Mehmet Ali Soytas

Abstract This paper estimates the impact of parents' time investment in young children, their socioeconomic status and family structure on long-term outcomes of children. We developed and estimated a model of dynastic households in which altruistic individuals choose fertility, labor supply, and time investment in children sequentially, using data on two generations from the PSID. Parental behavior affects the education outcomes of children, their labor market earnings, and also their marriage market outcomes.

Abstract In this paper we developed and estimated a model of dynastic households in which altruistic individuals choose fertility, labor supply, and time investment in children sequentially, using data on two generations from the PSID. We then use the estimates to quantify the quality-quantity trade-offs and the return to parental time investment in children.

The literature on college major choices and the first jobs after graduation shows that the college major determines not only the initial labor market returns for the female but also gives indication of the lifecycle earnings. The close link between the initial employment and the possible employment path is investigated by many authors. The choice of major is followed by a choice of occupation after the graduation and the current model captures the job choice as a subcategory under the possible occupations available to a college graduate individual. There are different returns to different jobs, both immediately and also in the lifecycle (the age experience profile of each job differs as well as the periodic payments). There is limited switching between occupations but the switching becomes more costly as age increases, and specific human capital on a particular job is accumulated. Therefore a job choice for a college educated female is not a choice with immediate returns but a choice which has consequences in the lifecycle. In fact this is true not only for female individuals but also for males. However what makes it different for females compared to males is the observation that female and males are not equally represented in the occupations that are available to a college graduate. This point is discussed by many authors from the perspective of labor market discrimination against women. The evidence suggests that the discrimination argument is less valid now then it was 3-4 decades ago. The other view supports the idea that women in fact are sorted into some occupations which are historically known as female friendly jobs  Following this second line of view, a dynamic discrete choice model starting with an occupation choice in first period can be used to identify and estimate the returns to different occupations from choices and observed characteristics.. For instance the cost of having a child can be different for two females with same observed demographics but different occupations.

In this paper we estimate an intergenerational dynamic discrete choice model with labor supply, fertility and investment decisions using the estimation framework developed by Gayle, Golan and Soytas (2010). Our model consists of finitely lived agents who maximize expected lifetime utility. In each period the agents take actions in order to maximize discounted sum of utility over their remaining lifetime. Agents care about the future generations since they can increase their utility by increasing the utility of future generations. Moreover the timing of actions they take in their lifecycle also affects the utility value of future generations. We use data from PSID to estimate the model. The estimated parameters are used to simulate different outcomes from the model in order to conduct policy experiments. The reduced form estimates suggest college educated women have an advantage in producing highly educated children if they choose to spend time with them while the children are young. This hypothesized claim that females with college education has an advantage in educating their children is incorporated into the model by an intergenerational transfer function which is affected by the lifecycle time with children decision and labor supply decision of females. The magnitude of the utility is affected both by the cost of investing and at the same time by the discounted utilities from the invested generation. Therefore the final effect will depend on the parameters of the model. The lifecycle simulations for females with different education levels can reveal the different utility responses to the same set of choices. If the returns from the invested generation are high enough to compensate the loss due to not participating in the labor market, a college educated female may increase utility by spending more time with the children by decreasing the labor hours supplied to the market. In this case, these phenomena can explain the slowdown in the convergence of gender gap in participation.

In this paper we estimate an intergenerational dynamic discrete choice model with labor supply, fertility and investment decisions using the estimation framework developed by Gayle, Golan and Soytas (2010). Our model consists of finitely lived agents who maximize expected lifetime utility. In each period the agents take actions in order to maximize discounted sum of utility over their remaining lifetime. Agents care about the future generations since they can increase their utility by increasing the utility of future generations. Moreover the timing of actions they take in their lifecycle also affects the utility value of future generations. We use data from PSID to estimate the model. The estimated parameters are used to simulate different outcomes from the model in order to conduct policy experiments. The reduced form estimates suggest college educated women have an advantage in producing highly educated children if they choose to spend time with them while the children are young. This hypothesized claim that females with college education has an advantage in educating their children is incorporated into the model by an intergenerational transfer function which is affected by the lifecycle time with children decision and labor supply decision of females. The magnitude of the utility is affected both by the cost of investing and at the same time by the discounted utilities from the invested generation. Therefore the final effect will depend on the parameters of the model. The lifecycle simulations for females with different education levels can reveal the different utility responses to the same set of choices. If the returns from the invested generation are high enough to compensate the loss due to not participating in the labor market, a college educated female may increase utility by spending more time with the children by decreasing the labor hours supplied to the market. In this case, these phenomena can explain the slowdown in the convergence of gender gap in participation.

Models of dynastic households have been traditionally used to analyze persistence in earnings and wealth across generations, more recently to study patterns of wealth and fertility, transfers to children and education choices. Using data of two generations from the PSID, this paper develops and estimates a dynastic life-cycle model with endogenous fertility, labor supply and inter-generational transfers. Specifically, individuals choose fertility, labor supply and time investment in children sequentially. The focus of the empirical applications is on the effect of maternal time investment on children's labor market outcomes, and the quantity-quality trade-offs involved in fertility decisions across education groups, and households' characteristics. The estimator developed can be applied to a rich set of models with altruistic preferences.
We propose a framework for estimation of dynastic models by developing a new representation of the problem in terms of the model's primitives and choice probabilities which allows for the estimation of the problem in several steps. The framework we developed can be used to estimate a large class of dynastic models with endogenous labor supply, transfers to children, fertility, and household bargaining. To the best of our knowledge, this is the first paper to structurally estimated dynastic model with altruistic preferences.