But Are They Competitive in Seattle? An Analysis of Educator and Comparable Non-educator Salaries in the State of Washington

Transcription

1 But Are They Competitive in Seattle? An Analysis of Educator and Comparable Non-educator Salaries in the State of Washington Submitted to: Submitted by: Compensation Technical Working Group Quality Education Council Dr. Lori Taylor Texas A&M University April 2012

2

3 Table of Contents Executive Summary... i Comparing Educators with Non-educators in Washington... i Comparing Base Teacher Salaries across States... v Comparing Fringe Benefits across Sectors... vi Conclusions... vi Introduction... 1 Comparing Educators with Non-educators in Washington... 1 Updating the NCES Comparable Wage Index... 2 Updating the High School CWI... 5 Estimating Occupation-Specific Wage Levels... 7 The Prevailing Salaries for School District Personnel Data and Estimation The Prevailing Salaries for Teachers The Prevailing Salaries for Beginning Teachers The Prevailing Salaries for Math and Science Teachers The Prevailing Salaries for Certificated Non-teaching Personnel The Prevailing Salaries for Classified Staff Summarizing the Evidence on Relative Salaries Comparing Base Teacher Salaries across States Comparing Fringe Benefits across Sectors Conclusions Acknowledgements Endnotes Bibliography Appendix A: Supplemental Data Tables... 57

4 Executive Summary Wages vary substantially from place to place and from occupation to occupation. In order to attract and retain a high quality workforce, Washington school districts must offer teachers a salary and benefits package that is competitive not only with teaching jobs in other states, but also with non-teaching jobs in the local community. This report examines the relative salaries and benefits of Washington educators using three different lenses. The first lens compares estimates of the prevailing salaries for educators with estimates of the prevailing salaries for non-educators. The second lens compares base teacher salaries in Washington with base teacher salaries in other states. The third and final lens examines the extent to which the fringe benefits teachers receive in the state of Washington are competitive with private-sector benefits. Whenever possible, the analysis has been conducted for each school district, metropolitan area, and non-metropolitan labor market in the state. Comparing Educators with Non-educators in Washington Average wages are typically low in communities where most of the workers are young and inexperienced, and high in communities where most of the workers are college-educated. Areas where most of the college graduates are health care workers will tend to have higher average wages than areas where most of the college graduates are social workers. Areas where most of the accountants are relatively inexperienced will have lower average accounting wages than areas where most of the accountants are highly experienced. Average teacher salaries can be high in a district that chooses to hire only experienced teachers with advanced degrees, and low in a district that can only afford to hire beginning teachers. None of these differences in average wages necessarily imply anything about differences in the competitiveness of educator salaries. To make fair comparisons between educators and non-educators in various locations, one needs to consider the demographically and occupationally adjusted or prevailing salaries. Variations in the prevailing salaries of educators reflect how much more or less each school district spends to recruit and retain similar school personnel. Meanwhile, variations in the prevailing salaries for each occupation indicate how much more or less employers pay in each location to employ the typical worker. Comparing prevailing salaries for educators with prevailing salaries for non-educators provides a particularly useful lens through which to view the relative competitiveness of educator salaries in the state of Washington. Baseline estimates of the prevailing salaries for non-educator occupations come from regression analyses of individual earnings data from the 2000 U.S. Census. Those baseline analyses were then updated using earnings data from the Occupational Employment Survey (OES), which is conducted annually by the U.S. Bureau of Labor Statistics (BLS). The methodology was adapted from the one used to generate the National Center for Education Statistics Comparable Wage Index (CWI), and generally follows the methodology Taylor (2008a) used in a previous analysis of educator salaries in Washington. Thus, I used the baseline regression model underlying the CWI to predict the prevailing salary in 1999 for each certified occupation under analysis. The prevailing salary in each labor market is the salary that Page i

5 would be expected for a college graduate who had the same educational and industrial profile as the average Census respondent in that occupational category, assuming that the person worked 40 hours a week and 52 weeks a year. I then used the OES data to calculate the growth in wages between 1999 and 2010 for each occupation and location, and adjusted the baseline prevailing salaries accordingly. For example, the baseline CWI regression model predicts that the prevailing salary for registered nurses in Seattle was $48,002 in Analysis of the OES data indicates that, on average, nurses salaries in Seattle rose percent between 1999 and Therefore, the prevailing salary for nurses in Seattle in 2010 was predicted to be $73,708 ($73,708 =$48,002*1.5355). Similarly, I estimated the prevailing salary for classified personnel using a baseline regression analysis of high school graduates without college degrees. Table E.1 indicates the occupations and prevailing salary estimates used in this analysis. Table E.1: State Average Prevailing Salaries in Washington, by Occupation State Average Predicted Salary 2010 Certified Occupations All College Graduates $67,515 ACM Teacher-Comparable Occupations $65,923 STEM Occupations $76,051 Registered Nurses $68,231 Social Workers $47,421 Psychologists $61,681 Counselors $43,606 Occupational Therapists $73,529 Physical Therapists $73,251 Speech And Language Pathologists $70,223 Audiologists $71,363 Selected Managerial Occupations $92,704 All Managerial Occupations Except Legislators $88,900 Classified Occupations All High School Graduates $41,958 Supervisors $45,943 Office And Administrative Support Occupations $36,344 Janitorial Occupations $30,353 Protective Service Occupations, Excluding Fire Safety Personnel $48,221 Information Technology Occupations $60,901 Facilities, Maintenance and Grounds $45,059 Warehouse, Laborers and Mechanics $42,572 Motor Vehicle Operators $38,039 Food Preparation and Serving Occupations $28,754 Note: The state average predicted salary is a pupil-weighted average of the salary predictions for each school district. The pupil-weighted state average is calculated using the FTE student counts from the school year. Page ii

