## China’s Current Account

## Source details

**Canonical URL:** [China’s Current Account](https://www.imf.org/-/media/files/publications/cr/2019/1chnea2019004.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2019/1chnea2019004.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2019/1chnea2019004.pdf.json)

---

### A. Trends in China’s Current Account
- China’s current account surplus has declined significantly from its peak in 2008; the decline in 2018 was partly cyclical but the decade-long trend is largely structural.
- Services deficit widened materially, driven mainly by outbound tourism.
- Goods surplus moderated with volatility from commodity prices, infrastructure-led demand, and global growth.
- Key statistics:
  - Tourism balance swung from around 5bn USD surplus in 2008 to nearly 250bn USD deficit in 2018.
  - Number of Chinese outbound visitors rose from 46mn in 2008 to 162mn in 2018.
  - Current account surplus was 0.4 percent of GDP in 2018 (down from 1.4 percent of GDP in 2017).
  - Imports increased by around 0.7 percent of GDP in 2018; imports of crude and refined petroleum and integrated circuits explain close to half of that increase.
  - After appreciating for much of the period, the REER has stabilized since 2016.

### B. Structure of Trade and Value Added
- Imports and exports composition:
  - Imports dominated by intermediate and capital goods; household consumption is a relatively small share of overall imports.
  - Import structure shifted toward raw materials and commodities (e.g., oil and iron ore).
- Trade in value-added (VA) terms:
  - OECD 2015 data: over 80 percent of VA in total gross exports is due to China.
  - U.S–China bilateral trade balance in VA terms was USD 219 billion in 2015, 13 percent lower than gross trade balance of USD 251 billion.
  - Averaging 2005–2015, VA trade balance is 19 percent lower than corresponding gross numbers.
  - China’s domestic share in final demand (share of total VA) has increased with rising VA shares in high-tech sectors.

### C. Structural Drivers Behind the Decline in the Current Account Surplus
- Normalization of saving and investment:
  - Decline primarily reflects normalization of the domestic saving rate after its surge between 2000 and 2008.
  - Both savings and investment declined since the peak, but savings declined faster, narrowing the savings-investment gap.
- Household saving dynamics:
  - China remains an outlier with a very high household savings ratio; household savings expected to continue downward.
  - Contributing factors: demographic effects of the one-child policy, social safety net transformation, housing reforms, and rising income inequality.
- Investment dynamics:
  - Fixed asset investment (FAI) has likely peaked and is falling; real estate investment growth declined from nearly two decades of above 20 percent growth.
  - Moderation expected due to high vacancy ratios, declining working-age population, and slowing urban migration.
- Trade policy and openness:
  - Tariff liberalization expected to raise imports; average tariff ratio fell from 9.8 percent in 2017 to 7.5 percent after cuts in November 2018.
- Structural shifts increasing CA volatility:
  - Decline in processing trade share and increase in primary product imports (about half fuel and petroleum) make CA more sensitive to commodity cycles.
  - Rise in outbound tourism increases CA sensitivity to currency fluctuations.

### D. Cyclical Factors Affecting Recent Changes
- 2018 decline partly cyclical:
  - 1 percentage point drop in the current account surplus between 2017 and 2018 was mostly due to a decline in the goods trade balance.
  - Weak global demand and trade tensions contributed to lower exports in 2018.

### E. Import Price Drivers and Magnitude (2018)
- Oil prices spent much of 2018 in the range of US$70-85/bbl, up from US$55-65 range in 2017.
- Petroleum imports increased by $80bn or around 50 percent from 2017; the price increase accounted for close to 80 percent of the increase in petroleum imports.
- Integrated circuit imports rose by around $50bn compared with 2017; the price increase accounted for around 60 percent of that rise.
- Estimated price impacts on GDP:
  - Oil import price impact: 0.4 percent of GDP.
  - Semiconductor price impact: 0.2 percent of GDP.
  - Combined impact on the current account is estimated at around 0.4 percent of GDP (smaller than import-side GDP impact due to higher export prices).

### F. Baseline Outlook and Scenarios (Medium Term)
- Baseline assumptions:
  - Continued rebalancing: imports rise as savings fall faster than investment; exports grow at pace similar to trading partners; benign commodity prices; tourism deficit increases with GDP; no significant change in income account or asset structure.
  - Unless trade tensions escalate markedly, they are not expected to have a large impact on CA in the baseline.
- Baseline projection:
  - Current account expected to remain close to balance; the small 2018 surplus is expected to turn into a small deficit as structural factors continue.
- Scenarios and risks that could shift the CA:
  - Upside shocks: higher commodity prices or upside growth surprises could shrink goods surplus and lead to higher CA deficits.
  - Higher-than-projected Chinese growth via domestic demand would increase imports and CA deficit.
  - Tourism rising faster than GDP would increase services deficit.
  - Real appreciation beyond fundamentals and higher-than-projected IP payments could increase CA deficit.
  - Slower growth or reversal in rebalancing could increase the CA surplus and return external imbalances (e.g., faster import substitution, “Made in China 2025”).

### G. China’s International Investment Position and Income Account (USD bn)
- International assets: 7,324
- International liabilities: 5,194
- Net international assets: 2,130
- Asset composition (USD bn — percent of international assets):
  - Reserve assets: 3,168 — 43.3%
  - Equity portfolio investment: 684 — 13.2%
  - Trade credit: 597 — 8.2%
  - Loans: 710 — 9.7%
  - Currency and deposits: 394 — 5.4%
  - Direct investment: 1,899 — 25.9%
  - Other: 59 — 0.8%
- Liability composition (USD bn — percent of international liabilities):
  - Direct investment in China: 2,762 — 53.2%
  - Debt portfolio investment: 412 — 7.9%
  - Currency and deposits: 483 — 9.3%
  - Trade credit: 393 — 7.6%
  - Loans: 417 — 8.0%
  - Equity portfolio investment: 270 — 3.7%
  - Other: 42 — 0.8%
- Asset-liability observations:
  - Less than 30 percent of China’s external assets consist of higher-yielding risky assets such as direct and equity portfolio investment.
  - 70 percent of external liabilities comprise riskier instruments such as direct and portfolio equity investments.
  - Share of direct investment abroad in total assets increased from less than 5 percent in 2007 to around 26 percent in 2018.

### H. International Implications
- As a share of global GDP:
  - China’s current account imbalance peaked at 0.66 percent of global GDP in 2008.
  - Declined to 0.06 percent of global GDP in 2018.
  - Projected to reach about -0.01 percent of global GDP in 2024.
- Decline reduces excess global imbalances; China’s excess CA surplus peaked at around 0.3 percent of global GDP in 2015.
- Implications of declining surplus:
  - Less global capital available for US debtors and other ODI — can contribute to higher long-term yields.
  - Loss of large surplus as a cushion reduces countercyclical tools and may increase volatility of portfolio flows and RMB exchange rate.
- Policy options to support stable inflows and RMB internationalization:
  - Sell more bonds to foreign investors, including pension funds.
  - Settle transactions in RMB instead of USD.
  - Increase transparency, diversify external asset portfolio, and allow greater two-way exchange rate flexibility.

### I. Policy Recommendations (Domestic and External)
- Continue rebalancing from investment and exports toward consumption to sustain trend to smaller CA surplus.
- Continue opening up the economy (including tariff cuts) to prevent excessive surpluses returning.
- Prepare the economy and financial system for more volatile capital flows as CA structure changes:
  - Continue financial de-risking and micro/macro-prudential reforms.
  - Encourage higher quality, stable, and diversified inflows by further opening capital account.
  - Increase availability of internationally traded instruments; continue RMB internationalization.
  - Allow currency to respond more to short-term moves and allow greater two-way flexibility.

---

### Household Debt, Macroprudential Policy, and Consumption Dynamics

### Overview and key statistics
- As of June 2018, total household debt stood at over 50.3 percent of GDP, 32 percentage points of GDP higher than in 2008.
- Since the global financial crisis, household leverage accumulation was the highest among all BIS-reporting countries.
- Housing-related debt accounted for just below two-thirds of household debt at end-2018, a 7-ppt increase since 2013.
- More than 93 percent of mortgages outstanding were first-home mortgages as of end-March 2019.
- Mortgage participation rate increased from 8 to 18 percent between 2010 and 2016.

### Geographic and distributional patterns
- Household indebtedness varies by province; richest coastal provinces most indebted.
- In some provinces debt > 60 percent of GDP (threshold cited for adverse impact on consumption growth in some studies).
- Housing assets ~60 percent of all household assets on average.
- Share of debt held by highly indebted households (DTI > 4) rose from ~25 percent to almost 50 percent between 2010 and 2016.
- Increase in DTI among lower-income households reached almost 600 percent (relative terms reported).

### Macro-financial risks
- High indebtedness increases household vulnerability, can force abrupt deleveraging, constrain housing demand, harm developers’ financial soundness, and increase probability of banking distress.
- Cross-country studies link higher household debt to lower GDP growth.
- Threshold effects: negative long-run effects on consumption intensify when household debt/GDP ratio exceeds thresholds (60 percent cited).

