Individual Differences in 401(k) Plan Allocation Choices

Abstract

Insufficient retirement savings behavior practiced by a significant number of employees underscores the importance of understanding the characteristics of those individuals most likely to choose the default contribution and asset allocation vehicles of an organization. We conducted an experiment of savings behavior involving 217 subjects asked to make choices about allocating 401(k) funds and investing those funds in investment vehicles. Individuals who specified their own allocation amounts invested significantly more money versus those who accepted a default. Psychological and demographic factors also predicted higher investment dollars and higher-risk investment choices. Changes in ambiguity tolerance significantly affected men’s decisions to choose high-risk investments for their 401(k) dollars, but these psychological variables did not affect women’s high-risk investment decisions. We offer implications for future research and practice.

Keywords

pensions individual differences decision making experimental study

Beginning in the early 1980s, a major shift occurred in the type of retirement plans that employers provided to their employees. Presently, defined contribution retirement plans, largely in the form of 401(k) offerings, have replaced defined benefit plans as the dominant form of workplace retirement plan.¹ Defined contribution plans fund an employee’s retirement account based on annual contributions the employee and/or employer make as a percentage of pay. The benefits paid at retirement are dictated by the contributions credited to the account plus the account’s accumulated investment earnings.² The movement from defined benefit pensions to defined contribution plans necessitates that employees take greater responsibility for pension contribution amounts and asset allocation decisions than they had to before. These pension contribution and allocation decisions strongly influence the employee’s ultimate retirement income.³

Historically, most employer-provided 401(k) plans have left to the employee the decision of whether to participate in the plan, the amount of income to contribute and the allocation of investment decisions. Apparently daunted, in part, by these complicated financial decisions, as of 2006, approximately one quarter of employees who were eligible for defined contribution plans in fact refrained from participation, thus impairing their prospects for retirement security.⁴ Motivated to increase 401(k) participation rates, Congress enacted the Pension Protection Act of 2006 to encourage the implementation of 401(k) automatic enrollment features.⁵ Automatic enrollment removes the burden from the employee to voluntarily opt into the plan and, in the absence of affirmative selections by the employee, establishes default contribution rates and asset allocations for the employee. Automatic enrollment features for 401(k) plans have currently been adopted by over half of the employers offering defined contribution plans and have experienced enormous success in bolstering employee participation rates.⁶

While automatic enrollment features have succeeded in boosting 401(k) participation, the likely adequacy of income replacement in retirement for those employees who automatically enroll in the default contribution options has been found to be far lower compared to employees who contribute through traditional voluntary enrollment procedures.⁷ The inadequacy of retirement savings for employees who are automatically enrolled in 401(k) plans derives primarily from the low default contribution rate made to retirement 401(k) accounts by the majority of companies offering this feature and the tendency for employees who are automatically enrolled to keep their contributions at the default rate.⁸ Overly conservative default contribution options by organizations and the accompanying inertia of automatically enrolled individuals to remain committed to these options are of concern because of the ultimate negative consequences they have for employee retirement income security.

Insufficient retirement savings behavior practiced by a significant number of employees accentuates the importance of understanding the characteristics of those individuals most likely to choose the default contribution and asset allocation vehicles of an organization. Determining the characteristics of those employees most likely to choose default investment options would help organizations identify those employees most needing financial counseling about pension choices that would improve their pension choices. Understanding the characteristics of those individuals most likely to accept automatic default options of 401(k) plans would benefit organizations that have yet to implement automatic enrollment procedures.

The purpose of this article is to highlight the results of an investigation to discern the demographic and psychological characteristics of individuals most likely to elect the automatic default contribution rate and asset allocation of a 401(k) plan. We administered questionnaires and investment performance tasks to undergraduate and graduate students. The study’s objective was to identify determinants leading employees to select 401(k) automatic enrollment features. To that extent, we tested several hypotheses.

Background

On January 1, 2008, Kathleen Casey-Kirschling of Cherry Hill, New Jersey, became the first member of the baby boom generation to apply for Social Security retirement benefits. Casey-Kirschling’s application for retirement benefits signaled the start of the “silver tsunami,” the mass exodus of the Baby Boom generation from the work force.⁹ In the next two decades of the 21st century, approximately 80 million Americans will likely become eligible for Social Security benefits.¹⁰

Historically, retirement income in the United States has been based on Social Security, personal savings and employer-sponsored retirement plans.¹¹ Concerns about the future solvency of Social Security and Americans’ low personal savings rates have placed greater pressure than ever on employer-sponsored pension programs as a source of retirement income.

Coinciding with employees’ greater reliance on industrial pension plans as a source of economic security in retirement has been the shift away from employers offering defined benefit retirement plans to their workforce to employers offering defined contribution pension plans, mostly in the form of 401(k) plans.¹² While 25 years ago more than half of the Fortune 500 firms offered new hires a defined benefit plan, currently less than 20% now offer new employees access to these plans.¹³ Among all employers, participation of workers in defined contribution plans has risen to 88 million total participants.¹⁴

Employers are replacing defined benefit plans with defined contribution plans principally because of the expense and long-term commitment required by a defined benefit plan.¹⁵ Employers’ substitution of defined contribution plans for defined benefit plans also eliminates cost volatility involved in retirement plan funding.¹⁶ Another frequent factor employers cite for the decline of defined benefit plans is that the long-term costs of defined benefit plans were masked when the baby boom cohort was young, but these costs rose as the baby boomers aged. The cost of defined contribution plans does not increase nearly as sharply when the average age of an employer’s workforce increases.¹⁷

Defined benefit plans promise a specific monthly benefit to employees at retirement. Defined benefit plans calculate retirement income using a formula requiring employers to pay employees a pension based on factors such as the worker’s age at retirement, final years of salary and years of tenure.¹⁸ Unlike defined benefit plans, defined contribution plans do not guarantee a specific benefit during retirement. Defined contribution plans provide income based on annual contributions, as a proportion of pay in the interest of the employee and accrued investment gains on the employee’s account.¹⁹

