Charter School Participation in State Teacher Pension Plans AEFP MARCH 2017

Transcription

1 Charter School Participation in State Teacher Pension Plans AEFP MARCH 2017 Authors Kevin Hesla National Alliance for Public Charter Schools Susan Aud Pendergrass National Alliance for Public Charter Schools Michael Podgursky Department of Economics University of Missouri Abstract Teacher pension costs have risen sharply over the last two decades, leading many states to consider alternatives to traditional teacher pension plans. In , charter schools can opt out of state plans and develop their own retirement plans in 19 states. This paper reports findings from an on-going project tracking charter school innovations in this area. This report focuses on the decision of charter schools to participate in the state plan and analyzes the adoption of other options available to charter schools in five states: Arizona, California, Florida, Louisiana, and Michigan. These states have large numbers of charter schools and flexibility regarding state plan participation. This analysis describes the institutional background for these decisions and some of the benefits and costs for charters that choose to opt out of the state pension plan. Descriptive data on participation rates are presented along with multivariate analyses of factors associated with participation. One of the most powerful factors associated with the decision to drop out is whether the charter school is affiliated with a management organization, suggesting that management organizations have better information regarding participation options and greater expertise in providing alternatives to state plans. In addition, charters that opened in recent years are significantly more likely to opt out. The authors wish to thank the Laura and John Arnold Foundation for research support, and Rebecca David and Amy Esker for excellent research assistance. The views expressed here are those of the authors and should not be attributed to their institutions, data providers, or the funders. Any and all errors are attributable to the authors. 1

2 Introduction Across the Unites States, defined benefit teacher pensions represent a large and growing cost in the public education sector. Estimates based on data from the Bureau of Labor Statistics (BLS) and the National Center for Education Statistics (NCES) show that on a per-student basis, inflation-adjusted teacher pension costs paid by employers (excluding Social Security) rose from $511 in 2004 to $1,108 in September 2016 (Figure 1). Further, these rising costs do not include teacher contributions. As a percentage of earnings, employer contributions towards pensions (including Social Security) have risen from 11.9 percent in 2004 to 21.0 percent in September For the sake of comparison, private sector managers and professionals contributed, on average, 10.7 percent towards retirement benefits in September Teachers primarily belong to defined benefit plans that tie their retirement annuity to the highest few years of earnings (usually the final few years). Researchers have shown that in these types of plans the accrual of pension wealth is heavily back-loaded (Costrell and Podgursky, 2009, 2010). That is, very little pension wealth accrues during the first decade or two of work, with pension wealth rising sharply as teachers reach conventional retirement age. After employed teachers reach a certain point in their careers (usually associated with retirement eligibility), pension wealth begins to fall as the reduction in the number of years they can collect their annuity dominates any increase in the value of the annual annuity. Moreover, to access this pension wealth the teacher must quit covered employment resulting in a considerably earlier average retirement age for teachers (Kim, et. al. 2016). Backloaded benefits that require two decades of continuous work with the same employer seem a very poor fit for a mobile, educated workforce. Young, educated workers tend to job shop and try out different opportunities to find a good job match. Since teacher turnover is higher in charter schools than traditional public schools, charters would seem to be exceptionally disadvantaged by this backloaded benefit structure. 2 Unlike most areas of education policy there is very little regulatory space for experimentation by districts. In many states, all public schools are required by law to participate in the state (or, in some cases municipal) pension plans. No school or district can opt out and experiment with an alternative arrangement. However, charter schools are an exception. In , charter schools are allowed to opt out of the state retirement plan in 19 states. 3 However, we know very little about what factors affect the decision of schools to participate in state plans, and importantly, what type of See Appendix. 2

