Final Course Major Assignment Research Paper Argumentative Research Essay Students will write a 5–6-pages research paper on one of the topics studied in class

Subject:SociologyPrice:15.89 Bought3

Share With

Final Course Major Assignment Research Paper Argumentative Research Essay Students will write a 5–6-pages research paper on one of the topics studied in class. This paper will include: • • • • • • an introduction that gives background information on the topic and has a thesis statement that takes a clear stand a concession arguments that are fully developed counterarguments and provide evidence to support the thesis the use of appropriate language to support the argumentative genre and a conclusion summarizes main points and reaffirms the strength of the argument. This paper must be informed by a minimum of 3 reliable and vetted sources. Most students will use 5+. You can (and should!) use the same sources you selected and evaluated for the Annotated Bibliography. You will be expected to submit an outline, working draft, and other process writing components before submitting the final paper. See some additional guidance on structure and design in this pdf: EAP 1850 UG Final Argumentative Research Paper (INTOUSF_KBR).pdf The Three Sources That you should use: First a PDF file included in the question. Second https://www.bls.gov/news.release/pdf/empsit.pdf Third https://data.oecd.org/unemp/unemployment-rate.htm The Labor Market Effects of US Reemployment Policy: Lessons from an Analysis of Four Programs during the Great Recession Marios Michaelides, University of Cyprus Peter Mueser, University of Missouri and Institute of Labor Economics (IZA) We present experimental evidence on four US reemployment programs targeting unemployment insurance (UI) recipients during the Great Recession. All programs reduced UI spells, produced UI savings that exceeded program costs, and increased employment rates. The services referral program had the smallest effects, occurring because of voluntary UI exits by participants to avoid requirements. The two programs reviewing participants’ UI eligibility produced greater effects because they induced voluntary exits and disquali?ed participants who did not meet UI requirements. The program requiring participation in both eligibility review and job-counseling services was the most effective, suggesting that services improved participants’ job search efforts. This paper was based on data collected by IMPAQ International as part of a study funded by the US Department of Labor, Employment and Training Administration (DOL/ETA). The authors are grateful to the European Research Council (ERC) for ?nancial support. The views expressed in this paper are those of the authors and should not be attributed to DOL/ETA, IMPAQ, or ERC. We thank Eileen PoeYamagata, Jacob Benus, Jeffrey Smith, David Kaplan, Oksana Loginova, participants in conferences of the Association for Public Policy Analysis and Management and the Society of Labor Economists, and seminar participants at the Department of Economics, the University of Cyprus, and IMPAQ. Contact the corresponding [ Journal of Labor Economics, 2020, vol. 38, no. 4] © 2020 by The University of Chicago. All rights reserved. 0734-306X/2020/3804-0006$10.00 Submitted December 2, 2017; Accepted June 27, 2019; Electronically published August 19, 2020 1099 1100 Michaelides/Mueser I. Introduction Over the course of the past 40 years, policy makers in the United States have made substantial investments in programs that provide job search assistance services to unemployed workers who collect unemployment insurance (UI) bene?ts. Programs have typically referred UI recipients to public employment of?ces in the early stages of their UI spells to learn about and receive services designed to help them improve their job search skills and connect to suitable jobs (Klepinger, Johnson, and Joesch 2002; Black et al. 2003; Wandner 2010). Experimental studies of US programs show that they are effective in reducing average UI spells and bene?ts collected but have limited effects on employment and earnings (Meyer 1995; Decker et al. 2000; Klepinger, Johnson, and Joesch 2002; Black et al. 2003). Reemployment programs in Europe—often referred to as activation programs—have included more intensive requirements, mandating regular participation in job search services and monitoring activities (Kahn 2012; OECD 2013). The effects of these programs have been studied extensively. Numerous experimental studies of European activation programs—including those in the Netherlands (Gorter and Kalb 1996), Denmark (Graversen and van Ours 2008; Maibom, Rosholm, and Svarer 2017), France (Behaghel, Crépon, and Gurgand 2012), and Sweden (Hägglund 2011)—?nd that these programs often lead to substantial improvements in job-?nding rates. After the start of the Great Recession, US policy makers focused on new strategies to facilitate the reemployment of UI recipients and reduce the burden they imposed on the UI programs. In addition to augmenting funding for existing programs, which referred selected UI recipients to job search services, new funding was established to implement programs with in-person eligibility reviews to identify and disqualify those who were not engaged in active job search or were otherwise ineligible for bene?ts. However, very little is known about whether the services referral and eligibility review programs implemented during the recession were effective. The most recent experimental studies focused on US programs implemented in the mid-1990s, a period characterized by low unemployment, and did not test many of the types of interventions in place during the recession. Similarly, most studies of European programs tested interventions that operated in strong labor markets and had more intensive requirements than the US programs operating during the recession. This paper presents experimental evidence on the effects of four reemployment programs, which correspond to a wide range of interventions supported by the