In the last decade, as the prevalence of defined contribution plans grew and it became clear that most future retirees would rely on defined contribution retirement plans as the primary source of retirement income, concerns were raised about the low participation rates of employees covered by defined contribution plans.²⁰ In 2006, one quarter of employees eligible to participate in defined contribution plans chose not to take part.²¹ Participation rates were particularly low among newer employees (typically in their 20s) who would have the most to gain from long-term compounded earnings accumulating tax-free.²²

The complicated financial decision making involved with defined contribution plans required participants to determine how much income to contribute, and in what manner to allocate investment choices, and it resulted in many employees abstaining from participation, thus impairing their long-term economic security.²³ Insufficient defined contribution participation during income-earning years preceding retirement has been identified as one factor largely responsible for the current workforce’s projected retirement shortfalls.²⁴

The Pension Protection Act of 2006 sought to encourage workers to properly save for retirement and to improve involvement rates among employees eligible for participation in 401(k) plans by encouraging employers to introduce automatic enrollment provisions.²⁵ Traditional 401(k) plans typically require eligible workers to enroll in the 401(k) plan by voluntarily choosing to invest a percentage of their income toward the plan. Automatic enrollment measures permit the employer to automatically enroll an employee when he or she becomes eligible for the 401(k) benefit at a default salary deferral rate absent the election of a different deferral rate or statement disavowing participation by the employee. Automatic enrollment provisions also permit employers to select default investment vehicles for employees absent specification of an investment choice by the employee. Automatic enrollment or negative enrollment removes the burden on the employee to actively opt into the 401(k) plan and creates an opt-out system where the 401(k) plan makes automatic deductions for employees at default contribution rates implemented by the employer.²⁶

Austin²⁷ estimated that over half of employers who offer 401(k) plans have automatic enrollment provisions for employees. Automatic enrollment has been found to increase participation rates in defined contribution plans significantly. A study of over 2.5 million employees found that 90% of the workers participated in a 401(k) plan if their company automatically enrolled them in the plan compared to a 68% employee enrollment figure at organizations without automatic enrollment provisions.²⁸ Unfortunately, while automatic enrollment has significantly bolstered participation rates in 401(k) plans, it has failed to enhance the quality of retirement savings among automatically enrolled employees because default contribution rates are typically set at a contribution level of 3% or less of employee wages.²⁹ This modest default level is particularly problematic because employer contributions to employee 401(k) accounts are often based on matches to the employee’s contribution toward the plan.³⁰

Several possibilities may explain why employers typically set automatic enrollment features so conservatively. For one, employers may set these contribution levels at low levels to encourage employees to remain in the 401(k) plan. Higher default rates are more costly for employers as they result in heightened employer matching costs. Finally, plan sponsors can avoid administrative requirements of the Pension Protection Act if they implement an automatic safe harbor design that requires employers to automatically enroll workers at a rate of as low as 3% of pay in their first year of employment.³¹

Automatic enrollment default contribution and asset allocation rates have a powerful effect on individual savings rates for retirement. Pension research has found that the majority of automatically enrolled employees accept the default contribution rate and asset allocation vehicle established by the employer.³² Workers presumably adopt the default contribution level and asset allocation because it represents the least amount of effort for the worker. Participants may also perceive the default rate and asset allocation choice as being investment advice initiated by the company. Furthermore, the establishment of a default amount and allocation choice may serve as a form of psychological anchoring for individuals and may make alternative investment selections less likely.³³

The large number of employees electing conservative default contribution rates threatens the retirement income security of workers. Without high 401(k) plan default contribution rates and diversified asset allocation portfolios established by the organization, employees need to make well-informed and appropriate investment decisions to prepare adequately for retirement. The employees most likely to voluntarily choose automatic enrollment contribution default rates and asset allocation selections appear to be the most apt to ultimately generate insufficient retirement savings. As an example, employees who were automatically enrolled in organizations that had 401(k) plans with relatively high employee default contribution levels of 7% or 8% of wages were found to accept less robust 401(k) default levels when they joined other organizations.³⁴ Consequently, it becomes important that employers establish automatic enrollment provisions and set default options at relatively higher rates to enhance employee defined contribution participation and savings. In exchange for offering these provisions, automatic enrollment features have been found to be popular among employees bolstering the prospects of those employers who have adopted these features to recruit and retain employees.³⁵

Hypotheses Development

Considerable literature on investment decisions and risk has been collected in varied disciplines such as marketing, behavioral economics, psychology, decision sciences, management, risk and insurance, public policy and finance.³⁶ Increasingly, studies have examined individual investment decisions in retirement planning.³⁷ To date, however, no research has specifically explored the influence of demographic and psychological characteristics of individuals most likely to elect the automatic default contribution rate or asset allocation of a 401(k) plan. The present study fills a need by examining attributes of individuals most likely to elect the automatic default features of a 401(k) plan as well as the characteristics of individuals who affirmatively select different contribution rates and asset allocations from the default settings of the 401(k) plan.

A tendency toward passive decision making leads to employees choosing their company default savings rate and increasing the impact of corporate-set contribution amounts.³⁸ In fact, the adoption of a default contribution usually requires no action by the employee and does not, therefore, engage the employee in an active decision process. Passive investors are “slow to join advantageous plans; they make infrequent changes; and they adopt naive diversification strategies.”³⁹ Therefore, we expected passive decision making to affect the amount of subjects’ 401(k) allocations, and also affect whether subjects actively choose investment options from among those offered for their 401(k) funds.

Hypothesis 1a: Individuals who specify a 401(k) allocation amount will invest significantly more money than those who choose the default amount.

