2001) found
that low-income households with bank accounts are more likely to have other types of
financial assets than households without ban
-
unts
k accounts.
4
Thus, access to a bank account can be
an im
s
nd
sured by
the lengt e
borro
e
ristics of those who once
had a
portant initial point for understanding how the mainstream financial services market
works and for obtaining financial accounts.
Without a transaction account, it is almost impossible to establish a credit history and
qualify for a loan. Households, especially low-income households, cannot obtain costly asset
such as homes and autos without credit. Hogarth and O'Donnell (1999) showed that owning a
bank account is a more significant factor than household net worth, income, or education in
predicting whether a household holds a mortgage and auto loan. In addition, Chakravarty a
Yilmazer (2005) showed that the relationship between the borrower and lender, mea
h of relationship and the number of transaction accounts, significantly affects th
wer's decision to apply for a loan and the lender's approval/rejection decision.
Characteristics of Unbanked Consumers Who Once Had a Checking Account
According to the 2004 SCF, 10.6% of families did not have a checking account and more
than 50% of households without a checking account reported that they once had a checking
account (Bucks et al. 2006). Therefore, their relationship with a financial institution must have
ended by either their own choice or the financial institution’s decision. Using data from th
2004 SCF, this section investigates whether the demographic characte
checking account and their reported reasons for not having a checking account are
significantly different than those who never had a checking account.
4
Gale and Carney (1998) acknowledge that having a bank account is endogenous to the asset holding behavior.
Therefore, their findings do not imply that giving a bank account to a household would cause the household’s
financial asset ownership to rise.
8
The SCF, sponsored by the Federal Reserve Board in cooperation with the Department o
the Treasury, is a triennial survey of U.S. families’ financial portfolios and includes detaile
information on families’ balance sheets, use of financial services, and demographics. Table 1
describes the demographic characteristics of those who did, and those who did not, have a
checking account in the 2004 SCF. Compared to ho
f
d
useholds with a checking account, those
witho
e
r,
ehold
e never had a checking account. Households that once had a checking
accou d a
g
is of
d
not
ut a checking account are more likely to be headed by younger and single adults, Blacks
and Hispanics, and have lower household income.
As the last two columns of Table 1 show, there are significant differences between thos
who once had a checking account and those who never had a checking account. In particula
race and income appear to play a significant role in whether or not the unbanked hous
once had a checking account. Those households that once had a checking account are less
likely to be headed by Hispanics and more likely to be headed by Whites than those
households that hav
nt are also likely to have higher household income than households that have never ha
checking account.
As shown in Table 1, consumers have a variety of reasons for not having a checkin
account. The reasons can be grouped into categories. For example, the cost/benefit analys
having an account (not writing enough checks to make having an account worthwhile),
financial management education needs (not being able to manage or balance a checking
account, problems related to credit history, and not liking to deal with banks), and lack of
financial resources, are types of reasons consumers gave for not having a checking account.
Reasons related to the cost/benefit analysis of having an account are less common and reasons
related to financial management education needs are more common among those who once ha
a checking account. For example, 24.4% of those who once had a checking account report
writing enough checks to make it worthwhile to have an account compared to 32.1% of those
9
who never had a checking account. In contrast, 10.5% of those who once had a checking
account report not being able to manage or balance a checking account, compared to 1.5% of
those who never had a checking account. These summary statistics provide evidence that for
those who once had a checking account, the benefit of having an account is greater than the
cost, however, financial management education issues are more li
kely to prevent them from
having and maintaining an account. Therefore, providing financial education to this group of
consu
n program, the
motiv ho
p
nse
rate t
t
whether they opened a checking and savings account since the completion of the program and
mers should have a positive effect on their account usage.
Data and Methods
From 2003-2005, 1,483 central Indiana consumers earned certificates through the Get
Checking
TM
program and responded to the end of session evaluations after three and six hours
of instruction. Most participants were referred to the program by the participating financial
institutions and some participants reported that they learned about the program from the media
or through friends. Therefore, as with any other financial management educatio
ation level of the participants can be assumed to be higher than those consumers w
were also reported to ChexSystems but did not register to attend the program.
In March 2006, a letter announcing the follow-up survey was mailed to 1,400 Get
Checking
TM
program participants that provided a complete address on their registration forms
and 212 of these letters were returned without a forwarding address. In April 2006, a follow-u
survey was mailed to 1,188 participants, and the number of surveys returned as undeliverable
was 63. Finally, 161 program participants responded to the follow-up survey. The respo
o the follow-up survey was 14.2%, and data used in this study includes information from
160 surveys. No incentives were used to enhance the response to the follow-up survey.
The follow-up survey included questions about the demographics of the respondents and
the financial management skills they acquired during the program. It included questions abou
10
whether these accounts were still open. It also asked whether they had opened other types of
asset and loan accounts since the completion of the program. The survey also asked what they
do di
e
hether
unt since the completion of the program and/or whether the
accou
-
ed
two
hecking
TM
program were male,
only
re no
of respondents lived in single-person households while nearly half of the respondents reported
fferently, if anything, to manage their accounts since the completion of the program.
Most of the respondents answered the questions on the survey completely. There wer
two missing values for the respondent’s race and two missing values for the respondent’s
gender. Respondents with missing demographic characteristics were not included in the
analyses that utilize these characteristics. In addition, six respondents did not indicate w
they opened a savings acco
nts were still open.
Table 2 presents the demographic and socioeconomic characteristics of the participants in
both the end of session evaluations and the follow-up survey. The information from the follow
up was not linked to the end of session evaluations. First, the analysis assesses whether those
who responded to the follow-up survey were significantly different than those who complet
the program. The majority of survey respondents in both surveys were between 25 and 44
years old. However, the age distribution of the respondents varied significantly across the
surveys and the average age of the participants in the follow-up survey was significantly
higher.
5
In addition, the distribution of gender varied significantly between the participants of
the two surveys. While 44.1% of the participants of the Get C
36.7% of the follow-up survey respondents were male.
In terms of race, household size, and household income of the participants, there we
significant differences between the respondents of the end of session evaluations and the
follow-up survey. In the follow-up survey, the proportion of Non-whites was about 46.8%.
