J Labor Res https://doi.org/10.1007/s12122-018-9270-2

Supplemental Retirement Savings Plans in the Public Sector: Participation and Contribution Decisions by School Personnel Robert L. Clark 1 & Aditi Pathak 2 & Denis Pelletier 2

# Springer Science+Business Media, LLC, part of Springer Nature 2018

Abstract Virtually all full-time state and local government employees are covered by a retirement plan, typically a defined benefit plan, in which they are required to participate. In addition, most school employees have the option of choosing to contribute to a voluntary retirement savings plan offered by their school district. Relative to private sector workers, public employees face an expanded choice of retirement savings plans. Federal tax policies allow state and local governments the opportunity to offer both 401(k) plans and 457 plans to their employees. In addition to these plans, public schools and certain other organizations can offer 403(b) plans to their employees. This paper examines the decision to participate in a voluntary savings plan and the level of contributions for those that enroll in at least one of the plans. The analysis begins by describing the savings options available to public school employees and how these plans differ. The findings indicate that the same economic and demographic factors that influence saving decisions by private workers also drive the decisions of school employees. The three savings plans offered to public employees have many similar characteristics; however, several differences in the plans imply that certain workers may prefer one plan type over the others. Probit and Tobit models of participation in any plan and total annual contributions are estimated. Finally, we estimate the determinants of the decision to choose any one or a combination of savings plans. Keywords Retirement saving plans . Saving decisions . Federal regulations Many Americans worry about having sufficient resources in order to achieve a desired level of well-being in retirement. One of the most efficient methods of saving for * Robert L. Clark [email protected]

1

Poole College of Management, North Carolina State University, Raleigh, NC, USA

2

Department of Economics, North Carolina State University, Raleigh, NC, USA

J Labor Res

retirement is through tax-qualified, employer-provided retirement savings plans. In the private sector, employers seeking to provide qualified retirement savings plans for their employees offer 401(k) plans. Over the past four decades, these plans have become the primary pension plan for millions of American workers as firms have frozen or terminated traditional defined benefit plans.1 Private sector workers are covered by Social Security which provides a base retirement benefit; however, most workers seeking to save more for retirement must participate in a 401(k) plan offered by their employer. In most cases, a firm selects a single vendor for its 401(k) plan and then negotiates with the chosen vendor the number and type of investment options to be included in the plan, fees for investment products, the communication and advice provided by the vendors to employees, and how the plan will be marketed to workers. In summary, the usual case in private firms is to have one retirement savings plan offered by a single vendor with plan characteristics and investment options agreed to by the employer and the vendor. Thus, the saving problem facing employees is to decide what percentage of salary they want withheld and contributed to the savings plan and then to manage their account by choosing from the investment options in the plan approved by the employer. The participation and investment decisions in 401(k) has been much studied. For example, see Beshears et al. (2011), Choi et al. (2009), Choi et al. (2002), Huberman et al. (2007), and Madrian and Shea (2001). These studies find significant differences in retirement saving patterns by income, age, tenure, and certain other economic and demographic characteristics. They also find that plan characteristics such as automatic enrollment and employer matches influence the probability of participation in a plan and the contribution rates (Papke 1995; Madrian and Shea 2001; Choi et al. 2004). Important defaults that are increasingly used by private firms include automatic enrollment, automatic escalation, and the use of target date funds as the default investment. Public employers are much less likely to have adopted such policies. Other studies have shown the need for financial literacy to understand the parameters of key decisions (Lusardi and Mitchell 2009, 2014 and Clark et al. 2014). The retirement saving landscape is much different and more complex in the public sector. Virtually all state and local employees are covered by a primary pension plan, typically a defined benefit plan, in which they are required to participate. In addition, most also have access to supplemental retirement savings plans. In comparison to the relatively large literature examining participation in 401(k) plans among private sector workers, relatively few studies have examined voluntary or supplemental retirement saving decisions of public employees. While there are several studies that focus on the choice between enrolling in a defined benefit plan, a hybrid plan, or a defined contribution plan as a primary pension option (Chingos and West 2015; Goldhaber and Grout 2013, 2016; Clark and Pitts 1999), we know of no research papers that have focused on the range of supplemental savings plan choices available to public employees and how public workers decide on which plan best meets their saving needs.2 1

The Bureau of Labor Statistics reports that information from the National Compensation Survey shows only 8% of establishments offered a defined benefit plan in March 2015 while 47% offered defined contribution plans. http://www.bls.gov/ncs/ebs/benefits/2015/ownership/private/table01a.pdf 2 Clark et al. (2016) provides considerable information about 403(b) plans and their use in the school districts in North Carolina. Clark and Hanson (2013) examined the use of 403(b) plans by school districts in all 50 states.

J Labor Res

Public employers have the option of offering alternative types of savings plans regulated under special tax codes. State governments can offer their employees the option of enrolling in either a 401(k) or a 457 plan. Local government employees may have the option of participating in a state-managed 457 plan, the state-managed 401(k) plan, and these agencies may also offer their own locally-managed 457 plan. In addition to these plans, school districts may offer employees locally-managed or in some cases state-managed 403(b) plans. Thus, the saving decision for many public employees consists of first deciding whether to contribute to a retirement savings plan offered by their employer. Having made this saving decision, employees then must select the most appropriate savings plan from the two, three, or even four plans that are offered. In some cases, especially in 403(b) plans, even after selecting the plan of choice, employees will then find themselves having to choose from a number of vendors in one or more of these plans.3 To examine retirement savings plans in the multi-plan public sector environment, this article reviews the plan offerings by school districts in North Carolina. All of 115 districts in the state enroll their employees in a defined benefit plan and offer their employees the option of enrolling in the state-managed 401(k) plan and a districtmanaged 403(b) plan. In addition, many of the districts also allow school personnel to contribute to the state-managed 457 plan and/or a locally-managed 457 plan. The analysis begins with a discussion of how these plans differ in terms of tax limits, withdrawals, and management by the district or state.4 Next, the current offerings of retirement savings plans are presented using data from a survey of school district administrators and payroll data provided by the districts. We then estimate the probability of participating in any savings plan and the level of annual contributions to all plans as a percent of salary. Following this analysis, we explore the choice of school personnel as they select among the various savings plans offered by their school district. Finally, we seek to determine what factors explain why individuals select one type of plan over another and why some individuals contribute to two or more plans at the same time.

