• Aucun résultat trouvé

Intensive margin: change in firm-level loan rate

3.6 Robustness check

3.6.2 Intensive margin: change in firm-level loan rate

Under the interest rate control regime in force during our sample period, China’s central bank sets a binding benchmark interest rate, so commercial banks cannot charge arbi-trarily high loan interest rates as risk premiums. Rational, profit-maximizing banks thus should take more deposits and lend to safer borrowers to earn the interest margin, which

is a no-loss business strategy. Due to this regime, the bank ownership reform is more likely to have affected the SOBs’ lending behavior by causing them to ration credit rather than charge more heterogeneous loan rates. Table 3.9 tests this conjecture by exploring the change in the firm-level loan rate caused by the bank ownership reform.

Since the ASIF does not report outstanding loans to allow for the calculation of interest rates, we only use listed companies for this analysis. Although the RESSET reports loan debt, it only reports total cash paid for distribution of dividends, profits, and interest expenses. Hence, the ratio of this variable to outstanding loans overestimates firm-level loan rates. Table3.1 shows that the mean level of the firm-level loan rate (interestt/loant) is high, at 14.8%, even after we drop observations in the bottom and top 5%. However, since column (1) in table 3.7 already shows that the bank ownership reform did not reduce or increase firm profitability for more-indebted listed companies, using the measure interestt/loant can at least provide some suggestive evidence on how the firm-level loan rate changed.

Column (1) shows that the bank ownership reform had little average effect on firm-level loan rates. The nonsignificant positive coefficient of Big Three Exposurec,t is consistent with column (1) in table 3.2 in showing that there is a nonsignificant negative effect on corporate credit access. Column (2) appears to suggest that more-indebted firms experienced an increase in the loan rate after the bank ownership reform. However, the coefficient of Big Three Exposurec,t×Debt ratioi,t−1 is statistically nonsignificant and has a small economic magnitude, indicating that loan rate adjustment was quite negligible.

Column (3) further controls for firm profitability (EBTi,t/assetsi,t1), as our measure of the firm-level loan rate contains profit distribution. Although firm profitability is indeed positively correlated with our dependent variable, the result for the interaction term Big Three Exposurec,t × Debt ratioi,t1 changes modestly, consistent with our previous finding that the bank ownership reform did not significantly increase firm profitability for more-indebted listed companies. Column (4) provides a robust result by adding industry-year fixed effects. Overall, the results suggest that the ownership reform did not cause SOBs to charge more dispersed loan rates for corporate borrowers with different levels of credit risk.

3.7 Conclusion

To make China’s four SOBs more efficient, the Chinese central government initiated a bank ownership reform, aiming to transform them into modern commercial banks with a joint-equity ownership structure. Specifically, the Bank of China (BOC) and the China Construction Bank (CCB) finished implementing the reform in 2004, and the Industrial and Commercial Bank of China (ICBC) finished the implementation in 2005.

Finally, the Agricultural Bank of China (ABC) finished the reform in 2008. Since our data cover 2000-2007, we exclude the ABC from our analysis. To identify the causal effect of the bank ownership reform on credit allocation, we use a city’s share of loans from the three SOBs in 1999 as our measure of ex ante local exposure, based on the assumption that corporate borrowers in cities with greater exposure have stronger relations with these SOBs and thus should be affected more strongly.

Using data on Chinese listed companies for 2000-2007, we show that the bank own-ership reform did not change total bank lending but reduced (increased) bank lending to more (less) indebted firms. These results imply that the bank ownership change caused loan officers to place much more importance on credit risk. The effect of the bank

own-Table 3.9: Intensive margin: firm-level interest rate

(1) (2) (3) (4)

Dependent Variable interesti,t/loansi,t

Big Three Exposurec,t 0.050 0.042 0.039 0.036

(0.042) (0.098) (0.097) (0.099) Big Three Exposurec,t×Debt ratioi,t1 0.033 0.040 0.021

(0.142) (0.141) (0.140)

EBTi,t/assetsi,t−1 0.132∗∗∗ 0.119∗∗∗

0.043 (0.045)

Controls Y Y Y Y

Year fixed effects Y Y Y

Firm fixed effects Y Y Y Y

Industry × Year fixed effects Y

Observations 6,612 6,612 6,612 6,612

R-squared 0.03 0.04 0.04 0.06

Notes: This table explores the effect of the ownership reform of three of China’s largest SOBs (Bank of China, China Construction Bank, and Industrial and Commercial Bank of China) on the firm-level loan rate for listed companies. The dependent variable is the ratio of cash paid for distribution of dividends or profits and interest expenses on current outstanding loans (interesti,t/loansi,t). The time-variant city-level exposure (Big Three Exposurec,t) is measured by the sum of the SOBs’ loan shares in 1999 interacted with the indicator for the postreform. All columns include the control variables from table 3.2, column (2), and all other terms (Debt ratioi,t1, Debt ratioi,t1×T{year >2003} and Debt ratioi,t1×Big Three Sharec,1999) are as in equation3.5. Robust standard errors (clustered at the city level) are in parentheses. We use ***, **, and * to denote statistical significance at the 1%, 5%, and 10% levels, respectively.

ership reform became prominent after a bank’s joint-equity ownership structure was es-tablished and grew stronger over time afterwards, so the observed decline in credit access for more-indebted listed companies is not attributable to other policy shocks, such as disposal of nonperforming loans before the bank reform. The (nonsignificant) increase in the firm-level loan rate for more-indebted firms suggests that risky corporate borrowers were affected more by credit rationing than by discriminating loan rates, and it also rules out the possibility of a decline in credit demand.

