Unemployment Insurance and Unemployment Spells - Literature review Example

Economic modeling a. Summary of the selected paper The aim of the study in the article was to establish the effect of the unemployment insurance benefit level and duration, on the duration of unemployment (Meyer, 1990:757). Different models have been applied to model the unemployment spells that have finite durations. Mortensen’s 1997 dynamic search model has shown that individuals maximize the present value of utility, such that they consume the unemployment insurance benefit without saving at the period far before the lapse of the unemployment benefit (Meyer, 1990:758). According to this model, the probability of job offers acceptance raises with the decline of the reserved wages, until at the point of exhaustion of the unemployment insurance benefit, where the environment facing an individual then remains constant (Meyer, 1990:758). Moffit and Nicholson’s 1982 Static labor-leisure model is another model that has been applied to model the unemployment durations with finite spells, indicating that individuals have two preferences; income and unemployment (Meyer, 1990:759). The value of unemployment derives from its leisure utility, while the value of income derives from the budget constraints utility. Thus, when an individual experiences a job loss, they choose weeks of unemployment based on their budget constraints, and the individual will maximize utility by returning to work the week after the insurance benefits exhaustion (Meyer, 1990:759). Therefore, according to these models, an individual will take time to rest during unemployment until they get a highly paid job, but will take on any job offer close to the lapse of the insurance benefit (760). However, the limitation with the two models is that they assume a fixed finite duration of unemployment insurance benefits. However, there are times when such unemployment insurance benefits duration extends (761). The results found that in the weeks just before the unemployment benefits lapse, a higher unemployment benefit is found to extend the duration of the unemployment for individuals, while the nearing of the duration of unemployment benefit lapse increases the probability of leaving unemployment for individuals (Meyer, 1990:757). This is because, with a long duration of enjoying Unemployment insurance benefits, the opportunity costs for job search and leisure are low (Meyer, 1990:771). However, on the event that the duration of the unemployment benefit is extended at a time close to when the benefit was supposed to lapse, the probability of unemployment spell ending is high, owing to the anticipated benefit lapse. The study applied semi-parametric techniques of time estimation, which were then compared with the non-parametric methods, and the results obtained indicated that the semi-parametric estimators gave the most plausible estimations (Meyer, 1990:758). b. Summary the paper that has cited the selected paper The difference between the two papers is that the selected paper holds that reducing unemployment insurance benefit level or duration can reduce unemployment, while this article that cites the selected paper holds that this argument is untrue. According to the article, “Unemployment Duration and the Interactions between Unemployment Insurance and Social Assistance” by Michele Pellizzari, the fact that the effect of the unemployment insurance benefit and the unemployment duration has been the subject of many studies is in no doubt. These studies have come out with two major accepted findings. First, according to Nickell [1979] and Lancaster [1979], higher unemployment insurance benefits results in extended unemployment spells (Pellizzari, 2005:1). Secondly, both Moffit [1985] and Meyer [1990] have managed to show that when sufficient information related to the duration of the unemployment benefits is at the disposal of the individuals, the probability of leaving unemployment will increase with the duration when the unemployment insurance benefit nears exhaustion (Pellizzari, 2005:1). Based on these findings, it would be expected that a reform that can either reduce the duration of the unemployment insurance benefit or reduce the level of unemployment benefit, would result in the decline in the rates of unemployment (Pellizzari, 2005:2). Nevertheless, despite numerous studies showing that reduction in the level of unemployment benefits or the reduction in the duration of offering the unemployment insurance benefit can reduce unemployment, evidence has shown that that theory is not true (Pellizzari, 2005:3). The reduction in the level of unemployment benefits and the shortening of the duration of enjoying the unemployment insurance benefits are some of the reforms that have been implemented in many parts of Europe, but the results have indicated that there is no correlation between these measures and the reduction of unemployment (Pellizzari, 2005:10). The major factor that account for this discrepancy is the social assistance programs. For the individuals who are eligible to both the unemployment insurance benefits and also the social assistance programs, they are less concerned about the reduction or the termination of the unemployment insurance benefits, and thus do not bother to seek for employment (Pellizzari, 2005:3). Therefore, there is a significant difference in the elasticity of the individual who are eligible for family assistance benefits and the probability of the individual seeking for a job (Pellizzari, 2005:6). Nevertheless, the data that was depended upon by the study under this article may not be entirely accurate, owing to the limitation of finding the true data that shows all the benefits obtained by individuals in the countries studied since such benefits are administered by different institutions (Pellizzari, 2005:11). c. Explain what data have been used for the analysis The data applied in the study under the article by Meyer (1990) is the data from the Wage and Benefit History Unemployment Insurance comprising of males from twelve states in the USA for the five-year period between 1978 and 1983 (Meyer, 1990:759). The data structure takes two sets, with the first data set being collected through random sampling from the available records of the Wage and Benefit History unemployment insurance. On the other hand, the second data set comprised of the data collected from the states whose duration of unemployment insurance benefit compensation were not likely to change (Meyer, 1990:762). The nature of the data collected include the rate of unemployment in the states, the level of unemployment benefit compensation and the average weeks of unemployment insurance benefit compensation for the different states from start to the exhaustion of the benefits. The dependent variable for the study is the unemployment rate, which is the variable being measured based on the variation of the independent variable. On the other hand, the independent variable is the unemployment insurance benefit, which is the factor that determines the unemployment rate that will eventually result, based on the level of the benefit that is compensated to the