Hypothesis 1b: Individuals who select a 401(k) allocation amount different from the default amount will be more likely to select 401(k) investment options that differ from the default investment option.

Identification of individual differences is essential to understanding why employees make specific retirement planning choices. Personal characteristics (including attitudinal and demographic characteristics) are important variables to investigate because they may influence preferences that lead to one’s pension choices.⁴⁰

Importance of Psychological Variables in Financial Decision Making

Self-Efficacy

Scholars have frequently employed self-efficacy, or self-perceived ability to perform a specific task, as a successful predictor of choice and risk-taking behavior in decision-making studies.^41,42 Past research has shown self-efficacy positively predicted risk behavior in employee’s choice of asset allocation options⁴³ and also predicted employee selection of pay plans.⁴⁴

Self-efficacy in decision making is formed from an individual’s evaluation of three elements: task attributes, the decision context and perceived personal ability.⁴⁵ The result of this evaluation is the individual’s own perception and, therefore, may differ from the decision maker’s objective ability as measured by skill tests. For example, individuals with the same level of objective skill may report different levels of self-efficacy because they have different types of environmental or personal supports.

Self-efficacy mediates the relationship between individual analysis of the task situation and outcomes such as performance and decision choice.⁴⁶ Analysis of the task encompasses perceptions of the task difficulty, personal abilities and situational support. Therefore, past experiences in similar tasks, perceptions of capabilities in the task and perceptions of resources at hand all affect the formation of self-efficacy.

Self-efficacy is an important predictor of outcomes in financial and complex decision-making studies.^47-56 In sum, self-efficacy influences choice in financial decision-making studies and acts as a mediator between the impact of individual difference and environmental variables on outcomes such as performance and choice. It was expected that self-efficacy would positively influence 401(k) investment choices.

Ambiguity Tolerance

People differ in how well they tolerate ambiguous scenarios.⁵⁷ Individuals with high ambiguity tolerance may prefer uncertain tasks and persist despite initial failures. People who have a high tolerance for ambiguity cope more effectively with major changes and anxiety.^58,59 In contrast, people with low ambiguity tolerance perceive tasks with a high degree of uncertainty as being threatening and negative, and they give up more easily when they fail in these tasks.⁶⁰

Individual resilience for ambiguous situations would also be expected to affect financial choices.^61,62 Ghosh and Ray⁶³ found that ambiguity tolerance positively affected decision-maker confidence. Markman and colleagues⁶⁴ showed that entrepreneurs may have significantly higher self-efficacy and perceived control over adversity. Furthermore, Teoh and Foo⁶⁵ discovered that ambiguity tolerance was related to entrepreneur’s self-perceived performance. In summary, it was expected that choice in 401(K) options relates to an individual’s tolerance for ambiguity.

Locus of Control

Locus of control is defined as an individual’s generalized expectancy that his/her personal outcomes are either dictated by his/her own behavior (internal locus) or by outside forces beyond his/her control (external locus). Research suggests that individuals with an external locus of control depend more on external sources of information and assistance when making decisions.^66-68 Conversely, individuals with an internal locus of control rely to a greater extent on their own knowledge and abilities and refrain from outside assistance in decision making.

Dulebohn⁶⁹ proposed that decision makers with an internal locus of control would make riskier investment decisions in their 401(k) participation compared to those with an external locus of control based on research supporting an association between successful risk-taking behavior and attribution to one’s own abilities.⁷⁰ It also would be expected that because individuals with an external locus of control are more attuned to external frameworks of information, they would be more likely to select the organization’s 401(k) default contribution level than individuals with an internal locus of control.

Perceived Financial Knowledge

Higher perceived financial knowledge is an important predictor of higher self-efficacy in business decision making,⁷¹ and of sound financial savings behavior.^72,73 Individual perceptions are important because self-efficacy in financial decision making is based on perceived ability combined with perceptions of success in past experiences.

Actual Financial Knowledge

Actual financial knowledge is an important measure because individuals may misperceive facts in pension investment situations.⁷⁴ People may incorrectly believe that they have factual knowledge and act on these perceptions when investing in company pension plans.⁷⁵ Measuring one’s actual knowledge and comparing it to perceived knowledge is an important check on this potential self-bias. Actual experiences, such as taking a financial course, also serve to increase perceived knowledge and therefore self-efficacy. It was expected that individuals who have higher levels of perceived and actual financial knowledge would be more prone to affirmatively select 401(k) contribution levels and asset allocations rather than submit to the organization’s 401(k) default options.

Based on the available literature regarding psychological influences on retirement saving and investment, we expected individual psychological variables to affect subjects’ 401(k) allocations and the choice of investment options among which they could dedicate their 401(k) funds.

Hypothesis 2a: Individuals who select a 401(k) allocation amount that differs from the default amount will have higher levels of self-efficacy, ambiguity tolerance, internal locus of control and perceived and actual levels of financial knowledge versus individuals who choose the default amount.

Hypothesis 2b: Considering only individuals who choose a 401(k) allocation amount that differs from the default amount: Dollar amounts invested will be positively associated with self-efficacy, ambiguity tolerance, internal locus of control and perceived and actual levels of financial knowledge.

Hypothesis 2c: Individuals who select 401(k) investment options that differ from the default investment option will have higher levels of self-efficacy, ambiguity tolerance, internal locus of control and perceived and actual levels of financial knowledge, relative to individuals who choose the default investment options.

Hypothesis 2d: Selection of higher-risk investments will be associated with higher levels of self-efficacy, ambiguity tolerance, internal locus of control and perceived and actual levels of financial knowledge.

Importance of Demographic Variables in Financial Decision Making

Demographic factors can be used to indicate employee needs and attitudes that may affect their 401(k) investment strategies. Considerable literature on retirement savings centers on the impact of demographic factors including gender,^76-78 age,^79,80 and work experience.^81,82 Each of these demographic factors has been related to retirement saving practices.

Age

Age is a frequent variable examined in the financial savings literature.^83,84 Higher age was found to lead to more goal clarity, which led to enhanced planning behavior and more retirement savings contributions.⁸⁵ Higher employee age was also positively associated with the likelihood of choosing a defined benefit versus a defined contribution plan.⁸⁶ This literature supports an age-difference effect occurring for 401(k) options.

Work Experience

Organization-specific experience may affect retirement investment decisions.^87,88 Workplace tenure was found to be positively associated with defined contribution participation and contribution rates.⁸⁹

Role Model Influence

The financial planning literature points to the positive effect of exposure to role modeling in enhancing investment goal setting, cash flow management and securities management.^90,91 Financial peer mentoring has been associated with enhanced financial confidence and a greater sense of investment competence.^92,93 These findings support the influence of the presence of role models on 401(k) practices.

Hypothesis 3a: Individuals who select a 401(k) allocation amount different from the default amount will have higher age, work experience and role model influence versus individuals who choose the default amount.

Hypothesis 3b: When considering only individuals who choose a 401(k) allocation amount different from the default amount, the dollar amounts invested will be positively associated with age, work experience and role model influence.

Hypothesis 3c: Individuals who select 401(k) investment options that differ from the default investment option will have higher age, work experience and role model influence, versus individuals who choose the default investment options.

Hypothesis 3d: Selection of higher risk investments will be associated with higher age, work experience and role model influence.

Gender

Gender is often studied in financial decision-making research due to the common finding that women report lower self-perceptions than men in these tasks.^94-97 Gender is also a common variable used to predict investment choice in employer-sponsored plans.^98,99 Prior research has shown that men have higher confidence than women do in trading stocks.¹⁰⁰ Male entrepreneurs were also found to report higher confidence than female entrepreneurs in their ability to make financial decisions.^101,102 These findings indicated that women were less likely to invest in their retirement than men. These findings suggested a gender effect exists for 401(k) elections.

Hypothesis 4a: Men will be significantly more likely than women to choose a 401(k) allocation amount that differs from the default amount.

Hypothesis 4b: Considering only individuals who choose a 401(k) allocation amount that differs from the default amount, men will allocate significantly more money than women.

Hypothesis 4c: Men will be significantly more likely than women to select 401(k) investment options that differ from the default investment option.

Hypothesis 4d: Considering only individuals who select 401(k) investment options that differ from the default investment option, men will choose higher risk investments than women.

Complex relationships between psychological and demographic variables have rarely been addressed in research of 401(k) investment behavior. Gender issues are particularly salient because women take lower levels of financial risk than men, but live longer and are paid less.¹⁰³ However, a study of financial decision making about stock investments found the influence of gender on financial decision making was moderated by ambiguity tolerance.¹⁰⁴ Although men have been found to be typically more confident than women in their ability to make complex financial decisions, the gender gap lessened as women’s ambiguity tolerance increased. Based on this related research, it was expected that the gender gap in the level of risk taken in 401(k) investments would be lessened when ambiguity tolerance increases.

Hypothesis 5: Considering only individuals who select 401(k) investment options that differ from the default investment option, gender will interact with ambiguity tolerance to predict percentage of total allocation to high-risk investments.

Methodology

Sample

Undergraduate and graduate students participated in the study for extra credit points during class time in six sections of a required management class (n = 217, 50.20% men). The average class size was 39 students. The average age was 23.44 years (SD = 3.63), with a maximum age of 35 years. The average work experience was 7.44 years (SD = 4.66), and average managerial experience was 1.48 years (SD = 2.05). Students had worked for an average of 4.78 employers (SD = 2.98) over their work career. A total of 64 participants (29.5%) reported that they had participated in a retirement plan at least once.

Participants were informed that they would participate in a class activity for in-class credit. The amount of class credit was consistent across each class in terms of the percentage of each student’s grade. The experimenters passed out the first sheet of the experiment (a single piece of paper that informed them of the task scenario), which the experimenter read to them while they followed along. Participants were instructed that they should assume that they were just hired at a new company with a salary of $50,000 per year. Each participant was then required to withhold some portion of this salary toward the company’s 401(k) retirement/pension plan. In the first questionnaire, subjects answered questions for the student independent variable measures and provided the demographic information detailed in the next section.

In Step 2, subjects were given a choice of 11 investment options, with labels such as “Growth Option,” “Balanced Option,” “Bond Option” and so on. A separate sheet was handed out to participants that provided details for each of the 11 options in table format with information provided for the following: “Investment Objective,” “Invests Primarily In,” “Primary Source of Return,” “Level of Risk,” “Investment Time Horizon” and “Amount of Risk Spread Over a 10-Year Period.” To gather realistic options for potential retirement allotment choices, we drew from the most common 401(k) investment options offered by the three largest third party retirement service administrators in the United States. From this, we selected 11 popular asset allocation options earmarked specifically for retirement investments differentiated by level of risk.

Following a review of the 11 investment options, we reminded participants of the scenario and instructed them to choose one of the two options in a two-step process: (1) a dollar amount of annual income to withhold or (2) the company default withholding level of 3% of annual income. After all participants responded to the first step, a second page explaining the second step was passed out. This second handout asked subjects now either to allocate their withholding amount to their own choice of the previously reviewed investment options or to allow the company to allocate the withholding amount in default investment options, which were referred to as “different investment options.” All participants were required to allocate their retirement funds, whether they chose their own allocation amount in Step 1 or whether they chose the company default allocation amount.

