Working Paper
WP 10-1
September 2010
Early-Life Schooling and Cognition and Late-Life Financial Literacy in the Wisconsin
Longitudinal Study
Pamela Herd and Karen Holden
Center for Financial Security WP 10-1
Early-Life Schooling and Cognition and Late-Life Financial Literacy in the
Wisconsin Longitudinal Study
Pamela Herd
University of Wisconsin-Madison
Karen Holden
University of Wisconsin-Madison
Center for Financial Security
University of Wisconsin-Madison
Sterling Hall Mailroom B605
475 N Charter St.
Madison, WI 53706
http://cfs.wisc.edu/
(608) 262-6766
Abstract
Using the Wisconsin Longitudinal Study, we examine the links between early-life cognition and
schooling experiences and late-life financial literacy. We find that early-life cognition, especially
for those with very low IQ scores, and schooling have a relationship with late-life financial
literacy. If the trend continues towards very individualized retirement planning, with definedcontribution plans and private savings dominating much of retirement income, we will need to
generate strategies to help these vulnerable populations with more limited cognitive functioning
manage this increasingly complex financial world.
The research reported herein was performed pursuant to a grant from the U.S. Social Security
Administration (SSA) funded as part of the Financial Literacy Research Consortium. The
opinions and conclusions expressed are solely those of the author(s) and do not represent the
opinions or policy of SSA, any agency of the Federal Government, or the Center for Financial
Security at the University of Wisconsin-Madison.
Center for Financial Security WP 10-1
Early-Life Schooling and Cognition and Late-Life Financial Literacy in the Wisconsin
Longitudinal Study
While Americans are increasingly left to save and invest on their own for retirement,
there is considerable evidence that many Americans are lacking the skills to effectively
complete this task. One-third of adults in their 50s say they have failed to develop any kind of
retirement saving plan at all (Lusardi 2003). Surveys of financial literacy find that most
respondents do not understand basic financial concepts, including topics such as bonds,
stocks, mutual funds, compound interest, loans, and mortgages (Agnew and Szykman 2005;
Bernheim 1995, 1998; Hogarth and Hilgerth 2002; Mandell 2004, 2007; Moore 2003; NCEE
2005). Consumers are also ignorant about social security and work-related pensions, two of
the most important components of retirement wealth (Chan and Huff Stevens 2003; Gustman
and Steinmeier 2004; Mastrobuoni 2005).
Consequently, in recent years, programs and initiatives have been developed to
promote financial literacy and provide financial education to United States consumers. Most
of these interventions have focused on providing specific financial skills to specific
subpopulations (Collins and O’Rourke, forthcoming; Lyons et al. 2006). The results from
these different programs and populations have been generally positive, but small (Bernheim
and Garrett 2003; Collins and O’Rourke, forthcoming; Bernheim, Garret, and Maki 2001;
Lyons et al. 2006; Way and Ang 2010). These studies are also only focused on short-term
interventions and measure short-term outcomes (Collins and O’Rourke, forthcoming).
Thus, while targeted and short-term financial literacy initiatives have blossomed and
been the subject of research, we know much less regarding whether and how more
generalized education (i.e., math and reading skills) and general cognition (i.e., IQ, which is
in part shaped by schooling) impact financial literacy. Does having a low IQ in early life lead
to poorer financial literacy and thus greater financial risk in late life? Consequently, this
study examines the links between early-life cognition and schooling experiences and late-life
financial literacy. Is early-life cognitive ability and schooling associated with financial skills
later in life? This study employs an innovative approach to measuring financial literacy. In
short, we capture individuals’ knowledge of their own financial situations, which is a
prerequisite for good financial behavior.
While our original goal was to look at a range of events across the life course, we
decided to focus more tightly on early-life experiences, especially schooling and cognition.
Because of this focus on early life, instead of emphasizing cognitive decline, we emphasize
the relationship between education and IQ, providing a rich picture of these earlier life
circumstances and their implications for late-life financial literacy. Future work will explore
how later-life events help mediate these early-life predictors.
Schooling and Financial Literacy
The contrast between focused financial literacy intervention and general schooling as a
means to producing better financial skills is based on a broader debate within human capital
theory regarding the relative importance of general versus specific skills. Human capital
theories emphasize that it is general skills, rather than specific skills, that schooling is so
effective at producing (Becker 1962; Schultz 1962). But a recent debate has emerged about
whether high rates of college completion are necessary or whether some students would be
better off focusing on more skills-specific training, such as trades, as opposed to the general
skill set offered by higher education (Lerman 2009; Vedder 2007). A similar, though less
contentious debate, has focused health research. Should interventions to improve health
emphasize general skills achieved via schooling or should interventions be targeted and
focused on improving healthy behaviors and effective medical care use (Link and Phelan
1995)? In short, there has been a tension between the effectiveness of more general training
received in primary, secondary, and postsecondary schooling versus more specific training
targeted to enhance a range of life outcomes.
To date, however, while the literature devoted to financial literacy can demonstrate a
growing level of information regarding task and subgroup-specific interventions, there is
much less research examining the broader effects of schooling and cognition. While
intervention studies have demonstrated links between short-term, targeted learning
intervention and positive outcomes, no studies have closely examined the secondary school
experience (Barron and Staten 2009; Bell et al. 2009; Devaney et al. 1996; Lyons et al. 2006;
Varcoe and Wright 1991; Wallace et al. 2010; Way and Ang 2010). And while nearly all
studies find that more years of schooling in early life are linked to financial well-being and
literacy across the life course (Cole and Shastry 2009; Lusardi and Mitchell 2007a), no
studies have examined how or whether things like academic performance and particular kinds
of coursework are linked to financial literacy.
And only recently have some studies begun to explore the relationship between
cognition and financial literacy. In part, this more recent literature has been catalyzed by
findings that link cognitive ability to wealth (Lusardi and Mitchell 2005; Banks and Oldfield
2007; McArdle et al. 2009). The findings from studies focused on the link between cognition
and actual financial literacy have been more mixed. Alan Gustman and colleagues (2010),
utilizing the Health and Retirement Study (HRS), found no link between cognitive numeracy
measures or fluid intelligence and knowledge of one’s pension plan and social security rules,
such as eligibility ages and plan values. Fluid intelligence is the general capacity to think
logically and solve problems in novel situations, while crystallized intelligence is the depth
and breadth of an individual’s knowledge and his or her ability to actually use that
knowledge. Contrastingly, Sumit Agarwal and Bhashkar Mazumder (2010) found that
individuals with higher Armed Services Vocational Aptitude Battery (ASVAB) scores, which
are more strongly correlated with crystallized intelligence, were less likely to make financial
mistakes. In short, those with higher scores were less likely to use a credit card for a
transaction after making a balance transfer on the account or inaccurately estimate the value
of a home on a home equity loan or line of credit.