3 retirement plans charters put into place when they choose to opt out. This paper is the first from an ongoing project that analyzes charter school innovation and practice in this area. This paper focuses on the charter school participation decision. The authors have constructed a unique national database on charter school participation in state retirement plans. For this paper, the authors focus on five states that: a) have a large number of charter schools, and b) by one means or another, permit charters to opt out of the state plan. It also draws upon a random survey of charter schools in these five states, with in-depth follow up phone interviews with a subsample of schools. Basic Plan Characteristics Traditional public school teachers in all five states participate in what is termed a final average salary (FAS) defined-benefit (DB) retirement plan. 4 It is useful to break up this definition into two parts. These plans are DB in that they provide the teacher with a retirement check payable until the member s (or surviving spouse s) death. This type of payment is called an annuity. The value of this annuity is tied to the member s work history and level of earnings. The second part of the definition is FAS, which means that the value of the annuity is based not on the earnings over the teacher s entire work history, but only on the highest several years (which are usually the final few years). The general formula for computing a FAS DB annuity is given below: Annual Annuity = FAS * Formula Factor * Service Years Table 1 reports several other factors that may influence the decision of a charter school to participate in a state plan. Vesting refers to the period after a teacher s initial hire that must elapse before they can become covered by the plan. For example, with five-year vesting a teacher who quits after four years will, upon separation, receive their contributions back (usually with interest) but will not receive the portion contributed on their behalf by their employer (nor will they have any claim on future pension benefits). A vested teacher with five or more years of employment has a claim on a future pension benefits. Since charter schools have a higher teacher turnover rate than traditional public schools, this means that fewer of their teachers will make it to vesting. Between 2007 and 2015, nine states passed laws that increased the vesting period for new teachers from 5 to 10 years. In addition, 29 states increased retirement eligibility rules by increasing the retirement age, required years of employment, or both. 5 4 New teachers in Florida and Michigan can choose a defined contribution (DC) alternative to the traditional pension plan. 5 National Association of State Retirement Administrators, Spotlight on Significant Reforms to State Retirement Systems, June 2016 ( 3

4 One factor that may influence charter school participation is whether public school teachers participate in Social Security. The original Social Security Act did not cover state and local employees (including teachers). However, amendments to the Social Security Act in 1951 permitted state and local government units to hold elections to determine whether employees would participate. Of the five states included in this analysis, teachers in three states (Arizona, Florida, and Michigan) participate in Social Security and two do not (California and Louisiana). Federal laws require that charter schools that do not participate in the state plans must participate in Social Security. Thus, in the latter case, teachers in charter schools that remain in the state plan do not participate in Social Security, but if they drop out, they must participate. NAPCS Charter School Database The National Alliance for Public Charter Schools (National Alliance) maintains a comprehensive database of charter schools that includes information on CMO/EMO status, year opened, grades served, and geographic location (amongst many other variables). The National Alliance builds this database by collecting data from both state Departments of Education and the Common Core of Data (which is collected and managed by the U.S. Department of Education). To identify whether a charter school participated in the state pension plan, the National Alliance contacted each state s respective pension fund for a list of charter schools that participated in fund. In Arizona, California, Florida, and Michigan this plan participation data was collected from the respective state pension plan (in the form of either a list of schools that participated in the plan or from the plan s Comprehensive Annual Financial Report that lists the members of the plan). In Louisiana, plan participation data was provided by the charter support organization (the Louisiana Association of Public Charter Schools). The collected school lists were matched to the National Alliance s database by school name. State Pension Law Overview Across all five states, there is a level of complexity and ambiguity that often makes it difficult for charter schools to understand what mechanisms they can utilize to opt out of (or in the case of Florida opt into) the state retirement plan. This complexity and ambiguity often leads to confusion and misinformation. Arizona Charter schools have the option to opt out of the Arizona State Retirement System (ASRS) at conception meaning when the school first opens. According to Arizona law, a charter school that is sponsored by a university, a community college district, a group of community college districts, the state board of education or the state board for charter schools is eligible to participate in the Arizona state retirement system pursuant to title 38, chapter 5, article 2. The 4

