efforts in Tanzania in the 1970s and more recent successes. Increased provision
of primary education in the 1970s and 1980s was frequently followed by regres-
sion during the 1990s. Filmer (2007) estimates that very large increases in the
availability of schools would have little impact on school attendance. Other
factors like the demand for education or the perceived quality of education are
likely to be contributing significantly to low levels of school attendance.
Improved transparency
Between 1991 and 1995 there was a three-fold real increase in public spending
on basic education in Uganda, which had no effect on official enrolment data
or the quality of education. The reasonable suspicion was that these funds were
being misappropriated. This extra expenditure had led to higher teacher
salaries which tripled in real terms between 1991 and 1995. About 20 per cent
of government spending on primary education was non-wage expenditure in
the form of per-pupil grants, of which schools received on average only 2 per
cent, and at best less than 30 per cent, the bulk being retained by local govern-
ment officials and politicians. In practice two-thirds of the non-wage expendi-
tures of schools in the early 1990s was raised through parental contributions.
While teachers were well aware of how much they should be earning, there
was little general knowledge about the existence of the per-pupil grant. The
information problem was tackled in a novel manner. The central government
published monthly reports in newspapers of how much had been transferred to
schools and required primary schools to post public notices on all inflows of
funds. This promoted accountability, and preliminary evidence from an evalu-
ation of the information campaign shows a sharp increase the share of grants
reaching schools and being spent on educational materials (Ablo and
Reinikka,1998; Reinikka and Svensson, 2004).
Teacher absence
Teacher absence undermines any efforts to ensure quality rises with greater
provision of school places and enrolment. Without teachers present pupil
attendance at school cannot translate into learning. Fieldwork suggests two
major explanations for teacher absence: poor quality of learning infrastructure,
and the politics of interest groups.
Surveys of education (and health centres) in Bangladesh, Ecuador, India,
Indonesia, Peru and Uganda show that good-quality infrastructure (presence of
a library, running water, electricity and covered classrooms) reduced absence
from schools. Presumably such schools were nicer places in which to teach.
Teachers were also less frequently absent in schools where the parental literacy
rate was high, which probably implies some combination of greater demand
for education, political influence and monitoring ability by parents, a more
pleasant working environment for teachers (children better prepared), or selec-
tion effects whereby educated parents choose schools with low absence rates
Education and Health 127