Leverage, Hand-to-Mouth Households, and Heterogeneity of the Marginal Propensity to Consume: Evidence from South Korea

Abstract

This study examines the link between households’ leverage, their liquid assets such as cash holdings or checking accounts, and the Marginal Propensity to Consume (MPC) out of income changes (the effect of an additional dollar of income on consumption) for South Korean households over 2012–2017. To build on previous studies, we precisely redefine the hand-to-mouth households (defined broadly as those who hold little or no liquid assets), and examine the asymmetric effects of the direction of income changes on MPCs. By using the methodology of Hansen (2000), we estimate the threshold of the liquidity ratio (defined as liquid assets to monthly after-tax income) and find that households with liquid assets less than about 2 months of after-tax income show higher MPCs than others. We also find that the leverage ratio and newly defined hand-to-mouth status show asymmetric effects on MPCs by the direction of income changes. The MPC of households with positive income changes is smaller than that of those with negative ones. Furthermore, hand-to-mouth households with high leverage are much more sensitive to negative income changes than positive ones. This result suggests that in economic circumstances where households are highly in debt and have insufficient liquid assets, consumption is likely to be vulnerable to negative income changes, which could hamper aggregate spending growth.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

Notes

  1. 1.

    Flodén et al. (2017) and IMF (2017) show that in economies, where a large share of households hold debt, monetary policy has the direct and indirect effects on household spending via its effects on households’ cash flow and disposable income.

  2. 2.

    Since the cash holdings of households are sensitive information, they are not provided in the publicly released HLC data. However, the bank of Korea members can access the raw dataset of the HLC survey for the purpose of research.

  3. 3.

    Since old retirees can affect the main results of this paper due to their tight financial conditions and relatively high MPCs, we perform all analysis with the households between the ages of 22–65. According to the results, although the estimated MPCs are slightly smaller than those with ages between 22 and 79, the general results are comparable.

  4. 4.

    Strictly speaking, whether the landlord can borrow against the house depends on the ratio of the tenancy deposit to the housing price. If the ratio is much lower than the loan-to-value (LTV) regulation of the mortgage market, the landlord can borrow more against the house in the financial market. However, since this is generally higher than or similar to the LTV regulation, it is unusual for landlords to borrow more by using the Jeonse house as collateral in the mortgage market.

  5. 5.

    Shin and Kim (2013) study the roles of the Jeonse rental system in the credit market of South Korea.

  6. 6.

    The main purpose of this policy is to promote the transparency of retail sellers’ sales.

  7. 7.

    According to a survey conducted by the Bank of Korea, about 71% of payment transactions were carried out with credit or debit cards in 2016, and this share shows little variation across age groups.

  8. 8.

    The definitions of liquid assets of Kaplan et al. (2014) differ across countries due to differences in survey designs. The definition of liquid assets for the euro area, which includes cash, sight (also called current, draft, or checking) accounts, mutual fund holdings, shares in publicly traded companies, and bond holdings, is similar to the definition used in this study.

  9. 9.

    Two types of loans are granted by credit card companies, card loans and cash advances. Card loans are similar to the unsecured loans granted by banks but have relatively small limits and high loan rates; cash advances are urgent loans with high loan rates that provide cash liquidity to customers. The credit limit of cash advances is generally one-third of the total limit of the credit card.

  10. 10.

    Park (2017) estimate the shares of HtM households using Korean Labor and Income Panel Study (KLIPS) data over the period 2001–2013. According the them, the share of HtMs is about 32% in 2013, which are slightly larger than our results. This may be due to the difference of definition of liquid assets.

  11. 11.

    According to Statistics Korea, the peak of housing rental prices was between the end of 2015 and the beginning of 2016 and the housing rental price has shown a decreasing trend since then. In this period, two phenomena are observed in our dataset. First, the illiquid savings of HtMs who move from wealthy to poor HtMs decreased. Second, their monthly pay increased significantly in the same period. From these two phenomena, we can infer that the fluctuations in housing rental prices affect the share of poor HtM households.

  12. 12.

    It is known that the relation between family size and consumption is non-linear (e.g. Atkinson et al. 1995). However, there is no concensus on how to capture the non-linear association. To reflect it roughly, we include family size squared as an independent variable.

  13. 13.

    The MPC differentials between expected and unexpected income shocks are also an important issue. The first-stage regression in Section 4.1 can be understood as a step toward ruling out expected income changes. However, as the residual term of Eq. (3) can have both components, as mentioned by Kaufmann and Pistaferri (2009), this approach is not perfect. In future research, this issue could be addressed using detailed data on income changes.

