پویایی بازار کار استرالیا در اعصار
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|17128||2013||11 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Economic Modelling, Volume 35, September 2013, Pages 453–463
Transition probabilities between four labor market states (full-time employment, part-time employment, unemployment and inactive) for three age groups (the young, mature and old) are calculated using monthly gross flow data for Australia from October 1997 to May 2012. We determine the responses of the different groups to phases of the business cycle by estimating four unobserved common dynamic factors that drive the age group transitions. Job-finding and job-losing factors were significantly affected by the business cycle variable. Impulse responses to a business cycle shock show that both job-finding and job-losing matter, with the former more important. Part-time job flow responses are more important cyclically than full-time ones. In business cycle downturns, the young have a disproportionate difficulty finding and keeping all jobs, the mature unemployed have more difficulty finding full-time work, and the old have difficulty keeping full-time work.
We consider age heterogeneity in gross flow data for Australian labor markets. We calculate transition probabilities, or hazard rates, between four labor market states (full-time employment, part-time employment, unemployment and inactive) for three age groups (the young, 15–24, the mature, 25–54, and the old, 55 plus) using monthly gross flow data for Australia from October 1997 to May 2012. Disaggregation in the age dimension allows a comparison of trends and an analysis of the responses of the different groups to phases of the business cycle. The results of such an analysis are useful for designing the details of macroeconomic and labor market policies over the business cycle, and for understanding the evolution of labor market aggregates. With twelve transitions between four states for three age groups, we develop a thirty six differential equation system from three sets of twelve equations, based on the model of intra-period transition probabilities proposed by (Shimer, 2012). In this paper, we address age demographics, but not gender for two reasons. First, the data for transitions is sparse for the old age group, and this problem intensifies with a gender breakdown. Second, if we had to add a gender breakdown, the number of equations and parameters to estimate would double from an already very large number. Since it is difficult to get a clear picture of what is going on at such a high dimension, one major contribution of this paper is the construction of a dynamic factor model that substantially reduces the dimensionality to a workable level. From the thirty six observed endogenous variables, we estimate four unobserved dynamic labor market factors—job-finding, job-losing, and labor market participation-in and participation-out the labor force. These factors are identified from relevant transition probabilities across the age groups. We explore how these factors are related to business cycle activity, and we find (as expected) that the job-finding and participation-in factors are significantly procyclical, while the job-losing and participation-out factors are significantly countercyclical. We then simulate the estimated model to establish the effects of a shock to the business cycle variable on the actual people transitions between the four labor market states for the three age groups. A critical question that we address is whether business cycle fluctuations affect job-finding or job-losing, and then how these in turn affect labor market transitions for the various age groups. We also consider the relative importance of part-time and full-time employment fluctuations over the business cycle. Our model is applied to Australian data because it is a typical developed economy with a flexible labor market and a rich source of relevant statistics,1 and thus provides an excellent application for our innovative modeling approach. Our sample includes two downturns in the Australian economy. The first began at the end of 2001 and persisted for about two years, while the second began towards the end of 2008 and began to stabilize at the end of the sample. These downturns affected the estimated dynamic labor market factors, and in turn the transition probabilities and gross flows between labor market states. Using a representative monthly business cycle variable – the aggregate employment-to-population ratio – we show that, through the significantly procyclical job-finding factor, the business cycle effect on transition probabilities from unemployment into full-time and part-time jobs was strong for the young and mature, but relevant only into part-time work for the old. For the young and mature, the probability of full-time job-finding was not significantly associated with exits from the inactive state, while for the old this effect was significant. Through the significant and countercyclical effects of the business cycle on the job-losing factor, we find that the probability of all job-losing is significant for all ages, except for exits from part-time jobs by the mature. The participation-in factor was acyclical, while the transition probabilities associated with the participation-out factor were surprisingly significantly procyclical. We suggest that this last result may be because they are likely to be voluntary separations driven by household income fluctuations. While these results for the business cycle effects on transition probabilities are interesting, they do not provide sufficient information about the gross flow quantities of actual people in the three age groups between the four states. As a major contribution, we recover this quantity information by simulating the estimated model for transition probabilities with a positive shock to the business cycle variable, reporting the deviations from the gross flow means of our sample. For job-finding, we find that the predicted gross flows into full-time and part-time jobs rise most for the mature, followed by the young. This result is unsurprising, given the relatively large size of the mature cohort. But given the relative size of the young group, their responses are disproportionately large. Part-time employment responds more significantly than full-time employment to business cycle shocks. While job-losing is also important in the business cycle, job-finding is much more of an issue. Part-time employment fluctuations are more important than full-time ones, with the young disproportionately represented in changes in part-time jobs. The next section surveys the literature on gross flows, and is followed by an explanation of the problems faced in dealing with gross flow data, how we resolved them and calculated the transition probability matrix. We graph the probabilities and discuss their key features. In Section 4, we present and estimate our dynamic factor model, and in Section 5, we provide impulse responses to understand the effects of a business cycle shock on the transitions between labor market states, and we discuss the policy implications of these results. The final section gives some concluding comments.
نتیجه گیری انگلیسی
In this paper we compute monthly transition probabilities between four different states in Australian labor market for three age groups. We find that these probabilities follow age-specific patterns. We also estimate a dynamic factor model assuming the presence of common job-finding, job-losing, participation-in and -out factors driving the relevant probabilities. We find that the job-finding and the job-losing factors have a significant effect on job-finding and job-losing transitions respectively for all age groups. We examine the impulse responses of the key labor market variables to a positive business cycle shock. First we work out the effects on the estimated factors, then on the transition probabilities, and then finally on the gross flows themselves. Our results show the importance of both job-finding and job-losing in the business cycle, but there is clear evidence that job-finding problems are in general more important than job-loss ones. Further, fluctuations in part-time employment over the business cycle appear more important than full-time fluctuations. The results suggest that in downturns of the business cycle, the main problems are the young unemployed having difficulty finding jobs, the young keeping jobs, the mature unemployed having more difficulty finding full-time work, and the old keeping full-time work. Job fluctuations in the business cycle are disproportionately associated with the young. Therefore labor market policies ought to be designed with all of these considerations in mind.