Almost 90% of the Non-whites in the follow-up survey were Black respondents. Nearly 30%
5
The follow-up surveys were mailed to the participants 1-3 years after they completed the program.
11
that their household consisted of two or three people. Nearly 20% had household incomes less
than $18,871 and more than 30% had household incomes between $18,871 and $31,450.
In addition, Table 2 shows how the demographic characteristics of the respondents of the
follow-up survey varied by race. Non-whites were significantly different than Whites in terms
of gender and household size. In particular, 21.9% of Non-white respondents were male,
compared to 50.6% of White respondents. Also, 20.2% of Non-whites lived in single-person
households, compared to 36.9% of Whites. Non-white respondents were not significantly
different than White respondents in terms of their age and income.
The end of session evaluation of the Get Checking
TM
program also asked the participants
how confident they felt about several financial management activities as a result of the
program. The majority of participants indicated the highest possible level of confidence about
each of these behaviors. For example, 84.2% of participants indicated they felt very confident
about maintaining a check register and 80.2% of participants indicated they were very
confident about reconciling their checkbook registers with a bank statement. Respondents were
relatively less confident about talking to a financial institution about savings goals or credit
needs. Nevertheless, 75.7% and 74.4% of participants indicated they were very confident about
talking to a financial institution representative about savings goals and credit needs,
respectively. The goal of the follow-up survey was to measure actual behavior change in terms
of account usage and asset-building. Therefore, questions on level of confidence about
financial management activities were not included in the follow-up survey.
Table 3 presents the summary statistics for changes in actual behavior in terms of account
usage and changes in financial management behavior since the completion of the Get
Checking
TM
program. In terms of account usage, 97.5% of the respondents opened a checking
account and 90.6% still had the checking account open at the time the survey was conducted.
At the same time, 56.4% of the respondents opened a savings account, and the retention rate
12
was quite high, as 54.3% of the sample reported still having the savings account open at the
time the survey was conducted. However, the proportion of respondents who opened another
asset account or a loan account was relatively lower than the proportion of respondents who
opened a checking or savings account. For example, 15.0% of respondents opened an asset
account. Most frequently opened asset accounts were retirement savings (6.2%) and
certificates of deposit (5.0%). In terms of loan accounts, 16.2% reported having opened a loan
account. The most frequently opened loan accounts were auto (9.3%) and mortgage loans
(7.5%).
The table in the Appendix shows that the age and race of the respondent had a significant
effect on the account usage. In the sample used in this study, all of those respondents below
age 25 opened a checking account after earning their certificate. In addition, those below age
25 were more likely to open an asset or a loan account. Specifically, they were more likely to
obtain an auto loan. After earning their certificate, Non-whites were significantly more likely
to open an asset account and less likely to open a loan account.
The follow-up survey included a question that asked whether or not respondents owed
any money to the financial institution before they participated in the Get Checking
TM
program
and how long it took them to repay their debts to the financial institutions. As reported in Table
3, 61.2% of the respondents owed money to a financial institution before they earned their
certificate. While 40.0% of the respondents repaid the amount they owed within one month
after earning the certificate, 5.6% made restitution within two or three months. In addition, at
the time the follow-up survey was conducted, 10.6% reported still owing money to a financial
institution and being in the process of making restitution.
In the follow-up survey, respondents were asked to indicate changes in behavior with
respect to managing their finances after completing the program. Table 3 presents the
percentage of respondents that reported a positive change in financial management behavior.
13
These changes were analyzed in four categories: recording transactions (keeping an up-to-date
check register or a record of ATM or debit card transactions), reconciling bank statements with
check register, planning a budget (working to achieve a written financial goal or managing
income and expenses to meet financial goals or using a written spending plan) and, finally,
communicating with financial institutions. Overall, 75.0% reported recording transactions,
53.1% reported reconciling bank statements with their check register, 67.5% reported planning
a budget, and 44.3% reported communicating with the financial institution since the
completion of the program.
Table 3 presents the account usage and financial management behavior by the
communication behavior of the respondents. All of the respondents who reported
communicating with financial institutions opened a checking account. Respondents who
reported communicating with financial institutions were more likely than others to open an
asset account and to obtain a loan. Those who reported communicating with financial
institutions were also more likely to be keeping an up-to-date check register or a record of
ATM or debit card transactions; reconciling bank statements with their check register; and
working to achieve a written financial goal, or managing income and expenses to meet
financial goals, or using a written spending plan.
In the follow-up survey, respondents were asked to indicate their experiences with
cashing checks and buying money orders. Almost 35% of survey participants reported that
they had no expenditures for cashing paychecks or buying money orders. This finding is
consistent with Dunham (2001). In her study, 27% of the unbanked survey population in New
York and Los Angeles did not incur any costs for check cashing and money orders. The
analysis of Get Checking™ participants indicated that for those who reported non-zero
expenditures, the median and average monthly costs of cashing paychecks and buying money
orders were $16.00 and $23.48, respectively. On average, Get Checking™ participants
14
reported cashing 2.86 paychecks and buying 5.48 money orders per month.
There were no
significant differences in the cost and the number of checks and money orders by age.
However, there were significant differences in the cost of cashing paychecks and buying
money orders by the race of the respondent. For example, while White respondents paid on
average $28.59 per month for these services, Non-white respondents paid only $15.91 per
month. The median values for the cost of these services for Whites and Non-whites were
$21.75 and $11.48, respectively. In addition, White respondents purchased a greater number
of money orders than their Non-white counterparts (6.35 vs. 4.58).
Results
Changes in Financial Management Behavior
Changes in financial management behavior analyzed in this study, which were recording
transactions, reconciling bank statements with check register, planning a budget, and
communicating with financial institutions, were recorded as dichotomous variables. Therefore,
probit models were used to analyze the role of demographic and socioeconomic characteristics
on reported changes in these financial management behaviors. For each category of financial
management, it was expected that younger respondents and those with higher household
income would be more likely to change their financial management behavior.
