The dissertation consists of two main parts. The first part presents a search model with non-stationary reservation wages. In the model the reservation wage declines over time and at a certain point is equal to utility of non-participation, after which it is not optimal to search and workers withdraw from the labour market. Thus, the potential time an unemployed may spend looking for a job is the solution to a maximisation problem and is treated in the model not as a random variable but as a choice variable. The potential search time is known at any point in time and it does not depend on whether an unemployed worker finds a job or not, but is observed only if a worker fails to find employment.
It is shown that changes in the variables which reduce reservation wages also shorten the potential search time. This establishes a trade-off in the model. On the one hand, lower reservation wages reduce unemployment duration, but on the other hand, agents have shorter spells of active search and hence higher exit rate from unemployment into nonparticipation. The effect of these measures on total employment is ambiguous in the model. Simulations show that there are cases when reduction of the unemployment benefits substantially reduces the potential search time and thereby the job-finding chances when arrival rate is low.

The second part presents a stationary search model where unemployed workers may look for jobs in two locations and have an option to change work place retaining their place of residence, i.e. commute. Introduction of commuting possibility into the model produces difference in search
intensities and reservation wages in origin and destination, i.e. agent sets different reservation wages for the "local" and "distant" region. Moreover, I allow for arrival rate to be the function of search intensity and therefore unemployed workers receive job offers only in those regions where their search intensity is nonzero. In this setting the maximal acceptable travel distance is determined by the condition of nonzero search intensity. The model shows that the circle of possible commuting destinations is bounded and the maximum distance a worker is willing to commute is determined by exogenous factors in the model. The paper demonstrates that positive commuter flows to a region are conditional on the location of the region within a maximal acceptable commuting distance thereby substantiating the use of "hurdle" models for empirical analysis of commuting. Empirical results based on estimation of the zero-inflated negative binomial model support the theoretical results.
The search model necessitates that search intensity and reservation wage depend on dispersion of wages. Therefore, unlike most of the empirical models on commuting and migration I use dispersion parameters of the wage distribution as additional regressors in an econometric specification. It will be shown that if the wage distribution is not symmetric and variance in two tails of the wages distribution can change independently of one another, the implications of the search theory slightly change. Namely, the dispersion in the left tail of the wage distribution reduces reservation wage and search intensity, and the dispersion in the right tail increases reservation wage and search intensity.

Die Dissertation besteht aus zwei Teilen. Der erste Teil entwickelt ein Modell mit nicht-stationären Reservationslöhnen. Es wird gezeigt, dass es einen Zeitpunkt gibt, nachdem es nicht mehr optimal weiter nach einem Job zu suchen, und die Suchenden verlassen den Arbeitsmarkt. Die potentielle Dauer, die man mit der Suche nach einem Jobangebot verbringt, ist somit die Lösung des Maximierungsproblems. Weiter wird gezeigt, dass Maßnahmen, die die Arbeitslosigkeit verringern sollten, die Partizipationsrate reduzieren. Der Effekt auf die gesamte Beschäftigung ist somit unbestimmt.
Im zweiten Teil wird das stationäre Suchmodell mit Pendeloption präsentiert. Das übliche Ergebnis von Suchtheorien ist, dass der Reservationslohn mit dem so genannten �Mean-Preserving-Spread� (Varianz, die den erwarteten Wert unverändert lässt) steigt. Eine Änderung der Varianz bei konstantem, Mittelwert, d.h. eine symmetrische Kompression oder Streckung der Lohnverteilung an beiden Enden der Verteilung, ist sehr unwahrscheinlich, wenn man mit realen Daten arbeitet. Als alternatives Maß wird der �Median-Preserving-Spread� untersucht, also die Varianz, die den Median unverändert lässt. Es wird gezeigt dass der Reservationslohn mit dem �Median-Preserving-Spread� im oberen Teil der Lohnverteilung steigt und im unteren Teil der Lohnverteilung sinkt. Die empirischen Ergebnisse auf Basis der Daten deutschen Pendlerströme belegen die Befunde des theoretischen Modells.