5 charter school is a political subdivision of this state for purposes of title 38, chapter 5, article 2. 6 During our survey and interview process, charter schools in Arizona expressed concerns that opting out of ASRS would impact their ability to recruit experienced teachers. However, they also expressed significant concerns about the cost of the plan and their ability to pay teachers a competitive salary. In addition, many schools reiterated that fact that once you opt in, you can t get out. California Under California law the State Teachers Retirement Plan provides a defined benefit to members of the program. The defined benefit is based on final compensation, credited service, and age at retirement, subject to certain variations. The State Teachers Retirement System (STRS) is administered by the Teachers Retirement Board. 7 Charter schools have the option to opt out of the State Teachers Retirement System at conception (and must state this intention in their charter petition). According to California law, a petition for the establishment of a charter school shall identify a single charter school that will operate within the geographic boundaries of that school district. A charter school may propose to operate at multiple sites within the school district if each location is identified in the charter school petition. 8 In addition, this charter petition must include the manner by which staff members of the charter school(s) will be covered by the State Teachers Retirement System, the Public Employees Retirement System, or federal social security. 9 The charter petition is often submitted long before the charter contract is formally signed. During our survey and interview process, California charter schools noted that the structure of the statute forced many charter schools to decide about their participation in STRS before they had all the information or before they had thought through all the consequences of participation. In addition, the fact that the charter petition can govern an entire geographic area often meant that charter schools that were expanding still did not have the ability to withdraw their new schools from STRS. Florida Charter schools must choose to opt in to the Florida Retirement System (thus they are automatically opted out at conception). According to Florida law, a charter school shall ⁱ chapter=2.&article 9 ⁱ chapter=2.&article 5

6 organize as, or be operated by, a nonprofit organization. A charter school may be operated by a municipality or other public entity as provided for by law. As such, the charter school may be either a private or a public employer. As a public employer, a charter school may participate in the Florida Retirement System upon application and approval as a covered group under s (34). If a charter school participates in the Florida Retirement System, the charter school employees shall be compulsory members of the Florida Retirement System. As either a private or a public employer, a charter school may contract for services with an individual or group of individuals who are organized as a partnership or a cooperative. Individuals or groups of individuals who contract their services to the charter school are not public employees. 10 During our survey and interview process, many Florida charter schools seemed unaware of their ability to opt into the Florida Retirement System. In the cases where charters had opted into the Florida Retirement System, teacher demand seemed to be the driving force behind this decision. Louisiana Louisiana law provides charter schools with the option to participate in the Teachers' Retirement System of Louisiana (TRSL), except that it requires Type 4 charters to participate. 11 Charter schools may choose to opt out of the TRSL at conception or upon each renewal. According to Louisiana law, a non-material amendment to a charter is an amendment that makes non-substantive changes to a school's charter. Non-material amendments may include changes in any option expressed in the charter contract exhibits with respect to Teachers' Retirement System of Louisiana. 12 Michigan Michigan law provides that participation in the state retirement system is dependent on employee status. If an employee is employed by a charter school board of directors, he or she must participate in the state retirement system. If an employee is not employed by a charter school board of directors, they are prohibited from participating in state retirement system. Thus, charter schools have access to the ability to opt out by virtue of how they hire their According to the Louisiana Department of Education, there was only one Type 4 charter school in (

7 employees. 13 Michigan s Public School Employees Retirement Act has not been updated since 1979 long before it had any charter schools. 14 Charter Participation Patterns Table 2 reports overall charter participation rates across the five states. The school weighted participation rate for all five states is 47 percent. However, the rate varies considerably by state from a low of 11 percent in Michigan to a high of 87 percent in California. Across the five states, California is clearly an outlier and we will explore one possible explanation for that below. In the previous section, we found that, for the most part, charter schools must make their participation decision at birth. Once they have hired staff and participated in a state plan, it is difficult to withdraw. Since the number of charters opening in some of our states in certain years was small, we have pooled the four low participation states and reported a single year value. We report California separately. In the pooled four-state group the participation rate exhibits a clear downward trend (Figure 3). In California, the participation rate held at roughly 90 percent throughout the period with a sharp drop in recent years. The most likely explanation for these declines are the rising employer costs for these plans. The costs generally rose during the 2000 s, but have increased sharply following the stock market decline in Nearly all pension plans failed to meet their target rates of return (typically 8 percent) in the years following the 2008 stock market decline. This has necessitated sharp increases in employer contribution rates. State teacher plans that seemed like a good deal to a startup charter schools in the 1990 s may be far less attractive today. Most employer contributions are being used to pay down the unfunded liabilities of the plan, and not to pay for benefits for current teachers, making them an unattractive deal for new teachers, and, by extension, new charter schools. 15 What explains California s outlier status? The employer contribution data in Table 1 provide a clue. Of our five states, only California provides a direct payment from the legislature for teacher pensions. Currently, the legislature contributes 8.9 percent of payroll toward teacher pension costs, and charter schools (like districts) contribute 14.4 percent. Until the financial crisis in 2008, teacher and employer contributions were 8.0 and 8.25 percent, respectively. However, if a charter school opts out, it loses the 8.9 percent legislative subsidy. This subsidy of-1980&version=txt 15 In 2015, the following percentages of employer annual required contributions went toward paying down unfunded liabilities: Arizona (83%), California (58%), Florida (42%), Louisiana (82%), Michigan (80%). National Council on Teacher Quality (2015, Figure 7). See also Backes, et. al., (2016). 7

