The effect of age discrimination laws on age-earnings profiles of postsecondary faculty.
The 1967 Age Discrimination in Employment Act (ADEA) intended to protect older workers against discrimination in hiring and layoffs. Amendments in 1978 and 1986 expanded the protected age group and finally banned the practice of mandatory retirement altogether. However, ADEA historically excluded several occupations from its protection, most notably faculty in postsecondary institutions. (1) Mandatory retirement for faculty was only abolished in 1994, eight years after the practice was banned for most occupations. This exclusion was based on the premise that this legislation would negatively impact the academic labor market.
Lazear (1979) points out that the ban of mandatory retirement may reduce efficiency by inhibiting long-term incentive contracts between workers and firms ("Lazear contracts"). In his model junior workers earn wages below marginal product, in return for wages above marginal product later in life. This compensation scheme creates incentives to put forth optimal effort, especially when monitoring is difficult or costly. Mandatory retirement is key in enforcing the implicit contract, since older workers earning more than the reservation wage have incentives to remain employed beyond the optimal contract time. This implies that a mandatory retirement ban may be inefficient, since optimal back-loaded implicit contracts can no longer be enforced.
These theoretical implications seem to resonate with higher education administrators, who were concerned that the ban of mandatory retirement would result in older faculty postponing retirement, thus reducing the number of positions for (lower-paid) junior faculty. This could place a burden on university budgets and limit affirmative action programs, since fewer openings would be available for women and minorities. Another concern was that the absence of mandatory retirement could undermine tenure by fueling more frequent performance evaluations and dismissals of tenured faculty. Finally, administrators were concerned that academic productivity could decline with age (Rees and Smith 1991; Hammond and Morgan 1991).
Neumark and Stock (1999) argue that the mandatory retirement ban may not necessarily reduce efficiency. Since ADEA prohibits age-based firings, it may encourage workers to enter Lazear contracts, because they are no longer concerned about the firm opportunistically replacing older workers with younger ones. The possibility of age-based termination makes bonding between the firm and the worker costly, and workers discount the future stream of rents and pensions at a higher rate than the firm. By protecting older workers, ADEA may have reduced bonding costs, thereby encouraging Lazear contracts, and resulted in efficiency gains.
This argument does not directly apply to academia because tenure precludes institutions from terminating existing contracts with older faculty. Tenure, however, does not guarantee specific salary levels or workloads. Therefore, bonding may be costly if there is a risk that the institution will alter workloads (by changing the relative load of research, teaching, and administrative duties over the life cycle), or reduce departmental support, perquisites, and wage growth rates for older faculty. Bonding could also be costly if the institution gains monopsony power over time. The leading explanation for the often-cited negative return to seniority in academia is that it represents evidence of the institutions' monopsony power gained over time as faculty become tied to the geographical area (Ransom 1993; Harnermesh 1988). However, since ADEA provisions extend beyond mandatory retirement and include broader protection for older workers, they prevent institutions from exercising their monopsony power by increasing workloads or reducing wage growth for older faculty. This reduces bonding costs, and may result in both longer employment and steeper wage profiles.
This study evaluates these conflicting views by investigating the effect of ADEA on faculty age-earnings profiles. The analysis employs exogenous variation in state mandatory retirement laws for faculty to identify the causal effect of the policy. The study supplements the evidence with data from nine Midwestern universities that indicate that changes in age-earnings profiles are not driven by productivity shocks. In doing so, this study generates implications for the effect on ADEA on earnings profiles for professions where productivity is not constant over the life-cycle. In addition, the results contribute to the recent literature on the effects of ADEA on the efficiency of the academic labor market.
ADEA and the Academic Labor Market
ADEA was first enacted in 1967. A 1978 amendment, which raised the mandatory retirement age from 65 to 70, did not take effect in academia until 1982. Similarly, a 1986 amendment, which abolished mandatory retirement, allowed institutions of higher education to continue to enforce this practice. In 1991, a Congress-mandated study by the National Research Council (Hammond and Morgan 1991) concluded that the fears of school administrators regarding the adverse effect of ADEA were unfounded. The study observed that most faculty retired before the age of 70, and only 1.6% of faculty in institutions where mandatory retirement was banned by state law were over the age of 70. In private research institutions, however, more faculty retired at 70, suggesting that the change in legislation could have a stronger impact on private universities. Finally, Hammond and Morgan asserted that tenure was unlikely to be undermined by the policy change, since dismissal of tenured faculty is a lengthy and costly process. Similarly, Rees and Smith (1991) found no systematic differences in the retirement age between institutions that had capped the retirement age and institutions that had not. As a result, Congress allowed the exemption for university professors to expire by January 1, 1994.
Recent studies, however, suggest that the impact of the policy change is far from negligible. A survey by the American Association of University Professors' (AAUP) Committee on Retirement (Survey of Changes in Faculty Retirement Policies 2000) including 608 institutions, shows that a greater proportion of faculty remain employed beyond 70, especially in doctoral and private institutions. Clark et al. (2001) find that retirements in three North Carolina universities have declined significantly after ADEA. Similarly, Ashenfelter and Card (2002) find that retirement rates for 70- and 71-year-olds fell to under 30% from the respective 75% and 60% rates prevalent during the mandatory retirement era. The age-discrimination legislation has also affected retirement plans in academia.