Comparison of Retirement Savings Plans While many economists and policy analysts are familiar with 401(k) plans and their basic characteristics, far fewer are knowledgeable of the basic characteristics of 403(b) and 457 plans.5 In general, these plans are very similar; however, some important tax and regulatory differences exist. A first distinction is which government agencies and public employers have the option of offering each of these retirement savings plans. 3

A survey of school districts in North Carolina found that the number of vendors in district-managed 403(b) plans ranged from 1 to 13. 4 Another important difference in these plans is that 403(b) plans often allow multiple vendors and provide much less management oversight. The result often is that that these plans include vendors and products with higher fees (Clark and Richardson 2010). 5 IRS discussion of 403(b) plans can be found at: http://www.irs.gov/publications/p571/index.htmland at http://www.irs.gov/Retirement-Plans/Retirement-Plans-FAQs-regarding-403(b)-Tax-Sheltered-Annuity-Plans Readers may also find the Department of Labor’s discussion of 403(b) interesting: http://www.dol.gov/ebsa/regs/fab2010-1.html

J Labor Res

While private employers can only offer their employees a 401(k) plan, state and local government employers have the option of offering 457 plans to their employees as well as being able to offer 401(k). In 1986, federal legislation removed the option for public employers to establish new 401(k) plans; however, existing plans were allowed to continue to operate. Local governments covered by state pension plans can still opt to offer the legacy state 401(k) plan.6 Public schools, hospitals, and charitable organizations are also allowed to offer 403(b) plans to their employees. The North Carolina retirement system provides public employees a detailed comparison of the provisions of the 401(k), 403(b), and 457 plans offered by the state of North Carolina (https://www. nctreasurer.com/ret/401k%20Plan%20Features/PlanComparisonChart.pdf. Table 1 presents an overview of the primary characteristics of these three plans. All of these plans allow employees to make pre-tax contributions; however, government employers can also offer Roth options in the various plans that allow employees to contribute after-tax dollars to the plan. The annual dollar limit in 2013 (year of the payroll data) on employee contributions is the same in all of the plans, $17,500, and this maximum is increased periodically. All of the plans have the same age 50 and over catch-up provisions that allow older workers to contribute an additional $5500 per year. Contributions to 401(k) and 403(b) plans count toward the same maximum annual contribution limit; however, contributions to 457 plans are viewed independently from the other two plans. In sum, public employees had the option of contributing up to $17,500 to any of these plans with individuals over age 50 being able to contribute up to $23,000 annually. The separate cap on contributions associated with 457 plans means that individuals desiring to have higher contributions could do so by participating in both a 457 plan and either a 401(k) or 403(b) plan. This strategy would allow them to exceed the annual limitation on contributions to a single plan. Other differences in the tax code affecting retirement savings plans may influence worker choices for their retirement saving. Distributions without tax penalties are allowed at retirement after age 55 in both 401(k) and 403(b) plans while the 457 plans allow such distributions at termination of employment at any age. This may be an important distinction as many public employees retire from their career state or local jobs in their early 50s. In-service distributions are allowed prior to age 59 and a half in both the 401(k) and 403(b) plans with a 10% tax penalty. In-service distributions are not allowed in 457 plans. Thus, workers who wish to maintain the ability to access their fund prior to retirement may be more likely to contribute to 401(k) or 403(b) plans. All of the plans have a required minimum distribution at age 70 and a half. All distributions in each of the traditional or non-Roth plans are treated as ordinary income for federal income tax purposes. In comparison, distributions from Roth plans are not considered taxable income. All of the plans allow rollovers to other retirement savings plans or IRAs. These differences in distribution rules may also provide incentives for public employees to prefer one type of plan over the others and give managers a reason for providing employees the option of selecting from several retirement savings plans. Prior to the passage of the Economic Growth and Tax Relief Reconciliation Act in 2001, 401(k), 403(b) and 457 plans were subject to different regulations related to 6

See the Tax Reform Act of 1986, which prohibited governmental employers from establishing new 401(k) plans. Governmental employers who had established 401(k) plans prior to the new legislation were allowed to continue offering these plans.

J Labor Res Table 1 Comparison of 401(k), 403(b) and 457 plans Provisions

401(k)

403(b)

457

Who can offer

• Private employers • Nonprofit employers • State and local government (only existing plans)

• Nonprofit employers (hospitals and charitable organizations) • Public schools

• State and local government

Pre-tax contributions

Yes

Yes

Yes

limits in 2013 on annual employee contributions

Up to $17,500

Up to $17,500

Up to $17,500

Combined annual contribution limits with other plans

403(b)

401(k)

No

Age 50+ catch-up contributions

$5500

$5500

$5500

Minimum distribution age

70½

70½

70½

Distributions without tax penalties

At retirement after age 55

At retirement after age 55

Termination of employment at any age

In-service distributions

Age 59½ (10% tax penalty)

Age 59½ (10% tax penalty)

Not allowed

elective deferrals, employer contributions and other areas. After the passage of new legislation, 403(b) plans were subject to greater oversight. Now, despite similar treatment by the IRS, remaining differences in plan characteristics may appeal to public employers and employees with different saving objectives. This overview of retirement savings plans has shown that in contrast to the rather simple choice facing firms in the private sector, public employers, especially local governments and school districts face a range of retirement saving options. An important difference in the management of these plans is choice of vendors. In most cases, state-managed plans select a single vendor through some type of competitive bidding process. In contrast, locally-managed 403(b) plans tend to allow any vendor that is able to meet certain criteria to be part of the plan. The result is that participants in 403(b) plans have a considerable choice of vendors; however, this choice may come at the expense of having less oversight by the employer and higher fees (Clark and Richardson 2010). This discussion raises several behavioral questions concerning retirement savings plans in the public sector that are examined in this paper. First, given that they are covered by Social Security and a defined benefit pension plan, what factors influence school personnel in their decision to participate in supplemental retirement savings plan and what personal characteristics affect how much employees contribute to these plans. Second, conditional on enrolling in a plan which type of plan is selected? Third, do public employees choose to contribute to more than one plan or spread their contributions across several plans? Among those who contribute to a 457 plan and either the 401(k) or 403(b), is there any evidence in this population that highly compensated individuals contribute to multiple plans in order to exceed the annual contribution limit of a single plan? To answer these questions, we use March 2013 payroll data from 53

J Labor Res

school districts in North Carolina covering over 70,000 employees along with a survey of administrative leaders in these districts.

