This paper argues that the conventional approach of data averaging is problematic for exploring the growth–inequality nexus. It introduces the polynomial inverse lag (PIL) framework so that the impacts of inequality on investment, education, and ultimately on growth can be measured at precisely defined time lags. Combining PIL with simultaneous systems of equations, we analyze the growth–inequality relationship in postreform China, finding that this relationship is nonlinear and is negative irrespective of time horizons.
