To justify studying Golden Albatross (i.e., stay-or-go) pensionable job decisions for my master’s program, I made an argument. I’m not talking about a Facebook or Twitter argument where everyone types in CAPITAL LETTERS and no one changes their mind. I’m talking about an academic argument. That’s right, I moved beyond my typical ranting via the interwebs and masqueraded as a social scientist for a few months. Let me tell ya, white lab coats for hairy knuckle draggers are hard to come by!

My thesis argued that human resources (HR) managers needed to know which pension design elements made their pensionable workers most likely to stay. Reasons they might need to know this included if pension plan re-design was required after a fiscal crisis — like the dot-com crash in the early 2000s. Since the main reason for offering a pension is to create worker retention, I reasoned that pensionable employers would want to avoid cutting design elements that most attracted workers towards staying.

Of course, the argument was hypothetical. I have neither the ear of HR managers anywhere nor the nerve to advocate cutting design elements from pensions. I simply made the argument to convince my advisor and those (un)lucky enough to grade my thesis. However, after collecting and analyzing the results from my pension survey, I was ready to declare ‘Don’t mess with healthcare!’ to any HR manager that would listen.

If you’ve read Pension Series Part 27, then you know why. Survey participants ranked ‘pension subsidized healthcare’ as the design element that made them consider staying at their job the most during their Golden Albatross decision. In fact, the final weighted score for healthcare was ten percentage points higher than the second-place design element, ‘immediate annuity.’ Therefore, the results appeared to support a ‘keep your money grubbing hands off of healthcare’ declaration.

That said, I’m glad I didn’t declare this straight away. As you’re about to find out, age, tenure, and gender are far more powerful elements during a Golden Albatross decision than any singular pension design element — even one as popular as healthcare.

What This Article Is

This article concludes the mini-series of Pension Series posts which showcase the findings of my master’s thesis. As such, it highlights results from the statistical analysis of the pension survey discussed in Part 27 while avoiding an extensive explanation of statistics (you can thank me later). As the culminating article in the mini-series, this post won’t make much sense if you haven’t read Parts 25 through 27 of the Pension Series. I say that because many of the pension design elements and terms I reference below are explained and examined in detail throughout those three other posts.

Findings: Overview

This section is for the ‘too long didn’t read’ (TLDR) crowd. It provides the highlights without any of the specifics and only a little context.

My statistical analysis showed that a survey participant’s age and tenure correlated significantly to their pension’s proportional size in their reasoning for staying during their Golden Albatross decision. The same cannot be said for any singularly ranked pension design element discussed in Pension Series Part 27, not even healthcare. In some cases, current age and gender were also significantly linked to participants’ Golden Albatross decision outcome (i.e., stayed, departed, or undecided). Again, the same cannot be said for any singularly ranked pension design elements from Part 27.

In other words, when a participant’s Golden Albatross decision occurred in terms of age and tenure, and in some cases, their gender mattered more than even the highest-ranked pension design element, which was healthcare.

That said, I keep repeating the word ‘singularly‘ on purpose! The statistical analysis techniques I used couldn’t test if a combination of pension design elements significantly impacted participants’ Golden Albatross decision cycle. I explain what that may mean later in this article. I also discuss how the age and tenure results match previous pension studies.

Conversely, the gender findings provide potentially new academic insight since pre-existing academic literature says little about the gender differences in retention rates for pensionable jobs. I also discuss this finding further later in the article.

Survey Structure

The survey was anonymous and had three main parts. The first part included mandatory demographic questions and ended with two discriminator questions. I cataloged participants’ demographics in Pension Series Part 27, so I won’t detail them here. Generally speaking, though, most participants were between 35 and 54 years old, female, white, married, well-educated, and current pensionable employees. As discussed in Part 27, the less you as a reader resemble these demographics, the less applicable these findings are to you.

