## 1. A Systemic Risk Dashboard

## Source details

**Canonical URL:** [1. A Systemic Risk Dashboard](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1375.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1375.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1375.pdf.json)

---

### Introduction and objectives
- The paper investigates three questions:
  - What is the potential systemic risk in China’s current macroeconomic and financial environment?
  - What macroprudential policies have the Chinese authorities introduced in recent years?
  - How effective have these macroprudential policies been in addressing the potential systemic risk?
- Empirical sample: 171 banks.
- Analytical streams: (i) surveillance; (ii) mitigation; (iii) crisis management.
- Focus: establishing and implementing a macroprudential policy framework as core to surveillance and mitigation.

### Identification of systemic risk (dimensions and China-specific drivers)
- Two dimensions of systemic risk:
  - Time dimension (procyclicality): excessive growth in credit and asset prices.
  - Cross-sectional dimension (interconnectedness/network risk): links between banking sector, informal financing, and local government financing platforms (LGFPs).
- China-specific circumstances:
  - Economic imbalances and regional heterogeneity:
    - Average investment-to-GDP ratio: 45 percent (2000–2011).
    - Average consumption-to-GDP ratio: 40 percent (2000–2011).
    - GDP per capita in Beijing is six times larger than in Guizhou province.
  - Unfinished reforms and capital flow dynamics:
    - De jure managed floating exchange regime with a narrow fluctuation band against the U.S. dollar since 1994.
    - De jure Chinn-Ito Index values for 2005–2010: -1.16 (each year listed).
    - De facto openness index values for 2005–2010: 0.92, 1.04, 1.09, 0.89, 0.82, 0.76.
    - Money supply (M2) increases driven by accumulated foreign exchange reserves; “hot money” proxied by net errors and omissions grew since the early 2000s.
  - Financial structure dominated by banks and regional fiscal disparities:
    - Informal financing accounts for over 20 percent of total social financing in late (period indicated in source).
    - Figure 11 (2009): loans borrowed by LGFPs outweigh local government revenue in most provinces.
- Major drivers of possible increases in systemic risk:
  - Loose monetary policy and fiscal stimulus (RMB 4 trillion stimulus package launched November 2008).
  - Bank-dominated financial system and relatively limited capital account openness.
  - Combined effects: excessive credit expansion, overcapacity, capital-intensive production, asset bubble tendencies, periodic need for public-funded bank recapitalizations.

### Macroprudential policies observed in China (tools, authorities, numeric specifics)
- Policy toolkit and lead authorities:
  - Dynamic adjustment of the differentiated reserve requirement ratio — People’s Bank of China (PBC).
  - Dynamic LTV requirement for first homes and second homes — PBC; China Banking Regulatory Commission (CBRC).
  - Dynamic provisioning requirement — CBRC.
  - Countercyclical capital buffer — PBC; CBRC.
  - Capital surcharge for SIFIs — PBC; CBRC.
  - Capital conservation buffer — CBRC.
  - Leverage ratio requirement — CBRC.
  - Liquidity surcharge — CBRC.
  - Enhancing supervision for SIFIs — CBRC.
  - Early warning system — PBC; CBRC.
- PBC differentiated reserve requirement:
  - Launched in 2004 to inhibit loan expansion by institutions with a CAR < 4 percent.
  - Dynamic adjustment formula: Ratio = the robust parameter * (the required CAR – the actual CAR).
  - Required CAR = regulatory minimum CAR (8 percent) + countercyclical capital buffer + capital surcharge for SIFIs.
  - PBC used different RRRs for large banks and small/medium banks and raised RRRs since 2010 to absorb liquidity.
- CBRC macroprudential instruments (launched 2011) — key numeric specifics:
  - Minimum CARs: core tier 1 CAR = 5 percent; tier 1 CAR = 6 percent; total CAR = 8 percent.
  - Permanent capital conservation buffer = 2.5 percent.
  - Countercyclical capital buffer = 0 to 2.5 percent.
  - Capital surcharge for the five largest banks = 1 percent.
  - Resulting total CAR requirements:
    - Systemically important banks: total CAR increased to 11.5 percent (from 8 percent).
    - Other banks: total CAR increased to 10.5 percent.
  - New provisioning requirement: provisions-to-loans ratio not lower than 2.5 percent; or provisions coverage ratio not lower than 150 percent (changed from 100 percent).
    - Rule effective since January 1, 2012; SIFIs and non-SIFIs to comply before 2013 and 2016, respectively.
  - New leverage ratio requirement: not lower than 4 percent; SIFIs and non-SIFIs to meet before 2013 and 2016, respectively.
  - Liquidity indicators: LCR and NSFR not less than 100 percent; monitored since 2009 and binding in 2013.
- House-related measures (selected dated actions and numeric measures):
  - April 2010: LTV caps lowered from 80 percent to 70 percent for primary homes and to 50 percent for second homes.
  - April 2010: Interest rates on mortgages for second homes raised to 1.1 times the benchmark lending rate.
  - October 2010: Mortgages for third homes suspended.
  - January 2011: LTV cap on second home mortgages lowered to 40 percent.
  - January 2010: Taxes increased on resale of properties within five years of purchase.
  - September 2010: Exemptions from stamp duty and income tax abolished except family’s only home.
  - January 2011: Business tax (5.5 percent) levied on 100 percent of sales proceeds if holding period < 5 years.
  - January 2011: Shanghai, Chongqing launch property tax.
- RRR and policy actions timeline:
  - PBC raised the RRR twelve times for a total of 600 basis points between January 2010 and June 2011.
  - Reserve requirement decreased three times for a total 150 basis points between December 5, 2011 and May 18, 2012.
  - Official benchmark lending rate: raised five times between October 2010 and July 2011 for a total of 125 basis points; decreased two times in 2012 for a total of 56 basis points (June 7, 2012 to July 5, 2012).
- Policy interaction notes:
  - RRR serves both monetary (sterilization of FX reserve growth) and macroprudential roles.
  - August 2011: margin deposits included in reserve requirements to remove arbitrage.
  - Some measures reversed since 2011 due to concerns about economic slowdown.

### Empirical approach and key variables
- Data and sample:
  - Panel fixed-effects models on annual data between 2000 and 2011.
  - Sample: 171 banks and 31 provinces/municipalities.
  - Separate analyses: five large banks vs. 166 small and medium-sized banks; regional groups east/middle/west.
- Dependent variables:
  - Loan growth (annual growth rate of loans in each bank).
  - House price growth (annual growth rate of city/provincial house prices where banks are located).
- Policy and control variables:
  - Monetary policy proxy: interest rates (bank rate (EOP)).
  - Macroprudential proxies: reserve requirement ratio (RRR, in first-differences d_), house-related policy dummy, capital adequacy ratio (CAR), liquidity ratio, reserves for impaired loans to total loans.
  - Controls: provincial GDP growth, land supply growth, inflation (CPI), NDF (annual change), provincial real effective exchange rate, foreign exchange reserve (annual change), VIX (annual change), lagged provincial house prices, lagged loans, etc.
- Estimation details:
  - VIX index and RRR are in first-differences (d_).
  - Robust standard errors; significance: ***, **, * denote 1 percent, 5 percent, 10 percent.

### Main empirical findings (policy effectiveness and heterogeneity)
- General results:
  - RRR and house-related policies are useful in containing excessive house price growth.
  - Mixed results for loan growth: RRR, interest rates, CAR, liquidity ratio, and reserves for impaired loans are negatively associated with loan growth in many specifications, but significance varies by bank size and region.
- Individual policy associations (summary):
  - Change in RRR:
    - Negatively associated with loan growth for large banks and banks in west China.
    - Negatively associated with house price growth in most samples; exceptions: positive association for large banks and banks in middle and west China in some specifications.
  - Interest rates:
    - Negatively associated with loan growth in most samples, though insignificant for banks in middle and east China.
    - Positively associated with house price growth in all sample groups (suggesting limited and unintended role for interest rates on house prices).
  - House-related policies:
    - Negatively associated with house price growth in all samples (insignificant for banks in middle and west China).
    - Ineffective in containing loan growth in all samples.
  - Capital adequacy ratio (CAR):
    - Higher CAR negatively associated with loan growth in full sample, small/medium banks, and east China.
    - Higher CAR positively associated with house price growth in all samples (insignificant for middle and west China).
  - Liquidity ratio:
    - Generally insignificant for loan growth.
    - Positively associated with house prices in some samples (e.g., significance at 10 percent in full sample).
  - Reserves for impaired loans / gross loans:
    - Negatively associated with loan growth in most samples; positively associated with loans in west China in some specifications.
    - Insignificant for house price growth in most samples.
- Multivariate (joint) policy interactions:
  - Loan growth:
    - When interest rates, RRR, and housing policy are included together, they tend to be insignificantly associated with lower loan growth, except higher interest rates are negatively associated with loan growth in the west region.
  - House price growth:
    - Higher interest rates are not associated with lower house prices in all samples.
    - Higher RRR is associated with lower house prices in most cases except for large banks and banks in middle and west China.
    - House-related policy is effective in all samples, though ineffective for banks in middle and west China.
  - By bank type among small/medium banks:
    - RRR’s effectiveness in containing house prices: highest for foreign banks, followed by city commercial banks, joint-stock commercial banks, and least for large banks.
    - House-related policy effective across all subgroups.
- Control-variable heterogeneity:
  - Small and medium-sized banks: loan and house-price growth positively associated with provincial GDP growth.
  - Large banks: these relationships turn negative.
  - East-region banks: loans and house prices more positively associated with expectation of RMB appreciation than middle and west.

