Contents Journal of Multinational Financial ManagementC. Vithessonthi / J. of Multi. Fin. Manag. 27...

J. of Multi. Fin. Manag. 27 (2014) 67–88

Contents lists available at ScienceDirect

Journal of Multinational FinancialManagement

journal homepage: www.elsevier.com/locate/econbase

Financial markets development and bank risk:Experience from Thailand during 1990–2012

Chaiporn Vithessonthi ∗

Department of Accountancy and Finance, School of Business, University of Otago, PO Box 56, Dunedin 9054,New Zealand

a r t i c l e i n f o

Article history:Received 26 February 2014Accepted 21 May 2014Available online 4 June 2014

JEL classification:G15G21G32O16

Keywords:Bank capitalizationBank riskFinancial markets developmentFinancial crisesThailand

a b s t r a c t

The relation between financial markets development and bank riskin Thailand during 1990–2012 is examined. After controlling formacro-level and firm-level variables, stock market development ispositively associated with banks’ capitalization ratio, and is nega-tively related to their beta. While banking sector development hasno effect on the banks’ capitalization ratio, it has a positive effecton their beta. In addition, banking sector development is negativelyrelated to the banks’ capitalization ratio when measured as theTier 1 capital to total risk-weighted assets ratio during 2000–2012.Overall, two dimensions of financial markets development seem tohave opposing effects on bank risk. While stock market develop-ment tends to lower the banks’ beta, banking sector developmentinduces the instability of the banking system by lowering the banks’capitalization ratio and by increasing the banks’ beta.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

In this paper, I analyze the relation between financial markets development and three dimensionsof bank risk – the capitalization ratio,1 revenue diversification, and market beta. The longitudinal

∗ Tel.: +64 3 479 9047; fax: +64 3 479 8171.E-mail address: [email protected]

1 It should be noted that leverage and capitalization of a bank are two sides of the same coin. In the banking literature, itis traditional to use a bank’s capitalization ratio as a measure of its leverage. Thus, a highly leveraged bank would typically

http://dx.doi.org/10.1016/j.mulfin.2014.05.0031042-444X/© 2014 Elsevier B.V. All rights reserved.

dx.doi.org/10.1016/j.mulfin.2014.05.003

http://www.sciencedirect.com/science/journal/1042444X

http://www.elsevier.com/locate/econbase

http://crossmark.crossref.org/dialog/?doi=10.1016/j.mulfin.2014.05.003&domain=pdf

mailto:[email protected]

dx.doi.org/10.1016/j.mulfin.2014.05.003

68 C. Vithessonthi / J. of Multi. Fin. Manag. 27 (2014) 67–88

study aims to understand the impact of the evolution of financial markets development bank behav-ior. The focus on Thailand during the period 1990–2012 therefore captures the events surroundingthe Asian financial crisis of 1997–19982 and the global financial crisis of 2007–2009. Several stud-ies have highlighted the benefits and costs of financial markets development. Essentially, financialmarkets development is thought to lead to economic growth by providing firms with better access tofinancing, thereby promoting domestic investments. Financial markets development (e.g., changes inbank regulations) can also affect the way in which banks operate and behave which, in turn, may haveeither positive or negative ramifications for the health of the banking system and the stability of thefinancial system.

This paper adds to the literature by analyzing the changes in financial markets development and thechanges in capitalization ratios of commercial banks in Thailand over the period 1990–2012. Duringthis period, there were two major financial crises – the Asian financial crisis of 1997–1998 and theglobal financial crisis of 2007–2009 – both of which directly affected emerging market countries. Whilethe former mainly affected Asian countries, the latter had global impact. Secondly, the paper addressesthe relationship between financial markets development and bank risk by addressing the followingthree questions: Does financial markets development encourage banks to be more leveraged? Doesfinancial markets development affect banks’ portfolio of assets? Did financial reforms implementedafter the Asian financial crisis of 1997–1998 cause banks to be better prepared for future crises?

I find that a measure of stock market development is positively associated with a bank’s capital-ization ratio (measured as total capital/total assets), after controlling for macro-level and firm-levelvariables and is negatively associated with a bank’s risk (implied by a bank’s market beta from themarket model). While the measure of banking sector development is not associated with a bank’scapitalization ratio when measured as the total capital to assets ratio (CTA), but it is negatively relatedto a bank’s capitalization ratio when measured as the Tier 1 capital to total risk-weighted assets ratio(CART1).3 In terms of the economic significance, my estimated slope coefficient on the banking sectordevelopment measure (BSD) of −0.075 (see Model 1 of Table 8) implies that a one-standard deviationincrease of BSD will result in a 0.133 (=17.608 × −0.075/9.954) fall in CART1 at the mean (of 9.954) dur-ing 2000–2012. I conduct further tests to show that the positive effect of stock market developmenton banks’ capitalization ratios mainly appears during 1990–1999, and that this linkage disappearsduring 2000–2012. In the context of my analysis, two dimensions of financial markets developmentseem to have opposing effects on bank risk. While stock market development tends to have no adverseeffect on banks’ capitalization ratios and lower the level of a bank’s beta, banking sector developmentinduces the instability of the banking system by lowering the CART1 ratios and by increasing the levelof the beta. In addition, I find that both stock market development and banking sector developmentmeasures are not associated with a bank’s revenue diversification (measured as the share of the non-interest revenue to net revenue). It is important to note that several studies provide evidence for apositive relation between non-traditional banking activities and profitability. For example, Apergis(2014) shows that there is a positive relation between the non-core banking activities and profitabil-ity (e.g., ROA) and that the non-core banking activities is positively associated with insolvency risk. Inaddition, in the context of the Philippines, Meslier et al. (2014) also show that the variation in bankrevenue diversification is positively associated with the level of profitability. Hence, the finding of norelation between both measures of financial markets development and a bank’s revenue diversifica-tion in this study seems to suggest that high levels of financial markets development do not necessarilydestabilize the financial system with respect to the banks’ non-core banking activities. However, theresults regarding the effect financial reforms following the Asian financial crisis are mixed. On theone hand, the depth of stock markets seems to lower the overall risk of the banking sector; on theother hand, the banking sector development tends to increase the risk of the banking sector. The net

refer to a bank with a low capitalization ratio (e.g., capital adequacy ratio). This notion is different to those used in the generalfinance literature where a firm’s leverage is usually measured as a ratio of long-term debt (or total debt) to assets; hence, ahighly leveraged firm would then refer to a firm with a high leverage ratio.

2 Thailand receives the IMF financial assistance programs during the Asian financial crisis of 1997–1998.3 The Tier 1 capital to total risk-weighted assets ratio for Thai banks is available from 2000 onwards.

C. Vithessonthi / J. of Multi. Fin. Manag. 27 (2014) 67–88 69

effect appears to be negative; that is, the adverse effect of the banking sector development tends tooutweigh the positive effect of the depth of stock markets during the sample period.

My sample period begins in 1990 and ends in 2012, thereby covering not only two financial crisis(i.e., the Asian financial crisis of 1997–1998 and the global financial crisis of 2007–2009) periods andtranquil periods, but also periods with different stock market conditions and different business cycleconditions. As noted by Noy (2004), the domestic financial liberalization, defined as the removal ofinterest rates control on bank deposits, in Thailand had taken place about 10 years before the onsetof the financial crisis in 1997. Furthermore, several Thai governments in the late 1980s and early1990s implemented several initiatives with the goals of making Thailand as a regional financial hub,thereby removing a number of restrictions during this period. Overall, financial markets in Thailandhave been liberalized during the study period, thereby providing an ideal setting to document therelation between financial markets development and a bank’s behavior in a small open emergingmarket economy. More importantly, this study will indirectly provide insights into the effectivenessof a series of financial and regulatory reforms after the Asian financial crisis of 1997–1998, in the sensethat the reforms should better mitigate the adverse effects of the global financial crisis of 2007–2009on banking system stability. To some extent, the findings and insights presented in this paper should beapplicable to emerging market countries that share similar key characteristics: (1) small and medium-sized open economics, (2) moderate or high degrees of financial openness, (3) moderate or high degreesof trade openness, and (4) bank-based economies.4 While prior studies on banks in emerging marketcountries, for example, Bourgain et al. (2012) who examine the risk-taking behaviors of banks in theMiddle East and Northern Africa region, Soedarmono et al. (2013) who investigate the impact of bankcompetition on bank risk taking of banks in Asia, and Meslier et al. (2014) who examine the effectof bank revenue diversification on bank profitability in the Philippines, provide new and importantinsights into the behaviors of banks in emerging market countries, they do not examine the impact offinancial markets development on bank risk. In this respect, this study provides additional insights tothe literature.

My paper builds on and contributes to prior literature on financial development (Bena and Ondko,2012; Calderon and Liu, 2003; Campos et al., 2012; Noy, 2004; Zagorchev et al., 2011) as well as onbank risk (Angkinand and Wihlborg, 2010; Beck et al., 2013; Calmès and Théoret, 2013; Claessenset al., 2013; De Jonghe, 2010; Iannotta et al., 2013; Stiroh, 2006). In a closely related study, Williamsand Nguyen (2005) find that the liberalization in the banking sector of Southeast Asian countries isassociated with higher efficiency performance of banks between 1990 and 2003. My main results arerobust to a number of different specifications. In addition, I address the extreme value concern by(1) removing observations with extreme values, at the cost of reducing the sample size and, moreimportantly, excluding some observations corresponding to the financial crises and (2) employing thequantile regression approach that deals with extreme values.

