## _wp09170

## Source details

**Canonical URL:** [_wp09170](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09170.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09170.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09170.pdf.json)

---

### I. INTRODUCTION AND SUMMARY
- Financial performance of Chinese banks:
  - Banks accumulated large profits, many have a relatively strong capital base, and their nonperforming loans are low.
  - Overall financial performance was comparable to banks in other large economies in 2007, and probably better in 2008, since Chinese banks were mostly insulated from the shock reverberating through the global financial system.
- Research approach:
  - Efficiency of Chinese banks analyzed using the stochastic production frontier approach on data spanning 2001–07.
  - Focus on banks’ use of deposits, capital, and labor to extend performing loans; loan quality controlled by focusing on net loans and correcting for nonperforming loans.
  - Hong Kong SAR banks included to make the production frontier more universally representative.
- Three key conclusions:
  - Chinese banks’ efficiency improved significantly during 2001–07; several joint stock commercial banks score high in efficiency, but a majority of banks remain quite inefficient, including several large state-owned banks and many city banks.
  - Large banks tend to attract too high level of deposits and operate at diminishing returns to scale; smaller banks operate at increasing returns to scale. Shifting deposits from larger to smaller banks would improve overall efficiency.
  - Banks’ profits are not correlated with their efficiency.

### II. FINANCIAL PERFORMANCE OF CHINESE BANKS
- Historical milestones:
  - Monobank system until the 1980s; big four banks carved out of the Peoples’ Bank of China in 1979 and 1984.
  - PBC designated central bank in 1983; first mortgage in 1985; Bank of Communications became first state-owned shareholding commercial bank in 1987; first foreign exchange forward in 1997 by Bank of China.
- 1990s reforms:
  - Lending/NPL cycle from aggressive directed lending led to a broad reform agenda in 1993 to allow banks to operate commercially.
  - Reforms included recapitalization, carving out bad assets, revamping supervision, and increasing foreign participation; IPOs of several large banks began in 2005.
- Recent indicators (as reported):
  - 2007 ROE of banking institutions: 16.7 percent.
  - 2007 ROA: 0.9 percent.
  - Profitability strengthened further in 2008 despite PBC rate cuts in the second half of 2008.
  - Listed banks’ CAR stood significantly higher than the minimum 8 percent requirement.
- Asset quality and capital adequacy:
  - NPL ratio declined from more than 20 percent in 2002 to 6.2 percent by end-2007.
  - NPL ratio declined to 2.5 percent by end-2008 (latter decline mostly reflecting the restructuring of Agricultural Bank of China).

### III. COMPARATIVE FINANCIAL PERFORMANCE (large banks, 2007)
- Return on average assets:
  - Chinese banks: higher than the G7.
  - Among BRIC, large banks in Brazil and Russia performed even better.
  - Correlation of returns with net interest margin: correlation coefficient 85 percent.
  - Banks in Hong Kong SAR had a higher return on asset despite narrower interest margin.
- Capital-asset ratio (CAR):
  - Average CAR in China: 13.4 percent.
  - Average CAR in G7 and BRIC countries: 12.3 percent.
  - Lowest average CAR in sample: Italy’s large banks at 9.7 percent.
- Nonperforming loan (NPL) ratio:
  - China (2007): 4.2 percent.
  - G7 (2007): 2.3 percent.
  - Brazil and Russia (2007): 6.7 percent.
  - Note: Chinese banks’ NPL ratio declined to the G7 level in 2008.
- Income and cost structure:
  - Interest earnings share of operating income: large Chinese banks 90 percent; G7 and BRIC about 70 percent.
  - Loans/Deposits ratio for large Chinese banks: 56 percent.
  - Fixed (overhead) costs as share of total costs: China 30 percent; G7 17 percent.
  - Net fee and commission income among listed Chinese banks in 2007: 12 percent of total bank income.

### IV. MODELING BANK EFFICIENCY (approach and definitions)
- Core intermediation focus: receiving deposits and making loans.
- Inputs: fixed assets (k), labor (l), equity (e), deposits (d).
- Output: total loans net of non-performing loans (nl).
- Rationale for netting NPLs: proxy for banks’ ability to resolve information asymmetries (example: adjusting Agricultural Bank of China’s lending lowers loan size by 18 percent).
- Stochastic frontier specification:
  - nl_it = p(k_it, l_it, e_it, d_it) + u_it – v_it
  - u_it ~ N(0, σ_u^2)
  - v_it ~ Truncated N^+(m_it, σ_v^2)
  - Inefficiency mean: m_it = m_0 + m_1 t + m_2 qm_i, where t = time trend, qm_i = proxy for quantitative macro measures.
- Production function form: translog
  - ln(nl_it) = a_0 + Σ a_j ln(x_jit) + 1/2 ΣΣ a_hj ln(x_hit) ln(x_jit) + u_it – v_it
- Returns to scale:
  - rts_it = Σ ∂ln(nl_it)/∂ln(x_jit)