Savings Options Offered by North Carolina School Districts To determine the plan offerings by each of the school districts in North Carolina, a survey was sent to each school district’s Chief Finance Officer, Director of Business Operations, or Director of Human Resources.7 In addition, requests were sent to all school districts asking for individual payroll records with 53 districts agreeing to provide the requested records. Table 2 provides sample means for all 115 school districts and similar data for the districts that provided employment information. Our sample districts are slightly smaller and had a somewhat higher proportion of students qualifying for subsidized lunch while the demographic characteristics of the school personnel are very similar. None of the values for the sample districts shown in Table 2 were statistically different from the state averages for all districts. The survey responses from the districts were then merged with individual-level payroll records from March 2013. The survey asked for information on all retirement savings plans offered by each district, the management and oversight of these plans and vendors, as well as for information about any employer-provided communications and other informational efforts related to retirement plans.8 The payroll data included each employee’s age, years of service, earnings, contributions to retirement savings plans, and an occupational code. Public school personnel in all school districts across North Carolina are enrolled in the Teachers’ and State Employees’ Retirement System (TSERS) and in the State Health Plan.9 TSERS is a defined benefit plan with a formula of 1.82% per year of service times an average salary over the 4 highest years of earnings. An employee retiring with 30 years of service would have a retirement income replacement ratio of about 55% of final salary. In addition to this pension benefit, retirees would also receive a Social Security benefit. Thus, most full career employees could expect to receive retirement income (TSERS plus Social Security) representing 70 to 80% of final salary. 7

The survey was conducted in coordination with the North Carolina Department of State Treasurer, Retirement System Division. A link to the online survey was sent by the Director of the North Carolina Retirement Systems to all 115 school districts. Responses were received 90 of the 115 districts between August 2013 and November 2013. We spoke with staff at 79 of the districts to verify responses and aid in completing the survey. In addition, the websites of each district was examined to assess the quality of information provided to school personnel concerning the plans offered by the district. Clark et al. (2016) describe the survey and the 403(b) landscape in North Carolina. 8 Additional data were gathered from districts’ websites and other public sources to: (1) validate districts’ survey responses, and (2) create an index of district website quality. A detailed discussion of the survey and its results is provided in Clark et al. (2016). 9 Currently, all retiring teachers who were hired before October 1, 2006 and those hired after that date who retire with 20 or more years of service are eligible to enroll in a 70/30 retiree health plan without any premium. Retirees hired on or after October 1, 2006 with 10 to 19 years of service must pay 50% of the premium and those with 5 to 9 years can remain in the state plan but must pay the full premium. No subsidy is given for spousal or dependent coverage, although access is permitted. The retiree health plan is subject to change by the General Assembly. A summary of employee benefits is available at: http://www.ncleg. net/FiscalResearch/fiscal_briefs/Fiscal_Briefs_PDFs/Fiscal_Brief_Value_of_Employee_Benefits.pdf, [accessed June 2015].

J Labor Res Table 2 Representativeness of sample districts All districts

Sample

N = 115

N = 53

Students ADM

12,325.7

11,445.3

% Needy

63.8%

66.0%

% Female

77.4%

77.7%

% Black

19.4%

20.7%

% White

77.5%

76.4%

% Other

3.0%

2.7%

Employees

Notes: Average Daily Membership (ADM) is a measure of count of students for a school year. For details about calculation of ADM in North Carolina http://www.dpi.state.nc.us/fbs/accounting/data/ % Needy defined as students receiving free or reduced price lunch as a percent of total students. None of the differences between the means for all 115 districts and the 53 districts in our analysis are statistically significant. Source: http://www.dpi.state.nc.us/fbs/resources/data/

All school personnel are eligible to contribute to the state-managed 401(k) plan and most also have the option of contributing to the state managed 457 plan. Prudential is the vendor for both plans. In addition, employees in all school districts can enroll in a district managed 403(b) plan. There are no employer match in any of these plans and none of the plans have adopted automatic enrollment. Because of the relatively generous defined benefit plan and coverage by Social Security, one would anticipate that compared to private sector workers who are not covered by a defined benefit plan, school personnel in North Carolina would have a lower need for additional retirement saving. Therefore, we expect lower participation rates and lower levels of contributions to these voluntary retirement savings plans by public employees. While most of the annual salary for school district personnel is paid by the state according to a fixed pay scale, most districts provide a local supplement to some employees’ state-funded base salary. Average annual district supplements for teachers vary from zero dollars in six districts to more than $6000.10 Other employee benefits (e.g. vacation) and job requirements (e.g., number of work days and number of hours worked per day) are similar for all North Carolina school districts. All school districts in the merged data offer a locallymanaged 403(b) plan and the state-managed 401(k) retirement savings plan. Thus, all districts in the sample offer employees the choice between at least two retirement savings plans and 48 of the 53 districts allow their employees to select from all three types of retirement savings plans.

10

The 2013 average supplement for teachers, principals, and assistant principals in each district are made publicly available through the North Carolina Department of Public Instruction: http://apps.schools.nc. gov/pls/apex/f?p=1:25:1259484044504001::NO::P25_SELECTYEAR:2014, [accessed June 2015]

J Labor Res

School Personnel and their Saving Decisions To estimate plan choice and saving decisions, the payroll records of 71,156 employees of school districts across North Carolina are examined. Using these records, we are able to determine whether workers contributed to the various plans in March of 2013 along with their contributions to each plan type offered by the school district. The means for the entire sample are presented in Table 3, Column 1. As expected, the labor force is more than three quarters female. The mean age of employees is 40 years and the employees have an average length of service of over 11 years. Educational professionals, composed of teachers, TAs, mentors, coaches, speech therapists, audiologists, tutors and instructors, represent 78.6% of the sample. Almost three quarters of the employees are white and 20.6% are black. About 60% of the sample is married but one third of the employees in the sample are single females. Table 3 Means of Sample Population by Participation Status Characteristics

Total Sample

Participating in at least one plan*

(1)

(2)

Age

40.06

46.62

Male

21.61%

19.26%

Female

78.39%

80.74%

Single

41.99%

40.92%

Married

58.01%

59.08%

Single Male

9.27%

7.90%

Married Male

12.34%

11.36%

Single Female

32.72%

33.02%

Married Female

45.67%

47.72%

White

72.90%

73.74%

Black

20.61%

19.63%

Hispanic

1.41%

1.03%

Other

5.08%

5.61%

Managers

3.09%

4.95%

Education Professionals

78.60%

82.36%

Support

18.30%

12.69%

Salary Q1

20.00%

16.73%

Salary Q2

20.00%

12.78%

Salary Q3

20.15%

20.27%

Salary Q4

19.96%

24.66%

Salary Q5

19.88%

25.56%

Tenure (years)

11.28

13.66

N

71,156

22,791

*All of the reported means for participants in Column 2 are statistically significantly different from the means for the entire population shown in Column 1 at the 0.1% level except for the mean of single females and Salary Q3 Payroll data from 53 school districts in North Carolina, March 2013

