Different ways of looking at old issues: a time-series approach to inequality and growth

Abstract

In this article we propose an econometric approach that steers clear of parameter heterogeneity, omitted variable bias and endogeneity problems, from which suffers the econometric analysis of economic growth. We propose to investigate the relation between income inequality and economic growth in a cointegrated VAR setting and present an application to Belgium, US and Finland.

Acknowledgements

The authors would like to thank Katarina Juselius, Rafael Doménech, David de la Croix, Freddy Heylen, Gerdie Everaert, Jozef Plasmans and two anonymous referees for useful comments and suggestions. We would also like to thank Christian Valenduc, Markus Jäntti and all others who helped with the dataset. Any remaining errors are ours. We acknowledge support from the Federal Public Planning Service Science Policy, Interuniversity Attraction Poles Program – Belgian Science Policy [Contract No. P5/21].

Notes

¹ We thank an anonymous referee for pointing this out to us (and for the useful reference).

² This comment is equally valid for the analysis of unemployment rates (see e.g. Papell et al., 2000). We circumvented the problem by testing for the presence of a unit root using logit-transformed data. These constructed variables vary between zero and plus infinity. The conclusions were unchanged.

³ Income inequality in Belgium (before taxes) and Finland; enrolment in secondary education in Finland and the US; enrolment in tertiary education in Belgium and the US.

⁴ We report the conclusion concerning the cointegrating rank on the basis of the maximum eigenvalue test (95% significance level) in a footnote to Table 3.

⁵ Based on the Akaike Information Criterion we would choose 2 lags for Belgium and 3 lags for Finland. The qualitative results of the cointegrating analysis are fairly robust to these changes.

⁶ Belgium with inequality before taxes:  = 3.306 (p: 0.192); Belgium with inequality after taxes:  = 2.941 (p-value: 0.230); US:  = 7.913 (p-value: 0.019); Finland:  = 16.619 (p-value: 0.000).

⁷ According to Van den Noord and Heady (2001), the tax system in Belgium is highly progressive while the tax system in the US has only limited redistributive effects.