People’s Republic of China: Selected Issues

International Monetary Fund. Asia and Pacific Dept

doi:10.5089/9798400201486.002.A002

Author:

International Monetary Fund. Asia and Pacific Dept

International Monetary Fund. Asia and Pacific Dept
Search for other papers by International Monetary Fund. Asia and Pacific Dept in
Current site
Google Scholar

Publication Date:: 04 Feb 2022

eISBN:: 9798400201486

Language:: English

Keywords:: savings rate; household saving; debt effect; guarantee program; household debt; Income; Consumption; Household consumption; Housing prices

Download PDF (986.4 KB)

Despite fast the recovery from the crisis, growth remains unbalanced with private consumption lagging and the household savings rate—which was high in international comparison even before the crisis— now well above pre-pandemic levels. Empirical analyses based on prefecture- and household-level data suggest that higher government spending on social security and health is associated with a lower savings rate, while rising household debt, particularly housing-related debt, is associated with a higher savings rate.2 Rising income inequality at the expense of lower-income households, which tend to spend more of their income than higher-income households, also contributes to the higher aggregate savings rate. These findings suggest that targeted policy efforts to strengthen the social safety net, contain housing-related debt, and address income inequality would lower household savings rate.

Abstract

Household Savings and its Drivers: Some Stylized Facts^¹

Despite fast the recovery from the crisis, growth remains unbalanced with private consumption lagging and the household savings rate—which was high in international comparison even before the crisis— now well above pre-pandemic levels. Empirical analyses based on prefecture- and household-level data suggest that higher government spending on social security and health is associated with a lower savings rate, while rising household debt, particularly housing-related debt, is associated with a higher savings rate.^² Rising income inequality at the expense of lower-income households, which tend to spend more of their income than higher-income households, also contributes to the higher aggregate savings rate. These findings suggest that targeted policy efforts to strengthen the social safety net, contain housing-related debt, and address income inequality would lower household savings rate.

A. Introduction

1. Despite China’s rapid recovery from the COVID-19 pandemic, private consumption lags other GDP components. Amid the pandemic and the associated strong containment measures to contain the spread of the virus, China’s private consumption as a share of real GDP has declined significantly in 2020. Although the virus was largely under control in China in the second half of 2020, the recovery in private consumption still lagged the other GDP components. Moreover, the decline in China’s private consumption share in 2020 (in percentage points of GDP) was larger than in most other major economies, with the share remaining well below the emerging market average (Figure 1, panel 1). The household savings rate was already well above major emerging market economies before the pandemic and has increased further since (Figure 1, panel 2). At the same time, households’ marginal propensity to consume in China estimated from prefecture-l evel data has also dropped to multi-year lows, particularly for rural households (Figure 1, panel 3).

^{Figure 1.}

China: Household Consumption and Savings During the Pandemic

Citation: IMF Staff Country Reports 2022, 022; 10.5089/9798400201486.002.A002

2. China’s household savings rate varies significantly across income levels and between urban and rural households.

Across income groups. Prefecture-level data suggest that the household savings rates increase with income levels (Figure 1, panel 4), which is consistent with the micro-level evidence from the Chinese Household Income Project data (Zhang and others, 2018) and reflects the different propensities to consume out of income. While savings rates increased substantially in 2020 for all income groups, high-income households saw a more prominent increase.
Between urban and rural households. The savings rate of urban households is significantly higher than rural households (Figure 1, panel 5), which is in line with the distribution of savings rate across income levels as urban households typically earn higher incomes. However, the savings rate of rural households had been declining since the global financial crisis (GFC), while the urban households’ saving rate has been rising rapidly, likely reflecting the widening urban-rural income gap in monetary terms: the difference between urban and rural households’ disposable income per capita nearly tripled from 11 thousand RMB in 2009 to 27 thousand RMB in 2020, even though the ratio between the two has declined.