Table 4 presents the results of probit regression for the determinants of changes in
financial management behavior since the completion of the Get Checking
TM
program. Table 4
also presents the marginal effects for the reference group of respondents which consists of
White females below age 25 years who live in one person households and who have household
income of $18,871- $31,450. The base probability of recording, reconciling, budgeting, and
communicating for the reference group was 63.0%, 29.2%, 42.7%, and 35.0%, respectively.
The marginal effect of each dummy variable represents the change in the base probability of
the reference group, all else being equal. For example, the base probability of recording is
15
equal to ?(0.332), where ?() is the standard normal cumulative distribution and 0.332 is the
intercept of the regression for recording. The estimated coefficient of Male is 0.522, and the
marginal effect of Male is ?(0.332+0.522)- ?(0.332), which is equal to 18.3%.
The results show that age had limited explanatory power on changes in financial
management behavior with respect to reconciling bank statements with check registers and
planning a budget. Respondents who were between age 45 and 55 and respondents over age 55
were more likely to reconcile checkbook registers and bank statements than those below age
25. Similarly, respondents between age 25 and 34 were more likely to plan a budget than those
below age 25. Gender had some limited influence on the change in financial management
behavior, and male respondents were more likely to record transactions than females.
Non-white respondents were more likely to record transactions and communicate with
financial institutions than White respondents. Finally, income had some impact on the change
in financial management behavior. Respondents who had household income over $50,321 were
more likely to reconcile bank statements with their check register and plan a budget than
respondents who had household income between $18,871 and $31,450.
Changes in Behavior in terms of Account Usage and Asset Building
Following the estimation of the changes in financial management behavior, three sets of
probit models were estimated to determine the factors influencing the decision to open a
savings, an asset, and a loan account, which were recorded as dichotomous variables. Each
model was estimated in three steps: first including only demographic factors as the explanatory
variables (Model I), second, adding the four dichotomous variables that measure changes in
financial management behavior (recording transactions, reconciling bank statements with
check register, planning a budget, and communicating with financial institutions) to Model I
(Model II), and third, adding an interaction term between race and communicating with
financial institutions variable to Model II (Model III).
16
Model II was estimated to capture the effect of changes in financial management
behavior on personal finances.
6
One of the explanations in the literature for not having a
banking relationship is that unbanked consumers mistrust banks due to their negative
experiences, and Non-whites are more likely to have negative experiences with banks than
Whites (Longhofer and Peters 2005). The data from the 2004 SCF confirms that claim. The
percentage of households in Table 1 citing not liking to deal with banks as the reason for not
having an account was the second highest reason (21.9%). In addition, previous studies show
that good banking relationships increase the probability of being approved for a loan and
accumulating assets (Chakravarty and Yilmazer 2005). Model III was estimated to investigate
whether improving the relationship with a financial institution affects the account usage of
Whites differently than the account usage of Non-whites.
Table 5 presents the estimation results of the probit models and marginal effects for the
likelihood of opening a savings account since the completion of the Get Checking
TM
program.
Marginal effects of the dummy variables are calculated using the same method as the marginal
effects presented in Table 4. In all three models that were estimated, probability of opening a
savings account was significantly affected by household size and income. However, other
demographic factors such as age, race, and gender of the respondent did not have a significant
impact on the likelihood of opening a savings account. Respondents who lived in households
with three or more people were less likely to open a savings account than those who lived in
one person households. Compared to respondents who had household income between $18,871
and $31,450, respondents who had household income less than $18,870 were less likely, and
respondents who had household income above $50,321 were more likely to open a savings
account. Changes in financial management behavior did not significantly affect the decision to
6
This model estimates the effect of financial management skills that were obtained during the course. If the
participants were already recording transactions, reconciling bank statements with their check register, planning a
budget, or communicating with financial institutions before they had participated the Get Checking
TM
program
and did not report a change of behavior, the estimated coefficients on these variables would underestimate the true
effect of these variables on opening a savings, asset, or loan account.
17
open a savings account. In addition, no significant differences in the impact of communicating
with financial institutions between Whites and Non-whites on the decision to open a savings
account were observed.
As presented in Table 6, age and race significantly influenced the probability of opening
an asset account since the completion of the Get Checking
TM
program. In comparison to those
below age 25, those between 25-34 and those above 55 were less likely to open an asset
account (Model II). Non-white respondents were more likely to open an asset account.
Changes in communication with financial institutions had a significant positive effect on the
decision to open an asset account.
In Table 6, Model III reports that Non-white respondents who communicated with
financial institutions were more likely to open an asset account than Non-white respondents
who did not communicate with financial institutions.
7
Compared to Whites who communicated
with financial institutions, Non-whites who communicated with financial institutions were
more likely to open an asset account.
8
In terms of opening an asset account, Non-whites
seemed to benefit more from the financial management content emphasized during the
program.
Finally, the determinants of obtaining a loan since the completion of the Get Checking
TM
program were estimated. Results are presented in Table 7. Similar to the findings for asset
accounts, respondents below age 25 were more likely to obtain a loan than older respondents.
Male respondents were also more likely to have a loan account than female respondents.
Unlike asset accounts, the probability of having a loan account was lower for Non-white
7
The coefficient estimate of Non-whites who communicated with financial institutions is the sum of the
coefficients of Non-white (0.279), Behaviour_Communication (-0.154) and Non-white*
Behaviour_Communication (1.319) and equals to 1.444 (p-value
The Get Checking™ program is a “second chance” program that aims to provide financial education to consumers who were reported to ChexSystems by a previous financial institution for account abuse or mismanagement
Effectiveness of Financial Education on Financial Management Behavior and
Account Usage: Evidence from a ‘Second Chance’ Program
Rebecca Haynes-Bordas, D. E. Kiss and Tansel Yilmazer
Acknowledgements We thank Bharathi Nagarajan for her assistance with the collection of the
data and Patryk Babiarz for excellent research assistance. We also thank Angela C. Lyons and
Jeanne Hogarth and the seminar participants at the 2006 Federal Reserve Research Forum for
Closing the Wealth Gap: Building Assets among Low-Income Households.