US government since the start of the Great Recession. The Florida Priority Reemployment Planning (PREP) program asked participants author, Peter Mueser, at mueserp@missouri.edu. Contact Marios Michaelides at mariosm@ucy.ac.cy. Information concerning access to the data used in this paper is available as supplemental material online. The Labor Market Effects of US Reemployment Policy 1101 to attend an orientation meeting at a public employment of?ce to receive information and referrals to job search services. The Florida Reemployment and Eligibility Assessment (REA) program conducted one-on-one eligibility reviews in which participants provided information on their job search activities and employer contacts. The Idaho REA program required participants to provide information on their job search activities and employer contacts using an online questionnaire. The program then followed up with a subset of participants, contacting employers to verify job search actions and requiring some participants to attend in-person reviews. The Nevada Reemployment and Eligibility Assessment/Reemployment Services (REA/RES) program required participants to meet individually with program staff to undergo an eligibility review and receive mandatory job-counseling services. Those services included a skills assessment to identify an appropriate job search plan, assistance in developing a resume and other job-application materials, and referrals to job vacancies. In all four programs, requirements were scheduled in the ?rst few weeks of participants’ UI spells, with no further requirements to meet with program staff, receive additional services, or undergo eligibility reviews. Our analyses of program effects rely on state administrative data with information on all UI recipients in the three states who started collecting bene?ts in the period August–November 2009—during the Great Recession—and were subject to random assignment for participation in the reemployment programs. These data are used to measure, for both program and control cases, several measures of bene?t receipt, including the number of UI weekly payments received and the total dollar value of bene?ts collected under the UI claim and quarterly employment rates and earnings in the 4 quarters following program assignment. To estimate program effects, we use regression models that estimate program-control differences in UI receipt, employment rates, and earnings, controlling for individual characteristics and random assignment criteria. We then undertake analyses to examine the underlying program mechanisms that may have affected participant job search behavior and outcomes. First, these programs may have produced “threat effects,” occurring because some participants exited UI prior to scheduled program activities to avoid the anticipated cost of program participation. Second, the eligibility reviews required by three of the programs led to the disquali?cation of participants who did not complete the review or were deemed ineligible for bene?ts based on the review, what we term “monitoring effects.” Third, the programs may have created “services effects” by providing services or motivating participants to obtain services that helped them develop more effective job search strategies. For the three programs that disquali?ed participants, we observe disquali?cations, providing a proxy for monitoring effects at each week of the UI spell. To identify the importance of threat and services effects, we estimate 1102 Michaelides/Mueser the overall program effects on the UI exit probability in each week and compare these effects with our measure of monitoring effects. Our analyses rely on the plausible assumption that threat effects likely occurred after participants were noti?ed of program requirements but prior to actual participation, while services effects likely occurred after participants had met program obligations. Based on this assumption and the time pattern of program effects, we assess the extent to which the results of each program can be attributed to threat, monitoring, and services effects. We then consider whether the time pattern of estimated program effects may be in?uenced by dynamic selection, occurring because program requirements may have pushed some participants to exit early in their UI spells. We also consider whether differences across states in unemployment conditions and the characteristics of program-eligible UI recipients may play a role in explaining differences in estimated program effects. The remainder of the paper is organized as follows. Section II provides an overview of US reemployment policy in the past three decades and changes occurring as a response to the Great Recession, followed by a discussion of existing evidence on the effectiveness of reemployment programs. Section III discusses the characteristics of the four programs and the design of the evaluation studies. Section IV presents our data sources, and section V describes our methods. Section VI presents the results, and section VII describes results of robustness tests. Section VIII summarizes our ?ndings and conclusions. II. Background A. US Reemployment Policy and the Great Recession Before the Great Recession, the primary programs for supporting worker reemployment were the Worker Pro?ling and Reemployment Services (WPRS) and the REA programs. Both programs targeted new UI recipients who were designated as “services eligible,” omitting those who were reported to be on temporary layoff, attached to a union hiring hall, or active in employment or training programs. WPRS was created in the early 1990s to address concerns that UI recipients were not participating in job search services offered at public employment of?ces. WPRS required serviceseligible UI recipients who were deemed most likely to exhaust regular bene?ts to (1) register in the state’s labor exchange system to match them to available employment opportunities, (2) attend workshops to learn basic job search skills, and (3) receive individualized job-counseling