6 Estimates of the prevailing salary for Washington educators come from hedonic wage analyses of data provided by the Office of Superintendent for Public Instruction (OSPI). The hedonic salary models for Washington educators describe each educator s salary as a function of his or her personal characteristics, his or her job assignments, and the school building and school district in which he or she works. I use these models to predict average full-time-equivalent salaries in each school district, holding constant the influence of demographic and job characteristics. Those predictions indicate the prevailing salaries in each school district. Variations in the prevailing salaries reflect how much more or less each school district pays to recruit and retain comparable school personnel. The prevailing salary for a labor market is just a weighted average of the prevailing salaries in its constituent school districts. This analysis applies hedonic wage analysis to two measures of educator salaries fulltime-equivalent base salaries and full-time-equivalent total salaries. Base salaries measure employee earnings during the school year under terms of the base employment contract and are paid for by the state. Total salaries measure the final gross pay of each employee from all sources, including the state, local levies, and federal monies. For purposes of this analysis, pay for extracurricular and public activities has been excluded from both base and total salaries. Figure E.1 illustrates the results of the analysis of all types of teachers. The figure indicates the ratio of teaching salaries to comparable non-teaching salaries in this case the prevailing salary for all college graduates. A relative salary greater than 100 percent indicates that the average teacher is paid better than the average college graduate, whereas a relative salary less than 100 percent indicates that the average teacher is paid less than the average college graduate. As the figure illustrates, relative total salaries are more than 83 percent, on average, in all of the labor markets in the state. The 83-percent threshold is important because it indicates that full-time-equivalent total salaries for teachers in Washington are at or above the 10-month salaries for college graduates (10/12=.833). Recall that the prevailing salary for college graduates was constructed assuming that non-educators worked 52 weeks per year. The typical school year is obviously shorter than that. A common rule of thumb is to assume that the school year is 10 months long. A relative salary above 83 percent suggests that teaching salaries are higher than the 10-month salaries for the average college graduate. Thus, the evidence suggests that total teacher salaries are competitive with non-teacher salaries throughout the state of Washington. The evidence on relative base salaries is much more mixed. Relative base salaries are above the 10-month threshold in all of the county clusters except Cowlitz, Klickitat and Wahkiakum counties in southwestern Washington. They are below the 83 percent threshold in all of the major metropolitan areas except Bellingham and Spokane. The base salary for an average teacher in the Seattle metropolitan area is only 71 percent of the average salary for a college graduate. Page iii

7 Figure E.1: Relative Teacher Salaries by Labor Market, Chelan, Douglas, Kittitas and Okanogan Asotin, Columbia, Garfield, Walla Walla and Whitman Adams, Grant, Ferry, Lincoln, Pend Oreille and Stevens Clallam, Jefferson and Mason Grays Harbor, Lewis and Pacific Bellingham Metropolitan Area Island, San Juan and Skagit Spokane Metropolitan Area Cowlitz, Klickitat and Wahkiakum Portland Vancouver Beaverton Metropolitan Area Olympia Metropolitan Area Yakima Metropolitan Area Tacoma Metropolitan Area Bremerton Silverdale Metropolitan Area Kennewick Richland Pasco Metropolitan Area Seattle Bellevue Everett Metropolitan Area Base Pay Total Final Salary 60% 70% 80% 90% 100% 110% Relative Teacher Salary Note: Relative teacher salaries are the full-time-equivalent teaching salaries divided by the prevailing salary for all college graduates in each location. Source: Author s calculations using OSPI s S-275 files and the updated CWI. The complete report presents similar analyses for a variety of school district personnel. The analysis examines two additional teacher groups beginning teachers and teachers who are certified in math and science. It also examines relative prevailing salaries for a variety of other certified and classified personnel, including school administrators, educational staff associates, teacher aides and food service workers. Together with the analysis of all teachers, these analyses support four key findings: 1. The salaries most Washington teachers actually receive (i.e. their total final salaries) meet or exceed the salaries received by comparable non-teachers in their communities. On average, teachers in Washington earn 91 percent of the annual salary for the average college graduate, despite working a substantially shorter year. Only 30 school districts, which serve only 4 percent of the school children in Washington, pay total teacher salaries below the average 10-month salary for a typical college graduate. Page iv