### Empirical findings on consumption
- Micro-level (CFPS 2010–16) and province-level (2015Q1–2018Q4) regressions:
  - A 100-percent increase in DTI is associated with a 3.5-ppt increase in contemporaneous consumption growth but a 4.3-ppt decline in consumption growth two years later.
  - When lagged DTI < threshold 6.3, a 1-ppt increase in real income growth could boost consumption growth by 0.1 ppt; not statistically significant if DTI above threshold.
  - All household survey regression estimates reported are statistically significant at the 95-percent confidence level.

### Scenario simulations (assumptions and outcomes)
- Assumptions:
  - Real income growth slows gradually to 5.6 percent by 2023 (staff baseline real GDP growth).
  - Scenario 1: DTI increases at average pace of last five years.
  - Scenario 2: DTI remains at end-2018 level.
- Simulated outcomes:
  - Rising household debt could reduce annual consumption growth from nearly 7 percent in 2017 to less than 5 percent by 2030.
  - Scenario 1: higher consumption initially due to borrowing, but becomes lower than Scenario 2 over medium-to-long term.
  - If debt maintained at current level, consumption growth could stabilize at a rate higher than baseline income or GDP growth, aiding rebalancing.

### Macroprudential stance, gaps, and recommendations
- Current rules and gaps:
  - Maximum LTV for first-home buyers: 70 percent (cities with purchase restrictions) and 80 percent (cities without); provinces can reduce ratios (some as low as 60 percent).
  - Regulation requires borrower DSTI < 55 percent (mortgage-service-to-income cap 50 percent) — high relative to international norm of 30-50 percent.
  - Absence of a comprehensive credit registry; no supplemental capital requirements on household lending.
- Recommendations — strengthen systemic risk assessment and data:
  - Monitor non-housing loans and build vulnerability indicators: DSTI, DTI, loan maturity, asset/liability composition, interconnectedness, distributional metrics (share of debt at risk).
  - Strengthen interagency information sharing (2017 FSAP recommendation).
- Recommendations — strengthen macroprudential toolkit:
  - Adjust DSTI caps to 30-50 percent and extend DSTI limits to nonbank household loans.
  - Use stress testing of household DSTI to interest rate and income shocks.
  - Consider sectoral capital requirements on bank real estate exposures, with caution due to leakages to nonbanks.
- Other measures:
  - Develop comprehensive credit registry.
  - Strengthen micro-prudential supervision, consumer protection, financial education, transparency of financial contracts, and regulation of financial innovation products.

### Personal insolvency, credit registry, and related institutional reforms
- Recommendation to develop a personal insolvency framework to enable collective resolution and fresh start for honest debtors; pair legislative reforms with institutional arrangements (secured transaction registry, disclosure systems, insolvency practitioners, judicial capacity).
- Strengthen credit information system per 2017 FSAP: capture all individual debt obligations including nonbanks and P2P lenders; critical for lenders and policy makers and for personal insolvency regime functioning.

---

### Panel Regression Evidence (Key Results)
- Household-level main estimates:
  - Contemporaneous HD: Debt/disposable income coefficient on ∆c = 3.54*** (0.68); Debt/total asset (%) = 0.33*** (0.12).
  - Lagged HD: Debt/disposable income HD coefficient = -4.39*** (1.09); Debt/total asset (%) = -0.51*** (0.12).
  - Sample sizes: # obs. 9,059; 8,559; 8,990; 8,434; # households 6,365; 6,112; 6,199; 5,905.
- Threshold estimates (household-level):
  - Estimated threshold HD*: Debt/disposable income = 6.3; Debt/asset (%) = 2.4%.
  - Real income growth ∆y when HD_{i,t−1} < HD*: Debt/disposable income = 0.14*** (0.03).
  - When HD_{i,t−1} ≥ HD*: Debt/disposable income = -0.04 (0.07).
- Province-level findings:
  - Estimated threshold HD*: Debt/GDP = 49%; Debt/disposable income = 131%.
  - When below threshold, real income growth ∆y positively affects retail sales growth; when above, effect is zero or negative.
- Interpretation:
  - Short-term positive consumption effects from higher indebtedness can be followed by negative effects after two years.
  - Thresholds identify levels of indebtedness beyond which income elasticity of consumption falls substantially.

---

### Improving the Allocation of Corporate Credit, SOEs, and Access to Finance

### Key findings on corporate leverage and credit allocation
- Corporate debt-to-GDP fell from 142 percent in 2016 to 129 percent in 2018 via de-risking measures.
- SOE leverage ratios have fallen; 2018 industrial SOE leverage decline driven more by asset increases (~7 percent y-o-y) than debt reduction (~6 percent y-o-y).
- Majority of bank loans still flow to SOEs; banks provide over 70 percent of corporate financing.
- Adjusted SOE return on equity falls from average 8 percent to about −1.3 percent during 2011–15 when accounting for implicit support.
- SOEs estimated to have credit ratings about two to three notches higher than comparable POEs; SOEs pay an estimated 150-200 bp lower interest rates on bonds, and over 100 basis points less after controls.

### Policy recommendations — competitive and debt neutrality
- Apply competitive neutrality: treat enterprises under all ownership forms equally; publish SOE lists with categories (social, strategic, competitive).
- Remove implicit guarantees and increase banks’ risk weights on loans with implicit guarantees.
- Require SOEs to make a commercial rate of return and increase dividend transfers (target: 30 percent of profits by 2020).
- Sequence reforms to manage fiscal and financial spillovers: allow more defaults, restructure viable firms, develop distressed debt markets, and adopt market-based debt-equity swaps with strict solvency criteria.
- Strengthen legal and institutional insolvency frameworks and judiciary capacity.

### Improving access to equity and SME finance
- Propose Allowance for Corporate Equity (ACE) to reduce debt bias (examples given from Belgium, Italy, Cyprus, Portugal, Turkey).
- Expand equity financing channels, lower regulatory rating thresholds for bond issuance, develop SME-targeted capital markets, and harmonize credit bond regulations.
- Implement “1-2-5” policy goal:
  - at least 1/3 of new corporate loans from large banks to private firms;
  - at least 2/3 of new loans from small and medium banks;
  - at least 50 percent of all new corporate credit across banking system to private sector over next three years.
- Strengthen credit culture and lending based on cash-flow and movable collateral, improve credit bureaus/registries, and encourage better credit ratings and transparency.

### Policy tools and recent initiatives (selected figures)
- Dividend payout target for SOEs: 30 percent of profits.
- RRR cuts: 250 bps in 2018; Effective May 15, 2019 RRR cut for small and medium-sized banks up to 3.5 ppt to 8 percent.
- PBC collateral expansion (June 2018) to AA+ and AA corporate bonds with priority on small and micro enterprise, green, and agricultural bonds.
- November 2018: PBC subsidy of RMB 10 bn to China Bond Insurance for private enterprise debt sales.
- March 2019: increase loans to SMEs by large state-owned banks by 30 percent and lower SME financing cost by 1 percentage point.
- SASAC target: reduce central SOEs’ liability-to-asset ratio by 2 pp in 2018-20.

---

### Trade Diversion, Exports-at-Risk, and Trade-Tension Impacts

### Key findings on a hypothetical China–U.S. rebalancing deal
- If China closes the bilateral deficit with the U.S. (USD 336bn in 2017) by purchasing U.S. goods while total imports unchanged, extra U.S. purchases would displace other exporters (trade diversion).
- Estimated exports-at-risk (main scenario, sample aggregates):
  - USD 61bn from EU countries in sample.
  - USD 54bn from Japan.
  - USD 46bn from Korea.
  - USD 45bn from ASEAN countries.
- Country-level exposures (exports-at-risk as share of 2017 GDP):
  - > 3 percent of GDP: Oman, Angola, Singapore, Korea.
  - > 2 percent of GDP: Malaysia, Vietnam, Thailand.
  - ~1 percent of GDP: Germany, Japan.
  - Brazil: 1.3 percent of GDP (mainly oil seeds such as soya beans).
- Product allocation (share of total extra purchase, In percent):
  - Electronics, 13
  - Machinery, 12
  - Vehicles, 11
  - Oil seeds, 8
  - Aircraft, 3
  - Optical, 11
  - Mineral fuels/oils, 13
  - Plastics, 12
  - Pearls, 11
  - Wood pulp, 4

### Methodology and robustness
- Scenario assumes perfect substitution of non-U.S. suppliers by U.S. suppliers across top-ten products; allocations capped by China’s total imports of each product and U.S. export capacity.
- Alternative allocation scenarios show:
  - Concentrating on a subset (soya beans, LNG, oil, manufactured items) increases exports-at-risk for Germany, Japan, and ASEAN; oil exporters show higher risk as share of GDP.
  - Proportional distribution to China’s import structure raises exports-at-risk for ASEAN and Korea due to electronics exposure.
- Robustness with HS 6-digit data and ~200 economies: for 39 economies reported, differences in exports-at-risk (share of GDP) are within 1 percentage point for 35 economies.

### Simulated global impacts of tariff escalation
- U.S.–China tariff increases in 2018 projected to lower global GDP by 0.2 percent in 2020.
- Additional tariffs announced in May 2019 could cause global GDP to decline by an additional 0.3 percent in 2020 (on top of the 0.2 percent), under simulation assumptions described.
- Sectoral spillovers can be large as global value chains are repositioned.

