## 1chnea2021002

## Source details

**Canonical URL:** [1chnea2021002](https://www.imf.org/-/media/files/publications/cr/2021/english/1chnea2021002.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2021/english/1chnea2021002.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2021/english/1chnea2021002.pdf.json)

---

### FURTHER STEPS TO IMPROVE MONETARY POLICY EFFECTIVENESS AND CREDIT ALLOCATION

- Reform progress and remaining scope
  - The reform of China’s monetary policy framework has progressed, but there remains important scope for further improvement.
  - Interest rate guidance policies continue to influence pricing of bank deposits and loans, skewing risk-adjusted returns in favor of low-risk firms and limiting pass-through of policy interest rates to bank funding costs.
  - Reforming interest rate guidance policies is argued to: strengthen the recovery by boosting market-based credit allocation to riskier firms, increase effectiveness of interest-rate based monetary policy, and raise household income.

- Timeline and instruments
  - 2013: PBC officially phased out bank lending rate floor.
  - 2015: PBC phased out deposit rate ceiling; informal adoption of 7-day repo as short-term policy rate; self-regulatory mechanism for deposit pricing introduced.
  - 2019: PBC introduced the Loan Prime Rate (LPR) system referencing the 1-year Medium-term Lending Facility (MLF) rate plus bank spreads.
  - Macroprudential Assessment (MPA) mechanism incorporated “interest rate pricing behavior” in 2015.
  - Since 2018, the State Council introduced targeted credit policies with lending rate requirements for MSEs, often targeting spreads of only 100 to 200 basis points above prime customer rates.

- Deposit pricing and lending guidance
  - After 2015, a self-regulatory mechanism for deposit pricing maintained an effective ceiling linked to benchmark deposit rates; transmission from short-term policy rates to bank deposit rates has been notably limited.
  - Targeted credit policies require medium and large banks to target average lending rates to MSEs at spreads only 100 to 200 basis points above rates for prime customers.

### A. Interest rate guidance and credit allocation — empirical observations and mechanisms

- Observed loan-rate patterns and statistics
  - Most Chinese bank loan rates are within 100-200 basis points of 3-year AAA-rated corporate bond yields.
  - Only about one quarter of Chinese bank loans have interest rates more than 250 points above the 3-year AAA-rated corporate bond yield.
  - 27 percent of loans have spreads above 250 basis points (June 2020 estimate).

- Causes and effects on allocation
  - Deposit rate guidance keeps bank funding costs unusually low relative to other funding costs, making it profitable to lend to low-risk borrowers who might otherwise use bond market funding.
  - The average spread between investment-grade bond yields and bank funding costs in China is said to be double that in major banking systems.
  - Low bank loan rates may reflect: deposit guidance, lending rate guidance, banks’ preference for collateral, penalties for underwriting nonperforming loans, and limited underwriting/distribution capacity for small business loans.
  - Credit policies mandating lending to high-risk borrowers can prompt banks to tighten non-rate lending terms (collateral, fees, maturities) or to target the highest-quality borrowers eligible for credit policies, potentially leading to non-productive uses of credit (e.g., real estate speculation).
  - Continued decline in the share of corporate bonds issued by privately owned enterprises and low-rated firms suggests targeted lending may not broadly ease financial conditions for riskier firms.

- COVID-19 effects
  - Authorities increased use of lending rate guidance during COVID, expanding targeted credit policies with explicit lending rate requirements and imposing similar requirements on central bank re-lending funding.
  - COVID-19 lockdowns disproportionately affected small businesses, increasing default risks and reducing risk-adjusted returns of loans to these firms more than for larger firms, worsening banks’ expected profitability on such lending.

### B. Monetary policy transmission, deposit stickiness, and instrument efficacy

- Deposit rate stickiness and funding-cost dynamics
  - Average bank deposit rates were largely unchanged from end-2019 to mid-2020, despite reductions in policy interest rates.
  - Banks representing over 80 percent of deposits reported small increases in deposit funding costs during this period due to deposit composition shifts.
  - Deposit regulation constrains banks from raising rates to compete with higher-yielding alternatives, resulting in persistent deposit outflows and reluctance, especially of smaller banks, to lower deposit rates.
  - Banks’ issuance of deposit alternative products surged and yields on such funding rose, limiting pass-through of short-term policy interest rates to yields on deposit alternatives.

- Implications for PBC interest-rate tools
  - In the context of sticky deposit rates, cuts to the MLF rate that lower the LPR would immediately shrink bank lending margins, reducing banks’ ability to sustain credit growth and build buffers.
  - Uniform reductions in bank loan rates via the LPR mechanism would weaken risk-adjusted returns of loans to risky firms, increasing lenders’ preference for low-risk firms.
  - A 100-basis point increase in bank funding cost shifts the loan-rate distribution upward by 100 basis points relative to corporate bond yields in the stylized example.

### C. Advantages of phasing out interest rate guidance and quantitative scenario

- Expected reallocation effects if guidance is eased
  - Easing deposit rate guidance could increase bank funding costs initially, squeezing lending margins on loans to low-risk corporates and prompting those firms to shift to the bond market.
  - If lending rate guidance were also phased out, banks could increase lending to smaller, riskier firms requiring specialized credit assessment and monitoring.
  - Freeing deposit rate restrictions would increase household incomes and spending power, particularly among lower-income segments with savings primarily in bank deposits, and reduce precautionary savings and investment-related housing demand.
  - Short-term policy rates would have greater impact on bank deposit rates and other money market yields, improving monetary policy transmission.

- Stylized quantitative illustration
  - Assumptions:
    - Ending deposit rate guidance increases bank funding costs.
    - Investment grade corporate bond yields assumed unchanged.
    - Competitive credit markets allow low-risk firms to shift between bank and bond finance and loans priced to reflect credit risk.
  - Results:
    - Low-risk firms migrate to the bond market; medium-risk firms see no change in bank credit volumes.
    - In the stylized example, the share of lending to high-risk firms—proxied by lending with interest rates at least 250 basis points above AAA-rated bond yields—increases from 27 percent to 45 percent.

