Analysis of Wyoming State Government Attrition for Selected Occupations
Legislative Services Office, Program Evaluation Section
pursuant to MOU with
Department of Employment, Research & Planning Section
Research & Planning Staff1
10 March 2000
There is a continuous flow of population and labor into and out of Wyoming. As required by state statute, the vast majority of both the stock and flow of labor are reported to the Wyoming Department of Employment (DOE) by employers on their quarterly Unemployment Insurance (UI) tax forms. This analysis uses those quarterly reports and other administrative databases 2 to describe the earnings and work experiences of employees found in a selection of four State of Wyoming occupations. Our study examines the labor market experiences of case workers for the Department of Family Services (DFS), corrections officers with the Department of Corrections (DOC), highway patrol officers with the Department of Transportation (DOT), and technology staff (IT). Data regarding dates of employment for these occupations were provided to DOE, Research & Planning, merged with UI Wage Records (WR) earnings and demographic data bases to produce the following analysis.
All four occupational classifications demonstrated above normal attrition rates with out-migration from the state as the most likely explanation.
The highest attrition rate--for former IT employees--suggests a much higher rate of out-migration.
For 38 former DFS employees, average wages two quarters after their exit represented a 20.1 percent decrease. These case workers' wages may or may not have been the primary impetus for leaving their jobs.
The analysis of highway patrol officers and technology staff show that on average those who exited their jobs and stayed in Wyoming to work in the private sector increased their wages.
Corrections officers have an increased likelihood of holding a secondary job. Highway patrol officers demonstrated the highest level of multiple-job holding.
For all four occupational groups, multiple-job holding rates within the normal range of labor attachment for the state as a whole.
Multiple-job holding by state employees is an indicator associated with the greater likelihood that these employees may exit state employment.
Employee Strategies and Labor Market Flow
For any given stock of employed workers at any given point in time, it is normal for that stock of workers to diminish, or experience attrition, over time. Attrition occurs due to worker death, injury, retirement, withdrawal to meet the needs of young or sick family members, new job opportunities in another state, and for other reasons. During the period of study, 1995-1998, all four occupational classifications demonstrated above normal attrition rates with out-migration from the state as the most likely explanation.
In the first quarter after exit from their state jobs only 50 percent of DFS staff were found working in Wyoming. This contrasts markedly with the normal attrition rate. The rate at which all workers in the 25-34 year age group (the interval containing the mean age of former employees for DOC, DOT, and many IT occupations) exit the Wyoming wage records (WR) file is 9.9 percent with 90.1 percent still working after one quarter. Therefore, the norm indicates that on average, among all occupations, only about 10 percent of employees of this age group fail to appear on WR in the following quarters (see Figure 1). Our interpretation is that the difference between 90.1 percent and 50 percent retention rate for DFS employees probably represents out-migration. The highest attrition rate difference among our selection of occupations--between 90.1 percent on the WR file and 33.7 percent for former IT employees--suggests a much higher rate of out-migration. Wage differentials for IT occupations between Wyoming and other states are substantial. For example, the wage for public and private IT systems analysts [Occupational Employment Statistics (OES) code 25102] was $19.53 per hour in Wyoming in 1998. For computer programmers (OES 25105), the 1998 wage was $16.67. In Utah, the hourly wages for systems analysts and computer programmers were $22.80 and $21.90 respectively, and in Colorado they were even higher (systems analysts, $28.71/hour; computer programmers, $25.26/hour). See Figure 2 for a broad-based index of wage differentials for all IT occupations.
Labor Market Outcomes for Those Exiting State Employment and Remaining in Wyoming
Tables 1a-1d show time lines of employment and wages for employees who exited each of the selected occupations in our analysis. Table 1a shows that all 76 DFS case workers were found in WR as being employed by the State in the quarter prior to their exit from DFS employment, and they earned on average $6,112 in that quarter. Following their exit, 38 (50%) were distributed throughout Wyoming's labor market among several industries, both public and private. The average private sector earnings for both the first and second quarters, following the exit quarter, represented decreases in wages to $3,947 and $4,263 respectively. These new earnings levels also represent significantly less than the $5,323 paid to the average wage earner with the private sector (fourth column of data).