Measures

Self-Efficacy

We assessed self-efficacy by asking participants to judge their capability in retirement planning (α = .86).¹⁰⁵ Participants responded to four questions with a percentage indicating how capable they perceived themselves (0% to 100%), and we averaged the four responses for an overall score. The four questions were the following: (1) “I can make financial decisions about my 401(k) without assistance,” (2) “I can determine the best allocation of 401(k) funds for my stage of life and family situation,” (3) “I can determine what amount of money I will need in retirement,” and (4) “I can control my current spending in order to save for retirement.” The questions used were modeled after other research that investigated self-efficacy’s role in making financial decisions.¹⁰⁶

Tolerance for Ambiguity

We used five questions to assess tolerance for ambiguity, using a scale of 1 (Strongly agree) to 5 (Strongly disagree).¹⁰⁷ These questions displayed acceptable reliability (.82). Examples of questions included (1) “I find it difficult to respond when faced with an unexpected event” and (2) “Problems that cannot be considered from just one point of view are a little threatening.”

Locus of Control

We measured locus of control with five questions on a scale of 1 to 5 (Almost never to Almost always), which we averaged for the overall score (α = .78). A higher average indicated an external locus of control and a lower average indicated an internal locus of control. The scale was taken from Levinson¹⁰⁸ and included such questions as (1) “There is really no way I can solve some of my problems” and (2) “I can do anything I set my mind to.”

Perceived Financial Knowledge

We used five questions to assess perceived financial knowledge on a scale of 1 to 5 (Poor to Excellent), and we averaged the question responses for an overall score. This scale was used in related past research with adequate reliability,¹⁰⁹ and in this study (α = .83). Participants were asked to rate their knowledge on the topics of (1) interest rates, finance charges, and credit terms; (2) credit ratings and credit files; (3) managing finances; (4) investing money and (5) credit report details.

Actual Financial Knowledge

Following Martocchio and Dulebohn¹¹⁰ and Allgood and Walstad,¹¹¹ we assessed actual financial knowledge using a four-question multiple-choice test. Subjects’ responses were recorded with an accuracy score equaling the number of correct responses. The four question topics were the following: (1) real return, (2) average annual rate of return, (3) quality of bonds based on annual interest and (4) broad diversification in mutual funds and related risk.

Demographics

For age, managerial experience and work experience, subjects indicated years by filling in a blank for each response line. Subjects indicated their gender by checking a box indicting “male” or “female.” Data were coded 1 for “male” and 2 for “female”.

Graduate or Undergraduate

Subjects indicated their university status, coded as 1 for undergraduate (n = 199, 91.70%) or 2 for graduate (n = 18, 8.30%).

Role Model Influence

We used two questions to assess the influence of role models in saving for retirement and analyzed these separately. Using a scale of 1 (Strongly disagree) to 5 (Strongly agree), Role Model Question 1 asked participants to rate agreement with the statement, “My parent(s)/guardian(s) taught me about saving for retirement.” Role Model Question 2 asked for rating in agreement with the statement, “Another role model (friend, family member, teacher) taught me about saving for retirement.”

Past 401(k) Training and Participation

Past enrollment in a 401(k) plan can affect current enrollment choices in a similar plan.¹¹² Subjects responded to the question, “Have you ever attended a training program or received instruction on 401(k) programs and retirement savings (inside or outside of work)?” We provided four options as answers: “Yes, once” (coded as 1), “Yes, more than once” (2), “No, never” (0) and “Not sure” (not coded or used in analyses). Subjects were also asked, “Have you ever participated in a retirement savings plan or 401(k) plan?” Response choices were “yes,” “no” or “not sure.” For the purpose of analysis, the items were coded 1 for “yes” and 2 for “no.”

Education

Education represents a common demographic variable explored in retirement savings studies.^113,114 We included educational level as a control because of its frequent inclusion in risk behavior research.

Results

Table 1 presents correlations for the study variables. Table 2 provides descriptives. Variable 1, “Specify Default Amount,” was coded as 1 = participant-specified allocation amount and 2 = default allocation level. Variable 2, “Amount for Non-Default Choice,” was equal to the dollar amount that was specified by those participants who chose an investment amount other than “default.” Variable 3, “Default Allocation or Choose Risk Levels,” was coded as 1 = allocation to participant’s own choice of investments or 2 = allocation of specified amount to default investment choices. “Level of Financial Risk” was coded as 1 = low risk investment, 2 = moderate risk investment and 3 = high risk investment. These codes are represented in the variable, “Risk Level.” Past 401(k) training and past 401(k) participation did not appear to influence study-dependent variables, nor did whether the student was a graduate or undergraduate.

Table 1.

Correlations of Study Variables.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
1. Choice 1	1.00
2. Dollar Amount	^a	1.00
3. Choice 2	.36**	.03	1.00
4. Risk Code	.11	.11	^a	1.00
5. Self-efficacy	−.20**	.28**	−.20**	.15	1.00
6. Ambiguity Tolerance	−.11	−.01	−.13	.26**	.18**	1.00
7. Locus of Control	.02	−.05	.10	−.04	−.10	−.35**	1.00
8. Perceived Fin Knowledge	−.11	.25**	−.16*	.06	.50**	.19**	−.15*	1.00
9. Actual Fin Knowledge	−.13	.17*	−.25**	.13	.29**	.18**	.00	.18*	1.00
10. Managerial Exp	−.01	.04	−.02	.09	.03	.11	−.05	.13	.11	1.00
11. Work Exp	.07	.13	.05	.01	.03	.03	−.11	.05	−05	.41**	1.00
12. Gender	.11	−.05	.06	−.17*	−.26**	−.16*	−.05	−.11	−.33**	−.06	.148*	1.00
13. Age	−.08	.24**	.01	.01	.19**	−.01	−.07	.13	.07	.36**	.43**	.03	1.00
14. 401K Training	.12	−.08	.09	−.04	−.06	−.14*	.08	−.21**	−.09	−.10	−.19**	−.07	−.157*	1.00
15. 401K Part	.06	−.01	.02	−.09	−.17*	−.01	.05	−.14*	−.11	−.16*	−.18**	.08	−38**	.45**	1.00
16. Grad/UG	−.04	.03	−.10	−.03	.04	.04	−.09	−.01	−.05	.08	.20**	.20**	.38**	−.10	−.24**	1.00
17. Role Model 1	−.11	.18*	−.07	.02	.30**	.07	−.08	.32**	−.02	−.13	−.21**	−.18**	−.17*	.03	.01	−.16*	1.00
18. Role Model 2	−.20**	.15	−.21**	−.04	.25**	.03	.00	.35**	−.01	−.06	−.04	.04	.03	−.22**	.00	−.07	.33**

Cannot be computed because at least one of the variables is constant.

