Massey Universityeconfin.massey.ac.nz/.../JScott_FSRet_Plan_Leak.docx · Web viewb Assistant Professor, University of Texas Pan American, College of Business Administration, Edinburg,

1

Financial Sophistication and Retirement Plan Leakage

Janine K. Scott, Ph.D.a,*, Terrance Martin, Ph.D., MBA b, Philip Gibson, Ph.D., CFP c

a Associate Professor, School of Economics & Finance, Massey University, Palmerston North,

New Zealand

* Corresponding author. Phone +64 6 3569099, ext. 83844; fax: +64 6 3505660; email address: [email protected]

b Assistant Professor, University of Texas Pan American, College of Business Administration, Edinburg, Texas; phone: Phone: 956-665-7358; email address: [email protected]

c Assistant Professor, Winthrop University, College of Business Administration, Rock Hill, South Carolina; phone: 803 323 2686; fax: 803 323 3960; email address: [email protected]

ABSTRACT

Today, American workers bear greater responsibility saving for retirement. To encourage and to

facilitate the accumulation of retirement savings, the U.S. government provides significant tax

incentives to households that participate in qualified and tax-advantaged accounts. However, the

benefits of these tax incentives are sometimes diminished when some individuals meet current

consumption needs with retirement plan loans and withdrawals. Using data from the 2009

FINRA National Financial Capability Study (NFCS), we examine the relation between financial

2

sophistication and the decision to make retirement fund withdrawals and utilize plan loans. The

results show that respondents who demonstrate a low level of financial sophistication are more

likely to withdraw funds from retirement accounts in addition to having plan loans. Additionally,

results suggest that a respondent’s financial risk willingness, income, and the number of children

may also affect the decision to make withdrawals and take loans against retirement accounts.

3

INTRODUCTION

Defined contribution plans such as 401(k)s and individual retirement plans (IRAs) have

become more prevalent than defined benefit plans. As a result, employees have more control

over the portion of their total wealth that they will need to draw on during the decumulation life

cycle phase. While the widespread adoption of defined contributions plans places less burden on

employers from a funding standpoint, many employees may be making ill-equipped and/or ill-

advised decisions which can severely impact retirement and terminal wealth. Using the 2009

FINRA National Financial Capability Study (NFCS) this paper examines the characteristics of

households who make retirement plan withdrawals and those who borrow from retirement plans.

Extending beyond existing research, we examine the motivation for those who withdraw as well

as take plan loans, and how these decisions are moderated by financial sophistication.

Based on current U.S. regulation, certain retirement plans may allow participants to

receive hardship withdrawals from their elective deferral account. A distribution can only be

made if there is an immediate and heavy financial need and is restricted to the amount necessary

to satisfy that financial need. Those who take hardship withdrawals are subject to income taxes

in addition to a ten percent tax penalty. Once a distribution is taken, it cannot be repaid.

Furthermore, those who take a hardship distribution are prohibited from making any

contributions to that account for six months. As an alternative to taking a hardship distribution,

it is possible for some individuals to take loans from defined-contribution plans such as a 401(k),

403(b), or a similar non defined-benefit plan. If a loan is taken for a general need it must be

repaid in five years; however, if the loan is purposed to finance the employees’ principal

residence, the repayment period is extended. When an individual takes a loan from their

4

retirement account, there is loss in capital appreciation, income and compounding returns,

potentially increasing shortfall risk in retirement.

Through our findings we are able to provide a better understanding of households who

are making financial mistakes by withdrawing funds and taking plan loans against their

retirement accounts. The insights may add value to the implementation of policies and programs

designed to aid and educate specific segments of the population.

LITERATURE REVIEW

Who Are More Likely to Withdraw and Why?

Despite the tax penalties and other employer specific penalties that are associated with

early withdrawal from retirement accounts such as IRAs and 401(k)s, households still make early

withdrawals from these plans. Understandably, households may need to withdraw funds from

retirement accounts for hardship reasons. This typically includes meeting expenses for: housing,

college, health issues, and the funding of household needs after job loss or change in occupation

(Argento, Bryant, and Sabelhaus 2013; Butrica, Zedlewski, and Issa 2010; Brady 2011;

Copeland 2009).

Significant factors influencing a household having an outstanding 401(k) loan include:

age, education, income, job tenure, employee contribution amount (Utkus and Young 2011).

Additionally, prior studies have included marital status, health, non-retirement wealth, existing

household debt and home ownership as part of the retirement withdrawal/loan decision (Argento,

Bryant, and Sabelhaus 2013; Utkus and Young 2011). Contradictory to previous findings

associated with income, Utkus and Young (2011) find greater loan usage among households with

5

income of $100,000 or more. Moreover, greater resources also increased the likelihood of having

a 401(k) loan.

Financial Sophistication and Retirement Savings Behavior

Financial knowledge coupled with positive financial behavior (or more specifically,

action) results in increased financial literacy. Simply knowing how to prepare for retirement will

not necessarily translate into employing an effective retirement strategy. There are not only

present and future saving and consumption preferences to be taken into consideration;

unpredictable or uncertain economic and investment variables (to a certain extent) such taxes,

inflation, and rates of return makes the planning processing a bit more complex than at first

glance (Campbell et al. 2011; Mitchell and Utkus 2004). The knowledge, discipline, and time

consumption involved in the retirement saving decision can be complex and oftentimes

overwhelming. It is not difficult to understand the reliance on heuristics to simplify the planning

decisions.

Utkus and Young (2011) examine both 401(k) loan usage and financial literacy. Citing

the threat to retirement security from 401(k) plan withdrawals, they also highlight that

unsophisticated investors may suffer more when compared to more informed investors. For

example, following job termination, employees with outstanding 401(k) loans may be

susceptible to distribution penalties if those loans remain unpaid. Moreover, plan loans are not

only costly due to interest expense but also costly due to an investment opportunity cost. Missed

return opportunities as a result of decreased investment assets can have long-term effects on

household wealth.

6

Alternatively, some investors who borrow from retirement plans may be viewed as

financially sophisticated; since retirement plan loans often carry a lower interest rate than

traditional sources of consumer loans, this type of loan selection is seemingly a more optimal

choice (Tang and Lu 2014; Utkus and Young 2011). Retirement plan borrowers may also be loan

savvy in that they are borrowing from their own wealth instead of from a profit-seeking entity.

Senior households (age 71 – 80) who make IRA withdrawals subsequently have increased

savings (e.g. Cds and government securities), whereas late boomer households (age 61 – 70) who

make IRA withdrawals subsequently spend most of the funds (Banerjee 2013). Wenger and

Weller (2013) conclude that households who have access to a defined contribution plan with a

loan option benefit more from having this option if they already exhibit stable borrowing habits.

These households are also more likely to have higher savings in their defined contribution plans.

Behavioral Mistakes as a Result of Low Financial Sophistication

The behavioral finance literature also provides instances of less than rational investor

behavior in retirement planning. Economists have pointed out that even though many individuals

try to save for retirment, unfortunately many are limited by the complexity of the decision-

making process, the will-power to persevere through the planning process and the ability to

execute intentions (Thaler and Shefrin 1981; Simon 1955). Furthermore, many individuals fail to

save due to time preference—placing greater value on present day consumption at the expense of

future consumption needs (Campbell 2006; Frederick, Loewenstein, and O'donoghue 2002).

FINRA (2009) cite that previous borrowing habits and/or current loan usage also play a

role retirement plan withdrawal. Utkus and Young (2011) also find that individuals with 401(k)

plan loans also fail to repay credit card monies and exhibit lower retirement contribution rates.

7

Present-oriented individuals will tend to have insufficient wealth at retirement (Campbell et al.

2011; Eckel, Johnson, and Montmarquette 2005). For an extensive review of behavioral

influences on both household retirement and non-retirement wealth see Campbell (2006),

Mitchell and Utkus (2004) and Thaler and Shefrin (1981).

THEORY

The life cycle model of consumption and savings (Modigliani and Brumberg 1954)

provides an adequate theoretical framework for this study. In order to maximize utility,

consumers will select a consumption path that suits them best. To maintain a constant marginal

utility of consumption, individuals borrow when earnings are low, save when earnings are high,

and dissave during retirement. Consumption across time is determined by expected changes in

real income, changes in future income, personal discount factor, household size, and the real

interest rate (Yuh and Hanna 2010).

The life cycle model makes the assumption that an individual’s optimal level of

consumption is related to their total lifetime wealth. If households dissave or reduce current net

assets during the accumulation phase, there will be a negative impact on current wealth but

moreover, on lifetime wealth. As mentioned earlier, this reduction in wealth can be attributed to

taxes and penalties associated with the early withdrawal of retirement account funds in addition

to the loss of compounded investment returns.

METHODS

Data

The data sample used for this study comes from the 2009 National Financial Capability

Study commissioned by FINRA Education Foundation. The study contains three surveys: a

8

national survey with 1,488 respondents, a state-by-state survey with 28,146 respondents, and a

military survey with 800 respondents. For this study the state-by-state dataset was used since it is

highly representative and asks relevant questions specific to our study. The data were censored to

respondents ages 18 to 64, within the accumulation stage of their life-cycle. The final sample

size used for the study is comprised of 23,229 respondents, of which 3.24% reported having

made a withdrawal in the past year and 4.47% reported taking a loan against retirement assets.

Table 1 provides more detailed information about the sample.

[Insert Table 1 about here]

Empirical Models

The study examines two questions provided in the 2009 NFCS. First, respondents were

asked, “In the last 12 months, have you (or your spouse/partner) taken a hardship withdrawal

from your retirement account(s)?” Additionally, respondents were asked, “In the last 12 months,

have you (or your spouse/partner) taken a loan from your retirement account(s)?” Dichotomous

variables were created from these questions. If respondents answered yes to either question, they

were assigned ‘1’ and if no, ‘0’. We also investigate the characteristics of individuals who made

both a withdrawal and used a loan from their retirement account in the past year. If respondents

said yes to both questions, they were assigned ‘1’ and ‘0’ otherwise.

Using a binary logistic regression model, the decision to withdraw from and/or take a

qualified retirement plan loan is modelled as a function of financial sophistication and other

factors found empirically related to the decision to save for retirement.

