The Flow of Labor In Wyoming:
Department of Family Services, Division of Vocational Rehabilitation and Job Training Partnership Act Clients
by: Tony Glover , Senior Analyst

". . . placing clients with employers who have a solid record of employee retention in higher paying industries should become a higher priority than merely getting them jobs."

The Workforce Investment Act¹ specifies required and additional partners for a variety of employment and training services to job seekers, employers and others into One Stop delivery systems. The One Stop service delivery system concept includes performance measurement based on objective quantifiable data. Performance measurement, in turn, is used by the federal government to provide incentives to states for good performance, sanction others and report to Congress on the effectiveness of human resource programs.

No one, however, has examined how performance measurement might look in Wyoming across programs and begun asking questions about how we might use performance measurement to improve services to clients. This article compares the performance of three programs: Department of Family Services (DFS), Division of Vocational Rehabilitation (DVR) and Job Training Partnership Act (JTPA). DFS works to reduce dependence on public welfare while developing and strengthening skills within families and maintaining safety for children. DVR's mission is to advance opportunities for persons with disabilities in Wyoming to be employed and independent. JTPA serves individuals facing serious barriers to employment by providing job training and other services. Our objective is to begin the discussion about how performance data can be used as a program management tool.

The findings of this study conclude DFS, DVR and JTPA clients' labor market interactions are extremely unstable (see related article "The Instability Index as a Measure of Labor Market Activity" when compared to the control groups with regards to maintaining a consistent attachment to an employer before and after program involvement. While the client and control groups with employment display a similar natural attrition from the work force prior to and following their program year, there is a mass entry and exit from Wyoming's covered employment of the clients in the year of program completion. Client groups earn 50.0 percent less annually in the two years following program exit, creating a situation in which two of the three programs' annual wages fall below the poverty level for a family of one.²

Data supplied by DFS, DVR and JTPA provided the base for this analysis. Although many years of program data are available, only groups with exits occurring in 1996 are considered. Using 1996 as the study year allows us to examine longer-term program effects. In this situation we focus on one and two year follow up periods.

Client and Control Groups for the Analysis

A client's eligibility for the analysis group relied on several factors. First, the client's demographic information was needed to match to a control group. Second, the client must have appeared in Wage Records³ with earnings in the quarter of exit. This last criterion helps define a group of clients that worked at least part of the quarter in which they exited the program and are considered successful outcomes under the current Workforce Investment Act performance standards.

Table 1 shows that the characteristics of the client groups used to select the matched control groups included age group, gender and industrial sector of wages in the exit quarter. This was done because differences between groups on employment outcomes (i.e., percent appearing in Wage Records, earnings and labor force stability) could be a function of age differences between the groups. Using a control group with the same age distribution as the client group allows us to discard age as a competing factor explaining employment outcomes. Refer to the article for more information on control group selection.

A long list of factors that influence employment outcomes would include disability, educational level, family size, motivation and barriers to employment. These factors are currently not available for control group selection. However, the variables chosen have a strong influence on an individual's labor market outcome. A stratified random sample was selected corresponding to each program group by quarter. This selection process yielded the twenty-four analysis groups shown in Table 2.

Percent of Client and Control Groups Employed 1992 - 1998

Figure 1 shows the percent of DFS clients and the matched control groups with wages in the Wage Records file by quarter from first quarter 1992 (92Q1) to fourth quarter 1998 (98Q4). As mentioned earlier, the quarter of exit is easily discernible because 100.0 percent of the client and control groups were working. The percentage of the four DFS client and four control groups employed are averaged in Figure 2. An additional line appears at the bottom of Figure 2 representing the difference in the percentage between the client and control groups with employment. This process is repeated for the DVR and JTPA groups and Figure 3 is an overlay of the percent with employment for all three client and control groups.