J Labor Res

Thirty-two percent of all school employees have made the decision to contribute to at least one of the savings plans. This participation rate is considerably lower than that reported in studies of the broader labor force. For example, Papke (1995), Madrian and Shea (2001), Munnell et al. (2002), Choi et al. (2004), and Huberman et al. (2007) report participation rates of 70 to 80% with rates being higher when automatic enrollment had been adopted and higher with more generous employer matches. Since North Carolina school employees are covered by a defined benefit plan and the retirement savings plans do not have an employer match nor have they adopted automatic enrollment policies, it is not surprising that the participation rate for this labor force is considerably lower. Column 2, Table 3 reports the means of these individuals who are having contributions deducted from their paychecks. The average age of individuals contributing to a savings plan is about 6.5 years greater than the sample at large. It is interesting to note that women are more likely to be participating in a plan as are managers and educational professionals while a much smaller proportion of the support staff are contributing to a savings plan. Participants have about 2.4 more years of tenure than the general labor force. The means for those who are participating in a savings plan are significantly different from those of the entire labor force. Individuals are not restricted to contribute to only one plan; in fact, they have the option of contributing to any or all of the plans offered by their district. Table 4 indicates the proportion of employees who contribute to the various plans. First, we note that 32.0% of the individuals are contributing to at least one retirement savings plan. Of those contributing to any plan, 20,388 or 18.8% of employees, are contributing to exactly one plan. This means that 89.5% of all participants contribute to a single savings plan. The district-managed 403(b) plan is the most popular plan with 14.7% of all employees (45.9% of all participants) contributing only to the 403(b) plan while 10.0% of employees (31.1% of participants) contribute solely to the state-managed 401(k) plan. Only 4.0% of all workers or 12.4% of participants contributed only to a 457 plan. Earlier, we discussed how combining either a 401(k) or a 403(b) plan with a 457 could have some financial advantages due to differences in tax regulations. In this sample of North Carolina school personnel, 971 individuals were contributing to both 401(k) and the 457 plans, another 292 employees participated in both 403(b) and 457 plans while another 176 workers contributed to all three plans. Thus, 6.3% of individuals contributing to a retirement savings plan (or 2.0% of all employees) have chosen a strategy that could allow them to exceed the 2013 contribution limit of $17,500 per year by contributing to a 457 plan plus another savings plan. Finally, 964 employees contributed to the 401(k) and the 403(b) plans which are to the same government rules limiting annual contributions. These aggregate data yield several important observations. First, almost 90% of all individuals contributing to any retirement savings plan have chosen to participate in a single plan. Second, only 6% of all participants have selected a saving strategy that could enable them to exceed the annual contribution limit by contributing to a 457 plan plus another retirement savings plan. The contributions of these individuals are examined in detail in Section V. Third, it is interesting to note that 964 individuals or 4.2% of participants are making simultaneous contributions to the state 401(k) plan and the district 403(b) which have the same annual contributions limit.

J Labor Res Table 4 Distribution of Employees by Retirement Savings plans Characteristics

Number of Individuals

Participation Rate Percent of Sample

Total Sample

71,156

100.0%

Not contributing to any plan

48,365

68.0%

Contributing to one or more plans

22,791

32.0%

Type of Coverage Only 401(k)

7095

10.0%

Only 403(b)

10,464

14.7%

Only 457

2829

4.0%

401(k) and 403(b)

964

1.4%

401(k) and 457

971

1.4%

403(b) and 457

292

0.4%

401(k), 403(b) and 457

176

0.2%

Payroll data from 53 school districts in North Carolina, March 2013

Choice of a Retirement Savings Plan The following analysis is based on the assumption that school personnel first determine if they want to allocate some of their monthly earnings from current consumption to a qualified retirement savings plan offered by their employer. This assumption is similar to that in lifecycles model in which individuals are first assumed to decide to save while working in order to finance consumption in retirement and then decide on the allocation of their investment portfolio. Available information is limited to that provided in the administrative records of the districts which do not include total household income and wealth or other household information. Having made the decision to save through an employer plan, individuals then consider the plans that are available to them and select the plan that best fits their saving preferences. This latter choice may be influenced by the various characteristics of the three plans and the marketing done by the various vendors. Not all individuals in our sample have the opportunity to enroll in a 401(k), a 403(b), and a 457 plan. Five of the 53 districts do not offer a 457 plan. Our analysis of the savings plan choice will take this feature into account, allowing us to study the full sample consisting of the 53 districts. The payroll data for these plans include information on 71,156 full-time employees. The Participation and Contribution Decisions The individual level means of the independent variables for the full sample of employees are shown in Column 1, Table 5. The decision to participate in any plan is estimated using a probit model with the dependent variable being one if the individual was participating in any of the three plans, zero otherwise. The marginal effects at the mean of the variables derived from the estimated probit coefficients are reported in Column 2, Table 5. The estimated results indicate that managers are 9.5 percentage

J Labor Res Table 5 Participation in Any Retirement Savings plan Means (1)

Salary Q2 Salary Q3

0.079***

0.593***

(0.008)

(0.067)

0.203***

1.278***

(0.008)

(0.065)

0.213

Salary Q5 11.28

Tenure Squared Managers

Marginal effect on contribution (3)

***

Salary Q4

Tenure (years)

Marginal effect on participation (2)

3.09%

Support

18.30%

Black

20.61%

1.305***

(0.008)

(0.064)

0.206***

1.399***

(0.008)

(0.068)

0.038***

0.245***

(0.001)

(0.006)

−0.001***

−0.006***

(0.000)

(0.000)

0.095***

0.286***

(0.010)

(0.066)

−0.113***

−0.905***

(0.005)

(0.040)

0.020***

−0.114***

(0.005)

(0.035)

Hispanic

1.41%

−0.030**

−0.176

(0.016)

(0.131)

Other Ethnic Group

5.08%

0.023**

0.264***

(0.009)

(0.072)

−0.010

−0.158***

Single Male

9.27%

Married Male

12.34%

Single Female

32.72%

(0.007)

(0.049)

−0.025***

−0.211***

(0.006)

(0.046)

0.028***

0.047

(0.004)

(0.031)

Observations

71,156

71,156

Mean of Dependent Variable

0.320

1.726

points more likely to participate in a retirement savings plan compared to educational professionals while the support staff are 11.3 percentage points less likely to enroll in one of the retirement savings plans. Holding other factors constant, blacks are slightly more likely to participate and Hispanics are slightly less likely to contribute to one of the plans. Married men are 2.5 percentage points less likely to enroll in one of these plans compared to married women while single females are 2.8 percentage points more likely to be participants in the employer plans. To allow for a nonlinear relationship between annual earnings and participation in any retirement savings plan, the model includes a series of dichotomous variables

