Comments on “Sources of Procyclicality in East Asian Financial Systems”
- Stefan Gerlach, and Paul Gruenwald
- Published Date:
- July 2006
As we have seen, structural weakness of the financial system is the major reason contributing to exacerbated procyclicality in Asia. This finding concisely builds a bridge between micro-prudential and macroeconomic benefit, with which I tend to agree. The argument is convincing since it is not only based on qualitative analysis but also supported by thorough econometric analysis. The latter substantially enhances the credibility of subjective inferences on causes of procyclicality and policy proposals.
I find that most of seven sources of procyclicality mentioned in Chapter 3 fall within the context of credit risk, whose management does need to be improved given the high NPL ratio in this region. It is agreed by the supervisors and monetary policy-setters that the roadmap to mitigating procyclicality is also largely the process of strengthening credit risk management in the financial sector.
As mentioned in Chapter 3, procyclicality is a normal feature of economic systems, which I believe is the fundamental basis for our discussion. With this in mind, only largely magnified or excessive procyclicality needs to be addressed. Additionally, policy measures adopted by authorities should be market-oriented, primarily in the form of prudential regulation and supervision. As a result, efforts such as attempting to smooth the profit of financial institutions, thereby mitigating procyclicality, should not be encouraged unless considered as the last resort.
Lastly, to cope with the procyclicality of the financial sector, monetary policy or credit policy in particular should also be included into our consideration, because monetary policy can play a role in addressing this issue in a balanced way in coordination with the banking supervision and other policies. Another option is to develop financial markets, in particular the active and well-functioned bond market, to mitigate the negative impact of the procyclicality of the banking sector on the economy.
I shall now spend some time on the Chinese experiences with banking credit procyclicality.
A brief review of procyclicality in China
The growth of banking credit in China has demonstrated strong procyclicality since the 1990s. Banking credit movement is positively correlated with GDP growth rate significantly with a 0.74 correlation coefficient. The credit cycle matches very precisely the business cycle (Figure 3.2). Credit is not only procyclical, but tends to grow faster than GDP growth during expansions and much slower than expected during recessions. This is reflected clearly in the behavior of the credit-to-GDP ratio (Figure 3.3). In the recession period from 1998 to 2001, average annual credit growth was just 10.4 percent, which strikingly contrasted with 18.4 percent annual growth from 1990 to 2003. In particular, credit growth fell to 6 percent in 1999, which was two percentage points lower than GDP growth in the same year. In the boom period from 1992 to 1995, annual credit growth exceeded 23 percent. In 2002, China witnessed 21 percent credit expansion, which was more than double GDP growth. Procyclicality was no greater in recessions in China since banks’ self-discipline was not sufficient and there was a hunger for credit demand.
Figure 3.2:Growth rate of banking credit and GDP 1991-2002 Figure 3.3:Banking credit/GDP 1991-2002
Factors contributing to procyclicality
Banking credit is strongly procyclical in China. Here I want to address five specific reasons contributing to procyclicality in China. First, banking credit plays a dominant role in allocating savings and investments in China. The role of credit is magnified as the capital market is underdeveloped. Currently, bank lending accounts for 60-70 percent of total fixed asset investment, while the stock of working capital loans has exceeded 70 percent of GDP. Credit is so essentially important for the whole economy that it has been consistently used as an effective macroeconomic adjustment tool during various business cycles. Not surprisingly, we saw that credit fluctuation is both the cause and result of the business cycle in China.
Second, the procyclical behavior of credit is endogenous in the economy. Empirical study has shown that credit demand in China is mainly driven by economic growth, the real interest rate, and the marginal return of capital while credit supply is mainly determined by borrowers’ net wealth, loanable funds in the banking sector, and credit policy. Obviously, the finalized credit is closely connected with economic growth.
Third, measurement difficulty in the time dimension of risk and its backward-looking nature is also an important source of credit procyclicality. Most of the risk measurement methodologies used by banks imply that risk falls during booms and increases during recessions. The measurement bias pushes banks to expand lending activity during boom periods, and to contract their lending activity when business turns down. However, the reality is that risk is building up in economic booms, and the materialization of bad loans in recessions does not necessarily mean an increase in risk. I have always believed that the worst loans are made at the top of the business cycle. Of course, the above limitations in risk measurement methodologies would inevitably be reflected in banks’ provisions and capital.
