The average salary of full-time faculty in the United States rose by 3.7 percent over the 1999–2000 academic year.1 For the sixth time in the past seven years, the increase in faculty salaries outpaced the rate of inflation at the consumer level, suggesting a sustained improvement in the purchasing power of full-time faculty.
This year’s increase in real (inflation-adjusted) faculty salaries—1.0 percent—is more modest than the rate of gain in each of the past two years (2 percent in 1998–99 and 1.6 percent in 1997–98). All told, however, faculty salaries have increased in real terms by 5.4 percent since 1993–94, and by 5.7 percent since the start of the 1990s. Among continuing faculty (those who were at the same institution in 1998–99 and in 1990–2000), the average salary increase was 4.8 percent.
That translates into real salary growth of 2.1 percent for continuing faculty—a rate that lags behind last year’s rise (3.2 percent) but one that matches the general pattern of increases in the 1990s.
This sustained growth in real salaries is surely good news. But set side by side with the raises paid to comparable professionals, these gains are modest. Unparalleled economic expansion and technological change over the past fifteen years have brought big economic rewards to the most highly skilled and educated workers. They have also led to rising income inequality, which has generated "winners and losers."
How we feel about our compensation depends partly on how much more or less we have to spend compared with the previous year (and in this sense we are happy). But our satisfaction also hinges on how much others like us have to spend (and in this sense we are blue). Thus, amid our optimism over recent faculty salary gains, we worry that we are losing ground in relative terms to the many professionals who are cashing in on talents not dissimilar to our own in private-sector nonacademic jobs. This concern is not only personal; it may also affect the future of quality higher education in this country.
Economists have shown that people’s career decisions lag behind salary trends, often for many years. Deterioration in the relative salary position of faculty therefore raises a troubling question. Could the growing opportunity cost of an academic career end up discouraging future generations of brilliant students from pursuing the trade? And is the well-documented salary inequality within our profession a response to the pressures of economy-wide trends?
This report analyzes the basic salary figures from this year’s survey with these questions as a backdrop. Specifically, I evaluate faculty salaries in relative terms—discussing both what other highly educated professionals earn and the growing salary differences among faculty members themselves. Key among these differences are the disparities in pay between public and private institutions, between research universities and four-year colleges, between "superstars" and the rest of us, and between men and women.
Trends in the Recent Past
Over the past fifteen years, the U.S. economy has benefited from low inflation, low unemployment, and a technological revolution that will likely continue to change the way our homes, businesses, and educational systems function. While most Americans have benefited from economic prosperity to some degree, the "new economy" has disproportionately rewarded the most highly skilled workers. So how have faculty, as well-educated, highly skilled professionals, fared in this economy?
A glance at table 1 tells a story of advancement—but only since the early 1980s. The All Rank salary category has risen in real terms 20.4 percent cumulatively since the 1981–82 academic year, or at an average annual rate of 1.25 percent over the past eighteen years.2 The average annual increase over the past three years—1.5 percent—is marginally greater than the eighteen-year average. But even these gains do not compensate for the salary losses of the 1970s; in real terms, faculty salaries are actually lower overall than they were in 1972.
Comparing the salary gains of faculty with those of other highly skilled professionals shows us how faculty have fared in relation to their counterparts outside academia. To do such a comparison, I looked at data from the Current Population Survey (CPS) produced by the U.S. Bureau of Labor Statistics for the years 1997 and 1985. From each of the two years of data, I included individuals who worked full time in occupations for which the mean educational attainment is at least two years of postgraduate education.3 I first compared the mean annual salaries of college and university teachers in the CPS to the mean salaries of workers in the other highly educated occupations. Columns 1 and 4 in table 2 report the mean salaries for the two occupational groupings for 1997 and 1985, and columns 2 and 5 translate these averages into percentage differences. The data show that the gap is growing between the earnings of college and university teachers and those of highly skilled workers in other occupations. Whereas faculty could expect to earn about 13.8 percent less than other highly educated professionals in 1985, by 1997 this disadvantage had nearly doubled to roughly 24 percent. Moreover, because the CPS data are capped at an annual salary of approximately $100,000, and because this cap affects a greater share of the nonacademic group than of the academic group, the reported differences in table 2 most likely understate the true gap between faculty and nonfaculty earnings.4
Of course, average figures by occupation do not control for differences in education, age, marital status, or gender, which, as labor economists know, may affect earnings. Adjusting the mean difference in salaries for these individual-specific factors required a more formal statistical analysis. Specifically, the natural log of individual earnings in the CPS high-education subsample in table 2 was regressed on a control for whether or not the individual was a college or university teacher, as well as on controls for the individual’s education, age, marital status, and gender. The adjusted salary differences from this calculation appear in columns 3 and 6 of the table. The adjustments increase the gap in pay in both years between college and university teachers on the one hand and their counterparts in other highly educated occupations on the other. The adjusted differences in pay rises from 23 percent in 1985 to nearly 32 percent in 1997.