8 2. In contrast, the teacher salaries funded by the state through the school finance formula (i.e. the base salaries) are not competitive in most major metropolitan areas. Although base salaries are competitive in some parts of the state, less than one quarter of the school children in Washington attend a school district where base teacher salaries equal or exceed the 10-month salary for a typical college graduate. The base salary for an average teacher in the Seattle metropolitan area is only 71 percent of the average salary for a college graduate. 3. As a general rule, non-teaching school district employees receive salaries that are competitive with or well above those received by their counterparts outside of the education sector. The only major exceptions are the instructional aides. Teacher aides earn substantially less than the typical high school graduate throughout the state. 4. The non-teaching salaries funded by the state are generally not competitive. In the Seattle and Kennewick metropolitan areas, for example, the salary allocation for school district administrators represents less than 63 percent of the prevailing salary for comparable managers, on average. Comparing Base Teacher Salaries across States The Schools and Staffing Survey (SASS) is conducted periodically by the National Center for Education Statistics. Public school districts, principals and teachers throughout the nation are surveyed about a variety of education topics, including teacher salaries and benefits. Those survey responses are the best available evidence for determining whether or not teacher salaries in Washington are competitive with those in other states and form the basis for the analysis in this section of the report. In all cases, salaries have been adjusted for regional differences in labor cost using the updated CWI. The most recent SASS covers the school year and surveyed school districts about their normal yearly base salary. Thus, this is an analysis of base salaries rather than total salaries. In , total final salaries (excluding extracurricular and public activities) exceeded base salaries by an average of 15 percent in Washington, so the salaries that teachers actually received were probably more competitive than their base salaries. Unfortunately, the SASS provides no information on the relationship between base salary and total salary in each state, so it is impossible to say how much more competitive. Analysis of the SASS suggests that base teacher salaries in Washington are low by national standards. After adjustments for regional differences in labor cost, only Colorado, North Dakota, Iowa and Washington DC had base salaries for starting teachers that were lower than those in Washington. Cost-adjusted base salaries for mid-career teachers were also near the bottom of the national distribution. Because base salaries in Washington have risen more slowly than salaries in other occupations since , it is unlikely that Washington s position relative to other states has improved substantially over the last few years. Page v

9 Comparing Fringe Benefits across Sectors The third and final lens compares the typical benefits packages in public education to those available in the private sector. The evidence presented here comes from the U.S. Bureau of Labor Statistics (BLS) and from analyses of survey data conducted by the Economic Policy Institute (EPI) and the Employee Benefits Research Institute (EBRI). Those sources can be used to compare benefits in Washington with benefits in the rest of the country and to compare benefits by occupation. Given the limitations in the data, it is not possible to reliably compare benefits by occupation within the state of Washington. However, the evidence suggests that the benefit patterns for the state of Washington largely mirror those of the nation as a whole. Here, the evidence is clear. Teachers in Washington are more likely to receive retirement and health insurance benefits than comparable private sector employees, and school districts pay more for teacher benefits than comparable employers pay for non-teacher benefits. Conclusions Each of the three lenses used in this analysis report provides a slightly different perspective on educator compensation in the State of Washington. All told, the evidence suggests that teacher base salaries are generally not competitive with teacher base salaries in other states or with comparable non-teacher salaries in metropolitan Washington. Base salaries are also not competitive for most non-teaching personnel. On the other hand, total salaries are competitive in most of the state, and the fringe benefits appear unusually generous. As such, the total compensation packages offered by Washington school districts appear sufficient to attract and retain a high quality workforce. Page vi

10 Introduction Wages vary substantially from place to place and from occupation to occupation. In order to attract and retain a high quality workforce, Washington school districts must offer teachers a salary and benefits package that is competitive not only with teaching jobs in other states, but also with non-teaching jobs in the local community. Four years ago, Taylor (2008a) compared teacher salaries with non-teacher salaries in Washington. She found that teaching was highly competitive with non-teaching occupations in rural and eastern Washington, but much less competitive with non-teaching occupations in the Kennewick, Bremerton and Seattle labor market areas. Her analysis suggested that in the gap between teaching salaries and non-teaching salaries exceeded 20 percent in those three markets. Much has changed since 2007, however. The relationship between teaching salaries and non-teaching salaries has shifted, and there is renewed interest in the relative salaries of school district personnel who are not teachers. Therefore, this report updates and extends the earlier report by Taylor (2008a). This analysis: Updates the National Center for Education Statistics Comparable Wage index (NCES CWI) through 2010 and uses it to generate estimates of the general wage level for college graduates in each Washington labor market; Updates the High School Graduate Comparable Wage Index (HS CWI) through 2010 and uses it to generate estimates of the general wage level for high school graduates in each Washington labor market; Develops occupation-specific wage estimates for each labor market using the CWI methodology; Develops estimates of the prevailing salaries for Washington educators, and compares those salaries with the salaries of comparable workers outside education, using wage estimates based on the updated CWIs; Compares teacher salaries in Washington with teacher salaries in other states; and Compares the typical benefits packages in public education to those available in the private sector. Whenever possible, the analysis has been conducted for each school district, metropolitan area, and non-metropolitan labor market in the state. Comparing Educators with Non-educators in Washington Average wages are typically low in communities where most of the workers are young and inexperienced, and high in communities where most of the workers are college-educated. Areas where most of the college graduates are health care workers will tend to have higher average wages than areas where most of the college graduates are social workers. Areas where most of the accountants are highly experienced will have higher average accounting wages than areas where most of the accountants are relatively inexperienced. None of these differences in Page 1