Academic pension plans typically take the form of defined-contribution (DC) or defined-benefit (DB) plans. DC plans fix the contribution amount and create incentives to delay retirement because additional employment years increase the amount of (interest-earning) contributions. DB plans fix the amount of the benefit and also create incentives to delay retirement because their value increases with years of service and with the highest salary earned. Traditionally, faculty in public institutions have been covered by DB plans sponsored by state governments, whereas faculty in private institutions have been covered by DC plans. Although under a DC plan the investment risk rests with the individual, this plan provides for portability of retirement benefits upon job changes. In contrast, the DB plan remains with the institution that owns the fund. Lahey et al. (2008) find that in recent years, public universities have increased the availability of DC plans alongside of state-funded DB plans. The AAUP survey finds that 60% of private and 42% of the public doctoral institutions surveyed have introduced financial incentives to encourage faculty to retire prior to age 70 (Ehrenberg 2001). In addition, as of 1999, 50% of private doctoral and 31% of public doctoral institutions made use of phased retirement programs, which allow faculty to work part-time, but retain full benefits. More recently, a survey by Keefe (2001) including 66 colleges and universities, finds that 80% of the institutions offer variants of early retirement plans.
To date, no study has investigated the effect of ADEA on implicit long-term contracting in academia. One possibility is that in the absence of mandatory retirement, age-earnings profiles become flatter, or mirror performance paths more closely, since retirement at a certain age can no longer be enforced. Lazear (1982), however, points out that appropriately structured plans whose present value declines with retirement age, can be used to induce retirement in the absence of mandatory retirement. ADEA may even encourage Lazear contracts by reducing bonding costs. If faculty discount future rents and pension benefits at a higher rate than the institution, to account for the possibility that the institution will alter workloads or exercise monopsony power, they may be less likely to enter long-term back-loaded contracts. By protecting against such actions, ADEA may increase the incidence of long-term contracting.
Increased mobility is another reason why recent changes may encourage Lazear contracts. According to Lazear (1979), long-term contracts carry an immobility cost if pension adjustments or severance pay are not permitted. Unexpected changes in the value of leisure (due, for example, to changing health), or shocks to the value of marginal productivity (due to changes in duties and workloads, or sudden shocks to productivity) may result in inefficiencies. Even if the worker would be better off to retire early or take an alternative offer, the worker does not leave because he is owed payment by the current firm (equal to the underpayment early in life). If the farm was to negotiate a buyout, or a lump-sum payment for the amount owed, the worker could retire early or take an external offer and be better off, without any detrimental effects to the farm. Since 1995, 72% of private doctoral and 38% of public doctoral institutions in the AAUP survey report negotiating buyouts, amounting to a lump-sum payment close to a 9-18 month salary for early retirement (Ehrenberg 2001). Furthermore, upon changing jobs, faculty in public institutions forfeit their benefits if they are covered by a state DB plan. Thus faculty may discount DB pension benefits at a higher rate. The recent move toward DC plans, or a choice between the two, has reduced mobility costs, at least for the younger cohorts. (2) By giving institutions flexibility to respond to positive shocks to the value of leisure or productivity, the post-ADEA changes in retirement plans may have further encouraged Lazear contracts.
Empirical Analysis Using Public Data
The first part of the analysis employs data from the Current Population Survey (CPS) outgoing rotations from January 1980 to September 2008. Using occupation identifiers, I restrict the sample to university teachers, who are older than 24, and hold an advanced degree (masters, PhD, or other professional degree). I exclude unpaid, part-time, retired, and federal workers.
The analysis follows closely Neumark and Stock with few modifications. (3) The identifying information comes from states that banned mandatory retirement for faculty before 1994. Table 1 compiles important changes in state and federal laws regarding faculty. Eight states eliminated mandatory retirement for faculty before 1994: Connecticut, Florida, Hawaii, Maine, Michigan, Texas, Utah, and Wisconsin. States abolishing mandatory retirement only for public employees, include Arizona, Delaware, Georgia, Iowa, Louisiana, Maryland, New Hampshire, North Carolina, Vermont, and Virginia. I assume that earnings in time t for individual i in state j follow the model:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)
where Y denotes earnings, X includes demographics (age, gender, race, marital status), a dummy indicating union coverage, and a dummy indicating employment in a public institution. T represents year dummies, and S represents state dummies. [MR.sub.jt] takes a value of one if state j banned mandatory retirement in time t. [PMR.sub.ijt] denotes an individual who enters the job market after the ban of mandatory retirement. According to the NSF Survey of Doctoral Recipients for the United States, the median age at the time of doctorate is slightly less than 34. Assuming that faculty enter their first long-term contract at this age, [PMR.sub.ijt] takes a value of one when faculty are 34 or less in time t. (4) Therefore, [MR.sub.jt] x [PMR.sub.ijt] equals one for faculty in the youngest age group (24-34) in a state that bans mandatory retirement in time t. For this state, in time t+1, [MR.sub.j, t+1] [PMR.sub.ij], t+1 receives a value of one for faculty age 35 or less, because 34-year-olds entering the job market in time t would be 35 in year t+1. For the states that had no special mandatory retirement provisions, and were first affected by the federal law, the treatment group includes all faculty 34 and younger in 1994, 35 or younger in 1995, and so on. I assume that changes in long-term contracting will be more evident for faculty in the youngest age group, since it is easier to implement contractual changes for new entrants. Therefore, the focus will be on the treatment group that includes faculty under 34 as of the year when mandatory retirement was banned by state or federal law. Since several states banned mandatory retirement in the 1980s, this procedure identifies age-earnings profiles for faculty through age 66. The control group includes faculty 35 and over as of the year when mandatory retirement was banned, and all those employed prior to the change in legislation, regardless of age. This will be referred to as the "unrestricted control group" in the analysis.