Thus, prior research has not fully evaluated how the detailed schooling experience (i.e.,
academic performance and coursework) and general cognition (which is affected by
schooling) may impact financial literacy. For example, does having a comprehensive slate of
math classes help facilitate financial literacy across the life course (Peters et al. 2006)? Do
strong basic literacy skills, as measured by the number of English courses one takes in high
school, bear any relationship to financial literacy skills later in life? Does general cognitive
ability, measured in early life, help predict later-life financial literacy skills. While Gustman
and colleages (2010) also focus on financial literacy in late life, their cognitive measure was
taken in late life. Cognition, however, especially fluid intelligence, changes as people age.
Perhaps developing general knowledge and cognition in early life is a critical part of
developing strong financial behaviors that are evidenced in late life.
The argument that general cognitive skills, most of which are developed through general
schooling, will positively affect financial literacy is based on the premise that what we need
to know to be effective at financial management changes over time. For example, today,
effective retirement financial planning requires an understanding of defined contribution
plans, private investment accounts, and the nuances of social security policy. But in 30 years,
policy and employment changes may require a very different set of skills. General cognitive
skills, as opposed to very specific skills training, allow people to adapt and pick up new
skills. In short, it provides people with the fishing rod, as opposed to the fish.
If there is a link between these early-life schooling and cognitive measures and later-life
financial literacy, this implies the importance of strong general schooling (including math and
general literacy skills) to help prepare students not only for the job market, but to effectively
manage their financial lives in an increasingly complex and ever-changing financial and
policy context.
The Challenge of Measuring Financial Literacy
In recent years, while efforts have been made to promote financial literacy and
provide better financial education to consumers, the term ‘financial literacy’ has been used
without a clear consensus of what it constitutes and how it will lead to greater financial
security. Researchers have attempted to create measures, however, that can be used to
measure program impact and understand generally what predicts ‘financial literacy’ (Fox,
Bartholomae, and Lee 2005; Lyons and Scherpf 2004; Lyons et al. 2006; Schmeiser et al.
2010; U.S. Government Accountability Office 2004).
Though measures of financial literacy employed have varied considerably, they
broadly fall into two groups, knowledge based and behavior based. Knowledge-based
measures include things such as whether individuals can calculate compound interest or know
the difference between a stock and a bond (Hilgert et al. 2003; Holden et al. 2010; FINRA
2003; Moore 2003; NCEE 2005; Lusardi and Mitchell 2007a, 2007b, 2007c; Lusardi 2008a,
2008b; van Rooij et al. 2007). This approach to measuring financial literacy assumes that if
consumers have basic tools, they can use them to make better financial decisions. For
example, a NASBE (2006) report argued that financial education is like all types of
education, in which teaching the basics helps to develop the building blocks individuals need
to make good financial decisions throughout their lives. Alternatively, measures can be action
based, measuring whether individuals exhibit good financial behaviors (Moore 2003; ANZ
Bank 2008; Lusardi and Tufano 2009). These kinds of measures capture things such as levels
of financial debt, participation in pension plans, or adequacy of retirement savings. To be
sure, some studies have examined the validity of the knowledge-based measures by using
them to predict behaviors like retirement savings and debt levels (Lusardi and Tufano 2009;
Lusardi and Mitchell 2009). Generally, however, the measures themselves do not overlap and
capture both knowledge and behavior, although there is some evidence that they are
correlated.
We utilize a measure of financial literacy that supplements these existing measures
with the virtue of being drawn from questions that are asked on many surveys. That is it does
not depend on ‘financial literacy’-specific variables. In short, we attempt to capture an
individual’s knowledge of his or her own financial situation. Our survey measures how
knowledgeable individuals are regarding their own assets, their retirement savings, and their
immediate financial resources. We know of only one other study that has attempted this
approach (Gustman et al. 2010).
Having knowledge of one’s own finances is a prerequisite for making good financial
decisions. Further, unlike a more general knowledge-based measure, this measure gets at the
knowledge individuals require to manage their own finances. A virtue and criticism of our
measure is that it measures not only knowledge, but behavior as well. This reflects one of the
controversies, discussed in this paper, about the appropriate measures of financial literacy—is
it about knowledge, which may never be activated, or about behavior that achieves the
literacy goals? Our measures reflect the efforts individuals have made to be precise and active
in their financial life. There are likely many well-educated individuals who can identify the
difference between a stock and a bond, but cannot say how much they have in their checking
accounts. There are likely many who formally understand compounding but, even though on
the verge of retirement, do not know how much their pension will be worth. That said, of
course having complete knowledge of one’s existing financial situation does not guarantee
good financial decisions. Individuals may know they have very low assets relative to their
debt and still overspend, for example, or may know about their debt precisely because they
have overspent in the past. One could argue, however, that these are still more financially
knowledgeable individuals. And, we would argue, it is nearly impossible to make good
financial decisions if one is not aware of one’s basic financial resources.
Data and Methodology
There are no better data available in the United States than the Wisconsin
Longitudinal Study (WLS) to test the links between early-life schooling and financial skills in
late life. The WLS encompasses a cohort of 10,317 Wisconsin high school graduates mainly
born in 1939. The sample is generally representative of white, non-Hispanic high school
graduates across the United States. Survey data were collected in 1957, 1964, 1975, 1992/3,
and 2003/5. It is the longest-running longitudinal sample in the United States, with the first
wave of data collected when the respondents were seniors in high school in 1957 and the
most recent wave of data collected in 2004 when the respondents were in their mid-60s.
It is the only United States data set with administrative and prospective data that
allows one to explore the links between early-life schooling and financial literacy. While
surveys such as the HRS collect measures at multiple time points for a sample of older
Americans, the WLS includes a broad array of early-life schooling characteristics (including
years of schooling, course content, and school performance and interest) not included in other
studies of older Americans.
One weakness with the WLS, however, is that it is a homogenous sample of white
Wisconsin high school graduates from 1957. While this has obvious disadvantages, a
relatively homogenous sample can help rule out unobserved variable effects that would arise
from birth cohort, education level, and geographic area correlates. Though many observed
variables can be accounted for, such as sex, there are numerous correlates, such as cultural
differences, which are harder to account for but which are still potential confounders.