11 average wages necessarily imply anything about differences in the competitiveness of educator salaries. Average teacher salaries are also a poor basis for comparisons across districts, locations and school years. Average salaries can increase over time as teachers move up the salary scale, even if the scale itself remains unchanged. Average salaries can be high in a district that chooses to hire only experienced teachers with advanced degrees, and low in a district that can only afford to hire beginning teachers. Such differences do not tell us whether or not a teacher with the same qualifications would expect to receive a higher wage from one district than from another. To make fair comparisons between educators and non-educators in various locations, one needs to consider the demographically and occupationally adjusted or prevailing salaries. Variations in the prevailing salaries of educators reflect how much more or less each school district spends to recruit and employ similar school personnel to do similar jobs. Meanwhile, the prevailing salaries for each occupation indicate how much more or less employers pay in each location to hire the typical worker. Comparing prevailing salaries for educators with prevailing salaries for non-educators provides a useful measure of the relative competitiveness of educator salaries in the state of Washington. It is the approach taken here. Updating the NCES Comparable Wage Index The NCES CWI measures the prevailing wage for college graduates in 800 U.S. labor markets. The baseline estimates (for 1999) come from a hedonic wage analysis of the individual earnings data from the 2000 U.S. Census. 1 Annual updates to that baseline come from regression analyses of occupational earnings data provided by the U.S. Bureau of Labor Statistics (BLS). 2 The baseline analysis yields predicted wages in each labor market, adjusted for regional differences in worker characteristics and the mix of industries and occupations in each location. As such, the NCES CWI does not indicate that the wage level is low in an area simply because most of the workers are young and inexperienced, nor does it indicate that the wage level is low in an area simply because there are a disproportionate number of low-skill jobs. Rather, the NCES CWI isolates the regional variation in wages that is attributable specifically to differences in location. The labor markets in the NCES CWI are based on place-of-work areas as defined by the Census Bureau for the 2000 Census. Census place-of-work areas are geographic regions designed to contain at least 100,000 persons. The place-of-work areas do not cross state boundaries and generally follow the boundaries of county groups, single counties, or censusdefined places (Ruggles et al. 2003). Counties in sparsely-populated parts of a state are clustered together into a single Census place-of-work area. Each labor market in the NCES CWI is either a single place of work, or a cluster of the places-of-work that comprise a metropolitan area. All local communities in the United States are part of a place-of-work area. There are 16 NCES CWI labor markets in the state of Washington. Nine correspond to metropolitan areas Bellingham, Bremerton, Kennewick, Olympia, Portland, Seattle, Spokane, Page 2

12 Tacoma, and Yakima while seven represent clusters of rural counties. Each Washington school district is associated with one of the 16 labor market areas. Figure 1 illustrates the baseline CWI for the labor market areas in Washington. Figure 1: The Baseline CWI CWI Source: Taylor and Fowler (2006). Taylor and Fowler (2006) used data from the Bureau of Labor Statistics Occupational Employment Survey (OES) to extend the baseline estimates of the NCES CWI and provide annual index values for 1997 through The OES is a BLS database that contains average annual earnings by occupation for states and metropolitan areas. Each year, the BLS samples and contacts approximately 400,000 civilian, nonfarm establishments for the OES survey. 3 Survey respondents each year employ more than 70 percent of the civilian, nonfarm workers in the United States. When extending the baseline CWI, Taylor and Fowler used the OES data to estimate an occupationally adjusted wage in each labor market area. 4 They then adjusted the baseline NCES CWI to reflect annual growth in those wage estimates for each location. For example, if their analysis of the OES data indicated that the wage level in Seattle increased by 5 percent between 1999 and 2001, they revised the baseline CWI for Seattle upward by 5 percent to generate an estimate of the Seattle CWI in Following the same methodology as in that earlier work, I have updated the NCES CWI to cover the five years from 2006 through Thus, I have used OES data for each year from 2006 through 2010 to estimate the occupationally adjusted wage level in each state and major Page 3