To summarize, the main empirical results are as follows: First, two measures of financial marketsdevelopment generally have the opposing effects on a bank’s capitalization ratio and beta. Second,stock market development, on average, has no “adverse” effect on banks. That is, it does not lowercapitalization ratios but reduces the market beta. The finding of no relation between stock marketdevelopment and the banks’ capitalization ratio is critical, as recent studies, for example, Imbierowiczand Rauch (2014), show that the overall risk in the banking industry, measured as the banking sector’saverage leverage, increases a bank’s probability of default. The finding of the negative relation betweenstock market development and the banks’ beta suggests that the depth of stock markets enhancesthe overall stability of the banking sector. Third, the measure of banking sector development is notassociated with a bank’s revenue diversification and is positively related to a bank’s beta. While ithas no effect on a bank’s capitalization ratios (measured as total capital to total assets) during the fullperiod sample, it has a negative effect on the CART1 ratios during 2000–2012. These findings pointto the notion that the development of the banking sector tends to increase the overall instabilityof the banking system. As Bourgain et al. (2012) show that higher degrees of financial openness are

4 There are a number of emerging market countries with these characteristics. Examples of these countries are Argentina,Malaysia, Mexico, South Africa, and Brazil.


associated with higher levels of bank financial leverage in emerging market countries during the period2005–2008, the finding of the negative relation between banking sector development and CART1 maybe a worrying sign for bank supervision authorities because it appears to suggest that the experienceof the Asian financial crisis does not induce banks to be more conservative/prudent with respect totheir capital structure. A potential explanation for this observation is the presence of the full depositinsurance coverage that appears to not only encourage a bank’s excessive risk-taking but also lowerthe presence and effectiveness of market discipline (e.g., monitoring by depositors and investors). Inthe case of Thailand, the full deposit insurance coverage was first introduced during the Asian financialcrisis and remains in place as of the end of 2012; however, the removal of the full deposit insurancecoverage in Thailand will soon be in effect.

2. Brief literature review

Several scholars, such as Chinn and Ito (2006), provide empirical evidence suggesting that financialopenness results in financial markets development. Campos et al. (2012) show that in the context ofArgentina during the period 1896–2000, financial liberalization has had a long run positive effect oneconomic growth albeit a short run negative effect on economic growth. Calderon and Liu (2003) andZagorchev et al. (2011) find that financial markets development leads to economic growth, while Islamand Mozumdar (2007) show that financial markets development seems to reduce the extent to whichfirms are dependent on internal capital for corporate investment. The latter suggests that financial mar-kets development promotes the use of external finance as a substitute for internal capital. In addition,Bena and Ondko (2012) show that financial markets development promotes the efficient allocation ofresources (e.g., external finance is being used more by firms in industries with growth opportunities incountries with higher levels of financial markets development). Nevertheless, prior studies also revealadverse effects of banking sector development. For instance, Gimet and Lagoarde-Segot (2011) findthat banking sector development leads to income inequality in a sample of 49 countries during theperiod 1994–2002. Furthermore, Festic et al. (2011) find that the growth of credit has a negative effecton the quality of the bank’s portfolio of assets (measured by the NPL ratio).

As noted by Iannotta et al. (2013), there are several ways to measure the risk profile of a bank. Tra-ditionally, a bank’s risk is measured by using accounting ratios that evaluate numerous aspects of thebank (e.g., liquidity, capitalization, and asset quality).5 The level of bank capitalization is traditionallyconsidered an important factor that helps reduce the probability of a bank’s failure.

While financial markets development promotes economic growth, it also provides banks withgrowth opportunities by increasing the demand for bank loans from the private sector. To make moreloans to firms, a bank should simultaneously increase its capitalization. A key question arises as towhether financial markets development results in a lower level of bank capitalization ratios, all otherthings being equal. In other words, does financial markets development have an adverse effect on thebank capitalization ratio?

After the Asian financial crisis of 1997–1998, several scholars attribute the banking crisis to severalfactors, such as, the moral hazard problem as well as the inefficient bank supervision in the crisiscountries. Some scholars, such as Stiglitz (2000), argue that financial liberalization in emerging marketcountries subject these countries to banking and/or financial crises. All of these arguments point tothe prediction that financial markets development has a negative effect on bank capitalization. Thatis, the degree of financial markets development is negatively associated with a bank’s capitalizationratio (measured as the ratio of total capital to total assets). The negative relation between financialmarkets development and bank capitalization suggests that financial markets development leads tobanking system instability.

As financial liberalization that unintentionally motivates and enables banks to take (excessiveor undue) risk in the absence of efficient supervision is likely to result in a banking crisis (Noy,2004), an important question is whether financial markets development contributes to the growth of

5 Several scholars, such as Haq and Heaney (2012), have measured a bank’s risk using the market-based variables (e.g., thevolatility of a bank’s stock return, a bank’s stock beta).


noninterest income. Several scholars, such as De Jonghe (2010), suggest that a bank’s diversificationof revenue streams (e.g., the share of the noninterest income to total income) tends to increase thesystematic risk6 of the bank. In addition, Bikker and Metzemakers (2005) note that the increased shareof noninterest income in a bank’s revenue portfolio increases the volatility of the bank’s accountingprofits. It is possible that financial markets development affects a bank’s revenue diversification byencouraging the higher growth of the noninterest income from the non-traditional banking activities,relative to the growth of the interest income from traditional banking activities. If the bank’s revenuediversification accurately represents the level of the bank’s risk, a positive linkage between financialmarkets development and bank revenue diversification would imply that financial markets devel-opment contributes positively to the banking system risk. Furthermore, as noted by Angkinand andWihlborg (2010), there exists full deposit insurance coverage in several countries such as Ecuador,Indonesia, Malaysia, and Thailand, which may in turn encourage a bank’s risk taking behavior and, atthe same time, reduce bank monitoring by depositors.

3. Methodology

3.1. The effect of financial markets development on bank capitalization

I test the hypothesis that financial markets development has an adverse effect on the level of bankcapitalization by encouraging a bank to make more loans that are financed by short-term funding(e.g., deposits). If a central bank or a government makes an implicit guarantee to bail out financialinstitutions during banking/financial crises, then the risk profile of a bank will be greater. In line withthe work of Calmès and Théoret (2013), my main equation is as follows:

LEVi,t = + ˇFDFDt−1 + ıMCONt−1 + �BCONi,t−1 + �i + εi,t, (1)

where LEVi,t refers to a proxy for a capitalization ratio for bank i at time t; FDt denotes the indicator of acountry’s financial markets development at time t. MCONt is a vector of country-level control variablesat time t that may affect a bank’s capitalization ratio; BCONt is a vector of bank-level control variablesfor bank i at time t that may be associated with a bank’s capitalization ratio; vi is a firm fixed-effectsvariable; and, εi,t is the zero-mean error term. I include individual firm fixed effects to control for anyunobserved time-invariant firm effects.

As values for the Tier 1 capital to total risk-weighted assets ratio (CART1) are not available for mysample of banks prior to 2000, in order to have an appropriate measure of a bank’s capitalization ratiofor the full sample period, I simply compute the capitalization ratio (CTA) as the ratio of total capitalto total assets (in %). A bank that is well capitalized is relatively more resilient to banking systemicshocks. The summary statistics for CTA is given in Table 1.

To measure the degree of a country’s financial markets development, I use two measures that havebeen frequently used in the literature.7 First, I use stock market development (SMD), measured asthe ratio of stock market capitalization to GDP (in %). The SMD variable measures the depth of stockmarkets. As discussed in the literature, more developed stock markets provide an alternative externalfinancing source to firms and lower the extent to which firms rely on banks as a source of externalfinancing. Second, I use banking sector development (BSD), measured as the ratio of domestic creditprovided by banking sector to GDP (in %).8 If financial markets development positively influencesa bank’s behavior, I would expect ˇFD in Eq. (1) to be positive and significant. That is, banks in moreadvanced financial markets become well capitalized, which would in turn enhance the banking systemstability. On the other hand, if financial markets development enhances a bank’s risk-taking behavior,

6 In recent years, scholars, such as De Jonghe (2010), define the bank’s systematic risk as the sensitivity of its stock prices tochanges in stock market returns.

7 See e.g., Chinn and Ito (2006), Gimet and Lagoarde-Segot (2011), Xiao and Zhao (2012), and Becerra et al. (2012). Somescholars (e.g., Becerra et al., 2012) use the ratio of private credit to GDP as a measure of a country’s financial development.

8 Given that a bank’s leverage is a variable of interest in this paper, the use of BSD as a main proxy for a country’s financialmarkets development may be problematic because CTA and BSD may be highly correlated. As shown in Table 1, the correlationcoefficient between the two variables is positive and statistically significant.


Table 1Summary statistics of key variables.

Variables Mean Median Std. dev. Min Max N

TRADE 118.34 124.92 25.93 75.78 150.33 120FDI 3.12 2.85 1.31 1.23 6.54 120SMD 58.45 54.87 25.87 15.60 104.79 120BSD 135.36 131.58 22.67 94.08 177.58 120SIZE 13.52 13.54 0.65 11.55 14.70 120LIQ 3.16 2.78 1.79 0.80 10.96 120NLOAN 99.20 98.83 10.95 73.20 137.38 120LLR 5.47 4.59 4.00 0.71 24.07 120ROA 0.87 1.49 2.19 −9.14 3.14 120ROE 15.97 21.09 94.09 −546.07 635.36 120CIR 92.38 82.37 31.56 60.28 262.57 120CTA 11.56 11.36 3.92 4.76 33.03 120SNONIN 18.07 17.11 9.69 3.20 48.01 120BETA 1.20 1.20 0.27 0.55 1.81 120

This table presents summary statistics on key variables. TRADE is a measure of a country’s trade openness, measured as the ratioof the sum of total exports and imports to GDP (in %); FDI denotes a country’s financial openness (FDI), computed as the ratioof the value of foreign direct investment to GDP (in %); SMD refers to stock market development, measured as the ratio of stockmarket capitalization to GDP (in %); BSD represents a measure of banking sector development, defined as the ratio of domesticcredit provided by banking sector to GDP (in %); SIZE is the logarithm of total assets; LIQ is the ratio of cash to total deposits(in %); NLOAN is the net loans ratio, measured as the ratio of net loans to total deposits (in %); LLR is the loan loss reserve ratio,computed as the ratio of loan loss reserve to gross loans (in %); ROA denotes return on assets, measured as the ratio of EBITto total assets (in %); ROE is return on equity, measured as the ratio of EBIT to equity (in %); CIR is the cost-to-income ratio,measured as the ratio of operating expenses to net revenue (in %); CTA denotes the capitalization ratio, measured as the ratioof total capital to total assets (in %); SNONIN denotes a bank’s revenue diversification, measured as a share of the noninterestincome to net revenue (in %); and BETA is the estimated market beta from the market model (ri,w = + ˇmrmi,w + εi,w).

