家庭住房需求和资产配置——基于包含人力资本和禀赋异质性的生命周期模型

Households’ Housing Demand and Asset Allocation:Based on a Life-cycle Model Including Human Capital and Endowment Heterogeneity

识别和区分家庭住房需求和资产配置受生命周期、代际和时间的不同影响,对于准确把握房地产和金融资产需求的长期趋势,推进相关市场的长效机制建设有重要意义。数据显示:房产在家庭资产配置中起主导作用,房产拥有率高、占比大,而其它金融资产比例较低;首套房购买者有明显的年轻化趋势,住房拥有率在40岁前后达到高点;房价对不同代际和户籍身份的家庭有不同影响,年轻人和新移民的购房支出高、房贷压力大。基于上述事实,本文构建了包含人力资本和禀赋异质性的分析住房需求决策的生命周期模型,研究发现:更好的人力资本和初始禀赋对住房拥有和房产配置均有正向影响,但是存在一定的替代作用;教育年限增加了收入,但是推迟了就业和婚育,导致首次购房年龄延迟;在房价快速上涨的背景下,禀赋提升会增加房产持有,挤出金融资产配置。

Housing is the most important asset owned by Chinese households. In addition to its investment value, it also has consumption attributes. Housing affects economic development and people’s lives. Clarifying the facts of housing is crucial to understand household asset allocation, macroeconomic policies, and social mobility. Moreover, identifying the life cycle characteristics of housing demand and asset allocation play an important role in understanding the relationship between demographic changes and housing demand, in designing policies to regulate the housing market, which reduces macro financial risks and improves economic efficiency and social welfare. This paper summarizes the latest stylized facts of households’ housing demand and asset allocation based on the five waves(from 2011 to 2019) of the China Household Finance Survey(CHFS) data, focusing on the special characteristics of Chinese households and how asset allocation is affected by heterogeneous human capitals and endowments.This paper studies the homeownership and asset allocation of Chinese households from the perspective of life cycle. The difficulty in this type of empirical studies is to distinguish the time effect, the cohort effect, and the life cycle effect. The data observed at any specific time point reflect the comprehensive impacts of the above three aspects. Without any further restrictions, it is impossible to distinguish the life cycle effects(only the part related to age and life stage) from the time effect and the cohort effect. For example, in terms of intergenerational differences, those born in 1961-1965 bought their first house at the median age of 37, while those born in 1981-1985 at the median age of 27, which was ten years earlier. The differences in property purchase in the cross-sectional data of these two groups are partially due to time changes and also mixed with intergenerational differences. In this paper, we adopt the method used in the life cycle income dynamic research(Deaton, 1997;Heckman et al., 1998;Lagakos et al., 2018). That is, we use the cross-section data(CHFS, 2011-2019) to construct an intergenerational panel, and analyze the financial characteristics of Chinese households to identify the life cycle characteristics of household asset allocation more precisely. We document the following stylized facts.(1) Housing plays a dominant role in household asset allocation, while the proportion of other financial assets is low and more likely to be the result of adjustments in property allocation. The most salient patterns show up among the young and the elderly. The young households may not be able to afford a house, while the elder may not need a house or want to refinance their real estate asset.(2) Homeownerships show significant life cycle characteristics. The growth trend of young people buying houses is obvious and the homeownership peaks during the ages of 30-40, which is affected by short-term fluctuations in the market.(3) The impacts of fast-growing housing prices on different generations vary greatly with migration status. Young people and new migrants have high spending on housing purchases and they have great pressure in paying mortgage.(4) Different human capital levels and endowments lead to large heterogeneity in housing allocations. Housing purchase rate, housing value and education change in the same direction. Compared with locals, the housing purchase rate and housing unit size of migrants in their place of residence are lower, while the mortgage participation rate and housing purchase expenditure are higher. The homeownership rate in cities and towns is higher for groups that have received family support when buying houses. In view of the fact that housing expenditure accounts for a large proportion of household assets and has a great impact on household consumption and investment, and given the short period of commercialization in real estate market, and its relationship with household registration status, housing demands have large heterogeneities in generations and people with different migration status and human capital. Therefore, we introduce heterogeneities in initial endowments and human capital into the standard life cycle model to capture the huge differences in housing consumption and household asset allocation among the same generation and groups with different educational levels. The life cycle model based on family heterogeneity can better describe and explain the dynamic changes of housing demand in China, which provides better theoretical support for the long-term real estate market forecasting and policy design. The empirical findings of this paper are consistent with the theoretical implications.(1) Housing purchase and property allocation show a strong life cycle characteristic.(2) An improvement of human capital and initial endowment has a positive impact on housing purchase and property allocation.(3) Compared with non-local households, the age of initial housing purchase of the young local household is 0.6 years earlier.(4) Although better education may increase the salary, it may also postpone the time point of marriage and childbirth, which leads to the delay effect. In the context of rapidly rising housing prices, the increase in households’ initial endowment will squeeze out their allocation of financial assets and increase the allocation of real estate, that is, the substitution effect. And to a certain extent, it will reduce the life cycle characteristics of financial assets.This paper makes contributions in four aspects. First, we comprehensively use the CHFS data, and distinguish the life-cycle effect, cohort effect, and time(short-term market fluctuations) effect of household asset allocation with the latest method in life cycle income dynamics research. Second, we find that households have distinct life cycle characteristics in housing purchase and asset allocation, and there are large heterogeneities in human capital and initial endowments. Third, based on the life cycle model and empirical facts, we introduce the heterogeneity of human capital and initial endowments into the housing demand decision-making on renting or buying a house. Fourth, the improvement of human capital and initial endowment has a positive impact on housing purchase and property allocation. We find that the average purchase age of the first house by young local households is earlier;increased educational levels will cause delays in initial housing purchases. The dynamic analysis of housing demand based on the above theoretical framework can help us better understand the impact of real estate-related policies and promote a better long-term mechanism design for the real estate market.