All client groups shown in Figure 3 display similar behavior as to the percent with employment in Wage Records. By separating the graph into three periods (i.e., before year of program exit, year of program exit and after year of program exit) you can see that the greatest increase and decrease in the percent of clients employed is during the year of exit. This indicates that clients enter employment after program participation ends, but do not maintain employment long. There is the possibility that the clients entered non-covered employment subsequent to appearing in the Wage Records file. However, there is an equal likelihood that individuals in the control groups would have done the same; therefore, this does not explain the clients' exodus from Wage Records. In contrasting the periods before and after the year of exit between the client and control groups, it becomes apparent that the same socioeconomic factors influence the client groups as the control groups (i.e., the seasonality of employment opportunities). Moreover, the natural attrition of those clients with employment occurring in the period after the year of exit follows the same pattern as the control group.

Figure 3 tells us about the gross percent of clients with employment prior to and following their exit from a program. However, Figure 3 does not tell us if the 72.0 percent of DFS clients with employment in first quarter 1997 (97Q1) are the same 71.0 percent that appear in second quarter 1997 (97Q2). As discussed in prior Wyoming Labor Force Trends articles, Wyoming's labor market experiences a substantial level of churning/turnover in employment..⁴

Earnings

Figure 4 shows the average quarterly wages of employed individuals (i.e., appearing in Wage Records) across the three client and control groups. The differences in wages between the three control groups are attributable to the differences in the characteristics presented in Table 1. Different demographic groups tend to occupy different niches in the labor market. For example, the control group matched to the DFS clients has predominantly younger (less labor force experience) females working in the Services Producing sector (generally associated with lower wages.)⁵ Whereas, the DVR control group has 50.0 percent of its members over the age of 35, predominantly male and a higher proportion working in the Goods Producing sector.

The client groups display a dip in average quarterly wage during the year of program exit. This dip corresponds to the increased number of clients entering employment during this period. Because entry into a job can occur at any time during the quarter, the average wage during that quarter tends to be lower than the subsequent quarter of full employment.

The goals of these programs include returning individuals who are in need of assistance to competitive employment in the labor force and fostering self-sufficiency. In terms of program assessment, the clients' average quarterly wages following their program exit should begin to approach the corresponding control groups'. While the gap in average quarterly wages went from 90.0 percent below the control group in first quarter 1992 (92Q1) to 71.0 percent below in fourth quarter 1998 (98Q4) for DFS clients; DVR and JTPA clients went from 77.0 percent below the control group to 114.0 percent below and 41.0 percent below to 54.0 percent below, respectively, during the same time period. These findings fail to identify a clear, positive link between program participation and increased earnings.

There are a few important factors to keep in mind when interpreting the results of the three distinct programs. First, DVR and JTPA both provide employment and training services. Perhaps the two-year follow up period does not span enough time to observe a close in the wage gap between the respective programs and their control groups. Another factor influencing the average quarterly wage is the natural attrition in the percent of clients with employment subsequent to exiting the program. If the training prepared individuals adequately to seek competitive employment with higher wages in another state, they may have left Wyoming. Lastly, data on the types of services provided by DFS, beyond knowing they provide public assistance to their clients, are not available to us at this time. Because the DFS clients are generally younger females, most likely with dependents and some family or social network (i.e., a daycare provider) in Wyoming, lower earnings may not have discouraged their attachment to Wyoming. We did not have access to data about other sources of public or private support available to DFS clients.

The longitudinal data are one way to look at the overall behavior of the groups. The longitudinal data provided help identify the percent with wages and the average quarterly wage of those appearing in Wage Records. An important limitation of interpreting data presented graphically in Figures 1 and 2 is that the clients appearing in first quarter 1997 (97Q1) Wage Records may not be the same clients showing up in second quarter 1997 (97Q2). This is further compounded when only those appearing in Wage Records are used for calculating the average quarterly wage, thus presenting a higher average quarterly wage than the group as a whole actually achieved. For example, in first quarter 1996 (96Q1) there were 446 DFS exits. If only 10 of the clients showed up in Wage Records, the 436 clients with no earnings do not affect the average quarterly wage.