While the original WLS sample contains over 10,000 respondents, this study analyzes
just over half of that original sample (6,276 cases). Cases lost to follow-up include almost
1300 respondents who had died by 2004; approximately 1400 cases who refused to answer
the phone survey in 2004, in which key questions for our analysis were asked. The remaining
missing cases were lost due to information missing on key covariates (educational attainment
and course content) or on the outcome measure. The WLS has been ongoing for over 50
years, making sample retention an obvious challenge. Yet, compared to the Panel Study of
Income Dynamics (PSID), which lost 50 percent of its sample to attrition over a period of 21
years between 1968 and 1989 (Fitzgerald et al. 1998), the WLS has had strong sample
retention.
Outcome Measures. We developed a set of measures to establish whether or not individuals
exhibit good financial skills and knowledge of their individual finances. We defined three
measures based on the 2003/5 survey data. Descriptions of the variables are presented here.
Appendix A discusses additional details regarding sensitivity analyses to test the validity of
these measures.
These measures are:
• Percent of asset questions to which the respondent gave a precise
answer
• Knowledge of amounts in checking, savings, and money market
accounts
• Knowledge of the value of retirement accounts
The first measure is intended to capture the respondent’s overall awareness of his or her
current financial situation. This WLS-constructed measure is a simple accounting of the
percent of total asset categories held by the respondent (and spouse, if married) for which the
respondent can provide an exact dollar amount when asked in a series of 12 questions. Asset
categories include property, account balances in retirement plans, and life insurance cash
values. We include in our sample only those cases that completed the full series of asset
questions. Respondents who refused to answer at least one asset question comprised about 10
percent of the full sample and were excluded from our analyses.
The second measure is intended to capture day-to-day financial awareness. We
construct a binomial variable from the answer to the question, ‘If you added up all of your
and your spouses’ checking accounts, savings accounts, or money market funds, about how
much would they amount to right now?’ The variable is coded as 1 if respondents provided a
value and 0 if they stated that they did not know the amount. Respondents also had the option
to refuse to answer. Just under 10 percent of the sample refused to answer this question and
were thus excluded from these analyses.
The final measure is intended to capture long-term financial planning skills. Are
individuals aware of where they stand in relation to their retirement income plans?
Individuals are asked whether they or their spouse ‘have any retirement plans that accumulate
an account balance—examples include IRAs, 401(k)s, and profit-sharing plans.’
Approximately 75 percent of the sample reported having such a plan. Respondents were then
asked ‘If you added up all of your and your spouse’s retirement plans that accumulate an
account balance, about how much would they amount to right now?’ About 20 percent of
those who reported that they or their spouse had such accounts reported that they didn’t know
the value. Another 10 percent refused to answer and were thus excluded from the analyses. A
binomial variable is defined for those with such a plan that is equal to 1 if the respondent
reports a value and 0 if the respondent reports that they don’t know the value.
Covariates. This first set of independent variables is intended to capture basic cognitive
ability and the early acquisition of skills that are likely to enhance the ability of individuals
over their lifetimes to read about, understand, rigorously critique, and act on complex
information.
Childhood Cognitive Ability/IQ. These scores, available through school district
administrative records, are derived from the Henmon-Nelson Test of Mental Ability, which
was administered to high school students in Wisconsin. The Henmon-Nelson test is
considered a general measure of intelligence, but a recent analysis indicated that, although it
captures both fluid and crystallized intelligence, it is more strongly correlated with
crystallized intelligence (Pallier et al. 2000). It is important to note that although high school
rank and IQ are correlated (r=0.58), there is still meaningful variation. We ran this variable as
a series of three splines. The first was for IQ scores below 100; the second spline was for IQ
scores from 100 to 120; and the third spline was for IQ scores above 120.
Educational Attainment. This measure calculates years of schooling derived from the highest
degree attained and number of years of higher educational attainment. The measure ranges
from 12 to 20 years, with 12 being a high school graduate and 20 indicating the attainment of
a PhD.
High School Rank. This measure is a percentile rank based on high school grades [100–(rank
in class/(# of students in class))*100]. Rank was then divided into quartiles. The general
correlation between grades and standardized test scores is high at 0.9, which reduces
concerns that grades or rank may reflect teacher bias (Willingham et al. 2002).
Course Content. Three measures capture the kinds of courses students took in high school.
Two math-related measures capture the presumed acquisition of advanced math skills, the
general computational skills required for financial literacy. One of the measures indicates
whether the student took physics or trigonometry. The second measure indicates whether the
student took the average number of semesters of algebra, fewer than the average, or more
than the average among the WLS cohort. The student could report up to four semesters of
algebra. Over half of the sample reported they had taken two semesters of algebra. The third
measure captures the acquisition of what we would label general literacy skills not specific to
financial literacy, but necessary for reading and understanding complex writing. It indicates
whether the student took the average number of semesters of English, fewer than the average,
or more than the average. Students could report whether they took 0, 2, 4, 6, or 8 semesters of
English.
Confounding Covariates. This set of variables includes those that, if excluded, would lead
to biased estimates of the influence of the previous set of covariates. This could be either
because they are themselves causal of those skill acquisition attributes or because the
confounding variables enhance the contributions of those skills. The most straightforward
example would be parents of high socioeconomic status with their own financial skills who
teach those to their children; as adults, those children would be more financially literate
regardless of their high school course selections, which would also be influenced by their
parents’ expectations.
Parental Socioeconomic Status. It is critical to control for parental socioeconomic status
because it may predict schooling measures (attainment, rank, coursework), cognition, and
outcome measures. The parental socioeconomic status measure is a WLS-created factorweighted score ranging from 1 to 97. The score is based on: 1) highest number of years of
schooling for respondent’s mother and father, 2) Duncan SEI occupational score for
respondent’s father’s occupation, and 3) four-year average of parental income between 1957
and 1960, based on Wisconsin tax records.
Basic Demographics. These measures control for the sex, age, and race of the respondent,
although there is little variation by age and race.
Analytic Techniques and Models
The analyses include logit and ordinary least square (OLS) regressions. Because of
differences in the relationship between the outcomes, covariates, and whether or not
individuals had college degrees, we ran separate models for those with at least a college
degree and those without a college degree. An OLS model was estimated of the predictors of
the percentage of asset questions respondents answered. A logit model was employed for the
binary outcome measures regarding whether individuals know the value of their pension
accounts and whether or not they know the value of their checking accounts. The models
included parental socioeconomic status (measured in 1957), basic demographics, childhood
IQ (measured in 1955–57), high school rank (measured in 1957) and coursework variables
(measured in 1957), and years of schooling (a composite measure based on reports in 1974,
1992/3, and 2004). Note that the years of schooling variable captures additional schooling
beyond high school, but also throughout adult years. The WLS cohort was one in which
women returned to school after child rearing. This subsequent schooling is important in
capturing effects of later schooling on current financial acumen as well as reducing the
greater influence in 1957 of gender (versus scholastic aptitude) on college and post-college
attendance.
