Diversification in Korean Banking Business: Is Non-interest Income a Financial Saviour?

Abstract

This article examines whether the diversification of operating income in Korean banks has persistently enhanced the performance of Korean banks. The results show that, despite Korean banks’ efforts to diversify their operating income, these banks do not gain any benefit from the diversification. Thus, bank managers in Korea focus on interest income revenue. The results also show that the increase in non-interest income revenue keeps pace with the growth in expenses, which offsets the diversification effect on the performance of Korean banks. As a result, Korean banks discourage banking diversification and focus on non-interest income revenues.

JEL Classification: G21

Keywords

Banking non-interest income diversification risk

1. Introduction

The 1997 Asian financial crisis and the era of the International Monetary Fund (IMF) made Korean Banks embrace operating income revenue diversification strategies. In addition, bank regulators enacted laws that restructured and expanded the operating income revenue to increase its diversification from conventional banking activities (i.e., interest income-earning activities) to non-bank activities (i.e., non-interest income-earning activities).^1,2 As a result, the ratio of non-interest income to total interest income increased from 34.92 per cent in 1998 to 58.29 per cent in 2004.³ However, after 2004 the ratio of non-interest income to interest income steadily decreased and settled at 4.5 per cent in 2011. In other words, it is apparent that Korean banks stopped their diversification through increasing non-bank activities. Rather, they have been returning to the so-called ‘core’ banking business, which focuses on traditional interest income-earning activities.

Previous studies (Berger, Hassan, & Zhou, 2010; Natalia & Firsty, 2016; Park & Weber, 2006; Sufian, 2011) provide evidence showing that diversification of Korea bank industry positively affect its revue. However, these studies could lead to a misleading conclusion since they do not employ the time-series variation observed after 2007. Our study pays attention to this observation of the Korean banking industry and investigates this unique feature, focusing on two issues. First, we investigate the root cause of Korean banks’ return to the traditional interest income-earning activities. These returns could be regarded as a failure of the bank diversification strategy since the revamped banking structure in Korea is favourable to the diversification of operating income. In this regard, we identify factors that influenced Korean banks to terminate their non-banking activities. Second, we examine if the interest income earning after 2008 positively or negatively affected the bank performance. Since there is a significant departure from non-interest income earning to interest income earning, we examine how this transition improves or harms the performance of Korean banks. Therefore, our research makes contribution to the literatures on the effect of banking diversification in Korea.

Using commercial and special-purpose banks in Korea from 2001 to 2014, we conduct both cross-sectional and panel analysis. Our empirical analyses reveal that, unlike previous literature that studies the link between diversification and revenue in Korea banking industry, non-interest income-earning activities propel an increase in both a bank’s revenue and its risk. In other words, we find evidence supporting the trade-off between risk and return in banking diversification. This is in line with other cross-country evidences. Baele, De Jonghe, and Vander Vennet (2007) show that European banks with a higher level of non-interest income as a percentage of total operating income are exposed to greater systematic risk. Busch and Kick (2009) find that non-interest income-earning activities positively affect the risk-adjusted performance of German banks but that they incur higher risk in terms of their operating income. Li and Zhang (2013) show that for the Chinese banking industry there are diversification benefits in terms of an increase in non-interest income. Nevertheless, non-interest income-earning activities tend to increase volatility in banks’ operating income, which implies that increased reliance on non-interest income may worsen the risk/return trade-off for a given bank. As Saunders and Walter (1994), Kwast (1989) and Stiroh (2006) suggest, using the minimisation of expected risks (MER) method, Baek, Lee, and Chan (2015) also find that banking diversification in non-bank activities also could improve the performance of individual banks, but they document that there is a positive relation between non-interest income and volatility of total operating income.

In addition to this, we further find that the diversification by Koran banks into non-interest income activities failed due to rising costs overweighed revenues from the diversification, and hence they have intentionally retreated to traditional banking activity.

The article proceeds as follows. Section 2 provides the data sources and proxies for the variables used in the study. Section 3 presents the empirical framework. Section 4 includes the empirical results, and Section 5 concludes the article.

2. Data Sources and Variables

2.1. Variable Construction

Three sets of measures are described in this section. The first set comprises measures of non-interest income and diversification. The second set includes several measures of analyst forecast variables. The last comprises a set of control variables.

2.1.1. Diversification Measures

In this study, we employ adjusted diversification measures to identify the effect of non-interest earning activities on the Korean banking industry. As in Stiroh and Rumble (2006), Chiorazzo, Milani, and Salvini (2008), Sanya and Wolfe (2011), Lee and Hsieh (2013), and Meslier, Tacneng, and Tarazi (2014), we use an adjusted Hirschman-Herfindahl index (HHI) based on interest income and non-interest income components. To provide the details of non-interest income activity, we decompose non-interest income into five components, which include (a) fee commission (FEE), (b) securities sales (SEC), (c) foreign exchange trading (FX), (d) trust management (TRUST) and (e) other sources of revenue (OTHERS), consistent with the definition of the Financial Supervisory Services Authority in Korea. More specifically, the share of aggregate non- interest income revenue (SH_NON)⁴ is given by Equation (1):

S H_{N O N} = \frac{N N I}{N N I + N I I} ​

(1)

where NNI is net non-interest income revenue, and NII represents net interest income. Then, SH_NON can be specified in terms of decomposed share of non-interest income revenue as follows:

S H_{N O N} = S H_{F E E} + S H_{S E C} + S H_{F X} + S H_{T R U S T} + S H_{O T H E R S} ​

(2)

where SH_FEE is the share of net fee commission revenue as a percentage of total non-interest income, SH_SEC is the share of net gains (loss) from securities sales as a percentage of total non-interest income, SH_FX is the share of net foreign exchange-trading revenue as a percentage of total non-interest income, SH_TRUST is the share of net trust revenue as a percentage of total non-interest income and SH_OTHERS is the share of net other revenue as percentage of total non-interest income. More precisely, they are defined as follows:

S H_{F E E} = \frac{F E E}{N N I + N I I} ​

S H_{S E C} = \frac{S E C}{N N I + N I I} ​

S H_{F X} = \frac{F X}{N N I + N I I} ​

S H_{T R U S T} = \frac{T R U S T}{N N I + N I I} ​

S H_{O T H E R S} = \frac{O T H E R S}{N N I + N I I} ​

(3)

Morgan and Samolyk (2003), Thomas (2002) and Stiroh and Rumble (2006) measure the revenue diversification (DIV) as seen in Equation (4):

D I V = 1 - (S H_{N E T}^{2} + S H_{N O N}^{2}) ​

(4)

In this measurement, the highest value for DIV is equal to 0.5, indicating an even split between non-interest income and interest income while the lowest value is zero, indicating that a bank is fully concentrated with a single source of revenue.

