Efficiency gains and deregulation policies: evidence from bank level data.

Abstract

This thesis uses bank level data from developing countries and emerging economies and the data envelopment analysis (DEA) approach to provide empirical evidence on the impact of deregulation policies in the banking industry on the banks’ efficiency. Since the banking industry in China is the largest and most complex among the developing countries and in transition from a centrally planned to a market economy, its transition process and the development of the banking industry is analysed first in order to provide the background information for the following econometric analyses and the thesis. To gather the empirical evidence from China’s banking industry on the correlation between the World Trade Organization (WTO) accession and efficiency gains by commercial banks, this thesis evaluates the efficiency of Chinese banks over the period 2000–09 (this is referred to as the adapting phase of the WTO accession). During this period of time, the restrictions on the foreign banks were removed gradually. The evolution of banks’ efficiency is computed by the DEA approach combined with the bootstrapping technique. All commercial banks are broken down into four groups: (1) all banks in China; (2) domestic banks; (3) private banks; and (4) city banks. Since the categories are mutually exclusive, the empirical results reveal that the efficiency of the banks in China’s banking industry increased over this period. In terms of profit maximising, city banks were the least efficient banks and the catch-up effect was highly significant in this group, since their efficiency increased dramatically compared with other banks. However, the empirical evidence from the Chinese banking industry cannot identify the efficiency effect from removing restrictions on banks in the market. In order to identify the efficiency impacts from different kind of deregulation policies, first, the impact of the deregulation policies to remove the restrictions on foreign banks and domestic banks are explored in six Asian banking industries over the period 1997–2006, namely China (data for mainland China and Taiwan presented separately), India, Indonesia, Malaysia, Pakistan and Thailand. In the first stage of the two-stage DEA model, the output direction DEA is employed for the selected countries to compute the efficiency of the banks. In the second stage, the estimated DEA score is regressed on the indices of the restrictions in the market. The values of indices are taken from Dinh (2008). The main reason to select her indices is that these indices are used to estimate the restrictions on foreign banks and domestic banks in the given market. The expectation is that the deregulation policies to remove the restrictions on foreign or domestic banks will lead to efficiency gains in the markets. In order to overcome the reverse causality issue between the dependent variable (the estimated DEA score) and the independent variable (restriction indices), the two-step first-difference regression model is used and bank efficiency in the previous period is included in the model as one of control variables. The main reason to use the first-difference model is to partial out unobservable time and country effects from the data panel. In the sensitivity analysis, a couple of different model specifications are utilised to confirm the baseline results. The regression results show that the deregulation policies related to the operation of foreign banks are positively correlated with efficiency gains of commercial banks, but the other key set of policies to liberalise the domestic banks has not resulted in significant efficiency gains in the selected banking industries. As alternative channels for increased competition in the market, the efficiency impacts of mergers and acquisitions (M&As) are explored. Theoretically, banks may be encouraged to enhance their efficiency due to the pressures that arise from the possibility of M&As. Most previous analyses use the case study methodology in this topic rather than the cross-country statistical methodology. In this thesis, the efficiency impacts are examined with a sample of banks from a range of emerging economies (China, India, Malaysia, Russia, Thailand and Vietnam) over the period 2002–09. All banks in the selected countries are divided into three groups, namely target banks, acquiring banks, and the banks not involved in the event (or incumbent banks). To compare the differences in the impacts on efficiency between the banks involved in the event (target and acquiring banks) and the banks not involved in the event (incumbent banks), the two-stage DEA is employed. In the first stage, the efficiency of the banks is calculated in the DEA model. The results from the DEA show that the efficiency of the banks increased in most of the countries, except India, in which the bank efficiency is neutral over the sample period. In the second stage, two different matching methods was utilised in this thesis: the regression method and propensity score matching. The empirical results are robust across a number of sensitivity analyses and identification methods and reveal that the M&As reduce the efficiency of the acquiring banks and target banks in the selected emerging economies.