3. Income inequality has likely increased also within household groups, especially within urban households, which could also contribute to a higher aggregate household savings rate. The decline in China’s overall income inequality measured by the Gini coefficient before the pandemic has likely been driven by lower within-rural household inequality, as the income gap between urban households’ top and lowest income groups widened significantly in the past two decades (Figure 1, panel 6). The widened high/low-income ratio within urban households and the widened urban-rural income gap in monetary terms (discussed above) point to a likely increase in income inequality in 2020.^³ Since higher-income groups tend to save more than lower-income groups, an increase in income inequality (e.g., resulting from a pure redistribution of income) could potentially lead to a higher aggregate savings rate due to the composition effect (Zhang and others, 2018).

4. These changes by the pandemic could threaten the transition towards a balanced and sustainable recovery. Continued high savings would risk reversing the rebalancing efforts towards private consumption-driven growth and create upward pressure on the current account. Meanwhile, there is growing evidence that rising income inequality is harmful for the pace and sustainability of growth (Jain-Chandra and others, 2018; Ostry and others, 2014; Cingano, 2014; Berg and others, 2012; Berg and Ostry, 2011), mainly through reducing the equality of opportunity and curbing investments in education by the poor and lower middle classes. Reducing the high household savings rate and income inequality is important to achieve balanced and inclusive growth.

5. This paper investigates some of the key drivers of China’s high household savings rate, and explores policy options to achieve balanced and inclusive growth. In particular, we use empirical analyses based on both prefecture-level macro data and household-level micro data, with a focus on the drivers that are potentially more relevant for the recent increase in the savings rate during the pandemic. In addition, we also combine prefecture-level with household-level data to examine income inequality and its key sources. Finally, we provide policy recommendations to reduce China’s household savings rate and income inequality.

Household Income and Consumption

(Year-on-year, percent change)

Citation: IMF Staff Country Reports 2022, 022; 10.5089/9798400201486.002.A002

Sources: CEIC; and IMF staff calculations.

B. Potential Drivers of Household Savings During the Pandemic

6. Household consumption and income growth have been highly correlated until the COVID-19 pandemic. The annual growth rate differential between household consumption expenditure per capita and disposable income per capita had been below 3 ppt since early 1990s before the pandemic, but has widened to -6.3 percent in 2020 (text chart). This suggests little consumption smoothing by households, possibly related to the unique nature of the pandemic shock where services-related consumption was limited by stringent restrictions to contain the spread of the virus. However, the recovery in private consumption continued to lag other GDP components even when the containment measures were gradually lifted in 2020H2, and household savings rate still remained elevated at 35 percent in 2021Q2—well above its pre-pandemic level of 30 percent.

7. There is a large body of literature on the potential reasons behind China’s high household savings rate. One strand of the literature focuses on demographic factors (Curtis and others, 2015; Choukhmane and others, 2014; Modigliani and Cao, 2004), arguing that the changing demographic structure resulting from the one-child policy has led to high household savings, through lower spending on children and less expected inter-generational support given the fewer children. Another strand focuses on the role of precautionary savings (He and others, 2017, 2018; Chamon and Prasad, 2010; Blanchard and Giavazzi, 2005), arguing that the lagged transformation of the social protection system and falling job security during China’s transition towards a market-based economy is the 1990s led to a rise in household precautionary savings. Others have pointed to low returns on household deposits as another potential driver of high savings, although empirical evidence has been scant. Some studies also analyzed the impact of house prices and housing ownership on savings but generally found mixed evidence (Chen and others, 2016; Wang and Wen, 2011), while others examining the impact of house prices on income inequality pointed to a significant positive impact for urban households (e.g., Zhang and Zhang, 2015). Zhang and others (2018) provide a comprehensive study of all these factors and found that the demographic changes and the lagged transformation of the social protection system and job security contributed the most to China’s high household savings rate, with rising house prices and income inequality also playing a role.

8. The rest of this section uses prefecture-level macro data and household-level micro data to take a fresh look at these potential drivers, paying particular attention to the short-term factors—that may have been amplified by the pandemic—including, the weaknesses in the coverage and adequacy of the social protection system, rising household debt, the role of household assets particularly housing ownership, and income inequality. As a contribution to the literature, the household-level micro data allow us to estimate the debt and wealth effects on household savings rate separately.