13 metropolitan area in the United States. Using those estimates, I have also calculated the implied average wage level in the non-metropolitan remainder of each state. I then calculated the annual rate of change in the OES wage estimates and adjusted the baseline CWI accordingly. Table 1 presents the updated values of the NCES CWI for the 16 labor market areas in Washington. 5 As the table illustrates, the wage differences among Washington labor market areas widened slightly between 2005 and Where wages were almost 28 percent higher in Seattle than in rural eastern Washington in 2005, they were 30 percent higher in On average, wages for college graduates in Washington increased 2.6 percent per year over the fiveyear period. The slowest wage growth was in rural Washington, where the wage level increased by 2.3 percent each year. The most rapid wage growth was in Bremerton, where the wage level increased by 3.2 percent per year between 2005 and Table 1: Comparable Wage Index Values and the Prevailing Salary for College Graduates NCES CWI 2005 Updated CWI 2010 Prevailing Salary Level for College Graduates 2010 Bellingham Metropolitan Area $60,390 Bremerton-Silverdale Metropolitan Area $70,590 Kennewick-Pasco-Richland Metropolitan Area $71,129 Olympia Metropolitan Area $66,055 Portland-Vancouver-Beaverton Metropolitan Area $65,809 Seattle-Bellevue-Everett Metropolitan Area $72,935 Spokane Metropolitan Area $61,930 Tacoma Metropolitan Area $68,878 Yakima Metropolitan Area $66,652 Island, San Juan and Skagit $60,728 Chelan, Douglas, Kittitas and Okanogan $56,149 Adams, Grant, Ferry, Lincoln, Pend Oreille and Stevens $57,113 Asotin, Columbia, Garfield, Walla Walla and Whitman $56,147 Cowlitz, Klickitat and Wahkiakum $64,179 Grays Harbor, Lewis and Pacific $59,437 Clallam, Jefferson and Mason $57,707 Pupil-weighted state average $67,515 Note: The pupil-weighted state average is calculated using the FTE student counts from the school year. The final column in Table 1 indicates the prevailing annual salary for college graduates implied by the updated CWI. Using the baseline salary analysis from Taylor and Fowler (2006), I predicted the national average full-time equivalent salary in 1999 as the salary that would be Page 4

14 expected for a college graduate who worked 40 hours a week and 52 weeks a year. 6 Multiplying the local CWI by the national average full-time equivalent salary in 1999 ($46,175) yields the 2010 comparable salary for college graduates in each Washington labor market. As the table illustrates, the predicted full-time equivalent 2010 salary for a college graduate in Washington ranged from $72,935 in the Seattle-Bellevue-Everett metropolitan area to less than $47,150 in several rural counties. Updating the High School CWI The NCES CWI measures regional variations in the prevailing wage for college graduates. As such, it is a credible price index for teachers, administrators and other professional staffers. However, school districts also hire a large number of workers such as clerical workers, teacher aides, cafeteria workers or custodians who typically are not college graduates. Arguably, the wages of workers without a college degree may have a different geographic pattern than do the wages of college graduates. If so, then it would be inappropriate to use the CWI to evaluate variations in the wages of classified workers. Instead, a more appropriate price index for classified staff would be a comparable wage index for high school graduates. Following the methodology used for the NCES CWI, Taylor (2010) estimated a baseline comparable wage index for high school graduates (HS CWI) using the Individual Public Use Microdata Sample (IPUMS 5-Percent) from the 2000 U.S. Census. 7 As with the NCES CWI, she extended the HS CWI through 2007 using OES data. This analysis further extends the HS CWI through Table 2 presents the updated values of the HS CWI for the 16 labor market areas in Washington. 8 It also presents the prevailing annual salary for high school graduates implied by the updated HS CWI. Using the baseline salary analysis from Appendix Table A.4, I predicted the national average full-time equivalent salary in 1999 as the salary that would be expected for a high school graduate who worked 40 hours a week and 52 weeks a year. 9 Multiplying the local HS CWI by the national average full-time equivalent salary in 1999 ($27,724) yields the 2010 comparable salary for high school graduates in each Washington labor market. As the table illustrates, the predicted full-time equivalent 2010 salary for a high school graduate in Washington ranged from nearly $46,000 in the Seattle-Bellevue-Everett metropolitan area to less than $36,000 in several rural counties. Page 5

15 Table 2: High School Comparable Wage Index Values and the Prevailing Salary for High School Graduates Prevailing Salary Level for High HS CWI 2005 HS CWI 2010 School Graduates 2010 Bellingham Metropolitan Area $38,926 Bremerton-Silverdale Metropolitan Area $43,789 Kennewick-Pasco-Richland Metropolitan Area $40,197 Olympia Metropolitan Area $40,921 Portland-Vancouver-Beaverton Metropolitan $40,771 Area Seattle-Bellevue-Everett Metropolitan Area $45,956 Spokane Metropolitan Area $37,321 Tacoma Metropolitan Area $42,501 Yakima Metropolitan Area $39,018 Island, San Juan and Skagit $39,260 Chelan, Douglas, Kittitas and Okanogan $36,644 Adams, Grant, Ferry, Lincoln, Pend Oreille $35,849 and Stevens Asotin, Columbia, Garfield, Walla Walla and $35,505 Whitman Cowlitz, Klickitat and Wahkiakum $39,286 Grays Harbor, Lewis and Pacific $38,164 Clallam, Jefferson and Mason $38,350 Pupil-weighted state average $41,958 Note: The pupil-weighted state average is calculated using the FTE student counts from the school year. Figure 2 compares the 2010 CWI and HS CWI for the 16 labor markets in Washington. The figure indicates the percentage difference in wages between the local wage level and the pupil-weighted state average wage level for each type of labor. As the table illustrates, the prevailing wage for college graduates does not track perfectly with the prevailing wage for high school graduates. In particular, wages in the Kennewick-Pasco-Richland Metropolitan Area are 5 percent above the state average for college graduates, but 4 percent below the state average for high school graduates. However, both the CWI and the HS CWI indicate that hiring costs in the Seattle metropolitan area are more than 29 percent greater than hiring costs in the lowest-cost Washington counties. Page 6

