Housing wealth and consumption among elderly Japanese

Abstract

This paper proposes an alternative channel for elderly consumption response to housing value changes by focusing on the informal financial markets. This channel may be important, particularly in a society where the elderly are less mobile and have limited access to the formal financial markets. In such a society, elderly homeowners may leave their housing assets to their children, and in exchange receive cash to fund consumption. Japanese household-level data on the elderly are used to examine this possibility. The results indeed suggest that the consumption of the elderly responds to changes in their house values through the intrafamily finance channel alone. The marginal propensity to consume out of housing wealth is only 0.01.

Keywords:

• Elderly
• house values
• bequest motives
• intergenerational transfers
• Japan

Acknowledgements

We would like to thank the editor, three anonymous referees, Kuang-Liang Chang, Tetsushi Homma, Takako Idee, Koji Karato, Takumi Motoyama, Oleksandr Movshuk, Takashi Unayama, seminar participants at GRIPS, Musashino University, and Nihon University, and conference attendees at AsRES and JEA for their valuable comments and constructive suggestions. We would like to acknowledge Beverley Searle who not only established the INTEGRATE network, but who also motivated us to write this paper. We would like to acknowledge Seko Miki for her support. We are grateful to the Panel Data Research Center at Keio University for access to its micro-data. We would also like to thank Michio Naoi and Kazuto Sumita for their useful advice in the analysis of these data. This research was part of a project of the Housing Research and Advancement Foundation of Japan.

Disclosure statement

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

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1 Following Campbell & Cocco (2007), we also estimate the OLS model including one-year house value changes ($N=6570$). However, there are no significant changes to the main findings in Table 2. We apply the fixed effects model because using lagged variables reduces the sample size.

2 Approximately 85.7 percent of the elderly in our sample ($N=7728$) are outright homeowners. Therefore, our empirical results are less likely to be influenced by the elderly with mortgage debt. Indeed, we find that the estimated coefficient of home equity (net measure) is slightly smaller but qualitatively similar to that of house value (gross measure).

3 During this period, Japan experienced two largely unexpected declines in house values because of the Great Recession and the Great East Japan Earthquake.

4 We incorporate the interaction terms between the region (seven regions and three city sizes) dummy and the year dummy into the benchmark model, which is shown in Table 2 in Section 3.3, to account for time series variation in region-specific macroeconomic shocks. This alternative specification demonstrates that the coefficient of interest is quantitatively similar to that in our benchmark result.

5 When the elderly and their children live together, consumption may include the consumption of the children. Similarly, income may be earned by children, and some housing and financial wealth may be owned by them. The JHPS, however, reported that only 3.3 percent of the elderly in our sample dwell in a house owned by other family members, including children.

6 In terms of household background variables, economic factors such as financial wealth changes and annual income changes have a significantly positive effect on consumption. The elderly consume a part of increases in liquidity. The coefficients of nonworkers are significant and negative. The elderly tend to decrease consumption because they appear to reduce work-related expenses after retirement (Wakabayashi, 2008). The demographic attributes of household size changes also matter, being significantly and positively associated with consumption. Namely, consumption expenditure decreases when household size declines.

7 Naoi (2014) and Hori & Niizeki (2019) demonstrated that elderly homeowners respond more to changes in house values than their younger counterparts. This supports the direct housing wealth channel. To understand whether the elderly in our sample behave differently from other age groups, we estimate the fixed effects model for household heads aged younger than 40 years ($N=3204$) and 40–59 years ($N=8841$), respectively. The results suggest that the coefficients of house value are not significant for either of these age groups.

8 If our benchmark model is subject to bias stemming from confounding factors, then our home value variable should also influence the consumption of elderly renters, as suggested by the literature (Lovenheim & Mumford, 2013). To examine the effects of house value changes on renters, we estimate the regional house price, which averages the fitted self-reported values of homeowners at the municipality level. The results suggest that regional house price changes have a statistically insignificant impact on the consumption decisions of elderly renters ($N=1543$).

9 Alternatively, some elderly parents possess an altruistic bequest motive; therefore, they leave housing assets without receiving financial help from their children. Horioka (2002), however, suggested that the selfish bequest motive is dominant in Japan.

Additional information

Funding

This work was supported by JSPS KAKENHI Grant Numbers JP16K03616 and JP17H00988.

Notes on contributors

Shinichiro Iwata

Shinichiro Iwata is a professor at University of Toyama. His current research focuses on housing, urban, and regional issues in Japan such as housing prices, housing transfers, housing wealth, urban sprawl, and compact cities.

Norifumi Yukutake

Norifumi Yukutake is an associate professor at Nihon University. His research interests include applied economics and econometrics with a specific focus on housing and land markets. His recent publication estimates the residential land damage of the Fukushima nuclear accident.