Cost efficiency, technological progress and productivity growth of Chinese banking pre- and post-WTO accession

Abstract

China has recently taken substantial steps to reform its banking sector, particularly after joining the World Trade Organization (WTO) in December 2001. This study examines the effect of recent banking reforms and WTO accession on the cost efficiency of Chinese banking and the efficiency differentials across different bank ownership groups. We use a nonparametric approach to investigate the efficiency trend and productivity growth of banks between 1998 and 2006 prior to and after joining the WTO. We find that, on average, domestic banks outperform their foreign counterparts over the sample period in terms of overall and allocative efficiencies, but they fall behind in terms of overall technical efficiency. A pre- and post-WTO accession analysis reveals that the efficiency of domestic banks has declined post-accession, while foreign banks have enjoyed an improvement in their cost efficiency post-WTO accession. The findings further suggest that the total factor productivity of Chinese banks has weakened over the period under study. However, total factor productivity has increased for both domestic and foreign banks after China joined the WTO, equally owing to efficiency improvement and technological progress.

Acknowledgements

We thank the Lebanese American University for funding and Susmitha Obulareddy for her valuable research assistance.

Notes

¹ After factoring in the price differentials and measuring the purchasing power parity adjusted GDP, many economists believe that China's GDP is actually the second largest economy in the world behind the US. It is expected that this growth momentum will continue at least for the next 20 years (Holz, 2006). As of this writing, all signs of macro-economic variables are pointing to further growth of the Chinese economy.

² The rural banking activities of PBC were transferred to the Agricultural Bank of China (ABC); the foreign currency transaction banking activities of PBC were assigned to Bank of China (BOC); the banking activities related to the construction sector were delegated to China Construction Bank (CCB) and finally the Industrial and Commercial and Bank of China (ICBC) took banking responsibilities related to commercial and industrial activities.

³ These policy banks are China Development Bank, Agricultural Development Bank of China and Export-Import Bank of China.

⁴ Currently, there are 12 joint-equity banks comprising 15% of total banking assets, with an average Nonperforming Loan (NPL) of 2.8% and with a capital adequacy ratio of 8.15% as of 2006. There are also about 112 city commercial banks representing only 6% of banking assets (USD 346 billion) as of 2006, and which maintain close ties with their respective city government. The latter holds a 75% share of these banks, and about 70% of their loans are granted to state and city-owned enterprises that operate within the city boundaries. Recently, the CBRC allowed city commercial banks to engage in cross-city geographic expansion, and some of them have also decided to go public.

⁵ Recent ratio analysis studies on banking sector in China include Li et al. (2001), Shirai (2002) and Shih et al. (2007).

⁶ One disadvantage of the DEA method is that some firms may be labelled as 100% efficient not because they dominate any other firms, but simply because no other firm produces as much or more of every output (given inputs) or uses as little or less of every input (given outputs).

⁷ We exclude noncommercial banking institutions such as policy-lending institutions, finance companies and investment banks to avoid comparison problems among different types of financial institutions that may be characterized by different objective functions or technologies.

⁸ This resulted in dropping banks whose operations were ceased due to a merger with other banks during the period under study.

⁹ As in Berger et al. (2009), we treat deposits as an output and do not include it in our set of inputs.

¹⁰ Chinese banks’ financial statements do not separately report expenses on salaries and benefits, expenses on fixed assets, and – in most of cases – the number of employees. To overcome this deficiency, researchers have taken different approaches in defining fixed assets and labour and their prices. For example, Berger et al. (2009) and Chen et al. (2005) have used the ratio of noninterest expense to fixed assets as a proxy for the cost of fixed assets, while the cost of labour is ignored. We think that employee expenses are one of the major costs of intermediation in banking, especially in China where most banks, particularly the state-owned banks, are suffering from over-employment. At the same time, we believe that the cost of labour and fixed capital should be approximated by the unit cost per dollar of total assets rather than per dollar of fixed assets. To overcome this problem, we assume that the cost of fixed assets and labour are reported in the noninterest expense items, and, therefore, we define the cost of fixed assets and labour as the ratio of noninterest expenses (sum of other operating expenses and nonoperating expenses) to TA.