Results
The results from these analyses generally find that early-life experiences related to
both schooling and cognition are linked to late-life financial literacy skills. But these effects
vary in important ways across the sample. In particular, the effects of early-life cognition
appear largely confined to those with IQ scores toward the bottom of the distribution of IQ
scores and among those without college degrees. In contrast, the effects of additional years of
school and early academic measures (including math classes taken in high school) are largely
significant only for those with college degrees.
Table 1 here
Table 1 shows simple bivariate correlations between individual characteristics and the
three outcome variables. Around 72 percent of respondents who had retirement savings
accounts reported that they knew the value of their balance-accruing retirement savings plans.
A slightly higher percentage of individuals who had such accounts knew the value of their
checking, savings, and mutual fund accounts (81 percent). For the sample, on average 89
percent of total assets were reported with precise answers, without requiring estimates
through the bracketing series. For most of the outcome measures, there are large and
significant differences by gender, educational attainment, and IQ, those with fewer than 6
semesters of English in high school compared to those with up to 8 semesters of English,
more versus fewer semesters of algebra in high school, and between those who took physics
or trigonometry in high school and those who did not. These basic patterns held across the
outcomes. The main variation from these themes was that the number of semesters of English
in high school was not correlated with the bank account knowledge variable.
Table 2 here
Table 2 presents findings from an OLS model for which the outcome is the percentage
of total assets for which respondents provided exact dollar values. We provide separate
estimates for those with and without college degrees, a distinction that from our early
estimates we concluded was important. We discuss the results for those with at least a college
degree. Having more semesters of algebra (borderline statistical significance) and having
taken trigonometry and/or physics were both positively correlated with the outcome. Both
indicated that individuals provided exact dollar values for 1.6 percent more of their total
assets. Being male had especially large effects on the order of an 8.6 percent difference
compared to women. While parental socioeconomic status was negatively correlated with the
outcome, its effect was small. Nevertheless, it is interesting to note that respondents who had
come from a higher socioeconomic background actually had less knowledge of their asset
levels.
The second set of results in Table 2 is for those without college degrees. Unlike the
effect for college graduates, the two bottom IQ categories were positively correlated with
respondents’ knowledge of their asset levels. A gain of approximately 10 points in IQ in the
less than 100 IQ range increased the percentage of assets reported with precision by about 3
percent. The effect of a 10-point gain for those in the 100–120 IQ range was 2 percent. High
school rank had an effect for this group. Among those who did not go to or complete college,
there was a 2.6 percent difference for those at the top compared to the bottom quartile of their
high school class. Cognition and academic performance were correlated with asset
knowledge for this sample, but coursework in high school was not. But like those with
college degrees, being male was strongly correlated with having more knowledge of one’s
asset levels (a 9 percent difference between men and women).
Table 3 here
The results in Table 3, in which the outcome is knowledge of the value of
respondents’ retirement plans that accumulate a balance (such as a 401(k)), parallel the
results in Table 2. We provide both odds ratios and marginal effects. For college graduates,
both additional levels of schooling and being male were correlated with knowing the value of
one’s retirement account accumulations. Each additional year of schooling increased the odds
of knowing the value by 25 percent, representing a 2 percent increase in the probability. The
odds that a man knew the value of his retirement savings account were 4.6 times higher than
a comparable woman, a 17 percent difference in the probability. In contrast, for those without
a college degree, being male had a smaller but still significant influence and cognition had
significant effects. The odds ratio for a 10-point difference in the less than 120 range was
about 1.02. In probability terms, each 10-point increase led to a 3–4 percent increase in the
probability of knowing one’s private pension account value.
Table 4 here
The results in Table 4 are for the outcome relating to whether the respondent knows
the value of his or her checking accounts. Both odds ratios and marginal effects are
presented. For college graduates, the only even marginally significant covariate, besides
being male, was the number of algebra courses taken in high school. The odds of knowing
one’s checking account balances were 34 percent higher (or a 2.5 percent difference) for
those who took the highest number of algebra courses compared to those who took the
average number. But for those without college degrees, the results roughly paralleled the
prior outcome variables (asset knowledge and retirement savings account knowledge). In
short, for those with IQs between 100 and 120, the odds ratio was about 1.28 (or 3 percent)
for each additional 10 points of IQ.
Sensitivity Analyses. Sensitivity analyses conducted on all three outcomes altered covariates
included in the model, but did little to change our main conclusions. In the various alternative
models, we excluded the English variables, based on the concern that coursework patterns
were correlated in high school; altered the algebra coursework variable so that the reference
was 0 semesters of algebra, in order to identify the key differences in effects among the
categories; and varied cutoff points for the IQ splines. None of these variations altered the
basic general findings across these models. We plan additional sensitivity analyses, including
variations in how we treat individuals with missing covariate values. For example, missing
values on high school rank were simply excluded, which if missing for individual-specific
administrative reasons (e.g., individuals who had moved recently) could bias our results.
Findings on Financial Literacy for Vulnerable Populations
What are implications from these analyses for vulnerable populations? First and
foremost, these analyses indicate that as the retirement context shifts toward one that requires
significant financial literacy skills, individuals with more limited cognitive functioning may
fare poorly. If the trend continues toward very individualized retirement planning, with
defined-contribution plans and private savings dominating much of retirement income, we
will need to generate strategies to help those with more limited cognitive functioning. About
20 percent of the sample fell into this category. And it is important to keep in mind that this is
a relatively well-educated sample. Consequently, it is likely that a larger portion of the
general population faces challenges in acquiring effective financial literacy skills due to
cognitive functioning.
Implications and Conclusions
The main findings from this paper are that early-life cognition, especially for those
with very low IQ scores, and schooling do have a relationship with late-life financial literacy.
In this paper, we have defined financial literacy as the level of knowledge one has regarding
one’s own personal finances. We use data that are often collected in surveys and thus can
provide insight into financial knowledge across populations, even when financial literacy is
not targeted in a survey. It seems reasonable to argue that these questions reflect levels of
financial literacy required by individual circumstances:
1. Can you precisely identify all of your assets and their value?
2. Can you identify the value of assets most immediate to day-to-day financial transactions,
such as your checking accounts?
3. Can you identify the value of assets important to life-course planning, such as your
retirement savings accounts?