The age-earnings profiles of faculty who were mid-contract at the time of the legislation change may adjust differently from the profiles of younger faculty who are employed under the new regime. Since the focus is the younger cohort, a more direct approach is to remove all faculty 35 and older at the time when mandatory retirement was abolished (hereafter, the "restricted control group"). Then the policy effect is obtained by comparing faculty in states that have banned mandatory retirement with same-age faculty in states that have not.
In model (1), [[delta].sub.1] represents changes in age-earnings profiles for faculty who enter contracts without a mandatory retirement clause. The parameter [[delta].sub.2] picks up the policy effect on earnings profiles of older faculty. One problem with estimating [[delta].sub.1] and [[delta].sub.2] is that wages can change both because of shifts in relative demands for older and younger faculty, and because of slope adjustments due to changes in long-term contracting. Assuming that the relative demand shifts equally affect all workers, slope changes can be identified by explicitly including [MR.sub.jt] x [PMR.sub.ijt] and [MR.sub.jt] x (1-[PMR.sub.ijt]) in model (1). The term [[gamma].sub.1]([[gamma].sub.2]) captures the effect of the relative demand shifts on the younger (older) cohorts. State and year dummies control for variation in unobservables common to faculty in a given year or state. I also add age-year and state-year interactions to account for changes in earnings profiles that are unrelated to ADEA.
The change in legislation may have had a different impact on faculty in public institutions compared to faculty in private institutions. Model (2) includes an interaction of mandatory retirement laws with age and an indicator of employment in the public sector ([PUB.sub.ijt]):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)
The coefficient [[phi].sub.1]([[phi].sub.2]) captures the policy effect on demand shifts for faculty in public (private) institutions entering contracts post-legislation. Coefficients [[zeta].sub.1] and [[zeta].sub.2] pick up the policy effect on faculty working in either public or private schools before the change in legislation. The focus in on [[delta].sub.1], which captures changes in age-earnings slopes for newly-hired faculty in public institutions and [[delta].sub.2], which captures the policy effect on earnings slopes for young faculty in private institutions. Coefficients [[theta].sub.1] and [[theta].sub.2] represent changes in age-earnings slopes for cohorts hired before the change in legislation. (5) The sample restriction removes post-legislation information from individuals who are mid-contract when the legislation changes, and who may experience different profile changes from the younger post- or pre-legislation cohorts.
Results Using CPS Data
Models (1) and (2) are estimated via ordinary least squares, and the results are tabulated in Table 2, panels 1 and 2, respectively. Based on estimates in panel 1, it appears that age-earnings profiles have steepened for young cohorts, but flattened for older cohorts. In contrast, Neumark and Stock find that age-earnings profiles have steepened for both older and younger workers, although the slope shift is considerably smaller for the older cohorts. The magnitude of the slope shift is quite large, with estimates ranging from 0.9-1.2 percentage points for the younger cohorts and -0.9 to -1 percentage points for the older cohorts. (6)
The estimated slope shifts do not appear to be due to unobservable shocks in age-earnings profiles, since adding year-age interactions in column (3) does not change the main results. One possibility is that observations within the same year and state may be correlated with each other due to state-level variation in economic conditions, or education funding. Estimates in column (4) add state-year interactions to remove such state-year fixed effects. The slope shift for the younger cohorts increases slightly in magnitude, but the qualitative results remain the same.
Next the analysis restricts the control group by dropping older faculty that have entered contracts in the old regime. After this restriction, the policy effects are obtained by comparing profiles of young faculty in states with no mandatory retirement with profiles of young faculty in states enforcing mandatory retirement. In columns (5)-(8) of Table 2 the estimated slope shift for the younger cohorts is comparable to previous estimates, but appears much larger in specifications that control for state-year fixed effects. This suggests that faculty wages within a state in any given year may be strongly affected by conditions in the state (economy, state budgets, etc.), and can, in turn, have a strong impact on the policy estimates.
Next I focus on differences between private and public institutions. Model (2) estimates are presented in Table 2, panel 2. It appears that young faculty in public institutions experienced a stronger slope shift than faculty in private institutions. These results are amplified after restricting the sample to young cohorts. In contrast, age-earnings profiles for older faculty appear to have flattened more for those employed in private institutions. Since more faculty are staying employed past the age of 70, especially in private institutions, it could be that the rents have been spread out over a longer employment period for these cohorts. In addition, the move from DB plans to DC plans in public schools may have reduced mobility costs, and, therefore, bonding costs, more for young cohorts in public schools versus young cohorts in private schools.