This revenue diversification measurement considers total non-interest income revenue but does not consider all idiosyncratic diversification activities because it is constructed to monitor overall non-interest earning activities. To determine how major components of non-interest income affect banks’ business, our study employs an adjusted DIV measure, following Engle, Moshirian, Sahgal, and Zhang (2014), Meslier et al. (2014), and Lee and Hsieh (2013). Specifically, we use decomposed variables in Equation (2) and measure the degree of the revenue diversification in Korean banks, as follows:

\begin{matrix} D I V_{A D J} = 1 - (S H_{N E T}^{2} + S H_{F E E}^{2} + S H_{S E C}^{2} + S H_{F X}^{2} \\ + S H_{T R U S T}^{2} + S H_{O T H E R S}^{2}) \end{matrix}

(5)

The adjusted diversification index takes values between zero and 0.7. The value of zero denotes full concentration, indicating that a bank’s revenue relies fully on either interest income revenue or non-interest income revenue, while the value of 0.7 represents complete diversification or an equal balance among five non-interest income activities.

2.1.2. Bank Performance Measure

To evaluate the performance of a bank, we use return on equity (ROE), defined as the ratio of annualised net income to total shareholders’ equity.⁵ In addition to ROE, following Stiroh and Rumble (2006) and Chiorazzo et al. (2008), we compute the risk-adjusted ROE, RAROE_it, for a bank i in a given year t and employ it as the other bank performance measure:

R A R O E_{i t} = \frac{{\overset{─}{R O E}}_{i t}}{S D R O E_{i t}} ​

(6)

where ${\overset{─}{R O E}}_{i t} $ represents the average ROE of a bank i over the four quarters in a given year t, and SDROE_it is the standard deviation of a bank i over the quarters in the same year t.⁶

2.1.3. Risk and Insolvency Measures

We calculate risk-adjusted capital asset ratio (RA_CAR) and Z-scores to evaluate the risk exposure of banks (Böninghausen & Köhler, 2015; Stiroh & Rumble, 2006). RA_CAR, as suggested by Lepetit, Nys, Rous, and Tarazi (2008), is defined as the average of capital asset ratio, ${\overset{─}{C A R}}_{i t} $ , divided by the standard deviation of quarterly CAR over the four quarters of a given year. More specifically, RA_CAR for a bank i in a year t is as follows:

R A C A R_{i t} = \frac{{\overset{─}{C A R}}_{i t}}{S D C A R_{i t}} ​

(7)

where ${\overset{─}{C A R}}_{i t} $ is the average equity-to-asset ratio for a bank i over the four quarters in a given year t, and SDCAR_it is the standard deviation of the level of capitalisation for a bank i over the quarters in the same period. In general, RA_CAR measures a bank’s leverage risk. Stated another way, RA_CAR is used to assess whether banks maintain sufficient capital relative to the risk of their capital structure. In Equation (7), CAR is the equity-to-asset ratio that explains banks’ debt exposure. Because the equity-to-debt ratio is equal to 1 – debt-to-asset ratio,⁷ the smaller negative CAR indicates the higher leverage on debt and the higher probability of the insolvency of a bank.

The other indicator of bank risk used in this study is the Z-score, which measures, essentially, the distance to default for a given bank. The Z-score, a measure of probability of insolvency, is widely used in the literature (Berger et al., 2010; Chiorazzo et al., 2008; Demirgüç-Kunt & Huizinga, 2010; Köhler, 2014; Laeven & Levine, 2009; Meslier et al., 2014; Stiroh & Rumble, 2006). The Z-score for a bank i in a year t is usually expressed as:

Z - S c o r e = \frac{{\overset{─}{R O A}}_{i t} + {\overset{─}{E / A}}_{i t}}{σ_{R O A}} ​

(8)

where ${\overset{─}{R O A}}_{i t} $ represents the average ROE of a bank i over the four quarters of a given year t. Specifically, a bank with a higher Z-score is the one with a lower risk of insolvency, whereas a bank with a lower Z-score has the higher risk of insolvency.

2.1.4. Control Variables

Following Berger et al. (2010), Čihák and Hesse (2010), Hsieh, Chen, Lee, and Yang (2013), and Engle et al. (2014), we employ related endogenous variables, such as logarithm of total bank real assets, growth rate of total real assets, and CAR. We use the logarithm of a bank’s asset, following Stiroh (2004a, 2004b), Stiroh and Rumble (2006), Behr, Kamp, Memmel, and Pfingsten (2007), and Chiorazzo et al. (2008), to address the size effect of a bank. For example, a large bank tends to have a greater ability to allocate funds to invest in human or technology resources to add new business lines and to develop sophisticated measures of risk management.

Stiroh (2004b), Chiorazzo et al. (2008), and Meslier et al. (2014) argue that the growth rate of total assets is a proxy for a bank’s risk preference. More specifically, a bank with a high level of risk tolerance shows a steady- or slow-growth figure that is a bank that takes high risk can grow faster than can a bank with a high level of risk aversion. In our study, the level of a bank asset and the bank growth rate are based on real assets to address the disparity of time value in banks’ assets.

In addition to the three conventional endogenous variables, we consider other exogenous variables that might affect profitability and the risk of banks: real GDP per capita, term spread and credit spread. We include real GDP per capita, as a higher GDP per capita is likely to lead to an increase in total assets in the banking systems. Further, a higher GDP growth is likely to have a significant positive impact on a bank’s total operating revenue, as suggested by Demirgüç-Kunt and Huizinga (1999) and Hsieh et al. (2013).

The term spread refers to the difference between a long-term and a short-term yield, which is regarded as a future economic indicator. Generally, when the term structure shows a wider spread, market participants expect bull markets. Conversely, when the spread narrows down, participants anticipate bear markets. A short-term interest rate is strongly associated with monetary policy, while a long-term yield is sensitively related to the expectation of short interest changes, liquidity premium and bond duration. In this respect, the term spread is a mixture of various economic information, such as monetary policy, anticipation of future monetary policy change and future economic conditions, and hence we incorporate term spread as one of the control variables.

Similarly, we use the credit spread, defined as the difference in yield between a corporate bond and a government bond, as credit spread incorporates liquidity and default risk, as documented by Elton, Gruber, Agrawal, and Mann (2001). When markets have an optimistic forecast for economic growth, the risk premium tends to fall, which in turn results in a decrease in credit spread. In contrast, when there is a pessimistic view of the market, the risk premium rises, and thus the credit spread becomes extended.⁸

2.2. Data Sources

In this study, financial statement data are derived from the Financial Statistics Information System (FSIS),⁹ and the sample period is from 2001 to 2014. When constructing the research dataset, we consider backfill bias and survivorship bias because it is likely that those biases lead to spurious results, as Fama and French (1992, 1993) note. Accordingly, to handle survivorship bias, our research data include all banks, comprising both dead and live banks during the sample period: 22 commercial banks and 6 special-purpose banks. To address backfill bias, researchers conventionally screen out companies that have been in business for less than 24 months. We do not need to follow this screening method, however, to handle backfill bias, as by nature none of the banks in Korea allowed us to delay filing their fiscal information. This means that all of the banks must report financial statements to the public; this is not the case for non-bank financial statements, which could be held back until the company demonstrates good performance. This implies that our estimates are free of backfill bias.