### Policy implications
- Targeting bilateral trade balances alone will not necessarily reduce a country’s overall current account deficit; macroeconomic policies affecting saving/investment are more effective.
- Tariff increases are unlikely to change trade balances much because reduced imports tend to be offset by currency appreciation.
- An agreement between the U.S. and China should be:
  - Comprehensive and quickly reached to limit global disruption.
  - Reinforcing WTO rules, non-discriminatory, and based on market mechanisms and macroeconomic fundamentals.
- Avoid managed trade to minimize trade diversion and global value chain disruption.

*Source: PEOPLE’S REPUBLIC OF CHINA, INTERNATIONAL MONETARY FUND.*

### 1. China’s Current Account _______________________________________________________________ 3

### China’s Current Account

### A. Trends in China’s Current Account
- China’s current account surplus has declined significantly from its peak in 2008; the decline in 2018 was partly cyclical but the decade-long trend is largely structural.
- Current account components:
  - Services deficit widened materially, driven mainly by outbound tourism.
  - Goods surplus moderated, with volatility from commodity prices, infrastructure-led demand, and global growth.
- Key statistics and observations:
  - Tourism balance swung from a small surplus of around 5bn USD in 2008 to a deficit of nearly 250bn USD in 2018.
  - Number of Chinese outbound visitors rose from 46mn in 2008 to 162mn in 2018.
  - China’s current account surplus was 0.4 percent of GDP in 2018 (down from 1.4 percent of GDP in 2017).
  - The rise in imports in 2018 was driven in part by imports of crude and refined petroleum and integrated circuits, which can explain close to half of the increase in 2018 imports (imports increased by around 0.7 percent of GDP).
  - After appreciating for much of the period, the REER has stabilized since 2016.

### B. Structure of China’s Trade and Value Added
- Imports and exports composition:
  - Imports dominated by intermediate and capital goods; household consumption is a relatively small share of overall imports.
  - Import structure shift: increase in raw materials and commodities (e.g., oil and iron ore) to feed domestic demand, while manufacturing surplus remains sizeable but has plateaued.
- Trade in value-added terms:
  - OECD data from 2015 shows over 80 percent of value-added (VA) in total gross exports is due to China.
  - In 2015, the U.S.–China bilateral trade balance in VA terms was USD 219 billion, which is 13 percent lower than the gross trade balance of USD 251 billion.
  - Averaging available years (2005–2015), the VA trade balance is 19 percent lower than corresponding gross numbers.
  - China’s domestic share in final demand (as a share of total value added) has increased, reflecting moves up the value chain and rapid increases in VA shares in high-tech sectors.

### C. Structural Drivers Behind the Decline in the Current Account Surplus
- Normalization of a previously extraordinary saving rate:
  - The decline in the current account surplus primarily reflects normalization of the domestic saving rate after its surge between 2000 and 2008.
  - Both savings and investment have declined since the peak, but savings declined faster, narrowing the savings-investment gap.
- Household saving dynamics:
  - China remains an outlier with a very high household savings ratio; household savings are expected to continue on a downward trajectory.
  - Contributing factors to high household savings include demographic effects of the one-child policy, transformation of the social safety net and job security during economic transition, housing reforms, and rising income inequality.
  - Normalization of savings will depend on pace and success of rebalancing toward consumption, improvements in the social safety net, and reductions in income inequality.
- Investment dynamics:
  - Fixed asset investment (FAI) has likely peaked and is falling as rebalancing continues; real estate investment growth has declined from nearly two decades of above 20 percent growth.
  - Investment is expected to continue moderating due to high vacancy ratios, declining working-age population, and slowing urban migration.
  - As investment growth moderates, imports for commodities should decline, but this may be offset by rising consumption and associated imports.
- Trade policy and openness:
  - Tariff liberalization expected to raise imports: government estimates show the average tariff ratio fell from 9.8 percent in 2017 to 7.5 percent after cuts in November 2018.
- Structural shifts increasing current account volatility:
  - Decline in processing trade share (where exports and imports move largely in tandem) and increase in primary product imports (about half fuel and petroleum) make the current account more sensitive to commodity cycles.
  - Rise in outbound tourism increases current account sensitivity to currency fluctuations.

### D. Cyclical Factors Affecting Recent Changes
- 2018 sharp decline in the current account surplus was partly cyclical:
  - The 1 percentage point drop in the current account surplus between 2017 and 2018 was mostly due to a decline in the goods trade balance.
  - Weak global demand and trade tensions contributed to lower exports in 2018.

### E. Implications and Policy Recommendations
- Policy priorities:
  - Continue rebalancing from investment and exports toward consumption to sustain the trend toward a smaller current account surplus.
  - Continue opening up the economy to ensure excessive surpluses do not return (including tariff cuts).
  - Prepare the economy and the financial system to handle more volatile capital flows as the current account structure changes.
- Expected medium-term outlook:
  - With China’s faster growth relative to trading partners and a large existing export market share, exports are likely to grow at a pace similar to trading partner growth, while imports are expected to outpace exports as consumption rises—leading to a smaller current account surplus or near balance.
  - China’s move up the value chain and increasing domestic VA shares should continue to alter trade patterns and reduce reliance on re-exports.

### F. Global Perspective
- Decline in China’s surplus has lowered global imbalances, with heterogeneous impacts:
  - Trade balances vis-à-vis China improved for Korea, Germany, and Brazil.
  - Trade balances vis-à-vis China deteriorated for Japan, India, and Indonesia.

*Prepared by Pragyan Deb, Albe Gjonbalaj, and Swarnali Ahmed Hannan; excerpted from “The Drivers, Implications and Outlook for China’s Shrinking Current Account Surplus.”*

### 13.      The rise in imports was driven by an increase in oil and integrated circuit prices. Oil

### 1chnea2019004 - 13.      The rise in imports was driven by an increase in oil and integrated circuit prices. Oil

### Import price drivers and magnitude
- Oil prices spent much of 2018 in the range of US$70-85/bbl, up from US$55-65 range in 2017.
- Petroleum imports increased by $80bn or around 50 percent from 2017; the price increase accounted for close to 80 percent of the increase in petroleum imports.
- Prices of semiconductors surged after years of declines, pushing up integrated circuit imports by around $50bn compared with 2017; the increase in prices accounted for around 60 percent of the increase in integrated circuit imports in 2018.

### Summary of current account developments (2018 and recent)
- Structural factors driving decline in China’s current account surplus:
  - rebalancing;
  - increase in outbound tourism;
  - moderation in goods surplus due to market saturation and growth differentials with trading partners.
- Cyclical component in 2018:
  - Price impact of oil and semiconductor prices on imports is estimated to be 0.4 and 0.2 percent of GDP respectively.
  - Impact on the current account (CA) likely to be smaller (due to higher export prices) and is estimated at around 0.4 percent of GDP.

### Baseline outlook (medium term)
- Under the baseline of continued rebalancing, the current account is expected to remain close to balance and the small current account surplus recorded in 2018 is expected to turn into a small deficit as structural factors continue:
  - Import demand increases as savings fall faster than investment – a rise in share of private consumption.
  - Export growth slows due to market saturation and continued higher growth in China relative to trading partners.
  - Benign outlook for commodity prices.
  - Tourism deficit increases in line with GDP.
  - No significant change in income account and the structure of assets.
  - Unless trade tensions escalate markedly, they are not expected to have a large impact on CA with offsetting effects on exports and imports. (Impact of tensions in high-tech exports is beyond scope and not taken into account in the baseline.)

### China’s international investment position and income account
- International assets: 7,324 (USD bn).
- International liabilities: 5,194 (USD bn).
- Net international assets: 2,130 (USD bn).
- Asset composition (USD bn and percent of international assets):
  - Reserve assets: 3,168 — 43.3%
  - Equity portfolio investment: 684 — 13.2%
  - Trade credit: 597 — 8.2%
  - Loans: 710 — 9.7%
  - Currency and deposits: 394 — 5.4%
  - Direct investment: 1,899 — 25.9%
  - Other: 59 — 0.8%
- Liability composition (USD bn and percent of international liabilities):
  - Direct investment in China: 2,762 — 53.2%
  - Debt portfolio investment: 412 — 7.9%
  - Currency and deposits: 483 — 9.3%
  - Trade credit: 393 — 7.6%
  - Loans: 417 — 8.0%
  - Equity portfolio investment: 270 — 3.7%
  - Other: 42 — 0.8%
- Less than 30 percent of China’s external assets consist of higher-yielding risky assets such as direct and equity portfolio investment; 70 percent of external liabilities comprise riskier and therefore higher (expected) return instruments such as direct and portfolio equity investments.
- Share of direct investment abroad in total assets increased from less than 5 percent in 2007 to around 26 percent in 2018.

