Page 13 - Education for Development:George Psacharopoulos University of Illinois at Urbana-Champaign, USA

Education for Development: What Policies?c11

of the GDP according to one study (EFILWC, 2012), or 8% according to another

(Varly et al., 2014). It would produce 40% higher lifetime earnings in Estonia
(Anspal et al. 2014), 120,000 pounds in the UK (Oreopoulos, 2006), $8.2 billion
in Australia (Applied Economics, 2002), or 50,000 euros per Roma graduate in

Hungary (EU, undated).

Education quality

Average years of education may not be a sufficient statistic to predict
growth (Pritchett, 2001). A year of primary schooling in the UK compared to

Brazil will provide different learning outcomes.
On the micro side, cost-benefit analysis of education quality is not as
plenty as for education quantity. Many econometric studies have found that

increased resources for education (an input measure of school quality) have
not led to statistically significant improvements in test scores – a standard
measure of education quality. In a survey of 376 education production functions

relating school resources to student achievement, most studies report negative
or insignificant effects of expenditure per student, teacher salaries or class size
(Hanushek, 2003).

A review of 30 randomized control trials designed to improve test scores
in the developing world, found that two-thirds of them report near zero or
insignificant effects of alleged school quality enhancing interventions such as

textbooks, improved buildings or smaller class sizes (Kremer et al., 2013). A
meta-analysis of 76 quality-improvement experiments in developing countries
concluded that there are insufficient data to assess the relative cost-effectiveness

of interventions (McEwan, 2013).
On the macro side, Hanushek and Woessmann (2009) report a one