J Labor Res

indicating in which earnings quintile of the distribution the individual’s earning falls; the lowest quintile is omitted from the equation.11 As expected, employees with a salary in the highest three quintiles are about 20 percentage points more likely to contribute to a retirement savings plan than the workers with a salary in the first quintile. All regressions include fixed effects for the school district where the individual was employed to allow for variation across the districts. In Table 5, we also present the results of the analysis of the contribution decision. The payroll data we obtained are for only a single month, March. Using these data, we estimate annual contributions by multiplying the March contribution times 12 months.12 The dependent variable is the total annual contribution to the various plans divided by the annual salary (times 100 so the dependent variable is reported in percentage points). A Tobit model with the same explanatory variables as the probit model is then estimated.13 The mean of the dependent variable over the entire sample is 1.7% and the mean over individuals who have a strictly positive contribution is 5.5%. Interestingly, this contribution rate for those who contribute is similar to that reported in studies of 401(k) in the general labor force (Goldhaber and Grout 2016; Huber, Iyengar, and Jiang, 2007). The marginal effects on contributions at the mean of the variables for the Tobit model are reported in Column 3, Table 5. The first thing to observe is that most explanatory variables are statistically significant at the usual significance levels and the sign of the marginal effects for the contribution is the same as for the participation decision. The salary variables have a large impact; on average holding everything else constant, individuals in the highest quintile contribute 1.4 percentage points more of their salary than those in the first quintile. Tenure has an important impact on average contribution. The maximum impact is around 20 years of tenure at which point individuals contribute 2.5 percentage points more than new hires. Managers contribute more (0.29 percentage points), while support personnel contribute less (0.91 percentage points). The average contribution of blacks and Hispanics is lower than whites (0.11 percentage points and 0.18 percentage points more respectively). The differences in contribution across gender and marital status are also meaningful. For example, married males contribute on average 0.21 percentage points less than married females.

11 Using alternative functional forms for earnings (such as deciles and quadratic function) results in no substantial changes in the estimated effects on participation or contribution. 12 We multiply March contributions by 12 since most school personnel are paid over 12 months even though they are considered 10-month employees. This estimate of annual contribution based on a single month of data could be an overestimate of annual contributions if the individual stops or lowers contributions in future months. Of course, it could also be an underestimate if individuals make year-end contributions or increase contributions due to salary increases during the year. 13 One reason employees contribute to retirement saving plans is the preferential tax treatment provided to this form of saving. The size of the tax saving from these contributions is a function of household income and not just earnings. The payroll data used in this analysis does not include any information on income other than earnings. To use earnings as a measure of taxable income and to calculate a tax incentive for contributions would be misleading. The inclusion of the earnings variables in the contribution equation provides a proxy for the tax effect as higher earnings are likely to be in a higher tax bracket.

J Labor Res

Choice of Retirement Savings Plans Having made the decision to contribute to an employer provided retirement savings plan, individuals must now decide among the three plan types offered to them. In these North Carolina school districts, every employee has access to 401(k) and 403(b) plan, and about 93% have access to a 457 plan. Among participants in retirement plans, 31.1% only contribute to a 401(k) plan, 45.9% only contribute to a 403(b) plan and 12.4% only contribute to a 457 plan. Individuals contributing to a 401(k) and a 403(b) make 4.2% of the sample, about the same percentage as individuals who selected a 401(k) and a 457 plan. Individuals investing in a 403(b) and 457 plans make about 1.3% of the sample. Finally, one should notice that 0.8% of participants are contributing to all three plans. We now examine the factors that influence the decision to enroll in each of these plan types. A feature of this sample is that not every individual has access to a 457 plan (local or state). As mentioned above, about 7% of the sample is from one of the five districts that do not offer a 457 plan so they can only enroll in the 401(k) or 403(b) plans. We can nonetheless jointly study the choice of pension plan(s) for all individuals in the 53 districts using a typical random utility framework. The utility provided by a given option is taken to be a linear function of independent variables and an error term. Of the possible options, the individual selects the one that provides the highest utility. This is the framework underlying the basic multinomial logit model. The same random utility framework can be applied to our problem with one small modification. For most of the individuals, they have seven possible choices: only 401(k), only 403(b), only 457, both 401(k) and 403(b), both 401(k) and 457, both 403(b) and 457, and all three plan types. The individuals in the 5 districts not offering a 457 plan only have three options: only 401(k), only 403(b), both 401(k) and 403(b). Assuming the error term follows an extreme value distribution we obtain the typical multinomial logit expression for the probability that individual i chooses optionj,pi(j):
0 exp X i β j
pi ð j Þ ¼ 0 ∑k∈K i exp X i βk where Xi is the vector of independent variables for individual i, βj is the vector of parameters for option j. The term Ki represents the set of options available to individual i (3 or 7 options, depending on whether or not the district of individual i is offering a 457 plan). In the typical use of the multinomial logit model, every individual would have the same set of options. Estimates of the parameters and standard errors are obtained by maximum likelihood. Standard errors for the marginal effects are computed with the Delta method. District fixed effects are included in all regressions to allow for variation across the districts in information provided to employees and the quality of HR policies. To ensure identification, as in the multinomial logit, we set to zero the parameters of one option common to everyone (i.e., we set β1 = 0). From this model, we can compute marginal effects, how the probability of choosing option j changes when an element of Xi changes, and since these probabilities sum up to one the marginal effects sum up to zero.

J Labor Res

The estimated marginal effects at the means of the variables are shown in Table 6. The analysis of participation in a retirement savings plan indicates that the choice of a plan is significantly determined by occupational and demographic characteristics. Managers are 14.7 percentage points more likely to only participate in the 401(k) and 13.6 percentage points less likely to only enroll in a 403(b) plan compared to classroom personnel. The retirement saving decisions of the support staff follow a similar pattern. These differences in saving choices may reflect that managerial and support personnel may be more familiar with 401(k) plans since they may have spent working years outside of public education while educational professionals typically will have spent more of their worklife in public education. These results indicate that, everything else equal, classroom personnel are the primary users of the 403(b) plan. Holding other characteristics constant, blacks are 10.5 percentage points more likely to select a 403(b)-only plan and 11.1 percentage points less likely to select a 401(k)-only plan. Compared to married women, single women and single men are about 5.6 and 2.9 percentage points respectively more likely to enroll in only a 403(b) plan and 5.2 and 3.4 percentage points less likely respectively to select only a 401(k) plan. With the exception of being less likely to contribute to only a 457 plan if salary is in the fifth quintile, annual salary has no significant effect on the choice to contribute by type or number of plans. These estimates indicate that highly compensated employees and managers are no more likely to contribute to 457 plans and either 401(k) or 403(b) plans than other employees Thus, in this analysis, we find no economically or statistically significant evidence that highly compensated employees and managers are more likely to select plan combinations that enable them to take advantage of tax laws to make annual contributions exceeding the maximum contribution limits. This could be due to the relatively small number of individuals who combine other plans with a 457 plan.

