Occupational Choice: An Application to the Market for Public School Teachers”, Quarterly

Most previous work on occupational choice does not satisfactorily treat the potentially important effects of future demand conditions. In contrast, this paper develops and estimates a rational expectations model in which agents look beyond expected starting salaries and take explicit account of future demand conditions. The empirical results demonstrate that future demand conditions are important determinants of the decision to acquire secondary school certification but are not important for the decision to acquire elementary school certification. The paper concludes with a comparison of the dynamic properties of the rational expectations model and a cobweb specification.... But it is very difficult to ascertain the causes that are likely to determine the distant future of the trades which they are selecting for their children; and there are not many who enter on this abstruse inquiry. The majority assume without a further thought that the condition of each trade in their own time sufficiently indicates what it will be in the future; and, so far as the influence of this habit extends, the supply of labor in a trade in any one generation tends to conform to its earnings not in that but in the preceding generation.-Alfred Marshall [1961], Book VI, Ch. V, Section 2. In studies of occupations requiring an investment in training, it is necessary to deal with expectations of the future in some way. The reason for this is that training entails substantial costs in terms of direct and indirect expenses such as tuition and for-gone opportunities, while the returns from this training accrue in the future. In this regard, occupational choice has a capital.. like feature much like an investment in a specialized machine. However, most previous work in this field does not satisfactorily treat the potentially important effects of future market conditions on occupational decisions. ' The purpose of this paper is to incor-*I thank especially Sherwin Rosen and Robert Topel for helpful discussions and suggestions. In addition, I would like to thank John Geweke, Kent Kimbrough