unemployed, as well as the duration for which the insurance benefit compensation is made (Meyer, 1990:764). The potential issues associated with the data range from advantages to disadvantages. First, there are various benefits that are associated with the use of data from the Wage and Benefit History. First, this data is accurate, since it offers accurate weekly receipts of the unemployment benefits by the individuals recruited for the study. The other great advantage associated with the WBH data is that the data offers both the duration and the level of the unemployment benefits, which is rare to find in other data sources (Meyer, 1990:760). However, the major limitation associated with the data is that only the data related to the duration compensated by the unemployment insurance is available, while censoring the data associated with the behavior of the individuals in the period after the exhaustion of the unemployment benefit compensation (Meyer, 1990:761). This set of data provides the disadvantage that it is not possible to assess the behavior of the study participants for the period after the unemployment insurance benefit exhaustion. d. Describe the econometric model used by the authors to analyze the issue The economic modeling model that has been applied in the study is the duration model. This model has been applied in order to solve the limitations associated with the Kaplan-Meier model, which assumes that the samples are homogeneous, thus failing to put into consideration the heterogeneous characteristic of the samples in determining the study outcome (Meyer, 1990:767). The benefit associated with the use of the duration model is that it offers the flexibility of observing the high and the low hazards. Additionally, the duration model also makes it possible to observe the pure effect of getting closer to the unemployment insurance benefit exhaustion, while holding the other factors constant (Meyer, 1990:767). The duration model allows for the use of the time-dependent covariates, while also allowing for the censoring of the data, to take into account the behavior of the participants in the unemployment insurance benefit compensation after the exhaustion of the insurance benefit compensation duration. In order to account for the length of benefit, the duration model applies a non-linear function of time until the time of the exhaustion of the benefits compensation (Meyer, 1990:767). The regression approach is applied in the censoring of data, through assuming the shape for the distribution of the spells, and then applying the Tobit type technique. The duration model takes a less parametric approach to data distribution, but still allows for censoring, although the data at this point becomes more sensitive to the shape of the distribution. The Prentice and Gloeckler [1987] estimation model is used to allow for the nonparametric estimation of the shape of the hazard (Meyer, 1990:768). When the exact spell lengths are unknown, but the intervals during which the spells end are known, a log-likelihood is formed as the basis of analyzing the benefit compensation duration. Alternatively, the duration model can also be interpreted as a convenient discrete model, which operates by interpreting the data as an incompletely observed hazard time model (Meyer, 1990:769). To interpret the duration model data using the convenient discrete approach, the typical approach is to model the group or the discrete time failures such that by applying the density function terms, likelihood is produced as if the spells were observed continuously (Meyer, 1990:768). Nevertheless, this approach is only applicable where ongoing observations of some large spell lengths that exceed 39 weeks are censored. The problem with this approach is that there are only very few number of observations that last for more than 39 weeks, requiring that very strong parametric assumptions are made, to be able to make inferences that would manifest the hazard behavior for the duration after the 39 weeks (Meyer, 1990:769). e. Critically reflect on the issues that could arise in using this particular econometric model to analyze the issue The use of this econometric model to analyze the relationship between the unemployment insurance benefit and the unemployment rates can raise several issues. First, the computation of the discrete computation is very difficult (Meyer, 1990:771). This simply means that the applicability of the duration model for the analysis of the relationship between the unemployment insurance benefit level or duration, on the duration of unemployment is highly restricted, and thus only applicable under certain circumstances. The results obtainable through this model might be therefore prone to errors. Secondly, the application of the duration model requires that the unemployment spell lengths that extend for a period of more than 39 weeks should be censored at the 39th week (Meyer, 1990:796). The effect of this censoring of the spell of unemployment is that it is not accurate to tell the behavior of the individuals receiving the unemployment insurance benefit, for the period after the 39 weeks. The other issue that arises from the application of the duration model in determining the relationship between the unemployment rate and the unemployment insurance benefit level or duration, is that it applies nonparametric estimators. The baseline hazard that is obtainable where the nonparametric estimations are applied produces a choice heterogeneity whose distribution is unimportant (Meyer, 1990:771). This simply means that; despite the fact that the duration model is applied in order to help overcome the homogeneity assumptions that other models such as the Kaplan-Meier model are associated with, the overall effect of the heterogeneity consideration is not helpful. This is because, the duration model may encompass heterogeneity characteristics of the unemployment insurance benefit data, but the nonparametric estimation makes the heterogeneity encompassment unimportant. The other issue likely to arise from the use of the duration model to establish the relationship between the unemployment behavior and the unemployment insurance benefit level or duration is the duration parameter instability (Meyer, 1990:771). The duration model may be advantageous due to the fact that it encompasses heterogeneity of data and also offers high flexibility of the observation of low and high hazards, but the usefulness of this model is limited by the fact that it assumes that the exact unemployment spell lengths are known (Meyer, 1990:770). Thus, this model applies parameters that estimates the outcome based on the assumptions that may not be entirely accurate, and thus the outcome of the duration model estimates may fail to give a true reflection of the relationship between unemployment insurance benefit level or duration, and the corresponding unemployment behavior. References Meyer, B.D. (1990). Unemployment insurance and unemployment spells. Econometrica 58(4), 757-782. Pellizzari, M. (2005). Unemployment Duration and the Interactions between Unemployment Insurance and Social Assistance. Bocconi University. 1-29. Read More