- Transition risks and complementary measures
  - Transition may cause temporary frictions: firms’ limited ability to shift to bond finance, temporary rise in bond yields given banks’ investor role in government bonds.
  - The PBC may need additional monetary easing to offset temporary borrowing-cost increases.
  - Clear communication is critical to limit market volatility.
  - Macro-financial stability measures required:
    - Stronger prudential regulation and supervision.
    - Sound governance and credit risk management.
    - Reforms to ensure competitive neutrality between private firms and SOEs.
    - Consideration of measures to support SME access and borrowing costs during phasing out of credit policies, e.g., a public credit guarantee scheme.

### CHINESE STATE-OWNED ENTERPRISES, RESOURCE (MIS)ALLOCATION, AND PRODUCTIVITY — key findings and policy recommendations

- Size, scope, and roles of SOEs
  - In 2018, total assets of Chinese SOEs stood at 194 percent of GDP.
  - SOEs operate in all sectors of the economy and have played roles in supporting the economy and employment during recessions, and provide health and pension services.
  - Since the 1990s the number of SOEs has declined significantly (a ⅔ decline among the industrial firms alone), along with the share of urban workers in SOEs.
  - SOEs likely benefit from implicit government guarantees, enabling access to a higher share of bank financing and lower interest rates on liabilities.

- Productivity gaps between listed SOEs and POEs (data and magnitudes)
  - Data: Wind database of listed firms covering over 3700 listed firms in the Shenzhen and Shanghai stock exchanges between 2002 and 2019.
  - Listed firms account for 6 percent of GDP and 10 percent of manufacturing value-added in 2019.
  - In 2019, SOEs accounted for 29 percent of listed firms and 57 percent of listed firm value-added.
  - The typical SOE employs more than twice as many workers as the typical POE.
  - Revenue productivity: the gap widened to 30 percent during the Global Financial Crisis (GFC) and reduced to 20 percent as of 2019.
  - Value-added per unit of fixed assets is almost 40 percent lower for SOEs than POEs; value-added per unit of labor of SOEs is similar to that of POEs.
  - Sectors with larger revenue productivity gaps (Utilities, Transportation and Steel Manufacturing) are also those where SOEs account for a larger share of output and inputs.

- Quantified impact of misallocation and macro gains
  - Within-listed-firms results (Hsieh & Klenow, 2009 model):
    - Reducing the average SOE-POE revenue productivity gap in every sector could increase aggregate productivity among listed firms by between 5 and 6 percent (gains reported for 2019).
    - A policy that also reduces distorted capital-intensity of SOEs could yield gains of over 6 percent.
  - Extrapolation to the whole economy:
    - Estimates for listed firms are scaled down to around 4 percent for the whole economy (based on SOE share of assets for the whole economy being smaller than for listed firms).
    - Given aggregate TFP growth averaged 0.6 percent between 2012 and 2017 (Penn World Tables), SOE reform could more than double the rate of TFP growth for five years—or likely more if sectoral reallocation were considered.
  - Simulations on credit reallocation:
    - Reallocating credit from highly-leveraged SOEs to POEs can increase aggregate investment and boost growth by 0.3-0.4 ppts annually when highly-indebted SOEs deleverage by 2 ppts and the freed-up credit is channeled to POEs.

- Role of credit misallocation and leverage
  - Negative correlation between firm productivity and the leverage ratio (debt-to-asset ratio) indicates credit is allocated to the least efficient firms.
  - On average SOEs have higher leverage ratios than POEs.
  - The leverage ratio of low-productivity SOEs is still more than 5 percentage points higher than that of high-productivity POEs.
  - Within SOEs, the leverage ratio of high-productivity SOEs was 10 percentage points lower than the low-productivity SOEs at the end of 2016 (gap has narrowed since).
  - Threshold effect model (Hansen, 2002) estimates:
    - The impact of a 1-ppt increase in a firm’s leverage ratio on investment is 0.1 ppt; this impact falls to 0.01 ppt if the leverage ratio exceeds a threshold of about 35 percent.

- Policy recommendations to reform SOEs and related financial-sector measures
  - Identify non-viable SOEs and open non-strategic sectors to private/foreign competition; allow non-viable firms to default and exit rather than being merged with more profitable SOEs.
  - Reform the social safety net to relieve SOEs of social functions (employment stabilization, health and pension benefits), transferring these obligations to the state.
  - Ensure equal access to credit and capital for private firms and allow SOEs to deleverage:
    - Recognize and remove implicit government guarantees that enable SOEs to access financing at lower rates.
    - Prepare the financial sector for removal of implicit guarantees by requiring banks to carry higher risk weights on SOE loans, build liquidity buffers, reduce reliance on short-term funding, and increase capital.
    - Promote market-based policies to allow credit to flow to its most productive use, establish competitive neutrality among firms, strengthen credit culture, improve credit ratings, strengthen credit registries, ensure adequate bank capitalization, and promote more risk-based vs. collateral-based lending.
  - Improve SOE governance:
    - Increase transparency of SOE group structures and activities.
    - Allow appointment of company managers with international/private sector experience.
    - Clarify the role of the Party in decision making.
    - Encourage corporate boards where missing, and reduce the practice of exchanging management staff between supervisors (SASAC) and central SOEs.

### 1. China’s monetary policy framework — empirical identification and main results

- Institutional context and framework features
  - Interest rate liberalization largely complete (removal of the ceiling on deposit rates in 2015); development of an interest rate corridor.
  - State Council (SC) is the decision-making body and the People’s Bank of China (PBC) does not have full operational independence.
  - The previous official intermediate target, M2, was de-emphasized but has not been replaced with a new intermediate target.
  - The MLF rate became the main instrument to influence bank lending rates since mid-2019.

- Sample and shock measurement
  - SC text releases sample: May 2013 to April 2020.
  - Daily market shock measure: daily close-to-close change in the rate on one-year interest rate swaps (IRS) based on the interbank 7-day repo rate, measured around policy announcements.
  - Summary statistics for the daily measure (Mean |ΔIRS| (bps) / Std Dev ΔIRS (%)):
    - RRR 7.3 11.9
    - PBC 7-day reverse repo rate 10.9 15.6
    - Benchmark lending and deposit rate 5.5 6.8
    - MLF rate 4.3 3.2
    - All policy events 6.9 10.4
    - Non-event days 3.1 5.0
  - The standard deviation of the IRS rates on days without policy events is 5 basis points, and higher on policy announcement days, at 10 basis points.

