Tóm tắt Luận án The impact of multimarket contact on the competition, credit risk and operating performance of Vietnamese commercial banks

not have to exert efforts to keep costs under control; Second, market power can allow managers to pursue goals other than profit or value maximization; Third, in the absence of competition, managers who spend resources to win and maintain the market share will unnecessarily increase costs and reduce efficiency; Fourth, if banks benefit from market power, poorly-performing managers can survive without trying to work more effectively. 2.2.3 Mutual forbearance hypothesis Concept: According to oligopoly theory, multimarket companies will not compete strongly with their competitors in a certain market if they are afraid of retaliation in all the remaining markets (Edwards, 1955; Sorenson, 2007). This is because the higher number of mutual markets in which businesses compete with each other, the higher likelihood that retaliation will occur in these markets, resulting in significant losses. In such a situation, the overall level of competition will decrease to avoid excessive losses across the markets. Impact: Bernheim and Whinston (1990) have shown that, when businesses have multimarket contact, they will have tacit agreements with each other and maintain cooperation instead of competition if there are imbalances among markets where businesses compete. On the contrary, there is also the view that multimarket contact with competitors is positively related to competition. Solomon (1970) argues that if the bank was highly competitive in certain markets within a certain region, the number of multimarket linkages in the same area could lead to increased competition, and reduce bank profits. 2.2.4 Theory on the impact of competition on the risk of commercial banks Concept: This hypothesis holds that excessive competition among banks may threaten the solvency of particular institutions and hinder the stability of the entire banking system (Keeley, 1990). Impact: Competition among banks arising from the deregulation of the banking system will reduce the value of bank charter capital by reducing the benefits of monopoly power, and encourage banks to opt for more risky policies to maintain profit levels in previous periods 8(Keeley, 1990). However, Boyd and De Nicoló (2005) show that less competition among banks creates favorable conditions for setting higher interest rates on business loans, leading to increased credit risk of borrowers (ethical issues), and can lead to high nonperforming loan ratios and increased bank instability. 2.2.5 Theory on bank diversification Definition: Diversification can be referred to as an act of a company to extend its product range or fields of business with the aim of increasing operational efficiency and profitability or minimizing risks or both of two purposes. Impact: The first view is that product diversification can help the bank mitigate the risk of bankruptcy because it has allocated resources for different products (Haugen, 2001); The second view is that banks should focus on one type of product or service as a better operating strategy, rather than diversifying products and services. 2.2.6 “Too-big-to-fail” hypothesis Concept: "Too-big-to-fail" hypothesis refers to the relationship between bank size and bank risk. This theory states that banks or financial institutions are large-scale organizations associated with many other economic sectors, thus having a great influence on a country's economy. As a consequence, their bankruptcy risk will spell catastrophic trouble for the whole economy, so they need to be supported by the Government when they face distress and constraints. Impact: Government protection of large-scale financial institutions and banks may increase their risk-taking behavior (Mishkin, 1999). 2.3. Determinants of competition, credit risk and operating performance of commercial banks 2.3.1 Subjective factors: Financial capability, Managing capability 2.3.2 Objective factors: economic environment, political environment, and society 2.4. Literature review There are numerous researches on the competition, credit risk and operating performance of commercial banks in Vietnamese and international contexts. Nonetheless, the results tend to differ due to the differences in research scale, data sources, and research methods. Recently, banks have increasingly become aware of the importance of risks as well as the relationship between risk and profitability. A bank with proper risk management will 9have better resistance against and be less affected by unforeseen impacts, and be able to take timely actions to minimize losses to the bank. Integrating into international financial markets, commercial banks have to compete fiercely with each other to make profits, so the risks also increase and affect the operational efficiency. Ho and Saunders (1981) opine that the profits made by banks when