### V. DATA AND ESTIMATION
- Panel: 7 years (2001–07) and 69 banks; data from Bankscope.
- Exclusions: banks lacking critical variables, policy banks, trust banks, Postal Bank not included.
- Hong Kong SAR banks: 8 large banks included; Hong Kong dummy in efficiency mean.
- Variable definitions and transformations:
  - Net loans: loan loss reserves netted from total loans; if loan loss reserves < 80 percent of NPLs, difference subtracted from total loans.
  - Deflators: net loans deflated with GDP deflator; fixed assets with fixed asset investment deflator; equity with GDP deflator.
  - All variables in natural logs.
- Estimation method: Maximum likelihood method of Battese and Coelli (1995).

### VI. COEFFICIENTS AND KEY ESTIMATION RESULTS (selected)
- Selected production-function coefficients (dependent variable: net loans; all variables in natural logs):
  - constant: coefficient -0.54, T-stats -0.63
  - fixed assets (x1): coefficient -0.48, T-stats -1.41
  - equity (x2): coefficient 0.20, T-stats 0.77
  - labor (x3): coefficient 0.51, T-stats 2.06
  - deposits (x4): coefficient 0.60, T-stats 1.94
  - x1^2: coefficient -0.09, T-stats -2.14
  - x4^2: coefficient -0.26, T-stats -4.44
  - x2*x3: coefficient 0.23, T-stats 4.18
  - x2*x4: coefficient -0.26, T-stats -2.74
  - x3*x4: coefficient -0.31, T-stats -4.44
- Variables in the (negative) efficiency term:
  - constant: 0.10, T-stats 0.69
  - time trend: -0.07, T-stats -1.79
  - proxy for window guidance: 0.05, T-stats 0.54
  - proxy for Hong Kong banks: -1.57, T-stats -1.26
- Variance parameters:
  - sigma-squared: 0.07, T-stats 2.26 (σ_u^2 + σ_v^2)
  - gamma: 0.861, T-stats 10.26 (σ_v^2/(σ_u^2 + σ_v^2))
- Interpretation of estimation:
  - Gamma close to 1 implies most variation attributable to inefficiency (σ_v^2 significantly larger than σ_u^2).
  - Time trend negative and statistically significant: efficiency improved during the sample period.
  - Proxy for quantitative macroeconomic policies positive (financial repression may hinder efficiency) but not statistically significant.
  - Hong Kong bank proxy negative (Hong Kong environment helpful for efficiency) but not statistically significant.

### VII. EFFICIENCY LEVELS AND CROSS-SECTIONAL PATTERNS
- Average technical efficiency (excluding idiosyncratic factors): about 90 percent.
- Comparisons:
  - Somewhat below banks in India (Ray (2007)); about the same as some banks in Europe (Brissimis et al (2008)); above banks in South Africa and Turkey (cited studies).
- Variation across banks:
  - Most efficient banks: about 98 percent.
  - Least efficient banks: as low as 60 percent.
  - Five largest state commercial banks’ efficiency: about average.
  - Agricultural Bank of China: consistently the least efficient state commercial bank.
  - Several joint stock commercial banks (Merchants Bank, Hua Xia Bank, Shanghai Pudong Development Bank, CITIC Bank) significantly more efficient and some as efficient as Hong Kong banks.
  - Many city banks were as inefficient or worse than Agricultural Bank of China, though some city banks (Hangzhou, Huzhou) were among the most efficient.
- Time trend caveat:
  - Average efficiency improved over time; improvement may be biased upward due to reporting selection.

### VIII. INEFFICIENCY MAGNITUDES AND INPUT ADJUSTMENT IMPLICATIONS
- 2006 estimate of potential loan increases from eliminating inefficiency (no change in technology or input levels):
  - Mainland Chinese banks on average could have increased net loans by 15 percent.
  - For large state commercial banks: 13.8 percent.
  - Excluding Agricultural Bank of China: average remains at 12 percent.
- Input-adjustment interpretations (Table 4 summary):
  - Distances represent percent declines in each input (keeping others constant) to eliminate inefficiency; radial distance is proportional percent change in all inputs jointly.
  - Large banks tend to have excess deposits given their level of loans.
  - Banks with more than Y 100 billion deposits: can reach the frontier by reducing deposits by on average 12.5 percent.
  - Smaller banks: can reach the frontier by reducing only deposits by 22 percent or reducing all inputs by 12 percent.
  - Marginal benefit of a 1 percentage point reduction in deposits is much larger in larger banks than in smaller banks.
- Time series of inputs (2001–07):
  - Ratio of fixed assets and labor to total loans declined steadily.
  - Ratio of deposits remained broadly flat and increased in 2007.
  - Share of equity increased.