We further use a universe of Chinese unlisted manufacturing firms as complementary data for analysis, as listed companies are on average larger and have better financial performance. The separate large sample confirms that the bank ownership reform reduced bank lending to more-indebted firms. More importantly, we find that SOEs and less productive firms also experienced a significant decline in credit access due to the bank ownership reform, which indicates an increase in resource reallocation efficiency.

Moreover, we find that although highly indebted firms saw their bank financing re-duced due to the bank ownership reform, they suffered no decline in investment and even saw an increase in profitability. In contrast, both SOEs and less productive firms suffered significantly in terms of investment and profitability. One reasonable explanation is that reduced bank lending may have pushed highly indebted firms to become more efficient, whereas SOEs and less productive firms did not have this capability.

Most of the banking literature focuses on how the bank ownership reform affected bank efficiency; our paper instead contributes to the understanding of its impact on cor-porate credit access and firm performance. Since we find that the bank ownership reform reduced bank lending to riskier and less productive corporate borrowers, the results, in turn, support previous findings that the ownership change indeed made the SOBs more prudent and thus improved their performance. Overall, our findings provide a clear pol-icy implication that both the corporate sector and SOBs themselves can benefit from ownership changes to allow market forces to determine bank lending.

Appendix to Empirical Illustrations

Figure A.1: The Distribution of External Financing for Chinese listed firms

Notes: This graph illustrates the aggregate amount of external financing in all Chinese listed firms from 1998 to 2015. The financing channels separate into three categories: commercial loans from banks, issuing corporate bonds, and issuing new shares via IPO. This graph also shows the proportions of each financing channel by year and demonstrates that commercial lending contributes the most significant percentage among all financing channels. Source: RESSET.

Figure A.2: Number of Firms participating RPT activities in 3 categories

Notes: This graph illustrates the number of firms participate in at least one transaction in one category of related-party transactions (RPT) from 1998 to 2015. The three categories are RPT in loan (direct lend-ing), RPT in guarantee (the listed firm plays the guarantor for its related parties), and RPT in operation (non-capital related transactions). Each year, public listed firms are required to disclose all transactions with related firms, including parent firms, subordinates, or sibling firms under the same business group.

The evolution of RPT participation shows a significant increase in RPT loan and guarantee participation proportion. Source: RESSET.

Figure A.3: RPT Aggregating Amount in 3 categories

Notes: This graph illustrates the aggregating amount of three categories of related-party transactions (RPT) from 1998 to 2015, named RPT in loan (direct lending), RPT in guarantee (the listed firm plays the guarantor for its related parties), and RPT in operation (non-capital related transactions). Each year, public listed firms are required to disclose all transactions with related firms, including parent firms, subordinates, or sibling firms under the same business group. This graph sums the RPT amount of all Chinese listed firms each year. Over the past two decades, all categories of RPT increase significantly, especially after 2007-2008. Source: RESSET.

Figure A.4: RPT in Loan Amount in China

Notes: This graph illustrates the aggregating amount of direct loan lending in related-party transactions (RPT) from 1998 to 2015. Each year, public listed firms are required to disclose all transactions with related firms, including parent firms, subordinates, or sibling firms under the same business group. As a proxy of shadow banking activities, it pictures the rapid expansion of one form of shadow banks, firm-to-firm lending. Before 2009, the direct lending through relationships only takes account of fewer than 0.2 trillion Yuan. After the policy of removing commercial bank loan quota, the shadow banking amount under this format is also experiencing a sky-rocket increase, from 0.2 trillion in 2009 to 1.1 trillion in 2015.

Figure A.5: Proportion of capitalization involved in Shadow Banking

(a) Proportion of capitalization involved in Shadow Banking

(b) Proportion of state-owned capitalization involved in Shadow Banking

Notes: This area plot illustrates the asset weights of firms participating in shadow banking activities in the overall listed firm. We aggregate total assets of firms that exist related-party transactions in loans, divided by the sum of total assets of all firms in the corresponding year. In figure (a), the rapidly growing proportion of market value involved in shadow banking increase the worry about market stability, especially after 2008, from less than 40% to more than 60%. In figure (b), within the shadow bank part, state-owned firms possess the majority of related-party transactions.