The primary predictor variable is financial sophistication. To estimate a respondent’s

level of sophistication, a factor score was created. Respondents who participated in the 2009

9

NFCS were asked five financial literacy questions which serves as a proxy for financial

capability. The questions as stated in the survey are below.

(1) Suppose you had $100 in a savings account and the interest rate was 2% per year. After

5 years, how much do you think you would have in the account if you left the money to

grow?

(2) Imagine that the interest rate on your savings account was 1% per year and inflation was

2% per year. After 1 year, how much would you be able to buy with the money in this

account?

(3) If interest rates rise, what will typically happen to bond prices?

(4) A 15-year mortgage typically requires higher monthly payments than a 30-year

mortgage, but the total interest paid over the life of the loan will be less.

(5) Buying a single company's stock usually provides a safer return than a stock mutual fund.

Various arrangements/modifications of these five questions have been used by researchers to

capture the level of financial literacy among respondents (Knoll and Houts 2012). The

aforementioned questions have been previously asked in national surveys such as the Health and

Retirement Study, the 1997 National Longitudinal Survey of Youth, and the RAND American

Life Panel (Allgood and Walstad 2013). Lusardi, Mitchell, and Curto (2009) assert that a

substantial percentage of U.S. respondents have difficulties selecting the correct response for

these five basic finance questions. In prior studies, researchers have employed the answers to

some or all of these questions as proxies for financial sophistication (Huston, Finke, and Smith

2012) by producing indices or other linear combinations (Allgood and Walstad, 2013). However,

this study ventures beyond previous studies by creating a factor to proxy for investor

sophistication using all questions.

10

Dichotomous variables were created from each of the five questions. We then used factor

analysis to capture the latent concept of investor sophistication. Factor scores were estimated for

each respondent after which they were separated into quintiles. Respondents who answered all

five questions correctly were placed in the highest quintile of investor sophistication. For

purposes of the multivariate analysis, each quintile of financial sophistication is converted to a

binary variable and coded as 1 if the respondent is classified to that group and ‘0’ if otherwise.

(See Appendix 1 for the construction of other predictor variables).

The independent variables, justified by life cycle theory, include financial sophistication

and socio-economic variables. Demographic variables include marital status, gender, age, race,

education level, and the number of dependent children living at home. Additionally, we examine

economic variables such as income, wealth, being a homeowner and self-employment status.

Lusardi and Mitchell (2011) show that greater financial sophistication increases the

likelihood of planning for retirement. Consistently, Lusardi and Mitchell (2007) find that being

financially literate leads to greater wealth. Individuals who possess greater financial

sophistication should be more future-oriented and thus more likely to save and make informed

financial decisions. Therefore we anticipate seeing a lower withdrawal rate from individuals with

greater financial sophistication.

Married households might have more complicated goals as compared to single

households (Yuh and Hanna 2010). Unlike single households, married households face

additional financial challenges. For instance, consideration must be given to the possibility of

divorce or the death of a spouse. Yuh and Hanna (2010) suggest that it is possible that married

households are more future-oriented compared to un-married households. Prior literature also

shows that married households often save more than other types of households (Chang 1994;

11

Avery and Kennickell 1991). Additionally, in a dual income household, there may be greater

financial reserves. In a situation where one spouse experiences an interruption in income, the

other spouse may take on the additional financial responsibilities. Thus we expect to see a lower

withdrawal rate among married households when compared to single households.

Lusardi and Mitchell (2011) find that women are generally less financially

knowledgeable than males. Bryant and Zick (2006) cite that human capital may be greater

among males since they enter and exit the labor force at a lesser rate than females. However,

controlling for income, education, financial sophistication among other variables, there should be

no theoretical difference in the behaviors of males versus females.

Age is an integral part of the standard life cycle model as it dictates consumption

preferences and needs. Younger households are more likely to have greater loan obligations such

as those related student loans, mortgages, credit cards, car notes, etc. As a result, we anticipate

that younger individuals are more likely to withdraw funds from their retirement accounts.

Rha, Montalto, and Hanna (2006) show that white households are more likely to save

than non-white households. Likewise, Yuh and Hanna (2010) find that blacks were less likely to

save than white households, ceteris paribus. Accordingly, we expect that non-white households

are more likely to borrow from their retirement account when compared to white households.

Soloman (1975) argues that an individual’s level of education is related to their level of

time preference. More educated individuals may be more future-oriented displaying a lower time

preference. We hypothesize a negative relation between an individual’s level of education and

the likelihood of withdrawing funds from a retirement account.

The presence of dependent children induces higher levels of consumption, affecting the

ability to save (Bryant and Zick 2006). In the absence of dependent children, marginal utility of

12

consumption is lower and the marginal propensity to save is therefore greater (Deaton 2012).

Given a more constrained demand on financial resources, we expect that households with

dependent children are more likely to make withdrawals from their retirement accounts.

When compared to the average income household, Yuh and Hanna (2010) find that

households with very high incomes are more likely to save, while those with very low incomes

are less likely to save. Additionally, higher income generally means more access to credit. If a

household’s income exceeds their marginal utility of consumption then it is quite possible for a

household to have additional assets outside of a retirement account such as in an emergency

fund. So as income increases individuals should be less inclined to borrow from their retirement

account.

Regarding other financial variables, provided all else equal, net worth should be

negatively related to withdrawing funds from a retirement account. It is expected that wealthier

households who experience a decrease in income will have sufficient funds to maintain a

constant level of consumption without having to withdraw funds from their retirement account.

With regards to renters versus homeowners, we suspect that renters are more likely to

have savings outside of a retirement account as they might be saving to buy a house. Hence, in

the event of a financial shock, renters will be able to draw from savings. Homeowners on the

other hand, faced with the risk of foreclosure might be forced to borrow funds from their

retirement account. Therefore, we expect that homeowners are more likely to withdraw funds

from their retirement accounts.

Prior empirical evidence reveals that individuals with lower cognitive ability are more

likely to be impatient and less risk tolerant (Dohmen et al. 2010). Given the association between

risk tolerance, impatience, and lower levels of cognitive ability, households with a lower level of

13

risk tolerance may prefer current income consumption rather than delay consumption for the

future. Accordingly, households with greater risk tolerance will be more likely to withdraw funds

from their retirement accounts.

Income stability is dependent on working status and occupation. Those who are self-

employed might potentially face a riskier income stream compared to those who are not self-

employed (Browning and Lusardi 1996). Since self-employed individuals are more prone to

fluctuations in income, we anticipate that they are more likely to take hardship withdrawals from

their retirement accounts.

Descriptive Analysis

Figures 1 and 2 display retirement withdrawals and plan loans by financial sophistication

and risk tolerance, respectively. Individuals captured in the lowest quintile for financial

sophistication display a greater withdrawal rate than those who make plan loans. Within the top

two quintiles, respondents utilize plan loans at a greater percentage than respondents who make

withdrawals in those quintiles. Figure 2 demontrates withdrawals and plan loans by risk

tolerance. Respondents reporting lower risk tolerance use plan loans moreso than those who

withdraw funds at the lowest levels of risk tolerance.

[Insert Figures 1 and 2 about here]

Table 1 separates respondents into four groups. Individuals who withdrew funds from

their qualified retirement plans are compared to those who made no withdrawals. Similarly,

repondents who used a plan loan from a qualified retirement plan are compared to those who did

not take a loan from the plan. We perform chi-square tests to find any signficant difference

between these groups. As it relates to financial sophsitication and making withdrawals, the only

significant difference is found among respondents in the 3rd quintile of financial sophsitication.

14

As age increases from 18 to 54, there is a monotonic increase in fund withdrawals. Among the

age 18-24 group, 9.77% of respondents withdrew from retirement accounts compared to 28.47%

of respondents between ages 45-54. The data show only 13.42% of respondents in the oldest age

group report taking a withdrawal. Among those who withdrew funds from retirement accounts,

16.91% respondents report being willing to take the highest financial risk. A higher percentage

of whites, married couples, respondents with two or more children withdrew from retirement

accounts.

Similar patterns are observed among respondents reporting take loans on their retirement

plans. Among those who withdrew, 67.93% were homeowners; additionally, 70.52% of those

taking plan loans were also homeowners.

RESULTS

Logistic Regression

Table 2 shows the results from the binomial logistic regression model used to identify

those who have withdrawn funds from their retirement accounts. As suspected, the results show

that for respondents with a low level of financial sophistication the odds of withdrawing funds

from their retirement account is 86% greater when compared to those who demonstrate a high

level of financial sophistication. Furthermore, individuals who exhibit a high level of risk

tolerance were almost 193% more likely to withdraw funds from their retirement account

compared to those who are risk averse. Race is also identified as a significant factor; households

with White respondents exhibit a lower propensity to withdraw funds from their retirement

account when compared to those who are non-White. In examining levels of education, we find

that those who are not college educated are more inclined to withdraw funds from their

15

retirement account. As expected, household income has a negative relation on the decision to

withdraw funds from a retirement account. The results confirm that as households experience an

increase in income, the chances of withdrawing funds from their retirement account significantly

declines.

[Insert Table 2 about here]

The number of dependent children in a household significantly impacts the likelihood of

withdrawing funds from a retirement account. As the number of dependent children within a

household increases, there is a corresponding increase in the likelihood of withdrawing funds

from a retirement account. Being a business owner increases the chances of taking an early

distribution from a retirement account. The odds of taking an early withdrawal are 99% greater

among business owners than among those who are not self-employed. Those who are employed

are less likely to withdraw funds from their retirement account.

        The results displayed in Table 2 also demonstrate the odds of taking plan loans.

Regardless of the level of financial sophistication, it appears that when in need all households are

likely to borrow funds from their retirement account. However, households characterized with

the lowest level of financial sophistication were 47% more likely to borrow funds from their

retirement account compared to those with a high level of financial sophistication. Similar to the

descriptive statistics displayed in Table 1, we see that respondents who are more risk tolerant

demonstrate greater odds (152%) of borrowing funds from their retirement account compared to

those who are highly risk averse. Relative to similar non-White households, White households

are less likely to borrow funds from their retirement account. Being more educated also reduces

the odds of borrowing funds from your retirement savings. Those who have a college degree or

higher were significantly less likely to take funds from their retirement account. As the number

16

of dependent children in a household rises, so does the odds of taking a plan loan. Being a

business owner is positively associated with the taking a plan loan. The odds of a business owner

taking a loan from their retirement account are 163% greater than someone who is not self-

employed.