### Scenarios and risks that could shift the current account
- Upside growth surprises or higher commodity prices (particularly oil) could materially shrink the goods surplus and lead to higher current account deficits.
- Higher-than-projected growth (via domestic demand or credit) would increase imports and push up the current account deficit; higher Chinese growth could push up commodity prices, further increasing imports.
- Tourism could pick up as per capita incomes rise and grow faster than GDP, increasing the tourism deficit.
- Real appreciation beyond fundamentals and higher-than-projected IP payments could increase the current account deficit.
- Lower growth or slowdown/reversal in rebalancing could increase the current account surplus and return external imbalances:
  - Lower growth can decrease import demand and increase the current account surplus.
  - Offset via tax cuts (lower import intensity) versus public/quasi-public investment (higher import intensity) matters for import demand.
  - Faster import substitution (e.g., Made in China 2025, semiconductors) can lower import demand and increase global market share, increasing the current account surplus.
  - Greater availability of high-end and luxury products domestically can reduce overseas spending by Chinese tourists, moderating the services deficit and pushing up the overall current account.

### Policy recommendations (domestic and external)
- Continue and accelerate rebalancing to increase consumption demand.
- Continue regulatory and supervisory reforms to stabilize leverage and pursue “debt neutrality” with SOEs.
- “Made in China” should focus on “comparative advantage” and not “import substitution”.
- Deepen and accelerate opening up and continue to support the international trading system:
  - Further reduction in import tariffs and increased trade openness.
  - Further opening up of the service sector.
  - Liberalization of restrictions to trade and investment regime.
  - Address structural issues such as intellectual property enforcement.
- Prepare the financial system for greater volatility and larger capital inflows:
  - Continue financial de-risking and associated micro and macro-prudential reforms.
  - Encourage higher quality, stable and diversified inflows by further opening up the capital account.
  - Increase transparency and regulatory reforms to adhere to international standards and encourage investment by institutional and long-term investors (e.g., reforms to the ratings industry).
  - Increase availability of internationally traded instruments; continue RMB internationalization; diversify external asset portfolio to generate higher returns.
  - Allow the currency to respond more to short-term moves and allow for greater two-way flexibility of the exchange rate.

### International implications
- As a share of global GDP, China’s current account imbalance has declined:
  - Peaked at 0.66 percent of global GDP in 2008.
  - Declined to 0.06 percent of global GDP in 2018.
  - Projected to reach about -0.01 percent of global GDP in 2024.
- Decline in China’s current account surplus reduces excess global imbalances; China’s excess current account surplus peaked at around 0.3 percent of global GDP in 2015.
- The decline in China’s surplus diminishes a source of global savings, with implications:
  - Less global capital likely available for US debtors and other ODI, which can contribute to higher long-term yields.
  - China’s large surpluses provided a cushion of foreign exchange reserves and countercyclical tools; without this backstop, more volatile components of the financial account (portfolio capital flows) will have larger impact as FDI becomes less important, implying less flexibility for economic and currency management and greater RMB exchange rate volatility.
- China could support more stable capital inflows by encouraging RMB internationalization:
  - Sell more bonds to foreign investors, including long-term investors such as pension funds.
  - Settle transactions in RMB instead of USD.
  - These steps would increase availability and use of RMB assets, but require more flexibility and transparency in RMB management.

*Source: PEOPLE’S REPUBLIC OF CHINA, INTERNATIONAL MONETARY FUND.*

### 1.      Household debt has been rising rapidly in China since the global financial crisis... As of

### 1chnea2019004 - 1.      Household debt has been rising rapidly in China since the global financial crisis... As of

### Overview
- As of June 2018, total debt of Chinese households stood at over 50.3 percent of GDP, above the emerging market average, and 32 percentage points of GDP higher than in 2008.
- Since the global financial crisis, the speed with which households accumulated leverage was the highest among all BIS-reporting countries since the global financial crisis.

### Drivers and composition of household debt
- Housing-related debt (mortgage loans and Housing Provident Fund lending) accounted for just below two-thirds of all household debt at the end of 2018, a 7-percentage point (ppt) increase since 2013.
- The remainder of household debt was almost evenly split between consumption loans (including credit card debt) and loans extended to households for commercial purposes (mostly SMEs).
- Most mortgages:
  - are extended on fixed-term rates, with maturity of about 10-20 years;
  - have a minimum down payment of 20 (30) percent for the first (second) home (can be adjusted upwards by regional authorities);
  - mortgage rates are usually set at a discount/premium of 0-15 percent below/above the central banks’ benchmark lending rate, with the adjustment depending on market conditions.
- More than 93 percent of mortgages outstanding were first-home mortgages as of end-March 2019.
- Mortgage participation rate increased from 8 to 18 percent between 2010 and 2016.

### Geographic and income distribution
- Household indebtedness varies significantly by province; richest, coastal provinces are the most indebted.
- In some provinces, debt has already risen above 60 percent of GDP—a threshold cited for adverse impact on consumption growth in some studies.
- On average, housing assets constitute around 60 percent of all household assets (declined over time for richer households).
- The share of debt held by highly indebted households (DTI ratio above 4) increased from around a quarter to almost half between 2010 and 2016.
- The increase in DTI ratio among the lower income households reached almost 600 percent (in relative terms reported in the source).

### Macro-financial risks from high household debt
- Financial stability risks:
  - High indebtedness increases household vulnerability to adverse shocks and can force abrupt deleveraging with significant macro-financial impact.
  - Deleveraging could constrain housing demand growth, put pressure on house price growth, and harm property developers’ financial soundness.
  - Severe house price corrections could reduce banks’ financial soundness through lower collateral valuation and increase funding pressure for property developers.
  - Higher household debt may increase the probability of banking distress and worsen recessions; recessions preceded by housing busts are longer and more severe (IMF, 2012).
- Macroeconomic risks:
  - Cross-country studies indicate higher household debt may lead to lower GDP growth (Mian et al., 2017; Jorda et al., 2016).
  - Higher debt repayments could constrain future consumption growth, reduce real estate investment, lower fiscal revenues, increase public support of home ownership, contribute to excess savings and growing external imbalances.
- Threshold effects:
  - Negative long-run effects on consumption growth intensify when household debt/GDP ratio exceeds certain thresholds (60 percent for effects on consumption growth cited).
  - Provincial evidence (Tian et al., 2018) links high household debt in certain provinces to lower consumption growth.

### Impact of high household indebtedness on consumption (empirical results)
- Data and methods:
  - Micro-level: China Family Panel Studies (CFPS) household survey (2010-16) used to compute DTI and debt/asset ratios.
  - Macro-level: Quarterly provincial data span 2015Q1 to 2018Q4 covering 24 provinces. Debt/GDP ratio and DTI ratio used in province-level regressions.
- Key regression findings:
  - An increase in household indebtedness is associated with higher contemporaneous consumption growth but lower consumption growth two years later.
  - A 100-percent increase in DTI ratio is associated with a 3.5-ppt increase in contemporaneous consumption growth but a 4.3-ppt decline in consumption growth two years later.
  - When lagged DTI ratio is below a threshold of 6.3, a 1-ppt increase in real income growth could boost consumption growth by 0.1 ppt; the estimate is not statistically significant if DTI ratio is above that threshold (about 10 percent of households in the sample).
  - All household survey regression estimates reported are statistically significant at the 95-percent confidence level.

### Scenario simulations: illustrative paths for consumption growth
- Assumptions:
  - Real income growth slows gradually to 5.6 percent by 2023 (in line with staff’s baseline projections for real GDP growth).
  - Scenario 1: DTI ratio increases at the average pace of the last five years.
  - Scenario 2: DTI ratio remains at the end-2018 level.
- Simulated outcomes:
  - Rising household debt could reduce annual consumption growth from nearly 7 percent in 2017 to less than 5 percent by 2030.
  - Consumption growth in Scenario 1 is higher in the first few years due to higher borrowing but becomes lower than Scenario 2 over the medium-to-long term.
  - If household debt is maintained at the current level, consumption growth could stabilize at a rate higher than the baseline income or GDP growth rate, aiding internal rebalancing toward consumption.

### International experience and current macroprudential stance
- Effective measures internationally:
  - Improved data quality and comprehensive indebtedness measures help monitor household vulnerabilities.
  - Demand-side macroprudential measures (DSTI, LTV limits) and supply-side measures (limits on bank credit growth, loan loss provisions) are effective in mitigating negative effects of household debt on consumption and growth.
  - Strong banking supervision, higher bank capitalization, and credit registries reduce crisis risk.
- China’s measures and gaps:
  - Maximum LTV ratio for first-home buyers: 70 percent for cities with home purchase restrictions and 80 percent for cities without purchase restrictions; provinces can reduce ratios (some as low as 60 percent).
  - Current regulation requires borrower’s monthly DSTI ratio to be less than 55 percent (with a cap on mortgage-service-to-income ratio of 50 percent), which is relatively high compared to the international norm of 30-50 percent.
  - Absence of a comprehensive credit registry hinders lenders’ ability to evaluate household total liabilities.
  - China has not put in place supplemental capital requirements of different risk weights on household lending.

