Most colleges and universities adopted budgets for the 2002-03 academic year in the spring and early summer of 2002. At that time, a pessimist might have cited several factors-negative rates of return from institutional endowments, a rising unemployment rate, an economic recession, and large increases in college and university enrollments, for example-to predict that faculty members would not see their earnings increase substantially in real terms in the coming year. The good news is that, overall and on average, the pessimists' worst fears proved incorrect. The bad news is that the overall averages don't tell the whole story.
Although higher education endowments grew at an unprecedented rate during most of the 1990s, the fiscal year ending June 30, 2002, marked the second straight year in which declines in the stock market caused endowments to earn negative rates of return. The average return rates for the years ending June 30, 2001, and June 30, 2002, were -3.6 percent and -6.0 percent, respectively.1
Most colleges and universities base their spending from their endowments on the average value of the endowments over a multiyear period (often three years). So the increase in spending from endowments was inevitably going to be lower (if endowment spending grew at all) in 2002-03 than in previous years. Of course, only a handful of American colleges and universities derive a substantial share of their operating budgets from endowment income, and it is these institutions that have been most seriously hurt by the stock market decline.
In July 2001 the average seasonally adjusted unemployment rate was 4.6 percent. Unemployment grew steadily over the next year to reach 5.9 percent in July 2002.2 When unemployment rates rise, the ability of families to afford college often decreases, while the need for financial assistance increases. Colleges and universities that base their financial aid decisions solely, or largely, on need therefore faced growing pressures in 2002 on their financial aid budgets-which usually compete for resources with faculty salaries in the overall institutional budget.
The recession caused many states to reduce their spending, and funding for public higher education suffered. Aggregate appropriations for higher education in 2002-03 rose by only 1.2 percent. This increase, which was the smallest since 1992-93, was only about one-quarter of the previous year's increase and lower than the rate of inflation. Indeed, thirteen states actually cut their appropriations to higher education.3 Moreover, in several states in which appropriations rose, the major growth was in direct state aid to students-not in funding for public institutions of higher education.
Large increases in enrollments in some state colleges and universities, coupled with small increases in per-student appropriations, led many of these institutions to push aggressively for tuition increases. As a result, tuition at public two- and four-year institutions rose by an average of 7.9 percent and 9.6 percent, respectively.4 These increases, although large in percentage terms, were actually substantially smaller in absolute terms than tuition increases at private colleges and universities. Consequently, administrators at public institutions had, on average, fewer resources to increase their faculty members' salaries than did their counterparts at private institutions.
A Year of Mixed Results
What actually occurred, as opposed to what was predicted? As I noted above, the pessimists' worst fears proved incorrect: faculty members' real salaries rose on average in 2002-03, although not by as much as they had in the previous year.5
However, faculty in private-independent (non-church-related) institutions did much better than their counterparts in public and church-related institutions.6 Because most American faculty are employed in public colleges and universities, and most American students are educated in this sector, the declining relative salaries of faculty in state-supported institutions is a matter of serious concern.
Table A (.pdf), which shows salary increases by rank since 1971-72, summarizes the good news. Data are presented in this table for two-year intervals prior to 1985-86 and for one-year intervals thereafter. The salary increases listed under the heading Nominal Terms are the actual percentage increases; those listed under the heading Real Terms show how faculty salaries have grown relative to the rate of increase in consumer prices.
A glance at the data for all faculty (which appear in the top half of the table) reveals that the average faculty salary (shown in the All Ranks column) was 3.0 percent higher in 2002-03 than it was in 2001-02. This increase was 0.8 percentage points lower than that of the previous academic year. The increase varied across ranks and, as in previous years, was highest for assistant professors.
The rise in average faculty salaries was smaller this year than last year even though the Consumer Price Index rose slightly more this year (2.4 percent) than last year (1.6 percent). As a result, the increase in average faculty salaries for all faculty members in real terms was only 0.6 percent this year. As the All Ranks column reveals, although this increase was small, 2002-03 is the sixth consecutive academic year in which the average faculty member's salary nationwide has increased in real terms.