Empirical Analysis Using Internal Data
The previous analysis does not include information on individual careers and productivity. Therefore, changes in age-earnings profiles obtained earlier may be due to unobserved productivity shocks rather than ADEA. This section supplements the analysis with data from economics faculty in nine Midwestern universities (7) for the 1991-92, 1995-96, and 1998-99 academic years. In this sample of about 240 faculty, 17% join after mandatory retirement was eliminated by state or federal law. To quantify publication output, I use pages per author weighted by the commonly-used "impact factor". (8) I also collect information on books, textbooks, articles in collective volumes, book chapters, technical reports, conference proceedings, and working papers. In addition, I collect individual citations from both the Social Sciences Citation Index and Sciences Citation Index databases. Finally, I construct five field dummies to control for field-specific factors influencing publications and labor market conditions. Compensation measures 12-month salary, whereas experience is measured as years since Ph.D. completion. This variable is also decomposed into prior experience in the job market and years of employment with the current department (seniority). I also include indicators for administrative duties, such as chair, director of graduate studies, or head of a research center. An indicator for teaching awards identifies exceptionally good teachers. Similar to Moore et al. (2001), I identify faculty obtaining their doctorate from a leading program. (9)
To investigate the effect of the mandatory retirement ban on experience-earnings profiles for Big Ten faculty, I propose the following model for faculty i in department j and year t:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)
X includes gender, experience, graduation from a top-ten program, and field of research, T includes year dummies, and D includes department dummies. Performance includes publication measures, a proxy for teaching quality, and administrative duties. In model (3), [a.sub.i] represents time-invariant individual heterogeneity.
The variable [MR.sub.jt] x [PMR.sub.ijt] takes a value of one if the faculty member joins the current department after the ban of mandatory retirement (as a new hire or lateral move). However, the law may have had an impact on faculty who were hired before the change in legislation and were mid-contract when the new law was passed. The coefficient [[delta].sub.1] picks up the policy effect on earnings profiles of newly-hired faculty, whereas [[delta].sub.2] picks up the policy effect on earnings profiles of faculty employed before the change in law. I also estimate a similar model with experience decomposed into prior job market experience and seniority. Arguably, adjustments in seniority-earnings profiles may be more relevant to the analysis of long-term contracting.
Results Using Big Ten data
The estimates in Table 3 support the previous findings that experience-earnings profiles for newly-hired faculty have steepened after ADEA. The slope change in column (1) is about 1.4 percentage points, remarkably close to the estimates obtained with the CPS data. One concern is that unobserved individual heterogeneity may bias the estimated policy effect. Perhaps new hires differ in their unobserved ability, degree of specialization, or mobility patterns. Therefore, I also obtain time- and individual-fixed effects estimates. While this method does not identify the effect of experience, it recovers the change in the experience premium before and after the change in legislation. Column (2) estimates suggest that the slope shift is even larger for newly-hired faculty (about 2.3 points). This estimate of the policy effect comes very close to the state-year fixed effects estimates obtained with CPS data. Since Lazear contracts involve long-term contracting between the same firm and the worker, the effect of the mandatory retirement ban may be more evident in seniority-earnings profiles. Results in column (3) estimate a slope shift of about 2.3 points in seniority-earnings profiles for newly-hired faculty.
This study analyzes the effect of ADEA on earnings profiles of postsecondary faculty. Using CPS synthetic cohorts from 1980-2008 and data from economics faculty in the Big Ten from 1991-1999, this study finds that the ban of mandatory retirement has resulted in stronger job attachment and steeper wage profiles for faculty hired after the policy change.
One explanation for the steeper earnings profiles is that the change in legislation has reduced bonding costs between faculty and institutions. Faculty may discount the future stream of payments at a higher rate than the institution if the latter can change workloads, administrative support, and wage growth rates, or gain and exercise monopsony power. However, ADEA protects against any such form of discrimination. Therefore, the policy may have reduced bonding costs, resulting in a more permanent job attachment.
Interestingly, wage profiles of older cohorts have become flatter, especially in private institutions. Since older faculty entered contracts stipulating retirement at 70, and a larger proportion of them are staying employed beyond 70 (especially in private schools), it could be that institutions are spreading the rents over a longer period of time. In fact, reducing wage gains may have been the only way for institutions to indirectly reduce pension benefits and deal with unanticipated budget shortfalls immediately after ADEA. (10) For the young cohorts, institutions had the time to adopt a proactive position and design compensation schemes including non-coercive retirement incentives. Overall, the impact of state and federal age legislation seems to have intensified Lazear contracts for new entrants, suggesting that the policy may have improved the efficiency of the academic labor market.
Acknowledgements I thank David Neumark, Charles Ballard, Jeffrey Wooldridge, and Jeff Biddle for the critical role they played in the advancement of this work. I also thank seminar participants at the Department of Economics, Michigan State University for their valuable comments. All errors are mine.