Figure A.6: Credit to Non-financial Corporations in China

Notes: This graph illustrates the annual aggregation of credits to non-financial corporations from 2006 to 2016. It covers credit amounts outstanding by the end of the year, and in terms of financial instruments, credit covers loans and debt securities. Non-financial corporations include both private and public owned firms, also cover their branches of foreign corporations. Quote from BIS, ”the dataset is constructed by combining several sources such as the financial accounts by institutional sector, the balance sheets of domestic banks and non-bank financial institutions, and the international banking statistics.” Before 2009, the level of credit to non-financial firms is less than 30 trillion RMB. After the change of credit supply policy on Nov 9, 2008, credit to firms experienced a rapid explosion, from 40 trillion in 2009 to 120 trillion in 2016. The significant credit expansion stimulates the business of shadow banking. Source:

CEIC, Bank for International Settlements. Unit: RMB (Yuan) in Billion.

Figure A.7: Aggregate entrusted loan in China

Notes: This graph pictures the aggregation of the entrusted loan amount in the Chinese market from 2002 to 2017 August, sourcing from the People’s Bank of China. When one firm wants to lend its extra cash to other firms, it can choose to delegate one bank (trustee) to lend the firm’s extra cash to another appointed firm. This transaction is the so-called entrusted loan, one form of shadow banking. The lending money, although passing through a bank, belongs to the lending firm. The credit supply policy change in 2008 makes a precise breakpoint separating two intervals. Before 2008, the overall entrusted loan amount is less than 2 trillion yuan, and it was only 500 billion RMB in 2002. However, starting in 2009, the level of entrusted loan scales from 2.5 trillion in 2009 to almost 14 trillion in 2017 August. Striking increment in the shadow banking market creates potential risks out of regulation under the deposit reserve requirement by the central bank. Source: CEIC, People’s Bank of China.

Table A.1: Summary Statistics of Key Variables

Variable Unit Obs Mean Median Std. dev.

Panel A: Related-Party Trading amounts

RPT in Loans Billion 6987 0.699 0.083 2.230

RPT in Guarantees Billion 11133 0.851 0.250 2.024

RPT in Operations Billion 21409 1.202 0.070 8.628

RPT to Financial Firms Billion 6841 0.764 0.088 2.407

RPT to Real-estate Firms Billion 5559 0.382 0.033 1.306

RPT via Other Receivables Billion 27525 0.121 0.028 0.482

Panel B: Bank-firm Relationship measurements

Proximity

Median distance to county banks Kilometers 18377 12.967 1.748 60.280 Median distance to city banks Kilometers 18445 51.292 27.741 82.959 Median distance to province banks Kilometers 18400 312.671 243.955 322.788 Median distance to bank headquarters Kilometers 18200 1092.802 1069.648 627.994

Loan size

New Borrowing Billion 24190 1.633 0.384 5.726

Net Borrowing Billion 25011 0.142 0.015 1.079

Total Debt Billion 24494 1.449 0.338 5.206

Long-term Debt Billion 15877 1.035 0.127 4.070

Short-term Debt Billion 23594 0.792 0.252 2.288

Equity exposure

Equity Exposure Percentage 4033 0.040 0.021 0.050

Panel C: Firm characteristics

Size Billion 27559 5.730 1.753 19.688

Leverage Fraction 14787 0.143 0.069 1.216

Cash Fraction 25532 0.265 0.160 0.878

Capital Expenditure Fraction 25380 0.084 0.047 0.567

ROA Ratio 25532 0.062 0.042 0.463

Market-to-Book Ratio 24824 7.280 3.127 123.134

Kaplan-Zingales Index 25496 -0.142 -0.175 19.638

Age Year 27406 12.330 12.000 6.472

Appendix to mathematical proof

B.1 List of Variables

Variable Description w Firm capitalization

g(w) The distribution of firm capitalizationw

wN Threshold of firm capitalization under which firms are credit rationed (w1, w2) The range of firms which borrow from both bank and shadow banking p0 Non monitoring effort of entrepreneur

p Total effort of entrepreneur/ Probability of success of investment project β Marginal cost of effort of entrepreneur

m Marginal cost of monitoring of the bank Q Return of the investment project if succeed rf Risk free return

rs Expected return of shadow banking

¯

rs Upper bound ofrs

IB Loan volume from the bank RB Return of the bank loan

IS Loan volume from the shadow banking RS Return of the shadow banking

IBS Equilibrium bank loan to the firms which become creditor of shadow banking RSB Equilibrium bank return from the firms which become creditor of

shadow banking

IBM Equilibrium bank loan to the firms which borrow from both bank and shadow banking

RMB Equilibrium bank return from the firms which borrow from both bank and shadow banking

ISM Equilibrium shadow banking loan to the firms which borrow from both bank and shadow banking

RMS Equilibrium shadow banking return from the firms which borrow from both bank and shadow banking

IBOB Equilibrium bank loan to the firms which only borrow from the bank ROBB Equilibrium bank return from the firms which only borrow from the bank T S Total supply of shadow banking

T D Total demand of shadow banking