### Policy implications and recommendations
- Strengthen systemic risk assessment and data:
  - Extend monitoring beyond mortgages to include non-housing loans (non-housing loans account for about half of household debt for lower-income households).
  - Collect and process data to build household vulnerability indicators: leverage (consistent and comprehensive DSTI and DTI ratios), liquidity (average maturity of household loans or assets by type), composition of assets and liabilities, and interconnectedness with banks and non-bank financial sector.
  - Incorporate distributional aspects: share of “debt at risk” and share of “risky” borrowers.
  - Strengthen interagency information and data sharing as recommended by the 2017 FSAP (IMF, 2017b).
- Strengthen the macroprudential toolkit:
  - Adjust DSTI caps to the international norm of 30-50 percent and extend DSTI limits to other household loans including those from non-bank financial institutions.
  - Use stress testing of household DSTI ratio to interest rate and income shocks.
  - Consider sectoral capital requirements on banks’ exposures to the real estate sector, but treat with caution due to potential leakages to non-banks and lower effectiveness versus demand-side tools.

*Prepared by Fei Han, Emilia M. Jurzyk, Wei Guo, Yun He, and Nadia Rendak. Source: IMF staff analysis in the provided chapter.*

### 16.      Combining different tools and increasing the scope of macroprudential policy may

### 16.      Combining different tools and increasing the scope of macroprudential policy may

### Macroprudential policy: combining tools and expanding scope
- Combining different tools can enhance policy effectiveness and reduce leakages from any single measure.
- More active use of DSTI limits backed by comprehensive analyses could enhance the effectiveness of LTV limits by restricting the use of short-term consumption loans for housing down payment.
- All leveraged providers of credit should be included in the purview of macroprudential policy to avoid migration of credit provision from banks to less-constrained nonbanks (Jacome and Nier, 2011).
- Given the importance of mortgages in household debt and China’s specific land ownership structure, land policies should be used to increase effective housing supply and promote a transparent and efficient secondary market for land transactions (Box 1).

### Personal insolvency and debt enforcement
- China should consider developing a legal framework for personal insolvency while ensuring effective debt enforcement of commercial claims.
- The Enterprise Bankruptcy Law that came into effect in 2007 only applies to companies and not to individual debtors.
- Chinese law includes provisions that allow certain protections for the debtor against enforcement by creditors; examples include:
  - Courts can allow debtors to pay their obligation in installments (Article 108 of the General Principles of the Civil Law of the People’s Republic of China).
  - Protections allowing debtors to keep certain minimum assets to take care of themselves and their families.
  - Some local courts are reportedly developing tools that would allow debt write-off for debtors with virtually no assets.
- Limitations of current mechanisms:
  - Existing mechanisms are not well suited for situations where the debtor may have multiple creditors.
  - Without a well-functioning personal insolvency regime, debtors cannot get a “fresh start” and may have to cut consumption more substantially to repay debt.
- Recommendation: develop a personal insolvency framework to address inevitable situations of over-indebtedness while mitigating moral hazard (Box 2).

### Credit information system
- A comprehensive credit information system should be developed as a prerequisite for strong policy frameworks.
- The credit registry’s quality and scope should be strengthened in line with the 2017 FSAP recommendation by capturing all individual debt obligations including those from nonbanks (such as P2P lenders) and other service providers (e.g., utilities and telecommunications).
- Benefits:
  - Helps lenders better assess the credit risk of borrowers.
  - Allows policy makers to better monitor and assess financial stability risks.
  - Important for bankruptcy prevention and effective functioning of a personal insolvency regime (Box 2).

### Financial sector supervision and consumer protection
- Strengthen micro-prudential supervision with stronger supervisory powers or more stringent capital regulation frameworks to better contain negative effects of rapidly rising household debt on macroeconomic and financial stability.
- Strengthen consumer financial protection through measures such as:
  - Expanding financial education.
  - Increasing the transparency of financial contracts.
  - Regulating certain financial innovation products.
- These measures would help unsophisticated consumers make wiser finance decisions and enhance overall financial stability (IMF, 2017a).

### Box 1 — Land policies and house prices in China: main points
- Four types of policies mainly affect house prices in China: monetary (mainly interest rates and credit volume), macroprudential (mainly down payment requirements), tax, and land policies.
- Land policies affect house prices mainly through the volume and composition of land supply.
- Land ownership and rights:
  - Land is owned by the state (urban land) or by rural collectives (rural land); land-use rights are privately owned, but those for residential use are only valid for 70 years.
- Supply and composition effects:
  - Higher land supply, without significant land hoarding by developers, should increase housing supply and alleviate upward pressure on house prices.
  - Composition of land supply (residential, commercial, office) affects residential housing supply and house prices.
  - Residential land is divided into four types: small- and medium-sized apartments, villa and deluxe apartments, ‘economic’ houses (subsidized houses), and others.
  - Current policies aim to increase land supply for small- and medium-sized apartments while restricting supply for villas and deluxe apartments.
  - Increases in land supply for economic houses may limit land supply for other houses, leading to higher market-determined house prices.
- Administrative measures:
  - Restrictions on floor area ratios (ratio of a building’s total floor area to the size of the land upon which it is built) can guide land supply toward small- and medium-sized apartments.
  - Example: floor area ratios in Shenzhen were recently raised in the future planning of the Great Bay Area to facilitate higher-floor buildings.
- Policy recommendations:
  - Reduce land hoarding and enhance market mechanisms in the secondary land transaction market.
  - Enforce policies to contain land hoarding including penalties for developers.
  - Adopt alternative sources for local government financing such as property taxes.
  - Recent positive steps: relaxation of Hukou requirements in smaller cities and allowing some collectively-owned rural land to be transacted in the land market (without having to be purchased by the state first).
  - Potential room for raising floor area ratios in Tier 1 cities (less than 2 in 2010) compared to New York, Singapore, and Seoul which were mostly above 3 in the same period according to CICC estimates.

### Box 2 — Is it time for China to develop a personal insolvency framework? Key elements and conditions
- Role of personal insolvency:
  - Personal insolvency is a tool to address consequences of over-indebtedness by establishing rules and procedures for fair burden sharing between debtors and creditors.
  - Personal insolvency is a collective proceeding with participation of multiple creditors and allows honest debtors to get a “fresh start”.
  - Some jurisdictions include individual entrepreneurs and unincorporated micro and small enterprises to reduce stigma of business failure and promote entrepreneurship (Bergthaler, et al. 2015).
- Cross-country reform context:
  - Many countries reformed insolvency frameworks in the past decade; reforms aimed at strengthening corporate restructuring and establishing new personal insolvency regimes.
  - Several Asian countries have personal insolvency frameworks; examples include Japan, Korea, Philippines, Thailand, Singapore, and Malaysia; India’s 2016 reform is pending implementation.
- Design recommendations for China:
  - Adoption of a personal insolvency law would be an important step; no international standards exist, but cross-country experience can guide design.
  - The law could apply to individual debtors, including individual entrepreneurs and unincorporated enterprises.
  - Procedures should allow insolvent debtors with sufficient repayment capacity to restructure their debt (including secured debt) over a period of time by making partial payments, while providing a swift discharge from debt (a “fresh start”) after liquidation of the debtor’s assets for honest debtors with no repayment capacity.
  - The law should be developed in broad consultation with stakeholders.
  - Consider developing out-of-court mechanisms to facilitate resolution of personal debt distress (e.g., debt counseling, Financial Ombudsmen, mediation, arbitration).
- Implementation conditions:
  - Legislative reforms should be paired with institutional arrangements, including but not limited to:
    - (i) an effective system of registration and enforcement of secured transactions, including mortgages;
    - (ii) a well-functioning system of disclosure and verification of information about the debtor’s financial situation;
    - (iii) establishment of a community of qualified and regulated insolvency practitioners to assist debtors, including preparation of repayment plans; and
    - (iv) increasing judicial knowledge and competency in handling insolvency matters, including development of specialized expertise in courts.
  - Consider establishing or assigning an agency to provide advice to debtors facing financial difficulties.
  - Public education is needed to explain personal insolvency to Chinese society so it can gain acceptance as a standard market-economy tool.