### IX. RETURNS TO SCALE AND DEPOSIT REALLOCATION SIMULATION
- Average returns to scale: 0.97 (close to constant returns to scale).
- Estimated returns to scale across banks: between 0.84 and 1.09.
- Size pattern (2006 example):
  - Larger banks on average operated at diminishing returns to scale.
  - Smaller banks operated at increasing returns to scale.
- Policy simulation (deposit reallocation):
  - Scenario: reduce deposits of banks operating below the industry average return to scale by 10 percent and redistribute those deposits proportionally to remaining banks (everything else constant).
  - Resulting effects:
    - Other banks’ deposits increase by on average 25 percent.
    - Total loans of the banking system increase by 2.6 percent.
  - Relationship nonlinear: as more deposits shift from diminishing-RTS banks to increasing-RTS banks, loan origination gains are larger.

### X. PROFITABILITY DRIVERS: THREE HYPOTHESES AND EVIDENCE
- Core puzzle: Chinese banks are very profitable despite wide inefficiencies; efficiency does not predict profitability once Hong Kong banks are excluded (correlation close to zero for Chinese banks).
- Hypothesis 1 — Financial repression:
  - Deposit interest rates constrained by benchmark ceilings; loan rates constrained by benchmark floors; both largely binding.
  - H1 2008 loan-deposit spread: 4.8 percent.
  - Sterilization operations, window guidance, and administrative loan restrictions limit lending volumes and return on assets.
  - Implicit tax on household savings estimated at 4 percent of GDP; about a quarter of this is captured by banks — equal to about half the after-tax profits of the banking system in 2008 (Lardy (2008) estimate cited).
- Hypothesis 2 — Market structure (segmentation and concentration):
  - Banking system shares in 2008 (In percent of total, monetary survey):
    - State commercial banks (SCBs): Assets 48.4, Loans 43.2, Deposits 53.3
    - Joint stock commercial banks (JSCBs): Assets 18.4, Loans 19.4, Deposits 17.5
    - Policy banks: Assets 9.0, Loans 15.6, Deposits 0.9
    - City commercial banks: Assets 6.5, Loans 7.0, Deposits 7.5
    - Urban credit cooperatives: Assets 0.1, Loans 0.1, Deposits 0.2
    - Rural credit cooperatives: Assets 7.8, Loans 6.0, Deposits 9.7
    - Foreign banks: Assets 2.1, Loans 2.5, Deposits 0.9
    - Finance companies: Assets 1.5, Loans 1.9, Deposits 1.7
    - Postal saving bank: Assets 3.4, Loans 1.4, Deposits 4.9
  - HHI measures (selected series):
    - All financial institutions (Loans): 2001 0.11; 2007 0.07.
    - All financial institutions (Deposits): 2001 0.16; 2007 0.10.
    - Banks competing in urban areas (Loans): 2001 0.18; 2007 0.11.
    - Banks competing in urban areas (Deposits): 2001 0.17; 2007 0.12.
  - Interpretation:
    - HHI for whole banking system in 2007: 0.07 (low concentration).
    - Revised HHI for banks competing in similar urban markets (HHI-R) in 2007: 0.11 (moderate concentration); HHI-R was 0.18 in 2001.
    - Lack of entry/exit among large/medium banks over the last 10 years suggests high barriers and conditions conducive to oligopolistic behavior.
  - Jiangsu Province Cournot-style comparative statics (2000–07) summary (trend loan growth and correlations):
    - Total, short term: Trend growth SCBs 6.7, Non-SCBs 19.9, Correlation of deviations 62.8, Correlation with output growth 23.6
    - Industry: SCBs 14.8, Non-SCBs 25.9, Correlation of deviations 4.1, Correlation with output growth 39.6
    - Commerce: SCBs -7.1, Non-SCBs 9.7, Correlation of deviations -25.8, Correlation with output growth 63.1
    - Construction: SCBs 10.5, Non-SCBs 41.1, Correlation of deviations -48.9, Correlation with output growth 11.5
    - Agricultural: SCBs -14.9, Non-SCBs 25.9, Correlation of deviations 30.8, Correlation with output growth -36.4
    - Foreign joint ventures: SCBs -1.4, Non-SCBs 8.3, Correlation of deviations -33.0, Correlation with output growth 12.4
    - Private enterprises and individuals: SCBs 31.4, Non-SCBs 55.9, Correlation of deviations 19.3, Correlation with output growth -19.8
    - Others: SCBs 15.1, Non-SCBs 23.3, Correlation of deviations 83.3, Correlation with output growth -33.2
  - Evidence interpretation:
    - Competition relatively strong in commerce and foreign joint ventures (negative correlation of deviations or sign switches).
    - Oligopolistic behavior likely in lending to private companies and individuals and parts of services (“others”) where SCBs and non-SCBs move in tandem while lending is negatively correlated with output.
    - For industry, construction, agriculture the evidence is inconclusive due to demand correlations.
- Hypothesis 3 — Cyclical position:
  - Loan growth versus nominal GDP and fixed asset investment (FAI):
    - During 2004–07, loan growth was lower than nominal GDP growth and lower relative to FAI growth.
    - 2007: loan growth particularly limited relative to GDP and FAI.
  - Conclusion: cyclical position not a major factor behind large profits of recent years.

