Figure 1 Percentages of the total labour force in unemployment, labour market
programs and adult education in Sweden 1990-2002.
Unemployment---Labor market programs - - - Adult education
Note: ‘Adult education’ denotes the yearly number of individuals. Unemployment and labour
market programs are measured as yearly averages.
Empirical studies of the effects of AE in Sweden are relatively few despite the
fact that high quality data is available. One explanation for the lack of research
in this area is that the register data in komvuxdatabasen are relatively difficult
to arrange. Studies so far have therefore simplified the reporting of studies to
binary variables indicating term-wise registration at komvux. A drawback then
is of course that this is not necessarily a good proxy of the amount of studies
completed. It may also be a partial explanation as to why previous research,
presented below, has reported rather mixed results. Another unexplored issue
concerns the large fraction of AE participants that continue to higher studies. If
their performance is markedly weaker than average, a policy maker could pre-
fer to restrict their access to university studies. This could arise if e.g. AE is of
insufficient quality as preparation for university studies (Heckman and LaFon-
taine 2006, and Boesel et al. 1998), if adults experience more family- work-
and/or other commitments, if there are decreasing returns with age (Ekstrom
2003, Light 1995, Monks 1997) or if AE individuals tend to pursue education
due to bleak employment prospect (Stenberg 2007, and Stenberg and Wester-
lund, 2007).
The contribution of this report has two facets. The first is to evaluate the ef-
fects on annual wage earnings of adult upper secondary education at komvux.
IFAU - Does adult education at upper secondary level influence annual wage earnings?