Correlation is significant at the .01 level (2-tailed). *Correlation is significant at the .05 level (2-tailed).

Table 2.

Study Variable Descriptives.

	M	SD	n
Choice 1: Default or participant-specified allocation amount	1.22	0.42	217
Dollar amount for nondefault choice	4493.99	2296.58	169
Choice 2: Default or participant-specified allocation to investment risk levels	1.26	0.44	217
Risk code (L, M, H)	3.25	1.60	155
Self-efficacy	50.09	24.49	216
Risk aversion	2.96	0.78	136
Ambiguity tolerance	3.47	0.61	217
Locus of control	1.97	0.65	217
Perceived financial knowledge	3.13	0.94	216
Actual financial knowledge	0.88	0.97	210
Years of managerial experience	1.48	2.05	211
Years of work experience	7.44	4.66	215
Gender	1.50	0.50	217
Age	23.44	3.63	215
Past 401(k) training	1.65	0.48	210
Past 401(k) participation	1.74	0.51	217
Grad or undergrad	1.08	0.28	217
Role Model Question 1	2.80	1.32	217
Role Model Question 2	2.50	1.35	217
Mood	4.44	1.38	196

Hypothesis 1a

Hypothesis 1 stated that individuals who specify their own 401(k) allocation amount will invest significantly more money than those who choose the default allocation amount. A t test confirmed the hypothesis (t = 9.37, p < .001). Individuals who chose their own allocation amount specified significantly more ($4493.99, n = 169) than those who accepted the default allocation of $1,500 (n = 48).

Hypothesis 1b

Hypothesis 1b stated that individuals who select their own dollar amount when contributing to a 401(k) would be significantly more likely to self-select specific investment options. The hypothesis was verified with the nonparametric test Cramér’s V (V = 0.363, p < .001). A total of 139 of the 169 participants (82.25%) who specified their own 401(k) allocation amount also self-selected specific investment options.

Hypotheses 2a and 3a

We expected individuals who specified their own 401(k) allocation amount to differ from those who chose the default allocation amount. Results confirm that those who chose their own allocation amount displayed higher levels of self-efficacy (t = 2.99, p < .003), role model influence (nonparent role model influence; t = 2.95, p < .004), perceived financial knowledge (t = 2.44, p < .02), and actual financial knowledge (t = 3.73, p < .001). Participants who chose their own amount reported marginally higher nonparent role model influence (t = 1.68, p < .09), but did not report higher ambiguity tolerance, internal locus of control, role model influence (parent influence), age, work experience or managerial experience.

Hypotheses 2b and 3b

Correlation results show that those who specified their own allocation amount invested more dollars when reporting higher levels of parent role model influence (r = .19, p < .005), nonparent role model influence (r = .21, p < .002), age (r = .24, p < .001), self-efficacy (r = .28, p < .001), perceived financial knowledge (r = .22, p < .001), and actual financial knowledge (r = .20, p < .003). We found no relationship between dollars invested and work experience, managerial experience, internal locus of control or ambiguity tolerance.

Hypotheses 2c and 3c

When comparing individuals who chose their own risk investment levels versus those who accepted default investment levels, results confirm that those who chose their own risk investment levels displayed higher self-efficacy (t = 2.97, p < .01), perceived financial knowledge (t = 2.44, p < .02), actual financial knowledge (t = 3.73, p < .001), and role model influence (nonparent; t = 3.08, p < .001). A marginal result showed those who chose their own risk investment levels had higher ambiguity tolerance (t = 1.88, p < .06). No significant difference existed for internal locus of control, age, work experience, managerial experience or role model influence (parent).

Hypotheses 2d and 3d

Hypotheses 2d and 3d proposed that investments in higher risk asset allocations will be associated with higher levels of self-efficacy, tolerance for ambiguity, internal locus of control, age, work experience, managerial experience, role model influence and perceived and actual levels of financial knowledge. For these hypotheses, participants were first classified as low-, moderate- or high-risk investors based on the percentage of their allocations among the funds offered as options in the study. Two independent evaluators determined the (1) classification of the investment vehicles as low, moderate or high risk and (2) subject classification based on the clustering of participant allocations. Results included classification of low (n = 41, 26.45%), moderate (n = 54, 34.84%) and high (n = 60, 38.71%) risk investors. ANOVA results indicate that the three investing groups differed significantly in their ambiguity tolerance (F = 5.60, p < .004). As expected, moderate- and high-risk investors reported significantly higher ambiguity tolerance than low-risk investors. The investor risk classifications did not differ significantly according to self-efficacy, locus of control, age, work experience, managerial experience, role model influence or perceived and actual levels of financial knowledge.

We also tested Hypotheses 2d and 3d with dollar levels and percentages of dollars that a subject placed in low-, medium- and high-risk investments. Because classification into distinct categories can be prone to human and statistical smoothing errors, we used this second operationalization for risk behavior. Table 3 shows these alternative risk categories’ descriptive statistics.

Table 3.

Alternative Risk Level Descriptives.