### Stylized outcomes and mechanisms
- Policy outcomes and timing:
  - Credit-to-GDP ratio surged after 2008 Q4 and exceeded the 2 percent benchmark by 2009 Q2; brought below the benchmark by 2011 Q3.
  - Mortgages significantly decreased between 2010 and 2011 and the trend of rising house prices was restrained following policy implementation.
- Mechanisms:
  - Expectations channel: tightening curbed expectations about future housing gains, reducing demand.
  - Liquidity constraint channel: example—an increase of 0.5 percent in the RRR would reduce by over RMB 300 billion the available liquidity in the banking system, affecting especially small and medium-sized banks.

### Limitations of macroprudential policies and complementary reforms
- Macroprudential tools cannot contain all systemic risks arising from the current environment.
- Constraints on effectiveness:
  - Imperfect macroprudential instruments.
  - Ongoing unfinished reforms in monetary policy framework and financial structure.
  - Institutional coordination challenges and limited data/information sharing.
- Required complementary reforms and adjustments:
  - Further commercialization of large banks to improve policy transmission.
  - Financial liberalization (interest rates, exchange rate, capital account) to reduce burden on macroprudential policies.
  - Adjust the macroprudential toolkit as liberalization proceeds so monetary policy focuses on price/output stability and macroprudential policy targets financial stability.
  - Strengthen data collection and coordination across institutions; consider establishing a permanent Financial Stability Committee chaired by a very senior official with clear mandate and membership including PBC, supervisory agencies, MOF, and other macroeconomic agencies.

### Annex highlights — estimation results and robustness notes
- Tables present panel fixed-effects regressions for 2000–2011 with robust standard errors; significance denoted by ***, **, * at 1 percent, 5 percent, 10 percent.
- Selected numeric results (examples preserved exactly as in source):
  - Full sample (Table 11): Reserve requirement ratio — Loans -1.502 (-0.468); House Price -3.362*** (-2.751).
  - Full sample: Bank rate (EOP) — Loans -14.296* (-1.820); House Price 7.078*** (2.857).
  - Large banks (Table 12): Reserve requirement ratio — Loans -11.568* (-1.959); House Price 3.498 (0.496). Number of banks = 5.000.
  - Small/medium banks (Table 13): Reserve requirement ratio — Loans -2.365 (-0.526); House Price -3.909** (-2.569).
  - West region (Table 16): Reserve requirement ratio — Loans -17.239* (-1.739); House Price -1.671 (-0.952).
  - Multivariate grouped results (Table 17): Reserve requirement ratio — House Price -4.160*** (-4.078); Bank rate (EOP) — House Price 6.707*** (3.344); House-related policy dummy — House Price -4.328*** (-7.021).
  - Subgroup small/medium (Table 18 excerpt): City Commercial Banks — Reserve requirement ratio: Loans 54.972*** (3.014); House Price -11.759*** (-6.797).
- Sample sizes and fit vary across specifications (examples preserved):
  - Number of banks and observations reported per table (e.g., full sample observations: 1,437 in some specs; R-squared overall up to 0.730 in large-bank regressions).
- Methodological notes:
  - VIX and RRR are differenced (d_).
  - House-related policy dummy excludes monetary policies.
  - Results show heterogeneity by bank size, bank type, and region; some coefficients change sign and significance across subsamples, underscoring the need for differentiated policies.

*Source: _wp1375 — IMF working paper (1. A Systemic Risk Dashboard; Section VI; Annexes I–II).*

### 1. A Systemic Risk Dashboard ...........................................................................................

### 1. A Systemic Risk Dashboard

### Introduction and objectives
- The paper investigates three questions:
  - What is the potential systemic risk in China’s current macroeconomic and financial environment?
  - What macroprudential policies have the Chinese authorities introduced in recent years?
  - How effective have these macroprudential policies been in addressing the potential systemic risk?
- The paper uses individual bank data for empirical analysis and examines economic, financial, and transitional features in China.

### Scope and approach
- Empirical sample: 171 banks.
- Analytical streams considered for enhancing financial stability: (i) surveillance—the early identification of potential threats to financial stability; (ii) mitigation—the measures to make the financial system more resilient to shocks; and (iii) crisis management—the principles and procedures for responding to distress or failures in the financial system.
- The focus is on establishing and implementing a macroprudential policy framework as core to surveillance and mitigation.

### Identification of systemic risk in China
- Systemic risk is identified in two dimensions:
  - Time dimension: excessive growth in credit and asset prices.
  - Cross-sectional dimension: links between the banking sector and informal financing and local government financing platforms.
- The paper emphasizes that China’s sheer economic size makes the effectiveness of its policies significant for both China’s and the world’s financial stability, and that China’s experiences may be useful to other economies.

### Macroprudential policies observed
- Chinese authorities have initiated a set of macroprudential policies, including:
  - Dynamic adjustment of the differentiated required reserve ratio (RRR).
  - Capital and liquidity buffers.
  - House-related policies.
- The RRR is highlighted as a macroprudential tool with observable variation across bank types (large banks and small/medium banks).

### Effectiveness and heterogeneity of policies
- Some macroprudential policy tools are effective in containing systemic risk associated with China’s current economic and financial environment.
- Specific findings:
  - The required reserve ratio (RRR) and house-related policies can help contain increases in house prices.
- Heterogeneity:
  - The role and effectiveness of macroprudential policies differ by bank location and size.
  - Policy could be differentiated between banks in east, middle and west China, and between large banks and small and medium-sized banks.
  - Less stringent macroprudential policies could be applied to banks in the west and to small and medium-sized banks.

### Limitations and complementary reforms
- Macroprudential policies cannot be expected to contain all systemic risk associated with the current environment.
- Some systemic risks can only be addressed through further financial, fiscal, and structural reforms, for example:
  - Further commercialization of large banks.
  - Financial liberalization to improve the effectiveness of macroprudential policies.
- With further liberalization, the macroprudential policy toolkit should be further adjusted.
- Coordination and synergies between monetary and macroprudential policies can help achieve both price stability and financial stability.

### Contribution relative to existing literature
- Distinctions of this paper:
  - Explores systemic risk by considering economic, financial, and transitional features specific to China.
  - Uses individual bank data rather than aggregate data.
  - Employs a tailored approach to assess the usefulness and constraints of policies within a macro-financial framework.

*Source: _wp1375 - 1. A Systemic Risk Dashboard*

### Section VI concludes with policy implications.

### _wp1375 - Section VI concludes with policy implications.

### Literature review: goals and issues
- Macroprudential policies address systemic risk in the financial system (Borio, 2009; Bank of England, 2011).
- Three central issues:
  - Identifying systemic risk.
  - Selecting macroprudential tools.
  - Determining appropriate institutional arrangements.

### Identifying systemic risk
- Definition and dimensions:
  - Systemic risk: risk of disruptions in the provision of key financial services that can have serious consequences for the real economy (IMF, 2009).
  - Two types of systemic risk:
    - Time dimension (procyclicality): procyclical bias where financial institutions take on excessive risk in upswings and become overly risk-averse in downswings; evidenced by procyclicality of credit and asset/collateral booms and busts.
    - Cross-sectional dimension (interconnectedness/network risk): common exposures and linkages that amplify and rapidly transmit shocks (counterparty credit risk; interbank freezes).
- Diagnostic guidance (FSB-IMF-BIS, 2011; IMF, 2011):
  - Use aggregate indicators of imbalances (e.g., bank credit, external imbalances) to signal time-dimension risk.
  - Use high-frequency market indicators, macro stress testing, and concentration metrics to identify cross-sectional risk.
  - IMF (2011) recommends:
    - Improve early-warning capability using indicators such as credit-to-GDP gap, bank stability index, systemic contingent claims analysis (CCA).
    - Construct integrated systemic risk measures by combining signals from individual tools.
    - Develop a systemic risk dashboard (Table 1 in source), including tools such as Systemic CCA, distress dependence (JPod, BSI), leverage measures, macro stress tests, EDF measures for main SIFIs, network models, CCA-related joint-loss measures.
- Potential impact/likelihood channels include asset quality/price deviation, credit/GDP deviation, house prices, interbank exposures, core/non-core liabilities.