Most faculty members are more interested in the percentage salary increases for continuing faculty (shown in the bottom half of the table) than in those for all faculty. Even if an academic institution keeps its overall budget for faculty salaries constant during a two-year period, those faculty members who remain at the institution over that period can usually expect to receive a salary increase between the two years. This occurs because some faculty members will leave the institution each year; they may move to another institution, take a nonacademic position, retire, or perhaps be denied tenure. To the extent that lower-paid junior faculty members replace those who leave, the salary funds saved can be used to augment the pay of those who remain. So, typically, the average salary increase received by continuing faculty members at an institution will be larger than the increase in the average salary observed at the institution.7
This past year was no exception to the rule. Continuing faculty members received salary increases that averaged 4.3 percent, which was 1.3 percentage points higher than the rise in the average faculty member's salary nationwide. On average, continuing assistant professors received greater percentage increases than did continuing associate and full professors. The differences between the assistant and the full professor ranks were, however, smaller this year than in recent years. In real terms, the average salary increase for continuing faculty exceeded the rate of inflation, as measured by the increase in the Consumer Price Index, by 1.9 percentage points. Although the purchasing power of the average continuing faculty member rose as a result, it did so at a rate less than two-thirds that of the previous year.
Behind the Averages
Survey report table 1 (.pdf) shows the percentage change in average salary levels and increases in the average salaries of continuing faculty members from 2001-02 to 2002-03, broken down by institutional category, affiliation (public, private-independent, or church-related), and academic rank. The table reveals that faculty members' economic gains differed considerably across institutional affiliations.
Continuing faculty employed at private-independent doctoral institutions received an average salary increase of 5.3 percent, which was substantially higher than the 4.2 percent and 3.8 percent increases received by their counterparts at church-related and public doctoral institutions, respectively. Indeed, the average salary increase received by continuing faculty at private-independent doctoral institutions exceeded the comparable increases received by continuing faculty at public doctoral universities by between 0.9 and 2.1 percentage points across professorial ranks.
At master's institutions, the differences between the private and public sector were somewhat smaller. Yet continuing faculty at public institutions received increases between 0.6 and 1.0 percentage points lower at each professorial rank than those conferred on their private-independent counterparts.
Continuing full professors at public baccalaureate institutions received slightly larger salary increases than their peers at private-independent institutions. But across all ranks, the average salary increase received by continuing faculty at public baccalaureate colleges and universities was about 1.1 percentage points lower than those obtained by continuing faculty at private-independent institutions.
Previous AAUP reports have documented that between the mid-1970s and the mid-1990s, the salaries of faculty at public institutions fell relative to the pay of faculty at private colleges and universities. This decline was particularly pronounced at the full professor level in doctoral institutions. Between 1978-79 and 1993-94, the average salary of full professors in public doctoral institutions fell from about 91 to 79 percent of that of full professors in private doctoral institutions. Between the mid-1990s and 2001-02, however, faculty in public and private institutions received roughly comparable average salary increases.8 During these years, therefore, faculty at public academic institutions did not see further erosion of their salaries compared with their private-sector counterparts. With the salary changes for 2002-03, the previous trend has resumed. I discuss some of the implications of this trend below.
Survey report tables 2 and 3 (.pdf)present data on the distributions of average changes in faculty salaries and of average increases in the salaries of continuing faculty members by institutional affiliation and category. These tables highlight how incomplete a picture focusing on averages provides. For example, survey report table 3 indicates that continuing faculty at 18.1 percent of all institutions received average salary increases of 6 percent or more and that continuing faculty at 15.6 percent of all institutions received average salary increases of less than 2 percent. Yet continuing faculty members at public institutions were much more likely to receive salary increases that were less than the rate of inflation than were their counterparts at private-independent institutions; some public-institution faculty received no increase at all. The current budget problems in many states undoubtedly generated this result.
In percentage terms, continuing faculty at 24.0 percent of public institutions received average salary increases below the inflation rate; continuing faculty at only 4.2 percent of private-independent institutions received similarly low raises. The public colleges and universities conferring increases of less than 2 percent employed 24.7 percent of all public-sector faculty, while the comparable private institutions employed only 1.9 percent of all faculty in that sector.