  14. 14.

    We conduct Wald tests for the MPCs of overall income changes to examine whether the MPC differentials among the groups are significant. The Wald statistics are 2.78 (columns (2) and (3)), 44.88 (columns (3) and (4)), and 29.00 (columns (6)–(8)). These are statistically significant at least at the 10% level.

  15. 15.

    We adopt the sample split method instead of interaction terms to analyze the heterogenous MPCs between households with positive and negative income changes for two reasons. First, the endogeneity problem is weak because the estimated income changes can be regarded as income shocks. In the first-stage regressions described in Section 4.1, we rule out the deterministic changes of income, and this process is for the extraction of the idiosyncratic income shock of each household. Second, there are no major differences in the results between the two methods. Since all the independent variables used in the analysis are of interest to us, to use the combined model, we have to include all the interactions between the sign of the income change and each independent variable. In this case, the two methods are equivalent. We thus use the sample split method for brevity of reporting and interpreting the results.

  16. 16.

    Among households with no debt, two kinds of extreme households coexist in terms of borrowing constraints: those who can borrow but have no need for debt and those who cannot borrow for some reason (e.g., a low credit score). The latter may show higher MPCs owing to their low credit accessibility than the former. If households with no debt are included, the association between MPC and leverage can be deteriorated due to these households with characteristics of the extremes.

  17. 17.

    The LTV regulation of Korea that varied with regions and financial sectors ranged between 0.5 and 0.7 over the analysis period. Therefore, we set high-leveraged households as households with a leverage ratio greater than 0.6.

  18. 18.

    One plausible criticism of this finding is that the significant and positive relationship between unsecured loan rates and MPCs comes from high correlation between leverage ratio, HtM status, and unsecured loan rates. The correlation coefficients between leverage ratio, liquidity ratio, and unsecured loan rates are only 0.025 and −0.075, respectively.

  19. 19.

    The Korea Federation of Banks has provided the interest rate spreads for banks, financial products, and credit grades every month since March 2013. The values in this figure are calculated by averaging the interest rate spreads of banks in South Korea by financial products and credit grades. The periods are December 2015, 2016, and 2017. Since the majority of collateral loans are mortgage loans, the difference in loan rates between collateral and mortgage loans is small.

  20. 20.

    Of course, if the negative income change is so large that the household has to consume its savings, the threshold would be around 0.5. If the negative income change is, however, sufficiently small and temporary to be handled without draining savings, the threshold would be larger than 0.5.

  21. 21.

    In Table 6 we investigate the effects of cash buffers on MPC heterogeneity, using the HtM definitions of Kaplan et al. (2014). Compared to the results with three thresholds, the \({R}^{2}\) and the coefficients of HtM-related variables in Table 6 are smaller than those in Table 5. This suggests that the newly defined HtM households reflect MPC heterogeneity better than Kaplan et al.’s (2014).

References

  1. Andersen, A. L., Duus, C., & Jensen, T. L. (2014). Household debt and consumption during the financial crisis: evidence from Danish micro data. Danmarks Nationalbank Working Papers, 2014-89.

  2. Atkinson, A. B., Rainwater, L., Smeeding, T. M., et al. (1995). Income distribution in OECD countries: Evidence from the Luxembourg income study. OECD Social Policy Studies, No 18, Paris.

  3. Baker, S. R. (2014). Debt and the Consumption Response to Household Income Shocks. SSRN Research Paper, 2541142, 1–46.

    Google Scholar 

  4. Bunn, P., LeRoux, J., Reinold, K., & Surico, P. (2017). The consumption response to positive and negative income shocks. Journal of Monetary Economics, 96, 1–15.

    Article  Google Scholar 

  5. Caballero, R. J. (1991). Earnings uncertainty and aggregate wealth accumulation. The American Economic Review, 81(4), 859–871.

    Google Scholar 

  6. Carroll, C. D. (1997). Buffer-stock saving and the life cycle/permanent income hypothesis. The Quarterly Journal of Economics, 112(1), 1–55.

    Article  Google Scholar 

  7. Carroll, C. D., Hall, R. E., & Zeldes, S. P. (1992). The buffer-stock theory of saving: some macroeconomic evidence. Brookings Papers on Economic Activity, 1992(2), 61–156.

    Article  Google Scholar 

  8. Carroll, C. D., & Kimball, M. S. (1996). On the concavity of the consumption function. Econometrica: Journal of the Econometric Society, 64, 981–992.

    Article  Google Scholar 

  9. Carroll, C. D., & Kimball, M. S., et al. (1996). On the concavity of the consumption function. Econometrica, 64(4), 981–92.