3. Empirical Framework

In this article, we identify the relationship between the diversification of a Korean bank’s operating income and performance. To investigate this relationship thoroughly, we consider two empirical approaches. First, we run regressions in which all dependent and independent variables are generated over the respective bank’s lifetime. In other words, we exploit the cross-sectional variation by taking the average of variables over time and then examining the between effects. This process quantifies risk exposure by the share of non-interest income and the volatility of profit for a Korean bank over its entire history. Second, we conduct a panel regression analysis, whereby all of the explanatory variables are computed for each bank in a given year. Although the first approach captures the cross-sectional relationship between banking diversification and risk-adjusted performance, it ignores how the two vary over time. To address this, we do a panel analysis, using bank fixed effects, and exploit both cross-sectional and time-series variations.

3.1. Cross-sectional Analysis

In this research, the empirical regression model is written as follows:

Y_{i} = α_{0} + α_{1} {\overset{─}{D I V}}_{i} + α_{2} {\overset{─}{S H}}_{N O N_{i}} + α \cdot {\overset{─}{Z}}_{i} + ε_{i t}

(9)

where Y_i is a measure of performance or risk of a bank i, ${\overset{─}{D I V}}_{i} $ represents average net revenue diversification, ${\overset{─}{S H}}_{N O N_{i}} $ represents the average of the share of non-interest income, Z_i represents a vector of the control variables for a bank i as described earlier and ε_it is an error term. All of the variables in this model are calculated over a bank’s lifetime to examine the cross-sectional relationship between a bank’s performance or risk and its diversification effort (e.g., Stiroh & Rumble, 2006). To examine how a change of non-interest income affects the change in a bank’s performance or risk, Stiroh and Rumble (2006) employ the first derivative of Equation (9):

\frac{\partial Y_{i}}{\partial {\overset{─}{S H}}_{N O N_{i}}} = {α^{'}}_{1} \frac{\partial {\overset{─}{D I V}}_{i}}{\partial {\overset{─}{S H}}_{N O N_{i}}} + {α^{'}}_{2} \frac{\partial {\overset{─}{S H}}_{N O N_{i}}}{\partial {\overset{─}{S H}}_{N O N_{i}}} = {α^{'}}_{1} \frac{\partial {\overset{─}{D I V}}_{i}}{\partial {\overset{─}{S H}}_{N O N_{i}}} + {α^{'}}_{2} ​

(10)

The first term on the right-hand side, ${α^{'}}_{1} $ , provides the magnitude of the indirect impact of the share of non-interest income change on the bank’s performance, which works through the bank’s diversification change.¹⁰ The second term on the right-hand side, ${α^{'}}_{2} $ , gives the magnitude of the direct effect of a change in the non-interest income share. Because the net effect is the sum of the indirect and direct effects, this equation is useful for determining the relationship between a change in a bank’s performance and a change in non-interest income. Note that this model describes only how a bank’s performance is affected by a bank’s aggregate non-interest income-share change.

To examine an individual effect of a change in each of the aforementioned five non-interest income activities on the bank’s performance through diversification efforts, we modify Equation (9), using Equations (2), (3) and (5) and suggest the modified equation as follows:

Y_{i} = β_{0} + β_{1} {\overset{─}{D I V}}_{A D J_{i}} + γ \cdot {\overset{─}{S}}_{i} + β \cdot {\overset{─}{Z}}_{i} + ε_{i t}

(11)

where ${\overset{─}{D I V}}_{A D J_{i}} $ represents the average of adjusted diversification for bank i and ${\overset{─}{S}}_{i} $ is a vector of shares of the non-interest income components: $S H_{F E E_{i}}, S H_{S E C_{i}}, S H_{F X_{i}}, S H_{T R U S T_{i}} $ and $S H_{O T H E R S_{i}} $ . Based on this basic model, we investigate how each of the five non-interest income components indirectly and directly changes with a bank’s risk-adjusted performance.¹¹

3.2. Panel Analysis

To conduct the panel analysis, we use the following fixed-effect regression, exploiting panel variation across banks and time:

Y_{i t} = δ_{0} + δ_{1} {\overset{─}{D I V}}_{A D J_{i t}} + θ \cdot {\overset{─}{S}}_{i t} + δ \cdot {\overset{─}{Z}}_{i t} + λ_{i} + u_{i t}

(12)

where ${\overset{─}{Z}}_{i t} $ is a vector of control variables, including both bank-level and macro-variables, λ_i is the bank fixed effect, u_it is an error term and other variables are defined as before, except that they vary over time. The bank’s fixed effect mitigates a potential unobservable omitted variable bias, as some time-invariant bank-specific factors, such as bank type, may affect performance and risk exposure of banks.

4. Empirical Results

4.1. Do Korean Banks Hesitate to Diversify Operating Income by Expanding into Non-banking Activities?

The Korean banking industry has reformed its banking business lines and legal system to enhance the diversification of the banking business through non-banking activities. Despite this strategic support, Korean banks hesitate to shift their primary source of revenue from interest income to non-interest income, as shown in Table 1. The table presents the mean of net interest income share and that of non-interest income share as a percentage of total operating income from 1990 to 2014 for commercial banks and special-purpose banks in Korea. For the commercial banks, the share of net interest income (80.2%) is about four times greater than that of net non-interest income (19.8%). Similarly, for the special banks, the share of interest income is approximately 11.5 times greater than that of non-interest income. These indicate that Korea banks concentrate on interest income activities, on average, regardless of the type of bank.

Table 1.
Percentage of the Shares of Net Interest and Non-interest Income

Banks Number of Banks SH_NET SH_NON

Commercial 22 80.2 19.8

Special 6 92.4 7.6

Total Banks 28 82.9 17.1

Banks	Number of Banks	SH_NET	SH_NON
Commercial	22	80.2	19.8
Special	6	92.4	7.6
Total Banks	28	82.9	17.1

Source: The authors.

Note: SH_NON represents the ratio of net non-interest income to sum of net interest income and non-interest income.

