房产预期回报率视角下的中国家庭资产配置

Subjective Belief on Housing Return and Its Implications on the Asset Allocations of Chinese Households

房产在中国家庭的资产配置中起绝对的主导作用,而风险金融资产特别是股票的占比过低。现有文献大多刻画了中国家庭资产配置“重房产、轻风险金融资产”的特征事实,但这种配置结构是否有效,还没有被深入研究过。利用2019年中国家庭金融调查数据和资产历史表现数据,本文首次发现了中国家庭实际资产组合收益率的均值—方差与经典投资组合理论中有效资产组合收益率的均值—方差并不一致,提出家庭对房产收益分布的预期回报率持有错误的认知可能是导致此现象的一个主要原因。进一步将这一假设引入标准的均值—方差投资组合理论框架,构建了一个具有主观认知的资产配置结构模型,实证检验了这一假设并定量研究了其对中国家庭资产配置的影响。结构估计的结果表明:中国家庭普遍高估了房产的预期回报率,且平均高估2.8个百分点。更为重要的是,家庭对房产预期回报率的认知偏差在城市级别层面存在差异:一线城市家庭低估了房产预期回报率,而非一线城市家庭则高估房产预期回报率最高达4.9个百分点。反事实分析表明,高估房产预期回报率使中国家庭的财富平均减少了相当于5.2个百分点的家庭年收入,而在非一线城市则减少了最高达8.9个百分点。

It has been often noted that Chinese households'asset portfolios are excessively dominated by their real estate holdings at the expense of financial assets especially equities.By some estimates,more than 70%of the household wealth in China is tied up in real estate(Zhang et al.,2021).Accordingly,academics and policymakers have long been concerned about the cause and effect of such a heavy concentration of wealth in real estate.However,despite a great deal of research that has been conducted on the investment behavior of Chinese households,whether their investment portfolios represent an optimal allocation of wealth across different asset classes is less understood.After all,given that housing prices in major Chinese cities such as Beijing and Shanghai had appreciated at an annual growth rate of 15.4%and 12.0%respectively from 2009 to 2019(according to anjuke.com),while the annual rate of return of Chinese equity market was only 8.6%during the same period(according to China Stock Market&Accounting Research Database),it may be that what appears to be an overreliance on housing investment,and conversely,a limited participation in the equity market,are actually the results of an optimal allocation rather a misallocation of household wealth.Investigating this issue requires a framework under which optimality or suboptimality of wealth allocation can be evaluated.To that end,we employ Harry Markowitz's modern portfolio theory which is also known as the mean-variance analysis.It provides guidance for how rational,risk-averse investors should allocate their wealth across different assets to achieve an optimal trade-off between returns(mean)and risk(variance)for the whole asset portfolio,so as to obtain the highest expected return for the level of risk they are willing to accept.A classical application of the mean-variance analysis assumes that investors exhibit constant absolute risk aversion(CARA),which seems unlikely because it implies that the amount of wealth they invest in a particular risky asset remains the same no matter how wealthy they are.Therefore,we instead assume that investors have a constant relative risk aversion,which implies only that they invest the same portion of their wealth at all wealth levels.By applying the asset allocation theory to the micro data from the 2019 China Household Finance Survey,we find that Chinese households'portfolios are not consistent with the modern portfolio theory.They are overinvested in housing and underinvested in equities.Interestingly,however,we do find that their behaviors are not merely random or irrationally biased.Rather,they are consistent with investors who optimally trade off returns for risk but overestimate the returns on their own city's housing.We therefore introduce Chinese households'concept of housing into Markowitz's framework and estimate the cognitive bias and the relative risk aversion coefficient that rationalize the observed household portfolios.Our estimation results show that on average Chinese households overestimate the annual return on the housing in their city by 2.8 percentage points.More importantly,we find that this cognitive bias and the resulting loss of wealth vary substantially across city-tier levels.Households in the first-tier cities in fact underestimate their city's housing return,while those in lower-tier cities overestimate it by as much as 4.9 percentage points.Counterfactual analysis based on our estimated risk aversion coefficient implies that the abovementioned cognitive bias has reduced the potential wealth of an average Chinese household by the equivalent of 5.2 percentage points of its annual income and by as much as 8.9 percentage points for households in lower-tier cities.The empirical results we have obtained show that excessive investment in housing,which has been typically framed as a macroeconomic problem,has an actual cost to the individual households as well,particularly to those in lower-tier cities who can least afford it.There have been many policy suggestions and attempts at moderating the households overreliance on housing investment.Some of these suggestions,like improving the financial literacy of Chinese households,are difficult to achieve in practice,and some,like imposing property tax,may have a large unintended effect on the economy.Our analysis suggests that a relatively simple-and-easy-to-implement solution like providing households with a more accurate knowledge of their city's housing returns,for example by widely publicizing the transaction prices,may substantially improve the misallocation of wealth.A more optimal allocation may also help to strengthen the Chinese equity market by both deepening and broadening the household participation.Moreover,because an optimal allocation would have a greater effect on lower-tier cities,it may contribute to improving the wealth inequality as well.