- Key empirical findings — same-day and short-run effects
  - Monetary policy shocks have immediate impacts on sovereign bond yields and state-owned enterprise bond spreads, but less impact on corporate bond spreads and other credit bond spreads.
  - Same-day estimates:
    - A shock accompanied by a 100 bps decrease in the IRS rate is associated with about a 20-30 bps decrease in sovereign yields.
    - The same shock is associated with about a 15-20 bps decrease in enterprise bond spreads (significant at 1- to 10-year maturities).
  - Cumulative impulse responses to a 1 percentage point expansionary monetary policy shock:
    - Interbank market: 1 percentage point shock leads to a 35 bps decrease in the 3-month SHIBOR on impact.
    - Sovereign yields: 0.7 percentage point decrease in the sovereign yield after 3 days (effect fades and is not statistically significant after one week).
    - Enterprise bond spreads: 1.3 percentage point decrease at the 3-day peak response.
    - Corporate bond spreads: 1.7 percentage point decrease (timing as reported).

- Role of monetary-fiscal coordination
  - The impact of monetary policy is significantly stronger when coordinated with fiscal policy.
  - Coordinated shocks generally produce larger effects than uncoordinated shocks.
  - Example magnitude: coordinated shocks reduce enterprise spreads by 3.3 percentage points 20 days after the shock.

- Interpretation, limitations, and policy implications
  - Collective decision-making and close monetary-fiscal coordination explain why monetary and fiscal policy often work in tandem to counter shocks.
  - A muted market response to uncoordinated monetary policy weakens the central bank’s ability to use interest-rate channels to affect the economy.
  - Policy recommendations include:
    - Formulating a clear inflation objective and granting the PBC operational (instrument) independence; the State Council should set overall goals but delegate responsibility to the PBC.
    - Streamlining and clarifying the monetary policy framework with a focus on one key policy interest rate; PBC should guide the short-term interbank rate clearly and let longer-term rates be market-determined.
    - Reforms to further improve interest rate pass-through, including continued progress on LPR reform and phasing out the benchmark deposit rate.
    - Steps to increase the financial robustness of the banking system to fluctuations in short-term interest rates, including (i) raising bank capital, and (ii) further developing interest-rate hedging instruments.

### Intergovernmental fiscal relations, provincial risk-sharing, and counterfactual transfers

- Pre-pandemic fiscal structure and pandemic impact
  - Most provinces rely on transfers from the central government (CG), net tax refunds, LG bond issuance within CG-allocated quotas, dividend transfers from local state-owned enterprises (LSOEs), withdrawals of LG deposits and land sales to fill fiscal gaps.
  - CG measures in 2020:
    - Increased ceiling on special LG bond issuance by 74 percent (or RMB1.6 trillion).
    - Issued RMB1 trillion anti-epidemic special Treasury bonds with proceeds allocated to LGs.
  - Provincial outcomes (January–April 2020):
    - LG fiscal expenditure declined in 24 out of 31 provinces by an average of 4 percent (y/y).
    - LG own fiscal revenue declined in 30 provinces by an average of 12 percent.

- Empirical strategy and key estimates
  - Methods: two-step estimation (Von Hagen (1992); Bayoumi and Masson (1995)) and one-step pooled mean group (PMG) estimator (Pesaran et al., 2016); data span 2006 to 2018.
  - Estimated systematic fiscal risk-sharing effect:
    - Coefficients: 0.87 (two-step) and 0.82 (one-step); average 0.85.
    - Interpretation: Disposable income of a province falls by about 85 cents in response to a temporary 1-RMB decline in aggregate income relative to the national average; the remaining 15 cents are smoothed by CG transfers. Reported risk-sharing effect of 15 percent on average.
  - Estimated redistribution effect:
    - Coefficient for redistribution ≈ 0.67 (statistically significant); magnitude of redistribution effect about 33 percent.

- Time and regional variation
  - Risk-sharing effect increased over time: from 14 percent during 2006-12 to 21 percent during 2012-18.
  - Regional differences: inland provinces ≈ 20 percent risk-sharing; coastal provinces ≈ 12 percent risk-sharing.

- Counterfactual rule-based transfers
  - Rule transfers calibrated so total alternative transfers ≤ total actual transfers over 2006-18; when calibration parameter τ is above a threshold (around 4), all idiosyncratic shocks could be smoothed.
  - Rule-based transfers could deliver significantly higher risk-sharing than realized ~15 percent while keeping redistribution similar to realized ~33 percent.

- Policy recommendations for intergovernmental reform
  - Reduce misalignment of central-local fiscal responsibilities (expenditure assignment based on economies of scale, equity, externalities; consider centralized pension and unemployment insurance systems).
  - Tax reforms to give LGs more authority over some tax rates and bases (e.g., PIT and property tax) to strengthen LG fiscal accountability.
  - Align LG borrowing limits with expenditure responsibilities and contain off-budget investment; ensure realistic LG financing arrangements; carefully assess and dismantle implicit guarantees on off-budget investments (e.g., allow defaults); bring non-commercial investment on budget.
  - Consider alternative automatic and non-regressive CG transfer mechanisms to increase fiscal risk-sharing while keeping redistribution roughly similar; a transparent rule would signal access to resources and help prevent sharp tightening of financial conditions.
  - Common borrowing (CG-issued bonds) as second-best when CG funding constrained: reduces degree of fiscal risk-sharing relative to transfers but can deliver more targeted support and break cycle of weaker LG health → higher borrowing costs under LG self-insurance.
  - Reduce LG-LSOE-bank interlinkages to contain moral hazard from enhanced fiscal risk-sharing: remove explicit and implicit guarantees for LSOEs; strengthen governance rules to prevent LGs using transfers to protect weak LSOEs.
  - Consider countercyclical rule for LSOE dividend transfers (e.g., discretionary increases in transfer ratios during downturns) to improve LG shock-smoothing capacity.
  - Increase interregional mobility of production factors by phasing out local protectionism and LG interventions in product and factor markets.