The use of Wage Records permits us to calculate average wages per job by quarter to compare state wages to private sector earnings.3 For those 38 former DFS employees who were found in WR, average wages two quarters after their exit as case workers represented a 20.1 percent decrease. For these case workers, wages may or may not have been the primary impetus for leaving their jobs.4
Table 1b provides a useful snapshot of the labor flow dynamic within the Department of Corrections. Twenty (20 or 9.8%) of the 204 corrections officers who exited from 1995 to 1998 were not found working anywhere in Wyoming the quarter prior to their exit quarter of employment from DOC. Also, these 20 worked for the DOC as corrections officers for less than three months. Like DFS case workers, only about half (54%, see Figure 1) reappeared in WR the quarter following exit from DOC. Those who found jobs in the private sector (earning on average $3,629 two quarters following their exit) or in government (earning $4,545 two quarters after) earned on average less than the $5,307 they earned as corrections officers. On average, their wages decreased substantially (-26.1%), indicating that circumstances and motivations other than salary probably led to the decision to leave their jobs.
By contrast, the analysis of highway patrol officers (see Table 1c) and technology staff (see Table 1d) show that on average those who exited their jobs and stayed in Wyoming to work in the private sector increased their wages each of the two quarters following their exit. For five (5) former highway patrol officers, private sector wages increased to an average of $8,026 per quarter. This wage level is about $2,700 above the average quarterly wage per job for all private sector workers in 1998 Quarter 4 ($5,323). It also represents an increase from the $7,397 earned quarterly as a state employee. Of the 12 who took another government job, quarterly wages decreased slightly to an average of $6,798. We cannot explain, from the limited data available to us, why highway patrol officers might leave state employment and subsequently return to it (perhaps, a state agency other than DOT), and assume new jobs with different duties for less earnings.
Among IT staff, 17 of 92 moved to private sector employment in Wyoming, on average earning quarterly wages of $7,887, above the $7,117 they earned the quarter prior to their exit. Those 14 who moved to other positions within Wyoming government earned slightly less than their former quarterly wages ($6,850). The 28 who were found in WR two quarters following their exit earned on average 3.5 percent more than they did when employed as IT technology staff in state government.
Employment Strategies of Those Who Exit During 1995-98: Multiple-Job Holding and Secondary Employers
In past editions of Wyoming Labor Force Trends, Research & Planning has published and applied a classification system to labor market research characterizing varying degrees of labor force attachment exhibited in wage records.5 Multiple-job holders, one category within this classification system, are defined as those employees found in wage records who worked for three or more employers in the same quarter or who worked for the same two employers for each of two consecutive quarters or more. Other categories use incidences of secondary employment on an itinerant or occasional basis which are often associated with employees' flow in and out of the labor market--or market churning (e.g., the 20 DOC staff not employed before working in DOC and remaining with DOC for less than 3 months--Table 1b). Tables 5a-5d show, for each occupation in our analysis, the incidences of state employees meeting the definition of multiple-job holder, during the year of exit, and provide the industrial sector where they held the multiple job. Our premise for studying this data is to determine whether or not there is a relationship between multiple-job holding and state government attrition for a given occupation.
The multiple-job holding rate for all state and local government workers in 1998 was 14.9 percent.6 For DFS and IT staff (Tables 5a and 5d), the data show normal multiple-job holding patterns of 14.5 percent and 14.1 percent respectively. Table 5b shows that corrections officers have an increased likelihood of holding a secondary job (19.6%). Highway patrol officers demonstrated the highest level of multiple-job holding (Table 5c), with 21.6 percent opting to work a second job. Multiple-job holding can be viewed as a strategy employed by state workers to augment their primary earnings. Of course, multiple-job holding may complicate one's domestic life.
Based on our published data, holding a secondary job (e.g., working in the private sector in the first quarter and then working the third and fourth quarters as a state employee) is normal for 22.6 percent of all state and local government workers. However, whether for reasons associated with exit (a form of market churning) or other reasons, three of the occupational exit groups displayed a higher than normal rate of holding secondary jobs (though not a multiple-job holder). DFS case workers held secondary jobs at a rate of 35.5 percent (see Table 5a). DOC corrections officers held secondary jobs at a rate of 35.3 percent (see Table 5b). DOT highway patrol officers who exited employment held secondary jobs at a rate only slightly higher than the norm, 24.3 percent (see Table 5c). Tables 6a-6d are crosstabulations of multiple-job holding, but only for those employees in our analysis who exited state employment in 1998. Due to the small numbers of employees who exited in this single year, the data were not generally useful in validating the continuing patterns in multiple-job holding.7
Although the data in Tables 6a, 6c and 6d are inconclusive, it appears that for corrections officers (Table 6b) multiple-job holding is not a constant but rather changes over time. Multiple-job holding increases to 26.3 percent in the year of exit in 1998 compared to 19.6 percent for the period 1995-98 (Table 5b). Without data by occupation for all of state government over time, we cannot determine the extent to which multiple-job holding represents occupation specific behavior. Without longitudinal data, we cannot determine how extensively multiple-job holding may be considered a function of change in general economic conditions. For example, in 1990, Wyoming ranked 37th among the 50 states in competitive wages.8 By 1996, our competitive wage ranking fell to 45. In 1998, Wyoming still ranked 45th in the nation. Only by placing occupational wages within the broader economic context can we begin to understand the labor market strategies of employees.