The effects of early-life cognition were significant, but confined to those at the very
bottom of the IQ distribution. In short, cognitive effects were found only for those without
college degrees and the differences were only statistically significant among those in the
lower IQ brackets, especially for those with IQ scores in the bottom portion of the
distribution. These findings differ from Gustman and colleagues (2010), who found in the
HRS that a numeracy measure was not predictive of knowledge of pension and social security
rules. There are a few possible explanations for these differences in findings. First, it is not
clear that Gustman and Steinmeier tested for nonlinear effects. Second, while their numeracy
measure is strongly correlated with fluid intelligence, the Henmon-Nelson used in our study
actually measures a mix of both fluid and crystallized intelligence. Fluid intelligence is the
general capacity to think logically and solve problems in novel situations, while crystallized
intelligence is the depth and breadth of an individual’s knowledge and his or her ability to
actually use that knowledge. So it may be that fluid intelligence is less important than
crystallized intelligence. Indeed, the Agarwal and Mazumder (2010) study also used a
cognitive measure that is more clearly correlated with crystallized intelligence, and they
found it to be quite predictive of financial literacy skills.
It is important to keep in mind that education has demonstrable effects on cognitive
scores. And because these scores were taken in high school we cannot be certain to what
extent these differences are actually driven by prior schooling experiences. We have
controlled, however, for family background characteristics, which are also correlated with IQ
measures. It is also possible that these early IQ scores have some predictive capacity for laterlife cognitive decline. We expect to explore whether and how current cognitive functioning
explains these early-life cognitive relationships.
For those with college degrees, the main findings, other than gender effects, were that
schooling experiences were linked to one’s financial literacy in late life. The number of years
of school beyond a college degree mattered. High school math coursework also mattered,
while English coursework never had a significant effect. This suggests that math-specific
skills may matter more than general reading comprehension to understanding financial
content. That coursework variables were not significant for those who did not go to or
complete college may be a product of students who did not continue onto college having that
decision shaped by their math performance. Though we do control for IQ, we cannot rule out
the possibility that the link between math classes and financial literacy is the product of
intrinsic quantitative capacity rather than something learned, although it is interesting that the
only prior study to look at a measure of numeracy, which captures this underlying proclivity,
did not find a link between it and financial literacy. While it would have been helpful to have
more detailed measures of coursework, including performance in those courses, to sort out
these issues, this analysis provides some evidence that the content and amount of early-life
schooling has some relationship to financial literacy skills in late life.
The findings also varied to some extent across the outcome variables, though this
variation was largely confined to those with college degrees. The relationship between
schooling characteristics and the outcome were strongest for the percentage of asset values
known and weakest for knowing checking account values. The asset variable measures
knowledge of one’s overall financial picture. We note, however, that a person with a more
complex portfolio would be more challenged to know all asset values precisely, while
someone with only a checking account would more likely score high on this variable. Thus,
the variation in findings could reflect the fact that the kinds of higher-order skills linked to
schooling are only important for more complicated outcomes. It is much easier to state the
value of one’s checking account as compared to knowing the value of a complex array of
assets.
We were not entirely surprised by the large gender differences in knowledge. On the
other hand, the variables were only defined for individuals who said they knew they held a
particular asset. We did conduct separate analyses of men and women by educational group
(college/noncollege). The years of schooling and cognitive effects were roughly comparable
to the findings reported here. The problem was simply that the samples became too small to
draw firm conclusions. Thus, while we were assured by these subgroup analyses, we hesitate
to draw strong conclusions from them because of small sample sizes. There remains a need to
explore what drives these gender differences, especially in a cohort where gender affected
additional schooling, marriage, and other intermediate outcomes.
There are some caveats to this study that should be addressed. First is the variation in
outcome variables. The percentage of people with knowledge ranges from 79 percent for the
retirement savings outcome to 88 percent for the asset knowledge outcome. As mentioned
above, this percentage is likely to be influenced by the complexity of portfolios, a fact that
confounds this as a measure of financial knowledge alone. We do find interesting
relationships even with this relatively small variation in the outcome variable. But ideally, it
would be useful to define variables that capture the greater variation in knowledge that has
been exhibited in financial literacy surveys. Hence, we are pursuing other measures of
financial outcomes to discover measures that vary more substantially in terms of knowledge.
Another caveat regards the sample selection. All of the analyses excluded those who
refused to answer. Prior research demonstrates that these individuals tend to be higher income
and wealth individuals. We are not certain of the impact on these analyses of this exclusion.
In the appendix, we describe some work on this issue and detail why we conclude that the
exclusion of these individuals does not invalidate the general findings presented here.
The two binary variables are defined only for those who hold the particular asset. This
of most concern for the pension variable; about 20 percent of the overall sample did not have
such an account. Clearly, these individuals are a select group, with those having accounts also
more likely to have characteristics that affect the probability of their knowing account values.
Those who do not have pension accounts are more likely to be women, to have low incomes,
and to have low educational attainment. These findings about financial knowledge predictors
must be understood to reflect relationships only for individuals who are likely to have these
kinds of market-based, individualized (defined contribution) retirement savings accounts.
It is also important to consider that the interpretation of these measures and analyses
must vary depending on the respondent’s age. For example, an individual age 40 being
unsure of the exact value of a retirement account is less policy problematic than is a 60-yearold individual contemplating retirement. It is only as retirement approaches that the precise
value of these accounts has a more immediate impact on individuals’ financial situation.
Thus, precision in answers is arguably an especially relevant measure of financial literacy for
those approaching retirement, although a more ‘ball park’ knowledge is important to younger
savers. One advantage of the WLS data is that it avoids this variation in relevant
knowledge—these are all retirement-age individuals.
As is always the case when using the WLS data, the findings must be understood in
the context of its population; these are all high school graduates and members of a very
specific cohort. These analyses cannot tell us how those without high school degrees fare,
although the link between lower cognitive scores and financial literacy gives some indication
of what that relationship may look like. Analysis of this cohort is valuable for understanding
influences on financial knowledge of younger individuals in the sense that the members of
the WLS cohort hold assets that prevail today and are faced with the same challenges of
understanding them. Retirement income is dependent on how careful and knowledgeable
individuals are as they face a financial landscape that changes for both young and old (e.g.,
new rules allowing conversions of IRAs and 401(k) assets into other forms). As already
noted, about 80 percent of the sample had defined-contribution retirement savings plans (like
401(k)s), which are accounting for a growing share of retirement assets for both younger and
older cohorts.
The final caveat is that while we see these associations between cognition, schooling,
and late-life financial literacy, we cannot claim these are entirely causal relationships.
However, our childhood measures are prospective and in some cases (like for cognition)
based on administrative data. So the study does provide a unique contribution for
understanding the links between early-life factors and late-life outcomes.