services (see Dickinson et al. 1999). WPRS became operational in all states in 1996 (Wandner 2010). By the early 2000s, UI systems had become highly automated, enabling unemployed workers to use telephone and internet systems to apply for UI, report their work search activities, and access basic job search services (Ridley and Tracy 2004; O’Leary and Wandner 2005). This automation The Labor Market Effects of US Reemployment Policy 1103 reduced administrative costs but raised concerns about the ability of agencies to monitor UI recipients (Wandner 2010). To address these concerns, the US Department of Labor established the REA program in 2005, which required services-eligible UI recipients to undergo an eligibility review. REA’s objective was to reduce UI fraud by disqualifying recipients who, during the meeting, were deemed ineligible for bene?ts (Benus et al. 2008; Poe-Yamagata et al. 2012). Nine states—including Florida, Idaho, and Nevada—adopted the program at that time (US Department of Labor 2005). At the end of 2007, the US economy entered the Great Recession, a period when the national unemployment rate climbed from 5% in December 2007 to a peak of about 10% at the end of 2009. From 2007 to 2009, new UI claims increased from 7.7 to 13.9 million, and registrations in labor exchange systems increased from 17.8 to 22.4 million (Wandner and Eberts 2014). To alleviate the added burden brought on by the recession, the American Recovery and Reinvestment Act of 2009 allocated $400 million—in addition to the annual $724 million allocation under the Wagner-Peyser Act for 2009 and 2010—to support job search services and expand WPRS. Partly as a result, WPRS referrals increased from 1.3 million in 2008 to 1.9 million in 2009 and to 2.1 million in 2010. The federal government also provided $76 million to the US Department of Labor to support the expansion of REA to 33 states and encourage states to offer job-counseling services to participants who passed the eligibility review (US Department of Labor 2009, 2010). B. Prior Research In the 1980s, the US Department of Labor implemented various experimental design demonstration programs to test the effectiveness of programs that required UI recipients to engage in job search services, including use of job banks, employment workshops, and job counseling. Most of these programs were found to reduce UI receipt, although they yielded small or no effects on participants’ reemployment rates and earnings (Anderson, Corson, and Decker 1991; Johnson and Klepinger 1991; Meyer 1995). These early studies did not provide any evidence on the underlying mechanisms that led to program effects on UI receipt and why these effects did not translate into substantive improvements in job-?nding rates and earnings. Several experimental studies published in the early 2000s provided evidence on the source of program effects. Black et al. (2003) showed that the Kentucky WPRS program in the period October 1994 through June 1996 reduced UI receipt by 2.2 weeks and bene?t amounts collected by $143. The study found that these effects occurred primarily because the program caused some participants to discontinue bene?ts following receipt of the letter informing them of the required meeting but prior to the meeting itself. They argued that most of the program’s impact was due to the threat effects of program requirements and that services played a limited role. 1104 Michaelides/Mueser Two other experimental studies found similar results. Decker et al. (2000) showed that two job search assistance demonstration programs implemented by the US Department of Labor in Florida and Washington, DC, in the mid1990s, which referred UI recipients to job search services, reduced UI duration by up to 1.1 weeks and UI bene?ts collected by up to $182. Using data from a demonstration program implemented in Maryland in 1994, Klepinger, Johnson, and Joesch (2002) showed that requiring UI recipients to participate in an employment workshop reduced UI duration by 0.9 weeks and bene?ts collected by $75. In each study, almost all effects occurred in the early stages of participants’ UI spells after they were noti?ed of program requirements and prior to engaging in program activities. In a recent study, Michaelides and Mueser (2018) showed that during the recession the Nevada REA/RES program reduced UI duration by more than 4 weeks and UI bene?ts collected by more than $1,000. Further analyses indicated that program effects were partly due to early participant exits but were also due to participant exits after program requirements had been met, providing suggestive evidence that job-counseling services were effective. There is other evidence based on US data that job-counseling services may help disadvantaged workers in the labor market, but results concern individuals receiving training (Perez-Johnson, Moore, and Santillano 2011) or disability bene?ts (Weathers and Bailey 2014). Although these studies provide valuable guidance on the effects of reemployment programs, they provide limited bases for inferring the relative ef?cacy of the types of interventions implemented during the Great Recession. First, many studies focused on programs that operated in the mid-1990s, more than a decade prior to the Great Recession, a period when the national unemployment rate was in the 5%–6% range and unemployed workers were less likely to exhaust their UI eligibility (extending up to 26 weeks). Second, none of those programs focused exclusively on formal eligibility reviews, where UI recipients were required to meet with program staff and provide proof of their job search activities as a condition for continued bene?t eligibility. Thus, there is no evidence on the potential monitoring effects stemming from the eligibility review programs operating in 33 states during the recession, including the Florida and Idaho programs examined here. Third, those who believe that services are of potential value have pointed out that the programs studied by Decker et al. (2000), Klepinger, Johnson, and