I would expect ˇFD in Eq. (1) to be negative and significant. The negative effect of financial marketsdevelopment on bank capitalization will have a negative effect on the health of the banking system.

I use two country-level control variables to control for macroeconomic effects on a bank’s cap-italization. First, I include a country’s trade openness (TRADE), measured as the ratio of the sum oftotal exports and imports to GDP (in %) to control for the effect of a country’s trade openness on bankcapitalization. Second, as it is possible that a country’s financial openness may affect bank capitaliza-tion, I include a country’s financial openness (FDI), computed as the ratio of the value of foreign directinvestment to GDP (in %), as a control variable. My relatively small sample size primarily dictates thechoice and number of control variables.

To measure the ex post risk of the bank, I intend to use a nonperforming loan ratio (NPLR), measuredas the ratio of nonperforming loans to gross loans (in %).9 Higher NPLR values will indicate that banksinvest in risky assets that eventually turn out to be nonperforming. However, due to missing valuesfor NPLR during the early period of the study, I instead use the loan loss reserve ratio (LLR), computedas the ratio of loan loss reserve to gross loans (in %). To control for the effect of size, I include firm size(SIZE), measured as the logarithm of total assets (in million THB), in all regressions.

Consistent with Cardone-Riportella et al. (2010), I use net loans to total deposits (NLOAN) ratio (in%) to proxy for a bank’s liquidity. In addition, I use a liquidity ratio (LIQ), measured as the ratio of cashto total deposits (in %), to measure the bank’s liquidity. Both measures are indicators of the risk profileof the bank (e.g., with respect to the ability to withstand a bank run or a liquidity shock). In line withthe banking literature, I use three measures to gauge bank performance, namely: (1) return on assets(ROA), measured as the ratio of EBIT to total assets (in %); (2) return on equity (ROE), measured as theratio of EBIT to equity (in %); and (3) cost-to-income ratio (CIR), measured as the ratio of operatingexpenses to net revenue (in %).

To control for the possible endogeneity issue and to establish the causality between financial mar-kets development and bank capitalization, all explanatory variables are one-period lagged. By lagging

9 The NPLR ratio and the ratio of risky assets to total assets have been used by Delis and Kouretas (2011) to measure therisk-taking behavior of a bank. Similarly, Fiordelisi et al. (2011) use the NPL ratio to measure a bank’s risk.


the explanatory variables, concerns arising from reverse-causality should be substantially alleviated.In addition, theoretically, it is unlikely that the measures of a bank’s risk (e.g., the capitalization ratio)at the firm level affect a measure of the financial markets development (e.g., the ratio of stock marketcapitalization to GDP) at the macro level. I estimate Eq. (1) using firm fixed-effect models with White’sheteroscedasticity-consistent cross-section standard errors and covariance.

It should be noted that using this methodology, I could not entirely differentiate the role of anunobserved factor unmasked (e.g., the effectiveness of banking supervision) by the financial marketsdevelopment variables from my hypothesis predicting that financial markets development affectsbanks’ risk.

3.2. The effect of financial markets development on diversification of bank revenue

Presumably, financial markets development not only influences a bank’s financing behavior but alsoaffects its revenue streams. Given new opportunities provided by more developed financial markets ina country, a bank may engage in other non-core banking activities, in addition to its traditional bankingactivities. Prior literature shows that a diversification of a bank’s businesses tends to be discounted bythe market (see e.g., DeLong, 2001; Laeven and Levine, 2007). To test the prediction that financial mar-kets development affects a bank’s revenue diversification (SNONIN), I estimate the following baselineregression:

SNONINi,t = + ˇFDFDt−1 + ıMCONt−1 + �BCONi,t−1 + �i + εi,t, (2)

where, similar to Stiroh (2006) and Calmès and Théoret (2013), SNONIN represents a bank’s revenuediversification (or its exposure to noninterest revenue), measured as a share of the noninterest incometo net revenue (in %). This measure is also known as the “market-oriented banking” variable in somestudies. If financial markets development affects a bank’s revenue diversification, I would expect ˇFD

in Eq. (2) to be positive and significant. The magnitude of this coefficient will indicate the degree ofeconomic significance of financial markets development on the revenue diversification of banks.

3.3. The effect of financial markets development on a market-based measure of bank risk

Recent literature suggests that the riskiness of a bank can be measured by using stock price infor-mation. For instance, Nijskens and Wagner (2011) use a bank’s beta estimated from an augmentedCAPM model to measure the risk, whereas Stiroh (2006) infers a bank’s risk using estimates from themarket model. For the purpose of this study, I estimate the following market model separately foreach back for each year:

ri,w = + ˇmrmw + εi,w, (3)

where ri,w is the return on stock i (i = 1,. . ., N) for week w, rmw is the stock market return for week w,and εi,w is an error term. By estimating Eq. (3) for each bank for each year, I construct a times-seriesof betas that represent the market-based measure of bank risk. If the market beta (BETA) accuratelyrepresents the perceived risk of a bank and financial markets development is associated with bankrisk, the significant effect of financials market development on bank risk, measured by BETA, shouldbe observed in the estimation of Eq. (1) using BETA as the dependent variable: a positive (negative)coefficient of the FD variable implies that financial markets development is associated with a higher(lower) degree of bank risk.

4. Data

My accounting data, macroeconomic data, and stock price data are primarily from Datastreamand Worldscope, supplemented by the Bank of Thailand. The sample period is 1990–2012. My goalis to include Thai commercial banks that were in operation during both the Asian financial crisis of1997–1998 and the global financial crisis of 2007–2009 in the sample. The cost is a rather small sampleof firms with potentially small cross-sectional variations, but this problem can be mitigated, to someextent, by using panel regressions. However, the key benefit is a relatively long time-series dataset


for which one can analyze possible structure changes in banks’ behavior after experiencing the Asianfinancial crisis of 1997–1998. I compile a data set with annual balance sheet and income statementinformation for all Thai banks that are in operation throughout the full sample period.10

To construct my sample, I first include all Thai banks that are publicly listed in the Stock Exchangeof Thailand (SET) at the end of 2012. In the second step, I remove banks that were not publicly listed inthe SET at the end of 1990. The beginning period is chosen for two reasons: First, it would allow for aseven-year period prior to the 1997 financial crisis. Second, several variables of data for my sample arenot provided by Datastream prior to 1990. After applying the selection procedure, I have unbalancedannual panel data from six Thai commercial banks (consisting of Bangkok Bank, Bank of Ayudhya,Kasikornbank, Siam Commercial Bank, Krung Thai Bank, and TMB Bank) over the 1990–2012 period.11

At the end of 1991, the six banks in the sample held a combined total of 1,498,349 million THB inassets, relative to the aggregate banking assets in Thailand of 2,360,135 million THB (or about 93,360million USD at the THB/USD exchange rate of 25.28 at the end of 1991), while at the end of 2012,these six banks combined held the total assets of 10,795,073 million THB, compared to the aggregatebanking assets of 14,773,937 million THB (or about 482,966 million USD at the THB/USD exchangerate of 30.59 at the end of 2012). This observation is in line with Soedarmono et al. (2013), who findthe Lerner index for Thailand is low during 1998–2004, in comparison with the pre-1998 period andthe post-2004 period.12 A low value for the Lerner index is associated with a decline in market powerin banking. Over the study period, the share of the aggregate banking assets in Thailand held by thesix banks combined is 70% on average, varying from the lowest of 62% in 1997 to the highest of 75%in 1992. Hence, the six banks in the sample reasonably represent the Thai banks’ behaviors during1990–2012. For comparison purposes, the six largest Canadian banks in the study of Gauthier et al.(2012) hold more than 90% of the aggregate banking assets in Canada.

Table 1 reports key summary statistics for my variables. The average value for the CTA variable is11.56, while the average value for the SNONIN variable is 18.07%. Apart from LLR, NLOAN, and ROE, allother variables appear to have no extreme values.13 Notably, large absolute values for these variablesare associated with the Asian financial crisis.

To provide a rough idea of the time-series variation in CTA, SMD, and BSD, I calculate the meanCTA for each year. Fig. 1 presents the time-series pattern of stock market capitalization (% of GDP),total commercial bank assets (% of GDP), total commercial banks’ equity to asset ratios (in %), andthe average ratio of total capital to assets for the sampled banks (in %) by year during 1990–2012. Iplot stock market capitalization (% of GDP) and total commercial bank assets (% of GDP) on the lefty-axis. The right y-axis measures total commercial banks’ equity to asset ratios (in %), and the averageratio of total capital to assets for the sampled banks (in %). This figure suggests an overall upwardtrend in the measure of banking sector development. However, it also points out the variation inthe level of stock market development. During the Asian financial crisis of 1997–1998, stock marketcapitalization (% of GDP) drops sharply. However, the global financial crisis of 2007–2009 appearsto have a relatively weak effect on stock market capitalization. This figure also shows that the Asianfinancial crisis of 1997–1998 does not seem to have a long-term effect on the development of thebanking sector. Indeed, the ratio of total commercial bank assets to GDP at the end of 2012 is almostthree times larger than that of the end of 1990, confirming the importance of the banking system tothe economic development of the country. With respect to bank capitalization, apart from the Asianfinancial crisis of 1997–1998 period, the average total capital to assets ratio for the six commercialbanks in my sample is closely in line with the average equity to asset ratio for all commercial banksin Thailand over the 1990–2012 period.

10 Datastream and Worldscope do not provide quarterly accounting data for the sample firms.11 Bangkok Bank and Kasikornbank are 2 of 22 banks included in a multicounty study of Huang et al. (2012).12 According to the work of Soedarmono et al. (2013), the Lerner index for Thailand not only increases during 2005–2009

but also is the highest among Asian countries during 2005–2009, implying that banks in Thailand increase their market powerduring this period which in turn helps them increase profitability.

13 That is, the minimum value and the maximum value of the variables are within the range of one standard deviation belowand above the mean value.


Fig. 1. Stock market development, banking sector development, and bank capitalization in Thailand during 1990–2012. Theleft vertical axis of this figure presents stock market capitalization (% of GDP) and total commercial bank assets (% of GDP) byyear, while the right vertical axis presents total commercial banks’ equity to asset ratios (in %) and sample banks’ total capitalto asset ratios (in %) by year.

