مدیریت ریسک و بهره وری مدیریتی در بانک های چینی : یک چارچوب شبکه تحلیل پوششی داده ها

Risk management in Chinese banks has traditionally been the Cinderella of its internal functions. Political stricture and developmental imperative have often overridden standard practice of risk management resulting in large non-performing loan (NPL) ratios. The training and practice of risk managers remain second class compared with foreign banks operating in China. This paper surveys Chinese bank risk managers and constructs metrics of risk management practice and risk management organisation. The metrics are used as intermediate inputs in a Network DEA framework to produce a measure of income efficiency. A statistical test is carried out to assess the importance of the risk metrics in evaluating bank income efficiency.

In recent years the Chinese banking system has made enormous strides in reform and deregulation. They have emerged relatively unscathed from the global banking epidemic that has infected the developed economies and the largest of them stand alongside the giants of global banking as first among equals1. However, despite the relative strength of the large listed Chinese banks in world banking, lingering doubts remain about the inherent fragility of the banking system in China. The past decade has seen a large volume of academic and professional papers expressing concerns about the safety and soundness of the Chinese banking system and their medium term viability in the face of increasing competition from foreign banks in the post WTO years. The common thread in many reviews of Chinese banking are: the large number of non-performing loans, the dominance of lending to state-owned enterprises, and the influence of local government and Communist Party officials in lending decisions. A particular area of concern for the regulatory authorities and strategic investors in the Chinese banks has been the quality of training of risk managers, the organisational culture and the misalignment of incentives associated with bureaucracy rather than commercialism [7]. The process of converting Chinese banks from state dominated bureaucracies to modern profit oriented banking institutions involves not just the training of decision makers in modern banking but also the transformation of the organisation. This transformation has been occurring but on an evolutionary rather than a revolutionary pace. With the encouragement of the regulatory authorities, Chinese banks have in recent years, had to restructure their balance sheet, develop modern risk management methods, improve capitalization, diversify earnings, reduce costs and improve corporate governance and disclosure2. This paper aims to evaluate the performance of the risk management function of Chinese banks in terms of its contribution to profitability. There are four parts to the research. The first part collects qualitative data on risk management practice and risk management organisation through a semi-structured questionnaire. The second part quantifies the qualitative data by constructing a metric of risk management practice, and risk management organisation in Chinese banks, using the foreign banks operating in China as a yardstick of best practice. The metric will measure how good the practice of risk management is in a Chinese bank and how well the risk management function organisation is relative to best practice. The third part uses the constructed metrics of risk management in a network DEA framework to evaluate bank income efficiency. The fourth part tests the hypothesis that the inclusion of the risk metrics in the network DEA improves the measurement of income efficiency and its link to bank profitability. The paper is organised in the following way. The next section outlines the results of interviews. Section 3 describes how the interview responses are converted into a relative score compared with two of the major foreign banks operating in China (HSBC and Citibank). Section 4 describes the method of performance evaluation based on Data Envelopment Analysis (DEA) and the use of network DEA to think of risk management and organisation as an intermediate output/input in the process. Section 5 tests the hypothesis that the use of the risk metrics in the network DEA improves the measurement of efficiency and its correlation with bank profitability. Section 6 summarises the results and concludes.

This paper has demonstrated an innovative application of utilising qualitative data in the efficiency evaluation of firms operating in the same market. Specifically we have constructed an index of risk management practice and risk organisational practice for a sample of Chinese banks from qualitative information. Risk management practice and risk organisational practice was confined to the classic retail banking functions. The metrics were constructed from scores provided by risk managers in domestic Chinese banks in responses to a semi-structured questionnaire. Scores on a Likert scale of 1–5 were translated into an index of practice from an asymmetric function that penalised downside deviations from best-practice more than up-side deviations. Best practice was defined from the interviews of four managers from two foreign banks operating in China. An aggregate score was constructed using principal components analysis. The metrics of risk management practice and risk management organisation obtained from questionnaire analysis may be used as a measure of performance however, organisation of the risk function is a management function and the risk management practice can be thought of as part of a culture of loan approval determined by a mixture of political as well as commercial interest. The organisation of the risk function metric can therefore be thought of as an internal intermediate input along with risk management practice as an external input, to produce interest earnings (as well as non-performing loans). The risk organisation and risk management metrics are a link in the production chain of revenue streams in Chinese banking. We found no significant direct relationship between the two constructed measures of risk management practice and risk management organisation and an objective measure of performance of the bank such as ROA. However, the input of these measures within a DEA network framework produced efficiency scores that explained ROA better than efficiency scores that excluded them. We have addressed the dimensionality problem in DEA and demonstrated that the improvement in average efficiency as a result of using the two risk measures as intermediate inputs is a valid exercise. We argue that the information content of the risk management practice and risk management organisation measures is indirect and is better revealed within a network DEA framework. The risk management practice and risk management organisational metrics constructed from interviews provide insight into the risk function in Chinese banks relative to best practice. The link between the two measures of risk management and that of bank revenue efficiency works only if the results from the qualitative analysis are representative of the organisation as a whole. There is no reason to believe that this is not the case. But, it is the combination of the risk practice and risk organisation with the other inputs and outputs of the banks that matter for final performance. Provided that the results from the interviews of risk managers for each bank are representative of broad practice nationwide, the metrics can be used as inputs in the intermediate stage of production.