16 Figure 2: Comparing the High School CWI with the Updated CWI, 2010 Seattle-Bellevue-Everett Metropolitan Area Kennewick-Richland-Pasco Metropolitan Area Bremerton-Silverdale Metropolitan Area Tacoma Metropolitan Area Yakima Metropolitan Area Olympia Metropolitan Area Portland-Vancouver-Beaverton Metropolitan Area Cowlitz, Klickitat and Wahkiakum Spokane Metropolitan Area Island, San Juan and Skagit Bellingham Metropolitan Area Grays Harbor, Lewis and Pacific Clallam, Jefferson and Mason Adams, Grant, Ferry, Lincoln, Pend Oreille and Stevens CWI HS CWI Chelan, Douglas, Kittitas and Okanogan Asotin, Columbia, Garfield, Walla Walla and Whitman 20% 15% 10% 5% 0% 5% 10% 15% Percentage Deviation from State Average, 2010 Source: Author s calculations using Ruggles et al. (2003) and the Occupational Employment Survey. The wage differentials indicated by the NCES CWI and HS CWI are large, but they are dwarfed by the differences in the cost of housing. According to the U.S. Department of Housing and Urban Development (HUD), the fair market rent for a two-bedroom apartment in Seattle was $1,056 per month in 2010, while the fair market rent for a comparable two-bedroom apartment in Adams County was only $637. Because housing costs are the primary determinants of the cost of living, the HUD data suggest that the cost of living in Seattle is nearly 67 percent higher than the cost of living in some other parts of Washington. A smaller variation in wages than in housing costs implies that the relatively high-rent parts of the state must also boast local amenities that make people willing to accept a lower real wage than they would otherwise require. In other words, the attractions of Seattle make people willing to accept salaries that are not high enough to fully offset the higher cost of housing. Estimating Occupation-Specific Wage Levels The CWI and HS-CWI describe the general wage level for college graduates and high school graduates, respectively. Arguably, additional information could be gleaned from directly comparing some school district positions such as nurses, audiologists or food service Page 7

17 workers with their corresponding occupations outside of education. Therefore, I also used the CWI methodology to predict the wage level for an array of certified and classified occupations. For each certified occupation in Table 3, I used the baseline regression model underlying the CWI to predict the prevailing salary for that occupation in The prevailing salary in each labor market is the salary that would be expected for a college graduate who had the same educational and industrial profile as the average Census respondent in that occupational category, assuming that the person worked 40 hours a week and 52 weeks a year. 10 I then used the OES data to calculate the occupation-specific wage level in each location for 1999 and Finally, I calculated the rate of growth in occupation-specific wages for each location and adjusted the baseline prevailing salaries accordingly. For example, I used the CWI regression model to predict that the prevailing salary for registered nurses in Seattle was $48,002 in Using the OES data, I calculated that, on average, nurses salaries in Seattle rose percent between 1999 and Therefore, I calculated that the prevailing salary for nurses in Seattle in 2010 was $73,708. By construction, these occupation-specific salaries have the same geographic pattern as the CWI. Thus, the prevailing salary level for each labor-market area (and each corresponding school district) can be calculated by multiplying the state average prevailing salaries in Table 3 by the local CWI in Table 1, and then dividing by the pupil-weighted state average CWI (1.462). Table 3: Prevailing Annual Salaries for College Graduates, by Occupation State Average Predicted Salary 2010 All College Graduates $67,515 ACM Teacher-Comparable Occupations $65,923 STEM Occupations $76,051 Registered Nurses $68,231 Social Workers $47,421 Psychologists $61,681 Counselors $43,606 Occupational Therapists $73,529 Physical Therapists $73,251 Speech And Language Pathologists $70,223 Audiologists $71,363 Selected Managerial Occupations $92,704 All Managerial Occupations Except Legislators $88,900 Note: The state average predicted salary is a pupil-weighted average of the salary predictions for each school district. The pupil-weighted state average is calculated using the FTE student counts from the school year. Most of the occupations in Table 3 are self-explanatory. One exception is the set of teacher-comparable occupations identified by Allegretto, Corcoran, and Mishel (2004). ACM identified 16 occupations in the Current Population Survey (CPS) that were particularly comparable to teaching on the basis of an evaluation of the skills required to do the job. Page 8

