Contributions published at Statistics and Empirical Economic Research (Rainer Winkelmann)

Preise, Einkommen, und Nachfrage – die zentralen zu erklaerenden Variablen in den Wirtschaftswissenschaften sind objektiv messbar. Daneben gibt es jedoch zwei nicht minder zentrale Groessen, Erwartungen und Praeferenzen, die sich einer direkten objektiven Messung entziehen. Es sind subjektive Variablen, also solche, die von persoenlichen Meinungen und Erfahrungen bestimmt sind. Allerdings bedeutet dies nicht, dass sie ueberhaupt nicht messbar waeren. Ganz im Gegenteil enthalten typische Haushaltsbefragungen eine ganze Reihe von subjektiven Fragen, die ueber verschiedene Aspekte von Erwartungs- und Praeferenzbildung informieren. Im folgenden werde ich darstellen, welche Arten von subjektiven Variablen bisher in der empirischen Wirtschaftsforschung betrachtet wurden. Dann werde ich auf Probleme bei der Interpretation von subjektiven Variablen eingehen. Und schliesslich werde ich die wesentlichen Aspekte der oekonometrischen Modellbildung besprechen und zwei neue Modelle vorstellen.

The paper studies the demand for foreign university graduates at the firm level. Using a unique dataset on recruitment policies of firms in four European countries, the determinants of demand for internationally mobile highly skilled employees are established. I investigate the number, origin, skills, and functions of foreign graduates, as well as the experiences of firms recruiting internationally. A number of hypotheses for the international demand are formulated and assessed. Foreign highly-skilled employees are recruited mainly because of their special skills that are not available domestically, be it international competence or technical know-how.

This paper is concerned with a comparison of immigrants and Swiss citizens with respect to level of education, labor market outcomes and healthcare utilization. The evidence is based on data for 1999 from the first wave of the Swiss Household Panel. In order to control for confounding influences, linear and non-linear (negative binomial) regressio nmodels are used. The main result is that differences in economic position between immigrants and Swiss nationals tend to be smaller than those found in other countries. The observed differences (higher employment levels of immigrant women, lower earnings of immigrant men, higher healthcare utilization rates of all immigrants) tend to be no larger than those observed between Swiss citizens living in different parts of the country.

This article is concerned with the analysis of correlated count data. A class of models is proposed in which the correlation among the counts is represented by correlated latent effects. Special cases of the model are discussed and a tuned and efficient Markov chain Monte Carlo algorithm is developed to estimate the model under both multivariate normal and multivariate-t assumptions on the latent effects. The methods are illustrated with two real data examples of six and sixteen variate correlated counts.

We present results of a negative binomial model on the determinants of the number of days of absence in a given year for a sample of 2049 workers drawn from three factories. We find evidence of the terms of the remuneration contract being important and we offer an interpretation of the differential effect of the company sickpay scheme on the behaviour of workers contracted to work four or five days a week.

This paper discusses the specification and estimation of seemingly unrelated multivariate count data models. A new model with negative binomial marginals is proposed. In contrast to a previous model based on the multivariate Poisson distribution, the new model allows for over-dispersion, a phenomenon that is frequently encountered in economic count data. Semi-parametric estimation is possible if some of the assumption of the fully specified model are violated.

This paper examines two competing explanations for workers' absenteeism, the shirking hypothesis and the adjustment-to-equilibrium hypothesis. Data on German workers for 1985-88 from the German SocioEconomic Panel are used in order to estimate the determinants of workers' absenteeism. The results indicate that firm size matters after wage effects are controlled for. This evidence supports the shirking hypothesis.

This paper reports quantitative information on the effects of tariffs and quotas on prices of individual goods. The analyses uses the natural experiment provided by a comprehensive unilateral trade policy reform in New Zealand to examine the response of foreign exporters to an incident of liberalisation that is unique in the developed world. The price effects of tariffs and quotas are estimated using a multidestination 7-digit longitudinal product-level dataset on export values and quantities. The effects are found to be by no means equivalent: whereas tariffs display no significant effect, the impact that quantitative restrictions have on the terms-of-trade of the country that imposes them are unequivocally detrimental and quantitatively important.

The macro evidence of increased adjustment pressure since the early 1970s suggests that job mobility should have increased. Hence, retrospective and spell data from the German Socio-Economic Panel are combined in order to test the hypothesis that job stability for German workers declined between 1974 and 1994. Using count data regression models in which we control for labour market experience, various demographic factors, and occupation, we find that job stability did not decrease, but if anything increased, between 1974 and 1994. Our finding suggests that labour market inflexibility is an important factor in explaining the European unemployment problem.

In this paper, I discuss the interpretation of qualification related income differentials from income functions that control for age rather than the theoretically more appropriate variable years of labour market experience. I reestimate income functions for New Zealand that were originally reported by Maani (1997) with changed specification, and I update her results by also estimating an income function with data from the 1996 census.