*Source: IMF staff compilation from the chapter "1. China’s monetary policy framework" and associated sections (PDF content unit: 1chnea2021002).*

### References ____________________________________________________________________________ 16

### References ____________________________________________________________________________ 16

### FURTHER STEPS TO IMPROVE MONETARY POLICY EFFECTIVENESS AND CREDIT ALLOCATION

- Reform progress and remaining scope
  - The reform of China’s monetary policy framework has progressed, but there remains important scope for further improvement.
  - Interest rate guidance policies continue to influence pricing of bank deposits and loans, skewing risk-adjusted returns in favor of low-risk firms and limiting pass-through of policy interest rates to bank funding costs.
  - Reforming interest rate guidance policies is argued to: strengthen the recovery by boosting market-based credit allocation to riskier firms, increase effectiveness of interest-rate based monetary policy, and raise household income.

- Timeline and instruments described
  - 2013: PBC officially phased out bank lending rate floor.
  - 2015: PBC phased out deposit rate ceiling; informal adoption of 7-day repo as short-term policy rate; self-regulatory mechanism for deposit pricing introduced.
  - 2019: PBC introduced the Loan Prime Rate (LPR) system referencing the 1-year Medium-term Lending Facility (MLF) rate plus bank spreads.
  - Macroprudential Assessment (MPA) mechanism incorporated “interest rate pricing behavior” in 2015.
  - Since 2018, the State Council introduced targeted credit policies with lending rate requirements for MSEs, often targeting spreads of only 100 to 200 basis points above prime customer rates.

### A. The Role of Interest Rate Guidance Policies

- Deposit pricing mechanism
  - After 2015, a self-regulatory mechanism for deposit pricing maintained an effective ceiling linked to benchmark deposit rates, with flexibility for smaller banks and certain wholesale/structured deposits.
  - Bank disclosures show deposit costs have drifted slightly higher, largely owing to a shift into time deposits and other instruments that can offer higher yields, but remaining relatively close to benchmark deposit rates.
  - Transmission from short-term policy rates to bank deposit rates has been notably limited.

- Lending guidance
  - Authorities maintain a self-regulatory mechanism for bank loan rate pricing; use of lending rate guidance remains important in the policy toolkit.
  - Targeted credit policies require medium and large banks to target average lending rates to MSEs at spreads only 100 to 200 basis points above rates for prime customers.

### B. Interest Rate Guidance Policies and Credit Allocation

- Observed loan-rate patterns
  - Most Chinese bank loan rates are within 100-200 basis points of 3-year AAA-rated corporate bond yields.
  - Only about one quarter of Chinese bank loans have interest rates more than 250 points above the 3-year AAA-rated corporate bond yield.
  - 27 percent of loans have spreads above 250 basis points (June 2020 estimate).

- Causes and effects
  - Deposit rate guidance keeps bank funding costs unusually low relative to other funding costs, making it profitable to lend to low-risk borrowers who might otherwise use bond market funding.
  - The average spread between investment-grade bond yields and bank funding costs in China is said to be double that in major banking systems.
  - Low bank loan rates may reflect: deposit guidance, lending rate guidance, banks’ preference for collateral, penalties for underwriting nonperforming loans, and limited underwriting/distribution capacity for small business loans.
  - Credit policies mandating lending to high-risk borrowers can prompt banks to tighten non-rate lending terms (collateral, fees, maturities) or to target the highest-quality borrowers eligible for credit policies, potentially leading to non-productive uses of credit (e.g., real estate speculation).
  - Continued decline in the share of corporate bonds issued by privately owned enterprises and low-rated firms suggests targeted lending may not broadly ease financial conditions for riskier firms.

- COVID-19 effects
  - Authorities increased use of lending rate guidance during COVID, expanding targeted credit policies with explicit lending rate requirements and imposing similar requirements on central bank re-lending funding.
  - COVID-19 lockdowns disproportionately affected small businesses, increasing default risks and reducing risk-adjusted returns of loans to these firms more than for larger firms, worsening banks’ expected profitability on such lending.

### C. Interest Rate Guidance Policies and Monetary Policy Transmission

- Deposit rate stickiness during COVID-19
  - Average bank deposit rates were largely unchanged from end-2019 to mid-2020, despite reductions in policy interest rates.
  - Banks representing over 80 percent of deposits reported small increases in deposit funding costs during this period due to deposit composition shifts.
  - Deposit regulation constrains banks from raising rates to compete with higher-yielding alternatives, resulting in persistent deposit outflows and reluctance, especially of smaller banks, to lower deposit rates.

- Interference from credit growth policies
  - Policy-driven supply constraints on deposits were exacerbated by sharply increased bank demand for funding due to higher aggregate credit growth targets and heavy net issuance of government bonds.
  - Banks’ issuance of deposit alternative products surged and yields on such funding rose, limiting pass-through of short-term policy interest rates to yields on deposit alternatives.

- Implications for PBC interest-rate tools
  - In the context of sticky deposit rates, cuts to the MLF rate that lower the LPR would immediately shrink bank lending margins, reducing banks’ ability to sustain credit growth and build buffers.
  - Uniform reductions in bank loan rates via the LPR mechanism would weaken risk-adjusted returns of loans to risky firms, increasing lenders’ preference for low-risk firms.

### D. Advantages of Interest Rate Guidance Policy Reform

- Efficiency gains from phasing out guidance
  - Easing deposit rate guidance could increase bank funding costs, tightening financial conditions initially, but would squeeze lending margins on loans to low-risk corporates, prompting those firms to shift to bond market finance.
  - If lending rate guidance were also phased out, banks could increase lending to smaller, riskier firms requiring specialized credit assessment and monitoring.
  - Expected benefits include eased credit constraints for smaller, riskier firms and better risk monitoring compared to smaller-scale nonbank lenders.

- Household and monetary transmission benefits
  - Freeing deposit rate restrictions would increase household incomes and spending power, particularly among lower-income segments with savings primarily in bank deposits, and reduce precautionary savings and investment-related housing demand.
  - Short-term policy rates would have greater impact on bank deposit rates and other money market yields, improving monetary policy transmission.
  - Reduced need for banks to compete via deposit alternatives would limit incentives for financial and regulatory arbitrage.

