Welfare-to-Work Outcomes: The Role of the Employer.

• Author: Lane, Julia

Julia Lane [*]

David Stevens [+]

This paper sets out to explore the role of the employer in successful welfare-to-work transitions. We demonstrate the uses of administrative records in identifying the types of firms that repeatedly hire welfare recipients. We also develop sets of criteria to identify "successful" outcomes. Our results suggest that there is substantial untapped potential for using administrative data to identify not simply which industries hire welfare recipients but those that are most likely to provide recipients with long-lasting jobs, or jobs accompanied by sustained removal from welfare dependency.

1. Introduction

   There are three major factors to be considered in the implementation of welfare-to-work policy--welfare recipients, employers, and the match between the two. Although both the firm and the match are obviously important to both politicians and the welfare community, much policy analysis has only been on the importance of personal characteristics in getting and keeping work. However, there is abundant evidence that firms differ in personnel practices--both in wage-setting and in turnover policies. As a result, worker outcomes could potentially differ dramatically depending on how workers are placed into different firms. The lack of research on this area is primarily due to the lack of data--there are very few data sets that have information on both firms and workers.

   This paper is one of the first to provide an analysis of the effect on welfare recipients of the interaction between workers and firms, as well as one of the first to demonstrate the value of administrative data in this area. We construct an integrated longitudinal database on firms and workers using the universe of wage record files from the state of Maryland. We merge these data with the universe of welfare records for the state. This enables us to document the impact of firm characteristics, such as previous hiring patterns and growth, on the ability of welfare recipients to successfully retain jobs and earn enough to stay off welfare. [1] In particular, we use administrative records to identify which industries primarily hire former welfare recipients, as well as their observable characteristics. We discuss whether there is firm-specific persistence in the hiring of welfare recipients. Finally we measure job quality by both how long the job lasted and whether the recipient returned to Aid to Families wit h Dependent Children (AFDC) rolls.

   The next section provides some background and context, followed by a description of the creation of the analytical data set from the administrative records. Our discussion of findings begins with a straightforward description of the baseline outcomes for firms, workers, and match durations. We then exploit the longitudinal nature and universal coverage of the unit-record database to estimate the effect of firm characteristics on worker outcomes.

2. Background

   Firm turnover and wage-setting policies can have an important impact on welfare recipients. If some firms deliberately choose high turnover policies, then this is likely to have a disparate impact on low-wage workers, who have shorter job tenure and a greater number of job spells than other workers. Job loss for least educated workers is particularly likely to lead to lower reemployment probabilities, higher probabilities of part-time work, and lower earnings (Farber 1999). This combination of low wages and numerous spells between jobs has important implications for welfare recipiency: The difference between poor and nonpoor low wage workers is that poor workers are employed for roughly the same number of hours per weeks, but 20% fewer weeks per year (Long and Martini 1991). There are longer-term consequences as well. Lost work time not only leads to lost income, but also skill depreciation (Topel 1993), and lower training (Royalty 1996). Finally, there is some evidence that new jobs have fewer benefits, suc h as health insurance and other fringe benefits.

   Firms are likely to have different turnover and wage-setting policies because there are different costs associated with hiring and firing workers. These are clearly dependent on both the type of worker and the nature of the production process, and are thoroughly surveyed by Mortensen and Pissarides (1999). In addition, Kremer (1993) has proposed a promising model of hiring behavior in the context of worker heterogeneity and complex tasks, which he calls an 'O-ring' production function framework. Kremer's model suggests that there will be variation in hiring patterns even across firms within an industry. In particular, different firms, with different production functions, will hire workers of different quality. This approach has since been extended by Kremer and Maskin (1996), who also suggest that, with complex tasks, employers will hire workers with a narrow range of skills. The Kremer-Maskin model predicts that the rising mean and dispersion of the skill distribution will cause organizations to specialize in hiring workers of one skill level or another. Hence persistent hiring of particular types of workers by employers is also predicted by this model. This has been observed empirically--firms, even within quite narrowly defined industries, have quite different, and persistent, workforce composition, productivity, and turnover patterns (Foster, Haltiwanger, and Krizan 1999; Haltiwanger, Lane, and Spletzer 1998). Within the retail trade industry, for example, a coffee franchise like Starbucks has explicitly chosen a relatively high-wage, high-benefit, low-turnover strategy to sell its coffee, whereas other shops will produce a different type of product with a different personnel strategy.

   There is little empirical evidence of the effect of these firm policies on the job outcomes of welfare recipients, despite a voluminous literature on supply-side behavior and consequences (see Moffitt 1992 for an excellent review). However, it is clear that firms do affect worker outcomes: It is well known that wages vary by type of firm (Groshen 1991), and that wages are particularly affected by firm size, industry, and turnover (Krueger and Summers 1988; Davis, Haltiwanger, and Schuh 1996; Lane, Stevens, and Burgess 1996). There is thus a potential link to welfare recipients, since it is well known that recidivism is affected by the wages that welfare recipients earn (Blank and Ruggles 1994; Blank 1997). Preliminary evidence on welfare recipients reinforces this--employment outcomes for welfare recipients are related to the type of firm and industry in which the worker is placed (Lane and Stevens 1995; Bartik 1997).

   However, a full analysis of the effect of firms on workers requires the use of a linked longitudinal database with information on both sides of the labor market. Such data sets have only recently begun to be exploited and have begun to generate a large literature (see Haltiwanger et al. 1998 and 1999 for a survey of some of the literature; Abowd and Kramarz 1999 for a related overview). A core finding of this emergent literature is that some firms pay idiosyncratically high wages, even after controlling for observable and unobservable worker characteristics (Abowd, Kramarz, and Margolis 1999; Abowd, Finer, and Kramarz 1999)--going some way to allay the concern raised by Gibbons and Katz (1992) that firm-specific wage differentials reflect sorting along unobservable characteristics.

   This paper draws on all of this literature in exploiting an unusually rich administrative linked employer-employee data set with longitudinal information on the universe of welfare recipients, the universe of workers, and the universe of firms in the state of Maryland. The following sections describe the basic methodology and creation of this data set in more detail.

3. Methodology

   The core issue is to examine the contribution of firm and worker characteristics to achieving "successful" work outcomes after a welfare recipient is matched with a firm. This can be looked at in two ways: a successful outcome for a worker, or the number of successful outcomes achieved by a given firm. Although "success" might be defined in many ways, in this paper we choose two readily measurable...