    Article  Google Scholar 

  10. Christelis, D., Georgarakos, D., Jappelli, T., Pistaferri, L., & Van Rooij, M. (2019). Asymmetric consumption effects of transitory income shocks.The Economic Journal, 199(622), 2322–2341.

    Article  Google Scholar 

  11. Deaton, A. (1991). Saving and liquidity constraints. Econometrica, 59(5), 1221–1248.

    Article  Google Scholar 

  12. Dynan, K. (2012). Is a household debt overhang holding back consumption? Brookings Papers on Economic Activity, 2012(1), 299–362.

    Article  Google Scholar 

  13. Fagereng, A. & Halvorsen, E. (2016). Debt and household consumption responses. Norges Bank Staff Memo, 2016(1).

  14. Flodén, M., Kilström, M., Sigurdsson, J. & Vestman, R. (2017). Household debt and monetary policy: revealing the cash-flow channel. Swedish House of Finance Research Paper, 16–18.

  15. Fuster, A., Kaplan, G. & Zafar, B. (2018). What would you do with $500? Spending responses to gains, losses, news and loans. NBER Working Paper No. 24386.

  16. Hansen, B. E. (2000). Sample splitting and threshold estimation. Econometrica, 68(3), 575–603.

    Article  Google Scholar 

  17. Hara, R., Unayama, T., & Weidner, J. (2016). The wealthy hand to mouth in Japan. Economics Letters, 141, 52–54.

    Article  Google Scholar 

  18. IMF. (2017). Household debt and financial stability. In Global Financial Stability Report Chapter 2. Washington, DC.

  19. Jappelli, T., & Pistaferri, L. (2014). Fiscal policy and MPC heterogeneity. American Economic Journal: Macroeconomics, 6(4), 107–36.

    Google Scholar 

  20. Kaplan, G., & Violante, G. L. (2010). How much consumption insurance beyond self-insurance? American Economic Journal: Macroeconomics, 2(4), 53–87.

    Google Scholar 

  21. Kaplan, G., & Violante, G. L. (2014). A model of the consumption response to fiscal stimulus payments. Econometrica, 82(4), 1199–1239.

    Article  Google Scholar 

  22. Kaplan, G., Violante, G. L., & Weidner, J. (2014). The wealthy hand-to-mouth. Brookings Papers on Economic Activity, 2014(1), 77–153.

    Article  Google Scholar 

  23. Kaufmann, K., & Pistaferri, L. (2009). Disentangling insurance and information in intertemporal consumption choices. American Economic Review, 99(2), 387–92.

    Article  Google Scholar 

  24. Kukk, M. (2016). How did household indebtedness hamper consumption during the recession? Evidence from micro data. Journal of Comparative Economics, 44(3), 764–86.

    Article  Google Scholar 

  25. Mian, A., Rao, K., & Sufi, A. (2013). Household balance sheets, consumption, and the economic slump. The Quarterly Journal of Economics, 128(4), 1687–1726.

    Article  Google Scholar 

  26. Park, K. Y. (2017). The wealthy hand-to-mouth households in South Korea. Global Economic Review, 46, 299–324.

    Article  Google Scholar 

  27. Shin, H. S. & Kim S. J. (2013). Financing growth without banks: Korean housing repo contract. In 2013 Meeting Papers, number 328. Society for Economic Dynamics.

Download references

Acknowledgements

I would like to thank the coordinating editor, Prof. Charles Horioka, and two anonymous referees. For comments and discussions, I am grateful to the participants at the interim seminar of the Bank of Korea and the joint workshop of BOJ-IMES and BOK-ERI. All errors are mine.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Sang-yoon Song.

Ethics declarations

Conflict of interest

The author declares no conflict of interest.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

Figure 6 and Tables 6, 7

Fig. 6
figure6

The composition of households with negative and positive income changes across net worth and income quintiles. This figure shows the composition of households with negative and positive income changes across net worth and after-tax income quintiles over the period 2013–2016. See Section 3.2 and 4.1 for the definitions of positive and negative income changers

Table 6 HtM status and MPC heterogeneity
Table 7 Aggregated models

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Song, Sy. Leverage, Hand-to-Mouth Households, and Heterogeneity of the Marginal Propensity to Consume: Evidence from South Korea. Rev Econ Household 18, 1213–1244 (2020). https://doi.org/10.1007/s11150-019-09470-1

Download citation

Keywords

  • Marginal Propensity to Consume
  • Leverage
  • Hand-to-Mouth Households

JEL Codes

  • D12
  • D14
  • D90
  • E21