### Annex I — Design of panel regressions: effects of household debt on consumption growth
- Two panel regression models estimated to explore trade-off between positive short-term and negative longer-term effects.
- Model for lagged effects (equation (1)):
  - ∆c_{i,t} = β_0 + α_i + β_1 ∆y_{i,t} + β_2 HD_{i,t−1} + β_3 (∆y_{i,t} * HD_{i,t−1}) + β_4 X_{i,t} + ε_{i,t}
  - ∆c_{i,t} and ∆y_{i,t} are household (real) expenditure growth and (real) disposable income growth.
  - HD_{i,t−1} is measure of household indebtedness with a two-year lag (CFPS survey conducted every two years).
  - X_{i,t} is vector of household characteristics; α_i is household fixed effect.
  - For a household with income growth ∆y, impact of lagged household indebtedness on consumption growth is β_2 + β_3 ∆y.
  - For a household with lagged household indebtedness HD, the income elasticity is β_1 + β_3 HD.
  - β_3 measures impact of household indebtedness on income elasticity.
- Short-term effects estimated with analogous model using contemporaneous household indebtedness (equation (2)):
  - ∆c_{i,t} = β_0 + α_i + β_1 ∆y_{i,t} + β_2 HD_{i,t} + β_3 (∆y_{i,t} * HD_{i,t}) + β_4 X_{i,t} + ε_{i,t}
- Threshold effects (equation (3)):
  - ∆c_{i,t} = β_0 + α_i + β_s ∆y_{i,t} + β_3 X_{i,t} + ε_{i,t}, where
    - β_s = β_1 if HD_{i,t−1} < HD^*
    - β_s = β_2 if HD_{i,t−1} ≥ HD^*
  - Expectation: when household indebtedness is lower than threshold HD^*, positive income elasticity (β_1 > 0); when higher than threshold, income elasticity may be zero (β_2 = 0).
- Macro-level models:
  - Province-level panel regression similar to equation (1) with controls including change in lending rate, house price growth, public consumption growth, fixed-asset investment growth, and (national) stock price growth (GFSR, 2017).
  - Dynamic panel models used to control for lagged dependent variable due to consumption inertia.
  - Threshold panel regression estimated to examine threshold effect at macro level.
  - Models estimated with quarterly database spanning from 2015Q1 to 218Q4 covering 24 provinces with available variables, and with one-year lagged household indebtedness measures (debt/GDP ratio and DTI ratio).
- Household characteristics used include changes in household size and the household head’s marital and education level; adding more characteristics (e.g., household location) does not alter main results.
- Some other macroeconomic variables (such as unemployment rate) were considered but do not alter main results.

*Source: 16. Combining different tools and increasing the scope of macroprudential policy may (IMF chapter content).*

### Annex II. Panel Regression Results

### Annex II. Panel Regression Results

### Household-Level Panel Regression Results — Main Estimates
- Dependent variable: Real expenditure growth ∆c_i,t
- Real income growth ∆y_i,t coefficients (lagged HD / contemporaneous HD):
  - Debt/disposable income (lagged HD): 0.13*** (0.02)
  - Debt/total asset (%) (lagged HD): 0.12*** (0.02)
  - Debt/disposable income (contemporaneous HD): 0.11*** (0.02)
  - Debt/total asset (%) (contemporaneous HD): 0.08*** (0.02)
- HD coefficients:
  - Debt/disposable income (lagged HD): -4.39*** (1.09)
  - Debt/total asset (%) (lagged HD): -0.51*** (0.12)
  - Debt/disposable income (contemporaneous HD): 3.54*** (0.68)
  - Debt/total asset (%) (contemporaneous HD): 0.33*** (0.12)
- Interaction ∆y_i,t * HD:
  - Debt/disposable income (lagged HD): -0.0002 (0.01)
  - Debt/total asset (%) (lagged HD): -0.002*** (0.001)
  - Debt/disposable income (contemporaneous HD): 0.01** (0.003)
  - Debt/total asset (%) (contemporaneous HD): -0.001 (0.001)
- Controls:
  - Change in family Size: 4.99*** (1.71); 4.91*** (1.78); 5.02*** (1.70); 6.69*** (1.88) across columns
  - Change in household head marital status: -9.90*** (3.74); -9.11** (4.31); -6.61** (3.22); -5.19 (3.60)
  - Change in household head education level: 3.38** (1.62); 4.06** (1.81); 2.86** (1.34); 4.36*** (1.47)
- Constants: 26.64*** (2.76); 31.74*** (3.25); 35.90*** (2.88); 41.70*** (3.46)
- Fixed effects: Household fixed effects Yes; Year fixed effects Yes
- Sample sizes and fit:
  - # obs.: 9,059; 8,559; 8,990; 8,434
  - # households: 6,365; 6,112; 6,199; 5,905
  - Overall-R^2: 0.04; 0.03; 0.03; 0.02

### Household-Level Threshold Panel Regression Results
- Dependent variable: Real expenditure growth ∆c_i,t
- Estimated threshold HD*:
  - Debt/disposable income: 6.3
  - Debt/asset (%): 2.4%
- Real income growth ∆y_i,t when HD_i,t-1 < HD*:
  - Debt/disposable income: 0.14*** (0.03)
  - Debt/asset (%): 0.62*** (0.09)
- Real income growth ∆y_i,t when HD_i,t-1 ≥ HD*:
  - Debt/disposable income: -0.04 (0.07)
  - Debt/asset (%): 0.03 (0.02)
- Controls and constants:
  - Change in family Size: 1.98 (2.38); 3.15 (2.76)
  - Change in household head marital status: -10.50* (6.03); -12.08 (8.33)
  - Change in household head education level: 1.10 (2.47); 0.89 (2.76)
  - Constant: 28.03*** (3.44); 33.77*** (3.65)
- Fixed effects: Household fixed effects Yes; Year fixed effects Yes
- Sample sizes and fit:
  - # obs.: 1,689; 1,455
  - # households: 563; 485
  - Overall-R^2: 0.03; 0.05

### Province-Level Panel Regression Results — Main Estimates
- Dependent variable: Real retail sales growth (qoq) ∆c_i,t
- Household indebtedness measures: Debt/GDP (%) and Debt/disposable income (%)
- Estimation methods: Fixed-effects with Driscoll-Kraay s.e.; Dynamic panel GMM (both for each HD measure)
- Key coefficients (Debt/GDP % specifications / Debt/disposable income (%) specifications):
  - ∆c_{i,t−1}: — ; -0.25** (0.10) ; — ; -0.26** (0.09)
  - Real income growth ∆y_{i,t}: 0.20*** (0.06); 0.23** (0.10); 0.13* (0.07); 0.20* (0.12)
  - HD_{i,t−4}: -0.10** (0.04); -0.12** (0.05); -0.05** (0.02); -0.08*** (0.02)
  - ∆y_{i,t} * HD_{i,t−4}: -0.002 (0.001); -0.003 (0.002); -0.0001 (0.001); -0.001 (0.001)
  - Real house price growth (lagged): 0.06** (0.02); -0.04 (0.03); 0.06** (0.02); -0.03 (0.03)
  - Real stock price growth (lagged): -0.02 (0.02); 0.004 (0.04); -0.02 (0.02); -0.01 (0.05)
  - Lending rate (lagged): -1.34 (1.02); 6.42 (3.99); -1.41 (1.11); 5.90 (3.92)
  - Real public consumption growth (lagged): 0.03 (0.14); -0.22 (0.19); 0.05 (0.15); -0.19 (0.19)
  - Real fixed asset investment growth (lagged): 0.02*** (0.004); 0.02 (0.02); 0.02*** (0.005); 0.02 (0.02)
- Sample sizes:
  - # obs.: 288; 240; 288; 240
  - # provinces: 24; 24; 24; 24

### Province-Level Threshold Panel Regression Results
- Dependent variable: Real retail sales growth (qoq) ∆c_i,t
- Estimation method: Fixed-effects threshold (for both HD measures)
- Estimated threshold HD*:
  - Debt/GDP: 49%
  - Debt/disposable income: 131%
- Real income growth ∆y_{i,t} when HD_{i,t−4} < HD*:
  - Debt/GDP: 0.17*** (0.05)
  - Debt/disposable income: 0.13** (0.06)
- Real income growth ∆y_{i,t} when HD_{i,t−4} ≥ HD*:
  - Debt/GDP: -0.02 (0.07)
  - Debt/disposable income: 0.03 (0.09)
- Other controls: Lending rate (lagged): -2.29 (2.04); -1.60 (2.27)
  - Real house price growth (lagged): 0.05 (0.04); 0.06 (0.04)
  - Real stock price growth (lagged): -0.02 (0.04); -0.01 (0.04)
  - Real public consumption growth (lagged): -0.01 (0.21); 0.02 (0.22)
  - Real fixed asset investment growth (lagged): 0.02 (0.01); 0.02 (0.01)
- Constants: 12.73*** (1.23); 12.79*** (1.30)
- Sample sizes:
  - # obs.: 288; 288
  - # groups: 24; 24

### Improving the Allocation of Corporate Credit — Key Findings
- China reduced corporate debt-to-GDP from a peak of 142 percent of GDP in 2016 to 129 percent of GDP in 2018 through de-risking measures.
- SOE leverage ratios have fallen; the decline in industrial SOE leverage in 2018 was driven more by asset increases (about 7 percent y-o-y) than by debt reduction (about 6 percent y-o-y).
- Financial regulatory tightening constrained credit to private corporates; SMEs were particularly affected by the contraction in shadow banking.
- Despite deleveraging:
  - A majority of bank loans still flows to SOEs.
  - Banks still provide over 70 percent of corporate financing.
  - Corporate bond and equity issuance has increased but remains a small portion of total liabilities.
- Adjusted SOE return on equity falls from an average of 8 percent to about −1.3 percent during 2011–15 when accounting for implicit support (Lam and Schipke, 2017).
- SOEs are estimated to have credit ratings about two to three notches higher than comparable POEs; SOEs pay an estimated 150- 200 bp lower interest rates on bonds, and over 100 basis points less after controlling for several factors (GavekalDragonomics, August 2018; Zhang and Wu, 2019).