The first discriminator question asked whether or not participants had ever seriously contemplated voluntarily departing their pensionable job before normal retirement age. In other words, I wanted to know if they had made a stay-or-go decision at a pensionable job. This decision could’ve taken place at any point throughout their working career, not just the recent past. Only participants who had made a voluntary turnover decision (regardless of outcome) were allowed to continue the survey.

The second discriminator question immediately followed. It asked whether or not participants had considered their pension as a factor during their voluntary departure decision. This was the real crux of the matter because it indicated that respondents had made a Golden Albatross decision (i.e., a stay-or-go decision based, in part, on their pension). Therefore, only the participants who had considered their pension were allowed to continue. In all, 144 of 306 participants screened through these two questions.

The second part of the survey asked the 144 remaining participants if they could identify the design elements in their pension. If they could, they then ranked the top three that made them consider staying at their job the most during their Golden Albatross decision. Again, this was regardless of the outcome. 130 of the 144 respondents ranked design elements. I detail and discuss those results in Pension Series Part 27. However, the results are also depicted in the graphic below:

The final section of the survey reunited all 144 respondents and asked a series of four questions. Two demographic questions asked participants their age and tenure at the time of their Golden Albatross decision. Another question asked about the outcome of the decision (stayed, departed, or undecided). The final question asked participants to judge their pension’s proportional role in their reasons for staying (again, regardless of outcome). It started with a ‘0% to 20%’ category and moved all the way to an ‘81% to 100%’ category in 20% chunks. These four questions were mandatory, and their answers are displayed in the figure below.

Statistical Analysis: Format, Definitions, and Caveats

I thought it best to simply highlight the significant findings in bullet format for the next section. Where necessary, I provide explanations of the significant factors involved in each result. I also explain how the findings conform to or diverge from the pre-existing pension studies within academic literature. Finally, I also placed some clearly labeled informed opinions to provide context for the results.

Two notes about vocabulary from someone who only learned about statistical techniques for his thesis. First, the words ‘significant’ or ‘significantly’ have a specific meaning. In this case, it means less than a 5% chance that the results I’m discussing occurred at random. In fact, in some cases, the calculations showed less than a 1% chance. Second, I use the term ‘population’ to describe participants who had access to a specific design element. In other words, 87 of 130 participants had access to the ‘pension subsidized healthcare’ design element. Therefore, the healthcare design element had a population of 87 for this survey. You can dig into all these details and more in the copy of my thesis that I posted in Pension Series Part 27.

Statistical Analysis Results

• Discriminator Question 1:
  • Finding: Only 50% (153) of the total (306) pensionable respondents ever seriously considered voluntarily departing their pensionable job before normal retirement age (NRA).
    • Informed opinion based on academic literature: The 50% to have never seriously considered voluntarily leaving their pensionable job throughout the entirety of their career seemed large. The explanation may be that I defined ‘seriously considered’ in the survey as taking action like talking to mentors, researching new jobs, or applying for them. This means participants had to do more than just daydream about leaving.
    • Academic literature: That said, the finding is in line with scholarly literature. Ippolito (1991, 1994, 2002), Lazer and Moore (1988), and Allen et al. (1993) documented the retention effect DBPs create with numbers ranging from 20% to 50% compared to non-pensionable counterparts. Even the more recent academic studies, like Haverstick et al. (2010) and Schuck and Rabe-Hemp (2018), found significant retention effects created by DBPs.
      • Numerous theories exist about how and why DBPs lead to less employee turnover, including employee self-selection, job satisfaction theory, psychological contract theory, and Transaction Cost Economics (de Thierry et al., 2014; Ippolito, 1991; Joo, 2017; Luchak & Gellatly, 2002).
      • As explained in Pension Series Part 26, I used continuance commitment and pension quit cost theory as the framework for this study.
  • Finding(s): Current age and gender significantly impacted the outcome of discriminator question one.
    • Respondents whose current age was 44 and younger considered voluntary departing significantly more than 55 and older respondents, who considered it significantly less.
      • Informed opinion based on academic literature: Age-based findings are not surprising as a large body of pension-focused academic work catalogs the effect. Younger workers in pensionable jobs are more likely to leave voluntarily than older workers, and vice versa (Haverstick et al., 2010; Kirkman, 2017; Llorens, 2015).
    • Females considered voluntary turnover significantly more, and males considered it significantly less.
      • Informed opinion based on academic literature: Gender difference was not an expected finding. Pension-connected voluntary turnover literature says little about gender (Wynen & Op de Beeck, 2014). The general job quit literature may explain gender’s results the best since it often points to the burden of familial obligations interfering with women’s careers (Wynen & Op de Beek, 2014).
      • Informed opinion based on academic literature: Alternatively, these results might reflect the current stress caused by COVID-19, which prompted many primary school teachers to quit or retire (Diliberti et al., 2021). Teaching is a pensionable profession traditionally dominated by women in the US (NCES, 2020). Many of the survey participants came from Facebook groups dedicated to teacher personal finance.

