2004). Child labor rates range from 0% to 70.9%, with OECD members and some Arab and former
Communist bloc states having no or very low child labor, and many African and Asian states
tending to have high rates of child labor.
Measuring compliance with the ERC is particularly challenging. Compliant states are those in
which men and women earn the same for “work of equal value” – i.e., for the same work performed
under the same conditions for the same number of hours, etc. Obtaining such data for OECD
countries alone is quite difficult, but it is impossible to obtain such information for most non-
OECD countries. The ILO has some wage data (ILO 2004b), but unfortunately they suffer from
a number of problems.
24
Only about 60% of ILO members report any wage data to the ILO,
and only half of those that do report disaggregate the data by gender. Developed countries are
more likely to report gender-disaggregated wage data, so the missingness of the data is clearly not
random. Finally, the unit reported (i.e., hourly, daily, weekly, or monthly earnings) varies between
and sometimes within countries, and it is generally difficult to obtain gender-disaggregated data on
the average number of hours worked for a large number of countries. For these reasons, measuring
compliance with the ERC by examining the female/male wage ratio is not possible here.
I instead examine women’s labor force participation. While this is clearly not a perfect proxy
for the wage ratio, the connection is logical based on the presumption that supply is price-elastic.
25
Indeed, as O’Neill (1981, 76) explains, “An increase in the market wage relative to the housewife’s
‘wage’ induces a substitution of market work for home work.” In other words, as the wage women
can earn outside the home increases, more women will join the labor force. While the strength
of this effect depends on how substitutable market goods and home goods are, the relationship is
likely to be strong. And indeed, much of the empirical work has found that female earnings (also
controlling for male incomes) explain much of the variation in female labor force participation (c.f.,
O’Neill 1981; Macunovich 1996). Hence, while female labor force participation does not measure
perfectly compliance with the ERC, it is the best available proxy. The variable Female Labor Force
Participation measures the female labor force as a percentage of the total labor force (World Bank
24
Data from other sources, such as the UN, are equally if not more problematic. The UN has some wage ratio data
(which it generally has obtained from the ILO) for only a few years. If data are not available for a particular country
in those years, the UN assumes the wage ratio to be 75%. This is clearly very problematic for my purposes.
25
I am thankful to Edward Leamer for suggesting this to me.
17
Convention