Publikationen

This dissertation is a collection of three stand-alone research papers containing theoretical and empirical research on household portfolio choice and consumption decision in the life time. Household finance is a new interesting field in financial economics that has attracted a lot of attention among financial economists recently. Campbell even gave his presidential address titled "Household Finance" at the 2006 AFA annual meeting. Modeling household finance is different from traditional finance because household financial problems have many special features such as long but finite horizon, non-traded human capital, illiquid housing, borrowing constraints, and complex taxation. Also, household asset demands are important in determining asset prices. Therefore, household portfolio choice is a valuable topic for a doctoral dissertation.
Empirical analyses show that low-wealth households tend to be non-participants concerning risky assets and tend to hold a large portfolio share of risky asset if they participate at all (gamblers). So far, few theoretical models have been able to satisfactorily provide a joint understanding of these two observations. Chapter Two solves a novel life-cycle model in which social links to relatives, friends, and public welfare programs play the key role in explaining both types of behaviors. Social and family transfers induce low-wealth households to over-consume and depress the savings motive since social transfers are available only if personal savings are exhausted. The depressed investment demand further increases the fixed participation cost and thus, increases non-participation. On the other hand, if low-wealth households decide to hold any risky asset, they tend to hold rather large and risky investments because the down side risk is insured by help from relatives and public welfare.
Chapter Three numerically solves the optimal life-cycle portfolio choice when the model is calibrated to match the empirical retirement age distribution: people tend to retire at markedly higher rates around the firm's early retirement age and the age at which the full pension can be received. The model shows that financial incentives for keeping investors in labor force and low leisure preference for young investors endogenously restrict investors from retiring early. Thus, this chapter suggests a novel effect of the early retirement option on portfolio choice. As opposed to results from earlier models, the optimal portfolio share of stock does not increase monotonically prior to retirement. Wealthy investors might find it optimal to reduce the stock share in their early stage of life in order to decrease the possibility of having insufficient wealth for retirement when they are older. Thus, the model predicts either an increasing or a hump-shaped pattern for life-cycle stock holding, consistent with empirical observations.
Chapter Four presents new evidence contradicting the existence of the portfolio composition puzzles concerning household finance: portfolio risk is empirically increasing in age and wealth which is contradicting Merton's (1971) solution. The puzzles cause serious problems in assessing the classical theoretical models that have been developed to rationalize households' portfolio choices. This chapter investigates the 2005 Panel Study of Income Dynamics data and shows that, when the household portfolio includes real estate and private business and allows for leverage, the portfolio risk for young and low-wealth households is in general higher than old and rich households, which is consistent with the predictions of classical models.

This dissertation is a collection of three stand-alone research papers on the composition of human capital in newly founded firms. The papers are all empirical but they are closely related to two theoretical approaches. The first approach is the jack-ofall-trades model by Lazear (2005) and the second the O-ring theory introduced by Kremer (1993) and applied to new firms by Fabel (2004a,b) and Fabel and Weber (2005). Besides contributing to the search of stylised facts about the effects of the composition of human capital in new firms, this dissertations aims at discovering to what extent the predictions of these theoretical approaches can be confirmed by the data. In the introduction (Chapter 1), it is motivated why there is interest in new firms. The three papers are included in Chapters 2 to 4.
In Chapter 2, it is analysed whether heterogeneity in the educational backgrounds of the founders matters for the success of academic spinoffs. Furthermore, it is examined whether team foundations are more successful than single entrepreneurs. These questions are analysed using a data set on academic spinoffs in Germany. Firm success is measured by employment growth. The results show that team foundations are more successful than single entrepreneurs. Team foundations of engineers perform better when they have a business scientist in the team. However, different subjects per se and homogeneity with respect to the academic origins of the founders do not play a significant role for the success of academic spinoffs.
In Chapter 3, it is investigated to what extent the predictions of the O-ring theory are supported by the data. The O-ring theory predicts that individuals sort between firms according to their level of ability and that a higher average ability level within firms is positively related to both the number of individuals in the firm and capital per head. For the analysis, a rich register data set is used, covering the whole population of firms founded in Denmark in 1998 as well as all individuals involved in these new firms in the start-up year and in the following three years. In order to analyse the extent of sorting of individuals between firms, statistical tests are constructed, which compare the actual distribution of individuals among firms with the distribution resulting from random assignment of individuals to firms. The results show that, contrary to the prediction of the theory, individuals with differentlevels of ability tend to team up in new firms. Also contrary to the prediction of the theory, firm size and average level of ability of the involved individuals turn out to be negatively related. The only hypothesis that is confirmed by the data is the positive relationship between capital per head and the average level of ability in a firm.
In Chapter 4, the implications of the O-ring theory for the survival of new firms are considered. The theory assumes that (given team size) average ability in a team is positively and (given ability) team size is negatively related to firm survival. Moreover, it can be inferred that a higher level of homogeneity with respect to ability and a higher level of heterogeneity with respect to the field of education leads to higher survival chances of new firms. Using the same data as in Chapter 3, it turns out that both the average level of ability in a team and team size have positive effects on a firms' probability to survive. Most important is the fact that a firm is founded by a team at all. In contrast, homogeneity with respect to ability and heterogeneity with respect to educations do not affect the probability of firm survival. It can beconcluded that the main reason why most of the hypotheses tested in Chapter 3 fail is that an additional person does not increase firm failure.