### Selecting macroprudential policy tools
- Tools mapped to dimensions of risk (summarized from IMF, 2011 and related sources):
  - Time-dimension tools (examples):
    - Countercyclical capital buffers.
    - Time-varying loan-to-value (LTV), Debt-To-Income (DTI), Loan-To-Income (LTI) caps.
    - Time-varying caps and limits on credit or credit growth.
    - Dynamic provisioning; stressed VaR; time-varying limits on loan-to-deposit ratio.
  - Cross-sectional-dimension tools (examples):
    - Systemic capital surcharges.
    - Levy on non-core liabilities.
    - Systemic liquidity surcharges.
    - Powers for resolution and higher capital charges for non-CCP cleared trades.
  - Recalibrated instruments (originally non-systemic, adapted to systemic risk): deposit insurance risk premiums sensitive to systemic risk; restrictions on permissible activities (e.g., ban on proprietary trading for systemically important banks); capital charge on derivative payables.
- International frameworks and calibrations:
  - Basel III (2010) mechanisms for time-dimension risk include:
    - Permanent capital conservation buffer of 2.5 percent of Common Equity Tier 1 Capital.
    - Countercyclical capital buffer that ranges from 0 to 2.5 percent of risk-weighted assets.
    - Minimum leverage ratio and new liquidity standards.
  - FSB framework for cross-sectional risks emphasizes additional loss-absorbency capability and resolution requirements for SIFIs.
- Role and prominence of reserve requirement ratio (RRR):
  - RRR characteristics and advantages:
    - System-wide measure (not sector-specific).
    - Can be effective against systemic risk; reserve requirements and other macroprudential instruments have a moderate and transitory effect and complement monetary policy.
    - Time-varying RRR rules can countervail adverse macro shocks and the financial accelerator mechanism.
    - Particularly useful in emerging economies:
      - Federico, A. Vegh, and Vuletin (2012): 74 percent of emerging economies use the RRR as a countercyclical tool compared with 38 percent in advanced economies.
      - Federico and others (2012) analyze 52 countries from 1970 to 2012 and conclude the RRR has replaced monetary policy as a countercyclical tool in most emerging economies for reasons including reluctance to change interest rates due to exchange rate concerns or capital inflows.
    - Policy interactions:
      - An increase in the RRR can reduce excessive credit growth without attracting new capital inflows or causing exchange rate appreciation.
      - RRR increases that dampen capital inflows can give more room for monetary policies to raise interest rates (Tovar and others, 2012).
  - Caution: RRR has also been used as a monetary tool; mixing macroprudential and monetary functions in RRR can confuse the signaling channel of monetary policy transmission.
- Complementarity with other policy tools:
  - Macroprudential instruments are typically less blunt than monetary tools and have shorter implementation lags than fiscal tools.
  - Effectiveness depends on interplay with monetary, fiscal, liquidity management, capital controls, and infrastructure policies.
  - IMF view: monetary and macroprudential policies are mainly complements, not substitutes; policymakers should employ monetary policy (focus on price stability) and macroprudential policy (focus on financial stability) simultaneously, with interaction depending on country-specific circumstances (financial structure, capital account openness).

### Institutional arrangements for macroprudential policy
- Key elements for effective institutional arrangements (IMF survey):
  - Institutional structure.
  - Mandate ownership.
  - Independent powers of key policymakers.
  - Accountability for financial stability.
- Seven models summarized (Nier and others, 2011a) analyzed across five factors:
  - Roles of the central bank and other supervisors.
  - Ownership of the macroprudential mandate.
  - Role of government (treasury).
  - Organizational separation of decision making and control over instruments.
  - Whether a committee without a macroprudential mandate can coordinate bodies.
- Assessment criteria: effective risk identification; timely and effective use of tools; effective coordination across policy.
- General findings:
  - Central bank should play a significant role in macroprudential policy.
  - Dominant participation of government (treasury) may give rise to significant risk.
- Strengths and weaknesses of institutional arrangements (summary from Table 4):
  - Full integration within central bank (Group of Full Integration / Twin peaks):
    - Strengths: delivery of prudential information to decision makers; use of existing expertise; effective coordination across macroprudential, monetary, microprudential functions; clear mandate and responsibility.
    - Weaknesses: concentrates power in central bank; perceived failures in prudential policy can affect monetary credibility; treasury typically excluded.
  - Partial integration (integration between central bank and prudential supervisor):
    - Strengths: limits reputational risks to central bank; allows treasury participation which may help garner political support.
    - Weaknesses: policymaker may lack immediate control over tools; separation between decision-making and control remains an issue.
  - Separation (institutional separation between central bank and supervisory functions):
    - Strengths: keeps agencies focused; strong institutional cultures; separate accountability.
    - Weaknesses: no single institution has all information for systemic risk analysis; risk of gaps, undetected risks, uncoordinated overlap; diluted accountability.
- Empirical patterns:
  - Central banks play a more significant role in macroprudential policy in emerging market countries than in advanced countries.
  - IMF (2011a) survey: countries where central banks are responsible for banking supervision account for two-thirds of emerging market countries but only one-half of advanced countries.
  - Merrouche and Nier (2010): stronger integration between central bank and banking supervisor more likely to reduce systemic risk associated with current account deficits.

### Potential systemic risk in China — China-specific circumstances
- Summary: Systemic risk in China has been rising due to macroeconomic and financial imbalances and credit expansion over the past four years (relative to source period).
- Three China-specific circumstances:
  1. Economic imbalances and regional heterogeneity:
     - Overreliance on investment and regional differences are structural weaknesses.
     - Average investment-to-GDP ratio: 45 percent between 2000 and 2011.
     - Average consumption-to-GDP ratio: 40 percent between 2000 and 2011.
     - China’s investment-to-GDP ratio is the highest among G-20 economies (Figure 1 in source).
     - GDP per capita in Beijing is six times larger than in Guizhou province (Figure 2 in source).
  2. Unfinished reforms and capital flow dynamics:
     - Exchange rate regime, capital account management, and interest rates are still being liberalized.
     - China has maintained a de jure managed floating exchange regime with a narrow fluctuation band against the U.S. dollar since 1994.
     - A decade-long current account surplus and accumulation of foreign reserves have driven RMB appreciation expectations and encouraged capital inflows; “hot money” proxied by net errors and omissions grew since the early 2000s (Figure 3).
     - Money supply (M2) increased mainly driven by accumulated foreign exchange reserves (Figure 4).
     - Capital controls remained tight; China’s de jure Chinn-Ito Index values for 2005–2010 are: -1.16, -1.16, -1.16, -1.16, -1.16, -1.16.
     - De facto openness index values for 2005–2010 are: 0.92, 1.04, 1.09, 0.89, 0.82, 0.76.
     - Capital controls and limited outflow channels may have amplified domestic credit boom.
  3. Financial structure dominated by banks and regional fiscal disparities:
     - Banking system dominates credit transmission; financial markets are underdeveloped (Figure 5).
     - Fiscal gaps (difference between local fiscal expenditure and revenue, in percent of local fiscal revenue) are large in some provinces (Figure 6), complicating synchronized policy responses.
- Informal financing and LGFPs:
  - Rapid growth in size and complexity of financial system, including informal financing and local government financing platforms (LGFPs), increases interconnectedness and common exposures.
  - Informal financing accounts for over 20 percent of total social financing in late (period indicated in source).

### Systemic risk in China — observed dynamics and vulnerabilities
- Time-dimension (procyclical) risk:
  - China experienced an unprecedented surge in credit during 2008–2010 following the RMB 4 trillion stimulus package launched in November 2008.
  - Local governments tapped financing through LGFPs to support local economies.
  - Credit-driven stimulus likely to worsen credit quality and negatively impact financial stability; domestic and global experience suggest such stimulus can be problematic.
  - Comparative evidence: recent credit expansion in China is among the highest in five housing boom episodes and close to Hong Kong SAR before the 1997 Asian financial crisis (Figure 7).
- Asset-price (housing) risks:
  - Rapid credit expansion has contributed to potential housing-market bubble risks.
  - Low and occasionally negative real deposit rates, large equities volatility, RMB appreciation expectations, and capital controls channeled savings into real estate, triggering house-price booms (Figure 8).
  - House prices in China increased more rapidly than during housing booms in Korea, the U.K., and the U.S. (Figure 9).
  - Strong procyclical dimension: easy credit and high risk appetite fueling exuberant asset-market behavior.
- Cross-sectional (interconnectedness) risk:
  - Increased size and complexity, including informal financing and LGFPs, enhance potential for rapid contagion due to interconnectedness and common exposures.
  - Informal financing accounted for over 20 percent of total social financing in late (period indicated in source), illustrating the scale of nonbank channels.