Rank and Gender
The other survey report tables accompanying this report present different ways to look at the data. Survey report table 4 (.pdf) provides information about average salary and average compensation levels for faculty members by rank in 2002-03. The table shows a slight compression in faculty salaries when one looks at faculty in the aggregate and takes into account information from previous annual salary reports. The ratio of the average full-professor salary to that of the average assistant professor declined somewhat in 2002-03 compared with the previous year, although the trend for that ratio over the past ten years has been generally upward. The full professor-associate professor ratio increased slightly for the current year, continuing a long-term trend of gradual increases. The associate professor-assistant professor ratio fell slightly this year, but it has remained remarkably consistent over the past three decades; associate professors earn about 20 percent more than assistant professors on average.
Survey report table 5 (.pdf) presents average salary data for men and women faculty by category, affiliation, and academic rank. Again, focusing on faculty in the aggregate reveals that women earned an average of 88.8 percent of what men earned at the full-professor level, 93.1 percent of what men earned at the associate-professor level, and 92.4 percent of what they earned at the assistant-professor level. At the full-professor level, the disparity is slightly greater than in 2001-02, while at the associate and assistant levels, it is slightly smaller.
Compared with last year's sample, the percentage of women among full professors rose from 21.4 to 22.3, the percentage among associate professors increased from 37.3 to 37.9, and the percentage among assistant professors decreased from 46.1 to 45.9. Great care must, however, be observed when drawing conclusions from these percentages, because the sample of institutions that report salary data by gender varies somewhat from year to year. The AAUP's Committee on the Economic Status of the Profession, perhaps working with the Association's Committee on the Status of Women in the Academic Profession, plans to analyze compensation by gender much more carefully in an upcoming project. Until we do so, we cannot conclude whether these changes reflect anything other than a fluctuation in the sample of institutions reporting data.
Over the past year, medical insurance cost increases nationwide far outstripped the increase in the Consumer Price Index. Survey report table 10 indicates that academic institutions were not immune to these cost increases. They had to deal with medical insurance costs that rose from 6.5 percent of the average faculty salary in 2001-02 to 7.3 percent of the average salary in 2002-03. Put another way, the jump in the cost of medical insurance added the equivalent of 0.8 percentage points to the increase in the average faculty member's salary. Many faculty members argue, however, that this additional outlay should not be considered an expansion of benefits to them, because increases in an academic institution's medical insurance costs typically reflect medical cost inflation, not an increase in the services covered by the institution's medical insurance plan.
In addition, at most colleges and universities, faculty members share with the institution premium costs for medical insurance. So whenever an institution's medical insurance costs rise as a percentage of faculty salaries, faculty members' premiums probably have also risen faster than their salaries have. Many faculty members believe that such a situation represents a decrease in their well-being; some say that if their health insurance premiums increase at faster rates than their salaries, their institutions should increase their salaries further to compensate them for the more expensive premiums.
Rising medical insurance costs leave neither faculty members nor their institutions very happy. Institutions scurry around, sometimes with the help of a faculty committee, to design plans that will reduce the rate of medical cost inflation. Faculty members worry that most plans that promise to do so will also reduce the real value of the medical services provided to them, often in the form of higher deductibles or greater copayments. Relations between faculty and administrators over economic matters would be much simpler if there were an easy solution to the problem of medical cost inflation.
Growing Public-Private Salary Differential
Several researchers have used AAUP data to document the decrease in the average salary of faculty members at public academic institutions relative to that of their peers at private institutions that took place between 1978-79 and 2001-02.9 Most of the decline occurred before the mid-1990s; the relative salaries of faculty in the public and private sectors remained roughly constant between 1996-97 and 2001-02. As the data in table A show, however, average salaries in public institutions of higher education dropped this past year relative to those in private institutions.
This relative decline in salaries at public colleges and universities probably makes it more difficult for them to hire and retain top faculty, especially at the senior level. Anecdotal stories tell of the "raiding" of public institutions by private colleges and universities that want to hire the public sector's tenured faculty members.10 s of yet, however, little systematic evidence exists to confirm such stories.