### XI. BANKS’ LENDING, PROFITABILITY, AND MARGINS (2008–2009)
- Reported metric: 16.7 percent while nominal GDP grew 21.4 percent.
- During accelerating GDP growth, there was no significant upward trend in banks’ loans.
- Late 2008 and Q1 2009:
  - Despite sharp deceleration in economic activity in Q4 2008, banks continued to make profits and sharply increased lending; pattern continued in Q1 2009.
  - Interest margin narrowed marginally during this period.
  - Higher lending expected to more than compensate for the decline in interest margin, provided non-performing loans do not increase.
- Interpretation:
  - These patterns suggest strong bank profits were not strongly linked to the cyclical upturn in economic activity.

### XII. POLICY CONCLUSIONS AND REFORM AGENDA
- Main recommendation:
  - Eliminating various controls over the financial system, while maintaining effective bank supervision, could significantly improve financial intermediation.
- Specific policy options highlighted:
  - Lifting the ceiling on deposit rates could:
    - Facilitate movement of deposits from large to smaller banks.
    - Allow smaller banks (more efficient at turning deposits into performing loans) to compete for funds.
    - Lower market concentration and boost competition among banks.
  - Uncertainty noted:
    - It is unclear whether higher deposit rates will increase household savings given the strong precautionary savings motive.
- Risks and recommended complementary measures:
  - Increased efficiency in intermediation could raise loan supply given the large deposit base; this needs careful macroeconomic and financial stability monitoring.
  - Complementary measures recommended:
    - Lower liquidity in the system through indirect monetary policy.
    - Channel some savings to other capital markets.
    - Ensure regulation/supervision is sufficiently strong to contain reckless loan growth.

*Source document content from _wp09170 - References.*

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

### _wp09170 - References

### I. INTRODUCTION AND SUMMARY
- Financial performance of Chinese banks has been impressive:
  - Banks have been accumulating large profits, many have a relatively strong capital base, and their nonperforming loans are low.
  - Overall financial performance was comparable to banks in other large economies in 2007, and probably better in 2008, since Chinese banks were mostly insulated from the shock reverberating through the global financial system (Section II).
  - These results reflect years of financial sector reforms, including restructuring and recapitalization of China’s largest banks and an increase in foreign participation.
- Research approach to bank efficiency:
  - The paper analyzes how efficiently Chinese banks use deposits, capital, and labor to extend performing loans using the stochastic production frontier approach applied to data spanning 2001–07 (Section III).
  - Loan quality is controlled by focusing on net loans and correcting for nonperforming loans.
  - Hong Kong SAR banks are included to have a more universally representative production frontier.
- Three key conclusions:
  - Chinese banks’ efficiency improved significantly during 2001–07, and a number of joint stock commercial banks score high in efficiency. However, a majority of banks remain quite inefficient, including several large state-owned banks and many city banks.
  - Large banks tend to attract too high level of deposits and operate at diminishing returns to scale. Smaller banks operate at increasing returns to scale. Shifting deposits from larger to smaller banks would improve the banking system’s overall efficiency.
  - Banks’ profits are not correlated with their efficiency.
- Investigation of reasons for large profits despite inefficiencies (Section IV):
  - Financial repression: Guarantees low cost of funds and a large interest margin for banks, but limits the scope for banks to increase loans. Net impact should raise profits, but likely not enough to explain all profits in the Chinese banking system.
  - Oligopolistic market structure: Market concentration in the banking system is high when measured appropriately, which could foster collusive behavior in lending in some market segments and support large profits.
  - Cyclical position: China’s cyclical position was strong during most of the sample period, but the link between output growth and loan growth was weak, suggesting the economic cycle was not a major factor behind the large profits of recent years.

### II. FINANCIAL PERFORMANCE OF CHINESE BANKS
- Historical evolution:
  - Until the 1980s, China had a monobank system with the Peoples’ Bank of China (PBC) handling virtually all banking activities.
  - The big four banks (Agricultural Bank of China, China Construction Bank, Industrial and Commercial Bank of China, and Bank of China) were carved out of the PBC in 1979 and 1984.
  - The PBC was designated as the central bank of China for the first time in 1983.
  - The first mortgage was issued in 1985.
  - Bank of Communications became the first state-owned shareholding commercial bank in 1987.
  - The first foreign exchange forward was written in 1997 by the Bank of China.
- 1990s challenges and reforms:
  - The 1990s saw a major lending and nonperforming loans cycle due to aggressive directed lending to industry.
  - A broad reform agenda was established in 1993 to allow banks to operate on a commercial footing (the State Council (1993)).
  - Reforms implemented since then include recapitalization of major banks, carving out bad assets, revamping bank supervision, and increasing foreign banks’ participation in domestic markets.
  - IPOs of several large banks began in 2005 and bank ownership structures began to change with board transformations.
- Recent financial indicators:
  - In 2007 the ROE of banking institutions reached 16.7 percent and the ROA was 0.9 percent.
  - Profitability strengthened even more in 2008 even though the PBC cut benchmark rates in the second half of the year.
  - Capital adequacy ratio (CAR) further improved in 2007, and CAR of all banks except one reached or exceeded 8 percent by end- (text truncated).

*Italic: Source document content from _wp09170 - References.*

### 2008. Listed banks’ CAR stood significantly higher than the minimum 8 percent

### 2008. Listed banks’ CAR stood significantly higher than the minimum 8 percent

### Asset quality and capital adequacy
- Listed banks’ CAR stood significantly higher than the minimum 8 percent requirement.
- NPL ratio developments:
  - Declined from more than 20 percent in 2002 to 6.2 percent by end-2007.
  - Declined to 2.5 percent by end-2008 (the latter decline mostly reflecting the restructuring of Agricultural Bank of China).

### Comparative financial performance (large banks, 2007 period before the global crisis)
- Return on average assets:
  - Chinese banks: higher than the G7.
  - Among BRIC, large banks in Brazil and Russia performed even better.
  - Correlation of returns with net interest margin: correlation coefficient was 85 percent.
  - Banks in Hong Kong SAR had a higher return on asset, even though their interest margin was narrower.
- Capital-asset ratio (CAR):
  - Average CAR in China: 13.4 percent.
  - Average CAR in G7 and BRIC countries: 12.3 percent.
  - Lowest average CAR among sample economies: Italy’s large banks at 9.7 percent.
  - Average CARs in all economies in the sample were high compared to the minimum required 8 percent.
- Nonperforming loan (NPL) ratio:
  - China (2007): 4.2 percent.
  - G7 (2007): 2.3 (percent).
  - Brazil and Russia (2007): 6.7 percent.
  - Note: The NPL ratio of these Chinese banks declined to the G7 level in 2008.
- Income and cost structure:
  - Interest earnings share of operating income:
    - Large Chinese banks: 90 percent.
    - Average for G7 and BRIC: about 70 percent.
  - Loans/Deposits:
    - Large Chinese banks’ loan to deposit ratio: 56 percent.
  - Fixed (overhead) costs as share of total costs:
    - China: 30 percent.
    - G7: 17 percent.
  - Labor cost share in total cost: "not out of line" with other countries (no numeric change).

### Core findings on banking activity structure
- Primary business: provision of loans to the corporate sector; fee-based businesses are small (net fee and commission income amounted to 12 percent of total bank income among the listed Chinese banks in 2007).
- Deposit services often charged little, if at all, to the public.
- Large banks lend primarily to large corporations and much less to consumers.
- Large fixed-cost share implies scale and operational structure differences relative to G7.