_______________________________
R. Haynes-Bordas
Purdue Extension Marion County, 6640 Intech Blvd., Suite 120, Indianapolis, IN 46278-2012, USA
e-mail:[email protected]
D. E. Kiss
Department of Consumer Sciences and Retailing, Purdue University, 812 W. State Street, West Lafayette, IN
47907-1262, USA
e-mail: [email protected]
T. Yilmazer (Corresponding author)
Department of Consumer Sciences and Retailing, Purdue University, 812 W. State Street, West Lafayette, IN
47907-1262, USA
e-mail: [email protected]
Effectiveness of Financial Education on Financial Management Behavior and
Account Usage: Evidence from a ‘Second Chance’ Program
Abstract The Get Checking™ program is a “second chance” program that aims to provide
financial education to consumers who were reported to ChexSystems by a previous financial
institution for account abuse or mismanagement. Using data collected from Indiana
participants of the program, the first goal of this study is to investigate the success of the
program in impacting financial management behavior of the participants. The second goal is to
investigate the change of participants’ actual behavior in terms of account usage and asset-
building after the completion of the program. The findings show that the program was
successful in positively influencing the financial management behavior of Non-whites in terms
of recording transactions and communicating with financial institutions. Also, financial
management skills emphasized in the program, especially communicating with financial
institutions, have a significant positive effect on the actual behavior of the participants in terms
of obtaining a loan. Among the heterogeneous group of the unbanked, findings shed light on
the demographic groups, such as Non-whites and young adults, that could benefit the most
from this type of financial management education.
Keywords Account Usage, Asset-building, Financial Management Education, Get Checking™
Program, Unbanked
Effectiveness of Financial Education on Financial Management Behavior and
Account Usage: Evidence from a ‘Second Chance’ Program
The unbanked – those who rely on check cashing outlets and pawn shops for their
financial transactions – usually pay high costs for check cashing and electronic money
transfers (Barr 2004). They can also become victims of financial scams and predatory lending
for their short-term loan needs (Barr 2001; Seidman and Tescher 2003; Stegman and Faris
2003). In addition to addressing these safety and security issues, establishing banking
relationships is a key element in establishing financial stability and building financial assets
(Barr 2001; Barr and Sherraden 2005). Finally, unbanked households can improve their credit-
risk profiles and gain access to lower-cost sources of credit by joining the mainstream financial
system and improving relationships with financial institutions (Belsky and Calder 2004).
A number of organizations (National Endowment for Financial Education, Cooperative
Extension System, National Foundation for Consumer Credit, social service and community
based-organizations, and financial institutions) have developed programs to teach personal
financial education to low- and moderate-income households. In general, the objective of these
programs is to increase financial knowledge, positively impact motivation, and lead to changed
behavior. Vitt et. al (2000) was one of the first studies to catalog these various programs.
In addition, financial institutions and educators have specifically developed programs to
bring the unbanked into the mainstream financial services market. One example, Money Smart,
an adult financial education program developed by the Federal Deposit Insurance Corporation
(FDIC), aims to help low-and moderate-income individuals develop financial skills and
banking relationships.
1
Through Money Smart, the FDIC was able to link many banks and
1
For more information on this program, seehttp://www.fdic.gov/consumers/consumer/moneysmart/
community groups and establish 1,200 private and public partnerships (Financial Literacy and
Education Commission 2006).
Lyons, Chang, and Scherpf (2005) investigated if – and how – financial education
translated into behavior change for low-income populations. They concluded that participants’
prior level of financial experience may be more important than the amount of education
received and that the financial management education may have the greatest impact on
financial behaviors that can be readily altered in the short run. Finally, they challenged
researchers to “focus less on outcomes tied to individuals’ financial situations and more on
whether individuals are able to make sound financial decisions regardless of their financial
situation” (p. 41). In addition, Xiao et al. (2004) investigated the changes in the behavior of
participants of Money 2000
TM
.
This program is unique as it was developed based on the
Transtheoretical Model of Change, which is a framework that has been widely used to study
health-related behaviors, like quitting smoking, and has established that the efficacy of
treatment interventions increases with a person’s readiness for action. Their findings suggested
that the same principle can be applied to financial management behavior.
Although researchers have begun analyzing what is known about effective financial
education and the impact of financial education programs on consumer skills and behaviors,
there is still much to learn (Borden et al. 2008; Collins 2007; Fox et al. 2005; Fry et al. 2008);
Hogarth 2006; Lyons et al. 2006; Peng et al. 2007). The effectiveness of the programs in
improving unbanked consumers’ financial management skills, influencing their attitudes
towards financial institutions, and changing their actual behavior in terms of account usage and
asset-building, has not been well documented. In addition, little is known about which sub-
groups of the unbanked population benefit most from these educational programs.
The goal of this study is to document the role of financial management education in
impacting financial management behavior and affecting the asset-building behavior of
2
consumers who once had a checking account but had it closed by a financial institution. This
study uses data collected from Indiana participants of the Get Checking™ program. This
program is a “second chance” program that aims to provide financial education to consumers
who were reported to ChexSystems by a previous financial institution for account abuse or
mismanagement.
Get Checking™ was developed in 1998 in Milwaukee, Wisconsin when a University of
Wisconsin Extension, Cooperative Extension educator and staff from financial institutions and
non-profit educational organizations collaborated to create the program. As a “second chance”
program, it emphasizes financial education, restitution if money is owed to a previous financial
institution, and the opportunity to open a checking or savings account upon completion of the
program. In 2001, eFunds Corporation became a national partner. Across the country, financial
institution partners market the program to consumers and open deposit accounts for program
graduates. Educational partners teach the curriculum.
The content of the course includes an introduction to the Get Checking Program
TM
and a
discussion of the importance of choosing an account that is right for the participant. It includes
teaching participants how to manage a checking account. Finally, the basics of financial
planning, and the importance of account ownership and credit rating are emphasized.
Consumers register and pay a fee for the six-hour class. When they complete all class
requirements, including successfully passing a quiz, they earn a certificate. After paying
restitution to any financial institutions they owe money to, they present the certificate to a
participating financial institution and are able to open an account.
2
A follow-up survey was mailed to all central Indiana participants who completed the
program in 2003, 2004, and 2005. The survey included questions about the checking, savings,
and other types of accounts that the participants had opened since their completion of the
2
Seehttp://www.getchecking.org/ for more information about the program.