There is no national data series on turnover rates among tenured faculty members. But each year the AAUP collects institutional-level information about the number of continuing faculty members. Continuing faculty members in a particular rank are defined as full-time faculty members employed in that rank in the previous year and employed by the institution in the current year, regardless of their rank in the current year. For example, a faculty member who was an associate professor last year and was promoted to full professor this year counts as a continuing associate professor this year. The AAUP's Committee on the Economic Status of the Profession uses information that institutions report on the number of continuing faculty members in each rank and their total salaries in both the previous and the current year to estimate the average salary increases for continuing faculty that appear in our annual report.
Subject to some qualifications, information about the number of continuing faculty members in a rank in one year, coupled with data about the number in the rank in the previous year, allows the computation of a continuation rate for faculty members in each rank at the institution. That is done by dividing the number of continuing faculty members in the rank in one year by the number in the rank in the previous year.11 The continuation rate-or, more precisely, one minus the continuation rate-is a measure of faculty turnover from year to year in the rank.
The continuation rate cannot, however, be used to measure voluntary turnover among assistant professors, because some of those who leave an institution do so involuntarily when they are turned down for tenure. Similarly, the continuation rate for full professors is "contaminated" by faculty departures due to retirement, disability, or death. The continuation rate for associate professors, most of whom are tenured, comes closest to approximating a measure of voluntary turnover influenced by average salaries at an institution.
I worked with two colleagues, Hirschel Kasper and Daniel Rees, more than a decade ago to analyze the relationship between the continuation rate among associate professors at particular institutions and the average salary of associate professors. We used institutional-level information about continuation rates from the 1988-89 AAUP compensation survey to do so.12 We found that, other factors held constant, institutions with higher average salaries tended to have higher continuation rates (that is, lower voluntary turnover rates) than their competitors. Moreover, the magnitude of the relationship was largest at doctoral universities. Given the pattern of public-private salary differentials in recent years, one would expect that private institutions of higher education would have higher average continuation rates among associate professors than their public-sector counterparts.
Cornell undergraduate student Matthew Nagowski and I examined data on continuation rates from the AAUP compensation survey for the academic years 1996-97 to 2001-02. Some institutions did not report any information on continuing faculty members during the period, and others reported only intermittently. To avoid distortion of our findings as a result of institutions' moving in and out of the sample, we confined our attention to a sample of fifty-seven doctoral, sixty-seven master's, and thirty-eight bachelor's institutions (public and private) that reported data on continuation rates for each year of the selected period.13 Figures 1, 2, and 3 (.pdfs) plot out the weighted average continuation rates among associate professors at these institutions over the period. The average continuation rate at private doctoral institutions was consistently greater than that at public universities. Similarly, at master's and bachelor's institutions, the private continuation rate was greater in four of the six years analyzed; in the other two years, the rates were approximately equal. Although these simple comparisons do not prove causation, they do suggest one cost to public institutions of having lower faculty salaries than their private competitors.
Growing Dispersion of Average Salaries
It is well known that average faculty salaries at public colleges and universities have fallen relative to those at private institutions. Less recognized is the fact that in both the public and the private sectors, dispersion of faculty salaries has increased. Figure 4 (.pdf) plots the variance of the logarithm of average real salaries of full professors across doctoral institutions (public and private) at five-year intervals between 1962-63 and 2001-02 for ninety-six institutions that reported data in every year.14 The variance of the logarithm of average salaries is a measure of dispersion that is invariant to the nominal level of salaries. For example, if each institution doubled its average faculty salary, the variance of the logarithm of average faculty salary would remain constant.
As the figure indicates, the dispersion of average faculty salaries across the sample institutions decreased between the early 1960s and late 1970s, but it increased fairly steadily thereafter. Figures 5 and 6 (.pdfs) plot the variance of the logarithms of salaries among associate and assistant professors, and the story is the same. To show that this increasing dispersion is not confined to doctoral institutions, figure 7 (.pdf) presents similar data for the variance of the logarithms of average faculty salary at the full, associate, and assistant professor ranks for between eighty-one and eighty-four (the number varies by rank) private bachelor's colleges that reported data. Again, starting in the mid-1970s, the dispersion in average faculty salaries across institutions rose for all ranks, with the greatest increase at the senior levels.