Joesch (2002), and Black et al. (2003) had relatively weak service components (Wandner 2010). Although many participants received basic services (such as registration in job banks, group orientations, and employment workshops), few participants received personalized job-counseling services. Thus, we may not be able to use these studies to make inferences about the effects of job counseling or to assess the ef?cacy of programs that mandated both eligibility review and job counseling. Michaelides and Mueser (2018), which focused exclusively on the Nevada program, one of the four programs The Labor Market Effects of US Reemployment Policy 1105 examined here, is perhaps the only study of a program with a strong jobcounseling services component. The study suggests that programs combining eligibility reviews with job-counseling services were effective during the Great Recession but provides no evidence on whether similar effects were achieved during the recession by interventions with less intensive requirements, such as the services referrals and eligibility review programs examined here. A substantial number of random assignment studies, many completed over the last decade, have evaluated programs implemented in European countries, including Denmark (Graversen and van Ours 2008), France (Behaghel, Crépon, and Gurgand 2012), Germany (Krug and Stephan 2013), the Netherlands (Gorter and Kalb 1996), Sweden (Hägglund 2011), and the United Kingdom (Dolton and O’Neill 2002). These studies—which mostly concerned programs implemented in periods of low unemployment—often found that programs were responsible for large effects on unemployment exits and reemployment rates. A meta-analysis of active labor market programs across a large number of countries by Card, Kluve, and Weber (2018) found that job search assistance programs generally had positive effects on employment. For the most part, European programs are more intensive than US programs (including those examined here), requiring continuous participation in monitoring activities and job counseling, so the incentives of participants to escape program requirements and monitoring effects cannot be distinguished from the bene?ts of job counseling. One partial exception is Maibom, Rosholm, and Svarer (2017), who show that an intensive counseling treatment in Denmark for newly unemployed workers had effects that continued even after program requirements had been satis?ed. Another partial exception is Cockx et al. (2018), who found that monitoring activities in the Belgian UI system had small positive effects on participants’ job search behavior. Importantly, none of the European studies determined whether programs had positive impacts on earnings, a key indication of whether the job counseling mandated by these programs led to improved labor market outcomes. III. Program Descriptions The four programs examined here were selected for two reasons. First, they used random assignment to allocate services-eligible UI recipients to either the program group (subject to program requirements) or the control group (not subject to program requirements). This ensures that program effects can be estimated through program-control comparisons in observed outcomes after controlling for the random assignment procedures. Second, they adopted different approaches, representing a wide range of programs that have been in operation in the United States since the start of the recession. This allows us to examine the effects of programs that have not been previously tested in the literature and provides an assessment of US 1106 Michaelides/Mueser reemployment policy during the recession. Below is a detailed description of the four programs; ?gure 1 summarizes program designs. A. The Florida PREP and REA Programs Florida implemented both the WPRS and REA programs during the recession. Each week, the UI agency uploaded the pool of new UI recipients who were eligible for program participation to an interface accessible by regional workforce boards. Each board randomly assigned services-eligible recipients to one of three groups: WPRS, called PREP; REA; and the control group.1 PREP participants were referred to a group orientation at a public employment of?ce to receive information and referrals to job search services. REA participants were required to undergo an in-person eligibility review at a public employment of?ce. Both PREP and REA participants were informed of program requirements through a noti?cation letter sent in week 2 of their UI spell (i.e., when they collected their second UI weekly payment). The letter speci?ed the date and time of the PREP orientation or REA review meeting, typically scheduled in weeks 4–6 of the UI spell. Control cases received no noti?cations and had no requirements under PREP or REA but were subject to the usual UI job search requirements. REA participants who were deemed ineligible based on the review were disquali?ed from collecting UI. Those who did not show up for and failed to reschedule the review within 3 weeks of the scheduled meeting were also disquali?ed, unless the employment service data system indicated that they had participated in job search services or were enrolled in training. In contrast, PREP participation was not strictly enforced and bene?ts were not cut off for those who failed to attend or reschedule the orientation. After the PREP or REA meeting, participants were informed that they had no additional requirements under the program. However, both program and control cases could participate in services on their own initiative. The design of Florida PREP is similar to that of most of the other WPRS state programs in effect during the recession, where participants were required to visit public employment of?ces to learn about and receive referrals to services. Unlike Florida PREP and other US programs studied to date, Florida REA did not include referrals to orientation meetings or to speci?c job search services. Instead, it focused exclusively on conducting eligibility reviews and disqualifying participants who were deemed ineligible during the reviews. Although this structure is similar to the structure used 1 The proportion assigned to PREP and REA varied on the basis of each regional board’s capacity to serve participants in a given week. Our analyses consider serviceeligible UI recipients in the 10 regions where substantial numbers of UI recipients were assigned to both WPRS and REA during the study period. Tabulations of the 2009 American Community Survey (ACS) show that these 10 regions covered 60% of unemployed workers in the state. FIG. 1.