18 Adjusted for differences in occupation codes between the Census and the CPS, those occupations are: Accountants and auditors; Insurance underwriters; Human resources, training, and labor relations specialists; Inspectors, testers, sorters, samplers, and weighers; Agricultural inspectors; Architects, except naval; Conservation scientists and foresters; Registered nurses; Occupational therapists; Physical therapists; Post-secondary teachers; Counselors; Clergy; Technical writers; Editors; News analysts, reporters, and correspondents; and Computer programmers. 12 Using the same strategy as in the calculation of nurses salaries, I calculate that the state-average predicted salary for workers in these occupations was $65,923 in The STEM occupations include all science, technology, engineering and mathematics occupations (Census occupation codes ). I calculate that the state average predicted salary for workers in these occupations was $76,051 in Another occupational category that is not self-explanatory is selected managerial occupations. This category contains occupations identified by Washington s Employment Security Department (ESD) as particularly comparable to certified, school-level administrators based on an evaluation of the work context and the knowledge, skills and abilities required to do the job. Appendix Table A.7 identifies these selected managerial occupations and reports each occupation s weight in the construction of the salary predictions. The weights correspond to the frequencies with which Census respondent with a college degree held each occupation in I calculate that the state-average predicted salary for workers in these selected managerial occupations was $92,704 in Finally, the occupational category all managerial occupations except legislators includes all occupations designated as management occupations by the 2000 Census (i.e. all occupation codes below 50) except for legislators (occupation code 3) plus all of the occupations included in the list of selected managerial occupations. Appendix Table A.7 identifies these managerial occupations as well, and reports each occupation s weight in the construction of the salary predictions. Again, the weights correspond to the frequencies with which Census respondent with a college degree held each occupation in I calculate that the state-average predicted salary for workers in this broader set of managerial occupations was $88,900 in For each classified occupation in Table 4, I used the baseline regression model underlying the HS-CWI to predict the prevailing salary for that occupation in The prevailing salary in each labor market is the salary that would be expected for a high school graduate who had the same educational and industrial profile as the average, high-schooleducated Census respondent in that occupational category, assuming that the person worked 40 hours a week and 52 weeks a year. 15 As with the occupation-specific wage estimates for college graduates, I then used the OES data to calculate the occupation-specific wage level in each location for 1999 and 2010, calculated the rate of growth in wages for each location, and adjusted the baseline prevailing salaries accordingly. Page 9

19 Table 4: Prevailing Annual Salaries for High School Graduates, by Occupation State Average Predicted Salary 2010 All High School Graduates $41,958 Supervisors $45,943 Office And Administrative Support Occupations $36,344 Janitorial Occupations $30,353 Protective Service Occupations, Excluding Fire Safety Personnel $48,221 Information Technology Occupations $60,901 Facilities, Maintenance and Grounds $45,059 Warehouse, Laborers and Mechanics $42,572 Motor Vehicle Operators $38,039 Food Preparation and Serving Occupations $28,754 Note: The state average predicted salary is a pupil-weighted average of the salary predictions for each school district. The pupil-weighted state average is calculated using the FTE student counts from the school year. Again, many of the occupations in Table 4 are largely self-explanatory. The occupational group Supervisors includes all non-military occupations with the term supervisor in the title. The groups Office and Administrative Support Occupations, Food Preparation and Serving Occupations and Protective Service Occupations, Excluding Fire Safety Personnel include all Census occupations in the corresponding Census groups except for first-line supervisors. The group Janitorial Occupations includes two Census occupations Janitors and Building Cleaners and Maids and Housekeeping Cleaners. The group Information Technology Occupations includes all Census designated computer occupations (i.e. occupation codes 100 through 119) and the group Motor Vehicle Operators includes all Census designated drivers of motor vehicles (i.e. occupation codes 911 through 915). Appendix Tables A.8a through A.8e report each occupation s weight in the construction of the salary predictions for these occupational groups. The weights correspond to the frequencies with which Census respondent with a high school diploma degree held each occupation in The two remaining occupational groups Facilities, Maintenance and Grounds, and Warehouse, Laborers and Mechanics contain occupations identified by ESD as particularly comparable to school district personnel in the respective categories based on an evaluation of the work context and the knowledge, skills and abilities required to do the job. Appendix Table A.8f reports each occupation s weight in the construction of the salary predictions for each occupational group. The Prevailing Salaries for School District Personnel This analysis relies on a series of hedonic wage models to estimate the prevailing salaries for educators throughout the state of Washington. Hedonic wage models use regression analysis to decompose the observed variation in salaries into that which is attributable to worker characteristics, that which is attributable to working conditions and that which is attributable to locational characteristics. 16 Page 10