In research reported elsewhere, Cornell graduate student Andrew Nutting and I used institutional-level data from 1972-73 to 1997-98 to estimate the logarithm of average faculty salary equations by rank separately for public and private doctoral institutions and private bachelor's institutions.15 The explanatory variables we used included endowment per student, tuition and state appropriations per student, and a set of institution-specific dichotomous variables for each institution. The inclusion of these latter variables makes what we did equivalent to specifying that the change in the logarithm of average faculty salary at an institution is a function of the change in endowment per student and in tuition and state appropriations per student at the institution. We used these estimates to understand which factors have contributed to the growing dispersion in faculty salaries across institutions.
Our models attribute most of the growing dispersion in average faculty salaries across private doctoral institutions and across private bachelor's institutions at each rank to the widening dispersion of endowment wealth during the period. To understand this relationship, it is important to realize that even if two institutions experience the same percentage increase in endowment per student during a period, the institution with the higher initial level of endowment per student will gain more in absolute terms than the institution with the lower initial level of endowment per student. If other sources of institutional income, such as tuition, grow at rates that are lower in percentage terms than the rate at which the endowment expands, the institution with the larger initial endowment per student will see its total income per student grow by a larger percentage than its relatively poorer counterpart. It could therefore increase its average faculty salary level by a greater percentage during the period.
Our estimates suggest that the growing variance of the logarithm of average faculty salaries at each rank at public doctoral universities arises from both increasing differences in endowment per student and widening differences in state appropriations per student. Nonetheless, changes in endowment per student played, at best, a minor role for all three professorial ranks. Most of the expansion in the variance of the logarithms of average real faculty salaries across public doctoral institutions results from increasing differences in the growth rates of state appropriations per student across institutions. Indeed, at the assistant and the associate professor levels, the widening dispersion of average faculty salaries can be explained solely by growing differences in the level of state appropriations per student across institutions.
The increased dispersion of average faculty salaries across institutions in the public and private sectors suggests that it is becoming more difficult for some institutions to attract and retain high-quality faculty. If faculty quality now differs more across institutions than it did in the past, where students choose to go to college may matter even more in the future than it has in the past.
Ambitious Agenda for the Survey
During our discussions this past year, members of the Committee on the Economic Status of the Profession set an ambitious agenda for our future annual reports. For example, we want to address the growing numbers of faculty who are employed part time or in non-tenure-track positions. To obtain information on these members of the profession, we must more carefully analyze and broaden our annual survey data and collect data from additional sources. Potential sources include the data on full-time lecturers that the AAUP collects (but which have never been fully analyzed); the biennial Fall Staff Survey that the National Center for Education Statistics has compiled as part of the Integrated Postsecondary Education Data System since 1987; and information from the administrative databases of the system offices of large state institutions of higher education.
Table B (.pdf) is an example of data from a state system, namely, the State University of New York (SUNY). It covers fall 1985 to fall 2001 and presents the ratios of full-time lecturers to full-time faculty with professorial ranks and of part-time faculty to full-time faculty. The faculty members are employed at the four state-operated university centers (doctoral institutions) and the thirteen state-operated university colleges (primarily master's institutions) of the SUNY system.16 ver the seventeen-year period, lecturers grew as a share of full-time professorial faculty at both the university centers and the university colleges, although the increase at the colleges was larger. Part-time faculty also grew relative to full-time faculty at these SUNY campuses; the ratio rose at each by about one-third. These increases were not, however, continuous. When resources permitted, the SUNY institutions increased their relative use of full-time tenured and tenure-track faculty.
Other issues the committee may explore include determining whether salary data can be obtained by gender, rank, and tenure status; revisiting how salary differentials by discipline have changed (a subject covered in previous annual reports); addressing in more detail issues relating to faculty health care; and analyzing institutional operating budgets to ascertain how their use has changed. We invite all members of the AAUP to suggest additional topics for us to consider. To do so, contact the AAUP's research director, John Curtis.
In 1991, in my role as chair of the Committee on the Economic Status of the Profession, I wrote the annual salary report. The next year, when one of my sons developed a life-threatening illness, I was forced to resign as chair. My son is now married and an attorney employed by the federal government in Washington, D.C., and it is with great pleasure that I return to serve another term as the committee's chair.