Published online: 10 December 2009
Ashenfelter, O., & Card, D. (2002). Did the elimination of mandatory retirement affect faculty retirement? American Economic Review, 92(4), 957-80.
Clark, R., & Pitts, M. (1999). Faculty choice of a pension plan: defined benefit versus defined contribution. Industrial Relations, 38(1), 18-45.
Clark, R., Ghent, L., & Kreps, J. (2001). Faculty retirement at three North Carolina universities. In R. L. Clark & B. P. Hammond (Eds.), To retire or not? Retirement policy and practice in higher education. Philadelphia: University of Philadelphia Press.
Conroy, M., & Dusansky, R. (1995). The productivity of economics departments in the U.S.: publications in the core journals. Review of 1995. Journal of Economic Literature, 33(4), 1966-71.
Dusansky, R., & Vernon, C. (1998). Rankings of U.S. Economics Departments. Review of 1998. Journal of Economic Perspectives, 12(1), 157-70.
Ehrenberg, R. (2001). Career's end: a survey of faculty retirement policies. Academe, 87(4), 24-29.
Graves, P., Marchand, J., & Thompson, R. (1982). Economics departmental rankings: research incentives, constraints, and efficiency. American Economic Review, 72(5), 1131-41.
Hamermesh, D. (1988). Salaries: disciplinary differences and rank injustices. Academe, 74(3), 20-24.
Hammond, P., & Morgan, H. (eds). (1991). Ending mandatory retirement for tenured faculty. Washington DC: National Academy Press.
Keefe, J. (2001). Survey of early retirement practices in higher education. In R. L. Clark & B. P. Hammond (Eds.), To retire or not? Retirement policy and practice in higher education. Philadelphia: University of Philadelphia Press.
Lahey, K., Michelson, S., Chieffe, N., & Bajtelsmit, V. (2008). Retirement plans for college faculty at public institutions. Financial Services Review, 17, 323-41.
Lazear, E. (1979). Why is there mandatory retirement? Journal of Political Economy, 87(6), 1261-84.
Lazear, E. (1982). Severance pay, pensions, and efficient mobility. NBER Working Paper No. 854.
Moore, W., Newman, R., & Turnbull, G. (2001). Reputational capital and academic pay. Economic Inquiry, 39(4), 663-71.
Neumark, D., & Stock, W. (1999). Age discrimination laws and labor market efficiency. Journal of Political Economy, 107(5), 1081-125.
Northrup, J. (1978). Old age, handicapped, and Vietnam-Era antidiscrimination legislation. Philadelphia: University of Pennsylvania.
Raish, D. (2001). Age-based retirement incentives for tenured faculty members: satisfying legal requirements. In R. L. Clark & B. P. Hammond (Eds.), To retire or not? Retirement policy and practice in higher education. Philadelphia: University of Philadelphia Press.
Ransom, M. (1993). Seniority and monopsony in the academic labor market. American Economic Review, 83(1), 221-33.
Rees, A., & Smith, S. (1991). Faculty retirement in the arts and sciences. Princeton: Princeton University Press.
Scott, L., & Mitias, P. (1996). Trends in rankings of economics departments in the U.S.: an update. Review of 1996. Economic Inquiry, 34(2), 378-400.
E. Pema ([email])
Graduate School of Business and Public Policy, Naval Postgraduate School, 555 Dyer Road, Monterey, CA 93943, USA
(1) Mandatory retirement is still a common clause in employment contracts for some public service employees (police, firefighters, judges, and the military).
(2) Using data from the North Carolina State University from 1971-1994, Clark and Pitts (1999) find that, given the choice, older new hires are more likely to enroll in a DB plan, whereas younger new hires are more likely to enroll in a DC plan to insure against mobility risk.
(3) Neumark and Stock use decennial Censuses.
(4) Because not all faculty are required to have a Ph.D., the age at the time of first hire may be less than 34. The analysis was replicated using age 31, 32, or 33 as the cutoff point, but the results were not sensitive to this change. Neumark and Stock similarly define the control and treatment groups based on the assumption that workers enter the labor market at 24.
(5) Coefficients [[gamma].sub.2] and [[delta].sub.2] in model (1) and [[zeta].sub.1], [[zeta].sub.2], [[theta].sub.1], and [[theta].sub.2] in model (2) are not identified when removing from the sample faculty who are mid-contract when the change in legislation occurs. In this setup, years after 1994 do not provide any identifying information. However, they provide longer profiles for faculty that start working after the legislation change.
(6) Full results are omitted for brevity. In regressions, the effect of age varies from 1.4-2.6%, females earn 16% less, married faculty earn 4-5% more, union members earn 3% higher wages, and faculty with Ph.D.-s earn about 23% more than faculty with other advanced degrees.
(7) Indiana University, University of Illinois, University of Iowa, Michigan State University, The Ohio State University, Purdue University, University of Michigan, University of Minnesota, and University of Wisconsin.
(8) The impact factor measures the frequency with which the average article in a journal is cited. I use 1999 values of the impact factor obtained from Journal Citation Reports.