### Policy Recommendations — Applying Competitive Neutrality to Financing
- Principles and first steps:
  - Implement competitive neutrality to treat enterprises under all forms of ownership on an equal footing.
  - Release a publicly available list of SOEs with assigned categories (social, strategic, competitive) to separate SOEs that should compete with POEs.
  - Recognize and realign the distinctive roles of government and market; require SOEs to make a commercial rate of return.
- Debt neutrality and regulatory neutrality actions:
  - Remove implicit guarantees given to SOEs.
  - Increase banks’ risk weights on corporate loans with implicit guarantees.
  - Adjust cost advantages provided to SOEs and harden SOE budget constraints.
  - Encourage equity financing and lower credit rating thresholds to allow small firms to issue bonds.
  - Strengthen credit ratings, credit registries, bank capitalization, and promote risk-based rather than collateral-based lending.
- Sequencing and implementation to dismantle implicit guarantees:
  - Ensure appropriate conditions: accept more defaults, rationalize and phase out implicit subsidies, and implement legal and institutional insolvency frameworks for exiting zombie firms and restructuring viable firms.
  - Anticipate fiscal and financial spillovers: local government fiscal gaps from exiting local SOE zombies; potential repricing of risk and disruptive withdrawals in financial markets.
  - Reinforce financial sector resilience: increase risk weights on loans to corporates with implicit guarantees, build liquidity buffers, strengthen oversight, and reduce reliance on short-term funding.
- Specific reform instruments and measures:
  - System-wide plan for restructuring and exit of local zombies including viability assessments, proactive bank recognition and workouts of NPLs, burden sharing, and developing distressed debt markets.
  - Market-based debt-equity swaps with strict solvency criteria, proactive bank roles as equity holders, limits on bank ownership duration, and conversion at fair value with recognition of losses.
  - Consider “TEMASEK-style” reforms (governing body holding shares and requiring commercial operation).
  - Legal reform: amend the Enterprise Bankruptcy Law to clarify scope, conditions for bankruptcy, and bankruptcy procedures; ensure tax neutral treatment for insolvency and debt restructuring; enhance judiciary capacity.
  - Tolerance of default events to strengthen market discipline and enable proper pricing of credit risks.
  - Dividend payout policy: central SOEs directed to increase dividend transfer to fiscal budget to 30 percent of profits by 2020 (dividends as share of profits declined to around 7 percent in 2018).
  - Continue rationalization of subsidies; SOEs still receive relatively more subsidies than POEs and a quarter of SOEs in 2015 remained loss making despite subsidies.
- Implementation priorities:
  - Sequence reforms carefully to manage fiscal, financial, and social impacts.
  - Ensure reforms are market-oriented, durable, and not circumvented.
  - Expand competitive neutrality coverage to enhance support for POEs and SMEs and improve overall credit allocation.

*Source: Annex II. Panel Regression Results; "IMPROVING THE ALLOCATION OF CORPORATE CREDIT IN CHINA" — extracted content from the supplied PDF content unit.*

### 16.      With SOEs receiving preferential treatment in credit/debt markets, other sources of

### 16.      With SOEs receiving preferential treatment in credit/debt markets, other sources of financing such as equity could help level the playing field.

### Improving access to equity financing
- Equity financing can help level the playing field because POEs rely more on equity financing.
- Policy proposal: an Allowance for Corporate Equity (ACE) providing an income tax deduction for a “normal” rate of return on equity.
  - ACE has been used by Belgium (2006), Italy (2012), Cyprus (2015), Portugal (2010-13), and Turkey (2015) to minimize the debt bias.
- Implementation caveat: ACE should be accompanied by measures to mitigate the impact on tax collection for the government.

### Paying the cost of implicit guarantees and debt neutrality
- Debt neutrality principle: all corporates should be given credit on the same terms; cost advantages received by any corporate should be adjusted.
- Transparency requirement: disclosure and remuneration of state guarantees should be transparent to ensure corporates adjust cost advantages.
- Possible approaches in China:
  - Require corporates that receive implicit advantages to explicitly pay the cost to the budget (in addition to targeted transfer of SOE profits to the fiscal budget), though estimating implicit guarantees is difficult when they are pervasive.
  - In the absence of corporates paying, increase banks’ risk weights on corporate loans that receive implicit guarantees to correct underpricing to banks.
  - Penalize SOEs that fail to meet a dividend payout target of 30 percent of profits (examples: reduce access to loans or increase their credit spreads).

- International examples where SOEs pay cost of guarantees to budget (Box 2):
  - Australia: adjust cost advantages through a debt neutrality payment to the Office of Public Accounts.
  - New Zealand: SOE loan documentation must explicitly disclaim government guarantee.
  - Spain: payments to the Treasury account for costs associated with advantages from public ownership (e.g., debt, guarantees, safeguards).
  - Switzerland: SOEs (for example the postal service) pay dividend to the government as reimbursement for lower interest rate.
  - Turkey: the government levies a guarantee fee for government guaranteed loans.

### Regulatory neutrality and prudential tools
- Apply the same financial regulations between SOEs and POEs to level the financial playing field (OECD guiding principles summarized in Box 3).
- Prudential tools to level the playing field:
  - Apply the same regulatory policies to all corporates; bank exposure limits can be equally applied to SOEs and POEs, though large bank exposures to SOEs mean transition may take time.
  - Near-term option: put caps on bank lending to SOEs with large exposure and gradually increase the cap over time to the same limits that apply to POEs (large and related party limits).
  - Increase sectoral risk weights to address credit misallocation and make lending to overleveraged sectors costlier.
  - For SME access, prefer targeted policy options (see section D) over lowering risk weights for SMEs or expanding public credit guarantees.
- Transparency: large state-owned banks should disclose policies and practices in providing services to enterprises of different ownership types.

### Structural measures to enhance access to finance by private enterprises
- Reduce barriers to entry:
  - Give all corporates free entry into all sectors; authorities announced opening financial sector, elderly care, education, and health care to private sector — extend to state-dominated services (logistics, telecommunications).
  - Break up administrative monopolies; ensure under the Company Law no government entity may use industrial policies or regulations to restrict POE access.
- Elimination of targeted lending:
  - Authorities announced quantitative targets to allocate credit towards the private sector:
    - Increase loans to SMEs by large state-owned banks by 30 percent.
    - Use multiple policy tools to lower the financing cost of SMEs by 1 percentage point.
  - “1-2-5” policy goal: at least 1/3 of new corporate loans from large banks to be extended to private firms; at least 2/3 of new loans to be extended from small and medium size banks; and at least 50 percent of all new corporate credit across the banking system to be extended to the private sector over the next three years.
  - Credit should be channeled market-oriented towards highly productive industries, avoiding moral suasion to reach allocated targets.
  - For SMEs:
    - Use state-owned/development banks to expand credit to SMEs based on clear mandates, sound governance, clear performance criteria, risk-based loan pricing, and qualified staff.
    - Develop specific capital markets targeted at SMEs (example: KONEX securities exchange platform in Korea).
- Lowering ratings threshold for bond issuance:
  - All credit bonds (excluding private placement) require minimum ratings such as AA or A–, or in some cases AAA.
  - Regulators can lower the regulatory threshold to encourage issuance by smaller private firms, while allowing bonds to default and investors to bear risks.
- Harmonizing credit bond scheme regulations:
  - China’s credit bond market is segmented with different regulators, issuance procedures, trading platforms, and depositories.
  - Harmonizing regulation would reduce segmentation and regulatory arbitrage, increase liquidity, and foster price discovery.
- Uniform application of laws:
  - The Commercial Bank Law does not differentiate corporates by ownership, but bank officers tend to be reluctant to lend to POEs due to perceived higher default probability.
  - Reform suggestion: protect bank staff who exercised proper due diligence by dropping “lifelong accountability” for loan defaults by private enterprises if proper due diligence was followed.

### Creating a conducive environment for lending and strengthening credit culture
- Credit culture beyond collateral:
  - SMEs often lack fixed asset collateral; policies to expand credit access include:
    - Foster use of movable collateral (machinery, accounts receivables) and leasing and factoring through NBFIs for SME financing; may require amending or formulating a new law on security interest.
    - Expand risk-based lending to SMEs based on potential profitability rather than collateral.
    - Improve credit reporting mechanisms such as credit bureaus (CB) and public credit registries; consider specialized CB for “small enterprises” leveraging alternative data (credit card sales slip, telephone/electricity bills, online shopping, commercial transactions, fintech lenders).
- More accurate credit ratings:
  - Over 90 percent of onshore bonds are rated AA to AAA by local rating agencies in China.
  - Causes: stringent regulatory threshold for issuance and a nascent rating industry.
  - Recent approval of S&P Global Inc. to provide credit-rating services is noted.
  - Improve domestic rating agencies and encourage more foreign ratings agencies to strengthen the sector.
- Credit bureau/registry improvements:
  - China has one public credit registry covering banking sector credit history.
  - Licenses granted to eight institutions to establish credit bureaus focusing on new financing types (P2P, etc.), but progress is limited.
  - Agreement to share key data between public and private credit bureaus/registry can ensure consistency and wider coverage.
  - Strengthen supervision and regulation on credit registry agencies (CRAs) by:
    - Enhancing disclosure to improve comparability of ratings among CRAs.
    - Strengthening assessment of CRAs focusing on adequacy/consistency of rating methodologies and timeliness of rating adjustment.
    - Requiring enhanced internal control to avoid conflicts of interest and strengthening enforcement actions by supervisors.
- Data quality and availability:
  - Accurate, high-quality data is needed to guide proper credit allocation and policies.
  - Currently limited: breakdown on credit by corporate ownership and information on subsidies and guarantees.