setting interest rates for deposits and loans depend on (i) banks' risk aversion, (ii) market structure, (iii) the economy of scale, and (iv) interest rate fluctuations. Ho and Saunder (1981) argue that banks are still profitable even in highly competitive markets. Previous studies have shown the positive impact of multimarket contact on improving efficiency, service quality, innovation and international competition (Claessens & Laeven, 2004). However, recent studies indicate that the relationship between competition and performance is less obvious in the financial sector compared to other sectors (Akins et al., 2016). Numerous studies find that multimarket contact modifies the relationship between competition and bank efficiency and stability (Maudos et al., 2004). Over the past few decades, banks have focused more on strengthening their position and competitiveness in the context of continuous technological progress, globalization and pro-monopoly regulatory provisions being removed. This phenomenon can have marked effects on the revolution of the level of competition, because in more concentrated markets, businesses will capture information more quickly, facilitating the comparison of interest rates. In addition, mergers and acquisitions (M&A) have led to more overlap between branches, thus increasing the number of markets in which banks will have to compete with each other. Issues related to multimarket contact have been analyzed in theoretical and empirical works for decades, and most of these studies have been conducted in developed countries. According to oligopoly theory, multimarket companies will not compete intensively with their competitors in the anticipation of retaliation in other markets. Because the greater the number of markets that businesses operate together, retaliation in these markets is likely to occur and will cause significant losses. In such a situation, the overall level of competition will shrink. Bernheim and Whinston (1990) show that, when multimarket encountering takes place, businesses tend to have tacit agreements to cooperate rather than compete if there are imbalances between markets in which firms compete. However, there are some studies that do not support this theory, even showing that competition in the market becomes stronger when businesses have multimarket contact (Degl’Innocenti et al, 2014). 10 Some authors have investigated the influence of multimarket contact on the profitability of manufacturing enterprises, and shown that businesses lever on multimarket contact to increase profitability (Scott, 1982; Hughes and Oughton, 1993). Gimeno (2002) studies American airlines and Busse (2000) studies the mobile phone industry in the U.S. market, while Fu (2003) looked at news businesses in Central America. All these studies have confirmed that the multimarket contact has a positive impact on the corporate profitability, in line with the "mutual forbearance hypothesis". In the banking sector, there have been a number of empirical studies providing inconsistent results on the influence of multimarket contact on bank performance. Coccorese and Pellecchia (2009) show that multimarket contact increases profitability of Italian banks. Using Italian bank data in the period from 1997-2009, Coccorese and Pellecchia (2013) show that companies operating in many markets are prone to collude with each other. Fuentelsaz and Gomez (2006) study Spanish savings banks, again showing an inverted U-shaped relationship between multimarket contact and market entry. A number of studies have analyzed the impact of diversification in the banking sector on risks and profitability. Several studies have found evidence that diversifying income sources increases risk (De Young and Roland, 2001; Stiroh, 2004; Stiroh and Rumble, 2006; Lepetit et al., 2008; Demirgüc-Kunt and Huizinga, 2010). Studies by Gallo et al. (1996), Rogers and Sinkey (1999) and Ashraf et al. (2016) have found the opposite. Income diversification is thought to have a positive effect on profitability in Stiroh and Rumble (2006), Chiorazzo et al. (2008), Lepetit et al (2008), Demirgüc-Kunt and Huizinga (2010) and Elsas et al. (2010). However, income diversification tends to exert a negative impact on profitability (DeYoung and Roland (2001); Stiroh (2004), DeYoung and Rice (2004), Stiroh (2004), Baele et al (2007), Berger et al (2010) and Fiordelisi et al (2011)). For the banking sector, the results of empirical studies are more ambiguous, and it is impossible to make a indecisive opinion on the role of multi-market contact in the relationship between competition and profitability. In these studies, multimarket contact is often measured by calculating the number of contacts between a certain group of banks (all the banks operating in the market or a subset of organizations) or the amount of deposits related to those links, while the level of competition is assessed through the changes in indicators such as profit, loan income, cost related to deposits, and market share. 