Decomposing the factors influencing household debt: the case of China

ABSTRACT

This study discusses the issue of household debt in China based on the data of China household finance survey completed in 2015. The quantile regression model is adopted as the key research method, and the conclusions are derived are as follows. The increase in household debt in China is significantly correlated with several financial factors, including household assets and income, as well as the socioeconomic factors of education, age and working unit, in addition to senses of security and happiness. The assets of highly indebted households influence their debt in a positive way; the influence of income on household debt decreases as the quantile increases; the householder age has a negative effect across all household quantiles, and the debt holdings are less at older ages. The influence of household educational background dramatically increases from negative to positive following the indebtedness level, suggesting that enormous debt burdens may be generated from the cross effect of disadvantages in income and educational background. By decomposing the results of the quantile regression estimations, the evidence suggests that sufficient financial knowledge is essential in avoiding the financial distress from household debt.

KEYWORDS:

• Household debt
• quantile regression
• quantile regression-based decomposition
• financial knowledge

JEL CLASSIFICATION:

• D14
• C21
• D12
• D31

Disclosure statement

No potential conflict of interest was reported by the author(s).

Declarations and ethics statement

All authors have agreed to the published version of the manuscript.

Data availability statement

Restrictions apply to the availability of the data, which was obtained from [Survey and Research Centre for China Household Finance], available [at https://chfs.swufe.edu.cn/datacenter/apply.html] with the permission of [Survey and Research Centre for China Household Finance].

Notes

¹. The BIS statistics are openly available at [https://stats.bis.org/statx/srs/table/f3.1?p=20202&c=].

². In fact, many institutions have warned that the surge of China’s household debts is indeed more severe than the predicted level.

[https://www.afr.com/world/asia/pboc-warns-as-China-s-household-debt-hits-record-20191127-p53ekk]

³. The data and survey results were obtained from the Survey and Research Centre for China Household Finance (https://chfs.swufe.edu.cn/datacenter/apply.html)

⁴. Indicated by Lusardi and Tufano (2015) and Kumhof, Rancière, and Winant (2015).

⁵. For missing and questionable values (such as a negative income), if applicable, throw-back interpolation was performed to complete the dataset; however, since household assets and income work as the key variables across the entire empirical estimation and cannot be substituted by another variable or reference value, in order to avoid bias in the overall estimation result, the remaining missing values or negative values were deleted.

⁶. Please refer to Koenker and Bassett (1978), Brown, Garino, and Taylor (2013) and Wu, Wang, and Zhang (2018) for a detailed mathematical induction.

⁷. This study jointly adopts the assessment methodologies used in Borden et al. (2008) and Brown, Garino, and Taylor (2013).

⁸. As (${\bar{Y}}^1-{\bar{Y}}^0)$ is used in the original method (Melly 2006), this study replaces the symbol of the dependent variable with (HD¹_τ -HD⁰).

Additional information

Funding

This research received no external funding.