Threshold effects of inequality on economic growth in the US states: the role of human capital to physical capital ratio

ABSTRACT

Theory suggests that the effect of inequality on growth varies with the level of economic development, as captured by the ratio of human capital to physical capital. In particular, the effect is shown to be positive at lower levels of this ratio, and turns negative beyond a threshold in such models. Using a comprehensive panel of annual data for the 48 contiguous US states over the period 1948 to 2014, we find overwhelming evidence in support of this theory, unlike prior work on this topic. Hence, our paper highlights the importance of accurately measuring the process of economic development using data on human capital and physical capital, instead of using proxies that are not theoretically consistent. Understandably, if not done so, policymakers would end up undertaking incorrect decisions.

KEYWORDS:

• Inequality
• economic growth
• ratio of human capital to physical capital
• panel threshold model
• state-level data of the United States

JEL CLASSIFICATION:

• C23
• C24
• O11
• O15

Acknowledgments

We would like to thank an anonymous referee for many helpful comments. However, any remaining errors are solely ours.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ We would like to thank the anonymous referee for directing us to clarify this issue.

² Lin et al. (2009) had earlier obtained similar results at the country-level, again based on per capita real income as a proxy for the process of development.

³ We would like to thank Professor Steven Yamarik for kindly sharing the updated version of the data with us.

⁴ The results from the instrumental variable regressions have been reported in Table A1 of the Appendix of the paper. As can be seen from Table A1, all variables, except for HKH and HKC, are statistically significant.

⁵ Note that the HKH and HKC variables are in logarithms, with the raw values of HKH and HKC being fractions (as physical capital stock is larger than human capital stock), and hence, the thresholds which provide the logarithmic values for HKH and HKC are negative, and translates into 0.0012 and 0.0002 respectively for the raw data.

⁶ As a preliminary analysis, we estimated a linear model with an interaction term involving inequality and the two HK variables separately. The results of regressions with interaction term strongly confirmed that the marginal effect of inequality on growth constantly changed as HKH or HKC increased. When HKH or HKC is relatively, inequality and growth were found to be positively related, but as HKH or HKC increased along the process of economic development, the relationship became less positive over time. In other words, the statistically significant negative coefficient on the interaction term suggested that at higher levels of economic development, the link between inequality and growth may eventually turn from positive to negative – an observation formally confirmed by the nonlinear model. Complete details of these results are available upon request from the authors.

⁷ The Gini coefficient take values between 0 (0%) (where every person has the same income), and 1 (100%) (where all income goes to one person). Thus, a 1 point reduction translates into a 10 percentage point decrease in the Gini coefficient.