	M	SD	n
Low-risk investment ($)	1214.34	1952.80	160
Moderate-risk investment ($)	945.94	1946.91	160
High-risk investment ($)	1804.22	2447.45	160
Low-risk investment (%)	0.31	0.32	156
Moderate-risk investment (%)	0.25	0.27	156
High-risk investment (%)	0.44	0.35	156

Correlation results show that the amount invested in low-risk investments was significantly associated with lower ambiguity tolerance (r = −.18, p < .03). The amount invested in high-risk investments was significantly associated with higher role model influence (nonparent; r = .16, p < .04) and higher self-efficacy (r = .18, p < .03). The amount invested in moderate-risk investments was significantly, but marginally, associated with a higher internal locus of control (r = .04, p < .08) and less work experience (r = −.13, p < .09). The percentage of a person’s total allocation to low-risk investments was significantly associated with lower ambiguity tolerance (r = −.27, p < .001). The percentage of a person’s total allocation to high-risk investments was significantly associated with higher self-efficacy (r = .16, p < .05) and higher ambiguity tolerance (r = .26, p < .001).

Hypothesis 4a to 4d

Men were expected to be more likely to specify their own 401(k) contribution amounts versus women. A total of 90 of the 109 men (82.57%) specified their own 401(k) contribution amount, while a total of 79 of 108 women (73.15%) specified their own amount. Hypothesis 4a was supported at a marginal level of significance (Cramér’s V = 0.11, p < .10).

Hypothesis 4b stated that, among individuals who specify their own 401(k) contribution levels, men would commit significantly more money than women. Results show that the amount specified by men who chose their own 401(k) amount was not significantly different from the amount specified by women (t = 0.70, p < .48). Men invested an average of $4250.60, and women invested an average of $3989.47.

We expected men would be significantly more likely than women to self-select specific investment options, versus accepting default allocations. Hypothesis 4c was not confirmed (Cramér’s V = 0.06, p < .42). A total of 83 of the 109 men (76.15%) chose to self-select specific investment options, while a total of 77 of 108 women (71.30%) chose to self-select specific investment options.

For Hypothesis 4d, we expected men to select more risky investment allocation options than women. We first classified participants as low-, moderate- or high-risk investors based on the percentage of their allocations among the funds offered as options in the study. Using these three risk types, Hypothesis 4d was confirmed (Cramér’s V = 0.20, p < .04). We found men were more likely than women to be classified as high-risk investors, with 43.03% (n = 34) of men in this classification and 34.21% (n = 26) of women in this classification. We also tested Hypothesis 4d with dollar levels and percentages of dollars in high-risk investments. Results showed that men did not allocate a significantly higher percentage of their 401(k) money to high-risk investments (t = 0.19, p < 0.85) nor did men allocate a significantly higher dollar amount to high-risk investments (t = 1.33, p < .19).

Hypothesis 5

Gender was expected to interact with tolerance for ambiguity to predict risk level of an individual’s self-specified investment (low, moderate, high risk). We tested Hypothesis 5 using percentage of total allocations in low-, moderate- and high-risk investments.

Gender interacted with ambiguity tolerance to predict percentage of 401(k) allocation participants placed in high-risk investments. Table 4 shows that the interaction term is marginally significant (t = −1.92, p < .06) and the overall model is significant (F = 5.09, p < .002). Figure 1 illustrates the interaction, showing that men’s percentage of 401(k) in high-risk investments increased significantly as ambiguity tolerance increased. Women’s percentage of 401(k) in high-risk investments did not change significantly based on ambiguity tolerance.

Table 4.

Hierarchical Regression Analysis.

DV: Percentage in high-risk investments
Predictor variables	Model 1	Model 2
Gender	−0.047	0.561^†
Ambiguity tolerance	0.141**	0.399**
Gender × Ambiguity tolerance		−0.173^†
R ²	0.069**	0.091**
ΔR²	0.069**	0.022^†
F	5.694	9.375
ΔF	5.694	3.681

Note. N = 198.

†

p < .10. *p < .05. **p < .01. ***p < .001 (all 2-tailed).

Figure 1.

Interactive effects of gender and ambiguity tolerance on percent of 401(k) allocation in high-risk investments.

Conclusions

This study examined the relationships between demographic and psychological characteristics of individuals and the election of 401(k) contribution rates and asset allocation vehicles through automatic enrollment default procedures. While past research has concentrated on the association between individual enumerations and attitudinal variables with retirement investment decisions, no previous study has specifically examined their relationship with enrollment in 401(k) default options.

Our findings in this study indicate that individuals who specify their own 401(k) allocation amounts invest significantly more money to this account compared to those who defer to a 401(k) default allocation rate. This finding is consistent with previous research, which has found that retirement savings of those selecting their own 401(k) contribution rates exceed the savings for those individuals whose 401(k) contributions are based on an organization’s default rates. In addition to contributing a higher percentage of their income toward their 401(k) amount, individuals electing to allocate their own dollar amount rather than choosing the organization’s salary default rate are significantly more likely to also self-select specific investment allocation options.

Our findings further suggest that some psychological and demographic factors are systematically related to retirement saving decisions. Individuals in our study specifying their own 401(k) allocation amount differ from those selecting the default allocation amount in that they have higher levels of self-efficacy, exposure to role models and financial knowledge. In addition, the findings suggest that, among those electing 401(k) account allocation amounts apart from the default amount, individuals with higher self-efficacy, tolerance for ambiguity, financial knowledge, role model influence and greater past employment history voluntarily invest more income into their account relative to others not having these characteristics.

With respect to selection of asset allocation vehicles, the findings suggest that those with higher tolerance for ambiguous situations select allocation options of higher risk. Furthermore, our findings suggest that relative to women, men’s percentage of 401(k) allocations in high-risk investments increased significantly as ambiguity tolerance increased.