3
program. Using data from the follow-up survey, the first goal of this study was to analyze the
success of the program in impacting the financial management behavior of the participants. In
particular, this study investigated whether demographic characteristics of the participant have a
varying effect on recording transactions, reconciling bank statements, budgeting, and
communicating with financial institutions. The second goal of this study was to investigate the
actual behavior of participants in terms of account usage and asset-building since the
completion of the program. Specifically, this study investigated the effect of financial
management skills emphasized in the program, especially communicating with financial
institutions, on the likelihood of opening a savings, an asset, or a loan account.
Determining the effectiveness of financial education programs targeted towards low- and
moderate-income households is difficult for a number of reasons. First, the effects of the
program on the participant’s behavior may be very different at the end of the program
compared to six months after it ends. Therefore, the point in time when the impact of financial
management education is evaluated plays a role when investigating whether or not a program
improves financial literacy and changes behaviors related to financial management. In this
study, the interval between completing the program and responding to the survey was long
enough to provide a reasonable indication of actual behavior change by the respondents.
Second, selection issues create difficulty in measuring the effectiveness of the program.
If the participants who enroll in the program are individuals who are already self-motivated to
change their financial practices, any changes in the observed behavior could be a result of the
motivation and aptitude of the participant, not the impact of the financial management course
(Caskey 2006). Since the financial institution partners require all participants to provide the
Get Checking™ certificate of completion as a requirement to opening a checking or savings
account, in this study the selection issue plays a very small role.
4
In the next section, the literature describing what is known about the unbanked and how
they could benefit from participating in the financial services market is reviewed. Then,
characteristics of the unbanked who once had a checking account are analyzed. Because such
an analysis has not been previously published, it is included here in order to place the results
of this study in context. Following that, the data and methods of this study are described.
Finally, the results are discussed and limitations, implications, and conclusions are presented.
Literature Review
What Do We Know About the Unbanked?
Despite the challenges of being unbanked, according to the 2004 Survey of Consumer
Finances, the proportion of families who did not do business with a financial institution did not
change much from 2001 to 2004 (Bucks et al. 2006). In 2004, 8.7% of American families did
not do business with a financial institution and were considered unbanked. Moreover, 10.6%
of families did not have a checking account in 2004, a slight decrease from 12.7% in 2001
(Bucks et al. 2006).
The unbanked community is not a homogenous group in terms of their financial and
demographic characteristics. Data from local and national surveys indicate that households
without banking relationships tend to have lower income and net worth; are younger,
unemployed, and less educated; are headed by Blacks, Hispanics, and single females; and rent,
rather than own, their residence (Berry 2004; Dunham 2001; Hogarth et al. 2003; Hogarth et
al. 2005; Paulson and Rhine 2008).
There are a variety of reasons cited by the unbanked for not having a checking account.
The Survey of Consumer Finances (SCF) asks respondents that do not have a checking account
to give a reason for not having an account. Commonly reported reasons include not writing
enough checks to make account ownership worthwhile, not having enough money, and not
liking to deal with banks (Bucks et al. 2006). Recent studies have discussed reasons why low-
5
income consumers choose to remain unbanked and use alternative financial service providers
(Belsky and Calder 2004; Berry 2004; Dunham 2001). A number of different reasons are
mentioned in these studies. First, consumers may prefer to use alternative providers for
convenience because these providers are more likely than banks to be open outside traditional
business hours and have staff that speak their language. Second, some consumers mistrust
banks due to past negative experience in the United States or another country. Third, low-
income consumers may not have a bank account because they do not perceive that they have
surplus cash to save. Fourth, poor credit scores prevent some households from having a
checking or saving account.
How Could the Unbanked Benefit from Participating in the Financial Services Market?
Consumers who use financial service centers to cash checks usually pay high fees for
these services. To cash a check, a consumer usually pays between 2% and 3.5% of the face
value of the check (Financial Service Centers of America n.d.).
3
These financial service
centers also offer services including money orders, money wire transfers, automatic teller
machine access, government benefit and payroll payments, payday loans, and electronic tax
preparation. According to the Survey of Non-Bank Institutions conducted in four markets
(Atlanta, Boston, San Antonio and San Diego) in December, 1999 and January, 2000, the
average fee for check cashing was between $4.10 and $6.68, and the average fee for money
orders was between $0.40 and $0.61 (Bachelder and Ditzion 2000). In 28 states, check cashing
fees are regulated. In Indiana, for example, a financial service center cannot charge check
cashing fees in excess of $5.00 or 10% of the face amount of the check (Indiana Department
of Financial Institutions n.d.). In a survey of low-income neighborhoods in New York City and
Los Angeles, those who used check cashing outlets to cash their checks paid $3.38 on averag
e
3
The industry reports that in approximately 11,000 financial service centers, 180 million checks totaling $55
billion are processed annually.
6
per ch
s
wrence (2001) reported that approximately 35% of the customers
surve
k
an
ry lending include targeting vulnerable populations (less educated and
unban
to
eck cashed and those who purchased money orders paid $1 for a money order (Dunham
2001).
Those that have a poor credit history or lack credit cards may use payday lenders for their
short terms credit needs. Payday loans carry high implicit annual interest rates that result from
an approximately $15 fee for each $100 borrowed. For an average size loan of $300, the fee for
a two-week loan translates into an APR of 380% (Stegman and Faris 2003). Payday lenders
extended 26 million to 47 million loans in 2000, totaling over $8 billion to 14 billion (Stegman
and Faris 2003).
Due to the high cost and the short term, many borrowers cannot repay their
original loan by their next payday and renew the loan by paying another fee. These borrower
take out payday loans repeatedly throughout the year and get caught in a “debt trap.” Using
data from a national survey conducted by the Community Financial Services Association of
America, Elliehaussen and La
yed renewed their payday loan one to four times; 20% of customers renewed their loans
nine or more times in a year.