Korean banks are highly dependent on interest income for two reasons. First, their business lines are sensitive to interest rates risk. Second, their performance is related to the spread between the deposit rate and loan rate. The revenue from interest income is closely related to the spread: the higher the spread the more profits that banks make and vice versa, as seen in Table 2. The table presents the average of ROE, the spread between lending rates and deposit rates and the net interest margin (NIM) for Korean banks over the 1999–2014 period. In general, there is a decreasing pattern in ROE, spread and NIM, indicating that the performance of Korean banks has steadily deteriorated. Close examination of ROEs and spreads shows two regimes: one from 2001 to 2007 and the other from 2008 to 2014.¹² During the first regime, when the Korean banking industry made a significant effort to promote banking diversification, ROEs and spreads were greater than 10.00 per cent and 3.00 per cent, respectively. In the case of second regime, however, ROEs decreased, beginning in 2008, when the financial crisis occurred. In 2007, the ROE of Korean banks was 14.60 per cent but decreases by almost half in 2008 due to an aftershock of the global financial crisis. Also in the second regime, there is a steadily decreasing pattern in the spread between loan rates and deposit rates. Following the global financial crisis, the Bank of Korea began to lower interest rates on both loans and deposits to offset the effect of the Federal Reserve Bank’s drive for quantitative earnings.¹³ The change in interest rates hurt their NIM and resulted in the poor performance of Korean banks over the 2008–2014 period.

As can be seen in Table 2, Korean banks’ deteriorating performance is caused by the decrease in low NIM. We cannot attribute this poor performance in Korean banks, however, only to the low NIM. Our analysis shows that the average shares of non-interest income and net interest income as a percentage of total operating income during the first period are 23.4 per cent and 76.6 per cent, respectively. The average shares of net non-interest income and net interest income during the second period are 12.7 per cent and 87.3 per cent, respectively. Comparing the values in the first period with those in the second period, we see that the share of net interest income increases by 10.7 percentage points, while the share of net non-interest income decreases by 10.7 percentage points. These results indicate that Korean banks hesitated to diversify into non-bank activities. Their reluctance to expand into non-interest income-producing activities makes them vulnerable to interest rates risk, and this in turn can adversely affect banks’ total operating revenue.

Table 2.

Percentage of ROE, Spread between Lending Rate and Deposit Rate and Net Interest Margin

Year	ROE	Spread	LR^(a)	DR^(b)	NIM^(c)
1999	−14.38	3.35	10.65	7.30	2.47
2000	−11.02	3.07	9.71	6.64	2.39
2001	12.76	3.40	9.01	5.61	2.64
2002	10.91	3.52	7.69	4.17	2.70
2003	13.41	3.34	7.07	3.73	2.56
2004	15.16	3.44	6.76	3.32	2.59
2005	18.42	3.70	6.66	2.96	2.81
2006	14.64	3.43	6.68	3.25	2.61
2007	14.60	3.25	7.01	3.76	2.44
2008	7.17	2.99	7.48	4.49	2.30
2009	5.76	2.59	6.03	3.44	1.98
2010	7.22	2.93	5.89	2.96	2.32
2011	8.40	2.88	5.91	3.03	2.30
2012	6.17	2.59	5.63	3.05	2.10
2013	2.69	2.31	4.80	2.49	1.87
2014	4.05	2.18	4.34	2.16	1.79

Source: The authors.

Notes: (a) LR represents a lending rate; (b) DR represents a deposit rate and (c) NIM stands for net interest margin.

To determine why Korean banks reduced the share of net non-interest income as a percentage of total operating income, we break down the total non-interest income into five components (i.e., SH_FEE, SH_SEC, SH_FX, SH_TRUST, SH_OTHERS) and look at each component. Table 3 presents the share of five components for two periods: 1999–2007 and 2008–2014. Overall, it appears that the major component of total non-interest income is the fee commissions. Our results show, however, that the decline in the share of non-interest income is accompanied by a decrease in the share of fee commissions because SH_FEE in the second period is reduced significantly, approximately 0.5 per cent decrease from SH_FEE in the first period. These results, as seen in Table 3, suggest that the decline of earnings from fee commissions in the Korean banking industry may be attributed to a decrease in overall net non-interest income.

Table 3.

Percentage of the Average of SH_FEE, SH_SEC, SH_FX, SH_TRUST and SH_OTHERS During Two Market Regimes: 1999–2007 and 2008–2014

Regimes	SH_FEE	SH_SEC	SH_FX	SH_TRUST	SH_OTHERS
Years 1999–2007	18.10	0.41	2.91	−3.66	−4.51
Years 2008–2014	7.76	5.21	2.82	0.98	−4.10
Overall	13.95	2.18	2.94	−2.09	−4.50

Source: The authors.

To further examine the sources of net fee commissions, we decompose net fee commissions into two basic elements: (a) revenue from fee commissions and (b) expenses associated with fee commissions. Table 4 presents the means and standard deviations of the annual growth rate of revenue and expenses for fee commissions. We see that the average growth rate for expenses associated with fee commissions far exceeds than that of revenue from fee commissions. The results presented in the table show that the average growth rate for revenue and expenses are 2.83 per cent and 10.48 per cent, respectively, within similar deviations: 19.34 per cent for revenue and 24.38 per cent for expenses. This provides evidence that the decrease in the share of net fee commissions, as shown in Table 4, is due to the substantial increase in the expenses of net fee commissions.

Table 4.

Growth Rate for the Revenue and Expense of Fee Commissions

Statistics	Growth Rate for Revenue	Growth Rate of Expense
Mean	2.83	10.48
Standard Deviation	19.84	24.38

Source: The authors.

Figure 1 depicts the trend of the share of net revenue from fee commissions, the share of expenses associated with fee commissions and the share of net interest income revenue between 2001 and 2014. The graph shows a downward trend in the shares of net fee commissions over the sample period and that this is accompanied by a downward trend for the share of net non-interest income, although there is a zigzag pattern in the share of expenses associated with fee commissions and the share of net interest income over the same period. The figure shows that the share of expenses associated with fee commissions has steadily increased, while the share of net interest income has gradually decreased over the same period. The negative correlation between net non-interest income and expenses associated with fee commissions (−0.72) suggests that the lack of control of fee commissions negatively affects the level of non-interest income.

Figure 1.

Source: The authors.

Based on our findings, we conclude that Korean banks fail to manage the expenses associated with fee commissions, even though the revenue earned from fee commissions is the key to banking diversification. The mismanagement of the expenses reduces the portion of the share of fee commissions and, thus, shrinks overall non-interest income for the bank. Therefore, Korean bank managers are more likely to focus on earning net interest income from traditional banking activities than on earning non-interest income from non-banking activities.

4.2. Does the Diversification Improve the Performance and Risk Management of Korean Banks?

Korean banks earn most of their revenues from interest income. They avoid earning non-interest income because they incur substantial expenses through the generation of fee commissions. Nevertheless, a concentration on interest income-producing activities can lower the diversification benefits of Korean banks.

We consider how traditional banking activities and non-bank activities affect the performance and the risk of Korean banks. We conduct an empirical analysis by taking the logarithms of all dependent variables because all variables are not normally distributed and skewed to the right, as shown in Figure 2. After logarithmic transformation of all dependent variables, we run cross-sectional and panel regressions.

Figure 2.

Source: The authors.