Participation in 457 Plan and 401(k)/403(b) Plans One policy concern is whether highly compensated individuals exploit the tax rules that provide separate contribution caps for 457 plans as compared to 403(b) and 401(k) plans. To further address this possibility, we examine the sample of only individuals who contributed to a retirement savings plan and had access to a 457 plan.14 Column 1, Table 7 presents the means for district personnel that contribute to a 457 plan and at least one other plan. These are individuals who potentially could make the maximum annual contribution to both the 457 plan and the maximum to either the 401(k) or the 403(b) plan offered by their district. These individuals who are contributing to plans that potentially could allow them to exceed the IRS cap on annual contributions to a single plan represent 6.8% of all employees contributing to any savings plan. Column 2 reports the means for those who contribute to only one plan or to a 14

This latter restriction means that individuals in the five districts that do not offer their employees the option of contributing to a 457 plan are dropped from this analysis. This results in a sample of 21,150 employees contributing to at least one plan instead of the 22,791 used in the earlier analysis.

J Labor Res Table 6 Choice of Retirement Savings plans Only 401(k) Salary Q2 Salary Q3

Only 403(b)

Only 457

401(k) & 403(b)

401(k) & 457

403(b) & 457

All 3 plans

−0.010

0.009

−0.006

0.009

−0.005

0.003

0.000

(0.020)

(0.021)

(0.008)

(0.008)

(0.005)

(0.007)

(0.002)

−0.026

0.014

−0.000

0.013

−0.006

0.004

0.001

(0.020)

(0.023)

(0.008)

(0.010)

(0.005)

(0.008)

(0.021)

−0.002

0.000

−0.009

0.017

−0.012

0.004

0.001

(0.020)

(0.022)

(0.008)

(0.012)

(0.007)

(0.010)

(0.016)

0.005

−0.016

−0.014*

0.022

−0.005

0.005

0.002

(0.023)

(0.026)

(0.008)

(0.015)

(0.005)

(0.012)

(0.029)

−0.001

0.002

−0.004***

0.004

−0.002*

0.000

0.000

(0.002)

(0.003)

(0.001)

(0.003)

(0.001)

(0.000)

(0.003)

0.000***

−0.000*

0.000***

−0.000

0.000

−0.000

−0.000

(0.000)

(0.000)

(0.000)

(0.000)

(0.000)

(0.000)

(0.000)

Managers

0.147***

−0.136***

0.001

−0.002

−0.008

−0.001

−0.000

(0.017)

(0.020)

(0.008)

(0.004)

(0.006)

(0.003)

(0.004)

Support

0.125***

−0.111***

−0.007

−0.011

0.006

−0.002

0.000

(0.013)

(0.014)

(0.005)

(0.008)

(0.004)

(0.005)

(0.005)

−0.111***

0.105***

−0.003

0.005

0.002

0.002

0.001

(0.011)

(0.016)

(0.005)

(0.004)

(0.003)

(0.004)

(0.016)

−0.024

0.018

0.006

−0.009

0.013

−0.004

−0.000

(0.038)

(0.039)

(0.016)

(0.013)

(0.010)

(0.009)

(0.006)

−0.014

0.001

0.004

0.008

−0.000

0.002

−0.001

(0.019)

(0.020)

(0.008)

(0.007)

(0.005)

(0.004)

(0.019)

−0.034**

0.029*

−0.004

0.002

0.003

0.002

0.000

(0.014)

(0.015)

(0.006)

(0.004)

(0.004)

(0.005)

(0.005)

Salary Q4 Salary Q5 Tenure Tenure2

Black Hispanic Other Single Male

0.008

0.002

−0.008

−0.005

0.003

0.000

0.000

(0.012)

(0.012)

(0.005)

(0.005)

(0.003)

(0.001)

(0.001)

−0.052***

0.056***

−0.003

−0.000

−0.001

−0.000

0.001

(0.009)

(0.011)

(0.004)

(0.002)

(0.002)

(0.001)

(0.010)

Observations

22,791

22,791

22,791

22,791

22,791

22,791

22,791

Mean of dep. Variable

0.311

0.459

0.124

0.042

0.043

0.013

0.008

Married Male Single Female

The sample includes all individuals who are contributing to at least one retirement savings plan. Individuals who do not contribute to a plan are not included in this analysis. Columns 1–7 show marginal effects at the mean from multinomial logit. District fixed effects are included, where the smallest four districts have been merged into one (27, 37, 44, 56 individuals respectively). Standard errors are between parentheses. The B457^ plan includes both the local 457 plan and the North Carolina 457 plan. In this sample, 1641 individuals do not have access to a 457 plan. Statistical significance: 10% (*), 5% (**), 1% (* * *). The salary variables are the same as in Table 5 (quintiles over the whole sample of 71,156 participants)

401(k) plan plus a 403(b), thus given their contributions to existing plans, all of the contributions of these individuals would be subject to the single IRS limit on annual contributions.

J Labor Res Table 7 Probit Model for Contributions to 401(k) and/or 403(b) with 457 Compared to Any Other Combinations of Contributions to Savings plans Means: 457 plus 401(k) and or 403(b) (1)

Means: all other contribution patterns (2)

Marginal effects (3)

Salary Q2

13.72%

12.55%

−0.002

Salary Q3

21.59%

20.14%

−0.002

Salary Q4

19.57%

24.63%***

−0.004

Salary Q5

21.59%

25.50%***

Tenure (years)

12.70

13.67***

(0.007) (0.006) (0.007) 0.008 (0.009) −0.001 (0.001) Tenure Squared

0.000 (0.000) −0.010

Managers

4.94%

4.95%

Support

14.83%

11.69%***

Black

26.25%

19.64%***

0.008

Hispanic

1.25%

1.06%

0.015

Other Ethic Group

5.57%

4.25%**

Single Male

7.90%

7.78%*

Married Male

9.33%

11.31%

(0.009) 0.006 (0.006) (0.007) (0.016) 0.004 (0.007) 0.011 (0.009) 0.006 (0.006) Single Female

32.31%

33.39%

0.000 (0.003)

Observations Mean of Dependent Variable

1436

19,714

21,150 0.0783

Comparing the means of these two groups provides no support that managers and highly compensated workers are more likely to take advantage of the separate IRS rules allowing the separate contribution caps in 457 plans. Relative to contributions to all other plans shown in Column 2, a smaller percentage of individuals in the highest quintile of earnings contribute to a 457 plans plus another retirement savings plan. Similarly, we find that managers are not more highly represented among individuals contributing to a 457 plus another plan. To further investigate this issue, we estimate a probit model for this sample with the dependent variable taking a value of 1 if the individual contributed to a 457 plan and at