### Modeling bank efficiency (approach and definitions)
- Focus: banks’ core intermediation activities — receiving deposits and making loans.
- Inputs in production function: fixed assets (k), labor (l), equity (e), deposits (d).
- Output: total loans net of non-performing loans (nl).
- Rationale for netting NPLs: proxy for banks’ ability to resolve information asymmetries; example—adjusting Agricultural Bank of China’s lending lowers loan size by 18 percent.
- Deposits treated as intermediate input (intermediation view).
- Stochastic frontier specification:
  - nl_it = p(k_it, l_it, e_it, d_it) + u_it – v_it
  - u_it ~ N(0, σ_u^2) (idiosyncratic stochastic error)
  - v_it ~ Truncated N^+(m_it, σ_v^2) (non-negative inefficiency)
  - Inefficiency mean: m_it = m_0 + m_1 t + m_2 qm_i, where t = time trend, qm_i = proxy for quantitative macro measures.
- Production function form: translog
  - ln(nl_it) = a_0 + Σ a_j ln(x_jit) + 1/2 ΣΣ a_hj ln(x_hit) ln(x_jit) + u_it – v_it
- Returns to scale measure:
  - rts_it = Σ ∂ln(nl_it)/∂ln(x_jit)

### Data and estimation
- Panel: 7 years (2001–07) and 69 banks, data from Bankscope.
- Exclusions: banks lacking critical variables, policy banks, trust banks, Postal Bank (not included in dataset).
- Hong Kong SAR banks: 8 large banks included; a dummy = 1 for Hong Kong banks is included in efficiency mean.
- Variables:
  - Net loans (loan loss reserves netted from total loans; if loan loss reserves < 80 percent of NPLs, difference subtracted from total loans).
  - Net loans deflated with GDP deflator; fixed assets deflated with fixed asset investment deflator; equity with GDP deflator.
  - All variables in natural logs.
- Estimation method: Maximum likelihood method of Battese and Coelli (1995).

### Coefficient and key estimation results (Table 2 highlights)
- Selected production-function coefficients and t-stats (dependent variable: net loans; all variables in natural logs):
  - constant: coefficient -0.54, T-stats -0.63
  - fixed assets (x1): coefficient -0.48, T-stats -1.41
  - equity (x2): coefficient 0.20, T-stats 0.77
  - labor (x3): coefficient 0.51, T-stats 2.06
  - deposits (x4): coefficient 0.60, T-stats 1.94
  - x1^2: -0.09, T-stats -2.14
  - x4^2: -0.26, T-stats -4.44
  - x2*x3: 0.23, T-stats 4.18
  - x2*x4: -0.26, T-stats -2.74
  - x3*x4: -0.31, T-stats -4.44
- Variables in the (negative) efficiency term:
  - constant: 0.10, T-stats 0.69
  - time trend: -0.07, T-stats -1.79
  - proxy for window guidance: 0.05, T-stats 0.54
  - proxy for Hong Kong banks: -1.57, T-stats -1.26
- Variance parameters:
  - sigma-squared: 0.07, T-stats 2.26 (σ_u^2 + σ_v^2)
  - gamma: 0.861, T-stats 10.26 (σ_v^2/(σ_u^2 + σ_v^2))
- Interpretation:
  - Gamma close to 1 implies most variation attributable to inefficiency (σ_v^2 significantly larger than σ_u^2).
  - Time trend negative and statistically significant: efficiency improved during the sample period.
  - Proxy for quantitative macroeconomic policies positive (financial repression may hinder efficiency) but not statistically significant.
  - Hong Kong bank proxy negative (Hong Kong environment helpful for efficiency) but not statistically significant.

### Efficiency levels and cross-sectional patterns
- Average technical efficiency (excluding idiosyncratic factors): about 90 percent.
- Comparisons:
  - Somewhat below banks in India (Ray (2007)).
  - About the same as some banks in Europe (Brissimis et al (2008)).
  - Above banks in South Africa and Turkey (cited studies).
- Variation across banks:
  - Most efficient banks: efficiency level about 98 percent.
  - Least efficient banks: efficiency levels as low as 60 percent.
  - Five largest state commercial banks’ efficiency: about average.
  - Agricultural Bank of China: consistently remained the least efficient state commercial bank.
  - Several joint stock commercial banks (Merchants Bank, Hua Xia Bank, Shanghai Pudong Development Bank, CITIC Bank) were significantly more efficient and some as efficient as Hong Kong banks.
  - Many city banks were as inefficient or worse than Agricultural Bank of China, though some city banks (Hangzhou, Huzhou) were among the most efficient.
- Time trend:
  - Average efficiency improved over time; improvement may be biased upward due to reporting selection.