11 Income diversification is one of a bank's competitive strategies during times of fierce competition, forcing the bank to seek a wide variety of sources of income. Another strategy that has received much less attention is that a bank diversifies its business areas within a territory (region or country level), resulting in repeated contacts between banks, leading to enhanced competition. Competition in the banking market can lead to banks competing for customers, thus increasing the risk of low-quality customers being granted the loans, affecting the bank's credit risk. According to the literature review, until recently there has been no research analyzing the relationship between multimarket contact, competition and credit risk and performance of banks, at least in developing countries. Banks in Vietnam have grown rapidly in terms of number, scale and location of operations. The competition with domestic commercial banks has become more and more intense, leading to the trend of banks diversifying their operational activities, mainly services, to maintain market share. The banking system plays an important role in Vietnam's economy and contributes 16 to 18 per cent to economic growth (Stewart et al., 2016). Studies in Vietnam have chiefly been conducted on bank performance, but these studies often focus on the impact of income diversification on bank performance (Canh and Minh, 2014; Vinh and Mai, 2015; Quynh and Hau, 2016). In Vietnam, there have been research topics on competition including Trung (2010), Vinh (2015), Vinh and Tien (2017). It can be seen that in Vietnam’s context, the link between competition in terms of multimarket contact, operational risks affecting the performance of commercial banks in Vietnam has not been dissected. Although Vietnamese commercial banks are trying to diversify services, the options are still not adequate; non-credit services such as payment, investment, foreign exchange business and financial consultancy are underdeveloped; Modern banking products and services are only in the experimental or pilot phase. In addition, following State Bank implementing tight monetary policy, commercial banks have restricted lending and have continuously increased lending interest rates, traditional banking services (mobilization and lending of money) could have little impact on the bank income, leading to increased competition in banks in the form of their increased presence in different markets. 2.5. Measures of competition, multimarket contact, credit risk and operating performance 12 2.5.1. Risk and credit risk: According to Kieu (2012), risk is an uncertain or an unstable condition. The basic feature of a bank is the pursuit of acceptable benefits and measurable risks. Credit risk occurs when a customer is unable to repay their loan. According to the Basel Committee (2004), credit risk is referred to as unexpected events that result in losses in the value of assets, reduction of profits compared to expected returns, or the generation of additional expenses to complete a particular transaction. Credit risk occurs when the borrowers are unable to pay both the principal and the interest on time as stipulated in the contract. 2.5.2. Competition and multimarket contact Competition is the fighting for more market share by banks to commercialize a wide range of products and services, as well as to dominate the market to attain the most favorable conditions for the production, consumption and market. Multimarket contact: occurs when two or more businesses are competing with each other at the same time and in different markets (product or location). When these businesses are exposed many times and in many places to each other, multimarket contact may lead to a relationship called "mutual forbearance". Businesses compete with each other at the same time in a number of different markets. Edwards (1955) conducted the seminal study, theorizing that when businesses compete with each other, they are likely to encounter in significant markets and repeated exposure could reduce competition. 2.5.3. Performance and operating performance of banks Performance is the relative relation between the outcomes and the total cost spent for that outcome (input), and the greater the difference between the two quantities, the higher the performance. Bank operating performance is a reflection of the bank's use of resources to achieve its goals, showing the relation between output and input. Banks minimize costs to increase its competitiveness versus other financial institutions. 