More generally, the results speak to the importance to later-life financial literacy of
general skills. Thus, while targeted interventions later in life are likely effective, these results
emphasize the importance of more general skills, which are achieved through schooling in
early life and perhaps in mid life as well. The specific skills needed to manage finances
effectively have changed dramatically over time. Having general skills that enable adaptation
to changing financial circumstances are likely critical for financial well-being.
Acknowledgements
The authors are grateful for feedback from discussants and participants at the summer CFS
FLRC workshop in Madison, WI. We are also grateful for the capable research assistance of
Yung-ting Su.
Appendix: The Validity of Financial Literacy Questions
Our financial literacy measures are based on several key assumptions, which we will
test in this section. One assumption is that individuals providing an exact response do
actually know the exact value of their assets. A second is that those who respond ‘I don’t
know’ are indicating lack of awareness of amounts even within some reasonable margin of
error, rather than simple uncertainty about the exact value of accounts at the time of
interview. Finally, we exclude from our variable definition those who refuse to answer
necessary questions. The implicit assumption is that these refusals are randomly distributed
across the sample rather than being more likely to have characteristics correlated with the
outcome variable. The results discussed below provide some assurance that these are
assumptions that, at least, do not lead to biased results.
We provide a range of estimates to help validate that individuals are giving accurate
answers if they provide an exact value of their assets. Because we do not currently have
administrative data to do a clear test, we perform a few sensitivity analyses to test the
potential accuracy of the responses. First, the fact that women are much less likely to report
specific asset values as compared to men helps confirm that individuals are answering
honestly. Tests of financial knowledge (i.e., how to calculate compound interest), like the
findings presented here, indicate that women are much less likely than men to exhibit strong
financial literacy skills (Lusardi and Mitchell 2008). Second, the IQ findings also provide
some assurance regarding response validity. Again, as do prior studies that test financial
knowledge, we find that those with lower cognition levels are much less likely to report
specific asset values.
We also test whether men, perhaps not wanting to appear unknowledgeable, report
values even if they do know the values of varying assets. To test this hypothesis we utilized a
measure of masculinity. Respondents were asked in the 2004 follow-up mail survey, ‘To
what extent do you agree that a man should always try to project an air of confidence even if
he doesn’t really feel confident inside?’ Responses ranged from strongly agree to strongly
disagree. The response distribution was a normal bell curve. We reran models only on men
and included this measure to test the hypothesis that men, especially those with very
traditional notions of what constitutes being a man, may report values for assets categories
even when they do not know them. We found no evidence to support the hypothesis. This
measure showed no ability to predict whether respondents reported a value versus reporting
that they don’t know. We would have expected that those who always felt like they had to
‘project an air of confidence even if [they don’t] really feel confident’ would be less likely to
report that they don’t know the value of an asset (such as their pension value or checking
account value). We found no evidence that this was true.
In addition to concerns about inaccurately reporting specific asset values, another
issue is whether those who respond ‘I don’t know’ are actually indicating a lack of awareness
of amounts rather than simple uncertainty about the exact value of accounts at the time of
interview. Again, our findings regarding gender differences and cognitive factors so closely
parallel prior studies predicting financial literacy; this provides some assurance that we are
getting at a true lack of knowledge. But we also performed some additional analyses to
further test the validity of this concern. In particular, we explored differences between those
who reported that they did not know the value of their checking accounts or their retirement
accounts and those who initially reported that they did not know but then provided answers in
the form of bracketed values or within a certain range of values. A similar pattern of findings
held across these groups, but the differences were much wider between those who never
provided an answer versus those who provided an exact answer, compared to those who
provided an answer through bracketing versus those who provided an exact answer. Those
with lower IQs, women, and those with more limited educational attainment were more likely
to provide no values at all or values through bracketing than to provide a specific value.
Finally, we explored the implications of excluding from the analyses those who
refused to answer. In particular, we wanted to be certain that those who refuse are actually
distinct from those who say that they don’t know the value of their asset categories. It is
possible, for example, that some individuals simply reported that they don’t know as a polite
way to refuse to respond. Thus, we compared those who refused to answer with those who
said they don’t know the value of their ‘retirement plans that accumulate an account balance,’
those accounts including IRAs, 401(k)s, and profit-sharing plans, and their checking account
values.
Table A1 shows the basic differences in the descriptive relationships between the
covariates and those who refused to answer versus those who reported that they don’t know
the value of their varying assets. In short, the table demonstrates that the main differences
(gender, cognition, and educational attainment) we saw between those who reported that they
didn’t know the value of their assets and those who reported a specific value were much
larger than when comparing the differences between refusals and those who provided an
exact asset value.
Insert Table A1
Indeed, we ran subsequent regression analyses to further explore the descriptive
differences presented in Table A1. The first set of analyses was for a sample that included
respondents who either reported a specific value of their retirement account balance or
refused to answer. We estimated the probability of refusal. An identical analysis was done for
a sample that included only those who either reported a value of these accounts or reported
they did not know that value. The test was whether the predictive variables were similar—
implying refusals and don’t knows are virtually the same answer—or different, which would
imply the validity of their distinct treatment. Gender played an important, different, and
expected role in these sensitivity analyses. First and foremost, the predictive value of gender
was much smaller for a refusal than a ‘don’t know’ response. This is consistent with the
literature that finds women possess lower financial literacy than do men. This leads us to
conclude that our ‘don’t know’ predictors do predict knowledge, while refusals are caused by
wealth-related factors. Refusing to answer questions on financial measures is
disproportionately high among those with very high asset levels (Juster and Smith 1997).
These gender differences inspired us to run these sensitivity models separately for
men and women. For men, IQ and academic measures were not predictive of refusal to
answer asset questions. In contrast, for men, education and IQ was predictive of ‘don’t’
know’ answers, and in the expected direction. Among women, however, we found a more
complicated story. We found evidence of reporting bias in the opposite direction than what
we initially assumed. It appears women may have been more likely to refuse when they did
not know asset amounts. Women who refused to answer had lower IQ scores and were less
likely to have taken advanced math courses in high school. In sum, the evidence of these
sensitivity analyses indicates, that if anything, our exclusion of refusals from our sample may
actually underestimate the effects of early-life schooling, especially for women.
References
Agarwal, Sumit and Mazumder, Bhashkar (2010). ‘Cognitive Abilities and Household Financial
Decision Making.’ Accessed June 30, 2010 at SSRN: http://ssrn.com/abstract=1651312
Agnew, Julie R., and Lisa R. Szykman (2005). ‘Asset Allocation and Information Overload: The
Influence of Information Display, Asset Choice, and Investor Experience,’ Journal of
Behavioral Finance, 6(2): 57–70.