J Labor Res

least one other retirement savings plan, zero otherwise. Being in the upper quintiles of the earnings distribution is not associated with being more likely to participate in both a 457 and another savings plan. In fact, none of the independent variables are statistically significant in this analysis which further indicates that among North Carolina school personnel, there is no evidence that senior administrators are adding 457 plans to their benefit mix in order to allow them to make contributions to retirement savings plans over the single plan limit. It is interesting to note that there are substantial differences in the proportion of employees contributing to multiple retirement savings plans across the school districts in our sample. Our analysis indicates that none of the independent variables have an impact on the likelihood of choosing a 457 plan on top of a 401(k) or 403(b) plan which indicates that differences in the school labor force does not explain the large variation in enrolling in multiple plans. The proportion of employees who contribute to multiple plans by school district is shown in Fig. 1. The columns in the figure indicate the percentage of all participants that contributed to 457 plan plus either the 401(k) or the 403(b) plan across districts. The bars are sorted from smallest to largest district in terms of employees. There is considerable variation across the districts with a few districts showing that over 30% of 0.4

Fraction contributing to multiple plans

0.35

0.3

0.25

0.2

0.15

0.1

0.05

0 0

5

10

15

20

25

30

35

40

45

District ID

Fig. 1 Proportion of Participants That Are Contributing to Multiple Plans Allowing Them to Exceed the Contribution Limit. The bars show the ratio of the number of individuals who are making contributions a 457 plus at least one additional retirement saving plans offered in each school district divided by the number of persons who are contributing to only a single plan plus those that are contributing to both 401(k) and 403(b) plans. Each bar represents an individual district. The districts are shown from left to right by smallest to largest district in terms of employees. The height of the bars indicate the proportion of all participants who are contributing to two or more plans that theoretically could allow their total contributions to exceed the annual cap on contributions of $17,500

J Labor Res

all contributors are contributing to both a 457 plan and another plan, while other districts below 5 % of employees being this type of dual contributors. These findings suggest that district management of the plans, oversight by central offices, and marketing differences may be an important determinant of plan choice by school employees.

Annual Contributions and Participation in Multiple Plans In order to provide additional evidence on whether high-income school managers are using the separate contribution cap on 457 plans to increase their tax-deferred saving beyond the annual contribution limit, we now examine predicted annual contributions for all employees contributing to any retirement savings plan. In 2013, the year of our payroll data, the contribution limit for each savings plan was $17,500. As described earlier, our payroll data is only for a single month so we do not have a precise estimate of annual contributions to each of the savings plans. We estimate annual contributions by multiplying the contribution in the observable month by 12.15 Of course, individuals can change their contributions from one month to another and also make one-time contributions throughout the year. Panel A, Table 8 shows the distribution of estimated annual contributions by four levels of contributions. Almost 94% of all school employees who made contributions to any retirement savings plan had estimated annual contributions of less than $5000. There were a total of 308 employees, or 1.4% of all participants, whose estimated contributions were more than $15,000 or near the contribution maximum. Among these, only 28 individuals had made contributions to a 457 plan and a 403b plan, a 401(k) plan, or both the 403(b) and the 401(k) plans available to them. Thus, relatively few individuals are making estimated annual contributions that approach the maximum allowable contribution. Interestingly, 1402 of the 1439 individuals (97.4%) who contributed to a 457 plan plus another retirement savings plan had annual contributions of less than $10,000. The payroll data indicate the most employees who contributing to multiple plans have total contributions well below the $17,500 annual limit on contributions in 2013 to either a 403(b) or a 401(k) plan. These individuals are obviously not contributing to multiple plans to evade the annual cap on contributions. These data indicate participating in multiple plans is mainly done by individuals whose total annual contribution is well below the contribution cap. It appears that virtually all individuals who choose to contribute to a 457 plan plus another retirement savings plan are doing so for other reasons besides trying to exceed the annual contributions cap on a single plan. Panel B, Table 8 takes a further look at individuals whose estimated annual contributions exceeded $15,000. We find that 163 individuals had estimated annual contributions between $17,500 and $23,000. Of these employees, 150 were over age 50 and thus eligible for the catch up provision that allowed contributions up to $23,000 in 2013. Only 10 of these high contributors were contributing to a 457 plan and either the 15

Salaries of school personnel are set by the General Assembly. Increases in annual salaries are typically made in July. Our data are from March, so we might expect individuals to consider contribution changes during the second half of the year in response to increases in salary. Of course, some employees have indicated that they want to contribute a percent of their salary so contributions may increase automatically to reflect increases in salaries.

153

21,350

401(k) + 403(b) + 457

Total

33

877

20

22

23

256

1

2

6

21

Cells shows number of individuals in each plan combination.

920

254

401(k) + 457

403(b) + 457

61

177

308

2

14

12

6

22

75

92

– 163

– 107

5

5

4

9

48

1

2

1

12

21

38

2

8

5

1

1

6

15

103

2681

876

457

401(k) + 403(b)

125 78

70

327

23,000+

311

17,500-23,000

6466

15,000-17,500

10,000

15,000+

401(k)

10,000-15,000

403(b)

5000-10,000

Panel B. Individuals with estimated annual contributions exceeding $15,000

0–5000

Panel A. All contributors

Yearly Contributions

Table 8 Estimated Annual Contributions by Plan Type

J Labor Res

J Labor Res

401(k) or the 403(b) plan. The mean salary of these 10 employees was only $37,990 with the highest salary being $58,500. Thus, most of these relatively high levels of annual contributions would seem to be due to older workers taking advantage of the catch-up provision and not using the separate limits on 457 plans to exceed the annual cap on contributions. A total of 38 individuals are projected to have annual contributions that exceeded $23,000, the annual limit including the $5500 catch-up. Only 15 of these participants are contributing to the 457 plan plus the 403(b) and the 401(k) plans. It seems likely that our estimates of annual contributions overestimate the annual contributions of some of these individuals as contributions in excess of $23,000 without some funds being contributed to the 457 plan would violate the tax code on contributions. In summary, there is little or no evidence that highly compensated North Carolina school personnel are making participation and contribution decisions with the objective of exceeding the annual contribution limit for retirement savings plans. This finding is not surprising given the moderate level of earnings for most of the sample.