—Program designs in Florida, Idaho, and Nevada. PREP 5 Priority Reemployment Planning; REA 5 Reemployment and Eligibility Assessment; REA/RES 5 Reemployment and Eligibility Assessment/Reemployment Services; UI 5 unemployment insurance. 1108 Michaelides/Mueser by the majority of the 33 states that implemented REA, it is notably different from the structure of most European programs, which mandate continuous participation in monitoring and job-counseling activities. B. The Idaho REA Program Idaho maintained both WPRS and REA programs during the study period, but WPRS was very small, serving only about 2% of services-eligible UI recipients.2 The remaining services-eligible recipients were randomly assigned to the REA program or the control group. Control group cases received no targeted communications beyond those associated with normal UI receipt. Those assigned to the REA program were sent a noti?cation letter in week 1 of their UI spell (when they collected their ?rst UI payment) asking them to complete an online review on the IdahoWorks website by week 4, providing information on their work search activities and employer contacts. In week 5, participants who were still receiving UI but had not completed the online review and those deemed ineligible based on their responses were disquali?ed from collecting UI.3 The Idaho UI agency then selected about 5% of those who completed the online review for telephone veri?cation of their employer contacts and about 20% for an in-person review. The remaining 75% had no further contact with the program. Participants who were selected for the telephone veri?cation were not informed of their selection except in cases where the information obtained warranted disquali?cation from collecting UI bene?ts. Those selected for the in-person review were contacted by phone in week 5 to set up an appointment; the in-person reviews were typically scheduled in weeks 6–7. Those who did not show up for the in-person review and those who were deemed ineligible during the review were disquali?ed. Those who passed the review were not required to receive any services and were explicitly informed that they did not have any further requirements under the REA program but that they were still subject to the usual UI job search requirements. The use of online tools for the eligibility reviews distinguishes Idaho REA from REA programs operating in other states, including Florida, which relied exclusively on in-person reviews. To our knowledge, the Idaho REA is the ?rst reemployment program in the United States or Europe that relied primarily on online tools and in which most participants were not required to have face-to-face interactions with program staff. 2 Idaho assigned services-eligible recipients to WPRS if they were deemed hard to employ based on whether they had low education, were previously employed in low-wage jobs, and had short prior job tenure. 3 Similar to the Florida REA program, participants who did not complete the online review but who participated in job search services or were enrolled in training, as shown in the employment service data system, were excused. The Labor Market Effects of US Reemployment Policy 1109 C. The Nevada REA/RES Program Nevada adopted a different approach from other states, essentially combining the WPRS and REA programs, requiring participants to attend a meeting with program staff in the ?rst few weeks of their UI spell to (1) undergo the REA eligibility review and, if deemed eligible, (2) receive mandatory staff-assisted job-counseling services (called RES).4 Each week, the Nevada UI agency randomly assigned program-eligible recipients into the REA/RES group or the control group. A noti?cation letter was sent to each REA/RES participant in week 1 of the UI spell (the ?rst week of UI payments) providing a speci?c date and time for the required meeting, typically in weeks 2– 4. REA/RES participants who did not show up for the meeting were disquali?ed from collecting UI unless they rescheduled the meeting. Exempted from the meeting were those who had received job search services, were active in training, or had discontinued UI bene?t receipt. Participants deemed ineligible for bene?ts during the meeting because they were not actively searching for a job or for other reasons were disquali?ed from collecting UI. Participants who passed the eligibility review were offered job-counseling services during the same meeting, including (1) an individual skills assessment and development of a work search plan targeting the types of jobs they should be pursuing based on their skills and experience, (2) assistance in developing a professional resume and other job-application materials, and (3) referrals to employers with job openings compatible with their skills and experience. Participants were offered job-counseling services based on individual needs, so not all participants received all services. Participants also received information about other available job search services, including job banks, orientation meetings, and employment workshops. Following the meeting, REA/RES participants were informed that they had completed program requirements, and—although they were not required to receive additional services or attend follow-up meetings—they were expected to continue to actively search for a job. Those assigned to the control group did not receive any noti?cations from the REA/RES program and were not required to receive any services, although they were subject to the usual UI work search requirements. The design of the Nevada REA/RES program is a departure from most US programs studied to date. Nevada was the only one of the 33 states adopting eligibility review programs since the start of the recession that combined the review with mandatory job-counseling services. Most US 4 REA/RES was implemented in the workforce regions covering the Las Vegas– Henderson–Paradise and Reno metropolitan areas. Tabulations of the 2009 ACS show that these regions covered 87% of unemployed workers in the state during the study period. The WPRS program continued to operate in the rest of the state. 