2.6. Research gap (i) There have been no studies analyzing the impact of the multimarket contact on the competition of Vietnamese commercial banks. (ii) There have been no studies analyzing the impact of multimarket on credit risk of 13 Vietnamese commercial banks. (i) There have been no studies analyzing the impact of the multimarket contact on the operating performance of Vietnamese commercial banks CONCLUSION OF CHAPTER 2 In this chapter, the author presented important theories related to the thesis, systematized the domestic and international studies, as well as related concepts related to multimarket contact, competition, credit risk and operating performance. CHAPTER 3: RESEARCH METHODOGY 3.1 Research model about Competition-related model, Risk-related model, Operating performance-related model 3.1.1 Competition-related model: The model is as follows: LERNERit = β0 + β1MMC1i,t + β2SIZEi,t + β3CAPi,t + β4SOCBi,t + β5LLPi,t+β6GDPt + β7INFt+ μit (MH1) - Dependent variable: competition (Lerner) - The independent variables that act as control variables are included in the model to reduce the noise in the model, and to simultaneously partition the impact of the multi- market contact on competition, including: multi-market contact (MMC1); Bank size (SIZE); Equity/total assets (CAP); State ownership (SOCB); Loan loss provision (LLP); Annual growth rate of GDP (GDP); Inflation (INF) 3.1.2 Risk-related model: The model is as follows : NPLit = β0 + β1MMC1i,t + β2SIZEi,t + β3DIVi,t + β4TLTAi,t + β5CAPi,t+ β6OETAi,t + μit (MH2) - Dependent variable (Risk): Credit risk is measured by NPL (Non-performing loan / Total loans). - The independent variables that act as control variables are included in the model to reduce the noise in the model and single out the impact of multimarket contact on risk, including: Multi-market contact (MMC1); Bank size (SIZE); Non-interest income / total income (DIV); Total Loans / Total Assets (TLTA); Equity / Total Assets (CAP); Total operating expenses / Total assets (OETA) 14 3.1.3 Operating performance-related model: The research model is as follows RAROAit = β0 + β1MMCi,t+ β2DEPOTAi,t+ β3DIVi,t+ β4TLTAi,t+ β4CAPi,t + β5OETAi,t+ μit (MH3) - Dependent variable (RAROA): Performance measure has been adjusted for risk. - The independent variables that act as control variables are included in the model to reduce the noise in the model and single out the impact of multimarket contact on operating performance, including: Multi-market contact (MMC1 and MMC2); Total deposits / total assets (DEPOTA); Non-interest income / total income (DIV); Total Loans / Total Assets (TLTA); Equity / Total Assets (CAP); Total operating expenses / Total assets (OETA). - The measurement of MMC1 and MMC2 is through the establishment of 3 matrices 3.1, 3.2, 3.3, in line with Coccorese and Pellechia (2009, 2013). First, calculate the matrix of the number of commercial bank branches in each provincial market, where k is the number of markets and dij is the number of branches of bank i in the market j. Next, matrix NxK is formed (C), and Cij representing whether bank i operates in market j. Regarding matrix C, cij = 1 if dij> 0 and cij = 0 if dij = 0. Therefore, cij = 1 means that bank i is active in market j, i.e., there is at least 1 branch in the market j. Then, matrix M (NxN) is calculated as follows: In which, CT is the inverse matrix of C. The components (mij) are the number of markets that banks i and j operate in, meaning that the number of markets that banks i and j concurrently have at least 1 branch. From these matrices, the author calculates an important indicator to measure the extent of multimarket contact as follows: (3.1) (3.2) (3.3) (3.4) 15 The lower and higher values of MMC1 depend on the distribution of banks among provinces. Theoretically, the minimum is zero, which occurs if a bank has the monopoly in the market where it operates, and maximum is equal to the number of provinces provided that all banks meet in all provincial markets. Therefore, for single-market banks, MMC1 is equal to 1 unless they are monopolies. Calculation of MMC2: Starting from the S matrix (NxK) with elements being the market share of bank i in province j The similarity index between the two banks is calculated by the sum of the absolute differences of market share in the provinces where the two banks meet: The SI index, in theory, runs from zero to the number of markets that banks come into contact with, and it will be small if the two banks are similar in terms of market share. To calculate an index that increases with the similarity of banks in terms of market share and runs from 0 to 1, we use the following conversion mechanism: The corresponding matrix for calculating the number of multi-market contacts becomes Finally, MMC2 is obtained as follows: 3.2.1. Comparison with the extant studies Similarity: The competition is measured by the Lerner index, in consistence with Vinh and Tien (2017) and Delis (2012), credit risk measured by NPL - similar to the study of Bana Abuzayed et al. (2018), and operating performance measured by RAROA of banks, (3.5) (3.6) (3.7) (3.8) (3.9) 16 which has been used in previous studies. Multimarket contact is measured using MMC1 and MMC2 according to