Social Protection System

9. Despite the significant policy efforts to strengthen the social protection system after the GFC, government spending in social security and health as a share of GDP had remained broadly unchanged until the pandemic. In 2009, the government strengthened the healthcare scheme for rural households by introducing the pay-as-you-go pension system heavily subsidized by direct government transfers and, for urban areas, established a mandatory insurance scheme for formal sector workers—mostly funded via social security contributions from employers and employees. An insurance scheme and a basic pension scheme for non-working residents were also established in 2009 and 2010, respectively, with significant government subsidies. However, the social protection system still remains incomplete and lacks adequate coverage and benefits (SIP 1). In particular, the pension benefits remain low for rural households and access to healthcare remains an issue for migrant workers (Jain-Chandra and others, 2018; Zhang and others, 2018). Moreover, government spending in social security and health as a share of GDP had remained stable or even declined somewhat in the case of health during 2016–19 (text chart) and remains low compared to international standards (Zhang and others, 2018), despite a minor rebound in 2020 due to the pandemic. Here, the social security spending is a component of the total social protection spending, including the central and local governments’ subsidies for the social insurance system (mostly social pension insurance) and social assistance; and the health spending includes the central and local governments’ spending on medical services (such as public hospitals) and subsidies for medical insurance.

Goverment Spending

(In percent of GDP)

Citation: IMF Staff Country Reports 2022, 022; 10.5089/9798400201486.002.A002

Sources: CEIC; and IMF staff calculations.

10. Our analysis finds that higher spending on social security and health is significantly positively associated with household consumption, with differing impacts on urban and rural households, suggesting a targeted policy focus. We study the impact of precautionary motives by observing regional variations in government spending on social security, health, and education. Using cross-sectional analysis based on all 296 prefectural municipalities monitored by the National Bureau of Statistics (NBS), we find a significant positive relationship between social spending and urban and rural consumption (see Appendix for data and detailed empirical analyses).^⁴ Our findings broadly confirm previous studies such as Zhang and others (2018).

Social security spending has a significant impact on urban household consumption…. Social security spending has a stronger association with household consumption expenditure than other types of social spending, with a 1-percent increase in social security spending per capita associated with a 0.05—0.08 percent increase in household consumption expenditure per capita and hence lower savings rates (Figure 2, panel 1). This suggests that, using the magnitude of the impact at each variable’s mean, a 100-RMB increase in annual social security spending per capita (representing about 6 percent of the average social security spending per capita in 2019) is associated with an increase in annual urban consumption per capita of about 90 RMB.
…but the health spending has a significant impact for rural households. While social security spending seems to have had a strong association with household consumption during 2015–17, health spending mattered significantly during 2018–19, with a 1-percent increase in the health spending per capita associated with a rise in rural consumption per capita by 0.1—0.2 percent (Figure 2, panel 2). Assessing the impact at the mean suggests a 100-RMB increase in the annual health spending per capita (about 9 percent of the average health spending per capita in 2019) is associated with an increase in annual rural consumption per capita of about 185 RMB on average.

^{Figure 2.}

China: Estimated Impact of Social Spending on Household Consumption

Citation: IMF Staff Country Reports 2022, 022; 10.5089/9798400201486.002.A002

11. The differences in urban and rural results could be related to the relatively weaker healthcare support in rural areas. Healthcare services and the social health insurance system are relatively more developed in urban areas, and hence less of an issue for urban household consumption. As a consequence, rural households might feel the need to accumulate relatively higher levels of precautionary savings—an effect that was likely heavily amplified by the pandemic.

Household Indebtedness

12. Household debt could constrain consumption and increase the savings rate. The literature suggests that the short-term and medium-term effects of higher household debt on growth differ (e.g., IMF, 2017; Mian and others, 2017; Lombardi and others, 2017; Jordà and others, 2016). For example, using cross-country panel data, Lombardi and others (2017) find that household debt increases consumption and GDP growth within one year but reduces them in the long run, with the negative long-run effects intensified when household debt-to-GDP ratio exceeds certain thresholds. The negative impact could come from the high marginal propensity to consume of heavily indebted households who cut spending rapidly following negative macro-financial shocks (Mian and others, 2013). In China’s case, however, this effect might be mitigated by the fact that more than half of all household debt is housing-related rather than driven by consumption (PBC, 2019).

13. China’s high and rising household debt may have started to constrain consumption growth. Recent empirical studies for China (e.g., Han and others, 2019; Tian and others, 2018) find that the rapid growth in household debt, particularly since 2015, may have already started to have a negative net impact on consumption growth. Moreover, higher household indebtedness could also reduce the income elasticity of consumption, particularly when facing a negative income shock, likely due to the higher debt service burden (Han and others, 2019).

14. Our empirical analysis confirms that higher household indebtedness is associated with higher savings rates for urban households. We use the longitudinal household survey data from the China Family Panel Studies (CFPS) to analyze the relationship between household indebtedness and urban and rural savings rates.^⁵ Using fixed-effects panel data regressions, we find that the debt-to-income (DTI) ratio linked to mortgages and non-housing debt, has positive associations with the savings rate, for both urban and rural households (Figure 3, panel 1; see Appendix for data and more detailed empirical analyses). These associations are statistically significant except that between non-housing debt and the rural savings rate, likely because rural households still have limited access to consumption credit and hence less non-housing debt service burdens.

^{Figure 3.}

China: Estimated Impact of Household Debt and Assets on Savings Rate

Citation: IMF Staff Country Reports 2022, 022; 10.5089/9798400201486.002.A002

Housing Ownership

15. Housing ownership could affect household savings behavior through various channels. These include the down payment effect, mortgage effect, and wealth effect (Zhang and others, 2018). The down payment channel implies that a tenant would save more if he or she decides to buy a house, and rising housing prices would make that incentive even stronger. The mortgage channel suggests that homeowners would need to save more to pay mortgages, and has been analyzed above in the estimated debt impact. Both the down payment and mortgage effects imply higher savings when house prices rise. The wealth effect implies the opposite: homeowners would increase consumption and reduce savings as they would feel wealthier with rising house prices. The overall impact of housing ownership on savings depends on the relative strength of these offsetting channels. Similarly, non-housing assets, such as equities or other risky financial assets, could also have both a debt effect (discussed above) if funded by borrowing and a wealth effect on household consumption and savings rate.

16. Our analysis suggests that housing ownership has a significant positive association with urban households’ savings rate. Similar to the breakdown of household debt into mortgages and non-housing debt, we include both net housing assets (by subtracting mortgages from gross housing assets) and the non-housing savings stock (as a proxy for non-housing assets) in the panel regressions. Estimation results suggest (Figure 3, panel 2):

Net housing assets have a significant negative association with the savings rate for both urban and rural households, suggesting a significant wealth effect through housing ownership. However, the magnitude of the wealth effect is much smaller than that of the debt effect for both urban and rural households, suggesting a more dominant mortgage effect on the savings rate than the wealth effect.
Non-housing assets have a significant wealth effect for rural households only. Combined with the significant impact of non-housing debt on the savings rate, this suggests that the non-housing wealth effect (which reduces the savings rate) dominates the non-housing debt effect (which increases the savings rate) for urban households, likely reflecting their easier access to credit and higher non-housing debt service burdens. While for rural households who have less access to non-housing credit, the non-housing wealth effect is more prominent.

Income Inequality

17. Income inequality translates into savings inequality. The prefecture-level data suggest that the household savings rate tends to be increasing with income levels as shown in Figure 1, reflecting the different propensities to consume out of income. Chinese household savings rates are higher than most countries at every income decile, but the difference is particularly large for the poor: the savings rate for the bottom 15 percentiles of Chinese urban households was about 11 percent in 2012 according to estimates based on data from the Chinese National Bureau of Statistics, while the savings rates for the bottom 10–20 percentiles are negative in many other countries due to the substantial social transfers to support the basic consumption (Zhang and others, 2018). Similarly, Jin and others (2011) find that income inequality significantly reduces households’—particularly low-income households’— consumption expenditures on non-education items, likely due to the savings motives to improve their social status.

18. Income inequality has multiple dimensions. In addition to the inequalities within urban households and between urban and rural households, we also use the CFPS survey data to break down the total income into different sources of income—including, wage income, housing-related income, business-related income, and other income. Higher inequality in non-wage income sources, particularly housing-related income, contributed to the overall income inequality (Figure 4, panel 1). In terms of wages, there is significant variation in the within-province wage inequalities, as shown by the wide distribution of the Lorenz curve across provinces relative to the national average (Figure 4, panel 2).

^{Figure 4.}

China: Income Inequality by Source and Region

Citation: IMF Staff Country Reports 2022, 022; 10.5089/9798400201486.002.A002

Sources: CFPS Household Survey data; and IMF staff calculations.

19. Income inequality appears positively associated with house price misalignment. The national Gini coefficient, only available until 2019 before the pandemic, shows a positive relationship with house price misalignment (text chart). This is in line with the finding that there is a significant positive association between housing prices and the Gini coefficient of the income of urban households as higher-income households benefit more from house price growth through housing-related income (e.g., Zhang and Zhang, 2015).

Gini Coefficient vs. House Price/Income Ratio

Citation: IMF Staff Country Reports 2022, 022; 10.5089/9798400201486.002.A002

Sources: CEIC; and IMF staff calculations.

C. Policy Implications

20. The empirical relationships discussed above suggest that measures to strengthen the social protection system and reducing income inequality would help reduce China’s elevated household savings rate. Specifically:

Increasing government spending on social security and health. Higher social security and health spending is associated with higher household consumption and lower precautionary savings, especially if targeted on health spending in rural areas and on social security spending in urban areas. Unemployment and social assistance reform could boost household economic security through, for example, better coordination between central and local governments, promotion of transfers of insurance and benefits between urban and rural areas, and a more generous minimum income guarantee program and non-contributory basic pension for rural and non-salaried urban residents (SIP 1). In addition, continued Hukou reform can help ensure that migrant workers have the same access to the social safety net as urban workers (Lam and Wingender, 2015). While spending on education does not appear to have a direct impact on household savings, it would help reduce income inequality in the future by providing equal education access to the poor (Zhang and others, 2018).
Further improving the macroprudential policy framework and toolkit to contain household debt and reduce house price misalignment. The larger mortgage effect of housing ownership on the savings rate than the wealth effect suggests that macroprudential policies targeted at reducing housing-related debt can help lower the savings rate in addition to the house price misalignment. International experience suggests that demand-side macroprudential measures such as limits on debt-service-to-income ratios and loan-to-value ratios are also effective in mitigating the negative effects of household debt, including mortgages, on consumption and GDP growth, and could enhance policy effectiveness (Han and others, 2019). In particular, the limits on debt-service-to-income ratio could be lowered and extended to fully cover household loans from non-bank financial institutions. Reducing the house price misalignment could help improve income inequality given that higher-income households typically benefit more from house price growth.
Increasing social transfers to poor households and further improving the efficiency of social transfers through improved risk-sharing mechanism. Further increasing social assistance spending in China, would also help reduce income inequality and low-income households’ savings rates, contributing to the reduction of the national savings rate. In addition, further increasing the efficiency of the central government transfer system by introducing an automatic and non-regressive fiscal risk sharing mechanism could help achieve higher fiscal risk-sharing and similar redistribution effects at lower costs, helping free up extra resources for social transfers to the poor (IMF, 2021).

References

Baker, S., 2014, “Debt and the Consumption Response to Household Income Shocks,” Working Paper, Stanford University.
- Search Google Scholar
- Export Citation
Berg, A., and J. D. Ostry, 2011, “Inequality and Unsustainable Growth: Two Sides of the Same Coin?” IMF Staff Discussion Note 11/08 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
Berg, A., and Ostry, J. D., and J. Zettelmeyer, 2012, “What Makes Growth Sustained?” Journal of Development Economics, 98(2), pp. 149–66.
- Search Google Scholar
- Export Citation
Blanchard, O., and F. Giavazzi, 2005, “Rebalancing Growth in China: A Three-Handed Approach,” CEPR Discussion Paper.
- Search Google Scholar
- Export Citation
Chamon, M. D., and E. Prasad, 2010, “Why are Saving Rates of Urban Households in China Rising?” American Economic Journal: Macroeconomics, Vol. 2(1), pp. 93–130.
- Search Google Scholar
- Export Citation
Chen, B., Yang, X., and N. Zhong, 2016, “Urban Housing Privatization and Household Saving in China,” Working Paper.
- Search Google Scholar
- Export Citation
Choukhmane, T., Coeurdacier, N., and K. Jin, 2014, “The One-Child Policy and Household Savings,” Working Paper.
- Search Google Scholar
- Export Citation
Cingano, F., 2014, “Trends in Income Inequality and its Impact on Economic Growth,” OECD Social, Employment and Migration Working Papers, No. 163, OECD Publishing.
- Search Google Scholar
- Export Citation
Curtis, C. C., Lugauer, S., and N. C. Mark, 2015, “Demographic Patterns and Household Saving in China,” American Economic Journal: Macroeconomics, Vol. 7(2), pp. 58–94.
- Search Google Scholar
- Export Citation
Han, F., Jurzyk, E., Guo, W., He, Y., and N. Rendak, 2019, “Assessing Macro-Financial Risks of Household Debt in China,” IMF Working Paper 19/258 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
He, H., Huang, F., Liu, Z., and D. Zhu, 2018, “Breaking the ’Iron Rice Bowl:’ Evidence of Precautionary Savings from the Chinese State-Owned Enterprises Reform,” Journal of Monetary Economics, Vol. 94, pp. 94–113.
- Search Google Scholar
- Export Citation
He, H., Ning, L., and D. Zhu, 2019, “The Impact of Rapid Ageing and Pension Reform on Savings and the Labor Supply: The Case of China,” IMF Working Paper 19/61 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
International Monetary Fund (IMF), 2021, “People’s Republic of China: Selected Issues,” IMF Country Report No. 21/12 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
International Monetary Fund (IMF), 2020, “People’s Republic of China: Staff Report for the 2020 Article IV Consultation,” IMF Country Report No. 2021/006 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
IMF, 2017, “Global Financial Stability Report October 2017: Is Growth at Risk?” (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
Jain-Chandra, S., Khor, N., Mano, R., Schauer, J., Wingender, P., and J. Zhuang, 2018, “Inequality in China – Trends, Drivers and Policy Remedies,” IMF Working Paper 18/127 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
Jin, Y., Li, H., and B., Wu, 2011, “Income Inequality, Status Seeking, and Consumption (in Chinese)”, China Economic Quarterly, Vol. 10, No. 3, pp. 887–912.
- Search Google Scholar
- Export Citation
Lombardi, M., Mohanty, M., and I. Shim, 2017, “The Real Effects of Household Debt in the Short and Long Run,” BIS Working Paper No. 607, Bank for International Settlements, Basel.
- Search Google Scholar
- Export Citation
Mian, A., Rao, K., and A. Sufi, 2013, “Household Balance Sheets, Consumption and the Economic Slump”, The Quarterly Journal of Economics, Vol. 128, Iss. 4, pp. 1687–1726.
- Search Google Scholar
- Export Citation
Mian, A., Sufi, A., and E. Verner, 2017, “Household Debt and Business Cycles Worldwide,” The Quarterly Journal of Economics, Vol. 132, Iss. 4, pp. 1755–1817.
- Search Google Scholar
- Export Citation
Modigliani, F., and S. L. Cao, 2004, “The Chinese Saving Puzzle and the Life-Cycle Hypothesis,” Journal of Economic Literature, Vol. XLII, pp. 145–170.
- Search Google Scholar
- Export Citation
Nakajima, J., 2018, “The Role of Household Debt Heterogeneity on Consumption: Evidence from Japanese Household Data,” BIS Working Paper No. 736, Bank for International Settlements, Basel.
- Search Google Scholar
- Export Citation
Ostry, J. D., Berg, A., and C. Tsangarides, 2014, “Redistribution, Inequality, and Growth,” IMF Staff Discussion Note 14/02 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
People’s Bank of China (PBC), 2019, China Financial Stability Report.
- Search Google Scholar
- Export Citation
Tian, G., Huang, X., Ning, L., and Y. Wang, 2018, “Safeguarding Household Debt Crisis and Associated Systemic Financial Risks (in Chinese),” Policy Research Report No. 58, Shanghai University of Finance and Economics.
- Search Google Scholar
- Export Citation
Wang, X., and Y. Wen, 2011, “Can Rising Housing Prices Explain China’s High Household Saving Rate?” Federal Reserve Bank of St. Louis Review, March/April 2011, 93(2), pp. 67–87.
- Search Google Scholar
- Export Citation
Zhang, L., Brooks, R., Ding, D., Ding, H., He, H., Lu, J., and R. Mano, 2018, “China’s High Savings: Drivers, Prospects, and Policies,” IMF Working Paper 18/277 (Washington: International Monetary Fund).
- Search Google Scholar
- Export Citation
Zhang, C., and F. Zhang, 2015, “The Effects of Housing Prices on Income Inequality in Urban China,” Critical Housing Analysis, Vol. 2, Iss. 2, pp. 11–18.
- Search Google Scholar
- Export Citation

Appendix I. Data and Empirical Analyses

Data

The prefecture-level data are obtained from the revamped household survey for 296 prefecture-level municipalities during 2015–19 published by the NBS. Key variables include urban and rural disposable income, house price, government spending in social security and employment, government spending in health, and government spending in education. Missing data are common so the total number of observations for each year is usually below the total number of prefectures.

The household-level data are used to study the role of households’ indebtedness and wealth in determining the household savings rate. In particular, we use fixed effects panel regressions with the household-level survey data from four waves of the China Family Panel Studies (every two years from 2012 to 2018 CFPS) conducted by Peking University. The CFPS data are collected through questionnaires. The survey contains detailed information on household income structure, expenditures breakdown, indebtedness, and assets.

Empirical Strategy

Prefecture-Level Analysis

Following Zhang and others (2018), we use the prefecture-level annual data to estimate the impact of government spending on social security (including employment), health, and education on household consumption expenditure. In particular, we use cross-section regressions and estimate the following consumption equation (based on the life cycle hypothesis approach) year by year for 2015–19 and for urban and rural households separately:

\log (c o n s u m p t i o n_{i}) = β_{0} + β_{1} \log (d i s p o s a b l e i n c o m e_{i}) + β_{2} \log (s p e n d i n g_{i}^{s s}) + β_{3} \log (s p e n d i n g_{i}^{h e a l t h}) + β_{4} \log (s p e n d i n g_{i}^{e d u c a t i o n}) + β_{5} H P I_{i} + ε_{i}

where i denotes prefecture; $s p e n d i n g_{i}^{s s}, s p e n d i n g_{i}^{h e a l t h}, s p e n d i n g_{i}^{e d u c a t i o n}$ represent the government spending in social security, health, and education in prefecture i respectively. HPI_i is the house price-to-disposable income ratio, which is a proxy for household wealth following Zhang and others (2018).

Household-Level Analysis

For this exercise, households with zero or negative reported income or expenditures are dropped from the sample. For the estimated households’ savings rate, only samples with values between -100 percent and 100 percent are kept. The household savings rate is defined as the residual of disposable income less consumption as a share of disposable income.

The model specification is as follows:

s a v i n g s ​ r a t e_{j, t} = β_{j} + α_{0} \log (d i s p o s a b l e ​ i n c o m e_{j, t}) + α_{1} D T I_{j, t - 1}^{m o r t g a g e} + α_{2} D T I_{j, t - 1}^{n o n - h o u \sin g} + α_{3} A T I_{j, t}^{h o u \sin g} + α_{4} A T I_{j, t}^{s a v i n g} + ε_{j, t}

In the panel regression, savings rate for household (j) is regressed against the disposable income, lagged mortgage to income ratio $(D T I_{j, t - 1}^{m o r t g a g e})$ , lagged non-housing debt to income ratio $(D T I_{j, t - 1}^{n o n - h o u \sin g})$ , the net housing asset’to income ratio $(A T I_{j, t}^{h o u \sin g})$ , and the saving to income ratio $(A T I_{j, t}^{s a v i n g})$ , for urban and rural households respectively.