### Inefficiency magnitudes and input adjustment implications
- 2006 estimate: mainland Chinese banks could have increased net loans on average by 15 percent without changing technology or input levels.
  - For large state commercial banks: 13.8 percent.
  - Excluding Agricultural Bank of China: average remains at 12 percent.
- Table 4 interpretation (distance from frontier):
  - Distances represent percent declines in each input (keeping others constant) to eliminate inefficiency; radial distance is proportional percent change in all inputs jointly.
  - Large banks tend to have excess deposits given their level of loans.
  - Banks with more than Y 100 billion deposits: can reach frontier by reducing deposits by on average 12.5 percent.
  - Smaller banks: can reach frontier by reducing only deposits by 22 percent or reducing all inputs by 12 percent.
  - Marginal benefit of a 1 percentage point reduction in deposits is much larger in larger banks than in smaller banks.
- Time series of inputs as share of output (Figure 5):
  - 2001–07: ratio of fixed assets and labor to total loans declined steadily; ratio of deposits remained broadly flat and increased in 2007.
  - Share of equity increased.

### Returns to scale and deposit reallocation simulation
- Average returns to scale: 0.97 (close to constant returns to scale).
- Estimated returns to scale vary between 0.84 and 1.09 across banks.
- Size pattern (2006 example):
  - Larger banks on average operated at diminishing returns to scale.
  - Smaller banks operated at increasing returns to scale.
- Policy simulation:
  - Reducing deposits of banks operating below the industry average return to scale by 10 percent and redistributing those deposits proportionally to remaining banks (everything else constant):
    - Other banks’ deposits increase by on average 25 percent.
    - Total loans of the banking system increase by 2.6 percent.
  - As more deposits shift from diminishing-RTS banks to increasing-RTS banks, loan origination gains are larger (nonlinear relationship).

### Profitability drivers: three hypotheses and evidence
- Chinese banks are very profitable, yet inefficiencies are wide and efficiency does not predict profitability once Hong Kong banks are excluded (correlation between efficiency and profitability close to zero for Chinese banks).
- Three potential explanatory factors:
  1. Financial repression
     - Deposit interest rates constrained by benchmark ceilings, and loan rates constrained by benchmark floors; both largely binding.
     - H1 2008 loan-deposit spread: 4.8 percent (Chu and Wen (2008)).
     - Sterilization operations, window guidance, and administrative loan restrictions limit lending volumes and return on assets.
     - Implicit tax on household savings estimated at 4 percent of GDP (Lardy (2008)); about a quarter of this is captured by banks — equal to about half the after-tax profits of the banking system in 2008 (Lardy (2008) estimate presented).
  2. Market structure (segmentation and concentration)
     - Bank categories and 2008 shares (In percent of total, monetary survey):
       - State commercial banks (SCBs): Assets 48.4, Loans 43.2, Deposits 53.3
       - Joint stock commercial banks (JSCBs): Assets 18.4, Loans 19.4, Deposits 17.5
       - Policy banks: Assets 9.0, Loans 15.6, Deposits 0.9
       - City commercial banks: Assets 6.5, Loans 7.0, Deposits 7.5
       - Urban credit cooperatives: Assets 0.1, Loans 0.1, Deposits 0.2
       - Rural credit cooperatives: Assets 7.8, Loans 6.0, Deposits 9.7
       - Foreign banks: Assets 2.1, Loans 2.5, Deposits 0.9
       - Finance companies: Assets 1.5, Loans 1.9, Deposits 1.7
       - Postal saving bank: Assets 3.4, Loans 1.4, Deposits 4.9
     - Herfindahl-Hirschman Index (HHI) measures:
       - All financial institutions (Loans): 2001 0.11, 2002 0.10, 2003 0.08, 2004 0.08, 2005 0.08, 2006 0.07, 2007 0.07.
       - All financial institutions (Deposits): 2001 0.16, 2002 0.14, 2003 0.12, 2004 0.11, 2005 0.10, 2006 0.10, 2007 0.10.
       - Banks competing in urban areas (Loans): 2001 0.18, 2002 0.15, 2003 0.13, 2004 0.12, 2005 0.12, 2006 0.11, 2007 0.11.
       - Banks competing in urban areas (Deposits): 2001 0.17, 2002 0.15, 2003 0.14, 2004 0.13, 2005 0.13, 2006 0.12, 2007 0.12.
     - Interpretation:
       - HHI for whole banking system in 2007: 0.07 (low concentration).
       - Revised HHI for banks competing in similar urban markets (HHI-R): 0.11 in 2007 (moderate concentration).
       - HHI-R was 0.18 in 2001, indicating higher concentration early in the decade.
       - Lack of entry/exit among large/medium banks over the last 10 years suggests high barriers and conditions conducive to oligopolistic behavior.
     - Cournot-style comparative statics using Jiangsu Province loan data (2000–07):
       - Table 8 summary (Trend growth columns are trend loan growth; correlation columns compare deviations and output growth):
         - Total, short term: Trend growth SCBs 6.7, Non-SCBs 19.9, Correlation of deviations 62.8, Correlation with output growth 23.6
         - Industry: SCBs 14.8, Non-SCBs 25.9, Correlation of deviations 4.1, Correlation with output growth 39.6
         - Commerce: SCBs -7.1, Non-SCBs 9.7, Correlation of deviations -25.8, Correlation with output growth 63.1
         - Construction: SCBs 10.5, Non-SCBs 41.1, Correlation of deviations -48.9, Correlation with output growth 11.5
         - Agricultural: SCBs -14.9, Non-SCBs 25.9, Correlation of deviations 30.8, Correlation with output growth -36.4
         - Foreign joint ventures: SCBs -1.4, Non-SCBs 8.3, Correlation of deviations -33.0, Correlation with output growth 12.4
         - Private enterprises and individuals: SCBs 31.4, Non-SCBs 55.9, Correlation of deviations 19.3, Correlation with output growth -19.8
         - Others: SCBs 15.1, Non-SCBs 23.3, Correlation of deviations 83.3, Correlation with output growth -33.2
       - Evidence:
         - Competition relatively strong in commerce and foreign joint ventures (negative correlation of deviations or sign switches).
         - Oligopolistic behavior likely in lending to private companies and individuals and parts of services (“others”) where SCBs and non-SCBs move in tandem while lending is negatively correlated with output.
         - Industry, construction, agriculture: inconclusive due to demand correlations.
  3. Cyclical position
     - Loan growth versus nominal GDP and fixed asset investment (FAI):
       - During 2004–07, loan growth was lower than nominal GDP growth and lower relative to FAI growth.
       - 2007: loan growth was particularly limited relative to GDP and FAI (Figure 8 shows divergent series through 2008 but no numeric series beyond qualitative statement).