As an example of how the national market for technology staff affect employment attrition rates at IT, Figure 2 shows the distribution of average annual wages of technology staff, both public and private sector for all states. The data were based on weighted averages of Occupational Employment Statistics (OES) survey data for 1996, fourth quarter. 9 The mean annual wage was $40,793, with a standard deviation of $5,707. The average annual wage for technology staff employed by the state of Wyoming was $28,468. The average annual wage for all technology staff working in both the public and private sectors in Wyoming was $31,366.
Tables 7a-7d are very useful in showing the contrast in patterns of multiple-job holding and holding secondary jobs between those who were actively employed in the respective agencies during 1998. For all four occupational groups, multiple-job holding and the holding of secondary jobs by state employees are within the normal range of labor attachment for the state as a whole. The comparison of these data with Tables 5a-5d indicate that multiple-job holding and the holding of secondary jobs on an itinerant basis by state employees are indicators associated with the greater likelihood that these employees may exit state employment, whether it be for financial reasons or driven by other circumstances.
Demographic Comparisons by Occupation of Those Who Exited State Government 1995-1998 with the Active Occupational Workforce 1998
Tables 3a-3d illustrate the distribution by age and gender of those exiting the respective occupations in this analysis. Characterizing those actively holding similar positions with the various agencies, demographic data in Tables 4a-4d permit us to identify the differences between those who work at a given occupation in state government and those who exit their job.
For case workers at DFS (Tables 3a and 4a), exiting employees are generally 4.2 years younger than our snapshot of the active 1998 DFS workforce; they have a mean age of 35.4 years compared to 39.6 years. In Table 4a, we can see that at least 73.8 percent of case workers were women (for some employees we did not have demographic data). Comparing the gender data of the two tables, we can see that male case workers represent 27.6 percent of those exiting DFS, but only 20.7 percent of the active workforce. This finding underscores the general observation confirmed by much of our published information that young males constitute a highly mobile segment of the labor market. Also, because 51.3 percent of employee exits are under the age of 35, family considerations (e.g., births, day care availability, education, employment of a spouse or other family member) probably weigh heavily in the decision to leave state employment.
An analysis of exits among correctional officers at DOC show a gap of 5.5 years between the mean age of those employed in 1998 (Table 4b; 37.6 years) and those who exited employment (Table 3b; 32.1 years). In contrast to DFS data, DOC data shows that in 1998, 75.3 percent of correctional officers were men. The proportion of males and females exiting DOC employment largely reflected the gender distribution within the occupation. Similarly to DFS, more than half of DOC exits fell among two age groups, those under 25 (16.7%) and those 25-34 (39.7%).
DOT highway patrol officers (see Tables 3c and 4c) who exit have a mean age of 29.7 years, compared to the 1998 DOT occupational mean age of 40.9 years for those in the active work force. Highway patrol officers are primarily male (89.2% in 1998; Table 4d). From the data, we cannot confirm whether male officers are more likely to leave employment than female officers. Interestingly, highway patrol officers who exited employment during 1995-98 were very evenly distributed among three age categories, under 35 (11 employees), 35-44 (10 employees) and 45-up (8 employees). The difference in mean age between exiting officers and those actively employed in 1998 (40.9 - 29.7 = 11.2 years) shows that young male officers are probably more likely to exit than older officers who are likely to have more years invested with the State.
Analysis of Table 4d shows that IT technology staff in 1998 had an employment ratio of 61 percent male to 29.9 percent female (no demographic data were available for 9.1%). Between 1995-98, female technology staff exited IT at a rate of 40.2 percent compared to the male exit rate of 48.9 percent. Thus, as a percentage of the IT workforce, younger women are probably more likely to leave state employment. The difference in mean age (8.5 years) between those exiting IT during 1995-98 (34.8 years) and those actively employed in 1998 (43.4 years) probably reflects the national market for technology staff and the likelihood that younger people are more likely to respond to it than older workers who, generally, have more family and community attachments.
Since the state is the primary purchaser (a near monopsony exists within Wyoming) of the services of some of these occupations, such as IT services, the choice is to either pick up the cost at the point of purchase with a highly visible market signal or pick up the cost in the less visible system of recruitment and training and the inefficiency in service to public and private customers associated with turnover--the price is the same, the question is how it is valued and how and who bears it.
