23.07.2013 Views

samlet årgang - Økonomisk Institut - Københavns Universitet

samlet årgang - Økonomisk Institut - Københavns Universitet

samlet årgang - Økonomisk Institut - Københavns Universitet

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

92<br />

NATIONALØKONOMISK TIDSSKRIFT 2005. NR. 1<br />

educational system could potentially also affect time to degree completion although<br />

there is unlikely to exist much variation in the universal financial aid packages (SU)<br />

available to Danish students. Two variables capture financial aid generosity: GRANT,<br />

which is the individual's (annual) grant entitlement, obtained from SU publications,<br />

which varies by age (threshold 18), residential arrangements (living with parents or<br />

not) and year, and WINDOW, an indicator of being in the eligible window to receive<br />

financial aid, which for most years and for most individuals in our sample, is a period<br />

of 70 months.<br />

With regard to student ability variables, as student test score information was<br />

unavailable in the 0.5% sample, we proxy student ability by type of high school degree<br />

obtained, i.e. a regular high school degree, a high school equivalent degree or a degree<br />

from a technical high school. Ideally, we would like to have for example verbal and<br />

math ability variables which previous studies have shown to be highly correlated with<br />

performance in higher-educational programs, see for example Albæk (2002), for Denmark,<br />

but these data are unavailable in the 0.5% sample. Of course, a more fundamental<br />

control for ability is made by allowing for the existence of unobserved heterogeneity<br />

using a three-point discrete distribution.<br />

Finally, labour market variables that are assumed to significantly affect education<br />

completion times are the individual's real annual labour market earnings in kroner, real<br />

average (log) wages of workers who have completed the same degree and average<br />

weekly unemployment degrees of similar degree holders. 10 While information on the<br />

individual’s own labour market earnings is available in the 0.5% sample data, the last<br />

two are aggregates created from a 2% sample of the population (Danish IDA data), for<br />

each type of degree (short, medium, long), for each year in the sample period which<br />

are merged to our estimation sample by educational type and year. In addition, cohort<br />

size effects are captured by year of high school graduation variable, which varies by<br />

definition from 1980-1988.<br />

Background variables such as gender, type of high school degree, year of high<br />

school graduation and taking a sabbatical year after high school are time invariant.<br />

The time-varying covariates are cohabitation status, presence of small children, the<br />

student’s real annual labour market earnings, real average wage of same degree holders,<br />

average unemployment degree of same degree holders, type of education (short,<br />

medium or long), time between programs, first program or not and the financial aid<br />

variables.<br />

10. In Denmark, unemployment is typically recorded as a degree (0-1000) indicating the fraction of the year<br />

that the individual has been unemployed.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!