- Stylized scenario illustrating impact (assumptions and conclusions)
  - Assumptions:
    - Ending deposit rate guidance is assumed to increase bank funding costs.
    - Investment grade corporate bond yields assumed unchanged.
    - Credit markets assumed competitive, allowing low-risk firms to shift between bank and bond finance and loans priced to reflect credit risk.
  - Mechanism:
    - Low-risk firms migrate to the bond market; medium-risk firms see no change in bank credit volumes.
    - Higher bank funding costs reduce profitability of lending at low interest rates, tightening supply to low-risk firms and prompting migration to bond issuance.
    - Banks increase supply of loans to high-risk firms as they can lend at higher rates and reduced lending to low-risk borrowers frees capacity.
  - Quantitative illustration:
    - A 100-basis point increase in bank funding cost shifts the loan-rate distribution upward by 100 basis points relative to corporate bond yields.
    - In this stylized example, the share of lending to high-risk firms—proxied by lending with interest rates at least 250 basis points above AAA-rated bond yields—increases from 27 percent to 45 percent.

- Transition risks and complementary measures
  - Transition may cause temporary frictions: firms’ limited ability to shift to bond finance, temporary rise in bond yields given banks’ investor role in government bonds.
  - The PBC may need additional monetary easing to offset temporary borrowing-cost increases.
  - Clear communication is critical to limit market volatility.
  - Macro-financial stability measures required:
    - Stronger prudential regulation and supervision.
    - Sound governance and credit risk management.
    - Reforms to ensure competitive neutrality between private firms and SOEs.
    - Consideration of measures to support SME access and borrowing costs during phasing out of credit policies, e.g., a public credit guarantee scheme.

### CHINESE STATE-OWNED ENTERPRISES, RESOURCE (MIS)ALLOCATION, AND PRODUCTIVITY — introduction excerpt

- Role of SOEs and potential gains from reform
  - SOEs play an outsized role in China; continuing SOE reform could provide a substantial boost to Chinese aggregate productivity growth over the medium run and help counter the downward trend of aggregate productivity amplified by the COVID-19 crisis.
  - Data show large revenue productivity gaps between listed SOEs and private firms (POEs) that reflect significant resource misallocation.
  - Credit misallocation is important in explaining distortions, affecting capital-intensity of SOEs relative to POEs.
  - Reforms that even the competitive playing field between SOEs and POEs could help drive potential output growth during the recovery from COVID-19.
  - Complementary reforms will be important to ensure inclusive distribution of gains and protect workers conducting social and non-economic functions within SOEs.

*1chnea2021002 - References ____________________________________________________________________________ 16*

### 1. State-owned enterprises command a large

### 1. State-owned enterprises command a large

### A. Size, scope, and non-economic roles of Chinese SOEs
- In 2018 (latest data available), total assets of Chinese SOEs stood at 194 percent of GDP—higher than in the early 2000s, and several orders of magnitude larger than in any other country.
- SOEs operate in all sectors of the economy, unlike other countries where SOE operations are usually concentrated in a few sectors (mostly transport, utilities, and finance).
- Since the 1990s the number of SOEs has declined significantly (a ⅔ decline among the industrial firms alone), along with the share of urban workers in SOEs.
- SOEs have played roles in supporting the economy and employment during recessions (including during the COVID-19 crisis), and provide health and pension services to the population.
- SOEs likely benefit from implicit government guarantees, enabling access to a higher share of bank financing and lower interest rates on liabilities.

### B. Productivity gaps between listed SOEs and POEs
- Data source and coverage:
  - Analysis uses the Wind database of listed firms covering over 3700 listed firms in the Shenzhen and Shanghai stock exchanges between 2002 and 2019.
  - Listed firms account for 6 percent of GDP and 10 percent of manufacturing value-added in 2019.
  - In 2019, SOEs accounted for 29 percent of listed firms and 57 percent of listed firm value-added.
  - The typical SOE employs more than twice as many workers as the typical POE.
- Revenue productivity definition:
  - Revenue productivity defined as the average product of capital and labor (value-added per unit of capital and labor), measured as value-added divided by a geometric average of capital and labor.
  - Revenue productivity varies dramatically across listed firms; the revenue productivity of firms at the 90th percentile is more than four times larger than that of firms at the 10th percentile.
- Key empirical findings:
  - There is substantial dispersion in revenue productivity for both SOEs and POEs, with significant overlap between their distributions; many SOEs are profitable and productive.
  - A large statistically significant average productivity gap exists between SOEs and POEs:
    - The gap widened to 30 percent during the Global Financial Crisis (GFC).
    - The gap reduced to 20 percent as of 2019.
  - Revenue productivity gaps are similar for central and local SOEs.
  - Gaps are pervasive across almost every sector, particularly large in Utilities, Transportation and Steel Manufacturing, and smaller in IT Services, Manufacturing of Pharmaceuticals, and Medical Equipment.
  - Sectors with larger revenue productivity gaps are also those where SOEs account for a larger share of output and inputs, amplifying distortionary impacts.
  - Evidence indicates a particularly low average product of capital for SOEs:
    - Value-added per unit of fixed assets is almost 40 percent lower for SOEs than POEs.
    - Value-added per unit of labor of SOEs is similar to that of POEs.
  - Low revenue productivity of SOEs is largely explained by an inefficiently high capital intensity of SOEs.

### C. Quantified impact of misallocation and extrapolated macro gains
- Quantitative model (Hsieh & Klenow, 2009) results for listed firms:
  - A policy that reduces the average SOE-POE revenue productivity gap in every sector could increase aggregate productivity among listed firms by between 5 and 6 percent (gains reported for 2019; historically larger gains found).
  - A policy that also reduces the distorted capital-intensity of SOEs could yield gains of over 6 percent.
- Extrapolation to the broader Chinese economy:
  - Estimates for listed firms are scaled down to around 4 percent for the whole economy (based on SOE share of assets for the whole economy being smaller than for listed firms).
  - Given aggregate TFP growth averaged 0.6 percent between 2012 and 2017 (Penn World Tables), SOE reform could more than double the rate of TFP growth for five years—or likely more if sectoral reallocation were considered.
- Comparison with prior studies:
  - Results are similar to Hsieh and Klenow (2009) who find 5.3 percent gains on average between 1998 and 2005 from reducing the average revenue productivity gap between state and private firms.
  - Brandt et al. (2013) estimate 10 percent gains for state vs. non-state capital reallocation (their estimate includes across-sector capital reallocation, which is outside the within-sector focus here).