Table 2 reports correlation coefficients for key variables. A measure of a bank’s financial capitaliza-tion (CTA) is positively correlated with trade openness (TRADE), bank size (SIZE), net loan rates (NLOAN),stock market development (SMD), banking sector development (BSD), and bank revenue diversifica-tion (SNONIN), but the magnitude of the coefficients is below 0.50. Trade openness is highly correlatedwith bank size (r = 0.67, p-value < 0.01) and bank revenue diversification (r = 0.70, p-value < 0.01), sug-gesting the possible collinearity issue when estimating Eqs. (1) and (2). As expected, the correlationbetween ROA and CIR is highly negative and statistically significant (r = −0.95, p-value < 0.01), suppor-ting the argument that one measure of performance should be separately entered into the regressionto avoid the multicollinearity issue.

5. Empirical results

5.1. The effect of financial markets development on bank capitalization ratio

I present regression estimates of Eq. (1), using the capitalization ratio (measured as total capital toassets) as the dependent variable, in Table 3. In all specifications, all explanatory variables are one-year lagged, and all firm-level and macroeconomic-level control variables are included. The results inTable 3 suggest that the net loan ratio (NLOAN) has a positive effect on the capitalization ratio, as thecoefficient on NLOAN is positive and statistically significant at the 1% level in all models. Consistentwith prior literature (Houston et al., 2010), bank size has a positive effect on the capitalization ratio. Allthree measures for bank performance are not associated with the capitalization ratio. With respect tothe model fit, the value of the model R2 in all specifications is around 0.47, which is relative higher thanthe model R2 of about 0.22 reported by Houston et al. (2010). While the direct comparison betweenthe two studies should be made with caution, it is important to note that a larger set of country-leveland bank-level control variables are included in their studies, relative to mine.

76

C. V

ithessonthi /

J. of

Multi.

Fin. M

anag. 27

(2014) 67–88

Table 2Correlation coefficients.

Variable TRADE FDI SMD BSD SIZE LIQ NLOAN LLR ROA ROE CIR CTA SNONIN

FDI 0.40***

SMD 0.32*** −0.11BSD 0.18* 0.13 0.02SIZE 0.67*** 0.23** 0.26*** 0.42***

LIQ −0.44*** −0.47*** 0.27*** −0.30*** −0.33***

NLOAN −0.09 −0.43*** 0.23** 0.20** 0.04 0.39***

LLR 0.32*** 0.49*** −0.11 0.16* 0.30*** −0.42*** −0.45***

ROA 0.05 −0.55*** 0.35*** −0.26*** 0.11 0.37*** 0.31*** −0.50***

ROE −0.02 −0.16* 0.03 −0.09 −0.01 0.06 0.02 0.00 0.08CIR −0.14 0.38*** −0.35*** 0.23** −0.21** −0.26*** −0.17* 0.39*** −0.95*** −0.03CTA 0.49*** −0.01 0.34*** 0.36*** 0.37*** 0.05 0.49*** 0.12 0.12 0.03 −0.07SNONIN 0.70*** 0.13 0.41*** 0.16* 0.60*** −0.21** 0.02 0.23** 0.23** 0.10 −0.29*** 0.47***

BETA 0.25** 0.41*** −0.17* 0.38*** 0.27*** −0.51*** −0.21** 0.48*** −0.33*** 0.01 0.29*** 0.16* 0.10

This table report correlation coefficients on key variables. TRADE is a measure of a country’s trade openness, measured as the ratio of the sum of total exports and imports to GDP (in %);FDI denotes a country’s financial openness, computed as the ratio of the value of foreign direct investment to GDP (in %); SMD refers to stock market development, measured as the ratioof stock market capitalization to GDP (in %); BSD represents a measure of banking sector development, defined as the ratio of domestic credit provided by banking sector to GDP (in %);SIZE is the logarithm of total assets; LIQ is the ratio of cash to total deposits (in %); NLOAN is the net loans ratio, measured as the ratio of net loans to total deposits (in %); LLR is the loanloss reserve ratio, computed as the ratio of loan loss reserve to gross loans (in %); ROA denotes return on assets, measured as the ratio of EBIT to total assets (in %); ROE is return on equity,measured as the ratio of EBIT to equity (in %); CIR is the cost-to-income ratio, measured as the ratio of operating expenses to net revenue (in %); CTA denotes the capitalization ratio,measured as the ratio of total capital to total assets (in %); and SNONIN denotes a bank’s revenue diversification, measured as a share of the noninterest income to net revenue (in %); andBETA is the estimated market beta from the market model. N = 120 bank-year observations.

* Significance at the 10% level.** Significance at the 5% level.

*** Significance at the 1% level.


Table 3Panel OLS regressions of a bank’s capitalization ratio during 1990–2012.

Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Intercept −44.234*** −44.090*** −45.226*** −39.305*** −40.186*** −40.757***

(13.635) (10.965) (11.651) (13.973) (10.994) (12.270)TRADE 0.032 0.034 0.029 0.030 0.027 0.027

(0.031) (0.022) (0.031) (0.031) (0.022) (0.031)FIN 0.224 0.207 0.251 0.207 0.246 0.244

(0.398) (0.287) (0.381) (0.424) (0.298) (0.406)SIZE 2.786** 2.761*** 2.883** 2.383* 2.445** 2.494**

(1.183) (0.984) (1.113) (1.253) (1.036) (1.210)LIQ 0.438** 0.456** 0.432** 0.277 0.278 0.270

(0.217) (0.218) (0.214) (0.216) (0.212) (0.211)NLOAN 0.119*** 0.120*** 0.120*** 0.121*** 0.122*** 0.120***

(0.043) (0.044) (0.043) (0.043) (0.045) (0.043)LLR 0.149 0.139 0.156 0.130 0.131 0.137

(0.113) (0.108) (0.105) (0.118) (0.112) (0.110)ROA 0.002 −0.063

(0.234) (0.247)ROE −0.003 −0.004

(0.005) (0.005)CIR −0.002 0.002

(0.015) (0.017)SMD 0.023** 0.024** 0.023***

(0.009) (0.010) (0.009)Firm fixed effects Yes Yes Yes Yes Yes YesAdjusted R2 0.462 0.466 0.472 0.470 0.476 0.479F-statistic 9.152*** 9.288*** 9.553*** 8.777*** 8.963*** 9.149***

Observations 115 115 115 115 115 115

This table presents the results of unbalanced panel OLS regressions of bank capitalization ratio (CTA), measured as the ratio oftotal capital to total assets (in %). TRADE is a measure of a country’s trade openness, measured as the ratio of the sum of totalexports and imports to GDP (in %); FDI denotes a country’s financial openness, computed as the ratio of the value of foreigndirect investment to GDP (in %); SMD refers to stock market development, measured as the ratio of stock market capitalizationto GDP (in %); SIZE is the logarithm of total assets; LIQ is the ratio of cash to total deposits (in %); NLOAN is the net loans ratio,measured as the ratio of net loans to total deposits (in %); LLR is the loan loss reserve ratio, computed as the ratio of loan lossreserve to gross loans (in %); ROA denotes return on assets, measured as the ratio of EBIT to total assets (in %); ROE is returnon equity, measured as the ratio of EBIT to equity (in %); CIR is the cost-to-income ratio, measured as the ratio of operatingexpenses to net revenue (in %). All explanatory variables are one-period lagged. All estimates are based on firm fixed-effectmodel with White’s heteroscedasticity-consistent cross-section standard errors and covariance. Robust standard errors arereported in parentheses.



The estimates in Table 3 show that financial markets development has a positive effect on thecapitalization ratio, as indicated by the positive and significant coefficient on SMD in Models (4)–(6).These results imply that as the depth of stock markets increases, banks become well capitalized. Theimprovement in the capitalization ratio of banks should subsequently enhance the banking systemstability. However, it must be noted that a highly capitalized bank may still be susceptible to a failuredue to, for example, (1) risky loans on its balance sheet and (2) systematic shocks to the banking system.Due to data limitations, I cannot empirically examine the effect of financial markets development onthe quality of bank loans in Thailand. Prior literature shows that bank-lending standards in emergingmarket countries, such as Thailand, are low and might be compromised under some circumstances.14

To test the robustness of the findings in Table 3, I replace the stock market development variablewith the banking sector development (BSD) variable as a measure of financial markets development.Table 4 presents the regression results of Eq. (1) using the BSD variable as a proxy for financial

14 For instance, Charumilind et al. (2006) discuss how firms with connections to banks or politicians receive preferentialtreatments from banks in terms of access to bank loans before the Asian financial crisis.


Table 4Panel OLS regressions of a bank’s capitalization ratio during 1990–2012.

Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Intercept −23.980 −24.262 −24.171 −17.171 −18.188 −17.898(17.328) (16.014) (15.104) (16.905) (15.075) (14.964)

TRADE 0.069** 0.071*** 0.067** 0.070** 0.068*** 0.068**

(0.032) (0.025) (0.031) (0.031) (0.023) (0.030)FIN 0.187 0.133 0.207 0.166 0.168 0.196

(0.360) (0.253) (0.330) (0.380) (0.261) (0.348)SIZE 0.580 0.624 0.623 −0.017 0.083 0.047

(1.527) (1.416) (1.406) (1.540) (1.393) (1.445)LIQ 0.553** 0.567** 0.551** 0.386* 0.385* 0.383*

(0.236) (0.234) (0.232) (0.225) (0.221) (0.220)NLOAN 0.117*** 0.117** 0.119*** 0.118*** 0.120** 0.119***

(0.043) (0.045) (0.044) (0.044) (0.045) (0.043)LLR 0.158 0.140 0.167 0.139 0.131 0.147

(0.112) (0.107) (0.103) (0.118) (0.112) (0.110)ROA 0.052 −0.014

(0.232) (0.244)ROE −0.003 −0.004

(0.005) (0.005)CIR −0.006 −0.002

(0.015) (0.016)SMD 0.025*** 0.026** 0.024***

(0.009) (0.010) (0.009)BSD 0.037** 0.035** 0.038** 0.040*** 0.038*** 0.040***

(0.015) (0.015) (0.015) (0.014) (0.014) (0.013)Firm fixed effects Yes Yes Yes Yes Yes YesAdjusted R2 0.472 0.474 0.482 0.483 0.487 0.493F-statistic 8.846*** 8.916*** 9.240*** 8.603*** 8.736*** 8.972***

Observations 115 115 115 115 115 115

This table presents the results of unbalanced panel OLS regressions of bank capitalization ratio (CTA), measured as the ratio oftotal capital to total assets (in %). TRADE is a measure of a country’s trade openness, measured as the ratio of the sum of totalexports and imports to GDP (in %); FDI denotes a country’s financial openness, computed as the ratio of the value of foreigndirect investment to GDP (in %); SMD refers to stock market development, measured as the ratio of stock market capitalizationto GDP (in %); BSD represents a measure of banking sector development, defined as the ratio of domestic credit provided bybanking sector to GDP (in %); SIZE is the logarithm of total assets; LIQ is the ratio of cash to total deposits (in %); NLOAN is thenet loans ratio, measured as the ratio of net loans to total deposits (in %); LLR is the loan loss reserve ratio, computed as theratio of loan loss reserve to gross loans (in %); ROA denotes return on assets, measured as the ratio of EBIT to total assets (in %);ROE is return on equity, measured as the ratio of EBIT to equity (in %); CIR is the cost-to-income ratio, measured as the ratioof operating expenses to net revenue (in %). All explanatory variables are one-period lagged. All estimates are based on firmfixed-effect model with White’s heteroscedasticity-consistent cross-section standard errors and covariance. Robust standarderrors are reported in parentheses.



markets development. While the coefficient on NLOAN retains its sign, magnitude, and significance,the coefficient on BSD is positive and statistically significant in all models. These results suggest thatthe measure of banking sector development has a positive effect on the banks’ capitalization ratio.

If the depth of stock markets, rather than the depth of the banking system, affects banks’ capital-ization, the coefficient on SMD should remain positive and statistically significant in the presence ofBSD in the regressions. The estimates of Models (4)–(6) of Table 4 show that not only the coefficienton SMD is positive and statistically significant in all models but also is the coefficient on BSD positiveand statistically significant. These findings imply that both the depth of the stock markets and thebanking system together enhance the capitalization ratio of banks. Overall, the results presented inthis section thus far suggest that financial markets development appears to exert no adverse effectson banks in terms of the level of capitalization ratio. As discussed earlier, a banking system in whichbanks are well capitalized is relatively well equipped to withstand a banking crisis.


Table 5Robustness tests: panel OLS regressions of a bank’s capitalization ratio during 1990–2012.

Variables Model 1 Model 2 Model 3 Model 4

Intercept −36.161** −32.332** −21.665 −15.777(14.395) (14.254) (20.774) (18.919)

TRADE 0.061** 0.054* 0.087*** 0.082***

(0.030) (0.029) (0.032) (0.029)FDI −0.174 −0.141 −0.181 −0.147

(0.407) (0.427) (0.387) (0.410)SIZE 1.813 1.479 0.261 −0.290

(1.259) (1.314) (1.870) (1.776)LIQ 0.495** 0.284 0.568** 0.358*

(0.203) (0.197) (0.217) (0.209)NLOAN 0.143*** 0.148*** 0.142*** 0.148***

(0.045) (0.047) (0.046) (0.047)LLR 0.192 0.185 0.187 0.178

(0.124) (0.132) (0.124) (0.132)ROE −0.009** −0.011** −0.008** −0.010**

(0.004) (0.004) (0.004) (0.004)SMD 0.028*** 0.030***

(0.010) (0.010)BSD 0.025 0.029

(0.021) (0.020)Firm fixed effects Yes Yes Yes YesAdjusted R2 0.497 0.513 0.498 0.516F-statistic 9.741*** 9.595*** 9.095*** 9.068***

Observations 107 107 107 107

This table presents the results of unbalanced panel OLS regressions of bank capitalization ratio (CTA), measured as the ratio oftotal capital to total assets (in %). TRADE is a measure of a country’s trade openness, measured as the ratio of the sum of totalexports and imports to GDP (in %); FDI denotes a country’s financial openness, computed as the ratio of the value of foreigndirect investment to GDP (in %); SMD refers to stock market development, measured as the ratio of stock market capitalizationto GDP (in %); BSD represents a measure of banking sector development, defined as the ratio of domestic credit provided bybanking sector to GDP (in %); SIZE is the logarithm of total assets; LIQ is the ratio of cash to total deposits (in %); NLOAN isthe net loans ratio, measured as the ratio of net loans to total deposits (in %); LLR is the loan loss reserve ratio, computed asthe ratio of loan loss reserve to gross loans (in %); and ROE is return on equity, measured as the ratio of EBIT to equity (in %).All explanatory variables are one-period lagged. All firm-year observations with the absolute value of ROE larger than 100%are excluded from the sample. All estimates are based on firm fixed-effect model with White’s heteroscedasticity-consistentcross-section standard errors and covariance. Robust standard errors are reported in parentheses.



To address the outlier issue associated with the extreme values of the ROE variable, I remove firm-year observations with the absolute value of ROE larger than 100%, thereby reducing the numberof observations to 107. Table 5 presents the estimation of Eq. (1) using the restricted sample. I findthat the coefficient on ROE remains negative but becomes statistically significant at the 5% level inall models. More importantly, as can be seen in Table 5, the estimated coefficient on SMD is positiveand statistically significant in all models (i.e., Models 2 and 4). However, the significance of BSD is nolonger evident.

5.2. The effect of financial markets development on diversification of bank revenue

In this section, I test whether financial markets development contributes to a bank’s revenue diver-sification. While a standard portfolio theory suggests that a bank will benefit from having a diversifiedportfolio of revenue streams, several scholars argue that this diversification increases the volatilityof a bank’s returns, but does not significantly increase the average return. Hence, a bank’s revenuediversification entails a greater level of risk. A priori, it is not clear whether a bank would diversify itsportfolio of assets more when financial markets are well developed than when financial markets are


Table 6Panel OLS regressions of a bank’s revenue diversification during 1990–2012.


Intercept −63.244 −64.581 −58.541 −62.456 −108.618 −110.071(48.315) (49.522) (46.259) (48.325) (66.785) (67.270)

TRADE 0.184** 0.187** 0.180** 0.184** 0.092 0.091(0.077) (0.077) (0.075) (0.076) (0.102) (0.103)

FDI −1.444 −1.464* −1.379* −1.453 −1.346 −1.335(0.929) (0.807) (0.826) (0.923) (0.952) (0.939)

SIZE 4.863 4.956 4.612 4.799 9.993 10.132(4.536) (4.630) (4.385) (4.597) (6.719) (6.820)

LIQ −0.108 −0.090 −0.123 −0.140 −0.411 −0.382(0.499) (0.497) (0.497) (0.367) (0.437) (0.356)

NLOAN −0.016 −0.015 −0.011 −0.015 −0.030 −0.031(0.071) (0.070) (0.072) (0.074) (0.071) (0.074)

LLR 0.315* 0.280* 0.344** 0.312* 0.263 0.264(0.177) (0.162) (0.172) (0.175) (0.186) (0.184)

ROA 0.116 0.100 −0.042 −0.0280.326) (0.341) (0.355) (0.366)

ROE 0.004(0.005)

CIR −0.018(0.018)

CTA −0.071 −0.081 −0.063 −0.073 0.030 0.033(0.184) (0.189) (0.184) (0.189) (0.220) (0.227)

SMD 0.004 −0.005(0.038) (0.034)

BSD −0.087 −0.088(0.075) (0.077)

Firm fixed effects Yes Yes Yes Yes Yes YesAdjusted R2 0.639 0.640 0.643 0.635 0.649 0.645F-statistic 16.358*** 16.474*** 16.799*** 15.044*** 15.907*** 14.703***

Observations 114 114 115 114 114 114

This table presents the results of unbalanced panel OLS regressions of a bank’s revenue diversification (SNONIN), measured asthe share of the noninterest income to net revenue (in %). TRADE is a measure of a country’s trade openness, measured as theratio of the sum of total exports and imports to GDP (in %); FDI denotes a country’s financial openness, computed as the ratioof the value of foreign direct investment to GDP (in %); SMD refers to stock market development, measured as the ratio of stockmarket capitalization to GDP (in %); BSD represents a measure of banking sector development, defined as the ratio of domesticcredit provided by banking sector to GDP (in %); SIZE is the logarithm of total assets; LIQ is the ratio of cash to total deposits(in %); NLOAN is the net loans ratio, measured as the ratio of net loans to total deposits (in %); LLR is the loan loss reserve ratio,computed as the ratio of loan loss reserve to gross loans (in %); ROA denotes return on assets, measured as the ratio of EBIT tototal assets (in %); ROE is return on equity, measured as the ratio of EBIT to equity (in %); and CIR is the cost-to-income ratio,measured as the ratio of operating expenses to net revenue (in %). All explanatory variables are one-period lagged. All estimatesare based on firm fixed-effect model with White’s heteroscedasticity-consistent cross-section standard errors and covariance.Robust standard errors are reported in parentheses.



less developed. As in the previous section, I include a number of macro-level and bank-level variablesas control variables.

Table 6 presents the regression estimates of Eq. (2). I find that the estimated coefficient on SMD ispositive but statistically insignificant in all models, thereby suggesting that financial markets devel-opment has no effect on a bank’s revenue diversification. Surprisingly, the results in all models ofTable 6 show that the coefficient on TRADE is positive and statistically significant at the 5% level infour out of six models. These results indicate that the level of trade openness of a country has a positiveand significant impact on the level of a bank’s revenue diversification. That is, a higher level of tradeopenness is associated with a higher share of the noninterest income to operating revenue of a bank.Again, as in the previous section, I now use the banking sector development measure, instead of thestock market development measure, as a proxy for the level of financial markets development. The


Fig. 2. Stock market development and bank market beta during 1990–2012. The left vertical axis of this figure presents stockmarket capitalization (% of GDP) by year; the right vertical axis presents the sampled banks’ mean beta estimated from themarket model, 1 standard deviation (SD) above the mean beta, and 1 standard deviation (SD) below the mean beta.

estimates of Eq. (2) using BSD as a proxy for financial markets development are presented in Models(4)–(6) of Table 6. Because the results are almost identical across three proxies for bank performance(ROA, ROE and CIR), I report only the estimates of regressions using ROA as the regressor for brevity. Theresults show that the banking sector development measure is not associated with a bank’s revenuediversification, as the coefficient on BSD is not statistically significant in all models.