¹¹ See Coelli et al. (1998) and Grigorian and Manole (2002).

¹² Since all other variables are ratio data, and in order to reduce scale bias, the logarithm value of total assets and GDP per capita are considered.

¹³ However, in the banking industry, a low cost efficient bank may not necessarily be a low profit bank. For example, a decision made by a manager to employ a high quality input (such as qualified labor), and hence incur high costs, results in a higher cost (low cost efficiency) but higher revenue and higher profit in the long term.

¹⁴ See, for example, Mester (1996).

¹⁵ The six governance indices provided by Kaufmann et al. (2006) include voice and accountability, political stability, government effectiveness, regulatory quality, rule of law and corruption.

¹⁶ Although based on somewhat different sets of inputs and outputs, the results are consistent with those reported by Berger et al. (2009) who report overall cost efficiency of 0.897 for a sample of 38 banks operating in China between 1994 and 2003.

¹⁷ This anomaly is also reported by Berger et al. (2009), and for which they provide two likely explanations; the skimping hypothesis and government subsidy on the cost side.

¹⁸ We are thankful to an anonymous referee for pointing out that our results may be driven by sample heterogeneity as a result of including the large Big Four state banks with smaller foreign and domestic banks in a pooled sample. According to Fiorentino et al. (2006), nonparametric methods are much more sensitive to sample heterogeneity than parametric techniques.

¹⁹ Further, foreign banks still suffer from foreign debt quotas above which they are subject to a 5% business tax, not to mention the lack of national treatment regarding capital requirements and limits on off-shore funding (PricewaterhouseCoopers, 2007).

²⁰ Later on, the Chinese authorities postponed the actual date implementation to April 2007, Wall Street Journal (23 April 2007, p. A8).

²¹ For example, the NPL ratio of the Big Four state-owned banks in 2006 declined to 9.22% from 10.49% compared to previous year. At the same time, the NPL ratio of joint-stock banks fell to 2.81% from 4.22% relative to a year earlier (CBRC website). A similar discrepancy between the NPLs of the Big Four and joint-stock banks prevails throughout the period of this study.

²² We distributed a questionnaire among Chinese business major students and asked their opinion if their parents’ deposits with major banks carry implicit government guarantee. They unanimously responded positively. They also backed their response by referring to the recent cash infusion by government into three of the four big state-owned banks to reduce their NPLs.

²³ The additional Tobit regression results are available from the authors upon request.

²⁴ To the best of our knowledge, Matthews et al. (2009) is the only published paper that examines the productivity of Chinese banks using Malmquist index.

²⁵ We are grateful to an anonymous referee for pointing our attention to the results of Table 7 (Panel C) that might be interpreted by the reader as opposing to those of Table 5. Please note that Table 5 presents average efficiency levels for each period (pre- and post-WTO), whereas Table 7 (Panel C) shows the change in OTE within the post-WTO period.

²⁶ To illustrate, the state-owned Industrial and Commercial Bank of China recently purchased a more than 50% stake in a bank in Indonesia, a country where foreign investors are allowed to own up to 99% in domestic banks (Wall Street Journal, 12 December 2006).

²⁷ One example is the Citigroup partnership with Guangdong Development bank, in which Citigroup owns 20% stake, as do China Life, State Grid and China International Trust and Investment Company (CITIC) trust. IBM holds 4.74% stake and Guangdong Financial Investment Holding will hold 0.85% stake (China Business Matrix, 2006).

²⁸ Kodjo (2007) studied the comparative advantages of Chinese local joint-stock banks with that of Big Four and small city banks as recipients of foreign investment and concludes that strategic partnerships between foreign banks and Chinese joint-stock banks are mutually beneficial.