我国商业银行逆周期资本监管的锚定指标选取

本文依据巴塞尔委员会提出的逆周期资本监管思想,以各类信贷指标为基础,通过绘制ROC曲线并计算AUC值,选取适合我国的具体锚定指标。本文认为,好的锚定指标需要满足及时性、稳定性、可解释性以及可获取性等标准。研究结果表明:房地产价格趋势缺口、房地产贷款趋势缺口、个人住房抵押贷款趋势缺口以及消费贷款趋势缺口等指标符合锚定指标的标准,可作为逆周期资本监管锚定指标使用,并且组合锚定指标的预警能力和稳定性优于单个锚定指标。本文从落实锚定指标和提高挂钩变量预警效果的角度,提出了更为精细的信贷类别指标,为推动我国逆周期资本计提及监管实施提供了决策参考。

Countercyclical Capital Regulation of Commercial Banks in China: Selecting an Anchoring Indicator

As an important part of macroprudential supervision, countercyclical capital requirements play a primary role in alleviating systemic risks caused by excessive bank credit and promoting financial stability. Although countercyclical capital regulation has been proposed for more than eight years, China has not yet promulgated the specific regulatory rules and requirements. A key reason is the lack of both a unified specific method and time-tested indicators as a reference to establish the anchoring indicator, which is an important basis for calculating, dynamically adjusting, and releasing countercyclical capital. Should the generalized “credit/GDP” guide recommended by the Basel Committee be adopted？ Is it in line with the risk characteristics of the Chinese banking industry？ Are there any other indicators and elements of particular concern？ What is the mechanism for selecting indicators？ Discussing these issues will enrich the discussion surrounding countercyclical capital supervision and promote policy implementation in China.
This paper first reviews the countercyclical capital regulation guidance proposed by the Basel Committee and its latest implementation in various countries, and chooses 15 candidates widely used in practice, together with three classes of anchor variables proposed by the Bank for International Settlements (Drehmann et al., 2011). Considering the risk characteristics of the Chinese banking industry and data accessibility, six credit variables are added to reflect the actual loan structure in China. Altogether, these 21 candidates provide 84 anchor variables through four forms of transformation and finally yield 105 candidate indicators.
The candidate indicators are then screened and verified. First, this paper replaces “systemic financial risk/financial crisis,” which has not happened in China, with the “downward stage” of the middle-term financial cycle. Based on the latest literature, the first quarter of 2011 is set as the starting point of the “downward stage”, which serves as a reference time node for operating the countercyclical regulation. Second, by making use of the method in Drehmann and Juselius (2014), this paper draws the Receiver Operating Characteristic (ROC) curve with the data at the time of the first quarter of 2010 and respectively calculates the Area Under the Curve (AUC) value, which measures the candidate indicators' ability to predict the outbreak of systemic financial risks one year later. Third, this paper ranks each indicator by examining whether it meets the criteria of static timeliness, dynamic stability, economic interpretability, and data accessibility when used for early warning of systemic risk, so as to complete the identification process of anchor indicators. Finally, this paper carries out robustness tests by determining the starting time nodes of the “downward stage” in three financial short cycles and checking whether the ROC curve method used obtains consistent indicators in different contexts in a financial downturn period.
The empirical results show that specific industry indicators such as the trend gap of real estate price, and specific categories of credit indicators, such as the trend gap of real estate loan, the trend gap of personal housing mortgage, and the trend gap of consumer credit, can track potential risks from multiple dimensions, and thus can be used as anchor indicators for countercyclical capital supervision. In addition, the combined anchor index has a higher early warning ability and stability than single anchor indicators. These conclusions improve the precision of optimal anchor indicators to reflect systemic risk by focusing on classified credit and industry credit instead of general credit. By exploring anchor indicators and improving the early warning effect of linked variables, this paper provides practical guidelines for decision making and promoting the operation of countercyclical capital regulation in China. Its findings may also provide inspiration for other countries or regions.
This paper contributes to the literature in several ways. First, the ROC curve method is used for the first time to identify the pros and cons of China's countercyclical anchor indicators and ensure the reliability of the conclusions. Second, the candidate indicators have been comprehensively arranged and scientifically transformed, which makes it possible to select the most effective anchor indicators. Third, the period from 2004 through 2018 contains relatively complete economic and financial cycles, and thus reflects rich information and offers a basis for empirical research on countercyclical policies. Finally, this paper links countercyclical capital regulation in China to personal housing credit and consumer credit as anchor indicators, which suggests avenues for viable countercyclical capital regulation mechanisms in China.