Thus the CPS data reveal a widening gulf between the earnings of professors and those of other highly educated professionals. The magnitude of this gap and its trend, which these data most likely underestimate, show that faculty members have been losing ground over time to similarly educated workers. This fact gives reason for salary blues.
Differences Among Us
For many of us, growing salary discrepancies among faculty over the past decade have diminished the significance of the All Rank pay increases, which do not tell us as precisely how particular groups of faculty fare. In the 1990s, as salary differences among faculty increased, the All Rank figures actually masked a widening gap.
The AAUP collects salary data at the institutional level. With such data, inequality is best described in terms of the differences in salaries between institutions. Table 3 summarizes the growing differences across institutions in the average salaries paid to professors and assistant professors since 1985–86. The table shows that the variation in salaries between institutions has risen over time, even after controlling for faculty rank and type of institution. The 1999–2000 data reveal that the level of variation in average salaries is greatest among private-independent (non-church-related) institutions, although the trend rate of increase in salary variation has been faster among public institutions. All told, the data show modest increases in institutional inequality on the order of 2 to 6 percent over the fifteen-year period since 1985–86. This modest increase compares favorably with a more accelerated growth trend in inequality across individuals, which was highlighted in last year’s report.5
Salaries of "Superstars"
What role do the salaries of so-called superstars play in increasing inequality among institutions? Because the AAUP gathers data in such a way as to preserve the anonymity of faculty members, salary information for individuals is not available. Despite this limitation, the data in Survey Report Table 8 on page 29, which gives the distribution of faculty salaries, can be used to shed some light on the role of very high and very low individual salaries in driving institutional averages over time. Specifically, changes in the distribution of faculty salaries, characterized in part by a widening or a narrowing in the spread of average and median wages, give clues about salary trends at the tails of the salary distribution.
The median salary in the AAUP data is the pay level that the middle individual in the faculty salary distribution earns—the level corresponding to the fiftieth percentile of the distribution in Survey Report Table 8. The salary paid to the middle person remains unchanged even if the earnings of several highly paid individuals in the institution double or triple. But the average salary will change, perhaps by a great deal. In other words, a superstar can have a big impact on average salaries and no impact on the median salary, so long as the superstar effect remains isolated and does not lead to trickle-down increases throughout the distribution. In the presence of superstars, we would expect faster growth in average salaries and slower growth in median salaries over time.
If we had data on individual faculty members, we could compare average (mean) and median salaries precisely and evaluate many distributional issues completely. Instead, table 4 compares the growth of institutional average salaries and individual median salaries over much of the past decade, by rank and institutional category. The table shows that average faculty salaries in each rank and category grew faster than did median salaries, consistent with the notion of isolated high wages at the top of the salary distribution.6 This effect was especially true for salaries of full professors at doctoral-level (Category I) institutions, where mean salaries grew about 50 percent faster than median salaries over the 1990s.7 It is at that rank and in those types of institutions that faculty superstars are most likely to be heavily recruited and retained.
The growth in the spread between mean and median salaries supports the notion that superstar salaries have caused a widening distribution of salaries at the very top of the faculty pay scale. A similarly detailed analysis of the individual salary distributions in survey report table 8 shows that increases in salary inequality in the 1990s arose mostly from the spread in salaries at the top end of the salary distribution.8
Salary Advantage at Research Universities
Among the different types of institutions, doctoral-level (Category I) universities pay the highest salaries. Moreover, the advantage, or premium, for affiliation with a research university has grown over time. Table 5 documents the premiums paid to faculty in research universities for selected years since 1984–85. As the table shows, these premiums have led to salary differences with other types of institutions on the order of 30 to 50 percent. The largest premium is at the rank of professor, where competition for the most highly touted researchers drives up salaries. The next largest premium is at the rank of assistant professor, where research universities seem to have an increasingly powerful recruitment advantage.
Importance of Institutional Quality
Do the salary premiums reflect the higher quality of faculty at research universities, or simply the ability of these institutions to pay more? Each year, U.S. News & World Report ranks U.S. colleges and universities. These rankings can be used to evaluate the importance of perceived institutional quality in faculty salary differences.
To do such an analysis, I selected a subsample made up of the top twenty-two private institutions and the top twenty public universities from the magazine’s 1999 report.9 The rankings of the institutions I selected have remained fairly stable over time in the magazine’s survey. These rankings are based, however imperfectly, on several factors, including institutional resources (ability to pay) and faculty quality.