Fourth, a well-functioning information collection system does not exist in the Chinese banking sector. Loan officers do not have strong discriminative power when selecting potential borrowers, as they are normally not able to access enough effective financial data from borrowers. Information asymmetry and herding behavior is extremely serious under this condition. Faced with economic climate change, banks tend to sharply increase or decrease their loans bluntly and collectively.
Finally, imbalanced and inconsistent performance evaluation through the economic cycle imposes distorted incentives on bank managers in China. Profitability and risk are the most important performance dimension for any bank, and profitability must be balanced with risk consistently. Unfortunately, banks tend to overemphasize profitability and market shares while increasing risk tolerance unconsciously during booms, and vice-versa. Interestingly, the lending line granted to the loan officer in branches increases during booms and decreases during recessions. Hence, imbalanced performance evaluation amplifies credit swings and procyclicality.
Supervisory responses to procyclicality
The procyclicality of banking credit may create a number of potential problems, including exacerbating the business cycle, increasing systemic risks and misallocating lending resources. As the economy presented certain overheating signals last year, the Chinese Banking and Regulatory Commission (CBRC) has been rethinking its role and positioning itself to cope with credit procyclicality in the course of economic adjustment. We believe that mitigating procyclicality is generally in the interest of effective banking supervision, since sound and sustainable economic development is an important precondition for effective banking supervision, and a severely fluctuating economy would inevitably pose a threat to the banking system.
Based on the above recognition and mindset, the CBRC has done its utmost to mitigate procyclicality, thereby coordinating with economic adjustment. Let me elaborate on some of our practices.
Fundamentally, we defined the point of intersection between banking supervision and addressing procyclicality as the promotion of improved measurement and control of risk in longer time horizons in the banking sector. Accordingly, we return to the basic prudential supervisory policies, particularly their implementation, such as capital adequacy, risk diversification, limits on risk concentration, limits on the loan-to-value ratio, sound risk management procedures and processes, and banks’ due diligence in lending. Our practice has demonstrated that provisioning and capital requirements can indeed be cautiously used as policy responses to cycles. As a result of strengthening capital regulation, commercial banks have fully recognized the size of their capital as a binding constraint for their balance sheet expansion.
In particular, emphasizing risk information infrastructure is another important feature of our responses. We required banks to establish management information systems, which are able to provide complete and well-categorized financial and risk data to improve their risk measurements and controls. We believe this is fundamental for enhancing risk measurement accuracy. As a supervisor, we issued risk advisory information when we saw the bubble in real estate as early as 2003. Our practice has shown that supervisory advice is also an effective tool for the promotion of better understanding of risks.
Challenges and concluding remarks
Supervisory policies have one distinct limitation in addressing procyclicality; namely, they could hardly be used to cope with recessions. I am also not sure that bank supervisors can accurately predict or anticipate the economic cycle or the potential financial distress of banks. Additionally, since China is a transition economy, banking supervision should always bear in mind its independence and credibility when coordinating with macroeconomic issues. Independence is as important as coordination for the newly established CBRC, if not more important. I personally tend to suggest that monetary policy should play an important role in addressing procyclicality together with other policies and measurement.
I firmly believe that building up a protective cushion in good times is an inherent requirement for banks’ risk management. If banks manage their risk diligently, the negative outcome arising from procyclicality can be substantially alleviated, if not eliminated. From the same consideration, I am not a strong proponent of dynamic provisioning. It seems to me that dynamic provisioning tries to use a new rule to solve the issue of imprudent provisioning. Probably, the dynamic provisioning does not make much sense for banks with sophisticated risk management skills. Furthermore, as long as the provision is adequate to cover the expected loss through the complete business cycle, the supervisors do not need to concern themselves with the precise time distribution of provision.
To conclude, from the perspective of a banking supervisor, the process of mitigating procyclicality is also the roadmap of building up risk measurement and control capacity. The business cycle and the credit cycle are “twin cycles,” and eliminating procyclicality is as difficult as killing the business cycle itself, if indeed that were possible. We must learn to live with it.