(9) The top ten programs were defined as those ranking consistently high in studies by Graves et al. (1982), Conroy and Dusansky (1995), Scott and Mitias (1996), Dusansky and Vernon (1998), and the National Research Council. The list consists of: the University of Chicago, Harvard University, Massachusetts Institute of Technology, Northwestern University, University of Pennsylvania, Princeton University, Stanford University, Yale University, University of California, and University of Wisconsin.
(10) Early retirement incentives did not take off until after 1998 because ADEA prohibited retirement incentives that reduced pension benefits based on age. In 1998, the Higher Education Amendments allowed institutions to offer retirement incentives that increased pension benefits during a window of time in return for early retirement, but did not reduce regular pension benefits outside this window. (Raish 2001).
Table 1 Important changes in federal and state laws regarding mandatory retirement of tenured faculty State Year Legislation US 1967 Age Discrimination in Employment Act (ADEA). Prohibits discrimination based on age for those 40-65. 1978 The protected age is extended to 70 years. Tenured faculty are exempt from the protection of these amendments. 1982 The ADEA amendments of 1978 take effect for faculty. 1986 Protection is extended to anyone over 40, thus banning mandatory retirement. Tenured faculty are exempt from this protection. 1994 The exemption from mandatory retirement protection for tenured faculty is allowed to expire. Arizona 1988 Removes requirement that members of the State retirement system retire at the age of 70. 1978 Raises mandatory retirement age to 70 for State employees, and allows employment after age 70 with employer approval. Arkansas 1989 Extends protection under age discrimination laws to State and local employees over 40. Permits mandatory retirement for executive and high- policymaking employees and, until January 1, 1994, for tenured employees in postsecondary institutions. 1979 Prohibits age discrimination in State and local public employment against persons age 40-70, with certain exceptions. Employees over age 70 may continue in employment with annual written authorization from their employers. California 1983 Abolishes mandatory retirement for State and local public employees except those in public safety jobs. Able employees may work beyond mandatory retirement date upon written request. Also mandatory retirement permitted for tenured faculty. 1979 Permits mandatory retirement of tenured faculty after age 65 prior to July 1, 1982, and age 70 thereafter, and mandatory retirement for persons in executive and high policy-making positions at age 65. 1978 Allows private and public sector employees to work beyond the normal retirement date if the employee so requests in writing and demonstrates continued adequate job performance. Colorado 1979 Raises mandatory retirement age to 70 for State employees but allows employment beyond 70 with employer approval. Connecticut 1989 Prohibits mandatory retirement of tenured employees at independent institutions of higher education, effective July 1, 1993. Delaware 1986 Allows state employees to work after mandatory retirement age of 70 on a year-to-year basis upon written approval. D.C. 1988 Removes mandatory retirement requirements (at the age of 70) for public school teachers. Georgia 1984 Abolishes mandatory retirement under the State employee's retirement system except for specified public safety personnel. Florida 1976 Enacts age discrimination law abolishing mandatory retirement for State and local government employees at age 65. Hawaii 1984 Bans mandatory retirement in both the public and private sectors. Illinois 1987 Removes the upper age limit for coverage of the prohibition against age discrimination in employment. Allows mandatory retirement for the employment groups specified in the Federal ADEA. Indiana 1979 Raises upper limit for State protection from 65 to 70. Mandatory retirement age for teachers is raised from 66 to 71 with provision for work beyond that age with a doctor's certificate. Iowa 1979 Bans mandatory retirement for State employees. Other public sector employees may work beyond age 70 with employer approval, except for police and firefighters, who must retire at age 65. Kansas 1988 Amends AREA to extend coverage to all persons over 18 (rather than those between 40 and 70). Allows mandatory retirement for high-policymaking employees and until January 1, 1994 for tenured employees at institutions of higher education. 1977 Abolishes mandatory retirement of State employees through an Executive Order, effective November 23, 1977. Louisiana 1987 Repeals mandatory retirement provisions at age 70 for public employees. Mandatory retirement of law enforcement personnel and firefighters continues to be required at age 65. Maine 1979 Prohibits mandatory retirement at any age in private employment (formerly applying only to public employment) from 1/1/80. Maryland 1988 Removes mandatory retirement at the age of 70 for those under the State employees and teachers retirement systems. Michigan 1991 Amends the Civil Rights Act to prohibit mandatory retirement for employees of institutions of higher education who are serving under unlimited tenure. Minnesota 1987 Eliminates mandatory retirement at age 70 for State employees in the executive branch under the State retirement system or the teachers' retirement system except for tenured faculty at institutions of higher education. 