4.2.1. Cross-sectional Analysis

Table 5 shows the direct and indirect effects of a change in the interest income share and non-interest income share on bank performance and bank risk. Risk-adjusted return on equity (RA_ROE) and risk-adjusted return on assets (RA_ROA) are two measures of bank performance, while the other two dependent variables (RA_CAR and Z-score) are measures of bank risk. Panel A presents the estimated indirect and direct impact of a 1 per cent change in the share of interest income on bank performance (i.e., RA_ROE, RA_ROA) as well as on bank risk (i.e., RA_CAR, Z-score). For example, in terms of indirect effect, a 1 per cent increase in the interest income share can lead to a significant decrease in RA_ROE and RA_ROA by 0.62 per cent and 1.50 per cent, respectively. Similarly, a 1 per cent increase in the share of interest income will result in a decrease in RA_CAR and Z-score by 0.255 per cent and 0.01 per cent, respectively. These effects on bank risk, however, are not statistically significant. In the case of the direct effect of the change in the interest income share on the performance measures, we find that a 1 per cent increase results in no significant effect on bank performance measures (RA_ROE and RA_ROA). In contrast, in the case of the direct effect of the change in the interest income share on the risk measures, the results show that a 1 per cent increase in interest income will lead to an increase in the risk measures of RA_CAR by 3.6 per cent and Z-score by 4.82 per cent. We also find that the estimates for the earlier two risk measures are statistically significant at the 0.01 level.

Table 5.
Estimated Impact of a Change in the Interest Income Share and Non-interest Income Share on RA_ROE, RA_ROA, RA_CAR and Z-Score

log(RA_ROE) log(RA_ROA) log(RA_CAR) log(Z-Score)

Panel A: A Change in the Interest Income Share on Dependent Variables

Indirect effect −0.6167 −1.4998 −0.2549 −0.0103

(0.2309) (0.5368) (0.1911) (0.1482)

Direct effect −0.0074 1.4582 3.6008* 4.8194*

(0.4959) (1.1526) (0.4103) (0.3182)

Panel B: A Change in the Non-Interest Income Share on Dependent Variables

Indirect effect −0.1006 −0.0756 −0.7286 −0.9385

(0.1227) (0.1282) (0.5192) (0.6453)

Direct effect 1.5831** 1.3233* 17.1641* 21.4564*

(0.6507) (0.6797) (2.7531) (3.4213)

	log(RA_ROE)	log(RA_ROA)	log(RA_CAR)	log(Z-Score)
Panel A: A Change in the Interest Income Share on Dependent Variables
Indirect effect	−0.6167**	−1.4998**	−0.2549	−0.0103
	(0.2309)	(0.5368)	(0.1911)	(0.1482)
Direct effect	−0.0074	1.4582	3.6008***	4.8194***
	(0.4959)	(1.1526)	(0.4103)	(0.3182)
Panel B: A Change in the Non-Interest Income Share on Dependent Variables
Indirect effect	−0.1006	−0.0756	−0.7286	−0.9385
	(0.1227)	(0.1282)	(0.5192)	(0.6453)
Direct effect	1.5831**	1.3233*	17.1641***	21.4564***
	(0.6507)	(0.6797)	(2.7531)	(3.4213)

Source: The authors.

Notes: Standard errors are given in parentheses. ***, ** and * indicate statistically significant at 1 per cent, 5 per cent and 10 per cent, respectively.

The results in Panel B of Table 5 show the indirect and direct effects of a 1 per cent change in the share of non-interest income on bank performance and risk measures. In the case of an indirect effect, our results show that a 1 per cent change in the non-interest income share has no significant effect on either bank performance or bank risk. However, what follows in the case of the direct effect of the change in the non-interest income share on bank performance and risk is that a 1 per cent increase in non-interest income can lead to increases in RA_ROE by 1.583 per cent, RA_ROA by 1.32 per cent, RA_CAR by 17.16 per cent and Z-score by 21.46 per cent.

In summary, an increase in the share of interest income has an indirect negative impact on bank performance, but an increase in the interest income share may lead directly to an increase in the bank’s risk. In contrast, an increase in the share of non-interest income is directly associated with a decrease in Korean bank risk and an increase in performance. Although the increase in the level of diversification does not generate statistically significant gains, Korean banks expect to reduce their risk through an increase in the share of non-interest income rather than through an increase in the interest income share because the impact of the change in the non-interest income share on bank risk is significantly greater than that of interest income (17.16% versus 3.60% for RA_CAR; 21.46% versus 4.82% for Z-score).

Based on this result, we conclude that diversification through an increase of the share of non-interest income had no significant influence on the Korean banking business. This may be attributed to two factors. First, diversification through non-banking activities can lead to an increase in the volatility of banking, thereby reducing the risk-adjusted returns. Second, an increase in non-interest income in Korea is more likely to be associated with a significant increase in non-interest expense. Thus, an increase in the level of diversification associated with an increase in the share of non-interest income appears to have no favourable impact on the performance of Korean banks.

Next, we study more thoroughly the effects of a change of non-interest income share on bank performance and risk. Table 6 presents the impact of a change in each non-interest income-share component and shows the respective direct and indirect effects on the return and risk of a bank.

Panel A of Table 6 provides the estimated impact of a 1 per cent change in the share of fee commissions. Specifically, the panel shows that the indirect effect of SH_FEE negatively affects risk-adjusted returns and bank risks, while the direct effect positively affects both risk and return measures. All of the estimates associated with indirect effect, however, are not significantly different from zero, indicating that Korean banks may not improve their performance and risk profiles through an increase in the share of fee commissions. Panel B shows the estimated direct and indirect impact of a 1 per cent change in the share of gains (or losses) on securities sales. The indirect estimation results in Panel B suggest that a shift towards non-interest income by increasing the share of security sales improves Korean banks’ risk-adjusted return. In contrast, the direct estimation results in Panel B show that an increase in non-interest income associated with securities-trading gains (or losses) can lead to a significant decline in RA_ROE and RA_ROA. Similarly, Panels C and E show the similar direct and indirect impact of a 1 per cent change in the share of foreign-exchange-trading gains (or losses) and other non-interest income-producing activities. In contrast, the results in Panel D show that diversification through a marginal increase in the share of trust management minimises the risk of Korean banks and enhances their profits. Although all estimates for the direct effect in Panel D provide positive signs, they are not statistically significant at the 10 per cent level.

Table 6.