ANZ Bank (2008). ‘ANZ Survey of Adult Financial Literacy in Australia.’ Accessed March
11, 2009, at
http://www.anz.com/Documents/AU/Aboutanz/AN_5654_Adult_Fin_Lit_Report_08_
Web_Report_full.pdf.
Banks, James and Oldfield, Zoë (2007). ‘Understanding Pensions: Cognitive Function,
Numerical Ability and Retirement Saving,’ Fiscal Studies, 28 (2): 143.
Barron, John M., and Staten, Michael E. (2009). ‘Is Technology-Enhanced Credit Counseling as
Effective as In-Person Delivery?’ Presented at the Networks Financial Institute at Indiana
State University Conference: Improving Financial Literacy and Reshaping Financial
Behavior, May 14-15, 2009. http://www.networksfinancialinstitute.org/News/Documents
/Staten.pdf.
Becker, Gary (1962). ‘Investment in Human Capital: A Theoretical Analysis,’ The Journal of
Political Economy, 70(5): 9-49.
Bell, Catherine, Daniel Gorin, and Jeanne M. Hogarth. (2009). ‘Does Financial Education Affect
Soldiers’ Financial Behaviors?,’ Presented at the 2009 Federal Reserve System
Community Affairs Research Conference, April 16-17, 2009.
Bernheim, Douglas (1995). ‘Do Households Appreciate Their Financial Vulnerabilities? An Analysis
of Actions, Perceptions, and Public Policy,’ Tax Policy and Economic Growth (Washington,
DC: American Council for Capital Formation), 1-30.
Bernheim, Douglas (1998). ‘Financial Illiteracy, Education and Retirement Saving,’ in O. Mitchell
and S. Schieber, Living with Defined Contribution Pensions (Philadelphia: University of
Pennsylvania Press), 38-68.
Bernheim, B. Douglas, and Daniel M. Garrett (2003). ‘The Effects of Financial Education in the
Workplace: Evidence from a Survey of Households,’ Journal of Public Economics, 87(7–8):
1487–1519.
Bernheim, B. Douglas, Daniel M. Garrett, and Dean M. Maki (2001). ‘Education and Saving: The
Long-Term Effects of High School Financial Curriculum Mandates,’ Journal of Public
Economics, 80(3): 435–465.
Chan, Sewin, and Ann Huff Stevens (2003). ‘What You Don't Know Can't Help You: Pension
Knowledge and Retirement Decision Making.’ NBER Working Paper No. 10185.
Cambridge, MA: National Bureau of Economics Research.
Cole, Shawn A. and Shastry, Gauri K. (2009). ‘Smart Money: The Effect of Education,
Cognitive Ability, and Financial Literacy on Financial Market Participation,’ Harvard
Business School Working Paper No. 09-071.
Collins, Michael, and Collin O’Rourke. (Forthcoming). ‘Financial Education and Counseling: Still
Holding Promise,’ Journal of Consumer Affairs.
DeVaney, Sharon A., Gorham, Elizabeth E., Bechman, Janet C. and Haldeman, Virginia A. (1996). ‘Cash
Flow Management and Credit use: Effect of a Financial Information Program. Financial
Counseling and Planning, 7:71-80.
Financial Industry Regulatory Authority (FINRA) (2003). NASD Investor Literacy Research:
Executive Summary. Accessed March 11, 2009, at
http://www.finrafoundation.org/surveyexecsum.pdf.
Fitzgerald, John, Peter Gottschalk, and Robert Moffitt (1998). ‘An Analysis of Sample
Attrition in Panel Data,’ The Journal of Human Resources, 33(2): 251–199.
Fox, Jonathan, Suzanne Bartholomae, and Jinkook Lee (2005). ‘Building the Case for Financial
Education,’ The Journal of Consumer Affairs, 39(1): 195–214.
Gustman, Alan and Steinmeier, Tom (2004), 'What People Don't Know about Their Pensions
and Social Security', in William Gale, John Shoven, and Mark Warshawsky (eds.),
Private Pensions and Public Policies (Washington, DC: Brookings Institution), 57125.
Gustman, Alan and Steinmeier, Thomas, and Tabatabai, Nahid. (2010). ‘Financial
Knowledge and Financial Literacy at the Household Level.’ Michigan Retirement
Research Center, Working Paper 2010-223, Ann Arbor, MI.
Hilgert, Marianne A., Jeanne M. Hogarth, and Sandra G. Beverly (2003). ‘Household Financial
Management: The Connection between Knowledge and Behavior,’ Federal Reserve Bulletin,
89(7): 309-322.
Hogarth, Jeanne M., and Marianne A. Hilgert (2002). ‘Financial Knowledge, Experience and
Learning Preferences: Preliminary Results from a New Survey on Financial Literacy,’
Consumer Interest Annual, 48.
Holden, Karen, Sara Kock, and Raghav Mohan (2010). ‘Making the Irrational Rational: Theories of
Human Psychology and What they Say about Financial Security.’ Presented at the Center for
Financial Literacy Research Conference, August 23, 2010, Madison, WI.
Juster, Thomas and Smith, James (1997). ‘Improving the Quality of Economic Data: Lessons from
the HRS and AHEAD,’ Journal of the American Statistical Association, 92(440):1268-1278.
Lerman, Richard (2009). ‘Apprenticeship in the United States: Patterns of Governance and
Recent Developments,’ pp. 125-140 in Erica Smith and Felix Rauner, Rediscovering
Apprenticeship: Research Findings of the International Network on Innovative
Apprenticeship (INAP). Dusseldorf, Germany: Springer-Verlag.
Link, Bruce G. and Phelan, Jo (1995). ‘Social Conditions As Fundamental Causes of
Disease,’ Journal of Health and Social Behavior, 35, Extra Issue: Forty Years of
Medical Sociology: The State of the Art and Directions for the Future: 80-94.
Lusardi, Annamaria (2003). ‘Planning and Savings for Retirement,’ Working Paper. Hanover, NH:
Dartmouth College.
Lusardi, Annamaria (2008a). ‘Household Saving Behavior: The Role of Financial Literacy,
Information, and Financial Education Programs.’ NBER Working Paper No. 13824.
Cambridge, MA: National Bureau of Economics Research.
Lusardi, Annamaria (2008b). ‘Financial Literacy: An Essential Tool for Informed Consumer
Choice?’ Working Paper. Hanover, NH: Dartmouth College.
Lusardi, Annamaria, and Olivia S. Mitchell (2005). ‘Financial Literacy and Planning: Implications
for Retirement Wellbeing.' Research Paper No. WP 2005–108. Ann Arbor, MI: Michigan
Retirement Research Center.