1979 Changes the date for prohibition of mandatory retirement from 6/1/1980 to 4/ 24/1979, for most private and public employees. Mississippi 1980 Raises mandatory retirement age for public employees to 70 applicable to all, except elected officials and Governor appointees. Nebraska 1987 Removes mandatory retirement requirements at the age of 70 for employees of public institutions of higher education, with the exception of tenured faculty members or law enforcement personnel. 1982 Allows county, state, and school system employees to work beyond mandatory retirement age with employer approval. 1979 Raises upper age limit for AREA to 70. Mandatory retirement age for state employees is raised to 70 also. New Hampshire 1979 Abolishes mandatory retirement for all public and private sector employees. New Jersey 1985 Prohibits mandatory retirement in the private and public sectors except for public safety employees, judges, tenured higher-education faculty, and certain executive, but employers may refuse to hire or promote any person over 70. North Carolina 1984 Bans mandatory retirement under the State and local government employees' retirement systems. Local boards of education and the Board of Governors of the University of North Carolina may provide for the retirement of certain personnel at the age of 70 provided they allow for continued service on a year- to-year basis beyond mandatory retirement age. 1979 Raises mandatory retirement age for public employees to 70 and allows employment after that with employer's approval. Ohio 1990 Changes provisions of mandatory retirement to coincide with the Federal AREA. South Carolina 1988 Allows mandatory retirement for tenured faculty as well as executive and high policymaking employees until January 1, 1994. Until this time allows age- related hiring and discharge decisions for firefighters and law enforcement officers. 1979 Extends the human affairs law previously applicable to public employees, to private firms of 15 or more workers. Mandatory retirement age for teachers was increased from 65 to 70, with an extension to 72 possible upon approval of annual requests. South Dakota 1987 Removes mandatory retirement clause (at age 70) for State employees and teachers, except for tenured faculty at State colleges. 1979 Raises mandatory retirement age from 65 to 70 for public employees except for police and firefighters (who should retire at 55). Tennessee 1988 Uncaps the age requirement for antidiscrimination laws applicable to both private and public employees. Allows mandatory retirement for tenured faculty as well as executive and high policymaking employees until January 1, 1994. 1979 Eliminates mandatory retirement age of 70 for State employees, except for firefighters, police, and State university employees. Texas 1989 Uncaps the mandatory retirement age. Mandatory retirement is also prohibited for tenured faculty. Utah 1987 Extends coverage of the age discrimination prohibition to those 70 and over (instead of only those between 40 and 70). Mandatory retirement banned, except for executives and high policymaking employees who are required to retire at 65. Vermont 1987 Removes mandatory retirement requirement at the age of 70 for municipal employees. 1981 Prohibits age discrimination for persons 18 or older. Bans mandatory retirement for everybody except police and firefighters and tenured faculty at higher education institutions who are required to retire at 65 (after July 1982, at 70). Virginia 1987 Bans mandatory retirement for state employees and county, city, or other local public school professional/ clerical employees. 1980 Sets mandatory retirement age at 70 for State employees and teachers under the State Supplemental Retirement Act. Formerly the employer could require mandatory retirement at any age between 65-70. Washington 1979 Raises mandatory retirement age from 65 to 70 for all public employees except police and firefighters. Mandatory retirement at 70 can be waived by an individual's employer. Wisconsin 1989 Bans mandatory retirement. Source: Monthly Labor Review (all years), Neumark and Stock (1999), Northrup (1978), Rees and Smith (1991) Table 2 The effect of mandatory retirement on age-earnings profiles Dependent Variable: Unrestricted Control Group log(Earnings) Panel 1. (1) (2) Change in age-earnings slopes for: young faculty cohorts (24-34) 0.009 0.010 (0.001) *** (0.001) *** older faculty cohorts (35-80) -0.010 -0.010 (0.001) *** (0.001) *** Panel 2. Change in age-earnings slopes for young faculty cohorts (24-34) 0.011 0.012 working in public institutions (0.001) *** (0.001) *** young faculty cohorts (24-34) 0.006 0.008 working in private institutions (0.002) *** (0.002) *** older faculty cohorts (35-80) -0.009 -0.010 working in public institutions (0.001) *** (0.001) *** older faculty cohorts (35-80) -0.011 -0.011 working in private institutions (0.001) *** (0.001) *** Other controls: age, race, gender, marital yes yes status, union coverage, public institution, PhD. degree year dummies yes yes state dummies no yes year-age interactions no no state-year interactions no no Observations 14,890 14,890 R-squared 0.35 0.37 Dependent Variable: Unrestricted Control Group log(Earnings) Panel 1. (3) (4) Change in age-earnings slopes for: young faculty cohorts (24-34) 0.010 0.012 (0.001) *** (0.002)*** older faculty cohorts (35-80) -0.010 -0.009 (0.001) *** (0.002) *** Panel 2. Change in age-earnings slopes for young faculty cohorts (24-34) 0.012 0.013 working in public institutions (0.002) *** (0.002) *** young faculty cohorts (24-34) 0.009 0.011 working in private institutions (0.002) *** (0.002) *** older faculty cohorts (35-80) -0.010 -0.009 working in