Estimated Impact of a Change in the Share of Non-interest Interest Components on RA_ROE, RA_ROA, RA_CAR and Z-Score

	ln(RA_ROE)	ln(RA_ROA)	ln(RA_CAR)	ln(Z-score)
Panel A: A Change in the Share of Fee Commissions on Dependent Variables
Indirect effect	−0.0283	−0.0369	−0.0357	−0.03582
	(0.04396)	(0.0587)	(0.1365)	(0.1584)
Direct effect	1.8169**	3.2200**	24.0200***	30.4419***
	(0.8691)	(1.1594)	(2.6991)	(3.1305)
Panel B: A Change in the Share of Security Sales on Dependent Variables
Indirect effect	2.1078***	2.3001***	13.3901***	15.9245***
	(0.3044)	(0.3119)	(1.8497)	(2.1815)
Direct effect	−13.5919*	−13.5341*	−50.8110	−51.7989
	(7.0117)	(7.1865)	(42.6098)	(50.2532)
Panel C: A Change in the Share of Foreign Exchange Trading on Dependent Variables
Indirect effect	0.2088*	0.3564*	1.6533	1.9468
	(0.1987)	(0.2052)	(1.2291)	(1.4449)
Direct effect	−14.7797	−2.9181	−48.8238	−50.6265
	(17.8861)	(18.4671)	(110.6212)	(130.0383)
Panel D: A Change in the Share of Trust Management on Dependent Variables
Indirect effect	1.0323***	1.0090***	6.5128***	9.3955***
	(0.1133)	(0.1339)	(0.7022)	(1.0293)
Direct effect	2.0961	2.2937	44.4183	38.0343
	(6.4544)	(7.6288)	(40.0079)	(58.6409)
Panel E: A Change in the Share of Other Non-Interest Income on Dependent Variables
Indirect effect	1.3132***	1.4139***	6.4696***	8.9339***
	(0.1540)	(0.1673)	(0.5107)	(0.6134)
Direct effect	−32.1400	−22.8323	0.1715	−62.9123
	(19.7901)	(21.4925)	(65.6224)	(78.8087)

Source: The authors.

Notes: Standard errors are given in parentheses. ***, ** and * indicate statistically significant at 1 per cent, 5 per cent and 10 per cent, respectively.

Taken together, we find that an increase in the share of interest income indirectly reduces performance measures, while directly reducing the risk of Korean banks. In contrast, an increase in the share of non-interest income helps to directly reduce Korean banks’ risk and improve performance. When we segregate the non-interest income share into five components and examine their effects on bank performance and risk, our results are mixed. We find that it is the fee commission that offsets the gains from the diversification, whereas the other components, in general, help to improve the risk-adjusted returns.

4.2.2. Panel Analysis

The previous results provide evidence in support of a fee commission that offsets the gains from diversification. These results, however, are based only on cross-sectional regression analysis. In other words, we do not utilise time-series variations within a bank that may affect the relationship among performance, risk and diversification. To address this, we conduct a panel analysis, exploiting variations of the results across banks over time. We include bank fixed effects to mitigate a potential omitted variable bias. We also focus on the disaggregated non-interest income share rather than the aggregated one, as the effect of the diversification through changes in aggregated income share is negligible. Thus, it is essential to find a reason that the banking industry in Korea does not gain from diversification. To this end, we scrutinise the disaggregated non-interest income shares.

We report fixed-effects regression results for Equation (12) in Table 7. There are several notable findings. First, an increase in SH_TRUSThas a statistically significant and positive impact on RA_ROE RA_ROA, while all other non-interest income activities do not have any effects on the two performance measures. Specifically, a 1 per cent increase in SH_TRUST results in a 13.28 per cent increase in RA_ROE and a 9.61 per cent increase in RA_ROA. Second, we find that changes in the non-interest income shares have mixed effects on risk measures. Although statistically insignificant, SH_FX and SH_OTHER are likely to have a positive influence on risk measures (RA_CAR and Z-score), while SH_TRUST is likely to have a negative impact on bank risk. In addition, SH_FEE and SH_OTHER can lead to an increase or a decrease in bank risk, depending on the risk measure. Third, CAR has a positive effect on the performance measures, and the effect is statistically significant at the 0.01 level. Specifically, a 1 per cent increase in CAR results in an approximately 14 per cent increase in both ROA and ROE. This might be attributed to the Basel III agreement by Banks for International Settlements (BIS), through which Korean banks strengthen the regulatory equity capital framework. Moreover, we find that the term spread is the other crucial factor that affects bank performance measures. As seen in Table 7 (Columns 2 and 3), a 1 per cent rise in the term spread increases the risk-adjusted ROE and ROA by 0.508 per cent and 0.687 per cent, respectively. These findings bolster our previous results, indicating that Korean banks are more likely to focus on traditional banking business. The operating revenue that is highly dependent on interest income, however, is more sensitive to fluctuation in interest rates. Finally, we find that the risk measures are affected by a macro-shock, especially the real GDP (RGDP) per capita.

Table 7.
Fixed-effect Regressions on RA_ROE, RA_ROA, RA_CAR and Z-score

ln(RA_ROE) ln(RA_ROA) ln(RA_CAR) ln(Z-score)

SH_FEE −2.079 −1.210 3.064* −0.366

(1.231) (1.231) (1.013) (1.066)

SH_SEC −0.236 −0.388 1.644 −2.615*

(1.020) (0.912) (0.605) (0.597)

SH_FX −0.257 0.0261 −0.332 −0.392

(0.949) (0.948) (0.514) (0.789)

SH_TRUST 13.28 9.611 2.776 2.199

(4.814) (4.228) (3.144) (3.397)

SH_OTHERS 0.456 0.589 −0.470 −0.705

(0.921) (0.922) (0.438) (0.765)

DIV_ADJ 0.301 0.331 0.0887 0.455

(0.319) (0.295) (0.145) (0.181)

CAR 14.66* 14.84* 0.182 2.896

(3.675) (3.958) (3.726) (4.598)

ln(real asset) 0.125 0.177 −0.557 −0.0405

(0.639) (0.627) (0.380) (0.500)

Asset growth 1.676* 1.419* 1.459 0.632

(0.939) (0.801) (0.611) (0.569)

RGDP per capita 0.0227 0.0404 0.263* 0.173**

(0.104) (0.105) (0.0744) (0.0727)

Term spread 0.508* 0.687 0.110 0.305

(0.295) (0.298) (0.243) (0.179)

Credit spread −0.190 −0.254 −0.129 −0.0795

(0.117) (0.115) (0.109) (0.0678)

Constant −1.676 −2.384 3.749 1.028

(5.238) (5.107) (3.293) (4.368)