### D. The role of credit misallocation
- Correlation between productivity and leverage:
  - There is a negative correlation between firm productivity and the leverage ratio (debt-to-asset ratio), suggesting credit is allocated to the least efficient firms.
  - On average SOEs have higher leverage ratios than POEs.
  - Although both low-productivity SOEs and low-productivity POEs have deleveraged since 2016, low-productivity SOEs remain the most leveraged.
  - The leverage ratio of low-productivity SOEs is still more than 5 percentage points higher than that of high-productivity POEs.
  - Within SOEs, the leverage ratio of high-productivity SOEs was 10 percentage points lower than the low-productivity SOEs at the end of 2016 (gap has narrowed since).
- Non-linear leverage–investment relationship:
  - Using the threshold effect model (Hansen, 2002):
    - The impact of a 1-ppt increase in a firm’s leverage ratio on investment is 0.1 ppt.
    - This impact falls dramatically to 0.01 ppt if the leverage ratio exceeds a threshold of about 35 percent.
    - Both SOEs and POEs have similar leverage thresholds and similar impact magnitudes.
  - Because a higher share of SOEs are highly indebted than POEs, deleveraging SOEs offers benefits:
    - Deleveraging highly-indebted SOEs can free resources or credit to be channeled to less-indebted POEs, boosting productive investment.
    - Simulations find reallocating credit from highly-leveraged SOEs to POEs can increase aggregate investment and boost growth by 0.3-0.4 ppts annually when highly-indebted SOEs deleverage by 2 ppts and the freed-up credit is channeled to POEs.

### E. Policy recommendations to reform Chinese SOEs
- Identify non-viable SOEs and open non-strategic sectors to private/foreign competition; allow non-viable firms to default and exit rather than being merged with more profitable SOEs to protect good performers and enhance market competition.
- Reform the social safety net in parallel to relieve SOEs of social functions (employment stabilization, health and pension benefits), transferring these obligations to the state to allow more productive private firms to hire workers.
- Ensure equal access to credit and capital for private firms and allow SOEs to deleverage:
  - Recognize and remove implicit government guarantees that enable SOEs to access financing at lower rates.
  - Prepare the financial sector for removal of implicit guarantees by requiring banks to carry higher risk weights on SOE loans, build liquidity buffers, reduce reliance on short-term funding, and increase capital.
  - Promote market-based policies to allow credit to flow to its most productive use, establish competitive neutrality among firms, strengthen credit culture, improve credit ratings, strengthen credit registries, ensure adequate bank capitalization, and promote more risk-based vs. collateral-based lending.
- Improve SOE governance:
  - Increase transparency of SOE group structures and activities.
  - Allow appointment of company managers with international/private sector experience.
  - Clarify the role of the Party in decision making.
  - Encourage corporate boards where missing, and reduce the practice of exchanging management staff between supervisors (SASAC) and central SOEs.

*Prepared by Wei Guo, Fei Han, Sarwat Jahan, Emilia Jurzyk (co-lead) and Cian Ruane (co-lead).*

### 1. China’s monetary policy framework

### 1. China’s monetary policy framework

### Transition and institutional setting
- China’s monetary policy framework continues its transition to a more market-based approach, with interest rate liberalization largely complete (with the removal of the ceiling on deposit rates in 2015) and the development of an interest rate corridor.
- The financial system remains largely bank-based; the government has regularly implemented measures to develop financial markets, including to open up China’s bond market.
- The institutional set-up is distinct from advanced countries:
  - The State Council (SC) is the decision-making body and the People’s Bank of China (PBC) does not have full operational independence.
  - The PBC’s recommendations carry disproportionate weight in the SC’s deliberations, but decisions on key monetary policy matters are collective and often taken in the context of larger policy decisions.
  - China’s 13th SC consists of 35 members: the Premier, Vice Premiers, heads of 25 ministries, and the Governor of the PBC. The Standing Committee of the SC consists of 10 members. The whole SC meets biannually or on an ad hoc basis, while the Standing Committee meets weekly.

### Monetary policy framework complexity
- Communications have improved, but lack of clarity remains on several parts of the framework, particularly on the nominal anchor (intermediate target).
- The previous official intermediate target, M2, was de-emphasized but has not been replaced with a new intermediate target.
- Multiple policy instruments appear aimed at several intermediate targets.
- The loan prime rate (LPR) reform in mid-2019 elevated the importance of the MLF rate as a medium-term policy rate; previously the PBC’s 7-day reverse repo rate was expected to be the key policy rate when benchmark lending rates were phased out.

### Main monetary policy instruments (identified from 2008 onwards)
- (i) Reserve requirement ratio (RRR): share of banks’ deposit kept in reserve with the PBC. Both broad-based changes and those targeted to a subset of banks are recorded.
- (ii) PBC’s 7-day reverse repo rate: policy rate at the center of the interest rate corridor (between interest on excess reserves and the 7-day standing lending facility rate).
- (iii) Benchmark deposit and lending rates: not changed since 2015 but used in earlier sample; when used they were adjusted in the same direction and by similar magnitudes.
- (iv) Rate on the PBC’s medium-term lending facility (MLF): since mid-2019, linked to LPR and became the main instrument to influence bank lending rates.