Our findings shed light on the characteristics of those most likely to accept the automatic enrollment default options of 401(k) options. Identifying how much income to contribute and which investment options to choose are key decisions individuals encounter when enrolling in their 401(k) plan. Accentuating the importance of automatic enrollment default options are findings that 80% of automatically enrolled individuals accept both the default contribution level and default allocation funds.¹¹⁵ Plan participants customarily set automatic enrollment 401(k) default options at very conservative contribution levels. Maintaining this default option for a prolonged time would impair an individual’s long-term retirement savings.

Companies practicing automatic enrollment have witnessed dramatic increases in employee participation in 401(k) plans. Vanguard, a leading investment management company, reported that among its clients 60% of newly hired employees were automatically enrolled in 401(k) plans and that participation in these plans had risen by two thirds in the past decade.¹¹⁶ Yet approximately one half of firms offering 401(k) retirement plans to their employees have not adopted features that automatically enroll employees in the plan unless the employees elect to do otherwise. Firms that have yet to install automatic enrollment features may resist doing so because of fears of appearing paternalistic as well as concerns over incurring increased costs associated with employer matching and ongoing administrative costs.

The findings of the present study appear to have most resonance for these firms that have yet to implement autopilot designs. Our research findings clearly suggest that demographic characteristics and psychological attributes of individuals influence retirement investment decisions. From a managerial vantage, these results imply that organizations with heavy concentrations of individuals lacking in both financial self-efficacy and knowledge as well as possessing low tolerance for ambiguity would most benefit from employing automatic enrollment features within their organization. Organizations considering starting 401(k) automatic enrollment procedures would be wise to determine whether a substantial segment of their workforce possessed characteristics most apt to accept default contribution rates and asset allocation options set by the organization.

Our study indicates that employees sharing specific individual characteristics would most likely take advantage of a firm’s pension plan only through opt-out procedures instituted by the employer. Identification of these individuals by the employer through assessment and the accompanying inauguration of automatic enrollment features to encourage their 401(k) plan participation would lead to enhanced retirement security for these employees. This action would also serve the purpose of promoting the value of retirement plans as a key component of a total rewards package among employees who are not likely to take advantage of this benefit. Recognition among members of the workforce of the centrality of retirement plans as an employee benefit would enhance organizational morale and improve an employer’s ability to attract and retain employee talent.

As is the case with many other studies, the limitations of findings and implications of these results must be considered. One limitation of this study is the use of students as a source of data, which limits its generalizability to employment settings. However, 401(k) automatic enrollment practices are particularly targeted to encourage pension participation among newer entrants to the workforce who are less experienced with retirement savings behavior than more mature workers. From this perspective, a student population may reasonably reflect characteristics of those employees who would especially benefit from automatic enrollment protocols. In addition, researchers have found that the use of students is appropriate when studying behavioral concepts¹¹⁷ because students often exhibit diverse attitudes found in the general society.¹¹⁸

Another limitation of the study is that the procedure for selecting the default option did not follow the typical 401(k) automatic enrollment situation. In most 401(k) plan automatic enrollments, eligible employees do not attend in-person enrollment meetings where they choose whether to select the employer’s default options. Individuals are generally allotted to these default options when they fail to choose other retirement income level and asset allocation choices. The consequence of the experimental procedure requiring study participants to make actual 401(k) elections likely resulted in fewer individuals choosing pension default options than would be the case in real 401(k) plan automatic enrollments. This is because factors such as procrastination and inattentiveness to complicated hard copy and/or emailed pension communications would lead many employees to drift into the pension default choices absent an event forcing an affirmative election. The nature of the task requiring individuals to elect 401(k) contribution rates and select asset allocation choices appeared to diminish the percentage of individuals typically participating in pension plans only through automatic enrollment features.¹¹⁹ The relatively smaller percentage of individuals electing pension default options in this study than are generally found in actual practice suggests the heightened importance of individual characteristics for those choosing automatic enrollment.

In addition, although we measured many different demographic and psychological attributes, which could affect retirement savings behavior, additional psychological constructs such as retirement goal clarity, attitudes toward retirement and income diversity could certainly influence 401(k) investment activities.

A final limitation of this study is the use of a cross-sectional design. It is not possible from these findings to infer that individual selection of 401(k) default options is based on any specific demographic characteristic or psychological attribute. One can only infer significant association between these variables. With longitudinal studies, future research would help our understanding of any causal relationships.

In conclusion, these findings suggest that individual differences influence selection of 401(k) default contribution rates and asset allocation options. Implications of this investigation should be further explored to generate additional organizational prescriptions. Future research would benefit from assessment of pension choices made by actual employees of organizations to determine whether this effect manifests itself in employment settings. We hope that future investigators with access to employee populations will add to this work into the employment domain to further clarify determinants of individual differences and employer-sponsored pension plans.

Footnotes

Acknowledgements

The authors wish to acknowledge Michael Weddell, Senior Consultant, Benefits Advisory and Compliance of Willis Towers Watson for his assistance with this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Eric R. Schulz, PhD, is a professor of Management at Eastern Michigan University. His research focuses on rewards management. His publications have appeared in such journals as Human Resource Management, Journal of Business Research and WorldatWork Journal.

Megan Endres, PhD, is a professor of Management at Eastern Michigan University. Her research focuses on knowledge sharing, self-efficacy, and decision making. Her publications have appeared in journals such as Academy of Management Journal, Journal of Managerial Issues, and Theory and Decision.

Sergei Kolomeitsev has experience working with clients such as Ford, Yahoo, ExxonMobil, Cisco, Medtronic, and the United Nations. He completed his undergraduate studies at the University of Michigan and graduated with high distinction while double majoring in Economics and Russian and East European Studies.

Rebecca Clark de Castillo is a process coordinator at Varroc Lighting Systems focused on training needs and business processes for the Customer Business Unit area. She completed her undergraduate degree at the University of Notre Dame and Master of Science in Human Resources and Organizational Development at Eastern Michigan University.