Academic journal article Economic Inquiry

Is There a "Double Bonus" from Reducing Inequality?

Academic journal article Economic Inquiry

Is There a "Double Bonus" from Reducing Inequality?

Article excerpt


In the 1990s, a number of empirical studies found an economically and statistically significant negative relationship between inequality and long-run growth. A huge literature that has investigated the effect of inequality on growth--much of it emphasizes the negative effect--can be broadly classified into three branches. (1) The first branch of the literature regards an endogenously determined redistributive policy as a catalyst (Bertola 1993; Persson and Tabellini 1994). It asserts that, in more unequal economies, greater distortion because of higher redistributive tax, which is determined endogenously by vote, is likely to impede more investment. The second branch of the literature argues that more inequality will lead to more sociopolitical instability, which in turn will discourage investment (Alesina and Rodrik 1994; Benhabib and Rustichini 1996; Grossman and Kim 1996). The third branch of the literature focuses on investment in human capital. (2) These papers show that, when credit markets are imperfect, more inequality implies less people being able to invest in human capital, which in turn implies slower growth.

The idea that low income inequality and high long-run growth are consistent objectives appeared to be a common agreement until the late 1990s when it started being challenged by the empirical literature. For instance, Forbes (2000) and Li and Zou (1998), although their focus is the effect in the short run (5-10 years), showed the positive effect of initial inequality on growth. Burro (2000) and Deininger and Squire (1998) separate their samples into two according to some criteria such as an income level or a degree of democracy, suggest the nonmonotonic effects of inequality on subsequent growth. Banerjee and Duflo (2003) also cast a doubt on the validity of a linear structure that is assumed in the empirical literature assessing the relationship between inequality and growth.

In the meantime, the third branch of the literature mentioned earlier has been extended by incorporating endogenous fertility. This new strand of the literature postulates that fertility and investment in human capital are jointly determined by the individuals' optimizing behavior, (3) and emphasizes the fertility difference between the rich and poor. It claims that more unequal economies have higher fertility differentials, accumulate less human capital, and hence have lower economic growth. (4) The endogenous fertility literature appears to reemphasize the negative effect of initial inequality on subsequent growth.

In this paper, we hope to further extend the literature by arguing that the effect of inequality on growth can be non-monotonic even when the fertility differential is taken into account. To this end, we employ an overlapping generations (OLG) model that follows a model by Ehrlich and Lui (1991) (hereafter E&L). A representative agent in their model maximizes her/his utility by choosing how much to invest in quality and quantity of human capital, caring about current and future consumption, as well as "companionship" by her/his children in the future. In contrast to E&L, who focus only on the representative agent, we look at heterogeneous dynasties that comprise an economy. By allowing education to have external effects, we define an economy as a set of dynasties that can affect one another through externality in education. It turns out that a dynasty's behavior differs according to the human capital stock it inherits, as well as the economy's average level of human capital. Hence, the initial distribution of human capital in an economy has strong implications for subsequent growth of that economy.

Our main interest is to assess, from our theoretical model, how inequality affects long run growth of an economy, say, over 25 years. However, this is a somewhat tricky exercise because 25 years correspond to only one generation in our three-period OLG model. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.