20 The hedonic salary model for Washington educators describes each educator s salary as a function of his or her personal characteristics, his or her job assignments, and the school building and school district in which he or she works. I use this model to predict the average full-timeequivalent salary in each school district, holding constant the influence of demographic and job characteristics. Those predictions indicate the prevailing salaries in the school district. Variations in the prevailing salaries reflect how much more or less each school district pays to recruit and retain comparable school personnel. The prevailing salary for a labor market is just a weighted average of the prevailing salaries in its constituent school districts. I estimated separate hedonic wage models for three sets of Washington teachers. The first model includes data for all types of teachers, and supports estimates of the prevailing wage for all types of teachers. The second model includes only data on beginning teachers, while the third model includes only data on teachers who are certified in math and science. I also estimated hedonic wage models for a variety of other certified and classified personnel, including school administrators, teacher aides and food service workers. The analyses of certified personnel largely mirror that of teachers. The analyses of classified personnel have been simplified somewhat because data on worker characteristics are more limited. Data and Estimation Data for this part of the analysis were provided by the Office of Superintendent for Public Instruction (OSPI). Data on earnings, worker characteristics and job assignments were drawn from the OSPI s S-275 files for the eight school years from through Data on teacher certification and endorsements come from OSPI s teacher certification files. The OSPI data provide two alternative salary measures base salaries and final total salaries. Base salaries measure employee earnings during the school year under terms of the base employment contract and paid by the state. 17 Final total salaries represent any and all earnings of school district personnel; in other words, the final gross pay of each employee from all sources, including the state, local levies, and federal monies. During , final total salaries for teachers exceeded base salaries by an average of $11,250, or 21 percent. For analysis purposes, I have transformed the standard OSPI salary measures. For the analyses of certificated personnel, I calculate the full-time equivalent base salary and full-time equivalent total salary, excluding pay for extracurricular and public activities and adjusting for differences in time worked using information on the educator s FTE. 18 The annualized total salary, which I use for the analyses of classified personnel, also excludes pay for extracurricular and public activities. It has been adjusted for time worked using the worker s calculated hourly rate of pay. 19 Following the method used in OSPI s school district personnel summary report, annualized salary is defined as the hourly salary times 2,080 hours (40 hours a week, 52 weeks a year). The personal and job assignment characteristics included in this analysis are outlined in Table 4. Most are self-explanatory, but a few require a bit of additional explanation. Page 11

21 Table 5: Explanatory Factors from the Hedonic Wage Model for Washington Educators Personal Characteristics Years of experience Highest degree held (BA, MA, PhD, H, G,V,S) Credit hours (academic, non-credit, in-service, other) Gender Ethnicity (Asian, Black, Hispanic, American Indian, Pacific Islander, Mixed, White) Continuing employee Certification endorsements Math English/Reading Science Bilingual/ESL Administration Health and Physical Education Elementary Early Childhood History Foreign Language Arts Social Science Special education Emergency certification Certificated Activities Board of Directors Superintendent s Office Business Office Human Resources Public Relations Supervision (Instruction) Learning Resources Principal s Office Guidance and Counseling Pupil Management and Safety Health Related Services Extracurricular Information Systems Teaching Classified Activities Food Services General Supervision Operations (Pupil Transportation) Maintenance (Pupil Transportation) Grounds Maintenance Operation of Buildings Maintenance Utilities Building and Property Security Information Systems Printing Warehousing and Distribution Motor Pool Public Activities Programs Special education Limited English proficiency Compensatory education Vocational education Community service Support Other Assignment Characteristics Percent FTE in teaching Duty code classification (elementary teacher, secondary teacher, elementary principal, audiologist, laborer, etcetera) Page 12

22 The highest degree held is one of the usual suspects bachelor s degree (BA), master s degree (MA) or doctorate (Ph.D.) plus a number of degree identifiers specific to the state of Washington. According to the S-275 Personnel Reporting Handbook, a highest degree of H indicates a person who obtained a bachelor s degree while employed in the state of Washington as a non-degreed vocational/career and technical education instructor. A highest degree of G indicates a person with a grandfathered bachelor s degree. 20 A highest degree of V indicates someone without a bachelor s degree who holds appropriate vocational or career and technical education certificates. Finally, a highest degree of S indicates someone without a bachelor s degree who is in special circumstances. Among the teachers under analysis for the school year, less than 1 percent had a highest degree of H, V or S. Most held either a master s degree (62 percent) or a bachelor s degree (34 percent). The analysis includes an indicator for whether or not the individual is a continuing employee. Recent analyses in Texas, New Mexico, and Wyoming indicate that new hires tend to earn less than employees with similar characteristics and assignments. The activity factors indicate whether or not the individual was assigned to the specific activity. Any person could have one or more activity assignments. Similarly, the program factors indicate whether or not the individual was assigned responsibilities in the designated educational program. Again, any person could have one or more program assignments. The certification endorsements are only used in the hedonic models for teachers and administrators. They indicate whether or not the educator is certified in the designated subject areas. In each case, an educator was considered certified if he or she held a current endorsement in the field. To be considered current for a particular academic year, a certificate must have been issued before September 1 st of the academic year, and must not have expired before September 1 st of that year. Any educator could hold endorsements in one or more of the designated fields. The Prevailing Salaries for Teachers To estimate the prevailing teacher salary in each NCES labor market and school district, I applied the hedonic salary model described above to data on the earnings and other characteristics of the teachers who taught at least half time in a Washington public school district during the school year. Complete data were available for 54,945 individual teachers from 295 school districts. 21 Appendix table A.9 presents the coefficient estimates and standard errors from two separate salary models, one for full-time equivalent base salaries and the other for full-time-equivalent total salaries. The hedonic models do a good job of capturing variations in teaching salaries. As expected, salaries increase with experience and educational attainment. Teachers with supervisory assignments earn nothing more in base salary but 6 percent more in total salary, on average, than other teachers. Individuals who teach in the compensatory education, bilingual education or vocational education programs earn more total salary than other teachers, all other things being equal. Teachers with an emergency certification earn 2 percent more than their peers, all other things being equal. All told, the model explains 95 percent of the variation in Page 13