|
|
|
| Asset Portfolios and Household Consumption in Urban China: Theory and Evidence |
| JIANG Tao, DONG Bingbing, ZHANG Yuan
|
Southwestern University of Finance and Economics;
Central University of Finance and Economics;
Postal Savings Bank of China |
|
|
|
Abstract Many countries, including China, have adopted fiscal policies such as tax cuts and subsidies to boost consumption. However, how effective these policies are is a matter of some debate. Previous studies have traditionally focused on relatively poor families because they have higher marginal propensity to consume (MPC) and therefore fiscal policy oriented to such families should be more effective (Galí et al., 2007, Morita 2015). Carroll et al. (2014) further show that the more inequality, the higher aggregate MPC in response to temporary income shocks. These studies follow Campbell and Mankiw (1989, 1990) in describing the poor as those who consume all their income, namely, hand-to-mouth (HtM) households, in contrast to households that do not do so and typically have greater wealth (non-HtM). Wealthy or non-HtM households have low MPC. However, such studies have neglected the possibility that even wealthy households may also have high MPC. The idea is that some wealthy households may have a large amount of illiquid assets but few liquid assets, versus households that have both. These families do not usually liquidify their asset holdings to make consumption smoother, and they therefore tend to consume more in response to a temporary income shock (due to tax cuts or subsidies). Kaplan and Violante (2014a, 2014b) formalize this scenario and re-categorize this group of households as “wealthy HtM households.” Such households are wealthy in terms of total wealth but are HtM because they consume almost all their income.
Our first contribution is to re-examine the strategy for selecting types of households and quantifying the shares of different households. We focus on the consumption behavior of urban households and exclude those living in rural areas, as the former have more reliable data and regular income flows. We differ from previous studies of China in that we use four rounds of China Households Finance Survey (CHFS) data (from 2011, 2013, 2015, 2017) and perform a more thorough analysis by comparing different selection criteria. We consider the following concerns. First, Chinese households are more prone to savings and keep more liquid assets. Second, the liquidity of assets is slightly different compared other countries. In particular, we take time deposits, bank issued investment products, and treasury bonds as liquid assets because they have much lower returns and higher liquidity than real estate, the main component of illiquid assets in China. We apply different thresholds of liquid assets over yearly income ratio to determine which households are HtM, complemented by the consumption over income ratio, which is also used in Kaplan and Violante (2014). We find a significantly higher threshold of liquid assets over yearly income ratio, 1/4, than has been used in the literature (1/24). Considering real estate as the main component of illiquid assets, we also use the net value of housing (total market value minus debt and mortgage) to determine whether a household is wealthy.
With this selection strategy, we find the following evidence from Chinese urban households. First, the wealthy HtM share is about 36.7%, higher than the upper bounds estimated by He and Zang (2016) and Cui and Feng (2017). We then estimate the heterogeneity of consumption behavior over different types of households. The estimated consumption income elasticities for poor HtM and wealthy HtM households are 4.4% and 5.9% higher than wealthy non-HtM households, respectively. The MPCs to temporary income for poor HtM and wealthy HtM households are 0.069 and 0.09, compared to 0.045 for wealthy non-HtM households.
Another contribution of the paper is its recognition of the fact that most households in China must save a high down payment to buy real estate. These households have relatively more liquid assets and negligible illiquid assets. They have smaller MPC compared to wealthy non-HtM households, as shown in a previous study (Zang and Zhang, 2018). Their consumption income elasticity is also 2.7% lower than that of wealthy non-HtM households.
Our study has several policy implications. First, we show that the share of wealthy HtM households is larger than previously reported in urban China. These households have high MPC and should be targeted by policies intended to boost aggregate consumption. Second, there is a significant group of poor non-HtM households in China that should not be ignored.
|
|
Received: 05 November 2018
Published: 29 November 2019
|
|
|
|
|
|
|
|