*Source: _wp1375 - Section VI concludes with policy implications.*

### 2012. Figure 11 shows that loans borrowed by LGFPs outweigh the local fiscal revenues in

### _wp1375 - 2012. Figure 11 shows that loans borrowed by LGFPs outweigh the local fiscal revenues in

### LGFPs and local fiscal capacity (Figure 11)
- Figure 11 shows that loans borrowed by Local Government Financing Platforms (LGFPs) outweigh local government revenue in most provinces in 2009.
- Note: some credit lines may not be fulfilled as a result of the policy tightening in 2011.

### Major drivers of possible increases in systemic risk in China
- Loose monetary policy and fiscal stimulus during the financial crisis encouraged credit expansion.
- Macroeconomic and financial environment, including unfinished interest rate and exchange rate liberalization, may have contributed to liquidity and credit expansion.
- A bank-dominated financial system and relatively limited capital account openness further added to credit expansion and house price increases.
- Combined effects produced downside risks in the form of: excessive credit expansion, overcapacity, a capital-intensive means of production, a tendency for asset bubbles, and periodic need for public-funded bank recapitalizations.

### Rationale for focusing on credit and asset prices
- The paper uses credit and asset prices as the major proxies for systemic risk in China.
- Global experiences indicate credit and asset prices can signal systemic risk buildup as early as two to four years before crises.
- Credit-fueled investment boom since the 2007 global crisis increased concern about banks’ asset quality and asset price deviations.
- Credit expansion has been extensively linked to other parts of the financial sector, such as informal financing and LGFPs.

### Macroprudential policy toolkit (summary)
- Policy instruments aimed at curbing credit boom and house price increases while avoiding cross-sectional systemic risk.
- Instruments and authorities (as listed):
  - Dynamic adjustment of the differentiated reserve requirement ratio — People’s Bank of China
  - Dynamic LTV requirement for first homes and second homes — People’s Bank of China; China Banking Regulatory Commission
  - Dynamic provisioning requirement — China Banking Regulatory Commission
  - Countercyclical capital buffer — People’s Bank of China; China Banking Regulatory Commission
  - Capital surcharge for SIFIs — People’s Bank of China; China Banking Regulatory Commission
  - Capital conservation buffer — China Banking Regulatory Commission
  - Leverage ratio requirement — China Banking Regulatory Commission
  - Liquidity surcharge — China Banking Regulatory Commission
  - Enhancing supervision for SIFIs — China Banking Regulatory Commission
  - Early warning system — People’s Bank of China; China Banking Regulatory Commission

### People’s Bank of China: differentiated reserve requirement
- Differentiated reserve requirement launched in 2004 to inhibit loan expansion by financial institutions with a low CAR and deterioration of asset quality.
- Purpose: reduce credit by decreasing loans and increasing reserves of banks with a CAR of less than 4 percent.
- Dynamic adjustment formula (as described):
  - Ratio = the robust parameter * (the required CAR – the actual CAR)
  - Required CAR = regulatory minimum CAR (8 percent) + countercyclical capital buffer + capital surcharge for SIFIs.
  - Robust parameter based on indicators: liquidity position, leverage ratio, provisioning, credit rating, internal risk management, payment and settlement cases, implementation of credit policy.
- PBC used different RRRs for large banks and small/medium banks and raised RRRs since 2010 to absorb liquidity.

### CBRC macroprudential instruments (launched in 2011) — key numeric specifics
- Revised CARs:
  - Minimums: core tier 1 CAR = 5 percent; tier 1 CAR = 6 percent; total CAR = 8 percent.
  - Permanent capital conservation buffer = 2.5 percent.
  - Countercyclical capital buffer = 0 to 2.5 percent.
  - Capital surcharge for the five largest banks = 1 percent.
  - Resulting total CAR requirements:
    - Systemically important banks: total CAR increased to 11.5 percent (from 8 percent).
    - Other banks: total CAR increased to 10.5 percent.
- New provisioning requirement:
  - Banks must meet the tighter of: provisions-to-loans ratio not lower than 2.5 percent; or provisions coverage ratio not lower than 150 percent (changed from 100 percent).
  - Rule effective since January 1, 2012; systemically important banks and non-systemically important banks to comply before 2013 and 2016, respectively.
- New leverage ratio requirement: not lower than 4 percent; systemically important banks and non-systemically important banks to meet before 2013 and 2016, respectively.
- Liquidity indicators:
  - LCR and NSFR set at not less than 100 percent for banks; monitored since 2009 and binding in 2013.
- Adjustments to LTV ratio were used repeatedly since 2007 to contain real estate risk.

### Stylized facts and observed policy actions (2010–2012)
- The PBC raised the RRR twelve times for a total of 600 basis points between January 2010 and June 2011.
- The PBC used dynamic adjustments to the differentiated reserve requirement since 2010.
- The CBRC employed stronger provisions-to-loans ratios to lower poor quality loans.
- Outcomes and indicators:
  - Growth rate of domestic credit was kept low in 2010 and 2011, contrasting with the dramatic increase between 2008 Q4 and 2009 Q3.
  - Credit-to-GDP ratio rapidly increased after 2008 Q4 and exceeded the threshold for launching counter-cyclicality buffers (i.e. 2 percent) by 2009 Q2; brought below the benchmark by 2011 Q3.
  - Mortgages significantly decreased between 2010 and 2011 and the trend of rising house prices was restrained following policy implementation.
- Mechanisms through which macroprudential policies work:
  - Expectations channel: tightening curbed expectations about future housing gains, reducing demand.
  - Liquidity constraint channel: example — an increase of 0.5 percent in the RRR would reduce by over RMB 300 billion the available liquidity in the banking system, affecting especially small and medium-sized banks.

### Interaction with monetary policy and reversals
- RRR serves both monetary and macroprudential roles: sterilizing reserve money from FX reserve growth and as a macroprudential instrument.
- August 2011 announcement included margin deposits in reserve requirements to remove arbitrage (margin deposits were included in loan-to-deposit but excluded from RRR calculations).
- Due to policy effectiveness and concerns about economic slowdown, some measures were reversed since 2011.

### Instruments for curbing house prices (Table 8) — dated actions and numeric measures
- April 2010: Caps on the LTV were lowered from 80 percent to 70 percent for primary homes and to 50 percent for second homes.
- April 2010: Interest rates on mortgages for second homes were raised to 1.1 times the officially administered benchmark lending rate.
- October 2010: Mortgages for third homes were suspended.
- January 2011: The LTV cap on second home mortgages was lowered to 40 percent.
- 2010–2011: The reserve requirement was raised twelve times for a total of 600 basis points between January 18, 2010 and June 20, 2011.
- 2011–2012: The reserve requirement was decreased three times for a total 150 basis points during the period December 5, 2011–May 18, 2012.
- 2010–2011: The official benchmark lending rate was raised five times between October 2010 and July 2011 for a total of 125 basis points.
- 2012: The official benchmark lending rate was decreased two times from June 7, 2012 to July 5, 2012 for a total of 56 basis points.
- January 2010: Taxes were increased on the resale of properties within five years of purchase.
- September 2010: Exemptions of home purchases from stamp duty and of home sales from income tax were abolished for all transactions except family’s only home.
- January 2011: Business tax (5.5 percent) will be levied on 100 percent of the sales proceeds if holding period is less than 5 years (previously levied on capital gains only).
- January 2011: Shanghai, Chongqing launch property tax.

### Empirical analysis — scope and methodology
- Objective: empirically investigate the effectiveness of macroprudential policies in addressing systemic risk (excessive credit expansion and asset price increases).
- Main questions explored:
  - Relationship between individual policies (macroprudential, monetary, housing-related) and proxies for systemic risk (credit and house prices).
  - Heterogeneous effects by bank size and region.
  - Joint relationship of macroprudential and monetary policies with credit and house price proxies.
- Data and sample:
  - Panel fixed-effects models on annual data between 2000 and 2011.
  - Sample: 171 banks and 31 provinces/municipalities.
  - Analysis separately for five large banks and 166 small and medium-sized banks; small/medium banks further divided into joint stock commercial banks, city commercial banks, and foreign banks.
  - Banks also separated into regional groups: east, middle, west. The five large banks mainly placed in the east group.
- Proxies and variables:
  - Monetary policy proxy: interest rates.
  - Macroprudential proxies: RRR, house-related policies (dummy summarizing accumulated measures), capital ratio (CAR), liquidity ratio, reserves for impaired loans to total loans.
  - Dependent variables: loan growth (growth rate of loans in each bank) and house prices (growth rate of house prices in each city where banks are located).
  - Control variables include: provincial GDP growth, provincial land supply growth, inflation (annual change in CPI), NDF (annual change in Non-Deliverable Forward), provincial real effective exchange rate, foreign exchange reserve (annual change), VIX (annual change), among others (see regression variable listing).
- Estimation strategy:
  - Two panel regression types:
    - Individual macroprudential and monetary policies entered separately to assess associations with loan growth and house price growth.
    - All macroprudential policies and monetary policy entered together in a single specification.

### Preliminary empirical findings (summary)
- Mixed results depending on bank size and location.
- RRR and house-related policies are useful in containing excessive growth in house prices.
- Further details and regression tables are provided in subsequent sections (referenced Tables 11 to 16 and Figures 12–18 in the source).

*Source: IMF working paper content provided in the supplied PDF excerpt.*

### Annex II).

### _wp1375 - Annex II)

### Individual policies: association with loan growth
- The RRR, interest rates, the ratio of reserves for impaired loans to gross loans, the liquidity asset ratio, and the CAR are negatively associated with growth in loans in most cases, although significance varies with different samples of size and location.
- Change in the RRR:
  - Negatively associated with loan growth for large banks and banks in west China.
- Interest rates:
  - Negatively associated with loan growth in all samples, although insignificant for banks in middle and east China.
- House-related policies:
  - Ineffective in containing loan growth in all samples.
- CAR:
  - A higher CAR is negatively associated with loan growth in the full sample, small and medium-sized banks, and banks in east China.
  - Insignificant in the large bank group, reflecting that loan growth in large banks is not sensitive to the capital buffer requirement.
- Liquidity ratio:
  - Insignificant in all samples, indicating its limited role in addressing loan growth.
- Ratio of reserves for impaired loans to gross loans:
  - Negatively associated with loan growth in all samples, although positively associated with banks in west China.
  - Indicates the ratio is not effective in managing loan growth in the west region.

### Individual policies: association with house price growth
- Among macroprudential policy measures, the RRR has the most significant negative association with house prices, although not for large banks and banks in middle and west China.
- Change in the RRR:
  - Negatively associated with house price growth in all samples, except it has a positive association for large banks and banks in middle and west China.
- Interest rates:
  - Positively associated with house price growth in all sample groups, indicating a limited and unintended role for interest rates in managing house prices.
- House-related policies:
  - Negatively associated with house price growth in all samples, although insignificant for banks in middle and west China.
- Reserves for impaired loans:
  - Insignificantly associated with house price growth in all sample groups, indicating a limited association.
- CAR:
  - A higher CAR is positively associated with house price growth in all samples, although insignificant for banks in middle and west China.
  - May reflect that banks with higher CARs tend to expand credit, contributing to increases in house prices.

- Note on statistical notation used in the source:
  - "+/- indicate the sign of estimated coefficients. ***, **, and * denote statistical significance at the 1 percent, 5 percent, and 10 percent levels of confidence based on robust standard errors."

### Multivariate policy interactions
- Summary: The RRR and house-related policies are useful in containing excessive house price growth.
- Multivariate policies in relation to loan growth:
  - All three policy tools—higher interest rates, higher RRR, and tougher housing policy—turn out to be insignificantly associated with lower loan growth except that higher interest rates are negatively associated with loan growth in the banks of the west.
  - Implication: When RRR and interest rate policy are implemented, higher interest rates slow loan growth only in the west region.
- Multivariate policies in relation to house price growth:
  - Higher interest rates are not associated with lower house prices in all samples.
  - A higher RRR is associated with lower house prices in all cases except in large banks and banks in middle and west China.
  - House-related policy is effective in all samples, although ineffective for banks in middle and west China.
  - By bank type (Table 18 / Figure 19):
    - In the group of small and medium-sized banks, the RRR is most negatively associated with foreign banks; coefficients for joint-stock commercial banks and city commercial banks turn insignificant.
    - Effectiveness of the RRR in containing house prices: highest for foreign banks, followed by city commercial banks, joint-stock commercial banks, and lastly the large banks.
    - House-related policy is effective across all subgroups of banks.

### Control variable results and cross-sectional differences
- Differences by bank size:
  - For small and medium-sized banks, growth rates in loans and house prices are positively associated with provincial GDP growth.
  - For large banks, these relationships turn negative.
  - Interpretation: small banks are more market-oriented than large ones.
- Differences by region:
  - Loans and house prices supported by banks in the east are more positively associated with the expectation of RMB appreciation than banks in the middle and west, reflecting the more export-oriented economy of the east region.

### Policy implications and recommendations
- Systemic risk in China exists in both time and cross-sectional dimensions:
  - Time: excessive growth in credit and asset prices.
  - Cross-sectional: links between the banking sector and informal financing and local government financing platforms.
- Macroprudential policies can be useful in containing financial stress:
  - When lower interest rates spur loan growth and risk-taking, the RRR and a house-related policy could be adopted to contain excessive house price increases.
  - Macroprudential policies targeted at excessive house price increases can contain undesirable side effects of monetary policy on financial stability.
- Current constraints on effectiveness:
  - Monetary policy framework and financial structure with ongoing unfinished reforms impose burdens on macroprudential policies.
  - Imperfect macroprudential tools, hard-to-capture financial stability concerns, limited tools, and institutional coordination challenges constrain effectiveness.
- Better-targeted and regionally differentiated policies:
  - Higher interest rates and tighter macroprudential policies are more associated with lowering loan growth in banks in west China, even though the west needs more financial support.
  - Policy could be differentiated between banks in east/middle and west China, with less tough macroprudential policies applied to banks in the west.
- Macroprudential policies can ease the burden on monetary policy:
  - Well-calibrated macroprudential tools can contain risks ex-ante and reduce pressure to raise interest rates, alleviating the dilemma between encouraging capital inflows and hurting economic recovery.
- Commercialization of large banks:
  - Further commercialization of the large banks would help improve policy transmission and the effectiveness of macroprudential policies.
  - Empirical evidence shows the RRR’s effectiveness in containing house prices is highest for foreign banks and least for large banks.
- Need for broader reforms:
  - Macroprudential policies should be complemented by financial, fiscal, and structural reforms.
  - The ineffectiveness of an increase in land supply in slowing house prices indicates attention should be paid to reducing speculative demand and slowing credit expansion.
  - Further liberalization of interest rates, exchange rates, and the capital account would reduce the burden on macroprudential policies and enhance their effectiveness.
- Toolkit adjustments and institutional framework:
  - With further liberalization, the macroprudential toolkit should be adjusted so monetary policy focuses on price and output stability while macroprudential policy targets financial-stability threats.
  - Establishing a supportive institutional framework is vital—well-coordinated arrangements among agencies are needed.
  - One proposed option: establish a permanent Financial Stability Committee, chaired by a very senior official with a clear mandate for monitoring systemic risk and making recommendations; membership would include PBC, the three supervisory agencies, the MOF, and other relevant macroeconomic agencies.
- Data and coordination:
  - China needs to strengthen data collection and improve coordination among various institutions, along with potential establishment of new institutions.

*Source: _wp1375 - Annex II).*

### Annex I.  List of Sample Banks

### Annex I.  List of Sample Banks

### List of Sample Banks (Numbers 1–39)
- 1 Agricultural Bank of China Limited East
- 2 Allied Commercial Bank East
- 3 Bangkok Bank (China) Co. Ltd. East
- 4 Bank of Anshan Co. Ltd. East
- 5 Bank of Beijing Co. Ltd. East
- 6 Bank of Changsha Co. Ltd. Middle
- 7 Bank of Chengdu Co. Ltd. West
- 8 Bank of China Limited East
- 9 Bank of Chongqing West
- 10 Bank of Communications Co. Ltd. East
- 11 Bank of Dalian East
- 12 Bank of Deyang West
- 13 Bank of Dongguan East
- 14 Bank of East Asia (China) Ltd. East
- 15 Bank of Fuxin Co. Ltd. East
- 16 Bank of Guangzhou Co., Ltd. East
- 17 Bank of Guilin Co. Ltd. West
- 18 Bank of Hangzhou Co. Ltd. East
- 19 Bank of Hebei Co. Ltd. East
- 20 Bank of Inner Mongolia Co., Ltd. Middle
- 21 Bank of Jiangsu Co. Ltd. East
- 22 Bank of Jilin Co. Ltd. Middle
- 23 Bank of Jinhua Co. Ltd. East
- 24 Bank of Jining Co. Ltd. East
- 25 Bank of Jinzhou Co. Ltd. East
- 26 Bank of Jiujiang Co. Ltd. West
- 27 Bank of Kunlun Co. Ltd. West
- 28 Bank of Liaoyang Co. Ltd. East
- 29 Bank of Liuzhou Co. Ltd. West
- 30 Bank of Luoyang Co. Ltd. Middle
- 31 Bank of Montreal (China) Co. Ltd. East
- 32 Bank of Nanchang Co., Ltd. Middle
- 33 Bank of Nanjing East
- 34 Bank of Ningbo East
- 35 Bank of Ningxia Co. Ltd. West
- 36 Bank of Qingdao Co. Ltd. East
- 37 Bank of Rizhao East
- 38 Bank of Shanghai East
- 39 Bank of Shaoxing Co. Ltd. East