• Discriminator Question 2:  
  • Finding: Of the 153 remaining respondents, only 144 made a Golden Albatross decision by considering their pension during their voluntary departure decision.
  • Finding: Most of the 144 respondents made their Golden Albatross decision between 5 and 14 years of tenure and between 25 and 44 years of age.
  • Finding(s): 96 of the 144 decided to stay at their pensionable job.
    • Current age correlated significantly to the stay-or-go outcome.
    • 34 and younger respondents departed or were undecided significantly more than their older counterparts.
  • Finding(s): 86 of the 144 signaled that their pension made up between 61% and 100% of the reason(s) why they considered staying.
    • Age and tenure at decision time correlated significantly with the importance a respondent placed on the pension.
    • Older (45 and up) and more tenured (15 years or more) respondents placed significantly more emphasis on their DBP than younger, less tenured workers.
  • Academic literature: These significant age and tenure findings are in line with the broader scholarly literature. Haverstick et al. (2010) specifically found that tenure gradually moves workers’ turnover decisions towards staying over time. Many others have also noted tenure’s effect (Asch, 2019; de Thierry et al., 2014; McCarthy et al., 2020). Ippolito (1991) captured the retention effect for older pensionable workers, which many others have also documented (de Thierry et al., 2014; Kirkman, 2017; Llorens, 2015).

• Pension design element selection:
  • Finding: A survey participant’s age at their Golden Albatross decision point significantly correlated to selecting the ‘pension subsidized healthcare’ design element.
    • 34 and younger respondents selected healthcare significantly less than older respondents.
    • Informed opinion based on academic literature: This result may reflect demographic trends among Americans. For example, the implementation of the Affordable Care Act in the US demonstrated that younger, healthier Americans often deemphasize healthcare insurance since they have less need for it (Deloitte LLP, 2014).
  • Finding: Gender and age at Golden Albatross decision time correlated significantly to the choice of ‘backloaded annuity.’
    • A significant percentage of men and 34 or younger respondents did not vote for a backloaded annuity. In contrast, women and 35 and older participants did.
    • Academic literature: Older workers’ emphasis on backloaded annuities is well-documented in pension literature because older workers are much closer to retirement than younger workers (Luchak & Gellatly, 2001; Yang, 2005). The gender-based results are unsupported by current academic literature on pensions and should be researched further.

• Pension design element ranking:
  • Finding: None of the pension design element ranking results correlated significantly to their population’s turnover decision outcomes (stayed, departed, undecided).
  • Finding: None of the pension design element ranking results correlated significantly to the proportional role (0% – 100%) a pension played in their population’s reasons for staying during a Golden Albatross decision.
  • Informed opinion based on academic literature: These findings are both surprising and unsurprising. They are surprising because healthcare’s dominant survey ranking result made it seem like a relationship between it and either question was possible. Conversely, the findings are unsurprising given the lack of previous studies examining the impact of individual pension design elements on employee retention.
  • Informed opinion based on academic literature: In other words, this study attempted to break new ground. As a result, and as cataloged in Pension Series Parts 26 and 27, I combined continuance commitment theory and pension quit cost theory to identify pension design elements to study and form hypotheses for my analysis. I assess these two theories’ usefulness in a follow-on section.