Given that the model’s R2 is high but almost all coefficients in Table 6 are not significant, it is possiblethat the estimates in Table 6 might be driven by extreme values of ROE and/or by multicollinearity (asevidenced by a highly positive and significant correlation coefficient between SIZE and TRADE). First, toaddress the outliers issue, I remove firm-year observations with the absolute value of ROE larger than100%, and re-estimate Eq. (2) using ROE as a measure for performance. I find that the coefficient onTRADE and LLR is positive and statistically significant in all specifications. However, the coefficient onboth SMD and BSD is not statistically significant. Second, to address the multicollinearity issue, I dropSIZE from all specifications, and find that that the positive coefficient on TRADE and LLR is statisticallysignificant. However, the coefficient on SMD and BSD remains statistically insignificant.

In sum, I find no evidence to suggest that both measures for financial markets development areassociated with banks’ revenue diversification. As a result, financial markets development does notappear to explain the revenue diversification of Thai banks during 1990–2012, after controlling forthe trade openness, the financial openness, and the bank-level characteristics.

5.3. The effect of financial markets development on bank market beta

As discussed in Section 3.3, I estimate OLS regressions of Eq. (3) for each bank for each year usingweekly returns data. The estimated beta (ˇM) from Eq. (3) is then used as a market-based measure ofa bank’s risk. Fig. 2 plots the time-series pattern of stock market development and bank market betaduring the 1990–2012 period. The left vertical axis of Fig. 2 presents stock market capitalization (% ofGDP) by year, while the right vertical axis of the figure presents the sampled banks’ mean beta, one


standard deviation (SD) above the mean beta, and one standard deviation below the mean beta, byyear.

Fig. 2 shows that the mean beta increases substantially from 0.97 in 1995 to 1.20, 1.30, and 1.60, in1996, 1997, and 1998, respectively, suggesting that the increase in the bank’s beta corresponds to theAsian financial crisis of 1997–1998. The mean beta falls to about 1.10 during 2003–2005 period15 andincreases substantially in 2006. The hike in the mean beta in 2006 corresponds to the military coup inThailand in 2006. The mean beta varies around 1.20 during the 2007–2012 period, implying either thatthe global financial crisis of 2007–2009 does not adversely affect the perceived risk of the Thai banksor that investors are irrational in the sense that the risk of Thai banks is not priced properly. However,given that this finding is also consistent with Galagedera (2013), who finds that Thailand is one of thetop five equity market performers (out of 40 equity markets) during 2007–2009, the low beta of theThai banks during the global financial crisis is rather more in line with the former explanation thanwith the latter. Overall, the pattern of the banks’ beta shown in Fig. 2 appears to correspond well withmajor events that affect the banking system stability in Thailand.

To test the hypothesis that financial markets development is associated with bank risk, I estimateEq. (1) using the estimated beta (BETA) as the dependent variable. Table 7 reports the unbalanced panelOLS estimations of Eq. (1). As the results in Models (1)–(3) of Table 7 show that all three measures(ROA, ROE and CIR) for performance are not significant, I only report the regression results using ROAas the performance measure in Models (4)–(6) of Table 7 to conserve space.16 Table 7 suggests thatthe coefficient on SMD is negative and statistically significant in all specifications, suggesting that thedepth of stock markets has a negative effect on a bank’s perceived risk, measured by the market beta.The results in Model (5) and (6) of Table 7 show that the coefficient on BSD is positive and statisticallysignificant, implying that banking sector development increases bank risk.

5.4. Additional robustness checks

5.4.1. Subperiod analysis using the Tier 1 capital to total risk-weighted assetsThus far, I proxy for a bank’s capitalization by using the ratio of total capital to total assets (in

%) because the value for the capital adequacy ratio (CART1), which is the ratio of Tier 1 capital tototal risk-weighted assets, is not available for the full sample period (i.e., CART1 is not available priorto 2000). To test the robustness of my results, I estimate Eq. (1) again using CAR1 and CTA as thedependent variable for the 2000–2012 period. Table 8 reports regression estimates. In comparisonwith the results in Table 4, the results in Table 8 show that the coefficient on SMD is not significantin all models. This finding suggests that the depth of stock markets has no effect on the capitalizationratio of the banks in my sample during 2000–2012 (i.e., in periods after the Asian financial crisis).The results also indicate that the coefficient on BSD is negative and statistically insignificant whenusing CTA as the dependent variable, but the coefficient on BSD is negative and statistically significantwhen using CART1 as the dependent variable. This finding suggests that since 2000, the measure ofbanking sector development has a negative effect on the capital adequacy ratio of the banks, therebyimplying that the development of the banking sector in Thailand seems to encourage the risk-takingof the Thai banks via lowering the capital adequacy ratios (i.e., the banks become more leveraged),after controlling for a number of macro-level and bank-level variables.

5.4.2. Inclusion of a lagged dependent variableAs an additional robustness check, I follow the approach of Claessens et al. (2013) by adding a

lagged dependent variable in regressions, so as to allow for natural convergence. Using the restric-tive sample (i.e., excluding firm-year observations with extreme values of ROE as discussed earlier)when estimating all regressions, my main conclusions remain generally unchanged after addingthe lagged dependent variable in the models.17 The untabulated results show that the measure of

15 Serwa (2010) estimates the extent of banking crises and suggests that the banking crisis in Thailand as a result of 1997–1998Asian financial crisis ends in 2002.

16 The pattern of the unreported results is almost identical to those in Table 7.17 To conserve space, I do not tabulate the results, but they are available upon request.


Table 7Panel OLS regressions of a bank’s beta during 1990–2012.


Intercept 0.707 0.721 0.478 0.101 3.584** 2.670*

(1.571) (1.471) (1.407) (1.428) (1.659) (1.586)TRADE −0.004 −0.004 −0.004 −0.004 0.002 0.001

(0.003) (0.003) (0.003) (0.003) (0.004) (0.004)FDI 0.006 0.008 0.004 0.013 0.000 0.007

(0.041) (0.024) (0.033) (0.039) (0.038) (0.037)SIZE 0.109 0.109 0.122 0.159 −0.216 −0.129

(0.144) (0.137) (0.134) (0.133) (0.164) (0.164)LIQ −0.057** −0.058** −0.057** −0.033 −0.038** −0.020

(0.023) (0.025) (0.023) (0.020) (0.017) (0.020)NLOAN −0.005* −0.005* −0.006** −0.006** −0.005 −0.005*

(0.003) (0.003) (0.003) (0.002) (0.003) (0.003)LLR −0.009 −0.007 −0.010 −0.007 −0.005 −0.004

(0.007) (0.005) (0.007) (0.007) (0.006) (0.007)CTA 0.027*** 0.027*** 0.027*** 0.028*** 0.021** 0.022**

(0.008) (0.009) (0.008) (0.008) (0.009) (0.009)ROA −0.006 0.006 0.004 0.013

(0.026) (0.023) (0.024) (0.022)ROE 0.000

(0.000)CIR 0.001

(0.001)SMD −0.003** −0.003**

(0.001) (0.001)BSD 0.006*** 0.005**

(0.002) (0.002)Firm fixed effects Yes Yes Yes Yes Yes YesAdjusted R2 0.190 0.192 0.208 0.259 0.261 0.309F-statistic 3.041*** 3.062*** 3.303*** 3.814** 3.837*** 4.370***

Observations 114 114 114 114 114 114

This table presents the results of unbalanced panel OLS regressions of a bank’s beta, measured as the estimated market betafrom the market model. TRADE is a measure of a country’s trade openness, measured as the ratio of the sum of total exports andimports to GDP (in %); FDI denotes a country’s financial openness, computed as the ratio of the value of foreign direct investmentto GDP (in %); SMD refers to stock market development, measured as the ratio of stock market capitalization to GDP (in %); BSDrepresents a measure of banking sector development, defined as the ratio of domestic credit provided by banking sector to GDP(in %); SIZE is the logarithm of total assets; LIQ is the ratio of cash to total deposits (in %); NLOAN is the net loans ratio, measuredas the ratio of net loans to total deposits (in %); LLR is the loan loss reserve ratio, computed as the ratio of loan loss reserve togross loans (in %); CTA is the capitalization ratio, measured as the ratio of total capital to total assets (in %); ROA denotes returnon assets, measured as the ratio of EBIT to total assets (in %); ROE is return on equity, measured as the ratio of EBIT to equity(in %); and CIR is the cost-to-income ratio, measured as the ratio of operating expenses to net revenue (in %). All explanatoryvariables are one-period lagged. All estimates are based on firm fixed-effect model with White’s heteroscedasticity-consistentcross-section standard errors and covariance. Robust standard errors are reported in parentheses.



stock market development has a positive effect on the capitalization ratio and a negative effect onboth beta and revenue diversification, and that the measure of banking sector development hasno effect on the capitalization ratio, a negative effect on revenue diversification, and a positiveeffect on beta. The coefficient on the lagged depend variables is positive and statistically signif-icant in all models, indicating the persistence of the dependent variables. By including a laggeddependent variable in the regression, the negative effect of financial markets development (i.e.,both stock market development and banking sector development measures) on revenue diversifi-cation becomes evident, implying that after controlling for past revenue diversification, financialmarkets development has the negative effect on the banks’ risk with respect to their revenuediversification.


Table 8Robustness test: panel OLS regressions of a bank’s capitalization ratio during 2000–2012.