Table 3 displays the results of the binomial logistic regression performed on those who

made an early withdrawal and utilized a loan from their retirement account. Overall, the results

are consistent with the prior findings. Respondents who display a low level of financial

sophistication were more prone to withdrawing funds from their retirement account compared to

those who are more financially sophisticated. Likewise, those respondents who identified

themselves as being risk tolerant exhibit a greater tendency of taking funds from their retirement

account. As expected, individuals experiencing economic hardship (especially without

emergency funds in place) are more likely to withdraw and use loans. Whites are less likely to

take funds from their retirement account compared to non-Whites. Similar to earlier findings, the

results also show that business owners are more likely to take funds from their retirement

accounts.

[Insert Table 3 about here]

CONCLUSION AND IMPLICATIONS

Saving for retirement should be an important financial goal for all households regardless

of age, race, or creed; however, can Americans afford to retire? This is a question that have been

posed by researchers (see Mitchell and Moore 1998; Skinner 2007). Both Mitchell and Moore

(1998) and Skinner (2007) find that most households face the risk of not having enough saved to

smooth consumption during retirement. Yet research shows that a number of Americans are

taking early withdrawals from their retirement accounts (Argento, Bryant, and Sabelhaus 2014).

17

Using the National Financial Capability study, we examine the relation between financial

sophistication and the decision to make early retirement withdrawals and/or utilize plan loans.

Similar to Argento, Bryant, and Sabelhaus (2014), we acknowledge that retirement withdrawals

or loans may allow households to increase consumption during periods when they need funds the

most. However, as employers (public and private) increasingly offer defined contribution plans,

taking early retirement withdrawals or plan loans may lead to greater disutility from decreased

consumption as these individuals move into retirement. For example, they may be forced to

extend their retirement age and/or be reduced to a lower standard of living for a number of years.

Our logistic regression results demonstrate those households captured in the lowest

quintiles of financial sophistication are more likely to make withdrawals and use plan loans. As

education and income rises, the likelihood of making withdrawals decreases among households

captured in our study. Additionally, as the number of children increase the odds of withdrawing

retirement funds increase also. This is expected as greater needs may arise in a household with a

greater number of children present.

Consistent with our hypothesis, self-employed households demonstrate greater odds of

making withdrawals and using plan loans than non-business owners. Given fluctuations in

income this is not surprising. Race is also a significant factor; Whites are less likely to make

withdrawals and use loans, consistent with prior research.

Even with tax penalties and commitment devices associated with defined contributions

accounts (e.g. automatic withdrawal and employer match), households still deviate retirement

funds from their primary purpose. We cannot ascertain if we are examining hyperbolic

consumers or just consumers who have hit a temporary financial snag. However, the primary

18

intent and goal of a retirement account is to provide income during retirement. Perhaps

increasing the withdrawal penalty and having stricter loan provisions is worth investigating

among policymakers.

Further research can also examine retirement withdrawal/loan decisions and the impact

on total wealth at retirement. Perhaps incorporating terminal wealth effects and providing this

information to consumers in a dynamic way can help consumers fully evaluate their decision to

tap into retirement savings.

In addition to households who save too little, have under diversified portfolios, practice

little self-control, based on the findings of this study, households with lower financial

sophistication are also at risk for having inadequate funds at retirement. Outside of higher

education (a significant factor in the retirement withdrawal/loan decision), personal finance

education providers alongside financial advisors may be able to aid these households realize

retirement goals. However, it can be debated whether or not those with lower financial

sophistication actually seek the aid of financial advisors. Thus, it is paramount that policymakers

and personal education providers continue to design and develop better means of reaching the

masses.

19

REFERENCES

Allgood, S., and William Walstad. 2013. Financial Literacy and Credit Card Behaviors: A Cross-Sectional Analysis by Age. Numeracy, 6(2): 3.

Argento, Robert, Victoria L. Bryant, and John Sabelhaus. 2013. Early withdrawals from retirement accounts during the Great Recession. Finance and Economics Discussion Series, Federal Reserve Board Working Paper, 22.

Banerjee, Sudipto. 2013. IRA Withdrawals: How Much, When, and Other Savings Behavior. EBRI Notes, 34(5).

Brady, Peter . 2011. The Role of IRAs in the Retirement Income Process.

Browning, Martin and Annamaria  Lusardi. 1996. Household saving: Micro theories and micro facts. Journal of Economic literature, 1797-1855.

Bryant, W. Keith and Cathleen D. Zick. 2006. The Economic Organization of the Household. New York, NY: Cambridge Univ. Press.

Butrica, Barbara A., Sheila R. Zedlewski, and Philip Issa. 2010. Understanding Early Withdrawals from Retirement Accounts. Washington, DC: The Urban Institute.

Campbell, John Y., Howell E. Jackson, Brigitte C. Madrian, and Peter Tufano. (2011). Consumer financial protection. The Journal of Economic Perspectives, 25(1):91-113.