This paper is concerned with the problems of posterior simulation and model choice for Poisson panel data models with multiple random effects. Efficient algorithms based on Markov chain Monte Carlo methods for sampling the posterior distribution are developed. A new parameterization of the random effects and fixed effects is proposed and compared with a parameterization in common use, and computation of marginal likelihoods and Bayes factors via Chib’s (1995) method is also considered. The methods are illustrated with two real data applications involving large samples and multiple random effects.

This paper tests for the importance of non-pecuniary costs of unemployment using a longitudinal data-set on life-satisfaction of working-age men in Germany. We show that unemployment has a large detrimental effect on satisfaction after individual specific fixed effects are controlled for. The non-pecuniary effect is much larger than the effect that stems from the associated loss of income.

The recent economic literature on the incidence of various forms of post-secondary on-the-job and off-the-job training in Germany and the United States, as well as on the effects of training on wages, inequality, and labor mobility is surveyed. Young workers in Germany receive substantially more company-based (apprenticeship) training than United States workers. In the United States, high turnover deters firms from investing in general skills while it results in improved job matches. The received literature consents that key institutional elements required to make the German apprenticeship system work are absent in the United States.

This article contributes to the ongoing debate on native wage impacts of immigration. I propose a mobile-fixed factor distinction as a framework in which to think about the differential impact of immigration on various labor market groups. Skilled workers are treated as a fixed factor of production since the strong reliance on skill certification in Germany inhibits mobility and shelters from competition. Unskilled workers, in contrast, receive competitive wages. Using data from the German Socio-Economic Panel for 1984–1989 I estimate panel wage regressions for groups of workers separated by skill certification. I find that university graduates‘ wages increase, and the wages of workers without postsecondary degree decrease, as the industry share of unskilled workers increases. The effect for apprentices is ambiguous.

A new approach for modeling under-reported Poisson counts is developed. The parameters of the model are estimated by Markov Chain Monte Carlo simulation. An application to workers absenteeism data from the German Socio-Economic Panel illustrates the fruitfulness of the approach. Worker absenteeism and the level of pay are unrelated, but absence rates increase with firm size.

Using data from the German Socio-Economic Panel for 1984-90, the author analyzes the entrance of young individuals into the German labor market, comparing the experience of apprenticeship graduates to that of graduates from universities, full-time vocational schools, and secondary schools. Apprentices experienced fewer unemployment spells in the transition to their first full-time employment than did non-apprentices. Among apprentices, those trained in large firms had the smoothest transition to employment; once employed, however, apprentices (whether they stayed in their training firm or not) and non-apprentices had similar job stability (as measured by tenure). An estimated 70% of apprenticeship trainees left their training firm within a five-year period. These findings are consistent with the view that apprenticeship training develops general, portable skills rather than firm-specific skills

We use data from the German Socio-Economic Panel to investigate how individual happiness is affected by unemployment. Unemployment has a large and negative effect even after controlling for individual specific fixed effects. Nonparticipation, in contrast, is much less harmful to happiness. Further, we decompose the total well-being costs of unemployment and find that well above three quarters are non-pecuniary, and below one quarter pecuniary. One implication is that income support programs for the unemployed do very little at mitigating the adverse effects of unemployment, and such transfers are unlikely to generate unemployment.

This paper explores the relation between non-exponential waiting times between events and the distribution of the number of events in a fixed time interval. It is shown that within this framework the frequently observed phenomenon of overdispersion, i.e. a variance that exceeds the mean, is caused by a decreasing hazard function of the waiting times, while an increasing hazard function leads to underdispersion. Using the assumption of i.i.d. gamma distributed waiting times, a new count data model is derived. Its use is illustrated in two applications: the number of births, and the number of doctor consultations.

Key demographic variables, such as the number of children and the number of marriages or divorces, can only take integer values. This papers deals with the estimation of single equation models in which the counts are regressed on a set of observed individual characteristics such as age, gender, or nationality. Most empirical work in population economics has neglected the fact that the dependent variable is a nonnegative integer. In the few cases where this feature was recognized, the authors advocated the use of the Poisson regression model. The Poisson model imposes, however, the equality of conditional mean and variance, a restriction which is often rejected by the data. We propose a generalized event count model to simultaneously allow for a wide class of count data models and account for over- and underdispersion. This model is successfully applied to German data on fertility, divorces and mobility.

The study introduces a distinction between two types of labor mobility: direct job to job changes (which are assumed to be voluntary) and job changes after experiencing an unemployment spell (assumed to be involuntary). Exploiting the close relationship between those two phenomena we adopt a bivariate regression framework for our empirical analysis of data on male individuals in the German labor market. To account for the non-negative and discrete nature of the two counts of job changes in a ten year interval a new econometric model is proposed: the bivariate Poisson regression proves to be superior to the univariate specification. Further, the empirical content of distinguishing between two types of mobility is subject to a test, and, in fact, supported by the data: The hypothesis that both measures are observationally equivalent can be rejected.