Subjective expectations as drivers of human capital nvestment

Many decisions are made under uncertainty, and people are likely to form subjective beliefs (expectations) about the probabilities of events that are relevant for their decisions. Preferences and expectations are then combined to reach a choice. Typically, researchers only observe the final outcomes and have data on choices. This leaves them with a basic identification problem when making inference on the decision-making process as many combinations of preferences and expectations can lead to the same observed choice. One way to address this problem is to ask decision-makers directly about their subjective expectations. This is particularly relevant for human capital investment, and MiSoc researchers are actively contributing to this line of research.

By collecting data on expectations on the return to a degree from youth in Pakistan, Delavande and Zafar (2019, published in the Journal of Political Economy) evaluated the role of future earnings, non-pecuniary outcomes and financial constraints in the choice of a university (between religious seminaries, Islamic universities or liberal universities). They find that future earnings and employment prospects play a small (but statistically significant) role. However, non-pecuniary outcomes, such as the school’s ideology, are the major determinants. Policy simulations suggest that implementing policies relaxing financial constraints would lead to large welfare gains and substantial switching in university choice.

Students’ effort at university is one of the primary inputs into the production function of higher education achievement. Yet, we still know relatively little about how university students spend their time, and even less on what determines their allocation choices, i.e. whether these choices are mainly driven by expectations about future outcomes, preferences or constraints. Delavande, Del Bono and Holford (2020, published in the Journal of Econometrics) take advantage of very detailed data on university time use and detailed subjective expectations about the academic and labour market returns of academic and non-academic inputs (such as time spent in lectures and volunteering). They document significant heterogeneity in student’s investment decisions along mainly the ethnic dimension, with non-white British students attending lectures and classes significantly less than white high-SES British students, and investing less time accumulating relevant work experience. Moreover, their results point to an important role of constraints in explaining differences in investment by ethnicity. Using simulations, the observed gaps in academic and non-academic investments would be substantially reduced if students had access to all feasible time allocation alternatives.

MiSoC Director Professor Del Bono and co-I Delavande are currently analysing 9 waves of expectations data about the education production function collected as part of the BOOST2018 cohort to understand the formation of educational expectations and the role of academic feedbacks and peers.

Expectations from parents about the most useful investments to make toward their children is also important in the production of new skills. Co-I Bhalotra and Co-I Delavande investigate the importance of subjective expectations of returns to and effort costs of the two main investments that mothers make in newborns: breastfeeding and stimulation. They find heterogeneity across mothers in expected effort costs and expected returns for outcomes in the cognitive, socio-emotional and health domains, and we show that thiscontributes to explaining heterogeneity in investments. Their findings, detailed here highlight the relevance of interventions designed to reduce perinatal fatigue alongside interventions that increase perceived returns to investments.