Table 6 highlights the gap between salary growth at top-ranked institutions and that at all research universities in the 1990s. It shows that salary increases at top-ranked public and private institutions modestly outpaced pay growth at the other research universities. It also reveals that salary increases at the top public universities lagged behind those at the top private universities. Indeed, as is clear from the table, growth at top public institutions actually trailed behind that at private universities generally for certain faculty ranks. These trends coincide with the theme of widening salary differences among us, because they reveal a rising premium to an association with a top-ranked university and because they are consistent with the public-private differences discussed above.
Salary trends at top-ranked universities may yield information about the role of institutional quality in explaining pay differences, since this small group is more homogeneous than public and private institutions overall.10 To the extent that a portion of the growing gap between salaries in public and private institutions arises from quality differences between institutions, the disparity between public and private institutions should be small in the top-ranked group.
Table 7 highlights the salary advantage of private institutions both before and after the "control" for institutional quality. The top panel provides data for all institutions by rank and shows a modest increase in the premium paid to faculty at private institutions over time. The middle panel, which compares private and public research universities, shows higher overall premiums paid to faculty in private institutions and a similar growth trend. The bottom panel compares the subsample of top-ranked private and public universities and shows smaller differences between salaries at these institutions. But even though the salary premium to faculty at private universities diminishes by roughly one-third in the top-ranked subset, it persisted, and, if anything, accelerated modestly in the 1990s.11 The remaining differences can be reasonably attributed to ability to pay and other factors.
Relative Salary Levels Among the Top-Ranked Universities
The salary differences at the rank of full professor between the top public and the top private-independent universities are indeed large. Figure 1 presents these differences in both nominal (actual dollars) and real (cost-of-living adjusted) terms. The top-ranked private universities are disproportionately represented at the left of the figure, where the highest nominal salaries appear, but their advantage decreases, and the ranking of institutions changes after adjusting for cost of living. For example, Harvard University, which pays the highest average salary in the group, slips to rank fifteen after adjusting for the cost of living in the Cambridge-Boston metropolitan area. Overall, the cost-of-living adjustment increases the relative salary rankings of the top public universities while reducing the salary premium to the top private institutions from 15.6 percent in nominal terms (see table 7) to 10.6 percent in price-adjusted terms, or by nearly one-third.
Figure 2 shows the salary advantage of professor-rank men at five-year intervals during the 1990s. The figure reveals two patterns. First, the largest salary differences between men and women faculty occur in research universities, while the smallest occur in four-year colleges. Second, the differences between men and women have not narrowed in all types of institutions, and in comprehensive (Category IIA) institutions, they actually increased over the 1990s. Research universities have been most successful in reducing gender-related salary differences, but the differences at such institutions remain substantial. The persistence of gender-related salary inequities among faculty is especially troubling at a time when the gap between men and women in the economy at large is narrowing, particularly among the highly educated.
Figure 3 shows the salary advantage of male professors by rank and category using this year’s salary data. The premium is greater at research universities across all ranks, and is most evident in all institutional categories at the professor level. Note the sizeable All Rank male premium, which is particularly evident at doctoral- level (Category I) and baccalaureate (Category IIB) institutions. Presumably, this advantage results from disproportionately large numbers of women at the lower ranks, which amounts to striking evidence of a distorted gender distribution by rank.
Figure 4 completes this analysis, comparing male salary premiums at top-ranked institutions with those at other research universities. The premium is higher at the professor and the associate professor ranks for the elite private universities, but lower at all ranks for the top-ranked public institutions. The male salary premium is slightly higher at the professor and the associate professor ranks in public universities than at private-independent institutions. Because salaries in the top-ranked colleges and universities are higher on average, and because women are disproportionately represented in lower-paying institutions, we would expect the gender differences to shrink in the top-ranked group.12 The fact that they do not do so uniformly gives us pause.
The persistence of the male salary advantage has to do partly with gender distribution by discipline. A study conducted at Harvard University found that women at Harvard held only 11 of the 162 tenured positions in the high-salaried natural sciences—just 6.8 percent. By comparison, 14.4 percent of the tenured professorships in the social sciences and 21.9 percent of those in the humanities were held by women.13 The Harvard study attributes the paucity of women scientists to an inadequate supply of women in doctoral programs in the sciences. Women now earn only a quarter of all science degrees awarded at Harvard, but more than a third of all social-science degrees and half of all humanities degrees.