Observations 240 240 240 240

R-squared 0.203 0.198 0.231 0.307

	ln(RA_ROE)	ln(RA_ROA)	ln(RA_CAR)	ln(Z-score)
SH_FEE	−2.079	−1.210	3.064***	−0.366
	(1.231)	(1.231)	(1.013)	(1.066)
SH_SEC	−0.236	−0.388	1.644**	−2.615***
	(1.020)	(0.912)	(0.605)	(0.597)
SH_FX	−0.257	0.0261	−0.332	−0.392
	(0.949)	(0.948)	(0.514)	(0.789)
SH_TRUST	13.28**	9.611**	2.776	2.199
	(4.814)	(4.228)	(3.144)	(3.397)
SH_OTHERS	0.456	0.589	−0.470	−0.705
	(0.921)	(0.922)	(0.438)	(0.765)
DIV_ADJ	0.301	0.331	0.0887	0.455**
	(0.319)	(0.295)	(0.145)	(0.181)
CAR	14.66***	14.84***	0.182	2.896
	(3.675)	(3.958)	(3.726)	(4.598)
ln(real asset)	0.125	0.177	−0.557	−0.0405
	(0.639)	(0.627)	(0.380)	(0.500)
Asset growth	1.676*	1.419*	1.459**	0.632
	(0.939)	(0.801)	(0.611)	(0.569)
RGDP per capita	0.0227	0.0404	0.263***	0.173**
(0.104)	(0.105)	(0.0744)	(0.0727)
Term spread	0.508*	0.687**	0.110	0.305
	(0.295)	(0.298)	(0.243)	(0.179)
Credit spread	−0.190	−0.254**	−0.129	−0.0795
	(0.117)	(0.115)	(0.109)	(0.0678)
Constant	−1.676	−2.384	3.749	1.028
	(5.238)	(5.107)	(3.293)	(4.368)
Observations	240	240	240	240
R-squared	0.203	0.198	0.231	0.307

Source: The authors.

Notes: Clustered standard errors are given in parentheses. ***, ** and * indicate statistically significant 1 per cent, 5 per cent and 10 per cent, respectively.

In the cross-sectional analysis, we find that the fee commission offsets the gains from the diversification. To examine whether this finding also holds when considering time-series variations within banks, we repeat the fixed-effect regression model of Equation (12) using two subsamples. Specifically, we divide the sample into two groups, based on the expense of fee commissions: (a) low-expense banks that are below the median value and (b) high-expense banks that are above the median value. As seen in Table 8, two findings are notable. First, the estimated effects of SH_FEE on the performance measures are significantly more negative in banks that are characterised by the high expense of fee commissions. This implies that the same increase in the share of fee commission activity would reduce the performance measures more in banks with high expenses. Second, the estimated effects of SH_FEE on the risk measures suggest that the higher the fee commission expense the higher the default risk. Taken together, the two results support our findings from the cross-sectional analysis.

Table 8.

Fixed-effect Regressions on RA_ROE, RA_ROA, RA_CAR and Z-score for the Two Subsamples

	ln(RA_ROE)		ln(RA_ROA)		ln(RA_CAR)		ln(Z-score)
	Below	Above	Below	Above	Below	Above	Below	Above
SH_FEE	−1.451	−6.234***	−0.820	−4.346**	4.548***	1.431	1.536	−3.418***
	(2.104)	(2.059)	(2.011)	(2.033)	(1.242)	(2.496)	(1.713)	(1.040)
SH_SEC	−0.242	−5.622***	−0.421	−5.019***	0.356	2.165	−2.301***	−6.995***
	(1.115)	(1.422)	(0.986)	(1.494)	(1.019)	(2.491)	(0.642)	(1.280)
SH_FX	0.858	−4.495*	0.900	−2.936	−0.355	−0.645	1.074	−3.500
	(1.190)	(2.552)	(1.109)	(2.633)	(0.779)	(1.449)	(0.785)	(2.982)
SH_TRUST	17.66	2.232	12.11	1.290	−7.032	6.981	5.066	−7.175
	(12.32)	(7.291)	(10.14)	(7.413)	(5.110)	(4.452)	(8.955)	(5.426)
SH_OTHERS	0.943	−2.912	0.922	−1.719	−0.918	−0.398	−0.0908	−3.005
	(1.028)	(2.571)	(0.956)	(2.576)	(0.862)	(1.659)	(0.606)	(2.505)
DIV_ADJ	−0.374	3.333***	−0.373	2.861***	0.636***	0.312	0.101	2.240***
	(0.600)	(0.887)	(0.529)	(0.777)	(0.162)	(1.055)	(0.295)	(0.611)
CAR	23.78***	1.041	23.54***	3.799	8.181*	2.592	11.03	−4.225
	(4.516)	(8.035)	(4.036)	(6.618)	(4.206)	(6.006)	(7.856)	(7.185)
ln(real asset)	−0.911	0.456	−0.865	0.587	0.00628	−0.290	0.142	−0.282
	(1.198)	(0.953)	(1.071)	(0.914)	(0.849)	(0.388)	(0.930)	(1.033)
Asset growth	1.317	1.712	0.968	1.806	1.278	2.232*	−0.0895	1.541*
	(1.384)	(1.305)	(1.146)	(1.231)	(0.898)	(1.230)	(0.640)	(0.876)
RGDP per capita	0.181	−0.00396	0.202	−0.00132	0.218	0.184**	0.134	0.234*
RGDP per capita	(0.171)	(0.167)	(0.152)	(0.163)	(0.164)	(0.0685)	(0.123)	(0.133)
Term spread	0.633	0.599	0.849*	0.716*	0.428	−0.268	0.261	0.438*
	(0.405)	(0.361)	(0.418)	(0.346)	(0.321)	(0.350)	(0.322)	(0.227)
Credit spread	−0.260	−0.188	−0.344**	−0.226	−0.195	−0.0252	−0.0493	−0.0974
	(0.156)	(0.142)	(0.153)	(0.137)	(0.180)	(0.142)	(0.123)	(0.0782)
Constant	5.032	−4.057	4.541	−5.782	−1.500	2.138	−0.656	3.149
	(8.854)	(8.606)	(8.119)	(8.139)	(5.921)	(4.471)	(7.132)	(10.10)
Observations	122	118	122	118	122	118	122	118
R-squared	0.246	0.255	0.254	0.236	0.312	0.288	0.396	0.335

Source: The authors.

Notes: Clustered standard errors are given in parentheses. ***, ** and * indicate statistically significant 1 per cent, 5 per cent and 10 per cent, respectively.

5. Conclusions

We examine whether the diversification of the operating income in Korean banks has consistently enhanced the performance of Korean banks. To this end, we conduct both cross-sectional and panel analyses, using financial statement data of 28 Korean banks from 2001 to 2014. Based on our empirical analysis, we find that despite bank managers’ efforts to diversify sources of operating revenue, Korean banks do not gain any benefits from the diversification and, rather, rely heavily on interest income-producing activities. The net results are weaker earnings and stronger balance sheets. We also find that the gains from the diversification effect are offset by expenses. As a result, Korean banks fail to diversify into non-interest income-producing activities. Instead, they stick to the traditional interest income-producing activities, which eventually cause them to be exposed to interest rate risk. Since the 2008–2009 global financial crisis, the low risk-taking behaviour of banks, combined with flattening yield curves, has had a significant negative impact on the performance of Korean banks.