Campbell, John Y. 2006. Household Finance. Journal of Finance, 61(4):1553-1604.

Choi, James, David Laibson, and Brigitte Madrian. 2006. Reducing the Complexity Costs of 401 (k) Participation Through Quick Enrollment (TM) (No. w11979). National Bureau of Economic Research.

Copeland, Craig. 2009. Lump-Sum Distributions at Job Change. EBRI Notes, 30(1).

Deaton, Angus. 2012. Franco Modigliani and the life-cycle theory of consumption. PSL Quarterly Review.

Department of Labor, Employee Benefit Security Administration . 2008. Private Pension Plan Bulletin Historical Tables.http://www.dol.gov/ebsa/pdf/privatepensionplanbulletinhistoricaltables.pdf

Department of Labor, Employee Benefit Security Administration . 2010. Private Pension Plan Bulletin: Abstract of 2007 Form 5500 Annual Reports. http://www.dol.gov/ebsa/pdf/2007pensionplanbulletin.pdf

20

Dohmen, Thomas J., Armin Falk, David Huffman, and Uwe Sunde. 2010. Are risk aversion and impatience related to cognitive ability?. The American Economic Review,100 (3), 1238-1260.

Eckel, Catherine, Claude Montmarquette, and Cathleen Johnson. Saving decisions of the working poor: Short-and long-term horizons (Vol. 10, pp. 219-260). Emerald Group Publishing Limited.

FINRA (Financial Industry Regulatory Authority). 2009. Initial Report on the National Survey Component of the National Financial Capability Study. Washington, DC: FINRA Investor Education Foundation. http://www.finrafoundation.org/web/groups/foundation/@foundation/documents/foundation/p120536.pdf

Frederick, Shane, George Loewenstein, and Ted O'donoghue. 2002. Time discounting and time preference: A critical review. Journal of economic literature, 40(2), 351-401.

Huston, Sandra J., and Michael S. Finke. 2003. Diet choice and the role of time preference. Journal of Consumer Affairs, 37(1), 143-160.

Hung, Angela, Erik Meijer, Kata Mihaly, and Joanne Yoong. 2009. Building Up, Spending Down: Financial Literacy, Retirement Savings Management, and Decumulation. RAND Labor and Population Document de travail WR-712. Santa Monica, CA, RAND Corporation.

Knoll, Melissa AZ, and Carrie R. Houts. 2012. The financial knowledge scale: An application of item response theory to the assessment of financial literacy. Journal of Consumer Affairs, 46(3), 381-410.

Lusardi, Annamaria, and Olivia S. Mitchell. 2007. 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. 2011. Financial literacy and planning: Implications for retirement wellbeing (No. w17078). National Bureau of Economic Research.

Lusardi, Annamaria, Olivia S. Mitchell and Vilsa  Curto. 2009. Financial literacy and financial sophistication among older Americans (No. w15469). National Bureau of Economic Research.

Mitchell, Olivia S., & Moore, J. F. 1998. Can Americans afford to retire? New evidence on retirement saving adequacy. Journal of Risk and Insurance, 371-400.

Mitchell, Olivia S., and Stephen P. Utkus . 2004. Lessons from behavioral finance for retirement plan design. Pension design and structure: New lessons from behavioral finance, 3-42.

Modigliani, Franco, and Richard Brumberg. 1954. Utility analysis and the consumption function: An interpretation of cross-section data. FRANCO MODIGLIANI, 1.

21

Papke, Leslie E. 1995. Participation in and Contributions to 401 (k) Pension Plans: Evidence on Plan Data (No. w4199). National Bureau of Economic Research.

Poterba, James M., Steven F. Venti, and David A. Wise. 2011. The Drawdown of Personal Retirement Assets (No. w16675). National Bureau of Economic Research.

Simon, Herbert A. 1955. A Behavioral Model of Rational Choice. Quarterly Journal of Economics, 69: 99 – 118.

Skinner, Jonathan. 2007. Are you sure you're saving enough for retirement? (No. w12981). National Bureau of Economic Research.

Solmon, Lewis C. 1975. The relation between schooling and savings behavior: An example of the indirect effects of education. In Education, income, and human behavior (pp. 253-294). NBER.

Tang, Ning, and Timothy Lu. 2014. Are Your Clients Making the Right Loan Choice?.Journal of Financial Planning, 27(10): 39-47.

Thaler, Richard H., and Hersh M. Shefrin. 1981. An Economic Theory of Self-Control. Journal of Political Economy, 89: 392 – 406.

Utkus, Stephen P., and Jean A. Young. 2011. Financial Literacy and 401 (k) Loans. Financial Literacy: Implications for Retirement Security and the Financial Marketplace, 59-75.

VanDerhei, Jack, Sarah Holden, Luis Alonso, and Steven Bass. 2013. 401 (k) Participants in the Wake of the Financial Crisis: Changes in Account Balances, 2007-2011. EBRI Issue Brief, (391).

Wenger, Jeffrey B., and Christian E. Weller. 2013. Boon or Bane: 401 (k) Loans and Employee Contributions. Research on Aging, 36(5): 527-556.