8 seems to keep most charters in, although the employer contribution has risen sharply and is projected to rise further in coming years. To further explore patterns of participation, we estimated a simple logit model of charter school participation. The model estimated is represented below where p is the probability of participating in the state pension plan and X is a set of school or other contextual variables:. ln = + The ratio p / (1 - p) is the odds ratio, and, as the name suggests, indicates the odds that an event, such as opting in to the state teacher pension plan, occurs. In California, with a charter participation rate of 87 percent, the odds that a randomly chosen charter school is a plan participant is 4 to 1. Basically, the logit model allows the variables in X (such as school age, size etc.) to have a proportionate effect on the odds of plan membership. Table 3 reports estimation results for our simple descriptive model of charter participation. Given the large difference in participation rates we estimated separate models for California and the other four states. The four-state results are presented in the left four columns, and the California results are in the right four columns. Each row represents one of the charter or contextual factors. The first column in each block represents the sample mean for that variable. The second column reports the effect of that covariate on the odds ratio. A coefficient greater than one implies that X raises the odds that a school will be in the state plan, whereas a coefficient less than one means that X lowers the odds. Coefficients that are statistically significant are indicated in bold. In these cases, we can reject the hypothesis that the coefficient is one (i.e., the covariate has no effect on the odds of participating in the state plan) in favor of an odds-ratio effect of greater or less than one. Included in the charter variables is a simple binary variable indicating whether the school opened in 2010 or later. The estimate for our four-state sample (0.245) indicates that, controlling for other school characteristics and the school in which the state operates, new charter schools are much less likely to participate in the state plans, as compared to schools opening in earlier years. This result seems reasonable, given the declining financial solvency of many state pension plans. Charter Management Organizations (CMOs) are non-profit firms that operate charter schools. Education Management Organizations (EMOs) are for-profit firms that operate charter schools. Given the fixed costs and expertise required to set up alternative retirement plans, one might expect these management organizations to be more likely to opt out, as compared to 8

9 independent charters. Moreover, EMOs and CMOs can operate in more than one state, and thus would likely favor mobile benefits that facilitate movement of staff from one school to another. State teacher plans impose sharp penalties with teachers or staff move from one state to another (Costrell and Podgursky, 2010). Thus, is it not surprising to find that CMOs and EMOs are considerably less likely to participate in state plans as compared to independent charter schools (the omitted group). EMOs seem particularly averse to participation, with the odds of an EMO participating in the state pension plan being 24 percent of a similarly situated independent charter school. For many charter schools, the benefit of participating in the state retirement plan is that it helps the school recruit experienced teachers from the traditional public school system. This may be a factor in interpreting some of the other coefficients. In the non-california states, combined, middle, and high schools are substantially more likely to participate in state plans as compared to elementary schools. 16 Survey data routinely show that the vast majority of public schools find it relatively easy to fill elementary school vacancies as compared to teachers in most other fields (Podgursky, 2010). It is thus not surprising that elementary charter schools would be more likely to find that the costs of pension participation exceed the benefits, since they find it easier to fill their teaching vacancies. For the four-state sample, charters schools operating in urban and suburban areas are considerably less likely to participate in state plans than rural or town charter schools. One interpretation is that charter schools operating in more populated urban and suburban markets find it easier to fill vacancies with teachers who do not have previous public school experience. They may find it easier to tap alternative sources of teachers like Teach for America (TFA) or other college graduates who are less inclined to spend a full career in teaching. Charter schools operating in more rural or smaller town environments may have a much smaller pool of such candidates, hence they are forced to more aggressively recruit teachers from traditional public schools and offer a retirement benefit package that is comparable to recruit qualified candidates. Survey Results and Charter Interviews To better understand the reasons for why charter schools chose to participate or opt out of state plans, we surveyed a random sample of roughly fifty schools that participated in state plans and fifty that opted out across five states to learn more about the reasons for their decisions. 17 For a subsample of these schools, we followed up with a phone interview as well. 16 Combined is a variable which includes K-8, 6-12, and K-12 schools (or any other less typical grade configuration). 17 We drew a stratified random sample of 400 schools across the five states with identical numbers of participants and non-participants. 73 schools (18 percent) responded. These 73 respondents represented 125 schools. 9