public institutions (0.001) *** (0.002) *** older faculty cohorts (35-80) -0.011 -0.009 working in private institutions (0.001) *** (0.002) *** Other controls: age, race, gender, marital yes yes status, union coverage, public institution, PhD. degree year dummies yes yes state dummies yes yes year-age interactions yes yes state-year interactions no yes Observations 14,890 14,890 R-squared 0.43 0.43 Dependent Variable: Restricted Control Group log(Earnings) (Faculty 34 or Younger at the Time of Policy Change) Panel 1. (5) (6) Change in age-earnings slopes for: young faculty cohorts (24-34) 0.008 0.010 (0.001) *** (0.001) *** older faculty cohorts (35-80) -- -- Panel 2. Change in age-earnings slopes for young faculty cohorts (24-34) 0.010 0.012 working in public institutions (0.001) *** (0.002) *** young faculty cohorts (24-34) 0.004 0.007 working in private institutions (0.002) ** (0.002) *** older faculty cohorts (35-80) -- -- working in public institutions -- -- older faculty cohorts (35-80) -- -- working in private institutions Other controls: age, race, gender, marital yes yes status, union coverage, public institution, PhD. degree year dummies yes yes state dummies no yes year-age interactions no no state-year interactions no no Observations 8,046 8,046 R-squared 0.39 0.41 Dependent Variable: Restricted Control Group log(Earnings) (Faculty 34 or Younger at the Time of Policy Change) Panel 1. (7) (8) Change in age-earnings slopes for: young faculty cohorts (24-34) 0.009 0.020 (0.001) *** (0.005)*** older faculty cohorts (35-80) -- -- Panel 2. Change in age-earnings slopes for young faculty cohorts (24-34) 0.012 0.021 working in public institutions (0.002) *** (0.005) *** young faculty cohorts (24-34) 0.007 0.017 working in private institutions (0.002) *** (0.005) *** older faculty cohorts (35-80) -- -- working in public institutions -- -- older faculty cohorts (35-80) -- -- working in private institutions Other controls: age, race, gender, marital yes yes status, union coverage, public institution, PhD. degree year dummies yes yes state dummies yes yes year-age interactions yes yes state-year interactions no yes Observations 8,046 8,046 R-squared 0.49 0.50 Earnings refer to usual weekly earnings in the current job. Faculty indicating part-time status or retired are excluded. I also remove self-employed and federal workers. The and average faculty member in 2008 earns about $70,000 (maximum $150,000). The average age is 46 years, 87% are white, 36% are female, 57% work in public institutions, 5% belong to a union * significant at 10%; ** significant at 5%; *** significant at 1% Table 3 Estimates of the effect of mandatory retirement on experience- and seniority-earnings profiles Fixed Dependent Variable: OLS effects log(earnings) (1) (2) Panel 1 Change in experience-earnings profiles 0.014 0.023 for faculty hired after mandatory retirement ban (0.005) *** (0.006) *** Change in experience-earnings profiles -0.005 -0.006 for faculty hired before mandatory retirement ban (0.003) (0.002) *** Panel 2 Change in seniority-earnings profiles -- -- for faculty hired after mandatory retirement ban Change in seniority-earnings -- profiles -- for faculty hired before mandatory retirement ban Productivity controls: impact-factor publication index 0.049 0.050 (pages per author, weighted by the impact factor) x 100 (0.013) *** (0.017) *** citations (x 100) 0.021 0.020 (0.007) *** (0.013) teaching awards 0.061 0.032 (0.041) (0.079) chair 0.239 0.130 (0.032) *** (0.040) *** administrative duties 0.133 0.075 (0.047) *** (0.039) * Observations 601 601 R-squared 0.63 0.71 Fixed Dependent Variable: OLS effects log(earnings) (3) (4) Panel 1 Change in experience-earnings profiles -- -- for faculty hired after mandatory retirement ban Change in experience-earnings profiles -- -- for faculty hired before mandatory retirement ban Panel 2 Change in seniority-earnings profiles 0.023 0.022 for faculty hired after mandatory retirement ban (0.011) ** (0.006) * Change in seniority-earnings profiles -0.004 -0.006 for faculty hired before mandatory retirement ban (0.003) (0.002) *** Productivity controls: impact-factor publication index 0.051 0.050 (pages per author, weighted by the impact factor) x 100 (0.013) *** (0.017) *** citations (x 100) 0.021 0.020 (0.006) *** (0.013) teaching awards 0.071 0.034 (0.040) * (0.079) chair 0.248 0.129 (0.034) *** (0.040)*** administrative duties 0.141 0.077 (0.048) *** (0.039)** Observations 601 601 R-squared 0.64 0.72 In addition to the above variables, all regressions include a variable measuring non journal publications (books, reports, etc.), gender, a dummy for whether Ph.D. was obtained from a top institution, indicators for field of research (5 categories), department dummies, and year dummies. Column (1) also includes experience, MR x PMR, and MR x(1-PMR). These variables are not identified in the fixed effects estimation in column (2). Column (3) includes a variable measuring job market experience prior to joining the current department, seniority, and interactions of each of these variables with PMR and (1-PMR). Similarly, these variables are not identified in column (4). Standard errors appear in parentheses and are robust to serial correlation and heteroskedasticity * significant at 10%; ** significant at 5%; *** significant at 1%
|Printer friendly Cite/link Email Feedback|
|Publication:||Atlantic Economic Journal|
|Date:||Mar 1, 2010|
|Previous Article:||Tax deductions for losses and equilibrium in competitive insurance markets.|
|Next Article:||Console price and software availability in the home video game industry.|