This paper uses recently published top 1% income share series in studying the inequality-development association. The top income shares data are of high quality and cover about a century for some countries and thus provide an interesting opportunity to study slow development processes. The empirical inequality-development studies have started to call into question the use of parametric (quadratic) specifications. To address the issue of functional form, this study exploits penalized spline methods. The association between top 1% share and development is found to experience a reversal at later stages of development and, thus, a positive link is observed in many advanced economies. Although this study is not taking a strong stand on causality, additional analysis in this paper advocates that more than sectoral shifts are needed to explain distributional changes.
inequality top incomes development nonlinearity longitudinal data