Another explanation for the male-female salary difference and its relative strength in research universities may be found in the difference in how academic men and women spend their working time. A recent study by economist Robert Toutkoushian using data from the 1993 National Survey of Postsecondary Faculty highlighted significant differences in time allocation by gender.14 Specifically, he found that women faculty spend relatively greater amounts of time at teaching and college or university service and proportionately less time at research than their male counterparts. Not surprisingly, the data show that the research output of women lagged behind that of the men.15
Finally, comparing the salaries of men and women faculty at the same rank may provide useful information, but such a comparison fails to detect another troubling trend, namely, an increase in the percentage of women faculty relegated to lower-status appointments. 16 The AAUP salary survey includes information on tenure rates by rank and on the number of men and women professors and associate professors. These data can be combined to create tenure rates for women at both ranks. The significance of associate professor status is not uniform across institutions—that is, a few universities confer tenure only at the full-professor rank. But at institutions that have both ranks, this distinction is often characterized by different status.17
Table 8 shows the tenure rates for women faculty by rank derived from using these combined data. The table reveals wide differences in the percentages of women tenured at the full and associate professor ranks. A disproportionate share of women hold tenure at the less-senior (and lower-status) rank of associate professor. Note as well that fewer women are tenured at research universities, regardless of whether the institutions are public or private-independent.
This year’s report is aptly titled. While it is foolhardy to pursue an academic career hoping to enjoy the same financial rewards as someone who launches a successful Internet start-up company, the fact remains that the differences have grown between what we earn and what they earn. And these differences may make us blue and cause us to ponder, and may perhaps even drive some potential faculty members away.
Less significant for the future of quality higher education in a global sense, but important nonetheless, are the growing differences among us. This report has highlighted the persistent and widening gap between salaries at public and private-independent institutions, between faculty at research universities and those at other types of institutions, and between women and men. Using "top-ranked" institutions as a crude control for institutional and faculty quality allowed us to see that although these differences diminish somewhat at reputedly elite institutions, they do not disappear. Not surprisingly, most salary differences are probably a function of an institution’s ability to pay. Finally, the evidence is consistent with (but not proof of) the notion that "superstar" salaries have been driving up average wages. This possibility and the other growing salary differences discussed in this report teach us that we must look increasingly beyond trends in average salaries to examine the trends for specific groups in our profession.
In this report, I have benefited from the help of Ernst Benjamin, the AAUP’s director of research, who not only was on constant call for last-minute data demands, but who suggested several of the topics pursued in this report and commented meticulously on an early version. Patrick McCauley, the associate director of the AAUP Faculty Salary Compensation Survey, and Clarice Evans, our new research associate, are also to be thanked for the many hours they put in to ensure the timely release of accurate salary numbers. In addition, I want to thank the members of Committee Z, who suggested topics to be pursued in this report and ways to improve on previous years of salary reporting within Academe. Special thanks go to those members of the committee who read the draft report in the eleventh hour and provided helpful and thorough review.
The Committee Z members are:
W. Lee Hansen (Economics, University of Wisconsin–Madison),
Anne Harrison (Finance and Economics, Columbia University),
James May (Communication Science and Technology, California State University–Monterey),
Lonnie Stevans (Business, Hofstra University),
Craig Swan (Economics, University of Minnesota–Twin Cities),
Jeffrey Waddoups (Economics, University of Nevada, Las Vegas).
