METHODOLOGY AND DATA

3.1. Method

We construct a panel regression model to study the effects of price-based and quantitative monetary policy instruments on bank risks as follows:

RK_t = β₀ + β₁ResRatio_t + β₂IR_t + β₃ln Open_t +β_iGDPG_t + β5INF_t + β₆ ln(PRI_t + β7 ln(SIZE_t) +β8ER_tt + β9Ownership_t + β ln(Regulation_t) + ε

(1) where the risk variable, RK, of bank i at time t is a function of monetary policy instrument variables (ResRatio, IR and Open); a set of macroeconomic, regulatory control variables(GDPG, INF, RPI and Regulation) and a set of bank-level control variables(SIZE, ER, and Ownership).

Table 5 provides descriptive statistics for all the variables used in the empirical analysis. Table 6 reports correlation coefficients between these vari­ables. According to Damodar N.Guj arati(2000),if zero-order correlation coefficient of each two regressors is over 0.8, the multi-collinearity problem will be a severe problem. Correlations in our study are acceptable levels as shown in Table 6. In what follows, we analyze the choice of our dependent and explanatory variables.

We proxy the risk behavior of banks using RiskAsset, LoanRatio and InvRatio. RiskAsset, a dependent variable, measures banks’ risks and is defined as the ratio of risk assets for bank i to total assets at time t. Bank risk assets are the as­sets that fluctuate with the money market or credit market and include all bank assets except cash, the deposits with Central Bank, the deposits with banks and interbank lending.

ResRatio represents the reserve requirement ratio, a main quantitative instrument used in China. We adopt the reserve requirement ratio for the large financial institutions with the reason that large financial institutions is the majority of Chinese banking industry and the reserve require­ment ratios for the large, small and medium-sized financial institutions have a basically uniform trend. As the reserve requirement ratio and the interest rate may be changed for several times within one year, the daily weighed means of the both is used. Since 2000, ResRatio has kept a rising

Table 5. Descriptive statistics

Note: RiskAsset is the ratio of risk assets for bank i to total assets at time t.

Table 6. Correlation matrix

trend from 6.0 percent in 2000 to 21.0 percent for the large financial institutions and 17.5 percent for small and medium-sized financial institutions in December 2011.

IR denotes the variables that relate to inter­est rate, the price-based instrument. We use two variables: DIR and Margin. DIR denotes the RMB 1-year benchmark deposit rate. Margin equals LIR (i.e. the RMB 1-year benchmark loan rate) less DIR. Figure 1 and Figure 2 shows the changes of DIR-LIR and Margin. The years of 1998-1999, 2000-2006, 2007-2008, 2009-2010 and 2011 are respectively high interest rate regime, low interest rate regime, high interest rate regime, low inter­est rate regime and high interest rate regime. The years of 1998-1999, 2000-2008 and 2009-2011 are respectively the periods of low interest rate margin, high interest rate margin and low interest rate margin.²

Open means open market operation, another main quantitative instrument in China. We use the turnover of T-Bond trading (including spot trading and repo.trading) to measure open market operation. Since 2006, total turnover of trading of T-Bond had been rising continually from 1063.35 billion Yuan in 2006 to 20084.133 billion Yuan in 2011.

We have chosen several control variables for our regression analysis. These variables are defined and discussed below, while summary statistics are reported in Table 5.

GDPG (realGDP growthrate) and /.∙V√∙'(ιnl'lalιon rate, denoted by CPI growth rate) is used to mea­sure the macroeconomic situation that may have an impact on bank assets and debts, thus affecting bank risk.

RPI, real estate price index, is used to measure the situations of real estate market. The real estate loans account for a large proportion of Chinese bank credits. Too much bank credits are input into the real estate market, which may give reasons to the potential risk.

SIZE represents size of the bank, which is measured by total asset of the bank. Banks with

Figure 1. The changes of the interest rates: Changes of LIR and DIR

Figure 2. The changes of the interest rates: The changes of Margin

the different sizes may have distinct motivation to engage in risky activities.

ER denotes equity ratio that is defined as the ratio of shareholders’ equity and total assets, which represent banks’ asset-liability structures.

Banks with the different asset-liability structures face different risk.

Ownership represents the bank ownership. Considering the differences of risk-taking behav­iour between state-owned banks and non-state-

owned banks, we construct a dummy variable BD on behalf of bank ownership. If the bank is state-owned (Industrial and Commercial Bank of China, Agricultural Bank of China, Bank of China, China Construction Bank and Bank of Communications), the value of BD is 1; if not, the value of BD is 0.

Regulation denotes the banking regulation intensity that is also an important factor that has an impact on bank risk. The value of the banking supervision intensity from 1998 to 2002 is 0.5, because China’s banking regulation was carried out by the People’s B ank of China before 2003 and separated regulation wasn’t implemented as well, which made the regulation relatively weak.

3.2. Data

We use annual data of nineteen Chinese banks whose assets account for over 64% of the assets of Chinese banking industry from 1998 to 2011. The sample contains the five large commercial banks, ten joint-stock commercial banks, and four city commercial banks. Among those, the large commercial banks include the Industrial and Com­mercial Bank of China (ICBC), the Agricultural Bank of China (ABC), the Bank of China (BOC), the China Construction Bank (CCB) and Bank of Communications (BOCOM); the joint-stock commercial banks include CITIC Bank, Ever­bright Bank of China, Huaxia Bank, Guangdong Development B ank, Shenzhen Development B ank, China Merchants Bank, Shanghai Pudong Devel­opment Bank, Industrial Bank, China Minsheng Banking Co., and Evergrowing Bank; the city commercial banks include Beijing Bank, Shanghai Bank, Nanjing Bank and Ningbo Bank.

The banks’ data of risk assets, total assets, loans, investments, equity ratio and the data of real estate price index, real GDP growth rate, inflation rate measured by CPI growth rate come from DRCNET Statistical Data Database System, Almanac of China ^,s Finance and Banking and the website of very bank. The data of the RMB 1-year benchmark deposit rate, the reserve requirement ratio, the total turnover of T-Bond trading comes from the website of People ^,s Bank of China.

4.