### List of Sample Banks (Numbers 40–79)
- 40 Bank of Suzhou Co. Ltd. East
- 41 Bank of Tianjin East
- 42 Bank of Tokyo Mitsubishi UFJ (China) Ltd. East
- 43 Bank of Weifang Co. Ltd. East
- 44 Bank of Wenzhou Co. Ltd. East
- 45 Bank of Xinxiang Co. Ltd. Middle
- 46 Bank of Yingkou East
- 47 Baoshang Bank East
- 48 Beijing Rural Commercial Bank Co. Ltd. East
- 49 BNP Paribas (China) East
- 50 Cangzhou City Commercial Bank East
- 51 Changshu Rural Commercial Bank East
- 52 Changzhi City Commercial Bank Middle
- 53 Chengde City Commercial Bank Co. Ltd. East
- 54 China & South Sea Bank Ltd., (The) East
- 55 China Bohai Bank East
- 56 China CITIC Bank Corporation Limited East
- 57 China Construction Bank Corporation East
- 58 China Development Bank Corporation East
- 59 China Everbright Bank Co. Ltd. East
- 60 China Guangfa Bank Co. Ltd. East
- 61 China Merchants Bank Co. Ltd. East
- 62 China Minsheng Banking Corporation East
- 63 China Resources Bank of Zhuhai Co. Ltd. East
- 64 China State Bank Ltd. East
- 65 China Zheshang Bank Co. Ltd. East
- 66 Chinese Mercantile Bank East
- 67 Chongqing Rural Commercial  Bank West
- 68 Citibank (China) Co. Ltd. East
- 69 CITIC Bank International (China) Limited East
- 70 Commercial Bank of Zhengzhou Middle
- 71 Credit Agricole CIB (China) East
- 72 Dah Sing Bank, Ltd. East
- 73 Datong City Commercial Bank Middle
- 74 DBS BANK (China) Limited East
- 75 Deutsche Bank (China) Co. Ltd. East
- 76 Dezhou Bank Co. Ltd. East
- 77 Dongguan Rural Commercial Bank Co. Ltd. East
- 78 Dongying Bank Co. Limited East
- 79 East West Bank (China) Limited East

### List of Sample Banks (Numbers 80–119)
- 80 Evergrowing Bank Co. Ltd. East
- 81 First Sino Bank East
- 82 Foshan Shunde Rural Commercial Bank Company Limited East
- 83 Fudian Bank Co. Ltd. West
- 84 Fujian Haixia Bank Co. Ltd. East
- 85 Ganzhou City Commercial Bank Middle
- 86 Guangdong Huaxing Bank Co. Ltd. East
- 87 Guangxi Beibu Gulf Bank Co. Ltd. West
- 88 Guangzhou Rural Commercial Bank Co., Ltd. East
- 89 Guiyang Commercial Bank Co. Ltd. West
- 90 Hana Bank (China) Company Ltd. East
- 91 Handan Commercial Bank Co Ltd. East
- 92 Hang Seng Bank (China) Limited East
- 93 Hankou Bank Middle
- 94 Harbin Bank Middle
- 95 Hengyang City Commercial Bank Middle
- 96 HSBC Bank (China) Co. Ltd. East
- 97 Hua Xia Bank Co., Limited East
- 98 Huangshi City Commercial Bank Middle
- 99 Huishang Bank Co. Ltd. Middle
- 100 Huzhou City Commercial Bank East
- 101 Industrial & Commercial Bank of China (The) - ICBC East
- 102 Industrial Bank Co. Ltd. East
- 103 Industrial Bank of Korea (China) Limited East
- 104 Jiangsu Jiangyin Rural Commercial Bank East
- 105 Jiangsu Wujiang Rural Commercial Bank East
- 106 Jiangsu Zhangjiagang Rural Commercial Bank Co. Ltd. East
- 107 Jiaozuo City Commercial Bank Co. Ltd. Middle
- 108 Jiaxing City Commercial Bank Co. Ltd. East
- 109 Jingzhou City Commercial Bank Co. Ltd. Middle
- 110 Jiujiang City Commercial Bank Co. Ltd. Middle
- 111 JP Morgan Chase Bank (China) Co. Ltd. East
- 112 Karamay City Commercial Bank Co. Ltd. East
- 113 KEB Bank (China) Co., Ltd. East
- 114 Kincheng Banking Corporation East
- 115 Kunshan Rural Commercial Bank East
- 116 Kwangtung Provincial Bank (The) East
- 117 Laishang Bank Co. Ltd. East
- 118 Lanzhou City Commercial Bank Co. Ltd. West
- 119 Linshang Bank Co. Ltd. East

### List of Sample Banks (Numbers 120–159)
- 120 Metropolitan Bank (China) Ltd. East
- 121 Mianyang City Commercial Bank West
- 122 Mizuho Corporate Bank China Ltd. East
- 123 Nanchong City Commercial Bank Co., Ltd. East
- 124 Nanhai Rural Credit Union East
- 125 Nanning City Commercial Bank West
- 126 Nantong City Commercial Bank Co. Ltd. East
- 127 Nanyang Commercial Bank (China) Limited Middle
- 128 National Commercial Bank Ltd. East
- 129 Ningbo Commercial Bank Corporation Ltd. East
- 130 OCBC Bank (China) Limited East
- 131 Ordos Commercial Bank Middle
- 132 Panzhihua City Commercial Bank Co. Ltd. West
- 133 Ping An Bank East
- 134 Ping An Bank Co. Ltd. East
- 135 Ping An Bank Co. Ltd. East
- 136 Qilu Bank Co Ltd. East
- 137 Qingdao International Bank East
- 138 Qishang Bank. East
- 139 Royal Bank of Scotland (China) Co. Ltd. (The) East
- 140 Shanghai Pudong Development Bank East
- 141 Shanghai Rural Commercial Bank East
- 142 Shangrao City Commercial Bank Middle
- 143 Shengjing Bank East
- 144 Shenzhen Rural Commercial Bank Co. Ltd. East
- 145 Shinhan Bank (China) Limited East
- 146 Sin Hua Bank Limited East
- 147 Societe Generale (China) Limited East
- 148 Standard Chartered Bank (China) Ltd. East
- 149 Sumitomo Mitsui Banking Corporation (China) Limited East
- 150 Suzhou City Commercial Bank Co. Ltd. East
- 151 Taizhou City Commercial Bank East
- 152 United Overseas Bank (China) Limited East
- 153 Weihai City Commercial Bank Co. Ltd. East
- 154 Wing Hang Bank (China) Ltd. East
- 155 Woori Bank (China) Ltd. East
- 156 Wuxi City Commercial Bank East
- 157 Xi'an City Commercial Bank West
- 158 Xiamen Bank East
- 159 Xiamen International Bank East

### List of Sample Banks (Numbers 160–171)
- 160 Xiaogan City Commercial Bank Co. Ltd. Middle
- 161 Xinyang City Commercial Bank Co. Ltd. Middle
- 162 Xuchang City Commercial Bank Co. Ltd. Middle
- 163 Yangzhou City Commercial Bank Ltd. East
- 164 Yantai Bank Co. Ltd. East
- 165 Yien Yieh Commercial Bank Ltd. East
- 166 Zhanjiang City Commercial Bank Co. Ltd. East
- 167 Zhejiang Chouzhou Commercial Bank East
- 168 Zhejiang Mintai Commercial Bank East
- 169 Zhejiang Tailong Commercial Bank Co. Ltd. East
- 170 Zhengxin Bank Company Limited East
- 171 Zhenjiang City Commercial Bank East