1110 Michaelides/Mueser programs studied prior to the recession did not include the eligibility review and did not require participation in individualized job-counseling services. Nevada REA/RES had requirements similar to those of many European activation programs, with the notable distinction that while the European programs mandated monitoring and job-counseling activities throughout the participants’ UI spells, Nevada required participants to engage in such activities only once, at the beginning of their UI spells, with no further requirements thereafter. D. Context of Operation This paper examines the effects of the four programs from August through November 2009, when the US economy was experiencing the consequences of the Great Recession, its worse economic downturn since the Great Depression. Figure A1 presents key indicators of the labor market for the three study states and the entire United States from 2005 through 2015, a period that encompasses the of?cial start and end date of the Great Recession (December 2007 through June 2009). As seen in the ?rst three plots, all three states experienced large increases in unemployment and sharp declines in housing prices and net job gains from 2007 through 2010. In the aftermath of the recession, from 2011 to 2015, unemployment rates gradually returned to their prerecession levels, and housing prices and net job gains rebounded. Nevada appears to be the most seriously affected of the three states, as it experienced the greatest changes in unemployment and housing prices during and following the recession. Unsurprisingly, demand for UI bene?ts rose dramatically during the recession—in 2009, about one in every four labor force participants in Nevada and Idaho started a new UI claim, compared with 14% in Florida and 19% in the entire United States. There was also a substantial increase in the average UI duration (number of weeks spent collecting UI) in all three states during the recession, followed by declines after the recession. It is also worth noting differences in the structure of the economy across the three states during the study period. As seen in table A1, average earnings in the three states were lower than the national average; Idaho had the lowest average earnings. A much higher proportion of the Nevada labor force (27.4%) was in other services (which includes casinos, entertainment venues, hotels, and restaurants) relative to Florida (15.5%) and Idaho (12.1%). Idaho was relatively more reliant on manufacturing (10.0%) and on mining and natural resources (6.1%). In Florida, professional services and health care services were slightly more important than in the other two states. There were also differences in the occupational distribution, although they were more modest. Overall, all three states experienced the adverse effects of the Great Recession and had similar recovery paths, although the Nevada economy was more volatile than the economies in the other two states. This may be partly The Labor Market Effects of US Reemployment Policy 1111 due to state differences in the industrial and occupational distribution of the labor force. It is thus possible that the ef?cacy of the programs considered here may be in?uenced by underlying differences in state economies. To examine this possibility, following our main analyses of program effects in each state, we undertake additional analyses of program effects based on weighting that accounts for differences in gender, age, education, occupation, and other characteristics across states. We also examine whether program effects in each state vary on the basis of the level of unemployment in a participant’s county of residence. These analyses show that our main results are not affected by underlying differences in the characteristics of the workforce and in unemployment rates, mitigating concerns about the comparability of ?ndings across states. IV. Data Description We use state UI claims data and wage records for all workers who started collecting UI bene?ts in August–November 2009 and were subject to random assignment for participation in the four programs. The UI agency in each state used information on employment and earnings in the base period (the ?rst four of the ?ve calendar quarters prior to the UI claim) to determine the weekly bene?t amount (WBA) and the number of regular UI weeks each recipient was eligible to collect. In all three states, bene?t payments were about half of prior earnings for low earners, but because of program caps on allowable weekly bene?ts ($275 in Florida, $362 in Idaho, and $393 in Nevada), the replacement rates were lower for higher-earning workers. Weeks of regular UI eligibility varied from 9 to 26 weeks in Florida, from 10 to 26 weeks in Idaho, and from 12 to 26 weeks in Nevada. Program rules were designed so that UI recipients with steady prior earnings would be eligible to receive the full 26 weeks, and those with both low and unstable prior earnings would be penalized with shorter UI eligibility.5 During the study period, the three states had their highest unemployment rates in more than 25 years—11.1% in Florida, 8.7% in Idaho, and 12.0% in 5 In Florida and Idaho, the WBA was equal to 1/26 of earnings in the quarter with the highest earnings during the base period, subject to a $32 minimum and a $275 maximum in Florida and a $65 minimum and a $362 