10 The key survey questions regarding participation are reported in Table 4 below. For charter schools that participated in the state plan, the primary reason drivers were recruiting and retaining teachers. However, for charter schools that opted out, the primary drivers were portability of benefits, flexibility of investment options for teachers, and reduced costs. Conclusion As the financial health of public sector defined benefit retirement plans continues to decline, the entities and their employees who participate in them have reason to be concerned about their promised benefits and the cost to provide them. In more than a dozen states, charter schools have been uniquely positioned to have an opportunity to innovate around this important and increasingly expensive component of the compensation package that they offer. This initial analysis of the opt-in/opt-out decision suggests that more charter schools are moving away from the state plans in recent years, that those schools with access to greater administrative capacity and knowledge are more likely to opt out, and that urban and suburban schools are leading the way. It is important to better understand these decisions and the impact, if any, that they are having on charter school operations, finances, and human resources. As a first step, this analysis has shed some light on the charter schools that are taking a more innovative approach. However, there is substantial further research needed. Next steps for this project are to develop an in-depth knowledge base regarding what types of retirement benefit packages charter schools are implementing as alternatives to the state plans and the impact they feel it is having. This knowledge base will be used to better understand charter school compensation package, but will also allow for those charter schools that want to change their status from the state plan to an alternative benefit package to have a source of information as to how to do so. 10

11 References Backes, B., D. Goldhaber, C Grout, C. Koedel, S. Ni, M. Podgursky, P. Xiang, Z Xu (2016). Benefit or Burden? Intergenerational Inequity of Teacher Pension Plans. Educational Researcher, 45 (6), Costrell, R. and M. Podgursky (2009). Peaks, Cliffs, and Valleys: The Peculiar Incentives of Teacher Retirement Systems and their Consequences for School Staffing. Education Finance and Policy, 4 (2), Costrell, R. and M. Podgursky (2010). Distribution of Benefits in Teacher Retirement Systems and Their Implications for Mobility. Education Finance and Policy. Vol. 5 No. 4 (Fall, 2010), Doherty, K, S. Jacobs, M. Lueken (2015). Doing the Math on Teacher Pensions. Washington, DC: National Council on Teacher Quality. Kim, D., C. Koedel, S. Ni, M. Podgursky, W. Wu (2016). Pensions and Late-Career Teacher Retention. Washington DC: CALDER. Olberg, A and M. Podgursky (2011). Charting a New Course to Retirement: How Charter Schools Handle Teacher Pensions. Washington DC: Fordham Institute (June). Podgursky, M. (2010). Teacher Compensation and Collective Bargaining. In R. Hanushek, S. Machin, and L. Woessman (eds.) Handbook of the Economics of Education. Vol. 3 Amsterdam: North-Holland, pp

12 Table 1 Characteristics of State Teacher Pension Plans State AZ CA FL LA MI Plan Type DB DB DB or DC DB Hybrid or DC Vesting Period Immediate (0 Yrs) 5 Yrs 8 Yrs 5 Yrs Hybrid: 10 DC: 0 Teacher Contribution 11.3% 9.2% 3% 8% Hybrid: 3-6.4% DC: 2%+ Employer Contribution 11.3% 14.4% 6.2% 25.5% 25.8% Social Security Participation Yes No Yes No Yes Source: National Council on Teacher Quality, Lifting the Pension Fog, February 2017 ( Notes: Plan characteristics and contribution Rates refer to new teacher hires. Arizona teachers who retire at 65 or later have no minimum service years. New teachers in Florida and Michigan can choose a defined contribution (DC) alternative to the traditional retirement plan. 12