Subprime lending businesses target those who cannot secure credit at prime rates. There
has been an increase in the size of the industry as home purchase and home refinance loans by
subprime lenders increased by 760% and 890%, respectively, from 1993 to 1999 (Immergluc
and Wiles 1999). Subprime firms typically charge borrowers higher fees and interest rates th
“prime” lenders and demonstrate predatory lending behavior (Immergluck and Wiles 1999).
Examples of predato
ked), packing loans with unnecessary fees, and concealing the true cost of financing with
balloon payments.
Increasing asset accumulation among low- and moderate-income households is critical
lowering the dependence on high-cost short-term credit, decreasing the risk of financial stress
due to income loss or unexpected expenses, and improving the prospects for asset-building
7
through homeownership. Without a checking or savings account, however, low- and moderate
income households face barriers to asset accumulation (Barr 2001; Barr and Sherraden 2005).
Dunham (2001) showed that across different income groups, individuals with bank acco
were more likely to save regularly than unbanked individuals. Gale and Carney (2001) found
that low-income households with bank accounts are more likely to have other types of
financial assets than households without ban
-
unts
k accounts.
4
Thus, access to a bank account can be
an im
s
nd
sured by
the lengt e
borro
e
ristics of those who once
had a
portant initial point for understanding how the mainstream financial services market
works and for obtaining financial accounts.
Without a transaction account, it is almost impossible to establish a credit history and
qualify for a loan. Households, especially low-income households, cannot obtain costly asset
such as homes and autos without credit. Hogarth and O'Donnell (1999) showed that owning a
bank account is a more significant factor than household net worth, income, or education in
predicting whether a household holds a mortgage and auto loan. In addition, Chakravarty a
Yilmazer (2005) showed that the relationship between the borrower and lender, mea
h of relationship and the number of transaction accounts, significantly affects th
wer's decision to apply for a loan and the lender's approval/rejection decision.
Characteristics of Unbanked Consumers Who Once Had a Checking Account
According to the 2004 SCF, 10.6% of families did not have a checking account and more
than 50% of households without a checking account reported that they once had a checking
account (Bucks et al. 2006). Therefore, their relationship with a financial institution must have
ended by either their own choice or the financial institution’s decision. Using data from th
2004 SCF, this section investigates whether the demographic characte
checking account and their reported reasons for not having a checking account are
significantly different than those who never had a checking account.
4
Gale and Carney (1998) acknowledge that having a bank account is endogenous to the asset holding behavior.
Therefore, their findings do not imply that giving a bank account to a household would cause the household’s
financial asset ownership to rise.
8
The SCF, sponsored by the Federal Reserve Board in cooperation with the Department o
the Treasury, is a triennial survey of U.S. families’ financial portfolios and includes detaile
information on families’ balance sheets, use of financial services, and demographics. Table 1
describes the demographic characteristics of those who did, and those who did not, have a
checking account in the 2004 SCF. Compared to ho
f
d
useholds with a checking account, those
witho
e
r,
ehold
e never had a checking account. Households that once had a checking
accou d a
g
is of
d
not
ut a checking account are more likely to be headed by younger and single adults, Blacks
and Hispanics, and have lower household income.
As the last two columns of Table 1 show, there are significant differences between thos
who once had a checking account and those who never had a checking account. In particula
race and income appear to play a significant role in whether or not the unbanked hous
once had a checking account. Those households that once had a checking account are less
likely to be headed by Hispanics and more likely to be headed by Whites than those
households that hav
nt are also likely to have higher household income than households that have never ha
checking account.
As shown in Table 1, consumers have a variety of reasons for not having a checkin
account. The reasons can be grouped into categories. For example, the cost/benefit analys
having an account (not writing enough checks to make having an account worthwhile),
financial management education needs (not being able to manage or balance a checking
account, problems related to credit history, and not liking to deal with banks), and lack of
financial resources, are types of reasons consumers gave for not having a checking account.
Reasons related to the cost/benefit analysis of having an account are less common and reasons
related to financial management education needs are more common among those who once ha
a checking account. For example, 24.4% of those who once had a checking account report
writing enough checks to make it worthwhile to have an account compared to 32.1% of those
9
who never had a checking account. In contrast, 10.5% of those who once had a checking
account report not being able to manage or balance a checking account, compared to 1.5% of
those who never had a checking account. These summary statistics provide evidence that for
those who once had a checking account, the benefit of having an account is greater than the
cost, however, financial management education issues are more li
kely to prevent them from
having and maintaining an account. Therefore, providing financial education to this group of
consu
n program, the
motiv ho
p
nse
rate t
t
whether they opened a checking and savings account since the completion of the program and
mers should have a positive effect on their account usage.
Data and Methods
From 2003-2005, 1,483 central Indiana consumers earned certificates through the Get
Checking
TM
program and responded to the end of session evaluations after three and six hours
of instruction. Most participants were referred to the program by the participating financial
institutions and some participants reported that they learned about the program from the media
or through friends. Therefore, as with any other financial management educatio
ation level of the participants can be assumed to be higher than those consumers w
were also reported to ChexSystems but did not register to attend the program.
In March 2006, a letter announcing the follow-up survey was mailed to 1,400 Get
Checking
TM
program participants that provided a complete address on their registration forms
and 212 of these letters were returned without a forwarding address. In April 2006, a follow-u
survey was mailed to 1,188 participants, and the number of surveys returned as undeliverable
was 63. Finally, 161 program participants responded to the follow-up survey. The respo
o the follow-up survey was 14.2%, and data used in this study includes information from
160 surveys. No incentives were used to enhance the response to the follow-up survey.
The follow-up survey included questions about the demographics of the respondents and
the financial management skills they acquired during the program. It included questions abou
10
whether these accounts were still open. It also asked whether they had opened other types of
asset and loan accounts since the completion of the program. The survey also asked what they
do di
e
hether
unt since the completion of the program and/or whether the
accou
-
ed
two
hecking
TM
program were male,
only
re no
of respondents lived in single-person households while nearly half of the respondents reported
fferently, if anything, to manage their accounts since the completion of the program.
Most of the respondents answered the questions on the survey completely. There wer
two missing values for the respondent’s race and two missing values for the respondent’s
gender. Respondents with missing demographic characteristics were not included in the
analyses that utilize these characteristics. In addition, six respondents did not indicate w
they opened a savings acco
nts were still open.