*Source: Author's calculations and excerpts from the provided content.*

### 16.7 percent while nominal GDP grew

### _wp09170 - 16.7 percent while nominal GDP grew

### Banks’ lending, profitability, and margins (2008–2009)
- Reported figure: 16.7 percent while nominal GDP grew 21.4 percent.
- During the period of accelerating GDP growth, there was no significant upward trend in banks’ loans.
- Late 2008 and early 2009:
  - Despite the very sharp deceleration in economic activity in Q4 2008, banks continued to make profits, and sharply increased their lending.
  - This pattern continued in Q1 2009.
  - Interest margin narrowed marginally during this period.
  - Higher lending is expected to more than compensate for this decline in the interest margin, provided non-performing loans do not increase.
- Interpretation: These factors together suggest that strong profits of banks was not strongly linked to the cyclical upturn in economic activity.

### Efficiency, market structure, and constraints
- Despite good recent financial performance, Chinese banks do not intermediate particularly efficiently.
- Large profits reflect:
  - More market concentration in the banking sector.
  - Administratively determined low deposit rates and large interest margins.
- Banks’ loan growth is quantitatively or administratively constrained, contributing to:
  - Inefficient financial intermediation.
  - Suppressed profits.
  - Little incentives for banks to improve efficiency.

### Policy conclusions and reform agenda
- Eliminating various controls over the financial system, while maintaining effective bank supervision, could significantly improve financial intermediation.
- Lifting the ceiling on deposit rates could:
  - Facilitate movement of deposits from large to smaller banks.
  - Allow smaller banks, which are more efficient in turning deposits into performing loans, to compete for funds.
  - Help lower market concentration and boost competition among banks.
- Uncertainty: Just as lower deposit rates have not lowered household savings, it is unclear whether higher deposit rates will increase savings, particularly given the strong precautionary savings motive behind household deposits.
- Risks and complementary measures:
  - Increased efficiency in bank intermediation could increase the supply of loans given the large deposit base, which will need to be monitored carefully both from a macroeconomic and financial stability perspective.
  - Recommended complementary measures:
    - Lower liquidity in the system through indirect monetary policy.
    - Channel some of the savings to other capital markets.
    - Ensure regulation/supervision is sufficiently strong to contain reckless loan growth.

*Source: _wp09170 - 16.7 percent while nominal GDP grew (IMF PDF chapter/section).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09170.pdf_