### Policy tools and recent initiatives (selected key figures)
- Dividend payout target for SOEs: 30 percent of profits.
- “1-2-5” policy goal details:
  - at least 1/3 of new corporate loans from large banks to private firms;
  - at least 2/3 of new loans to be from small and medium size banks;
  - at least 50 percent of all new corporate credit across the banking system to private sector over the next three years.
- RRR cuts and monetary support (Annex I highlights):
  - RRR cuts: 250 bps of RRR cuts in 2018 (April, July, and October); Effective May 15, 2019 RRR cut for small and medium-sized banks up to 3.5 ppt to 8 percent.
  - In June 2018, PBC expanded MLF-eligible collateral to include AA+ and AA corporate bond, with priority on small and micro enterprise, green, and agricultural bonds.
  - In November 2018, PBC announced a subsidy of RMB 10 bn to China Bond Insurance to support debt sales by private enterprises.
  - Regulatory actions: CBIRC announced the “1-2-5” policy goal in November 2018.
  - March 2019 announcement: increase loans to SMEs by large state-owned banks by 30 percent and use multiple policy tools to lower financing cost of SMEs by 1 percentage point.
  - SASAC target: reduce central SOEs’ liability-to-asset ratio by 2 pp in 2018-20.
  - Competitive neutrality initiatives: March 2019 NPC commitment to equal treatment; May 2019 State Council regulation effective July 1st 2019 on equal treatment for government investment fund allocation.

### Concluding recommendations and principles
- Continue measures to reduce corporate leverage with focus on improving credit allocation to POEs, especially SMEs.
- Pursue high-quality reforms sequentially under the mandate of competitive neutrality.
- Establish principles of debt and regulatory neutrality to eliminate market distortions and promote efficient credit allocation.
- Use proactive, market-friendly measures to support SMEs and ensure policies are market-oriented to achieve efficient resource allocation.

*Source: IMF chapter content (People’s Republic of China).*

### 2.      In the event of a new deal that involves China purchasing U.S. goods to reduce the

### 1chnea2019004 - 2.      In the event of a new deal that involves China purchasing U.S. goods to reduce the

### Key findings
- If China closes the U.S.—China bilateral trade deficit (USD 336bn in 2017) by purchasing U.S. goods while China’s total imports do not change, the extra purchases from the U.S. will be at the expense of cutting purchases from the rest of the world, generating large trade diversion effects.
- Estimated “exports-at-risk” (main scenario, sample aggregates):
  - USD 61bn from European Union countries included in the sample (exposure in vehicles, machinery, and aircraft).
  - USD 54bn from Japan (machinery, vehicles, and electronics).
  - USD 46bn from Korea (electronics, opticals, plastics).
  - USD 45bn from ASEAN countries (electronics, plastics, machinery).
- Country-level exposures (exports-at-risk as a share of 2017 GDP):
  - Higher than 3 percent of GDP: Oman, Angola, Singapore, and Korea.
  - Higher than 2 percent of GDP: Malaysia, Vietnam, Thailand (some ASEAN economies).
  - Around 1 percent of GDP: Germany and Japan.
  - Brazil: 1.3 percent of GDP (predominantly due to oil seeds, e.g. soya beans).
- Energy and agricultural exporters (oil exporters and soya bean producers) face substantial exports-at-risk in dollar terms and especially as a share of GDP for smaller oil exporters.

### Methodology and assumptions
- Scenario: China eliminates the bilateral deficit with the U.S. by increasing purchases of U.S. goods; China’s total imports remain unchanged (perfect substitution from rest of world to U.S.).
- Product focus: Top-ten products imported from the U.S. are considered; trading relationships are assumed difficult to change overnight, so China scales up where it already imports from the U.S.
- Allocation rules for extra purchases:
  - Allocation across the top-ten products is proportional to where there is more scope for U.S. catch-up and China’s capacity constraints.
  - Scale of increases is capped so total purchase of each product does not exceed China’s total imports of that product.
  - For most products, the allocation of purchases does not exceed total U.S. exports.
- Distribution of extra purchases across top-ten products (main scenario, share of total extra purchase, In percent):
  - Electronics, 13
  - Machinery, 12
  - Vehicles, 11
  - Oil seeds, 8
  - Aircraft, 3
  - Optical, 11
  - Mineral fuels/oils, 13
  - Plastics, 12
  - Pearls, 11
  - Wood pulp, 4
- Identification of affected countries: For each product, the loss in exports of other countries is proportional to their share in China’s imports of that product; top-ten non-U.S. exporters per product were considered in the main analysis.

### Alternative scenarios and robustness checks
- Two alternative allocation scenarios were considered:
  1. Purchases scaled up for a subset of the ten products (major items under discussion: soya beans, liquefied natural gas (LNG) and oil, manufactured items). Result: more pronounced exports-at-risk for Germany, Japan, and ASEAN economies; oil exporters show higher exports-at-risk as a share of GDP.
  2. Extra purchases distributed in proportion to China’s import structure (excluding the U.S). Result: ASEAN economies and Korea have higher exports-at-risk due to exposure on electronics; small oil exporters remain highly exposed as a share of GDP.
- Robustness using HS 6-digit data and an expanded sample (~200 economies):
  - Of 39 economies reported, differences in exports-at-risk (as a share of GDP) are within 1 percentage point for 35 economies and within 0.1 percentage points for 8 economies.
  - Example: Germany remains at 1 percent of GDP in both 2-digit and 6-digit analyses; France: 0.3 (2-digit) versus 0.2 (6-digit) percent of GDP.
  - Aggregate USD totals can differ due to inclusion of more countries in the robustness sample.

### Sectoral and supply-chain considerations
- Product exposures determine which countries are affected:
  - Increased purchases of oil seeds (soya beans) affect Brazil, Canada, and Argentina.
  - Increased purchases of vehicles affect Germany, Japan, and the U.K.
  - Increased purchases of aircraft affect France and Germany.
  - Increased purchases of machinery affect Japan, Germany, and Korea.
- China’s gross exports to the U.S. include substantial value added from other countries:
  - In 2015, China’s gross exports to the U.S. included 82.3 percent of China’s domestic value added (up from 76.7 percent in 2008); the remainder of value added came from Germany, Japan, Taiwan Province of China, Korea, and U.S.
  - The U.S. bilateral trade balance with China in 2015 was 13 percent lower in value added compared to gross terms (OECD TiVa database, 2015).
- The exports-at-risk metric is not equivalent to potential value-added losses; actual value-added impact depends on ease of switching to other markets and supply-chain rigidities.

### Escalated trade tensions and global implications
- Already implemented tariffs are holding back global trade and slowing industrial production; additional uncertainty weighs on investment and business sentiment.
- Simulated global GDP impacts:
  - The U.S.-China tariff increases in 2018 were projected to lower global GDP by 0.2 percent in 2020.
  - Additional tariffs announced by U.S. and China in May 2019 could cause global GDP to decline by an additional 0.3 percent in 2020 (on top of the 0.2 percent).
  - Simulation assumptions include: (i) increase in tariffs from 10 percent to 25 percent on USD 200 billion of U.S. imports from China (as of May 8, 2019), (ii) possible 25 percent tariffs on roughly USD 267 billion of U.S. imports from China (envisaged in May 2019), with assumed retaliatory actions by China.
- Sectoral spillovers can be large as global value chains are repositioned; multi-sector CGE model evidence suggests significant sectoral disruptions and large positive and negative spillovers to highly exposed by-stander economies.
- China’s exports in affected industries (example value-added origins, expressed as a share of China’s gross exports to the U.S. in the relevant industry, 2015):
  - Electrical Equipment: CHN, 81.2; USA, 2.0; JPN, 2.0; KOR, 1.9; AUS, 1.4; TWN, 1.2; ROW, 10.3.
  - Vehicles*: CHN, 83.7; USA, 2.1; JPN, 1.9; DEU, 1.6; KOR, 1.4; AUS, 0.8; ROW, 8.5.
  - (*Motor vehicles, trailers, and semi-trailers.)

### Policy implications and recommendations
- Targeting bilateral trade balances will not by itself reduce a country’s overall current account deficit; changes in current account balances are best achieved through macroeconomic policies that influence saving/investment decisions.
- Tariff increases are unlikely to significantly change the trade balance because lower imports tend to be offset by currency appreciation.
- An agreement between the U.S. and China should be:
  - Comprehensive and quickly reached to limit further global disruption.
  - Reinforcing WTO rules, non-discriminatory, and based on market mechanisms and macroeconomic fundamentals.
- Avoid managed trade and aim to resolve disputes in ways that minimize trade diversion and GVC disruption.

*Source: IMF staff analysis as presented in the provided chapter/section.*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2019/1chnea2019004.pdf_