• Pension design element populations:
  • Finding(s): Statistically significant relationships were found in four design element populations between a participant’s age and tenure and the proportional role a pension played in their reasons for staying.
    • Those design element populations included the ‘immediate annuity,’ ‘low-risk income,’ ‘non-portability,’ and ‘backloaded annuity.’
    • Generally speaking, the older (35 and above) and more tenured (15 years or more) participants within these populations placed significantly more emphasis on their DBP as a reason for staying (61% or above).
    • Younger (34 and below) and less tenured workers (14 years or less) did the opposite.
    • Informed opinion based on academic literature: These age and tenure findings align with the broader scholarly literature. As discussed above, age and tenure correlations were found across all 144 respondents who answered this question. Therefore, seeing the correlations replicated within specific design elements’ populations is not surprising. I have no opinion as to why only these four design elements’ populations demonstrated these relationships.
  • Finding(s): Statistically significant correlations between a participant’s gender and their voluntary turnover decision outcome were found within the ‘pension subsidized healthcare’ and ‘cost of living adjustments (COLAs)’ populations.
    • In both populations, significantly more males stayed after making their turnover decision, while considerably more women departed or were undecided.
    • Informed opinion based on academic literature: As stated previously, gender correlations to turnover decision outcomes were not expected because pension-connected turnover literature says little about gender. However, these correlations mirror the gender findings from the larger survey population (306) who answered discriminator question one. In that case, men seriously contemplated a departure from their pensionable job significantly less than women. I have no opinion as to why the gender correlations were only found within the populations of healthcare and COLA design elements.

Age, Tenure, and Gender vs. Ranked Design Element Results

To repeat, statistical analysis revealed that when (in terms of age and tenure) a survey participant experienced their Golden Albatross moment, and in some cases, their gender, was far more significant than any singular pension design element’s impact on their overall Golden Albatross decision process. In general, the older and more tenured the participant, the more significantly they considered their pension as a reason to stay. The statistical analysis also showed that age and gender correlated to participants’ Golden Albatross decision outcomes. In general, younger and female participants considered departing significantly more, and they either departed or were undecided significantly more than their male and older counterparts.

What does this mean for the ranked results? They shouldn’t be dismissed. Ranked questions are a legitimate method of determining what respondents value because their results reveal preference (Jacoby, 2011; Stonebraker, 1981). Thus, the ranked results show that participants valued pension design elements like ‘pension subsidized healthcare,’ ‘immediate annuities,’ and ‘low-risk retirement income’ the most. As such, the ranked results provide a new level of fidelity about which pension design elements participants considered important enough to incorporate into their Golden Albatross decision-making process.

That said, it’s clear that no singular design element, not even healthcare, played a more substantial role than age, tenure, or gender in participants’ Golden Albatross decision. Therefore, my hypothetical ‘don’t mess with healthcare‘ advice is more of a suggestion and less of a mantra for HR managers. I believe organizations that reduce or cut their pension subsidized healthcare program in a time of fiscal crisis would have a lot of unhappy workers on their hands. However, I’m not convinced there would be mass resignations. The same goes for the other top-ranked pension design elements on the list from Pension Series Part 27.

Results Do Not Preclude Impacts From Multiple Design Elements

As a master’s student with limited time and resources, I had to limit the scope of work for my thesis. Studying the effects of distinct pension design elements and demographics on the Golden Albatross decision cycle was that limit. In other words, my data collection and analysis methods were not designed to study the impact of more than one pension design element at a time on the Golden Albatross decision process.

In hindsight, that decision seems somewhat short-sighted. Pension-providing organizations almost always bundle various pension design elements into defined benefits packages to meet their needs and goals (and hopefully their pensionable workers). Those goals are typically retention-related. Thus, the impact of pension plan design on retention is probably best studied in a manner that can examine the combined effect of multiple pension design elements on retention outcomes.

