From  Raj Nallari and Breda Griffith's lecture notes.

 

Empirical studies of gender inequalities and economic growth

Turning to the empirical studies, we note that the real value-added lies in the derivation of a correctly specified econometric model that examines the relationship between gender equalities and economic growth.  The simultaneity bias makes it difficult however to fully isolate the effect of gender equality on economic growth and vice versa, given the likely coincidence of variables explaining both growth and gender equality. Varying estimation techniques offers a way of getting around the simultaneity problem but often many of the techniques, for example the use of two stage least squares (2SLS) depends on rich sources of data that may not always be available, especially in developing economies, see Box below.

 

 

 

                                                                                                                                                                  

The results from the empirical studies suggest a three-fold categorization. First, there are a number of studies that show a positive relationship between gender equality and economic growth that suggests a win-win scenario, improve gender equality-improve growth. At the other end of the spectrum are those studies that posit a positive relationship between gender inequality and economic growth, a lose-win scenario. In between are those studies that examine facets of gender discrimination, which act as a brake on economic growth and once released economic growth ensues.  It is unsurprising that such a classification exists. The nature of gender inequalities/equalities rely on social norms and conventions as well as economic factors and are subject to change over time.  We examine an empirical study from each scenario below.   

 

The basic story that emerges from Dollar and Gatti (1999) is that gender equality and economic growth are mutually reinforcing.  “Societies that have a preference for not investing in girls pay a price for it in terms of slower growth and reduced income” (Dollar and Gatti, 1999). The authors find a ‘win-win’ scenario for gender inequality and growth – improve gender inequalities ð improve growth.  Using data from over 100 countries over three decades, the authors investigate the relationships between gender inequality, income and growth. Having noted that only 5 percent of adult women had any secondary education in the poorest quartile of countries in 1990, only half the level of men, compared with 51 percent of women of adult women with at least some secondary education in the richest quartile, or 88 percent of the level of men. The authors wanted to know what the macroeconomic data revealed about the following questions:

    1. is lower investment in girls’ education simply an efficient economic choice for developing countries?

    2. does gender inequality reflect different social or cultural preferences about gender roles; and,

    3. is there evidence of market failures that may lead to under-investment in girls, failures that may decline as countries develop?  

The first two questions aim to explain gender inequality across countries and over time, while the third question considers the possibility that gender inequality affects growth. Based on this, the authors specify two equations to be estimated: 

    (1)  git=α+βyit+Zγ+εit (examining the relationship between gender inequality and economic growth) 

    (2)  [˙y/y]it=δ+ψgit+Xπ+ui (examining the relationship between economic growth and gender       inequality)

 

 
where:

    g is some measure of gender inequality;

    y is per capita income and [˙y/y] is per capita income growth

    Z are exogenous variables that affect gender inequality

    X are exogenous variables that affect growth;

    ε  and u are error terms 

Both equations also allow for country specific fixed effects, based on the estimation technique. Providing we ensure that the variables that affect gender inequality (Z) and those that affect growth (X) are not the same, we can use instruments to ensure that OLS (ordinary least squares) estimation of both equations are not biased. 

Beginning with Equation 1, Dollar and Gatti (1999) build on the previous efforts to estimate such an equation by including several different measures of gender inequality, a panel of countries rather than a single cross-section countries and the fact that income is treated as endogenous.  The gender equality variables considered are secondary attainment, life expectancy, economic equality, equality in marriage and women in parliament. The explanatory variables include civil liberties or economic policy, religious preference and religion factors and regional effects.   

Dollar and Gatti (1999) find for gender inequality in education as measured by differences in female secondary attainment and controlling for male achievement and other variables, a convex relationship suggesting that increases in income lead to a narrowing of gender inequality in education. The shape of the relationship implies that the effect is minor as countries move from very low-income ($500 per capita) to lower middle income ($2,000 per capita). It is only when countries move from lower middle income to higher income ( that the positive relationship between income and educational attainment for females kicks-in. The authors find that differences in female educational attainment can in part be explained by regional differences, religious preferences and the extent of civil liberties in society.   

 

Source:  Dollar and Gatti (1999); p. 39.