作者: Lars Johannessen Kirkebøen
DOI:
关键词: Demographic economics 、 Economics 、 Rank ordered logit 、 Identification (information) 、 Choice set 、 Higher education 、 Preference heterogeneity 、 Actuarial science 、 Earnings 、 Logistic regression 、 Differential (mechanical device)
摘要: Expected earnings are considered to influence individuals' choice of education. However, the presence nonpecuniary attributes and different set available prospective students make identification this relationship difficult. This paper employs a conditional logit model on exceptionally rich application data, which likely reflect actual preferences applicants, given their individual sets. Controlling for several attributes, average lifetime is shown strongly educational choice. A one-percent increase education increases number male applicants by about 5 percent female 2 percent. other also matter, in particular risk. Increasing both risk as they correlate cross section has essentially no effect applicants. Difference contribute gender differential. Finally, there some preference heterogeneity chosen.