compulsory school or a two-year upper secondary level. One may argue that
these individuals experienced a positive payoff from AE but not from their en-
rolment in higher education.
Now, based on these results, let us consider the groups of AE participants
that were not associated with positive effects. Those with three-year upper sec-
ondary school and no registration in higher education represented 12 per cent,
while 11 per cent of the AE participants enrolled at university but completed
less than two years of higher education. Based on the reasoning in the preced-
ing paragraph, one could say that 23 per cent of the participants in AE did not
experience positive returns. However, as was just noted, within the 11 per cent
more than half (57 per cent) of the individuals came from groups that experi-
enced a positive payoff from AE per se. This would mean that the fraction with
non-positive returns would be reduced to something like 12 + 11*(1 - 0.57) =
17 per cent.
To sum up, the majority of the participants seem to have gained from AE.
However, large groups, around one fifth according to the back-of-the envelope
calculation above, appear to have experienced zero or even negative returns
from their AE. To reach any sort of conclusion about the total effects, it would
be necessary to resort to a social cost benefit analysis, a task that is fraught with
numerous measurement difficulties and beyond the scope of this report.
6 Concluding remarks
The purpose of this report is to analyze the effects on annual wage earnings of
upper secondary adult education (AE) in Sweden, conducted at the municipal
adult education centers komvux. As more than 40 per cent of the enrollees in
AE continued to higher education, a second purpose is to study whether the
returns to university studies differ between prior AE participants and non-par-
ticipants. Their respective accomplishments in higher studies are also consid-
ered.
The results in this study are all obtained in a setting where selection issues
must be addressed. Similar to earlier studies in this field, the data offers no ex-
ogenous variation in AE enrolment, so causal interpretations of the estimates
should therefore be made with caution. For the sample that did not continue to
higher education, a fixed effects regression model is used to control for time-
40
IFAU - Does adult education at upper secondary level influence annual wage earnings?