Field Experiment on the Willingness of the Unemployed to Accept a Job, With a Mixed Logit Estimation

Field Experiment on the Willingness of the Unemployed to Accept a Job, with Mixed Logit Estimation 1 Introduction Since people spend anywhere from 20−25% of their lives working, their workplace takes on considerable importance. Not only is it where they work, it is also where they socialize with other people. For the unemployed, finding a job is not just a matter of securing financial independence. After all, they are already receiving income from their unemploy− ment benefits. What they are also looking for is their place in society. In this study, we will estimate the willingness of the unemployed to accept a job offer on the basis of measurable variables such as salary, length of commute and profession. Certain aspects of this area of the labor market, involving willingness of the unemployed, have already been previously analyzed by Namkee et al. (1999), who studied the will− ingness to move for work. Also Blanchard et al. (1995) presented a Probit estimate of accepting jobs requiring functional or geographic mobility. Our goal is to assess three types of mobility: financial (i.e. willingness to accept a lower salary), territorial (i.e. willingness to commute) and finally professional (i.e. willingness to change professions). Since there is a lack of data, we decided to conduct a field experiment and gather data on stated preferences. The unemployed taking part in the experiment would be presented with hypothetical job situations and asked whether they would be willing to accept the conditions or not. Respondents would be asked to make a binary choice where accepting the job offer would mean no longer being unemployed and declining the job offer would mean remaining unemployed for a longer period of time. Their decisions were then analyzed using discrete choice models, particularly the Mixed Logit model. This model accepts random parameters, which allows us to test the sensi− tivity to each variable. Mixed Logit on the labor market was also used by Alba and Esguerra (1998) to analyze the labor supply decision of individuals and more recently by Barros et al. (2005) to estimate low−wage employment in Portugal. The paper is organized as follows: Chapter 2 describes some necessary theoretical un− derpinnings so that concepts like “choice models, stated preference data, experimental design, utility model” and the properties of the statistical models are clarified. Chapter 3 describes the population and sampled data. Chapter 4 highlights the most interesting findings. Finally, Chapter 5 concludes. 2 Theoretical framework 2.1 Analysis of choice models In the following subsections we define some concepts that are used throughout this study. First, we give a definition of what conjoint analysis is and how this discipline relates to choices. Then, the concepts of revealed and stated preferences are clarified. The next subsection shows how experimental design is used to construct sets of multi−attribute alternatives. The final subsection ties in with Economic Theory, presenting the Random Utility Model. 2.1.1 Conjoint analysis and choice models While conjoint analysis is a statistical technique that originated in mathematical psy− chology, it is now widely used in marketing as well as in many of the social sciences, like transport, behavioral, or environmental economics. Conjoint analysis assumes that a product, or a service, can be broken down into its component attributes. The objective of 3