Table 2 presents the demographic and socioeconomic characteristics of the participants in
both the end of session evaluations and the follow-up survey. The information from the follow
up was not linked to the end of session evaluations. First, the analysis assesses whether those
who responded to the follow-up survey were significantly different than those who complet
the program. The majority of survey respondents in both surveys were between 25 and 44
years old. However, the age distribution of the respondents varied significantly across the
surveys and the average age of the participants in the follow-up survey was significantly
higher.
5
In addition, the distribution of gender varied significantly between the participants of
the two surveys. While 44.1% of the participants of the Get C
36.7% of the follow-up survey respondents were male.
In terms of race, household size, and household income of the participants, there we
significant differences between the respondents of the end of session evaluations and the
follow-up survey. In the follow-up survey, the proportion of Non-whites was about 46.8%.
Almost 90% of the Non-whites in the follow-up survey were Black respondents. Nearly 30%
5
The follow-up surveys were mailed to the participants 1-3 years after they completed the program.
11
that their household consisted of two or three people. Nearly 20% had household incomes less
than $18,871 and more than 30% had household incomes between $18,871 and $31,450.
In addition, Table 2 shows how the demographic characteristics of the respondents of the
follow-up survey varied by race. Non-whites were significantly different than Whites in terms
of gender and household size. In particular, 21.9% of Non-white respondents were male,
compared to 50.6% of White respondents. Also, 20.2% of Non-whites lived in single-person
households, compared to 36.9% of Whites. Non-white respondents were not significantly
different than White respondents in terms of their age and income.
The end of session evaluation of the Get Checking
TM
program also asked the participants
how confident they felt about several financial management activities as a result of the
program. The majority of participants indicated the highest possible level of confidence about
each of these behaviors. For example, 84.2% of participants indicated they felt very confident
about maintaining a check register and 80.2% of participants indicated they were very
confident about reconciling their checkbook registers with a bank statement. Respondents were
relatively less confident about talking to a financial institution about savings goals or credit
needs. Nevertheless, 75.7% and 74.4% of participants indicated they were very confident about
talking to a financial institution representative about savings goals and credit needs,
respectively. The goal of the follow-up survey was to measure actual behavior change in terms
of account usage and asset-building. Therefore, questions on level of confidence about
financial management activities were not included in the follow-up survey.
Table 3 presents the summary statistics for changes in actual behavior in terms of account
usage and changes in financial management behavior since the completion of the Get
Checking
TM
program. In terms of account usage, 97.5% of the respondents opened a checking
account and 90.6% still had the checking account open at the time the survey was conducted.
At the same time, 56.4% of the respondents opened a savings account, and the retention rate
12
was quite high, as 54.3% of the sample reported still having the savings account open at the
time the survey was conducted. However, the proportion of respondents who opened another
asset account or a loan account was relatively lower than the proportion of respondents who
opened a checking or savings account. For example, 15.0% of respondents opened an asset
account. Most frequently opened asset accounts were retirement savings (6.2%) and
certificates of deposit (5.0%). In terms of loan accounts, 16.2% reported having opened a loan
account. The most frequently opened loan accounts were auto (9.3%) and mortgage loans
(7.5%).
The table in the Appendix shows that the age and race of the respondent had a significant
effect on the account usage. In the sample used in this study, all of those respondents below
age 25 opened a checking account after earning their certificate. In addition, those below age
25 were more likely to open an asset or a loan account. Specifically, they were more likely to
obtain an auto loan. After earning their certificate, Non-whites were significantly more likely
to open an asset account and less likely to open a loan account.
The follow-up survey included a question that asked whether or not respondents owed
any money to the financial institution before they participated in the Get Checking
TM
program
and how long it took them to repay their debts to the financial institutions. As reported in Table
3, 61.2% of the respondents owed money to a financial institution before they earned their
certificate. While 40.0% of the respondents repaid the amount they owed within one month
after earning the certificate, 5.6% made restitution within two or three months. In addition, at
the time the follow-up survey was conducted, 10.6% reported still owing money to a financial
institution and being in the process of making restitution.
In the follow-up survey, respondents were asked to indicate changes in behavior with
respect to managing their finances after completing the program. Table 3 presents the
percentage of respondents that reported a positive change in financial management behavior.
13
These changes were analyzed in four categories: recording transactions (keeping an up-to-date
check register or a record of ATM or debit card transactions), reconciling bank statements with
check register, planning a budget (working to achieve a written financial goal or managing
income and expenses to meet financial goals or using a written spending plan) and, finally,
communicating with financial institutions. Overall, 75.0% reported recording transactions,
53.1% reported reconciling bank statements with their check register, 67.5% reported planning
a budget, and 44.3% reported communicating with the financial institution since the
completion of the program.
Table 3 presents the account usage and financial management behavior by the
communication behavior of the respondents. All of the respondents who reported
communicating with financial institutions opened a checking account. Respondents who
reported communicating with financial institutions were more likely than others to open an
asset account and to obtain a loan. Those who reported communicating with financial
institutions were also more likely to be keeping an up-to-date check register or a record of
ATM or debit card transactions; reconciling bank statements with their check register; and
working to achieve a written financial goal, or managing income and expenses to meet
financial goals, or using a written spending plan.
In the follow-up survey, respondents were asked to indicate their experiences with
cashing checks and buying money orders. Almost 35% of survey participants reported that
they had no expenditures for cashing paychecks or buying money orders. This finding is
consistent with Dunham (2001). In her study, 27% of the unbanked survey population in New
York and Los Angeles did not incur any costs for check cashing and money orders. The
analysis of Get Checking™ participants indicated that for those who reported non-zero
expenditures, the median and average monthly costs of cashing paychecks and buying money
orders were $16.00 and $23.48, respectively. On average, Get Checking™ participants
14
reported cashing 2.86 paychecks and buying 5.48 money orders per month.
There were no
significant differences in the cost and the number of checks and money orders by age.