Variables CART1 CART1 CART1 CTA CTA CTAModel 1 Model 2 Model 3 Model 4 Model 5 Model 6

Intercept 5.933 12.175 6.740 8.863 −3.018 8.341(36.805) (35.507) (39.144) (30.797) (24.365) (27.839)

TRADE 0.071** 0.076** 0.064* 0.055 0.055* 0.058(0.035) (0.029) (0.032) (0.045) (0.030) (0.044)

FDI −0.778* −0.877* −0.667* −0.960 −1.000* −1.047(0.414) (0.457) (0.388) (0.646) (0.517) (0.728)

SIZE −0.639 −1.201 −0.706 −1.332 −0.442 −1.309(3.121) (2.985) (3.245) (2.554) (2.155) (2.373)

LIQ 0.501 0.411 0.516 1.198** 1.344** 1.204**

(0.528) (0.536) (0.521) (0.564) (0.587) (0.567)NLOAN 0.124* 0.137** 0.123* 0.137* 0.163** 0.134*

(0.071) (0.067) (0.073) (0.077) (0.074) (0.077)LLR 0.209** 0.262** 0.215** 0.128 0.186* 0.117

(0.096) (0.108) (0.097) (0.103) (0.105) (0.105)ROA −0.101 −0.118

(0.225) (0.384)ROE −0.005*** −0.014**

(0.001) (0.005)CIR −0.001 0.012

(0.013) (0.027)SMD 0.022 0.022 0.019 0.038* 0.042* 0.041*

(0.019) (0.018) (0.017) (0.020) (0.021) (0.023)BSD −0.075*** −0.078*** −0.069*** −0.010 −0.040 −0.016

(0.027) (0.029) (0.026) (0.034) (0.032) (0.042)Firm fixed effects Yes Yes Yes Yes Yes YesAdjusted R2 0.329 0.371 0.326 0.457 0.509 0.459F-statistic 3.274*** 3.733*** 3.247*** 5.146*** 6.110*** 5.181***

Observations 66 66 66 70 70 70

This table presents the results of unbalanced panel OLS regressions of a bank’s capitalization ratio during 2000–2012. CART1,measured as the Tier 1 capital to total risk-weighted assets ratio (in %) is the dependent variable in Models (1)–(3), whereas,CTA, measured as the total capital to total assets ratio (in %), is the dependent variable in Models (4)–(6). TRADE is a measure ofa country’s trade openness, measured as the ratio of the sum of total exports and imports to GDP (in %); FDI denotes a country’sfinancial openness, computed as the ratio of the value of foreign direct investment to GDP (in %); SMD refers to stock marketdevelopment, measured as the ratio of stock market capitalization to GDP (in %); BSD represents a measure of banking sectordevelopment, defined as the ratio of domestic credit provided by banking sector to GDP (in %); SIZE is the logarithm of totalassets; LIQ is the ratio of cash to total deposits (in %); NLOAN is the net loans ratio, measured as the ratio of net loans to totaldeposits (in %); LLR is the loan loss reserve ratio, computed as the ratio of loan loss reserve to gross loans (in %); ROA denotesreturn on assets, measured as the ratio of EBIT to total assets (in %); ROE is return on equity, measured as the ratio of EBIT toequity (in %); and CIR is the cost-to-income ratio, measured as the ratio of operating expenses to net revenue (in %). All firm-year observations with the absolute value of ROE larger than 100% are excluded from the sample. All explanatory variables areone-period lagged. All estimates are based on firm fixed-effect model with White’s heteroscedasticity-consistent cross-sectionstandard errors and covariance. Robust standard errors are reported in parentheses.



5.5. Quantile regressions

Thus far, I rely on panel OLS regression to examine the effect of financial markets development onthe bank capitalization ratio, bank revenue diversification, and bank market beta. While I carefullyaddress the potential influence of extreme values on empirical results by removing extreme valuesfrom regressions, the cost of this approach is a smaller sample size.

As another robustness check of my results, I perform a series of panel quantile regressions, because,as first developed by Koenker and Bassett (1978), and as noted by Koenker and Hallock (2001), Sula(2011), and Beine et al. (2010), the quantile regression (QR), which estimates the conditional median, is

C. V

ithessonthi /

J. of

Multi.

Fin. M

anag. 27

(2014) 67–88

85

Table 9Panel quantile regressions of a bank’s capitalization, revenue diversification, and beta during 1990–2012.

Variables CTA CTA CTA SNONIN SNONIN SNONIN BETA BETA BETAModel 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9

Intercept −0.962 12.985 6.390 −17.045 −26.000 −24.957 −0.426 2.026 1.511(12.572) (12.480) (12.066) (23.376) (21.246) (20.845) (1.193) (1.248) (1.166)

TRADE 0.047** 0.070*** 0.060*** 0.170*** 0.160*** 0.163*** −0.002 0.000 0.001(0.022) (0.018) (0.018) (0.047) (0.041) (0.041) (0.002) (0.002) (0.002)

FDI −0.089 −0.189 −0.133 −1.128 −0.475 −0.532 0.041 −0.002 0.003(0.337) (0.216) (0.242) (0.686) (0.585) (0.596) (0.027) (0.029) (0.029)

SIZE −0.350 −1.034 −0.838 0.494 1.219 1.176 0.104 −0.073 −0.043(0.740) (0.722) (0.684) (1.308) (1.253) (1.226) (0.065) (0.077) (0.070)

NTLOAN 0.123*** 0.066* 0.099** 0.029 0.029 0.030 −0.003 −0.002 −0.001(0.046) (0.039) (0.040) (0.108) (0.091) (0.089) (0.004) (0.005) (0.004)

LLR 0.204 0.124 0.175 0.763** 0.742** 0.765** −0.003 0.001 −0.001(0.129) (0.096) (0.109) (0.333) (0.307) (0.301) (0.007) (0.008) (0.007)

ROE −0.005 −0.002 −0.004 0.002 0.001 0.000 0.000 0.000 0.000(0.004) (0.003) (0.003) (0.005) (0.005) (0.005) (0.000) (0.000) (0.000)

CTA 0.008 0.329 0.328 0.019* 0.002 0.003(0.294) (0.424) (0.455) (0.011) (0.012) (0.011)

SMD 0.017 0.019 0.016 −0.008 −0.005*** −0.003**

(0.014) (0.012) (0.026) (0.027) (0.001) (0.001)BSD 0.035** 0.027** −0.065** −0.065** 0.006*** 0.005***

(0.015) (0.013) (0.027) (0.030) (0.002) (0.002)Adjusted pseudo R2 0.166 0.181 0.189 0.333 0.350 0.344 0.074 0.089 0.139Quasi-LR statistic 6.254*** 49.895*** 5.275*** 96.604*** 111.572*** 112.378*** 21.003*** 23.804*** 34.191***

Observations 115 115 115 114 114 114 114 114 114

This table presents the results of unbalanced panel quantile regressions of a bank’s capitalization, revenue diversification, and beta during 1990–2012. CTA, measured as the total capitalto total assets ratio (in %), is the dependent variable in Models (1)–(3); SNONIN, revenue diversification, measured as the share of the noninterest income to net revenue (in %), is thedependent variable in Models (4)–(6); BETA, measured as the estimated market beta from the market model, is the dependent variable in Models (7)–(9). TRADE is a measure of a country’strade openness, measured as the ratio of the sum of total exports and imports to GDP (in %); FDI denotes a country’s financial openness, computed as the ratio of the value of foreigndirect investment to GDP (in %); SMD refers to stock market development, measured as the ratio of stock market capitalization to GDP (in %); BSD represents a measure of banking sectordevelopment, defined as the ratio of domestic credit provided by banking sector to GDP (in %); SIZE is the logarithm of total assets; LIQ is the ratio of cash to total deposits (in %); NLOANis the net loans ratio, measured as the ratio of net loans to total deposits (in %); LLR is the loan loss reserve ratio, computed as the ratio of loan loss reserve to gross loans (in %); and ROEis return on equity, measured as the ratio of EBIT to equity (in %). All explanatory variables are one-period lagged. All estimates are at the 50th (median) quantile using Huber Sandwichstandard errors and covariance. Robust standard errors are reported in parentheses.




not as sensitive to outliers as the OLS regression, which estimates the conditional mean.18 As suggestedby Sula (2011) that the use of a fixed effects model is not appropriate for the QR method due to theissues of implementation and interpretation, the QR estimation is therefore not based on a fixed firmseffect model. By employing the QR framework, I do not have to remove extreme values from thesample, and thus can use the full bank-year observations data set, which improves the statisticalpower of the estimation approach. Another advantage of using the QR approach in this study is thatit provides information about a potentially asymmetric impact of financial markets development ongood (upper quantiles) or bad (lower quantiles) capitalization ratios of the banks under study.

I estimate unbalanced panel QR of Eqs. (1) and (2), using CTA, SNONIN, and BETA as the dependentvariables in Models (1)–(3), Models (4)–(6), and Models (7)–(9), respectively; all QR estimates are atthe 50th (median) quantile using Huber Sandwich standard errors and covariance. Because the QRresults are very similar across models using different measures for bank performance (ROA, ROE andCIR), I only report the QR results using ROE as a proxy for performance in Table 9 for brevity.

Results in Models (1)–(3) of Table 9 indicate that the effect of stock market development on banks’capitalization ratio is positive but statistically insignificant, whereas the effect of banking sector devel-opment is positive. Consistent with the OLS regression results, estimates of QR in Models (4)–(6) ofTable 9 indicate that stock market development has no effect on a bank’s revenue diversification mea-sure. The findings also show that banking sector development has a negative effect on a bank’s revenuediversification measure. Similar to the OLS results shown in Table 7, the QR results in Models (7)–(9)of Table 9 show that stock market development has a negative effect on a bank’s beta, while bankingsector development has a positive effect on a bank’s beta.

The QR results suggest that the positive linkage between stock market development and a bank’scapitalization ratios obtained in the OLS estimations is potentially driven by observations in periodsprior to the Asian financial crisis. Recall that the results in Table 8 show that stock market developmenthas no effect on the capitalization ratio during the 2000–2012 period. To check for this explanation, Iestimate OLS regressions of Eq. (1) for the 1990–1999 period. The results show that the coefficient onSMD is positive and statistically significant. Taken together, empirical evidence indicates that the depthof stock markets is associated with higher total capital to total asset ratios during 1990–1999, and thislinkage is no longer evident during 2000–2012. To summarize, the basic conclusions concerning withthe relation between financial markets development and a bank’s risk are robust to correcting forextreme values.