*Annex I.  List of Sample Banks*

### Annex II. Detailed Panel Estimation Results

### Annex II. Detailed Panel Estimation Results

### Full sample (Table 11 — Individual Variables Based on the Full Sample)
- Provincial GDP growth:
  - Loans: 0.168 (0.292)
  - House Price: 0.588** (2.042)
  - Additional model estimates include: Loans 0.304 (0.543); House Price 0.463 (1.495); Loans 0.220 (0.416); House Price 0.361 (1.145); Loans 0.361 (0.833); House Price 0.695** (1.990); Loans 0.059 (0.113); House Price 0.830*** (2.820); Loans 0.531 (1.045); House Price 0.870** (2.483)
- Inflation (CPI):
  - Loans: 0.338 (0.187)
  - House Price: 2.890*** (8.449)
  - Additional estimates include House Price 1.700*** (3.670), Loans 1.484 (0.947), House Price -0.665 (-0.928), Loans -1.457 (-1.615), House Price 0.975* (1.695), Loans 0.212 (0.163), House Price 0.545 (1.251), Loans -2.015** (-2.169), House Price 0.535 (1.140)
- NDF (HK):
  - Loans: -1.911 (-1.037)
  - House Price: -2.857*** (-3.609)
  - Other significant estimates: Loans -1.415** (-2.470); House Price -0.841*** (-3.710); Loans -1.914*** (-2.600); House Price 0.232 (1.272); Loans -0.721* (-1.815); House Price -0.725*** (-2.621); Loans -1.098** (-2.036); House Price -0.928*** (-4.523); Loans -0.286 (-1.243); House Price -0.913*** (-4.367)
- VIX Index:
  - Loans: -0.019 (-0.086)
  - House Price: -0.537*** (-2.847)
  - Additional estimates show consistently negative and often significant house-price coefficients (e.g., -0.503*** (-3.193); -0.646*** (-3.828); -0.573*** (-3.293); -0.448*** (-3.407); -0.503*** (-3.369))
- Land supply area:
  - Loans: -0.036 (-1.375)
  - House Price: 0.076** (2.327)
  - Other estimates show House Price coefficients 0.084*** (2.592), 0.067** (2.080), 0.064*** (2.794), 0.069*** (3.456), 0.060*** (3.029)
- Provincial house price with one lag:
  - Loans: -0.027 (-0.916)
  - House Price: -0.056** (-2.035), and in other specifications: -0.123*** (-3.549); -0.160*** (-3.241); -0.173*** (-4.369); -0.122*** (-2.865)
- Policy variables (selected significant coefficients):
  - Reserve requirement ratio (Macroprudential policy): Loans -1.502 (-0.468); House Price -3.362*** (-2.751)
  - Bank rate (EOP) (Monetary policy): Loans -14.296* (-1.820); House Price 7.078*** (2.857)
  - House-related policy dummy: Loans 3.438 (1.323); House Price -4.693*** (-7.703)
  - Capital adequacy ratio: Loans -0.402*** (-4.942); House Price 0.112*** (3.810)
  - Liquidity ratio: Loans -0.275 (-1.199); House Price 0.139* (1.684)
  - Reserves for Impaired Loans/ Gross loans: Loans -4.286*** (-3.461); House Price -0.352 (-0.701)
- Sample and fit:
  - Number of banks varies by specification (e.g., 118.000, 171.000, 103.000, 147.000, etc.)
  - Number of observations varies by specification (e.g., 485; 1,437; 372; 534; 484; 710; 441; 649)
  - R-squared (overall) reported from 0.000 up to 0.125 depending on specification

### Large banks (Table 12 — Individual Variables Based on the Sample of Large Banks)
- Provincial GDP growth:
  - Loans: -2.387* (-1.936)
  - House Price: -2.101*** (-10.821)
  - Other specification results include Loans 0.731 (1.461); House Price -1.972*** (-78.646); Loans -2.451* (-1.933); House Price -0.950 (-1.111); Loans -1.865* (-1.695); House Price -0.605*** (-3.385); Loans -1.405 (-0.892); House Price -2.114*** (-2.656); Loans -2.083** (-2.455); House Price -1.156** (-2.310)
- Inflation (CPI):
  - Loans: 10.233* (1.760)
  - House Price: 0.868 (0.239)
  - Additional significant entries: House Price 2.211*** (12.947); House Price 3.586*** (3.052); Loans 2.271* (1.687); House Price 2.150** (2.359)
- Foreign reserve growth:
  - Loans: -0.393* (-1.869)
  - House Price: 0.741** (2.292)
  - Several specifications show significance for House Price (e.g., 0.385*** (10.381); 0.658*** (16.471); 0.517*** (3.540); 0.497*** (3.821); 0.552*** (28.603))
- NDF (HK):
  - Loans: -6.634** (-2.006)
  - House Price: -1.789 (-0.694)
  - Other significant entries include Loans -1.042*** (-2.918); House Price -2.712*** (-57.049); Loans -3.238** (-2.020); House Price -1.356* (-1.892); Loans -1.438* (-1.758); House Price -1.896*** (-4.483); Loans -1.138 (-1.376); House Price -3.246*** (-11.685); Loans -1.434** (-2.459); House Price -2.953*** (-7.313)
- VIX Index:
  - Loans: -0.454 (-1.093)
  - House Price: -0.741*** (-3.037)
  - Additional significant negative house-price coefficients (e.g., -0.394*** (-4.116); -0.458*** (-4.185); -0.655*** (-5.827); -0.745*** (-4.368); -0.579*** (-2.609))
- Land supply area:
  - Loans: 0.055* (1.862)
  - House Price: -0.035 (-0.520)
  - Other specifications show Loans 0.086** (2.291); Loans 0.088** (2.293); Loans 0.076** (2.352)
- Provincial house price with one lag:
  - Loans: 0.233 (1.504)
  - House Price: 0.184 (1.504)
  - Significant entry: Loans 0.042*** (3.113)
- Policy variables (selected significant coefficients):
  - Reserve requirement ratio: Loans -11.568* (-1.959); House Price 3.498 (0.496)
  - Bank rate (EOP): Loans -41.346** (-2.068); House Price 19.172* (1.843)
  - House-related policy dummy: Loans 7.083* (1.957); House Price -5.162** (-2.016)
  - Capital adequacy ratio: Loans -1.032 (-0.512); House Price 2.893*** (10.368)
  - Reserves for Impaired Loans/ Gross loans: Loans -11.988*** (-4.425); House Price -0.668 (-1.012)
- Sample and fit:
  - Number of banks: 5.000 across specifications
  - Number of observations: ranges e.g., 28; 50; 36; 38
  - R-squared (overall): ranges from 0.506 to 0.730 depending on specification

### Small- and medium-sized banks (Table 13 — Individual Variables Based on the Sample of Small- and Medium-Sized Banks)
- Results largely mirror the full sample patterns:
  - Provincial GDP growth: House Price coefficients often significant (e.g., 0.677** (2.453); 0.522* (1.681); 0.818** (2.217); 0.919*** (3.003); 0.928** (2.524))
  - Inflation (CPI): House Price 2.958*** (8.839) and 1.676*** (3.314); Loans and other estimates reported with the same numeric values as in the full sample table
  - NDF (HK): House Price -3.122*** (-3.215); Loans significant entries include -1.441** (-2.329)
  - VIX Index: House Price typically negative and significant (e.g., -0.526*** (-2.586); -0.518*** (-3.129); -0.654*** (-3.731); -0.597*** (-3.137); -0.473*** (-3.358); -0.542*** (-3.389))
  - Land supply area: House Price 0.077** (2.240); 0.085** (2.553); 0.070** (2.063); 0.068*** (2.782); 0.071*** (3.431); 0.061*** (2.957)
  - Provincial house price with one lag: House Price negative and significant in several specs (e.g., -0.066** (-2.362); -0.129*** (-3.608); -0.190*** (-3.514); -0.197*** (-4.683); -0.140*** (-3.031))
- Policy variables (selected significant coefficients):
  - Reserve requirement ratio: Loans -2.365 (-0.526); House Price -3.909** (-2.569)
  - Bank rate (EOP): Loans -13.436* (-1.663); House Price 6.350** (2.560)
  - House-related policy dummy: Loans 3.445 (1.293); House Price -4.529*** (-7.118)
  - Capital adequacy ratio: Loans -0.400*** (-4.835); House Price 0.099*** (4.170)
  - Reserves for Impaired Loans/ Gross loans: Loans -4.244*** (-3.469); House Price 0.582 (1.346)
- Sample and fit:
  - Number of banks: e.g., 113.000; 166.000; 98.000; 142.000; 108.000; 152.000
  - Number of observations: e.g., 457; 1,387; 344; 498; 456; 672; 413; 611
  - R-squared (overall): low in many specifications (e.g., 0.001 to 0.125)