maximum in Idaho. In Nevada, the WBA was equal to 1/25 of earnings in the highest-earnings quarter, with a $16 minimum and a $393 maximum. In Florida, weeks of eligibility were equal to 25% of earnings in the base period divided by the WBA, with a 9-week minimum and a 26-week maximum. In Idaho, weeks of eligibility were equal to one-third of the base period earnings divided by the WBA, with a 10-week minimum and a 26-week maximum. In Nevada, weeks of eligibility were equal to one-third of the base year earnings divided by the WBA (12-week minimum and 26-week maximum). Additional conditions for eligibility also applied. Source: US Department of Labor, Employment and Training Administration, https://workforcesecurity .doleta.gov/unemploy/comparison/2000-2009/comparison2009.asp. 1112 Michaelides/Mueser Nevada, compared with the 9.8% national unemployment rate.6 The state rates exceeded the threshold for activating the Emergency Unemployment Compensation (EUC) and Extended Bene?t (EB) programs. Thus, in accord with federal rules, recipients who exhausted regular UI bene?ts and had a minimum of 20 weeks of employment in the year prior to the start of the UI spell were eligible to apply for up to 53 more weeks of bene?ts under EUC.7 In addition, those who exhausted EUC bene?ts could apply for up to 20 additional weeks of bene?ts under EB.8 WBAs for EUC and EB were identical to those under regular UI. The UI claims data used here provide the socioeconomic characteristics of services-eligible recipients at the start of their UI claims and their program/ control status for the reemployment programs. The data also report the number of bene?t weeks and the dollar value of bene?ts collected under the regular UI and EUC programs; bene?ts collected under EB were not available because states had not yet developed a system to track EB payments.9 Using UI claims data, we construct the following measures of UI receipt: (1) an indicator of whether the individual exhausted regular UI bene?ts, (2) an indicator of whether the individual started collecting EUC, (3) an indicator of whether the individual exhausted EUC bene?ts, (4) the number of regular UI and EUC weeks collected, and (5) the dollar value of regular UI and EUC bene?ts collected. Our data do not provide information on employment transitions, so they cannot be used to measure unemployment and employment spells. Wage records are maintained by states under the UI program and provide individual information on quarterly earnings received from employers in the state. For this study, we linked wage records with the UI claims data at the individual level. Using the wage records, we construct two measures of employment at the individual level for each of the 4 quarters prior to UI entry and the 4 quarters following UI entry: (1) a dichotomous variable indicating whether earnings in a quarter are positive and (2) the dollar value of total earnings in a quarter, with those without earnings included with values of zero. These data do not provide information on earnings from 6 Source: US Bureau of Labor Statistics, https://www.bls.gov/web/laus/ststdn sadata.txt. 7 Recipients who exhausted regular UI and applied for EUC were eligible for (1) up to 20 weeks of EUC tier 1 bene?ts (lesser of 80% of regular UI and 20 weeks), (2) up to 14 weeks of EUC tier 2 bene?ts (lesser of 50% of regular UI and 13 weeks, plus 1 week), (3) up to 13 weeks of EUC tier 3 bene?ts (lesser of 50% of regular UI and 13 weeks), and (4) up to 6 weeks of EUC tier 4 bene?ts (lesser of 25% of regular UI and 6 weeks). UI recipients had to exhaust one tier of bene?ts before qualifying for the next tier. 8 Recipients who exhausted EUC and applied for EB were eligible for up to 20 weeks of EB (lesser of 80% of regular UI or 20 weeks). 9 The implication is that we do not observe the entire UI spell for those who exhausted both regular UI and EUC and applied for EB. The Labor Market Effects of US Reemployment Policy 1113 employers outside the state, informal or illegal employment, or federal and military jobs. Also, the data do not include employment length or hourly wages. Despite these limitations, wage records appear to provide valid measures of program impact and have been used widely in the program evaluation literature in lieu of survey data (Black et al. 2003; Wallace and Haveman 2007; Perez-Johnson, Moore, and Santillano 2011). Table A2 presents sample means of UI receipt, employment, and earnings outcomes for program and control groups. We also have access to program data that provide partial information on participants’ program-related activities. The Florida data provide the exact date when the PREP and REA meetings were initially scheduled and whether the participant attended the meeting.10 The Idaho REA data report which participants were selected for the phone veri?cation and for the in-person interview but do not report whether participants completed the online and in-person reviews. The Nevada REA/RES program data provide the date when the REA/RES meeting was scheduled and whether the participant attended the meeting.11 The data for the three programs with the eligibility review report whether a participant was disquali?ed for failure to complete the review or because of issues identi?ed during the review. The Nevada data also report the speci?c job search services received by treatment and control cases, so we can explore whether the program increased services take-up. Unfortunately, we were not able to obtain job search services data for the other three programs. Table 1 summarizes the characteristics of program-eligible UI recipients in the three states. In Florida, there were 58,416 eligible recipients; 32% were assigned to PREP and 40% to REA, with 28% in the control group. In Idaho, 79% were assigned to REA, whereas in Nevada, 16% were assigned to REA/RES. There are some differences across states in the characteristics of eligible recipients. Blacks made up about a sixth of recipients in Florida but only 1% of those in Idaho. (Nevada does not have a reliable measure of race.) In Florida, a relatively smaller proportion reported some college education. Differences in occupational distribution re?ect the aggregate differences reported earlier—there were relatively more UI recipients in Florida reporting white-collar, high-skill jobs, and there were relatively more UI recipients in Idaho and Nevada reporting blue-collar, low-skill jobs. Prior earnings in Florida were slightly higher than earnings in Nevada and were 30% higher than earnings in Idaho. Moreover, UI recipients in Florida and Nevada were eligible for more weeks of regular UI bene?ts than UI 10 Florida’s REA participants could reschedule the meeting to occur within 3 weeks after the original date. The data do not report the actual week when the meeting took place but indicate whether the meeting did not occur. 