However, there were significant differences in the cost of cashing paychecks and buying
money orders by the race of the respondent. For example, while White respondents paid on
average $28.59 per month for these services, Non-white respondents paid only $15.91 per
month. The median values for the cost of these services for Whites and Non-whites were
$21.75 and $11.48, respectively. In addition, White respondents purchased a greater number
of money orders than their Non-white counterparts (6.35 vs. 4.58).
Results
Changes in Financial Management Behavior
Changes in financial management behavior analyzed in this study, which were recording
transactions, reconciling bank statements with check register, planning a budget, and
communicating with financial institutions, were recorded as dichotomous variables. Therefore,
probit models were used to analyze the role of demographic and socioeconomic characteristics
on reported changes in these financial management behaviors. For each category of financial
management, it was expected that younger respondents and those with higher household
income would be more likely to change their financial management behavior.
Table 4 presents the results of probit regression for the determinants of changes in
financial management behavior since the completion of the Get Checking
TM
program. Table 4
also presents the marginal effects for the reference group of respondents which consists of
White females below age 25 years who live in one person households and who have household
income of $18,871- $31,450. The base probability of recording, reconciling, budgeting, and
communicating for the reference group was 63.0%, 29.2%, 42.7%, and 35.0%, respectively.
The marginal effect of each dummy variable represents the change in the base probability of
the reference group, all else being equal. For example, the base probability of recording is
15
equal to ?(0.332), where ?() is the standard normal cumulative distribution and 0.332 is the
intercept of the regression for recording. The estimated coefficient of Male is 0.522, and the
marginal effect of Male is ?(0.332+0.522)- ?(0.332), which is equal to 18.3%.
The results show that age had limited explanatory power on changes in financial
management behavior with respect to reconciling bank statements with check registers and
planning a budget. Respondents who were between age 45 and 55 and respondents over age 55
were more likely to reconcile checkbook registers and bank statements than those below age
25. Similarly, respondents between age 25 and 34 were more likely to plan a budget than those
below age 25. Gender had some limited influence on the change in financial management
behavior, and male respondents were more likely to record transactions than females.
Non-white respondents were more likely to record transactions and communicate with
financial institutions than White respondents. Finally, income had some impact on the change
in financial management behavior. Respondents who had household income over $50,321 were
more likely to reconcile bank statements with their check register and plan a budget than
respondents who had household income between $18,871 and $31,450.
Changes in Behavior in terms of Account Usage and Asset Building
Following the estimation of the changes in financial management behavior, three sets of
probit models were estimated to determine the factors influencing the decision to open a
savings, an asset, and a loan account, which were recorded as dichotomous variables. Each
model was estimated in three steps: first including only demographic factors as the explanatory
variables (Model I), second, adding the four dichotomous variables that measure changes in
financial management behavior (recording transactions, reconciling bank statements with
check register, planning a budget, and communicating with financial institutions) to Model I
(Model II), and third, adding an interaction term between race and communicating with
financial institutions variable to Model II (Model III).
16
Model II was estimated to capture the effect of changes in financial management
behavior on personal finances.
6
One of the explanations in the literature for not having a
banking relationship is that unbanked consumers mistrust banks due to their negative
experiences, and Non-whites are more likely to have negative experiences with banks than
Whites (Longhofer and Peters 2005). The data from the 2004 SCF confirms that claim. The
percentage of households in Table 1 citing not liking to deal with banks as the reason for not
having an account was the second highest reason (21.9%). In addition, previous studies show
that good banking relationships increase the probability of being approved for a loan and
accumulating assets (Chakravarty and Yilmazer 2005). Model III was estimated to investigate
whether improving the relationship with a financial institution affects the account usage of
Whites differently than the account usage of Non-whites.
Table 5 presents the estimation results of the probit models and marginal effects for the
likelihood of opening a savings account since the completion of the Get Checking
TM
program.
Marginal effects of the dummy variables are calculated using the same method as the marginal
effects presented in Table 4. In all three models that were estimated, probability of opening a
savings account was significantly affected by household size and income. However, other
demographic factors such as age, race, and gender of the respondent did not have a significant
impact on the likelihood of opening a savings account. Respondents who lived in households
with three or more people were less likely to open a savings account than those who lived in
one person households. Compared to respondents who had household income between $18,871
and $31,450, respondents who had household income less than $18,870 were less likely, and
respondents who had household income above $50,321 were more likely to open a savings
account. Changes in financial management behavior did not significantly affect the decision to
6
This model estimates the effect of financial management skills that were obtained during the course. If the
participants were already recording transactions, reconciling bank statements with their check register, planning a
budget, or communicating with financial institutions before they had participated the Get Checking
TM
program
and did not report a change of behavior, the estimated coefficients on these variables would underestimate the true
effect of these variables on opening a savings, asset, or loan account.
17
open a savings account. In addition, no significant differences in the impact of communicating
with financial institutions between Whites and Non-whites on the decision to open a savings
account were observed.
As presented in Table 6, age and race significantly influenced the probability of opening
an asset account since the completion of the Get Checking
TM
program. In comparison to those
below age 25, those between 25-34 and those above 55 were less likely to open an asset
account (Model II). Non-white respondents were more likely to open an asset account.
Changes in communication with financial institutions had a significant positive effect on the
decision to open an asset account.
In Table 6, Model III reports that Non-white respondents who communicated with
financial institutions were more likely to open an asset account than Non-white respondents
who did not communicate with financial institutions.
7
Compared to Whites who communicated
with financial institutions, Non-whites who communicated with financial institutions were
more likely to open an asset account.
8
In terms of opening an asset account, Non-whites
seemed to benefit more from the financial management content emphasized during the
program.
Finally, the determinants of obtaining a loan since the completion of the Get Checking
TM
program were estimated. Results are presented in Table 7. Similar to the findings for asset
accounts, respondents below age 25 were more likely to obtain a loan than older respondents.
Male respondents were also more likely to have a loan account than female respondents.
Unlike asset accounts, the probability of having a loan account was lower for Non-white
7
The coefficient estimate of Non-whites who communicated with financial institutions is the sum of the
coefficients of Non-white (0.279), Behaviour_Communication (-0.154) and Non-white*
Behaviour_Communication (1.319) and equals to 1.444 (p-value