Our findings for the Korean banking industry are consistent with the global trend towards lower interest rates and lower operating income among commercial and investment banks throughout the post-financial crisis period. Although Korean banks are not yet experiencing negative interest rates, the prospect of persistently low interest rates is likely to continue to depress bank profit margins. This implies that to mitigate this expected decline in Korean banks’ profits, they should restore non-traditional business income activity. However, in order to motivate them to refocus on non-interest income business, Korean government super- visory and banking associations should make efforts to curtail non-interest income expenses and to lift regulations so that Korean banks can freely create various non-interest income products and charge-realised fees and commissions to their customers.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Footnotes

Notes

References

Baek

, Lee

N. H.

, & Cha

S. Y.

(2015). The effect of the diversification in Korean banks: The impact on profit and risk. Journal of Accounting and Finance, 15(2), 51–69.

Baele

, De Jonghe

, & Vander Vennet

(2007). Does the stock market value bank diversification? Journal of Banking and Finance, 31(7), 1999–2023.

Behr

, Kamp

, Memmel

, & Pfingsten

(2007). Diversification and the banks' risk-return-characteristics: Evidence from loan portfolios of German Bank, Discussion Paper Series 2: Banking and Financial Studies 2007,05. Germany: Deutsche Bundesbank.

Berger

A. N.

, Hasan

, & Zhou

(2010). The effects of focus versus diversification on bank performance: Evidence from Chinese banks. Journal of Banking and Finance, 34(7), 1417–1435.

Böninghausen

, & Köhler

(2015). Diversification and determinants of international credit portfolios: Evidence from German banks. International Review of Economics and Finance, 15, 57–75.

Busch

, & Kick

(2009). Income diversification in the German banking industry. Frankfurt am Main: Dt. Bundesbank.

Chiorazzo

, Milani

, & Salvini

(2008). Income diversification and bank performance: Evidence from Italian banks. Journal of Financial Services Research, 33(3), 181–203.

Čihák

, & Hesse

(2010). Islamic banks and financial stability: An empirical analysis. Journal of Financial Services Research, 38(2–3), 95–113.

Demirgüç-Kunt

, & Huizinga

(1999). Determinants of commercial bank interest margin and probability: Some international evidence. World Bank Economic Review, 13(2), 379–408.

10.

Demirgüç-Kunt

, & Huizinga

(2010). Bank activity and funding strategies: The impact on risk and returns. Journal of Financial Economics, 98(3), 626–650.

11.

Elton

, Gruber

, Agrawal

, & Mann

(2001). Explaining the rate spread on corporate bonds. Journal of Finance, 56(1), 247–277.

12.

Engle

, Moshirian

, Sahgal

& Zhang

(2014). Banks non-interest income and global financial stability. Research report (Working Paper No. 015). Australia: CIFR..

13.

Fama

E. F.

, & French

K. R.

(1992). The cross-section of expected stock returns. Journal of Finance, 47(2), 427–465.

14.

Fama

E. F.

, & French

K. R.

(1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics, 33(1), 3–56.

15.

Hsieh

M. F.

, Chen

P. F.

, Lee

C. C.

, & Yang

S. J.

(2013). How does diversification impact bank stability? The role of globalization, regulations, and governance environments. Asia-Pacific Journal of Financial Studies, 42(5), 813–844.

16.

Köhler

(2014). Does non-interest income make banks more risky? Retail-versus investment-oriented banks. Review of Financial Economics, 23(4), 182–193.

17.

Kwast

(1989). The impact of underwriting and dealing on bank returns and risk. Journal of Banking and Finance, 13(1), 101–125.

18.

Laeven

, & Levine

(2009). Bank governance, regulation and risk taking. Journal of Financial Economics, 93(2), 259–275.

19.

Lee

C. C.

, & Hsieh

M. F.

(2013). The impact of bank capital on profitability and risk in Asian banking. Journal of International Money and Finance, 32(1), 251–281.

20.

Lepetit

, Nys

, Rous

, & Tarazi

(2008). Bank income structure and risk: An empirical analysis of European banks. Journal of Banking and Finance, 32(8), 1452–1467.

21.

, & Zhang

(2013). Are there diversification benefits of increasing noninterest income in the Chinese banking industry? Journal of Empirical Finance, 24, 151–165.

22.

Meslier

, Tacneng

, & Tarazi

(2014). Is bank income diversification beneficial? Evidence from an emerging economy. Journal of International Financial Markets, Institutions and Money, 31(2), 97–126.

23.

Morgan

, & Samolyk

(2003). Geographic diversification in banking and its implications for bank portfolio choice and performance. New York, NY: Federal Reserve Bank of New York.

24.

Natalia

M. R. K.

, & Firsty

R. R.

(2016). Bank income diversification from stock market perspective: Evidence from ASEAN+3. Indonesian Capital Market Review, 8(1), 32–45.

25.

Park

, & Weber

(2006). Profitability of Korean banks: Test of market structure versus efficient structure. Journal of Economics and Business, 58(3), 222–239.

26.

Sanya

, & Wolfe

(2011). Can banks in emerging countries benefit from revenue diversification? Journal of Financial Services Research, 40(1–2), 79–101.

27.

Saunders

, & Walter

(1994). Universal banking in the United States, What could we gain? What could we lose? New York, NY: Oxford University Press.

28.

Stiroh

K. J.

(2004a). Do community banks benefit from diversification? Journal of Financial Service Research, 25(2–3), 135–160.

29.

Stiroh

K. J.

(2004b). Diversification in banking: Is noninterest income the answer? Journal of Money, Credit and Banking, 36(5), 853–882.

30.

Stiroh

K. J.

(2006). A portfolio view of banking with interest and noninterest activities. Journal of Money, Credit, and Banking, 38(5), 1351–1361.

31.

Stiroh

K. J.

, & Rumble

(2006). The dark side of diversification: The case of US financial holding companies. Journal of Banking and Finance, 30(8), 2131–2161.

32.

Sufian

(2011). Profitability of the Korean banking sector: Panel evidence on bank-specific and macroeconomic determinants. Journal of Economics and Management, 7(1), 43–72.

33.

Thomas

(2002). Firm diversification and asymmetric information: Evidence from analysts’ forecasts and earnings announcements. Journal of Financial Economics, 64(3), 373–396.

Diversification in Korean Banking Business: Is Non-interest Income a Financial Saviour?

Abstract

Keywords

1. Introduction

2. Data Sources and Variables

2.1. Variable Construction

2.1.1. Diversification Measures

2.2. Data Sources

3. Empirical Framework

3.1. Cross-sectional Analysis

4.1. Do Korean Banks Hesitate to Diversify Operating Income by Expanding into Non-banking Activities?

Table 1. Percentage of the Shares of Net Interest and Non-interest Income Banks Number of Banks SHNET SHNON Commercial 22 80.2 19.8 Special 6 92.4 7.6 Total Banks 28 82.9 17.1

Declaration of Conflicting Interests

Funding

Footnotes

Notes

References

Table 1.
Percentage of the Shares of Net Interest and Non-interest Income

Banks Number of Banks SH_NET SH_NON

Commercial 22 80.2 19.8

Special 6 92.4 7.6

Total Banks 28 82.9 17.1