### East region (Table 14 — Individual Variables Based on the Sample of Banks in the East Region)
- Provincial GDP growth: Loans 0.179 (0.273); House Price 0.419** (2.011); other specs include House Price 0.957** (2.292), 1.041** (2.105)
- Inflation (CPI): House Price coefficients often significant (e.g., 3.137*** (10.170); 1.705*** (8.531); 1.566*** (2.601); 1.453*** (3.460); 1.297*** (2.808)); some negative significant entries (e.g., -1.325*** (-3.627))
- NDF (HK): House Price -3.603*** (-6.311); other significant entries: Loans -1.452** (-2.125); House Price -1.147*** (-6.694); Loans -2.042** (-2.426); House Price 0.156 (0.730)
- VIX Index: House Price negative and significant across specifications (e.g., -0.441*** (-5.110); -0.414*** (-4.598); -0.564*** (-6.662); -0.645*** (-3.313); -0.526*** (-3.621); -0.570*** (-3.370))
- Land supply area: Loans negative and sometimes significant (e.g., -0.064* (-1.680)); House Price positive and significant (e.g., 0.080*** (6.178); 0.092*** (7.656); 0.121*** (3.331); 0.130*** (5.098); 0.118*** (4.280))
- Provincial house price with one lag: House Price negative and significant in multiple specs (e.g., -0.050** (-1.963); -0.126*** (-4.344); -0.130*** (-2.609); -0.148*** (-3.731); -0.093** (-2.159))
- Policy variables (selected significant coefficients):
  - Reserve requirement ratio: Loans -0.574 (-0.171); House Price -4.136*** (-4.026)
  - Bank rate (EOP): Loans -14.192 (-1.561); House Price 8.404*** (3.336)
  - House-related policy dummy: Loans 3.762 (1.335); House Price -5.610*** (-10.017)
  - Capital adequacy ratio: Loans -0.412*** (-5.305); House Price 0.098*** (3.594)
  - Reserves for Impaired Loans/ Gross loans: Loans -5.010*** (-3.677); House Price -0.552 (-1.047)
- Sample and fit:
  - Number of banks: e.g., 96.000; 130.000; 83.000; 109.000; 92.000; 119.000
  - Number of observations: e.g., 408; 1,121; 307; 411; 407; 562; 366; 505
  - R-squared (overall): ranges (e.g., 0.001 to 0.223)

### Middle region (Table 15 — Individual Variables Based on the Sample of Banks in the Middle Region)
- Loan dynamics show greater persistence and significance of lagged loan:
  - Loan with one lag: -0.606*** (-4.574); -0.601*** (-4.351); -0.625*** (-4.936); -0.338** (-2.081); -0.616*** (-4.674); -0.645*** (-4.629)
- Provincial house price with one lag:
  - House Price significant negative in several specs: -0.225** (-2.511); -0.314*** (-2.704); -0.288** (-2.403)
- NDF (HK) and VIX show mixed signs and occasional significance in some specifications (see numeric entries in table)
- House-related policy dummy:
  - Loans 5.453** (2.560); House Price -0.414 (-0.151)
- Reserves for Impaired Loans/ Gross loans:
  - Loans -5.794* (-1.792); House Price 4.631 (1.274)
- Sample and fit:
  - Number of banks: e.g., 12.000; 24.000; 11.000; 22.000
  - Number of observations: e.g., 35; 180; 29; 61; 35; 78
  - R-squared (overall): low to moderate across specs (e.g., 0.003 to 0.218)

### West region (Table 16 — Individual Variables Based on the Sample of Banks in the West Region)
- Land supply area:
  - House Price positive and often significant (e.g., 0.023** (2.108); 0.026** (2.505); 0.028*** (2.768); 0.023** (1.975); 0.029*** (3.341); 0.022 (1.607); 0.023* (1.806); 0.027** (2.088))
- Loan with one lag and provincial house price lag show negative significant effects in some specifications:
  - Loan with one lag: -0.207** (-2.310); -0.302*** (-2.820); -0.225* (-1.650); -0.256*** (-3.982)
  - Provincial house price with one lag: -0.122* (-1.857); -0.143** (-2.143); -0.172** (-2.215); -0.176 (-1.556); -0.170 (-1.494); -0.215** (-1.978)
- Policy variables (selected significant coefficients):
  - Reserve requirement ratio: Loans -17.239* (-1.739); House Price -1.671 (-0.952)
  - Bank rate (EOP): Loans -30.678** (-2.231); House Price 5.338 (1.060)
  - Liquidity ratio: Loans -0.214 (-0.538); House Price 0.384** (2.045)
  - Reserves for Impaired Loans/ Gross loans: Loans 5.026*** (6.423); House Price 2.075 (1.211)
- Sample and fit:
  - Number of banks: e.g., 10.000; 17.000; 9.000; 16.000
  - Number of observations: e.g., 42; 136; 36; 62; 42; 70; 41; 69
  - R-squared (overall): variable across specs (e.g., 0.000 to 0.180)

### Multivariate policies: grouped by size and regions (Table 17)
- Full sample multivariate estimates (Grouped by Size of Banks and Grouped by Regions) — selected coefficients:
  - NDF (HK): negative and significant for House Price in full-sample multivariate: -1.963*** (-2.882)
  - VIX Index: House Price -0.462*** (-2.702)
  - House-related policy dummy: House Price -4.328*** (-7.021)
  - Bank rate (EOP): House Price 6.707*** (3.344)
  - Reserve requirement ratio: House Price -4.160*** (-4.078)
  - Provincial house price with one lag: Loans -0.118*** (-3.775)
- Grouped results show heterogeneity across Large Banks, Small/Medium Banks, and regions (East, West, Middle) with varying significance and sign patterns (see numeric entries under each grouped column)
- Sample sizes and fit reported for each grouped specification (e.g., Number of banks 118.000 and 171.000 in full sample; R-squared (overall) value reported per column)

### Multivariate policies: subgroups in small- and medium-sized banks (Table 18)
- Subgroup estimates for City Commercial Banks, Foreign Banks, and other subgroups in the small/medium bank sample — selected significant results:
  - Reserve requirement ratio:
    - City Commercial Banks: 54.972*** (3.014) for Loans; House Price -11.759*** (-6.797)
    - Foreign Banks and other subgroups show varied signs and significance (e.g., -8.627; -1.919)
  - Bank rate (EOP):
    - City Commercial Banks: -110.685** (-2.147) for Loans; House Price 5.602 (1.570)
    - Some subgroups show Bank rate positive and significant for House Price (e.g., 17.532*** (6.157) in one subgroup)
  - House-related policy dummy:
    - City Commercial Banks: 13.099* (1.691) for Loans; House Price -2.854*** (-3.539)
    - Other subgroups: e.g., House-related policy dummy 8.000* (1.744) for Loans; House Price -6.415*** (-4.260)
  - NDF (HK):
    - City Commercial Banks: 21.705*** (3.187) for Loans; House Price -7.604*** (-6.644)
- Sample and fit:
  - Number of banks and observations vary by subgroup (e.g., City Commercial Banks: Number of banks 22.000; Number of observations 76; R-squared (overall) 0.142; Foreign Banks: Number of banks 29.000; Number of observations 284; R-squared (overall) 0.409; other subgroups reported with their sample sizes and R-squared)

*Source: IMF staff estimates. Note: Tables present panel fixed-effects regressions on policies and factors affecting credit (Loan) and house price in China between 2000 and 2011. Standard errors are shown under the coefficients. ***, **, and * denote statistical significance at the 1 percent, 5 percent, and 10 percent levels of confidence based on robust standard errors. For the definition of dependent variables and independent variables, see table 6. For sample banks, see annex I.*

### 2011. The results are presented for the three subgroups of banks in small- and medium-sized banks. Dependent variables a

### _wp1375 - 2011. The results are presented for the three subgroups of banks in small- and medium-sized banks. Dependent variables a

### Estimation setup and dependent variables
- Results presented for the three subgroups of banks in small- and medium-sized banks.
- Dependent variables:
  - annually loan growth in each bank
  - house price in each province
- House-related policy dummy exclude the monetary policies.
- VIX index and Reserve requirement ratio are in first-differences (d_).
- T-statistics are in parenthesis.
- ***, **, and * denote statistical significance at the 1 percent, 5 percent, and 10 percent level of confidence based on robust standard errors.

### Bank subgroup highlighted
- Joint -stock Commercial Banks

### Key methodological notes (explicit from source)
- Use of first-differences for VIX index and Reserve requirement ratio, indicated by prefix d_.
- Robust standard errors used for inference; significance denoted by ***, **, and * at 1 percent, 5 percent, and 10 percent levels respectively.
- House-related policy dummy excludes monetary policies (explicit exclusion).

*Source: _wp1375 - 2011. The results are presented for the three subgroups of banks in small- and medium-sized banks. Dependent variables a*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp1375.pdf_