Despite the limitations of the administrative records approach to the analysis of market outcomes for state employees who exit employment with the state, it is clear that the approach has promise, especially given the pending confidentiality agreements between states that will permit market based tracing of former workers for statistical purposes.
While it is possible to place occupational exit analysis in the context of such broader issues as the "brain drain"10 and historic trends in the level of turnover by industrial sector,11 the time constraints imposed upon the study did not permit more comprehensive analysis.
In addition, since the state lacks comprehensive analysis of exit behavior in general, it is difficult to determine how extensively the behavior of the four occupations is unique within state government, to what extent other occupations are beginning to exhibit comparable patterns of exit, or to what extent pay or other policies are effective in enhancing employee retention and cost control. Absent a comprehensive, ongoing program of market based occupational analysis, the value of narrow, episodic and retrospective analysis is questionable.
All of the occupations in this study require some post-high school training. We cannot account for the migration decisions of families when the spouse may be unable to use his or her post-high school education or training in finding suitable work.
Contextual variables for the family (e.g., the availability and cost of day care) are not available to us, nor are non-monetary costs and benefits of the particular type of work part of this study's components. Given that case workers average annual wages during the study period are 173 percent above the poverty level for a family of three ($24,448 compared to $14,150), for example, it may be logical for a DFS case worker to care for children at home rather than work and pay for day care.12
Finally, additional limitations of the administrative data approach to analysis are discussed at length in the publication cited in end note 2.
1 Tom Gallagher, Manager; Craig Henderson, One Stop Program Supervisor; Tony Glover, Senior Analyst; Norman Baron, Economist.
2 Wyoming Department of Employment. Research & Planning. Wyoming Wage Records 1992-1998: a Baseline Study, November 1999. This publication contains a comprehensive bibliography of all publications and articles published by the agency that relied on wage records research, principally articles published in Wyoming Labor Force Trends.
3 Wyoming Department of Employment. Research & Planning. Wyoming Wage Survey 1998. (10 March 2000).
4 At this time, Research & Planning does not have interstate agreements in place with other states to track workers earnings and employment or commuting patterns across state lines using wage records. Future analysis of this type will permit a more complete interpretation of employees' labor market strategies.
5 Wyoming Department of Employment. Research & Planning. Wyoming Wage Records 1992-1998: a Baseline Study, November 1999.
6 Ibid., p. 64.
7 These tables were run as part of our contractual agreement with the Wyoming Legislative Service Office.
8 Carol Kjar, " Competitive Wage Ranking: Retaining Wyoming's Workforce," Wyoming Labor Force Trends, March 2000.
9 The weighted average was based on the following OES occupational codes: 25102, 25103, 25104, 25105,25108, 25111, 25199.
10 Steven Butler, Tracking University of Wyoming Graduates into the Wyoming Work-force: a report prepared for the Research & Planning Section of the Employment Resources Division, State of Wyoming (September 17 1995). See also Wyoming Department of Employment. Research & Planning. Under the Lamppost: a Report to the Wyoming Workforce Development Council, November 1998.
11 Mike Evans, " Job Turnover and Hire Rates in Wyoming: Which is Greater: Job Creation or Job Destruction?," Wyoming Labor Force Trends, June 1999.
12 Tony Glover, "The Flow of Labor in Wyoming: Department of Family Services, Division of Vocational Rehabilitation and Job Training Partnership Act Clients," Wyoming Labor Force Trends, March 2000.
On February 29, 2000, We received four files from Legislative Service Office (LSO) that contained Social Security numbers (SSN) of individuals that had terminated their employment from certain occupations within state government. These four files contained the following information:
Effective Date of Termination
An examination of the data contained in these files revealed the following:
Termination dates outside the date range specified in the MOU with LSO
SSNs appearing more than once in the file
LSO was contacted and informed about the inconsistencies in the data. An agreement was reached on how to deal with the duplicate SSNs and other inconsistencies that were found by R&P and LSO. The final files contained the following number of Records:
Department of Family Services 76
Department of Corrections 204
Department of Transportation 37
Information Technology 92
These files were then matched by SSN to both Quarterly and Yearly Unemployment Insurance (UI) Wage Record files based on the individuals termination date. Information on industry and county of employment was linked to these files from the Quarterly Unemployment Insurance (QUI) files. QUI files contain information about employers were as Wage Record files contain information on the individual. Demographic information was obtained by matching these files to R&P's master demographic file on SSN. The master demographic file is a combination of demographic information from several files including the following:
Employment Service files
UI Claims records
Last modified on August 10, 2001 by Valerie A. Davis.