This report could not have been written without the hard work of John Curtis, who assumed responsibility for the annual compensation survey this year, or Galina Lewis, the AAUP's research associate. Nor could it have been completed without the many institutional representatives who take the time each year to respond to our annual survey. The members of our committee participated in two meetings (one virtual) that spelled out the issues to be discussed in this year's report, and many also commented on early drafts of the report. The committee members are Linda A. Bell (Economics), Haverford College; Daniel S. Hamermesh (Economics), University of Texas at Austin; George E. Lang (Mathematics and Computer Science), Fairfield University; Steven London (Political Science), Brooklyn College of the City University of New York; Jim Monks (Economics), University of Richmond; Saranna Thornton (Economics), Hampden-Sydney College; Craig Swan (Economics), University of Minnesota, consultant.
RONALD G. EHRENBERG
Committee on the Economic Status of the Profession
1. National Association of College and University Business Officers press release, January 6, 2003. Note that the rate of return on an endowment is not equal to the change in its market value. The latter also depends on an institution's spending from its endowment during the year and the additions to the endowment that it receives from gifts.Back to text
2. U.S. Bureau of Labor Statistics <www.bls.gov>. .Back to text
3. Michael Arnone, "State Spending on Colleges Increases at Lowest Rate in Decade," Chronicle of Higher Education, 13 December 2002. The data used in this article come from the annual survey of state support for higher education conducted by the center for the Study of Education Policy at Illinois State University and can be found on the center's website. .Back to text
4. Jeffrey R. Young, "Public-College Tuition Jumps at Highest Rate in Ten Years," Chronicle of Higher Education, 1 November 2002. .Back to text
5. Most of the information about faculty salary increases in this report is based on the AAUP survey of higher education institutions in the United States. In 2002-03, 1,454 institutions (representing 1,732 campuses) are represented in the survey. Data from these institutions are included in the basic results in table A and in many of the other tables in this report. AAUP staff compiled the data on which the tables in this report and the appendices that follow are based. .Back to text
6. Unless otherwise specified, the designation "private" in this article refers to private-independent (non-church-related) institutions..Back to text
7. There are exceptions to this statement. For example, an academic institution may divert all the saved salary funds to other uses, such as the startup costs for scientists and engineers or deficit reduction. In such a case, the average increase in all faculty salary would equal that in continuing faculty salary. Or, to take another example, more expensive "star" faculty may replace departing faculty. In that case, the average increase in all faculty salary could be higher than that in continuing faculty salary. Back to text
8. Daniel Hamermesh, "Quite Good News-For Now: The Annual Report on the Economic Status of the Profession 2001-02," Academe (March-April 2002)..Back to text
9. F. King Alexander, "The Silent Crisis: The Relative Fiscal Capacity of Public Universities to Compete for Faculty," Review of Higher Education 24 (winter 2001); Ronald G. Ehrenberg, "Studying Ourselves: The Academic Labor Market," Journal of Labor Economics (April 2003); Daniel S. Hamermesh, "Quite Good News-For Now," Academe (March-April 2002); and Cynthia Zoghi, "Why Have Public University Professors Done So Badly?" Economics of Education Review 22 (2003). .Back to text
10. See, for example, Scott Smallwood, "The Price Professors Pay for Teaching at Public Universities," Chronicle of Higher Education, 20 April 2001..Back to text
11. These qualifications relate to faculty who serve as administrators or who are on leave in either the current or the previous year. The presence of such individuals introduces possible measurement error into the calculation of the continuation rate..Back to text
12. Ronald G. Ehrenberg, Hirschel Kasper, and Daniel I. Rees, "Faculty Turnover in American Colleges and Universities," Economics of Education Review 10 (1991)..Back to text
13. Nonetheless, when we used a sample that consisted of any institution that reported data on continuing faculty members in any year, our findings were similar to those reported here..Back to text
14. Using a larger sample of all institutions that reported data in any year yields similar results..Back to text
15. Ronald G. Ehrenberg, "Studying Ourselves: The Academic Labor Market," Journal of Labor Economics (April 2003). The estimation period is limited by the availability of data from other sources..Back to text
16. Precise definitions of these variables are found in the notes to table B. The definition of who is a faculty member differs in these two sets of calculations; the part-time faculty data are head-count data and do not represent the number of full-time equivalent part-time faculty members..Back to text