Linda A. Bell, chair (Economics, Haverford College)
Committee Z on the Economic Status of the Profession
1. Most of the information in this report is based on the AAUP survey of higher education institutions in the United States. In 1999–2000, 1,769 institutions (representing 2,107 campuses) are represented in the survey. Data from these institutions are included in the basic results in Table 1 and many of the other tables in this report. AAUP staff compiled the data on which the tables in the report and the appendixes that follow are based. Back to Text
2. Differences in salary growth at public and private-independent (non-church-related) institutions since the 1980s imply cumulative pay losses for public university faculty and modest gains for professors at private-independent institutions when the average figures are disaggregated. In this article, the designation "private-independent" does not include church-related institutions, which are listed separately in the survey report tables that follow this article. Back to Text
3. Full-time workers are defined as those reporting at least thirty-five hours per week of usual work. The samples are then truncated further to exclude occupations in which the mean education is less than two years of postgraduate study. Back to Text
4. Specifically, the annual earnings figures are imputed from weekly earnings, which are truncated at $999 in the 1995 data and at $1,923 in the 1997 data. Truncated data have been adjusted by a factor of 1.5 in both years. The earnings of college and university faculty are disproportionately less likely to be truncated in both years, but the difference is increasingly profound in the 1997 data. Back to Text
5. The increase in inequality across individual faculty in the CPS data is about 16 percent since 1979. See Table 4 on page 19 of the March– April 1999 issue of Academe. Back to Text
6. Some caution is advised in interpreting these results. The precise comparison would involve average and median individual-level wages. But because the AAUP data do not provide information on average individual wages, I use average institutional wages. In so doing, I make several strong assumptions, namely that the average salary reported by the AAUP and derived from institutional average salaries is a reasonable proxy for an average salary derived from individual salaries, that the relative distribution of salaries is constant across institutions, and that the relative weights of institutions are fixed through time. Back to Text
7. For definitions of institutional categories, see Explanation of Statistical Data on page 37. Back to Text
8. A common way to describe inequality is to take the range of reported salaries and to characterize the difference in salary between institutions at the top ninety-fifth percentile of the salary distribution and at the bottom fifth percentile of the salary distribution. This overall difference can reasonably be separated into the portion attributable to differences at the top (the difference between what the ninety-fifth and the fiftieth percentile in the salary distribution earn) and to differences at the bottom (the difference between what the fiftieth and the fifth percentile in the salary distribution earn). For the AAUP ninety-five to five percentile spread, the larger share is attributable to the ninety-five to fifty percentile spread. Back to Text
9. Because Carnegie Mellon University was excluded from the 1999 ranking but had been included in several previous years, I chose to include it in my analysis. Similarly, a twentieth-rank tie led to a total of twenty-two private institutions in my sample. My highest ranked private universities include Harvard University, Princeton University, Yale University, the Massachusetts Institute of Technology, Stanford University, Cornell University, Duke University, University of Pennsylvania, Brown University, Columbia University, Dartmouth College, Northwestern University, Johns Hopkins University, University of Chicago, Emory University, Washington University (St. Louis), Rice University, University of Notre Dame, Georgetown University, Vanderbilt University, Carnegie Mellon University, and the California Institute of Technology. The highest-ranked public universities include the University of California, Berkeley; Rutgers, the State University of New Jersey, New Brunswick; the University of California, Los Angeles; the University of Virginia; the University of California, San Diego; the University of California, Irvine; the University of California, Santa Barbara; the University of Michigan, Ann Arbor; the University of California, Davis; the University of Texas at Austin; the University of North Carolina at Chapel Hill; the University of California, Santa Cruz; Penn State University, Main Campus; the Georgia Institute of Technology; the University of Illinois at Urbana-Champaign; the University of Minnesota–Twin Cities; the College of William and Mary; the University of Wisconsin–Madison; Texas A&M University; and the University of Washington. Back to Text
10. This is a somewhat dubious assumption, at least according to several statistics related to institutional quality that show that the top public institutions trail the top private institutions considerably. For example, the top public institutions are in serious deficit in terms of average Scholastic Assessment Test scores, acceptance and graduation rates, the percentage of first-year students in the top 10 percent of their high school classes, and student-faculty ratios. See Kevin Schroeder, "Public and Private Elite University Salary Differences," Undergraduate Thesis, Haverford College, 1998. Back to Text
11. The Schroeder thesis, cited above, decomposes the private-public premium into quality and ability-to-pay factors. It suggests a prominent role for ability-to-pay factors in explaining the public-private difference. See the manuscript for the statistical tests supporting this conclusion. Back to Text
12. On the disproportionate representation of women in lower-paying institutions, see Ernst Benjamin, "Disparities in the Salaries and Appointments of Academic Women and Men," Academe (January–February 1999): 60–62 . Back to Text
13. See Harvard Magazine (January–February 2000): 66–68. Back to Text
14. See Robert K. Toutkoushian, "The Status of Academic Women in the 1990s: No Longer Outsiders, but Not Yet Equals," Quarterly Review of Economics and Finance 39 (special issue, 1999): 679–98. Toutkoushian based his analysis on data from the 1993 "National Survey of Postsecondary Faculty" by the U.S. Department of Education’s National Center for Education Statistics. Back to Text
15. Interestingly, the same qualitative differences exist between faculty in public and private research institutions: faculty in public institutions are likely to spend more time on teaching and less time on research than are faculty in private-independent institutions. Back to Text
16. Benjamin, "Disparities in the Salaries and Appointments of Academic Women and Men." See note 12. Back to Text
17. Specifically, using the AAUP raw data from last year, I summed the numbers of men and women in each rank and category for public and private-independent institutions to arrive at an aggregate figure for the total number of male and female faculty. I then used the tenure rates from Suvery Report Table 11 to estimate the numbers of tenured faculty at each rank and by institutional control. Note that the table figures do not offer specific male and female tenure rates but an overall tenure rate average. Back to Text