11 In Nevada, the speci?ed interview date includes any postponement, so unless the individual is coded as missing the interview, the interview occurred on the date indicated. Table 1 Characteristics of Services-Eligible Unemployment Insurance (UI) Recipients Florida Idaho Nevada N (%) Total Program: PREP REA REA/RES Control Sex: Men Women Race: White Black Other Education: No high school diploma High school diploma Some college College degree Age: 16–24 years old 25–34 years old 35–44 years old 45–54 years old 55–64 years old ≥65 years old Occupation: White collar, high skill White collar, low skill Blue collar, high skill Blue collar, low skill 58,416 (100) 12,645 (100) 21,898 (100) 18,510 (32) 23,471 (40) 9,950 (79) 16,435 (28) 2,695 (21) 3,496 (16) 18,402 (84) 33,422 (57) 24,994 (43) 7,769 (61) 4,876 (39) 12,426 (57) 9,472 (43) 35,194 (60) 10,194 (17) 13,028 (22) 10,336 (82) 79 (1) 2,230 (18) 7,843 (13) 31,352 (54) 9,955 (17) 9,266 (16) 1,694 (13) 4,827 (38) 4,510 (36) 1,614 (13) 3,480 (16) 9,475 (43) 6,286 (39) 2,657 (12) 6,550 (11) 13,663 (23) 13,605 (23) 13,672 (23) 8,421 (14) 2,505 (4) 1,956 (15) 3,442 (27) 2,603 (21) 2,671 (21) 1,559 (12) 414 (3) 2,767 (13) 5,526 (25) 4,972 (23) 4,782 (22) 2,807 (13) 1,044 (5) 17,796 (30) 16,749 (29) 12,830 (22) 11,041 (19) 2,236 (18) 3,049 (24) 2,986 (24) 4,374 (35) 4,219 (19) 7,040 (32) 4,980 (23) 5,659 (26) Mean (SD) Regular UI entitlement ($) Regular UI weeks allowed Earnings: First quarter prior to entry Second quarter prior to entry Third quarter prior to entry Fourth quarter prior to entry 5,428 (1,973) 22.5 (4.6) 5,198 (2,603) 20.0 (5.6) 7,054 (3,031) 22.9 (4.5) 7,412 (8,089) 7,718 (10,305) 7,878 (8,129) 7,775 (8,591) 5,582 (5,210) 5,446 (5,271) 5,204 (5,850) 5,986 (5,754) 7,108 (7,038) 7,267 (7,102) 7,422 (7,305) 7,482 (7,440) NOTE.—Reported are the number of services-eligible recipients with sample proportions in parentheses; for regular UI entitlement, regular UI weeks allowed, and prior earnings, sample means with standard deviations in parentheses are reported. PREP 5 Priority Reemployment Planning; REA 5 Reemployment and Eligibility Assessment; REA/RES 5 Reemployment and Eligibility Assessment/Reemployment Services. The Labor Market Effects of US Reemployment Policy 1115 recipients in Idaho,12 while the total dollar value of the regular UI entitlement was higher in Nevada than in the other two states. Using observed characteristics and prior labor market outcomes, we investigate whether random assignment was properly undertaken in each state. In Florida, random assignment occurred on a weekly basis within each region, at the level of the regional workforce board, and the number of eligible recipients assigned in each program depended on the program capacity in each region at the time of assignment. To test Florida’s random assignment process, we ran two linear probability models, predicting PREP and REA assignment with individual characteristics and measures of prior labor market outcomes, region and week indicators, and regional workforce board-week interactions. In Idaho and Nevada we ran similar regressions, but since random assignment occurred at the state level on a week-by-week basis, the inclusion of region indicators and region-week interactions was not necessary. Only 3 of 74 coef?cients are statistically signi?cant at the 10% level and none are signi?cant at a higher level, a count that could easily be due to chance (table A3). These results suggest that observed characteristics and prior labor market outcomes do not contribute to the program assignment prediction, consistent with the view that participants were randomly selected. Importantly, estimated parameters are very small, implying that any treatmentcontrol imbalances in characteristics do not cause substantive differences in outcomes once week (and region by week effects in Florida) are controlled. Table 2 provides information on the scheduling of program requirements and completion rates. The vast majority of PREP and REA meetings in Florida were scheduled in weeks 4 –6. In Idaho, all participants were required to complete the online review by week 4, and about one-?fth were selected for an in-person interview; most in-person interviews were scheduled in weeks 6–7. In Nevada, about a ?fth of REA/RES meetings were scheduled in week 2, with 95% of meetings scheduled by week 6.13 Florida and Nevada data show that most participants completed the in-person eligibility reviews, and about 60% of Florida PREP participants attended the orientation. Completion data were not available for Idaho. Program data also provide information on disquali?cations occurring because treatment cases failed to complete the eligibility review or because of ?ndings of ineligibility during the review. These data show that a small proportion of program cases were disquali?ed in the three programs with an eligibility review—1.4% in Florida REA, 2.6% in Idaho REA, and 2.1% in Nevada REA/RES. The timing of disquali?cations for each program 12 Note that while Florida UI rules allow 9–26 weeks of eligibility, in practice we observe that Florida UI recipients in our sample were eligible for 10–26 weeks; none was eligible for 9 weeks.