Concluding Observations Public employees face an expanded choice of retirement savings plans relative to similar workers in the private sector. Federal tax policies allow state and local governments the opportunity to offer 401(k) plans that dominate the private sector, along with 457 plans and public schools and certain other organizations also can offer 403(b) plans to their employees. Our analysis of public school employees in North Carolina illustrates a range of economic and demographic factors that influence the decision to enroll in one of these plans in a similar fashion as private sector employees. To our knowledge, this analysis of plan choice is the first attempt to examine how public employees decide among alternative retirement savings plans. The three alternative plans have similar characteristics and IRS guidelines; however, several differences in the plans imply that certain workers may prefer one plan type over the others. The multinomial probit results indicate that the significant determinants of plan choice are occupation, race, gender, and marital status. These findings might reflect how individuals respond to the marketing efforts of vendors or to work and retirement plans. Finally, we find no indication that managers and highly compensated employees are systematically using the availability of multiple plans to exceed the annual contribution limit on retirement saving. These results indicate that approximately one-third of North Carolina public school personnel decide that they would like to save additional monies for retirement over and above being covered by a defined benefit plan and Social Security. The cost of saving a dollar for these employees is one dollar as there is no employer match for employee contributions to the retirement savings plans. Since the state has not adopted automatic enrollment into one of these plans, participation is entirely due to workers’ desire for additional retirement saving. While there are no direct comparisons to our results for public employees contributing to supplemental retirement savings plans, several related studies seem to suggest that public employees already have sufficient expected retirement income and do not value additional benefits from defined benefit plans at their cost. For example,

J Labor Res

Fitzpatrick (2015) finds that public school employees value the present value of $10 of annuity income at only $2 in current wages. At first, her findings may seem at odds with the fact that many teachers actually engage in additional retirement savings; however, contributions to a retirement savings plan which build retirement wealth can be viewed differently from a promise of an increased retirement annuity. There is ample evidence individuals often have a bias toward lump sums over annuities (Chalmers and Reuter 2012; Warner and Pleeter 2001; Brown 2001) or as Fitzpatrick notes her findings are consistent with individuals favoring lump sum distributions over illiquid annuitized future benefits or in our case, the desire to contribute to a savings plan and accumulate personal wealth is different than the desire to have a larger retirement annuity.16 Marginal effects are calculated at the mean values based on probit estimates of participation in any plan. District fixed effects are included in the estimation equation. Statistical significance: 10% (*), 5% (**) and 1% (***). The salary variables represent dummy variables for the participant’s salary being in a given quintile of the salary distribution. Marginal effects are calculated at the mean values based on probit estimates of participation in a combination of plans that allow doubling of retirement contributions (401(k) and 457, 403(b) and 457, or 401(k) and 403(b) and 457) with the alternative being a combination of plans that does not (only 401(k), only 403(b), only 457, 401(k) plus 403(b)). District fixed effects are included. This sample does not include individuals from the five districts that do not offer a 457 plan (1641 individuals). Statistical significance: 10% (*), 5% (**) and 1% (***). For column 2, the stars represent statistical difference between the means of columns 1 and 2. The salary variables are the same as in Table 5 (quintiles over the whole sample of 71,156 participants). Funding This study was funded, in part, by a grant from the TIAA Institute to Robert Clark. Compliance with Ethical Standards Conflict of Interest

The authors declare that they have no conflict of interest.

References Beshears J, Choi J, Laibson D, Madrian B (2011) Behavior economics perspectives on public sector pension plans. Journal of Pension Economics and Finance 10(2):315–336 Brown J (2001) Private pensions, mortality risk, and the decisions to annuitize. J Public Econ 82:29–62 Chalmers J, Reuter J (2012) How do retirees value life annuities? Evidence from public employees. Rev Financ Stud 25(8):2601–2634 Chingos MM, West MR (2015) Which teachers choose a defined contribution pension plan? Evidence from the Florida retirement system. Education Finance and Policy 10(2):193–222 Choi J, Laibson D, Madrian B, Metrick A (2002) Defined contribution pensions: plan rules, participation decisions, and the path of least resistance. In: Poterba J (ed) Tax policy and the economy. MIT Press, pp 67–113 16

The retirement system of North Carolina has a transfer benefit option that allows retirees with funds in the statement managed retirement savings plans to convert these to an annuity. A very small number of retirees utilize this option.

J Labor Res Choi J, Laibson D, Madrian B (2004) Plan design and 401(k) saving outcomes. Natl Tax J 57(2):275–298 Choi J, Laibson D, Madrian B (2009) Reducing the complexity costs of 401(k) participation: the case of quick enrollment. In: Wise DA (ed) Developments in the economics of aging. University of Chicago Press, Chicago, pp 57–82 Clark R, Hanson E (2013) 403(b) Plans for Public School Teachers: How They Are Monitored and Regulated In Each State? TIAA-CREF Research Dialogue, Issue No. 107. https://www.tiaa-crefinstitute. org/public/pdf/institute/research/dialogue/107b.pdf Clark R, Pitts M (1999) Faculty choice of a pension plan: defined benefit vs. defined contribution. Ind Relat 38(1):18–45 Clark R, Richardson D (2010) Who’s Watching the Door? How Controlling Provider Access Can Improve K12 Teacher Retirement Outcomes. TIAA-CREF Research Dialogue. https://www.tiaa-crefinstitute. org/public/institute/research/dialogue/rd_98.html Clark R, Maki J, Morrill M (2014) Can simple informational nudges increase employee participation in a 401(k) plan. South Econ J 80(3):677–701 Clark R, Hanson E, Morrill M, Pathak A (2016) Supplemental plan offerings and retirement saving choices: an analysis of North Carolina school districts. Journal of Pension Economics and Finance 15(3):333–355 Fitzpatrick M (2015) How much are public school teachers willing to pay for their retirement benefits? American Economic Journal Economic Policy 7(4):165–188 Goldhaber D, Grout C (2013) Which plan to choose? The determinants of pension system choice for public school teachers. Journal of Pension Economics and Finance 12(1):1–25 Goldhaber D, Grout C (2016) Pension choices and the saving patterns of public school teachers. Education Finance and Policy 11(4):449–481 Huberman G, Iyengar S, Jiang W (2007) Defined contribution pension plans: determinants of participation and contributions rates. J Financ Serv Res 31(1):1–32 Lusardi A, Mitchell O (2009) How Ordinary Consumers Make Complex Economic Decisions: Financial Literacy and Retirement Readiness, NBER Working Paper #15350, September 2009 Lusardi A, Mitchell O (2014) The economic importance of financial literacy: theory and evidence. J Econ Lit 52(1):5–44 Madrian B, Shea DF (2001) The power of suggestion: inertia in 401(k) participation and savings behavior. Q J Econ 116(4):1149–1187 Munnell AH, Sunden A, Taylor C (2002) What determines 401(k) participation and contributions? Soc Secur Bull 64(3):64–75 Papke L (1995) Participation in and contributions to 401(k) pension plans: evidence from plan data. J Hum Resour 28(2):311–325 Warner J, Pleeter S (2001) The personal discount rate: evidence from military downsizing programs. American Economics Review 19(1):33–53