Publications

The principal objective of this paper is to investigate the effects of changes in standard working hours and relative factor prices on the demand for workers, average hours worked and capital input. As opposed to neoclassical testbook models which treat hours and the number of persons employed as homogeneous input factors with equal elasticities of production and the same factor prices, a demand system is derived accounting for the technological difference between hours and workers as well as for the nonlinearity of the labor cost scheme. The "nonlinear" system can be expressed in terms of the standard linear system. This approach yields a simple test with respect to the second order conditions. An econometric factor demand system for employment, average hours worked and capital input generated by a nonlinear labor cost scheme is estimated on the basis of three flexible functional forms: the translog, the generalized Leontief and the generalized Cobb-Douglas system. In general, the estimates show that employment and hours are complements. A reduction of standard hours as well as the taxation of overtime work does not seem to be the appropriate policy to increase employment. The results are independent of the estimation procedure and the functional form chosen.

When a pay-as-you-go financed public pension system is replaced by a funded system or by private savings, it is generally impossible to compensate the pensioners in the transition generation for the loss incurred without making at least one later generation worse off. This is shown in a standard one-sector overlapping-generations model, both for a "small" open and for a closed economy. Our findings imply that the choice between financing methods for social security can not be reduced to the intergenerational Pareto-criterion but necessarily involves distributional conflicts among generations.

dc.title:

dc.contributor.author: König, Heinz

dc.title:

In periods of demographic change, pay-as-you-go financed social security systems imply transfers of lifetime income not only among generational cohorts, but also between families of different size and generational composition. Whereas previous models of voting on social security in democratic societies focused on the first type of transfer and assumed homogeneity of interests within each generation, we treat the family as the relevant decision-making unit. It is then analyzed how the results of majority voting on public pension and sickness funds depend on the rate of time preference, the overall rate of population growth and the distribution of children across families. Not surprisingly, opposition to mandatory social security turns out to be greatest when children are most unevenly distributed.

In the empirical estimation of hospital cost functions, two radically different types of specifications have been chosen to date, ad-hoc forms and flexible functional forms based on neoclassical production theory. This paper discusses the respective strengths and weaknesses of both approaches and emphasizes the apparently unreconcilable conflict between the goals of maintaining functional flexibility and keeping the number of variables manageable if at the same time patient heterogeneity is to be adequately reflected in the case mix variables. A new specification is proposed which strikes a compromise between these goals, and the underlying assumptions are discussed critically.

Die Hypothese der "anbieterinduzierten Nachfrage" auf den Gesundheitsmärkten behauptet, dass zunehmende Verfügbarkeit von Ärzten zu steigender Inanspruchnahme ärztlicher Leistungen führe, da die Ärzte als Anbieter die Möglichkeit hätten, die Nachfrage nach ihren Leistungen wesentlich selbst zu bestimmen und sich dabei ein vorgegebenes "Zieleinkommen" zu sichern. Ein positiver empirischer Zusammenhang zwischen Arztdichte und Inanspruchnahme kann jedoch außer auf gezielte Nachfrageausweitung noch auf mehrere andere Gründe zurückgehen. In diesem Aufsatz wird ein Modell des Arztverhaltens entwickelt, das explizit zwischen den verschiedenen Versionen des "Verfügbarkeitseffekts" unterscheidet und somit einen empirischen Test zwischen den alternativen Hypotheses für sein Zustandekommen ermöglicht.

dc.title: