## 1. Consumption in China: Low and Falling

## Source details

**Canonical URL:** [1. Consumption in China: Low and Falling](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1016.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1016.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1016.pdf.json)

---

### I. Introduction — problem and research focus
- Consumption in China is unusually low and has continued to decline as a share of GDP over the past decade; rebalancing away from reliance on exports and investment toward consumption is urgent amid the collapse in global demand.
- Low household consumption (high saving) is often linked to precautionary motives due to under-developed government health, education, and pensions systems; households self-insure against uncertainty.
- Paper’s narrow focus: empirical relationship between government health and education spending and household saving.
- Main empirical conclusion:
  - Government spending on health, but not on education, reduces urban household saving.
  - Estimated magnitude: a 1 yuan increase in government health spending associated with a 2 yuan increase in household consumption.
  - Implication: Total (household plus government) consumption could increase by as much as 3 yuan depending on whether government health spending takes the form of consumption instead of transfers.
- For rural households: evidence is mixed; increased government health spending appears to affect savings only in higher-income provinces.
- Paper does not attempt to explain all drivers of consumption decline; household income is an important driver (Aziz and Cui, 2007).

### II. Consumption and saving in China: stylized facts
- Consumption trends:
  - Consumption as a share of GDP has steadily declined and is quite low relative to other countries.
  - Household consumption share has been falling over time; mechanically offset by rising investment and an expansion in net exports.
- Saving trends:
  - Saving rates have risen and help explain the decline in the consumption ratio.
  - Urban saving rate has steadily risen; rural saving is more volatile but has also risen.
  - Urban households account for the lion’s share of economic activity and drive aggregate saving increases.
- Income and consumption per capita:
  - Urban and rural household incomes have fallen steeply in per capita terms relative to GDP per capita.
  - From the early 1990s to 2008, urban household per capita income dropped from nearly 90 percent of GDP per capita to just under 70 percent.
- Attribution of consumption decline:
  - For urban households, roughly 60 percent of the decline in consumption since the early 1990s is attributable to a drop in disposable income as a share of GDP.
  - For rural households, falling income explains around three-fourths of the decline since the early 1990s.
- Additional stylized facts and magnitudes:
  - Real household consumption (CPI deflated) grew at an average annual rate of 8¾ percent over the past 10 years; GDP growth was 9¾ percent.
  - Urban consumption accounts for three-fourths of household consumption; urbanization increased urban share of consumption.
  - Urban per capita income (consumption) is about 3 times rural per capita income (consumption).
  - Saving is positively related to income; higher income groups have gradually increasing saving rates, with acceleration around 2003.
  - No clear sign of consumption smoothing: volatility of consumption growth is greater than income growth for both urban and rural households.

### III. Reducing precautionary saving: econometric analysis and results
- Data and approach:
  - Provincial pooled data used to exploit variation in provincial health and education spending and differences in saving rates.
  - Data period: 1994–2007; robustness checked with 2003–07 sample.
  - Urban and rural households analyzed separately.
  - Government spending data do not distinguish urban vs. rural allocation; interpretation of urban/rural differences requires caution.
- A. Urban households — key econometric findings:
  - Statistically significant relationship between government health spending and urban saving.
  - Parameter estimate around -2 across samples and specifications:
    - Each 1 yuan of government health spending → 2 yuan decrease in saving (equivalently, 2 yuan increase in consumption).
    - Implies a 1 percent of GDP increase in government health spending would boost private consumption by 2 percent of GDP and yield a total demand effect of 3 percent of GDP for every 1 percent of GDP increase in health spending.
  - Interpretation: negative sign suggests government health spending reduces precautionary saving rather than merely substituting for private health spending.
  - Empirical evidence on private health care spending:
    - Urban health care spending (household survey) fell by more than ½ percent of urban household consumption over the past four years; government spending on health care more than doubled to the equivalent of 3 percent of total household consumption.
    - National accounts data show higher urban health care spending historically, but decline in recent years.
    - Rural healthcare spending increased over the past four to five years by ½–¾ percent of rural household consumption (survey and national accounts).
  - Government education spending:
    - Little evidence that higher government education spending decreases urban saving.
    - Coefficient for education never statistically significant; sign changes between full sample and latter period.
    - Education spending per capita is more than five times larger than health spending and grew rapidly over the past four years.
- B. Rural households — key econometric findings:
  - Little evidence that higher health spending reduces precautionary saving for rural households.
  - Health coefficient always positive and never statistically significant in main samples, suggesting higher government health spending may increase saving or has little effect.
  - Possible explanations:
    - Rural households may be subsistence consumers and save a substantial part of any freed resources.
    - Rural areas may receive a smaller share of provincial health spending.
    - Increase in public spending small relative to expected rural health needs.
  - Education spending for rural households:
    - Little evidence of impact; coefficients positive and not statistically significant, though near significance in 2003–07 sample, consistent with near one-for-one substitution.
- C. Robustness checks:
  - Interaction with highest income provinces (Beijing, Tianjin, Shangdong, Jiangsu, Fujian, Guangdong, Zhejiang, and Shanghai) tested.
  - Rural results: government health spending in high income provinces is associated with a decline in rural household savings; magnitude similar to -2 found in urban regressions.
  - Urban results: no strong evidence that the impact of health spending differs in higher income provinces; interaction term usually not statistically significant though always negative.
  - Additional controls (real income growth, inflation, unemployment) included in robustness checks; main results on urban health spending remain substantively unchanged.

### IV. Conclusion — implications and policy relevance
- Main robust finding: higher government health spending reduces urban household saving; each additional yuan in government health spending boosts urban consumption by 2 yuan.
- Rural households: except in higher income provinces, no evidence that government health spending reduces saving.
- No evidence that higher government education spending affects urban or rural saving.
- Mechanism explanation:
  - Government spending can have competing effects: substitution (increasing saving by freeing household resources) versus reducing precautionary motives (decreasing saving).
  - Precautionary motives likely stronger for health than education; older generations drive much of the savings and are more concerned about healthcare costs.
- Policy relevance:
  - Broadening public health care coverage could have important effects on household precautionary savings and thereby on consumption.
  - Reorientation of government spending toward poorer rural areas and increased efficiency in public health care delivery will be key to improving health outcomes and impacting consumption behavior.
  - The government’s 2009–11 health reform strategy and priorities in the 11th five-year plan could improve rural health coverage and raise rural consumption, though challenges and inequalities in government health spending remain.
  - World Bank (2009) study: progress toward universal rural coverage is noted, but inequalities exist, with government health expenditure disproportionately benefiting the better off; over half of general government health spending supports urban insurance schemes whose members are disproportionately higher income.

### V. Key empirical tables and estimates (selected)
- Urban households (Table 1):
  - Health coefficient estimate ≈ -2.10 (1994–2007 sample, Model (1)); (Standard error) (0.72); [P-val] [0.00].
  - Health coefficient estimate ≈ -1.92 (various specifications); (Standard error) (0.86); [P-val] [0.03].
  - R-squared values reported in Table 1 range from 0.18 to 0.25 across specifications.
  - # Obs. listed: 285, 304, 285, 285, 150, 150, 150, 150 (depending on specification).
- Rural households (Table 2):
  - Health coefficient estimates positive (e.g., 0.51, 0.22, 0.37, 0.06) and not statistically significant; (Standard errors) (0.59), (0.58), (0.67), (0.64); [P-val] [0.39], [0.70], [0.58], [0.93].
  - Education coefficients positive (e.g., 0.45, 0.49, 0.91, 0.90) with (Standard errors) (0.36), (0.38), (0.61), (0.54); P-values not always below conventional significance thresholds.
  - R-squared values in Table 2 range from 0.31 to 0.37.
  - # Obs. listed: 285, 304, 285, 285, 150, 150, 150, 150 (depending on specification).
- Notes on tables:
  - All variables in tables are in first differences.
  - Dependent variable: saving rate.
  - Government spending variables: per capita spending expressed as a share of per capita urban disposable income (lagged one period).
  - Pooled provincial data used with fixed and time effects.

*Source: _wp1016 - References (PDF chapter/section).*

### 1. Consumption in China: Low and Falling ...............................................................................

### 1. Consumption in China: Low and Falling ...............................................................................

### Urban and Rural Saving Rates
- Section title: "2. Urban and Rural Saving Rates" (page 5)

### Urban and Rural Income
- Section title: "3. Urban and Rural Income" (page 5)

### Indicators of Urban and Rural Consumption and Income
- Section title: "4. Indicators of Urban and Rural Consumption and Income" (page 6)

### Household Income and GDP per Capita
- Section title: "5. Household Income and GDP per Capita" (page 7)

### Urban Household Saving Rate by Income Group
- Section title: "6. Urban Household Saving Rate by Income Group" (page 7)

### Health and Education Spending
- Section title: "7. Health and Education Spending" (page 9)

### Tables
- "1. Urban Households: Saving and Government Spending" (page 12)
- "2. Rural Households: Saving and Government Spending" (page 12)

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1016.pdf*

### References .............................................................................................................

### References

### I. Introduction — problem and research focus
- Consumption in China is unusually low and has continued to decline as a share of GDP over the past decade; rebalancing away from reliance on exports and investment toward consumption is urgent amid the collapse in global demand.
- Low household consumption (high saving) is often linked to precautionary motives due to under-developed government health, education, and pensions systems; households self-insure against uncertainty.
- Chamon and Prasad (2008) argue the breaking of the “iron rice bowl” in the 1990s increased household uncertainty and precautionary saving.
- Paper’s narrow focus: empirical relationship between government health and education spending and household saving.
- Main empirical conclusion:
  - Government spending on health, but not on education, reduces urban household saving.
  - Estimated magnitude: a 1 yuan increase in government health spending associated with a 2 yuan increase in household consumption.
  - Implication: Total (household plus government) consumption could increase by as much as 3 yuan depending on whether government health spending takes the form of consumption instead of transfers.
- For rural households: evidence is mixed; increased government health spending appears to affect savings only in higher-income provinces.
- Paper does not attempt to explain all drivers of consumption decline; household income is an important driver (Aziz and Cui, 2007).

### II. Consumption and saving in China: stylized facts
- Consumption trends:
  - Consumption as a share of GDP has steadily declined and is quite low relative to other countries.
  - Household consumption share has been falling over time; mechanically offset by rising investment and an expansion in net exports.
- Saving trends:
  - Saving rates have risen and help explain the decline in the consumption ratio.
  - Urban saving rate has steadily risen; rural saving is more volatile but has also risen.
  - Urban households account for the lion’s share of economic activity and drive aggregate saving increases.
- Income and consumption per capita:
  - Urban and rural household incomes have fallen steeply in per capita terms relative to GDP per capita (Figure 3 referenced).
  - From the early 1990s to 2008, urban household per capita income dropped from nearly 90 percent of GDP per capita to just under 70 percent.
- Attribution of consumption decline:
  - For urban households, roughly 60 percent of the decline in consumption since the early 1990s is attributable to a drop in disposable income as a share of GDP.
  - For rural households, falling income explains around three-fourths of the decline since the early 1990s.
- Additional stylized facts:
  - Real household consumption (CPI deflated) grew at an average annual rate of 8¾ percent over the past 10 years; GDP growth was 9¾ percent.
  - Household income rose quickly but not as fast as GDP.
  - No clear sign of consumption smoothing: volatility of consumption growth is greater than income growth for both urban and rural households.
  - Urban consumption accounts for three-fourths of household consumption; urbanization increased urban share of consumption.
  - Urban per capita income (consumption) is about 3 times rural per capita income (consumption).
  - Saving is positively related to income; higher income groups have gradually increasing saving rates, with acceleration around 2003.
  - Urban household saving rate by income group plotted (Figure 6 referenced).

### III. Reducing precautionary saving: econometric analysis and results
- Data and approach:
  - Provincial pooled data used to exploit variation in provincial health and education spending and differences in saving rates.
  - Data period: 1994–2007; robustness checked with 2003–07 sample.
  - Urban and rural households analyzed separately.
  - Government spending data do not distinguish urban vs. rural allocation; interpretation of urban/rural differences requires caution.
- A. Urban households — key econometric findings:
  - Statistically significant relationship between government health spending and urban saving.
  - Parameter estimate around -2 across samples and specifications:
    - Each 1 yuan of government health spending → 2 yuan decrease in saving (equivalently, 2 yuan increase in consumption).
    - Implies a 1 percent of GDP increase in government health spending would boost private consumption by 2 percent of GDP and yield a total demand effect of 3 percent of GDP for every 1 percent of GDP increase in health spending.
  - Interpretation: negative sign suggests government health spending reduces precautionary saving rather than merely substituting for private health spending.
  - Empirical evidence on private health care spending:
    - Urban health care spending (household survey) fell by more than ½ percent of urban household consumption over the past four years; government spending on health care more than doubled to the equivalent of 3 percent of total household consumption (Figure 7 referenced).
    - National accounts data show higher urban health care spending historically, but decline in recent years.
    - Rural healthcare spending increased over the past four to five years by ½–¾ percent of rural household consumption (survey and national accounts).
  - Government education spending:
    - Little evidence that higher government education spending decreases urban saving.
    - Coefficient for education never statistically significant; sign changes between full sample and latter period.
    - Possible reasons: increases concentrated in primary/secondary education while saving targets higher education; education spending per capita is more than five times larger than health spending and grew rapidly over the past four years.
- B. Rural households — key econometric findings:
  - Little evidence that higher health spending reduces precautionary saving for rural households.
  - Health coefficient always positive and never statistically significant in main samples, suggesting higher government health spending may increase saving or has little effect.
  - Possible explanations:
    - Rural households may be subsistence consumers and save a substantial part of any freed resources.
    - Rural areas may receive a smaller share of provincial health spending.
    - Increase in public spending small relative to expected rural health needs.
  - Education spending for rural households:
    - Little evidence of impact; coefficients positive and not statistically significant, though near significance in 2003–07 sample, consistent with near one-for-one substitution.
- C. Robustness checks:
  - Interaction with highest income provinces (Beijing, Tianjin, Shangdong, Jiangsu, Fujian, Guangdong, Zhejiang, and Shanghai) tested.
  - Rural results: government health spending in high income provinces is associated with a decline in rural household savings; magnitude similar to -2 found in urban regressions.
  - Urban results: no strong evidence that the impact of health spending differs in higher income provinces; interaction term usually not statistically significant though always negative.
  - Additional controls (real income growth, inflation, unemployment) included in robustness checks; main results on urban health spending remain substantively unchanged.

### IV. Conclusion — implications and policy relevance
- Main robust finding: higher government health spending reduces urban household saving; each additional yuan in government health spending boosts urban consumption by 2 yuan.
- Rural households: except in higher income provinces, no evidence that government health spending reduces saving.
- No evidence that higher government education spending affects urban or rural saving.
- Mechanism explanation:
  - Government spending can have competing effects: substitution (increasing saving by freeing household resources) versus reducing precautionary motives (decreasing saving).
  - Precautionary motives likely stronger for health than education; older generations drive much of the savings and are more concerned about healthcare costs.
- Policy relevance:
  - Broadening public health care coverage could have important effects on household precautionary savings and thereby on consumption.
  - Reorientation of government spending toward poorer rural areas and increased efficiency in public health care delivery will be key to improving health outcomes and impacting consumption behavior.
  - The government’s 2009–11 health reform strategy and priorities in the 11th five-year plan could improve rural health coverage and raise rural consumption, though challenges and inequalities in government health spending remain.
  - World Bank (2009) study: progress toward universal rural coverage is noted, but inequalities exist, with government health expenditure disproportionately benefiting the better off; over half of general government health spending supports urban insurance schemes whose members are disproportionately higher income.

### V. Key empirical tables and estimates (selected)
- Urban households (Table 1):
  - Health coefficient estimate ≈ -2.10 (1994–2007 sample, Model (1)); (Standard error) (0.72); [P-val] [0.00].
  - Health coefficient estimate ≈ -1.92 (various specifications); (Standard error) (0.86); [P-val] [0.03].
  - R-squared values reported in Table 1 range from 0.18 to 0.25 across specifications.
  - # Obs. listed: 285, 304, 285, 285, 150, 150, 150, 150 (depending on specification).
- Rural households (Table 2):
  - Health coefficient estimates positive (e.g., 0.51, 0.22, 0.37, 0.06) and not statistically significant; (Standard errors) (0.59), (0.58), (0.67), (0.64); [P-val] [0.39], [0.70], [0.58], [0.93].
  - Education coefficients positive (e.g., 0.45, 0.49, 0.91, 0.90) with (Standard errors) (0.36), (0.38), (0.61), (0.54); P-values not always below conventional significance thresholds.
  - R-squared values in Table 2 range from 0.31 to 0.37.
  - # Obs. listed: 285, 304, 285, 285, 150, 150, 150, 150 (depending on specification).
- Notes on tables:
  - All variables in tables are in first differences.
  - Dependent variable: saving rate.
  - Government spending variables: per capita spending expressed as a share of per capita urban disposable income (lagged one period).
  - Pooled provincial data used with fixed and time effects.

*Italicized bibliographic references and full citations are included in the References section of the source document.*

*Source: _wp1016 - References (PDF chapter/section).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1016.pdf_