### Sample and shock measurement
- Text releases from the English version of the State Council website are used to classify monetary-fiscal coordination; website releases begin in May 2013, thus the sample goes from May 2013 to April 2020.
- A daily market measure captures the unexpected component of monetary policy events: the daily close-to-close change in the rate on one-year interest rate swaps (IRS) based on the interbank 7-day repo rate, measured around policy announcements.
  - Advantages: (i) directly captures the ‘surprise’ component of policy announcements, and (ii) comparable across instruments.
  - The daily shock measure is validated by a shorter sample of higher-frequency data.
- Summary statistics for the daily measure of monetary shocks (as reported):
  - Mean |ΔIRS| (bps) / Std Dev ΔIRS (%)
    - RRR 7.3 11.9
    - PBC 7-day reverse repo rate 10.9 15.6
    - Benchmark lending and deposit rate 5.5 6.8
    - MLF rate 4.3 3.2
    - All policy events 6.9 10.4
    - Non-event days 3.1 5.0
  - The standard deviation of the IRS rates on days without policy events is 5 basis points, and higher on policy announcement days, at 10 basis points.

### Types of monetary-fiscal coordination identified (from SC releases)
- Two main types:
  - (i) Joint measures focused on micro and small enterprises (MSEs).
  - (ii) Accommodative monetary policy to provide supportive conditions for fiscal stimulus.
- An alternative coordination measure (used as a cross-check) flags when multiple monetary instruments are used within a two-week span; results are similar to the text-based coordination measure.

### Empirical approach
- High-frequency identification of monetary shocks follows recent empirical literature (e.g., Nakamura and Steinsson 2018; Kamber and Mohanty 2018).
- Local projections (Jordà 2005) are estimated to study impacts over horizons h = 0, 1, 2, ...:
  - Dependent variables: interbank market rates, sovereign bond yields, and credit bond spreads (various maturities).
  - Controls: monthly measures of cyclical economic activity (IMF staff construct) and inflation.
  - To gauge coordination effects, local projections are estimated separately for coordinated and uncoordinated shocks; interaction terms with the output gap and inflation are included to control for state-dependent effects.
- Monetary policy shocks are classified by quarter as coordinated or uncoordinated based on SC text searches for monetary and fiscal keywords.

### Key empirical findings — same-day and short-run effects
- Monetary policy shocks have immediate impacts on sovereign bond yields and state-owned enterprise (enterprise) bond spreads, but less impact on corporate bond spreads and other credit bond spreads.
- Same-day (h = 0) estimates (selected results reported):
  - Sovereign yields:
    - Coefficient estimates are positive and significant for 1- to 10-year maturities.
    - A shock accompanied by a 100 bps decrease in the IRS rate is associated with about a 20-30 bps decrease in sovereign yields.
  - Enterprise bond spreads:
    - The same monetary shock is associated with about a 15-20 bps decrease in enterprise bond spreads (significant at 1- to 10-year maturities).
  - Corporate and other credit bond spreads:
    - Coefficient estimates are positive but smaller and not generally statistically significant.
- Cumulative impulse responses to a 1 percentage point expansionary monetary policy shock (decrease in the IRS rate):
  - Interbank market: 1 percentage point shock leads to a 35 bps decrease in the 3-month SHIBOR on impact.
  - Sovereign yields: 0.7 percentage point decrease in the sovereign yield after 3 days (effect fades and is not statistically significant after one week).
  - Enterprise bond spreads: 1.3 percentage point decrease at the 3-day peak response.
  - Corporate bond spreads: 1.7 percentage point decrease (timing as reported in the figures).
  - Commercial paper: effect not statistically significant for any timeframe.

### Role of monetary-fiscal coordination
- The impact of monetary policy is significantly stronger when coordinated with fiscal policy:
  - Coordinated shocks generally produce larger effects than uncoordinated shocks.
  - Only the 1-year sovereign yield is significantly affected by both uncoordinated and coordinated shocks.
  - For credit bond spreads, significant impacts depend on fiscal policy moving in the same direction as monetary policy.
  - Example magnitude: coordinated shocks reduce enterprise spreads by 3.3 percentage points 20 days after the shock.
  - Similar amplification by coordination is found for the 3- and 5-year maturities of sovereign and credit bonds (not pictured).

### Interpretation and limitations
- The PBC’s monetary policy has generally been countercyclical (IMF staff cyclical activity measure compared with the shock series).
- Collective decision-making and close monetary-fiscal coordination explain why monetary and fiscal policy often work in tandem to counter shocks.
- However, a muted market response to uncoordinated monetary policy weakens the central bank’s ability to use interest-rate channels to affect the economy, particularly when rapid, independent monetary responses are required.

### Policy implications and recommendations
- Continued reforms to strengthen the interest-rate based framework are needed, including:
  - Formulating a clear inflation objective and granting the PBC operational (instrument) independence.
    - The State Council should set overall goals for monetary policy (but not specific interest rate targets) and delegate responsibility to meet them to the PBC to avoid multiple stakeholders steering policies toward conflicting objectives and to limit fiscal dominance in non-crisis periods.
  - Streamlining and clarifying the monetary policy framework with a focus on one key policy interest rate.
    - The PBC should guide the short-term interbank rate clearly and let longer-term rates be market-determined so markets can form a yield curve reflecting expectations of future policy rates and inflation.
  - Reforms to further improve interest rate pass-through, including continued progress on LPR reform and phasing out the benchmark deposit rate.
  - Steps to increase the financial robustness of the banking system to fluctuations in short-term interest rates, including:
    - (i) raising bank capital, and
    - (ii) further developing interest-rate hedging instruments.

*Source: IMF staff compilation from the chapter "1. China’s monetary policy framework" (PDF content unit: 1chnea2021002).*

### 1. Already before the COVID-19 outbreak, China’s LGs were heavily relying on transfers

### 1. Already before the COVID-19 outbreak, China’s LGs were heavily relying on transfers

### Pre-pandemic fiscal structure and funding sources
- Most provinces rely on funding sources other than local tax and non-tax revenues to fill the local revenue-expenditure gap (fiscal gap) resulting from a long-standing misalignment between LGs’ limited revenue sources and high expenditure responsibilities.
- Major funding sources to fill fiscal gaps:
  - Transfers from the CG budget (major source, especially for provinces with large fiscal gaps).
  - Net tax refund from shared taxes between CG and LGs.
  - LG bond issuance within CG-allocated quotas.
  - Dividend transfers from local state-owned enterprises (LSOEs).
  - Withdrawals of LG deposits and land sales.
- Note: The fiscal gap is defined as the difference between the local revenues and expenditures of LGs, which would be largely closed if transfers and tax refund from the CG are also included in the revenues.

