top

Asiasn review of Financial research

Past Issues

HOMEPast Issues Past Issues

Asian Review of Financial Research Vol.26 No.4 pp.527-559
The Association of Household Debts and Financial Asset Holdings
Wonnho Choi Visiting Researcher, Institute of Finance and Banking, Seoul National University
Key Words : Household Debts,Financial Asset Holdings,Household Income Class,Real Estate,Asset Allocation

Abstract

Household debts have become an increasingly influential factor in holding financial assets (Cocco, 2005; Yao and Zhang, 2005; Chetty and Szeidl, 2009; Becker and Shabani, 2010). This paper therefore investigates the impact households' debts have on their holding status of various financial instruments by examining the data of 10,000 households from the 2011 and 2012 Household Financial Survey conducted by Statistics Korea. In exploring the relations between household debts and household financial asset holdings, this study specifically addresses two issues. One is about the extent to which household debts across household income classes have an impact on the shareholdings of each financial instrument. The other is about the sway of household debts on the choices of three financial instruments at the moment households make asset allocation decisions. Some evidences are found on the relation of households'debt and financial asset holdings. First, while financial asset holdings as well as household debts increase in proportion to a certain age up to 60, they drop sharply afterwards. The ratio of financial asset over total asset exhibits a U-shaped distribution. This result can be explained by the fact that Korean households in general have strong tendency to demand real estate relative to financial assets, and this preference is more pronounced as age increases. In addition, both financial asset holdings and household debts increase proportionally to the education level of households, while the ratio of financial asset or financial debts over total asset shows a moderately inverted U-shaped pattern. These results seem to imply that the better educated and the older households, the more capable they become in employing debt instruments either in financial or real assets as tools to accumulate wealth. To explore the relation of household debt and financial asset holdings, dependent variables such as fin, risky assets, and cash are normalized by dividing financial wealth, risky assets, and cash by total assets. With the scaling of dependent variable being equal to or greater than zero, Tobit regressions are conducted while, for the purposes of model setting, controlling for either one or combination of categories of different levels of education, industries, and occupations. Besides, to control for the endogeneity of household debts with financial asset holdings, I use the predicted household debt ratio as an instrument variable estimator obtained from running 2SLS. The second finding is that the impact of household debt ratio on financial wealth holdings is plainly pronounced among household income classes based on the analysis of the samples of 2010 in quintiles, although the household debt ratio in the whole sample negatively dominates the holdings of three financial instruments. In particular, the household debt ratio negatively affects the holdings of financial wealth in middle income class, whereas it positively does in high income class. This result indicates that the debt ratio of high income households plays a complementary role to financial wealth holdings, but, in contrary, the debt ratio of the middle income households substitutes financial asset holdings. Further, this finding implies that high income households take advantage of debt to accumulate wealth by investing in both financial assets (Pollin, 1988, 1990) and real estate, while middle income households do not hold both together, probably owing to credit constraint. However, high income households owning real estates are more inclined to reduce the holdings of both risky assets and cash holdings. Third finding is about asset allocation across household income classes. The financial debts in middle income households shrink the holdings of financial wealth, risky assets, and cash, whereas the disposable income in middle income class raises the holdings of the three major financial vehicles. But the financial debts in high income class increase the financial wealth, but reduce the cash holdings. Additionally, when the sample is restricted to households with debts, then financial debts reduce both financial wealth and risky assets, but raise the level of cash holdings. Taken together, these results generate two possible interpretations: the extent of the impact of financial debts over asset allocation varies depending on either household income classes or the levels of household debts. As a robustness check, I carry out the Tobit models using the 2011 sample. Unfortunately, the effect of household debts on financial wealth for the high income class only is consistent with the result in the prior year sample. This research is restricted to the two survey period samples and there are, therefore, limitations to fully take in the features of households behaviors on asset holdings and asset allocation in relation to household characteristics over time. I have a good guess about the asset holdings and choices by households in 2010 and 2011, which were unduly influenced by the 2008 Subprime Mortgage disaster. One of future research is possibly the comparison with changes in households asset holdings in terms of household debts and income before and after 2008 financial crisis.
LIST
Export citation