Aside from the complexity issues, I genuinely thought a few pension design elements, like healthcare and immediate annuities, would produce singularly significant statistical results. That thought was reinforced by all the literature I read about continuance commitment, as discussed in Pension Series Part 26. Honestly, I was shocked that healthcare didn’t significantly correlate to participants who decided to stay, especially after seeing its dominant position in the weighted ranking results. BUT, that’s why we have the scientific method. It allows humans to test their beliefs and theories to see if they’re well-founded. In this case, mine were not.

Accuracy and Applicability 

How accurate and applicable are these results?

The results are as accurate as the participants’ memories and understanding of the requirements and definitions that I provided in the survey. Most respondents had to choose which design elements their pension plan featured before ranking them. In some cases, depending on the respondent, they remembered a situation from years ago. Thus, not only did this study require some financial literacy, but it also required accurate recall of the events. This was a drawback of running a questionnaire that polled hundreds of workers from different DBP plans. Whereas I collected a far more diverse set of answers, it required reliance on fallible human memory.

As far as applicability, I did not gather survey participants randomly. Therefore, the results do not apply to pensionable workers everywhere. That’s why I mentioned the participants’ demographics early in this article and in Pension Series Part 27. The less that you as a reader resemble those demographics, the less applicable these findings will be.

A great example of that would be if you’re a Black pensionable worker. Few participated in the survey. Moreover, as I discussed in Pension Series Part 24, Black pensionable workers have a different set of concerns than their white counterparts during their Golden Albatross decision. Thus, my survey results probably don’t address those concerns satisfactorily.

That said, I think the tenure and age-related findings apply to most pensionable workers. I say that because so much available academic research catalogs similar age and tenure findings. The less you resemble the demographics of my survey’s participants, the less the average age and tenure breakpoints in my study may mean to you. However, generally speaking, age and tenure at the time of a person’s Golden Albatross decision are significant factors that contribute substantially to outcomes. I now understand why pension-providing organizations put so many age and tenure-related provisions and penalties into their plans!

This study’s gender-related findings are the real wild-card due to the general lack of other academic studies. What research exists shows mixed results from the standpoint that gender matters/doesn’t matter to pension-related retention (Wynen & Op de Beeck, 2014). At best, my results indicate that participating women probably had a different set of considerations than the men during their Golden Albatross decision. Those considerations probably overlapped in many areas, but were also probably distinct in a few key areas.

That’s a lot of probably! I’d need to dig further into the existing research to try and identify what those key areas of difference might be. Maybe that should be the next Pension Series article? Bottom line, though, if I was a white middle-age female with an advanced degree, I’d sit up and take notice of this survey’s findings because they made up the largest demographic of participants.

Continuance Commitment Theory vs. Results

As I stated in Pension Series Part 27, continuance commitment and pension quit cost theory appeared to explain the results of the design element ranking rather well. Most of the top-ranked design elements are lucrative for employees, creating stay pension value. One, ‘non-portability,’ imposes high quit pension costs. All contribute to the idea that pensionable workers are reluctant to leave their jobs due to the high costs of departure. The only noted aberration was the ‘non-contributory design element’s low ranking.

Nevertheless, the statistical analysis results paint a different picture. Continuance commitment and pension quit cost theory did not explain the lack of significant correlation for design elements. Furthermore, continuance commitment theory doesn’t accommodate the age and tenure relationships identified in my analysis. In fact, neither age nor tenure is considered an essential ingredient in developing continuance commitment within workers (Meyer & Allen, 1991; Meyer et al., 2002). Thus, despite how well continuance commitment and pension quit cost theory explained the weighted ranking results, I found severe limitations in both theories’ ability to explain the statistical analysis results.

However, I wouldn’t write-off continuance commitment or pension quit cost theory. Both theories work well to explain how pensions retain some workers, but only at a pension plan level. In fact, my research showed that continuance commitment existed at the pension plan level for my participants. Remember, two-thirds of the 144 signaled that their pension made up between 61% and 100% of the reason(s) why they considered staying.

