Exploring the link between household debt and income inequality: an asymmetric approach

ABSTRACT

We investigate the relationship between household debt and income inequality in the USA, allowing for asymmetry, using data over the period 1913–2008. We find evidence of an asymmetric cointegration between household debt and inequality for different regimes. Our results indicate household debt only responds to positive changes in income inequality, while there is no evidence of falling inequality significantly affecting household debt. The presence of this asymmetry provides further empirical insights into the emerging literature on household debt and inequality.

KEYWORDS:

• Household debt
• income inequality
• credit asymmetry
• nonlinear ARDL

JEL CLASSIFICATION:

• C22
• C51
• C52
• D31

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ See Treeck (2014) and Mian and Sufi (2015, pp.4–30) for detailed discussions on household debt and crisis.

² See Neftçi (1984) and Falk (1986).

³ See http://www.wid.world/#Database:

⁴ Intuitively, the ‘inverted Pareto coefficient’ measures the fatness of the upper tail of the income distribution and is generally represented as b = a/(a−1). For example, a coefficient of b = 2.78 in 2008 for the USA means that the average income above 1 million is equal to 2.78 million USD.

⁵ A simple scatter plot indicates a strong negative relationship between household debt and inequality under this regime, which we show in the Appendix as Figure A1. In addition, there is a change in the frequency of data collection for the household series in 1953 and so we will not include this regime in our post-war sample.

⁶ During this period, the overall inequality levels in the USA fell due to rises in the income of the bottom 90%, while the income of the top of the income distribution remained stagnant (see Goldin and Margo (1991)).

⁷ We exclude the post-war regime, where the relationship seems to have changed as discussed above.

⁸ Although 1980 is generally perceived as the period of liberalization, the tendency towards liberalization of the financial sector did begin in 1970s as discussed in Orhangazi (2015). We consider this regime for the purpose of comparison with the 1980s sample, representing a more wide spread liberalization of the economy.

⁹ A Wald test is used to test for the presence of an asymmetric cointegration.

¹⁰ See Pesaran, Shin and Smith (2001) for a detailed discussion.

¹¹ Moreover, we perform unit root analyses for all the selected samples, finding that none of the series are I(2). This implies that the asymmetric cointegration test is valid.

¹² This result has been well reported in the literature for the US households (Cox and Jappelli 1993; Crook 2001)

¹³ See Mian and Sufi (2012) for an extensive discussion on household deleveraging and the severity of financial crises.

Additional information

Funding

This work was supported by RANNIS (Iceland) and Institute for New Economic Thinking (INET).