Lusardi, Annamaria, and Olivia S. Mitchell (2007a). ‘Baby Boomer Retirement Security: The Roles
of Planning, Financial Literacy, and Housing Wealth,’ Journal of Monetary Economics,
54(1): 205–224.
Lusardi, Annamaria, and Olivia S. Mitchell (2007b). ‘Financial Literacy and Retirement
Preparedness: Evidence and Implications for Financial Education,’ Business Economics,
42(1): 35–44.
Lusardi, Annamaria, and Olivia S. Mitchell (2007c). ‘Financial Literacy and Retirement Planning:
New Evidence from the Rand American Life Panel.’ Research Paper No. WP 2007–157.
Michigan Retirement Research Center.
Lusardi, Annamaria, and Olivia S. Mitchell (2008). ‘Planning and Financial Literacy: How Do
Women Fare?,’ American Economic Review, 98(2): 413–417.
Lusardi, Annamaria, and Olivia S. Mitchell (2009). ‘How Ordinary Consumers Make Complex
Economic Decisions: Financial Literacy and Retirement Readiness.’ NBER Working Paper
No. w15350. Cambridge, MA: National Bureau of Economics Research.
Lusardi, Annamaria, and Tufano, Peter (2009). ‘Debt Literacy, Financial Experiences, and
Overindebtedness’, NBER Working Paper Series, No. w14808. Cambridge, MA: National
Bureau of Eonomics Research.
Lyons, Angela C., and Eric Scherpf (2004). ‘Moving from Unbanked to Banked: Evidence from the
Money Smart Program,’ Financial Services Review, 13(3): 215–231.
Lyons, Angela C., et al. (2006). ‘Are We Making the Grade? A National Overview of Financial
Education and Program Evaluation,’ Journal of Consumer Affairs, 40(2): 208–235.
Mandell, Lewis (2004). ‘Financial Literacy: Are We Improving?’ Results of the 2004 National
Jump$tart Survey, Washington, DC: Jump$tart Coalition for Personal Financial Literacy.
Mandell, Lewis (2007). ‘Financial Literacy of High School Students,’ in Jing Jian Xiao, Handbook
of Consumer Finance Research. New York, NY: Springer, 163–183.
Mastrobuoni, Giovanni (2005). ‘Do Better-Informed Workers Make Better Retirement Choices? A
Test Based on the Social Security Statement.’ Unpublished Mimeo. Princeton, NJ: Princeton
University.
McArdle, John J., James P. Smith and Robert Willis (2009). ‘Cognition and Economic Outcomes in
the Health and Retirement Survey,’ NBER Working Paper No.15266. Cambridge, MA:
National Bureau of Eonomics Research.
Moore, Danna (2003). ‘Survey of Financial Literacy in Washington State: Knowledge, Behavior,
Attitudes, and Experiences.’ Technical Report n. 03-39. Social and Economic Sciences
Research Center. Pullman, WA: Washington State University.
National Association of State Boards of Education (NASBE) (2006). ‘Who Will Own Our Children?
The Report of the NASBE Commission on Financial and Investor Literacy.’ Executive
Summary. Arlington, VA: NASBE.
National Council for Economic Education (NCEE) (2005). ‘What American Teens and Adults Know
about Economics.’ Accessed February 25, 2009, at
http://www.ncee.net/cel/WhatAmericansKnowAboutEconomics_042605-3.pdf.
Pallier, Gerry, Roberts, Richard, and Stankov, Lazar (2000). ‘Biological versus Psychometric
Intelligence: Halsteads (1947) Distinction Revisited.’ Archives of Clinical Neuropsychology,
15(3):205-226.
Peters, Ellen, et al. (2006). ‘Numeracy and Decision Making,’ Psychological Science, 17(5): 407–
413.
Schmeiser, Maximilian, Karen Walsh, Fang Xiaodong, and Wang Xiao (2010). ‘Trigger Events and
Financial Outcomes across the Life Course.’ Presented at the Center for Financial Literacy
Research Conference, August 23, Madison, WI.
Schultz, Theodore (1962). ‘Reflections on Investment in Man.’ The Journal of Political Economy,
70(5), 1-8.
United States Government Accountability Office (GAO) (2004). ‘Highlights of a GAO
Forum: The Federal Government’s Role in Improving Financial Literacy’. Retrieved
from
http://www.gao.gov/docdblite/summary.php?rptno=GAO-05-93SP andaccno=A13412.
Van Rooij, Maarten, Annamaria Lusardi, and R. J. M. Alessie (2007). ‘Financial Literacy and Stock
Market Participation.’ Research Paper No. 2007-162. Ann Arbor, MI: Michigan Retirement
Research Center.
Vedder, R. (2007). Over-invested and Over-priced: American Higher Education Today. Center
for College Affordability and Productivity: Washington DC.
Wallace, Geoffrey, Robert Haveman, Karen Holden, and Bobbi Wolfe (2010). ‘Trigger Events and
Financial Outcomes in Older Households.’ Presented at the Center for Financial Literacy
Research Conference, August 23, Madison, WI.
Way, Wendy, and Ang, Ling See (2010). ‘Harnessing the Power of Technology to Enhance
Financial Literacy Education and Personal Financial Well-Being: A Review of the
Literature, Proposed Model, and Action Agenda.’ Presented at the Center for
Financial Literacy Research Conference, August 23, Madison, WI.
Willingham, Warren, Judith Pollack, and Charles Lewis. (2002). ‘Grades and Test Scores:
Accounting for Observed Differences,’ Journal of Educational Measurement, 39(1): 1–
37.
The Financial Literacy Research Consortium
The Financial Literacy Research Consortium (FLRC) consists of three multidisciplinary
research centers nationally supported by the Social Security Administration. The goal of this
research is to develop innovative programs to help Americans plan for a secure retirement.
The Center for Financial Security is one of three FLRC centers and focused on saving and
credit management strategies at all stages of the life cycle, especially helping low and
moderate income populations successfully plan and save for retirement and other life events,
including the use of Social Security's programs.
The Center for Financial Security
The Center for Financial Security at the University of Wisconsin-Madison conducts applied
research, develops programs and evaluates strategies that help policymakers and practitioners
to engage vulnerable populations in efforts which build financial capacity. The CFS engages
researchers and graduate students through inter-disciplinary partnerships with the goal of
identifying the role of products, policies, advice and information on overcoming personal
financial challenges.
For More Information:
Center for Financial Security
University of Wisconsin-Madison
Sterling Hall Mailroom B605
475 N Charter St.
Madison, WI 53706
(608) 262-6766
cfs@mailplus.wisc.edu
http://www.cfs.wisc.edu/