That said, much like the above discussion about packaging defined benefits, it’s probably the cumulative effect of all the design elements working together that create continuance commitment. That isn’t just my opinion either. As Pension Series Part 26 notes, several important pension studies captured continuance commitment’s effect at the pension plan level. My findings show that continuance commitment and pension quit cost theory have limits in explaining the relationships between pensions, retention, and voluntary turnover. One of those limits is a bright red line drawn between the pension plan and pension design element level.

Grumpus’ Overall Thoughts

My thesis research shows something I already knew. In fact, I know I knew it because I spent a lot of time in my book talking about it.

Say that three times!

What I knew is that something more complex is going on during people’s Golden Albatross decisions than simply tallying up a pension’s valuable design elements. Among other things, age and tenure correlations mean people are concerned about starting over with another employer later in their career. Those concerns could represent multiple issues in their personal lives.

For example, a pensionable worker may be weighing up what’s best for their family in the present versus what’s best for their retirement in the future. Moreover, the gender-related findings may have little to do with money and retirement savings concerns altogether. They may reflect gender-based familial norms. Alternatively, they could reflect the difficulties of working in an industry (teaching) so heavily impacted by COVID-19.

All of this is a long way of saying that the decision to stay at or leave a pensionable job will always include the personal circumstances of the person making the decision. To be more specific, personal finance is personal! And, the decision to stay at or leave a pensionable job is both a personal and personal finance decision.

One Last Plug for Pension Plan Design

That said, let’s not throw the baby out with the bathwater. Much like I point out in my book, determining pension plan generosity should play a role in the Golden Albatross decision … just not the only role. As pointed out in Pension Series Part 25, pension plan generosity is determined through pension plan design. So, in my mind, pension plan design matters, especially if several valuable pension design elements are stacked on top of each other.

Unfortunately, my research can’t definitively point a reader towards a stack of pension design elements (or lack thereof) that correlate strongest to staying or going during a Golden Albatross moment. The ranked results from Pension Series Part 27 hint at which ones might create the most pull to stay, but lack of statistical support mean they are far from definitive. Thus, no one should consider the ranked results more important than the significant age, tenure, and gender-based correlations highlighted in the statistical analysis.

References

Allen, S. G., Clark, R. L., & McDermed, A. A. (1993). Pensions, bonding, and lifetime jobs. The Journal of Human Resources, 28(3), 463–481. https://doi.org/10.2307/146155

Asch, B. (2019). Setting military compensation to support recruitment, retention, and performance. RAND Corporation. https://doi.org/10.7249/RR3197

Deloitte LLP. (2014). Young adults and health insurance: Not invincible — but perhaps convincible. www.deloitte.com/us/youngadultenrollment

Diliberti, M. K., Schwartz, H. L., & Grant, D. (2021). Stress topped the reasons why public school teachers quit, even before COVID-19. https://doi.org/10.7249/rra1121-2

Haverstick, K., Munnell, A. H., Sanzenbacher, G., & Soto, M. (2010). Pension type, tenure, and job mobility. Journal of Pension Economics and Finance, 9(4), 609–625. https://doi.org/10.1017/S1474747209990321

Ippolito, R. A. (1991). Encouraging long-term tenure: Wage tilt or pensions? Industrial and Labor Relations Review, 44(3), 520–535. http://www.jstor.com/stable/2524159

Ippolito, R. A. (1994). Pensions and indenture premia. The Journal of Human Resources, 29(3), 795. https://doi.org/10.2307/146253

Ippolito, R. A. (2002). Stayers as ‘workers’ and ‘savers’: Toward reconciling the pension-quit literature. The Journal of Human Resources, 37(2), 275–308. http://www.jstor.com/stable/3069648

Jacoby, W. G. (2011). Measuring value choices: Are rank orders valid indicators? http://projects.iq.harvard.edu/cces/files/jacoby_measuring_value_choices-1.pdf