Coccorese and Pellecchia (2009, 2013). The thesis also uses the intrinsic variables of banks related to credit activities and operating performance including: Bank size (SIZE), Non-interest income / total income (DIV), Total loans / Total assets (TLTA), Equity capital / Total assets (CAP), Total operating expenses / Total assets (OETA), state ownership (SOCB) measured as the ratio of state capital to total ownership; loan loss provision (LLP), annual growth rate of GDP (GDP), Inflation (INF) are used as the control variables in the model. Differences: The biggest difference of this study compared to previous studies is the application of the factor of multimarket contact (Coccorese and Pellecchia, 2009, 2013) to study how banks competing in the same market can exert influence on credit risk and operating performance measured by RAROA (Performance measure that has been adjusted for risks of Vietnamese commercial banks). Therefore, compared with the prior studies, the thesis focuses on understanding the impact of multimarket contact and the factors related to competition, credit risk and operating performance that no previous studies on Vietnamese commercial banks discussed. 3.2 Estimation method Study using unbalanced panel data and Stata 12.0 for data processing and the estimation using GMM for all the 3 models. Conclusion for Chapter 3: Chapter 3 discussed the research models examining the link between multimarket contact, credit risk and operating performance of Vietnamese commercial banks. 17 CHAPTER 4 : RESEARCH RESULTS AND DISCUSSION 4.1 Descriptive statistics Table 4.1:Descriptive statistics of variables Variable Obs Mean Std. Err Min Max RAROA 319 1.9561 1.509 0.012 6.293 NPL 311 0.0099 0.00993 -0.0099 0.1097 LERNER 311 0.424 0.091 -0.173 0.634 MMC1 319 14.600 7.542 1.000 34.148 MMC2 319 13.931 6.958 0.910 31.275 DIV 311 0.099 0.080 -0.195 0.468 TLTA 319 0.516 0.141 0.114 0.852 SIZE 319 17.616 1.403 13.135 20.590 LLP 310 1.409 2.547 0.013 43.969 OETA 319 0.016 0.007 0.000 0.069 CAP 319 0.126 0.094 0.011 0.661 SOCB 319 0.154 0.361 0.000 1.000 GDPGR 319 6.004 0.529 5.247 6.812 INF 319 8.193 6.305 0.879 23.116 Source: Author’s calculation using Stata 12 The variable RAROA is the risk-adjusted measure of operating performance in Table 4.1. The average value of the variable in the sample is 1.9561 with the lowest value of 0.012 of Tien Phong Commercial Joint Stock Bank in 2013 and highest value of 18,416 of Vietnam Joint Stock Commercial Bank for Industry and Trade in 2011, showing that there is a significant difference in risk-adjusted operating performance between commercial banks in the sample. Credit risk variable (NPL) has the average value of 0.0099, and standard deviation being 0.00993, the largest value 0.1097 and the smallest value -0.0099. This shows that the data have high homogeneity, and there is low variation in credit risk measured by NPL. The variable LERNER has the average value of 0.424, standard deviation 0.091, the largest value 0.634 and the smallest value -0.173. This shows that the data have poor similarities, and there are significant differences in competition variable between banks. Multi-market contact (MMC): The MMC1 variable measures the impact of multi-market exposure of banks, with a minimum value of 1 and a maximum value of 34,148 with a mean 18 value of 14,600 and standard deviation of 7,542. The gap between the largest and the smallest values is relatively large at 33,148 along with the high degree of dispersion around the mean value, suggesting that the commercial banks in the sample have a large difference in the level of multimarket contact. The MMC2 variable has the smallest value of 0.910 and the largest value 31.275 with an average value of 13,931 and a standard deviation of 6,958. MMC2 shows that the number of markets in which banks are exposed to each other is 13,931 (nearly 14 provinces). The MMC1 variable shows that the banks are exposed on average in 14,600 (more than 14) provinces. MMC2 is lower than MMC1, suggesting that the increasing presence of banks in different regions has caused a general decline in their average market share. 4.2 Results on the impact of multimarket contact on competition (MH1) Table 4.3: Regression result of MH1 Coeff Std. Err t P>t LERNER _L1. -0.1326 0.011 -11.52 0.000 MMC1 0.0037 0.001 1.96 0.049 SIZE 0.0248 0.013 1.79 0.073 CAP 0.3786 0.155 2.43 0.015 LLP 0.0015 0.001 0.54 0.124 SOCB 0.1468 0.039 0.37 0.712 GDP -0.0051 0.013 -0.38 0.706 INF 0.0047 0.000 6.52 0.000 _cons -0.1273 0.201 -0.47 0.640 AR(2) 0.96 Sargan-Hansan test 0.456 Source: Author’s calculation using Stata 12 4.3 Regression results on the impact of multimarket contact on credit risk (MH2) Table 4.5: Regression result of MH2 Variable Coeff Std. Err t P>t Biến Hệ số Sai số t P>t NPL_ L1 -.0975672 .0108035 -9.03 0.000 MMC1 .4788295 .0809334 5.92 0.000 DIV -7.41919 3.607396 -2.06 0.046 TLTA 13.89682 4.050023 -3.43 0.001 SIZE