| Abstract: | Abstract. This paper presents methods to analyze convergence in cross‐sectional data collected over time using distribution free statistics that are not sensitive to the magnitudes involved. Measures of concordance and discordance are employed in the empirical analysis of real personal income per capita for 48 U.S. States over the period 1929–2002. Although most States are converging with each other, some are converging faster than others. The methods used have the flexibility to focus on specific characteristics such as convergence in absolute differences or convergence in the ratio of rewards. The methods may also be used to consider convergence without switching and additionally be applied to other features such as the percentiles of the distributions. |