Employer Attachment

Another way to look at performance is to present follow up data that includes information on the entire group at a subsequent point in time. Tables 3 and 4 present data on one and two year follow up periods for a limited set of labor market participation indicators.

The first two follow up measures are the average number of quarters worked and the average number of employers following exit from a program. The client and control groups are similar on these two variables with the clients having a slightly lower average duration of employment and a higher average number of employers.

The instability index combines the two previous indicators to produce a clearer understanding of how the number of quarters worked and the number of employers for whom one worked, taken together, influence an individual's success in the labor market. For example, consider a client who works all four quarters following exit but has nine different employers during that period. This would have a positive impact on the number of quarters worked and a negative impact on a stable relationship with an employer. To see how an instability index is calculated, see the related article in this month's Trends ("The Instability Index as a Measure of Labor Market Activity"). The instability index represents the percent of labor market interaction that is unstable. For example an individual that enters employment with an employer in first quarter 1996 (96Q1) and exits employment from the same employer in second quarter 1996 (96Q2) would have an index of 100.0 percent. Whereas, an individual that enters employment in first quarter 1996 (96Q1) and continues to be employed with the same employer for the remaining three quarters, would have an index of 25.0 percent. The client groups are significantly higher on the instability index of the individual as well as the instability index of employers they work for.

The last indicator, average total wages during the follow up periods, is significantly different as well. The client groups earn about half of the wages of the control groups, regardless of program. The DVR clients have the highest annual average wage of $9,281, which is slightly above the poverty level of $8,350 for a family of one and below the poverty level for a family of two ($11,250) - (see Table 5).

Conclusions

This article is intended as a first step in understanding what happens to clients leaving programs, some of which will be centralized in a One Stop environment. It is also intended to generate more questions than answers. For example, with an extended set of data we could ascertain if the clients departing the DFS program were receiving additional aid from some other program. With interstate agreements for Wage Records sharing, we could determine where the people go and what they earn in those states. Interstate agreements will allow us to better understand the flow of Wyoming's labor market and if and why clients we train for employment leave Wyoming. Further, they will allow us to understand some of the reasons people have for attaching themselves to Wyoming's labor market. With a better understanding of the issues surrounding the supply and demand of Wyoming's labor market, perhaps we can assist the programs in training clients for occupations needed in our state.

Our analysis and the points in the related article support the idea that placing clients with employers who have a solid record of employee retention in higher-paying industries should become a higher priority than merely getting them jobs. Placing clients in low wage, high turnover jobs does little to increase their stability and future earnings in the job market over the two year follow up period. The identification of employers with high instability rates and low wages is still in the development stage, but may offer a partial solution to competitive employment outcomes. The benefits of this process will help us secure and sustain our most valuable asset, our own people.

1The Workforce Investment Act of 1998, Pub. L. No. 105-220 (1998). Sec. 121 (b) (1) and (2).
2 "Annual Update of the Health and Human Services (HHS) Poverty Guidelines," Federal Register.
3 Wyoming Wage Records 1992-1998: A Baseline Study, Research & Planning, Wyoming Department of Employment.
4 G. Lee Saathoff, "Separation from the Wyoming Labor Market," Wyoming Labor Force Trends, March 1999; Krista R. Shinkle, "Wyoming-Attached Workers: Living and Working in Wyoming," Trends, April 1999; Gregg Detweiler, "Industry Variations in Wyoming's Steady Workers," Trends, May 1999; Mike Evans, "Job Turnover and Hire Rates in Wyoming," Trends, June 1999; Valerie A. Davis, "Who Are Wyoming's New Hires?," Trends, July 1999.
5 Where are the Jobs? What Do They Pay? 1998 Annual Covered Employment & Wages, Research & Planning, Wyoming Department of Employment.

Table of Contents | Labor Market Information | Employment Resources | Send Us Mail

These pages designed by Gayle C. Edlin.
Last modified on by Valerie A. Davis.