6. Conclusions

I examine a bank’s risk (the capitalization ratio, the revenue diversification, and the market beta)relating to financial markets development in Thailand during 1990–2012. I find that the level of stockmarket development (measured by the ratio of stock market capitalization to GDP) is associated withhigher bank capitalization ratios as measured by the ratio of total capital to assets, after controlling fortrade openness, financial openness, firm size, liquidity ratio, loan rates, loan loss rates, and performance(measured by return on assets, return on equity, or cost-to-income ratio). But this positive relation onlyholds during 1990–1999 and is no longer evident during 2000–2012. The finding indicating that stockmarket development has no adverse effect on banks’ capitalization ratios contradicts related resultsin prior literature on the positive relation between financial markets development/liberalization andbanking sector risk/instability (see e.g., Carbó-Valverde et al., 2012; Ranciere et al., 2006). I also findthat the depth of stock markets has a negative effect on a bank’s perceived risk, measured by themarket beta, while banking sector development tends to increase the bank’s market beta. During2000–2012, the measure of banking sector development is not associated with a bank’s capitalizationwhen measured as the total capital to assets ratio, but it is negatively related to a bank’s capitalizationwhen measured as the Tier 1 capital to total risk-weighted assets. With respect to the influence offinancial markets development on a bank’s revenue diversification (measured as the share of the

18 From a methodological point of view, I refer interested readers to prior literature using the quantile regression approach(e.g., Baur et al., 2012; Baur and Schulze, 2009; Edgerton, 2012; Koenker and Bassett, 1978; Koenker and Hallock, 2001).


noninterest revenue to net revenue), neither the measure of banking sector development nor themeasure of stock market development has any effect on the bank’s revenue diversification.

Overall, my findings suggest that the depth of stock markets enhances, rather than deteriorates, thebanking system stability by lowering the bank risk implied by the market beta. However, the devel-opment of the banking sector appears to induce the instability of the banking system by lowering theTier 1 capital adequacy ratio and by increasing the level of the market beta of a bank. There exists alarge body of literature on the relation between financial markets development and economic devel-opment. In particular, prior literature offers relatively mixed results with respect to the question ofwhether financial markets development/liberalization enhances or harms the stability of the financialsystem. In this paper, I compliment the existing literature by providing new insights into the impactof financial markets development on the degree of risk-taking of the six Thai banks that have beenthrough two financial crises over the past two decides.

References

Angkinand, A., Wihlborg, C., 2010. Deposit insurance coverage, ownership, and banks’ risk-taking in emerging markets. J. Int.Money Financ. 29, 252–274.

Apergis, N., 2014. The long-term role of non-traditional banking in profitability and risk profiles: evidence from a panel of U.S.banking institutions. J. Int. Money Financ. 45, 61–73.

Baur, D.G., Dimpfl, T., Jung, R.C., 2012. Stock return autocorrelations revisited: a quantile regression approach. J. Empir. Financ.19, 254–265.

Baur, D.G., Schulze, N., 2009. Financial market stability – a test. J. Int. Financ. Market. Inst. Money 19, 506–519.Becerra, O., Cavallo, E., Scartascini, C., 2012. The politics of financial development: the role of interest groups and government

capabilities. J. Bank. Financ. 36, 626–643.Beck, T., De Jonghe, O., Schepens, G., 2013. Bank competition and stability: cross-country heterogeneity. J. Financ. Intermed. 22,

218–244.Beine, M., Cosma, A., Vermeulen, R., 2010. The dark side of global integration: increasing tail dependence. J. Bank. Financ. 34,

184–192.Bena, J., Ondko, P., 2012. Financial development and the allocation of external finance. J. Empir. Financ. 19, 1–25.Bikker, J.A., Metzemakers, P.A.J., 2005. Bank provisioning behaviour and procyclicality. J. Int. Financ. Market. Inst. Money 15,

141–157.Bourgain, A., Pieretti, P., Zanaj, S., 2012. Financial openness, disclosure and bank risk-taking in MENA countries. Emerg. Market.

Rev. 13, 283–300.Calderon, C., Liu, L., 2003. The direction of causality between financial development and economic growth. J. Dev. Econ. 72,

321–334.Calmès, C., Théoret, R., 2013. Market-oriented banking, financial stability and macro-prudential indicators of leverage. J. Int.

Financ. Market. Inst. Money 27, 13–34.Campos, N.F., Karanasos, M.G., Tan, B., 2012. Two to tangle: financial development, political instability and economic growth in

Argentina. J. Bank. Financ. 36, 290–304.Carbó-Valverde, S., Marques-Ibanez, D., Rodríguez-Fernández, F., 2012. Securitization, risk-transferring and financial instability:

the case of Spain. J. Int. Money Financ. 31, 80–101.Cardone-Riportella, C., Samaniego-Medina, R., Trujillo-Ponce, A., 2010. What drives bank securitisation? The Spanish experience.

J. Bank. Financ. 34, 2639–2651.Charumilind, C., Kali, R., Wiwattanakantang, Y., 2006. Connected lending: Thailand before the financial crisis. J. Bus. 79, 181–217.Chinn, M.D., Ito, H., 2006. What matters for financial development? Capital controls, institutions, and interactions. J. Dev. Econ.

81, 163–192.Claessens, S., Ghosh, S.R., Mihet, R., 2013. Macro-prudential policies to mitigate financial system vulnerabilities. J. Int. Money

Financ. 39, 153–185.De Jonghe, O., 2010. Back to the basics in banking? A micro-analysis of banking system stability. J. Financ. Intermed. 19, 387–417.Delis, M.D., Kouretas, G.P., 2011. Interest rates and bank risk-taking. J. Bank. Financ. 35, 840–855.DeLong, G.L., 2001. Stockholder gains from focusing versus diversifying bank mergers. J. Financ. Econ. 59, 221–252.Edgerton, J., 2012. Agency problems in public firms: evidence from corporate jets in leveraged buyouts. J. Financ. 67, 2187–2213.Festic, M., Kavkler, A., Repina, S., 2011. The macroeconomic sources of systemic risk in the banking sectors of five new EU

member states. J. Bank. Financ. 35, 310–322.Fiordelisi, F., Marques-Ibanez, D., Molyneux, P., 2011. Efficiency and risk in European banking. J. Bank. Financ. 35, 1315–1326.Galagedera, D.U.A., 2013. A new perspective of equity market performance. J. Int. Financ. Market. Inst. Money 26, 333–357.Gauthier, C., Lehar, A., Souissi, M., 2012. Macroprudential capital requirements and systemic risk. J. Financ. Intermed. 21,

594–618.Gimet, C., Lagoarde-Segot, T., 2011. A closer look at financial development and income distribution. J. Bank. Financ. 35,

1698–1713.Haq, M., Heaney, R., 2012. Factors determining European bank risk. J. Int. Financ. Market. Inst. Money 22, 696–718.Huang, X., Zhou, H., Zhu, H., 2012. Assessing the systemic risk of a heterogeneous portfolio of banks during the recent financial

crisis. Journal of Financial Stability 8, 193–205.Houston, J.F., Lin, C., Lin, P., Ma, Y., 2010. Creditor rights, information sharing, and bank risk taking. J. Financ. Econ. 96, 485–512.Iannotta, G., Nocera, G., Sironi, A., 2013. The impact of government ownership on bank risk. J. Financ. Intermed. 22, 152–176.Imbierowicz, B., Rauch, C., 2014. The relationship between liquidity risk and credit risk in banks. J. Bank. Financ. 40, 242–256.

http://refhub.elsevier.com/S1042-444X(14)00014-0/sbref0005






















































































































































































































































































































































































































































































































































Islam, S.S., Mozumdar, A., 2007. Financial market development and the importance of internal cash: evidence from internationaldata. J. Bank. Financ. 31, 641–658.

Koenker, R., Bassett Jr., G., 1978. Regression quantiles. Econometrica 46, 33–50.Koenker, R., Hallock, K.F., 2001. Quantile regression. J. Econ. Perspect. 15, 143–156.Laeven, L., Levine, R., 2007. Is there a diversification discount in financial conglomerates? J. Financ. Econ. 85, 331–367.Meslier, C., Tacneng, R., Tarazi, A., 2014. Is bank income diversification beneficial? Evidence from an emerging economy. J. Int.

Financ. Market. Inst. Money 31, 97–126.Nijskens, R., Wagner, W., 2011. Credit risk transfer activities and systemic risk: how banks became less risky individually but

posed greater risks to the financial system at the same time. J. Bank. Financ. 35, 1391–1398.Noy, I., 2004. Financial liberalization, prudential supervision, and the onset of banking crises. Emerg. Market. Rev. 5, 341–359.Ranciere, R., Tornell, A., Westermann, F., 2006. Decomposing the effects of financial liberalization: crises vs. growth. J. Bank.

Financ. 30, 3331–3348.Serwa, D., 2010. Larger crises cost more: impact of banking sector instability on output growth. J. Int. Money Financ. 29,

1463–1481.Soedarmono, W., Machrouh, F., Tarazi, A., 2013. Bank competition, crisis and risk taking: evidence from emerging markets in

Asia. J. Int. Financ. Market. Inst. Money 23, 196–221.Stiglitz, J.E., 2000. Capital market liberalization, economic growth, and instability. World Dev. 28, 1075–1086.Stiroh, K.J., 2006. A portfolio view of banking with interest and noninterest activities. J. Money Credit Bank. 38, 1351–1361.Sula, O., 2011. Demand for international reserves in developing nations: a quantile regression approach. J. Int. Money Financ.

30, 764–777.Williams, J., Nguyen, N., 2005. Financial liberalisation, crisis, and restructuring: a comparative study of bank performance and

bank governance in South East Asia. J. Bank. Financ. 29, 2119–2154.Xiao, S., Zhao, S., 2012. Financial development, government ownership of banks and firm innovation. J. Int. Money Financ. 31,

880–906.Zagorchev, A., Vasconcellos, G., Bae, Y., 2011. Financial development, technology, growth and performance: evidence from the

accession to the EU. J. Int. Financ. Market. Inst. Money 21, 743–759.






































































































































































































































































































Contents Journal of Multinational Financial ManagementC. Vithessonthi / J. of Multi. Fin. Manag. 27...

Documents

Transcript of Contents Journal of Multinational Financial ManagementC. Vithessonthi / J. of Multi. Fin. Manag. 27...