Yuh, Yoonkyung, and Sherman D. Hanna. 2010. Which Households Think They Save?. Journal of consumer affairs, 44(1): 70-97.

22

FIGURE 1

Withdrawal and Plan Loan Usage by Financial Sophistication

Low In

vesto

r Sop

histic

ation

2nd quin

tile

3rd qu

intile

4th qu

intile

High In

vesto

r Sop

histic

ation

0%5%

10%15%20%25%30%35%40%

WithdrawalPlan Loan

23

FIGURE 2

Withdrawal and Plan Loan Usage by Risk Tolerance

Low Fi

nancia

l Risk

Willi

ngess

Risk2

Risk3

Risk4

High Fina

ncial

Risk W

illingn

ess0%5%

10%15%20%25%30%35%

WithdrawalPlan Loan

24

TABLE 1

Descriptive Statistics for Withdrawals and Plan Loans

ParametersWithdrawa

l

No Withdrawa

lPlan Loan

No Plan Loan

Low Investor Sophistication 23.83 21.78 18.75 ** 21.992nd quintile 23.48 23.22 21.91 23.293rd quintile 16.83 ** 19.95 19.77 19.854th quintile 21.32 20.21 25.07 *** 20.02High Investor Sophistication 14.55 14.84 14.5 14.85Male 52.95 50.07 51.38 50.11Age 18-24 9.77 *** 16.01 4.53 ***. 16.33Age 25-34 23.34 ** 19.98 25.42 *** 19.84Age 35-44 25 ** 21.47 28.15 *** 21.28Age 45-54 28.47 ** 22.92 26.92 ** 22.93Age 55-64 13.42 *** 19.62 14.99 ** 19.62Low Financial Risk Willingness 15.34 *** 28.83 16.61 *** 28.94Risk2 19.66 21.91 21.35 21.86Risk3 29.31 ** 24.41 26.69 24.47Risk4 18.79 17.86 20.79 ** 17.75High Financial Risk Willingness 16.91 *** 6.99 14.57 *** 6.98White 53.26 *** 65.55 58.96 *** 65.44<=High School 23.7 *** 31.82 25.83 *** 31.82Some College 47.65 ** 42.06 46.95 ** 42.02College 18.57 17.04 18.23 17.04Graduate 10.09 9.08 8.99 9.12Married 58.81 ** 52.51 66.46 *** 52.08Single 26.07 ** 32.34 20.77 *** 32.67Separated 2.62 1.9 2.02 1.92Divorced 11.04 11.31 9.85 11.37Widowed 1.47 1.93 0.9 1.97Less than 15000 5.66 *** 15.42 3.29 *** 15.6515K-25K 7.24 *** 12.5 3.65 *** 12.7425k-35k 11.05 12.34 10.56 * 12.3835k-50k 21.44 *** 15.53 16.03 15.7150k-75k 21.61 ** 18.4 26.82 *** 18.1275k-100k 15.02 ** 11.12 18.11 *** 10.93100k-150k 11.05 9.35 13.95 *** 9.19150k+ 6.92 * 5.33 7.59 ** 5.28One Child 19.35 17.59 18.6 17.61Two Children 21.54 ** 15.32 22.1 *** 15.22Three Children 10.96 *** 6.68 10.65 *** 6.64Four or more Children 6.82 *** 3.84 6.66 *** 3.81No child 41.32 *** 56.56 41.99 *** 56.72Homeowner 67.93 *** 54.63 70.52 *** 54.35Self Employed 10.12 8.61 6.75 ** 8.75Fulltime 54.68 *** 41.68 31.36 *** 59.13Part-time 9.01 10.02 7.2 ** 10.11No Labor Income 26.19 39.69 17.42 *** 40.27

25

*p<.10;**p<.05;***p<.01

26

TABLE 2

Logistic Regression Results for Withdrawals and Plan Loans

*p<.10;**p<.05;***p<.01

Withdrawal Plan Loan

Odds Ratio

Pr > ChiSq

Odds Ratio

Pr > ChiSq

Financial Sophistication (Ref: High )

Low Investor Sophistication 1.86 *** 1.47 ***

2nd quintile 1.29 * 1.31 **3rd quintile 1.11 1.40 **4th quintile 1.18 1.35 **

Gender: FemaleMale 1.17 0.97

Age (ref: 55-64)Age 18-24 0.90 0.55 **Age 25-34 0.85 1.00Age 35-44 0.90 0.97Age 45-54 1.13 0.93

Risk Tolerance (Ref: Low Willingness)

Risk2 1.28 1.35 **Risk3 1.45 ** 1.33 **Risk4 1.33 * 1.39 **High Financial Risk

Willingness 2.93 *** 2.52 ***Race (ref: non-white)

white 0.59 *** 0.65 ***Education (Ref: no college)

Some College 1.25 * 0.91College 0.97 0.59 ***Graduate 0.97 0.43 ***

Marital Status (ref: single)Married 0.80 0.92Separated 1.40 1.06Divorced 1.03 0.97Widowed 0.79 0.85