13 Table 2 Number of Charter Schools and State Plan Participation Rates State Year Schools State Plan Participation Arizona % California , % Florida % Louisiana % Michigan % Total 3, % Source: NAPCS Charter School Database. 13

14 Table 3 Sample Means and Logit Estimates All States Except California California mean odds ratio std error p-value mean odds ratio std error p-value New CMO EMO Combined Middle High City Suburb AZ MI LA Sample Size Notes: The omitted characteristic for CMO/EMO is an independent charter school, for combined/middle/high it is an elementary charter school, for city/suburb it is a rural charter school, and the omitted state is Florida. New is a binary variable indicating that the charter school opened in 2010 or later. 14

15 Table 4 Reasons Why Charter Schools Opted in or Out of State Teacher Retirement Plans A. Charter Schools That Participate Please indicate how important each of the following factors was in influencing your school's decision to remain in the state retirement plan for teacher retirement benefits: Very Unimportant Somewhat Unimportant Neither Somewhat Important Very Important Advantages in recruiting teachers 5.7% 3.8% 9.4% 5.7% 75.5% 53 Advantages in retaining teachers 5.7% 1.9% 11.3% 11.3% 69.8% 53 Policies of my school's management organization (if applicable) 14.0% 0.0% 34.0% 8.0% 44.0% 50 Participation in the state plan is legally required for my school 20.4% 0.0% 40.8% 2.0% 36.7% 49 Input regarding retirement plans from staff and/or board members 4.0% 0.0% 32.0% 28.0% 36.0% 50 Lower estimated costs 15.7% 5.9% 41.2% 11.8% 25.5% 51 Additional Social Security requirements for schools opting out of state plans 13.5% 9.6% 57.7% 1.9% 17.3% 52 Lack of resources and/or required expertise to implement an alternative plan 11.5% 3.8% 57.7% 9.6% 17.3% 52 N B. Charter Schools that Opt Out Please indicate how important each of the following factors was in influencing your school's decision to opt out of the state retirement plan for teacher retirement benefits: Very Unimportant Somewhat Unimportant Neither Somewhat Important Very Important Increased portability of retirement funds for teachers 9.6% 1.9% 15.4% 28.8% 44.2% 52 Wider range of investment options for teachers 9.8% 0.0% 15.7% 31.4% 43.1% 51 Lower estimated costs 3.8% 0.0% 11.5% 44.2% 40.4% 52 Additional control over total compensation for teachers 3.8% 5.8% 30.8% 30.8% 28.8% 52 Policies of my school's management organization (if applicable) 12.0% 0.0% 44.0% 18.0% 26.0% 50 Input regarding retirement plans from staff and/or board members 9.6% 1.9% 46.2% 21.2% 21.2% 52 My school is legally restricted from participation in the state plan 32.7% 1.9% 48.1% 1.9% 15.4% 52 Advantages in recruiting teachers 11.5% 9.6% 40.4% 25.0% 13.5% 52 Advantages in retaining teachers 11.5% 1.9% 32.7% 44.2% 9.6% 52 N Source: NAPCS Charter Pension Survey. 15

16 Figure 1 Public Education Contributions Per Pupil for Retirement Benefits Source: University of Arkansas, Department of Education Reform ( 16

17 Figure 2 Charter Schools Participating in State Teacher Retirement Plan 100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% 87.2% 35.8% 27.9% 11.2% 13.4% MI FL AZ LA CA Source: NAPCS Charter School Database. 17

18 Figure 3 Trends in Charter Participation in State Plans by Year School Opened 100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% not CA CA Source: NAPCS Charter School Database. 18

19 Appendix Participation Options by State State Alaska Arizona Arkansas California Colorado Connecticut Delaware Florida Georgia Hawaii Idaho Illinois Indiana Iowa Kansas Louisiana Maine Maryland Massachusetts Michigan Minnesota Mississippi Missouri Nevada New Hampshire New Jersey New Mexico New York North Carolina Ohio Oklahoma Oregon Pennsylvania Rhode Island Must Participate Have Some Mechanism to Opt Out 19

20 South Carolina Tennessee Texas Utah Virginia Washington Wisconsin Wyoming Total Source: NAPCS Model Law Rankings ( 20