Joo, M. K. (2017). Retirement benefits as a two-way street: Investigating psychological mechanisms of retirement benefits on employee outcomes – a conceptual model. Journal of Organizational Psychology, 17(2), 61–75. https://www.articlegateway.com/index.php/JOP/article/view/1684/1598rg/atoztitles/li

Kirkman, L. (2017). The implications of rising turnover and voluntary turnover within the Federal government [University of Southern Mississippi]. https://aquila.usm.edu/honors_theses/508

Lazear, E. P., & Moore, R. L. (1988). Pensions and turnover. In Z. Bodie, J. B. Shoven, & D. A. Wise (Eds.), Pensions in the U.S. Economy (pp. 163–190). University of Chicago Press. http://www.nber.org/books/bodi88-1

Llorens, J. J. (2015). Fiscally driven compensation reform and threats to human capital capacity in the public sector. International Journal of Organization Theory & Behavior, 18(1), 22–46. https://doi.org/10.1108/IJOTB-18-01-2015-B003

Luchak, A. A., & Gellatly, I. R. (2001). What kind of commitment does a final-earnings pension plan elicit? Relations Industrielles, 56(2), 394–417. https://doi.org/10.7202/000030ar

Luchak, A. A., & Gellatly, I. R. (2002). How pension accrual affects job satisfaction. Journal of Labor Research, 23(1), 145–162. https://doi.org/10.1007/s12122-002-1023-5

Luchak, A. A., Pohler, D. M., & Gellaty, I. R. (2008). When do committed employees retire? The effects of organizational commitment on retirement plans under a defined-benefit pension plan. Human Resource Management, 47(3), 581–599. https://doi.org/10.1002/hrm.20233

McCarthy, I. O., Moonesinghe, R., & Dean, H. D. (2020). Association of employee engagement factors and turnover intention among the 2015 U.S. federal government workforce. SAGE Open, 10(2). https://doi.org/10.1177/2158244020931847

Meyer, J. P., & Allen, N. J. (1991). A three-component conceptualization of organizational commitment. Human Resource Management Review, 1(1), 61–89. https://www-sciencedirect-com.nmit.idm.oclc.org/science/article/abs/pii/105348229190011Z

Meyer, J. P., Stanley, D. J., Herscovitch, L., & Topolnytsky, L. (2002). Affective, continuance, and normative commitment to the organization: A meta-analysis of antecedents, correlates, and consequences. Journal of Vocational Behavior, 61(1), 20–52. https://doi.org/10.1006/jvbe.2001.1842

National Center for Education and Statistics. (2020, May). The condition of education – preprimary, elementary, and secondary education – teachers and staff – characteristics of public school teachers – indicator May (2020). https://nces.ed.gov/programs/coe/indicator_clr.asp

Schuck, A. M., & Rabe-Hemp, C. E. (2018). Investing in people: salary and turnover in policing. Policing: An International Journal, 41(1), 113–128. https://doi.org/10.1108/PIJPSM-09-2016-0137

Stonebraker, P. W. (1981). A multiple-variable evaluation of employee benefit preferences [Arizona State University]. https://search-proquest-com.nmit.idm.oclc.org/docview/303111774/abstract/25D00510ABD048AAPQ/1?accountid=40261

de Thierry, E., Lam, H., Harcourt, M., Flynn, M., & Wood, G. (2014). Defined benefit pension decline: the consequences for organizations and employees. Employee Relations, 36(6), 654–673. https://doi.org/10.1108/ER-02-2013-0020

Wynen, J., & Op de Beeck, S. (2014). The impact of the financial and economic crisis on turnover intention in the U.S. federal government. Public Personnel Management, 43(4), 565–585. https://doi.org/10.1177/0091026014537043

Yang, T. (2005). Understanding the defined benefit versus defined contribution choice. In Pension Research Council Working Paper Pension (No. 2005–4). https://doi.org/10.2139/ssrn.755064