Income (ref: <$15000)15K-25K 0.68 0.50 **25k-35k 0.67 0.7735k-50k 0.74 0.8250k-75k 0.57 ** 0.9075k-100k 0.54 ** 0.95100k-150k 0.48 ** 0.90150k+ 0.32 *** 0.79

Children (ref: no children)One Child 1.25 * 1.18 *Two Children 1.53 *** 1.39 ***Three Children 1.81 *** 1.48 **Four or more Children 1.79 *** 1.71 ***

Homeowner 1.11 1.02Business Owner 1.99 *** 1.63 ***Employment (ref: no labor income)

Self Employed 0.77 0.79Fulltime 0.73 ** 1.48 ***Part-time 0.70 * 1.06

Pseudo R-Square 0.0688 0.0529

27

TABLE 3

Logistic Regression Results for Hardship Withdrawals and Plan Loans

Withdrawal with Hardship

Plan Loan with Hardship

Odds Ratio

Pr > ChiSq

Odds Ratio

Pr > ChiSq

Financial Sophistication (Ref: High )Low Investor Sophistication 1.95 *** 0.912nd quintile 1.31 ** 1.103rd quintile 1.12 1.40 **4th quintile 1.15 1.28 **

Gender: FemaleMale 1.30 ** 1.00

Age (ref: 55-64)Age 18-24 0.84 0.33 ***Age 25-34 0.80 0.78 *Age 35-44 0.86 0.79 *Age 45-54 1.07 0.75 **

Risk Tolerance (Ref: Low Willingness)Risk2 1.27 * 1.43 **Risk3 1.52 *** 1.37 **Risk4 1.36 ** 1.47 **High Financial Risk Willingness 2.57 *** 1.54 **

Race (ref: non-white)white 0.64 *** 0.80 **

Education (Ref: no college)Some College 1.19 0.83 *College 0.97 0.54 ***Graduate 0.98 0.36 ***

Marital Status (ref: single)Married 0.80 * 0.92Separated 1.41 1.08Divorced 1.04 0.93Widowed 0.76 0.84

Income (ref: <$15000)

28

15K-25K 0.66 * 0.6625k-35k 0.69 * 1.0435k-50k 0.73 1.1250k-75k 0.59 ** 1.4575k-100k 0.57 ** 1.63100k-150k 0.58 ** 1.70150k+ 0.43 ** 1.76

Children (ref: no children)One Child 1.10 0.97Two Children 1.34 ** 1.22 *Three Children 1.45 ** 1.05Four or more Children 1.34 1.11

Homeowner 1.20 * 1.08Business Owner 1.80 *** 1.01Employment (ref: no labor income)

Self Employed 0.74 * 0.84Fulltime 0.65 *** 1.67 ***Part-time 0.74 * 1.35 *

Economic Hardship

Overdraft 1.99 *** 1.65 ***Deficit Spending 1.57 *** 1.14No Emergency Fund 1.40 *** 2.44 ***Maxed out Credit Cards 2.25 *** 1.75 ***

Calculated Retirement Income Need 1.70 *** 1.37 ***Pseudo R-Square 0.137 0.0989

*p<.10;**p<.05;***p<.01

29

APPENDIX A

Coding of Independent Variables

Marital StatusMarried If yes then 1; otherwise 0Single If yes then 1; otherwise 0Separated If yes then 1; otherwise 0Divorced If yes then 1; otherwise 0Widowed If yes then 1; otherwise 0

GenderMale If yes then 1; otherwise 0

Age18-24 If yes then 1; otherwise 025-34 If yes then 1; otherwise 033-44 If yes then 1; otherwise 045-54 If yes then 1; otherwise 055-64 If yes then 1; otherwise 065 and older If yes then 1; otherwise 0

RaceWhite If yes then 1; otherwise 0

EducationNo College If yes then 1; otherwise 0Some College If yes then 1; otherwise 0College Graduate If yes then 1; otherwise 0Graduate School If yes then 1; otherwise 0

Dependent ChildrenNo Children If yes then 1; otherwise 0One Child If yes then 1; otherwise 0Two Children If yes then 1; otherwise 0Three Children If yes then 1; otherwise 0Four or More Children If yes then 1; otherwise 0

IncomeLess than $15,000 If yes then 1; otherwise 0$15,000-$24,999 If yes then 1; otherwise 0$25,000-$34,999 If yes then 1; otherwise 0$35,000-$49,999 If yes then 1; otherwise 0$50,000-$74,999 If yes then 1; otherwise 0$75,000-$99,999 If yes then 1; otherwise 0

$100,000-$149,999 If yes then 1; otherwise 0$150,000 and above If yes then 1; otherwise 0

HomeownershipHomeowner If yes then 1; otherwise 0

Employment StatusUnemployed/Temporary If yes then 1; otherwise 0

30

Laid Off (Ref)Self-Employed If yes then 1; otherwise 0Working Full-Time If yes then 1; otherwise 0Working Part-Time If yes then 1; otherwise 0

Risk Tolerance1-2 = Most Risk Averse (Ref)

If yes then 1; otherwise 0

3-4 If yes then 1; otherwise 05-6 If yes then 1; otherwise 07-8 If yes then 1; otherwise 09-10 = Most Risk Tolerant If yes then 1; otherwise 0