### Pandemic impact on provincial finances (January–April 2020)
- CG measures in 2020:
  - Increased ceiling on special LG bond issuance by 74 percent (or RMB1.6 trillion).
  - Issued RMB1 trillion anti-epidemic special Treasury bonds with proceeds allocated to LGs.
- Provincial outcomes:
  - LG fiscal expenditure declined in 24 out of 31 provinces during January–April 2020 by an average of 4 percent (y/y).
  - LG own fiscal revenue declined in 30 provinces by an average of 12 percent.
  - Provinces worse hit by the pandemic experienced larger widenings in their fiscal gaps, constraining LGs’ ability to support growth.

### Empirical strategy
- Objective: Disentangle risk-sharing (short-term smoothing of idiosyncratic shocks) from redistribution (permanent transfers from richer to poorer regions).
- Methods:
  - Two-step estimation approach (Von Hagen (1992); Bayoumi and Masson (1995)): cross-section regression on average levels (redistribution) and panel regression on changes (risk-sharing).
  - One-step approach (Poghosyan et al. (2016)) using pooled mean group (PMG) estimator by Pesaran et al. (2016) for robustness.
- Data: Province-level fiscal and macroeconomic data spanning from 2006 to 2018.
- Key variables: net CG transfers, pre-transfer aggregate provincial income, post-transfer disposable provincial income. All variables in real per capita terms. CG transfers at province-level available until 2018.

### Empirical results — risk-sharing and redistribution effects
- Estimated systematic fiscal risk-sharing effect:
  - Coefficients: 0.87 (two-step) and 0.82 (one-step); average 0.85.
  - Interpretation: Disposable income of a province falls by about 85 cents in response to a temporary 1-RMB decline in aggregate income relative to the national average; the remaining 15 cents are smoothed by CG transfers. Reported risk-sharing effect of 15 percent on average.
  - LSOE dividend transfers: risk-sharing effect not statistically significant.
- Estimated redistribution effect:
  - Coefficient for redistribution ≈ 0.67 (statistically significant) using both approaches.
  - Interpretation: Magnitude of redistribution effect about 33 percent — a province with a 1-RMB permanently lower aggregate income would have disposable income only 67 cents below the national average, with 33 cents covered by CG transfers.
- Time and regional variations:
  - Risk-sharing effect increased over time: from 14 percent during 2006-12 to 21 percent during 2012-18.
  - Regional differences: inland provinces ≈ 20 percent risk-sharing; coastal provinces ≈ 12 percent risk-sharing.
- Comparison with literature / advanced economies:
  - Text table entry for China (average of the two-step and one-step approaches) — sample period 2006–17: Risk-sharing effect 16% and Redistribution effect 34%.
  - Earlier China estimate (Du et al., 2011) for 1980–2007: 9% risk-sharing.
  - Note: risk-sharing of 15 percent falls within ranges reported for some advanced economies, while redistribution average of 33 percent for China is on the high end.

### Business cycle synchronization and need for shock-smoothing
- China exhibits lower business cycle synchronization than the U.S.; average divergence of provincial GDP growth is higher in China than in the U.S.
- Lower synchronization implies stronger need for idiosyncratic shock-smoothing via fiscal transfers.

### Counterfactual: rule-based automatic non-regressive transfers
- Rule construct: redistribute total CG transfers over provinces based on provincial shocks (ε_it), relative provincial size (post-transfer income), total transfers, and calibration parameter τ.
- Provincial shock regression used to derive ε_it: ∆log(GDP_it) = α_i + Σ β_ij ∆log(GDP_i,t−j)^2 + ε_it (estimated province-by-province).
- Simulation findings:
  - When τ is above a threshold (around 4), all idiosyncratic shocks could be smoothed.
  - Rule-based transfers could deliver significantly higher risk-sharing than realized ~15 percent while keeping redistribution similar to realized ~33 percent.
  - Total size of alternative transfers constrained to be ≤ total actual transfers over 2006-18; implicit assumption CG could borrow intertemporally.

### Considerations and policy recommendations for reform
- Broad objective: Reform intergovernmental framework to address misalignment between LGs’ revenues and expenditures and increase fiscal risk-sharing efficiency.
- Recommendations and considerations:
  - Reduce misalignment of central-local fiscal responsibilities (expenditure assignment based on economies of scale, equity, externalities; consider centralized pension and unemployment insurance systems).
  - Tax reforms to give LGs more authority over some tax rates and bases (e.g., PIT and property tax) to strengthen LG fiscal accountability.
  - Align LG borrowing limits with expenditure responsibilities and contain off-budget investment; ensure realistic LG financing arrangements; carefully assess and dismantle implicit guarantees on off-budget investments (e.g., allow defaults); bring non-commercial investment on budget.
  - Consider alternative automatic and non-regressive CG transfer mechanism to increase fiscal risk-sharing while keeping redistribution roughly similar; a transparent rule would signal access to resources and help prevent sharp tightening of financial conditions.
  - Common borrowing (e.g., CG-issued bonds) as second-best when CG funding constrained: reduces degree of fiscal risk-sharing relative to transfers but can deliver more targeted support and break cycle of weaker LG health → higher borrowing costs under LG self-insurance.
  - Reduce LG-LSOE-bank interlinkages to contain moral hazard from enhanced fiscal risk-sharing: remove explicit and implicit guarantees for LSOEs; strengthen governance rules to prevent LGs using transfers to protect weak LSOEs.
  - Consider countercyclical rule for LSOE dividend transfers (e.g., discretionary increases in transfer ratios during downturns) to improve LG shock-smoothing capacity.
  - Increase interregional mobility of production factors by phasing out local protectionism and LG interventions in product and factor markets to allow market and financial channels to smooth shocks and reduce need for fiscal risk-sharing.

*Source: IMF staff chapter — "Already before the COVID-19 outbreak, China’s LGs were heavily relying on transfers."*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2021/english/1chnea2021002.pdf_