March 2008 working paper no. 330 Issn 1975-5163 Joon-Ho Hahm

 
 
Keywords: Commercial Bank; Non-interest Income; Bank Profitability; Bank Risk; 
Macroeconomic Volatility 
JEL Classification: G2; E5 

---

---

1 
Ⅰ. Introduction 
The evolution of Korea’s banking industry since the 1997 financial crisis can be 
characterized by both financial consolidation and conglomeration. While bank 
consolidation is a result of merger and acquisitions within the same industry, 
conglomeration is a result of strengthened cross-sector activities between bank and non-
bank financial institutions. The conglomeration trend in the banking industry is a global 
phenomenon driven by deregulation and increased competition, both of which have 
eroded traditional comparative advantages of banks. Banks in the US and Europe have 
reacted to declining market share and profitability by expanding non-traditional 
businesses. Korean banks are also making effort to diversify their revenue sources toward 
non-interest income such as fee-based financial services. However, it is uncertain whether 
these shifts can lead to higher and more stable earnings for commercial banks. 
The profitability and risks of banks can be impacted by financial conglomeration 
through various channels. First, profitability and earnings potential can be enhanced if 
banks exploit and realize scope economies from product diversity and cross-selling 
opportunities. For instance, banks in a financial group can share marketing and 
distribution channels as well as database and IT systems, spreading out their fixed costs. 
With enhanced profitability and cost efficiency, bank insolvency risk can be reduced, and 
higher charter value in turn reduces moral hazard incentives of banks. 

---

2 
Second, conglomeration may lower the riskiness of individual banks since it 
provides greater opportunities for risk diversification. Cross-industry diversification may 
contribute to reductions in earnings variability if returns are sufficiently heterogeneous 
across different financial services. On the other hand, banks can be exposed to unexpected 
shocks, arising from the capital market and other financial sectors, through their non-bank 
financial affiliates. 
Third, while large and complex banking organizations may be able to benefit from 
the scale and scope economies and risk diversification, operational risk may substantially 
increase with growing organizational complexity, inefficiencies in management and 
internal control, and difficulties of harmonizing risk management. (Group of 10, 2001) 
Furthermore, for the financial system as a whole, banking sector consolidation and 
conglomeration may increase systemic risk potential. Although the extent of 
diversification may increase at individual bank level, large banking conglomerates tend to 
share increasingly similar characteristics in their business portfolios and asset structures. 
This paper studies implications of the changing revenue structure of commercial 
banks in the era of financial conglomeration. On a bank level, we can take the amount of 
revenue generated through non-bank activities as a proxy for bank conglomeration. By 
focusing on the non-interest income share of commercial banks, this paper empirically 
assesses benefits, risks, and potential determinants of bank revenue diversification. Since 
the shift toward non-interest income has been a global phenomenon caused by recent 
progress in financial deregulation and disintermediation, we investigate multi-country 

---

3 
evidence by utilizing firm level data of 662 relatively large commercial banks in 29 
OECD countries.
More specifically, this paper addresses the following questions: What are the 
factors that influence the degree of non-interest income diversification on a bank level? In 
addition to bank specific factors, are macroeconomic factors important determinants of the 
non-interest income ratio? What are the impacts of the expansion toward non-interest 
income on bank profitability, capital adequacy, profit variability and insolvency risks? 
What are the impacts of the increasingly homogeneous bank revenue structure within 
countries on macroeconomic volatilities? The paper is organized as follows: Section 2 
overviews the existing literature. Section 3 describes data, methodology and empirical 
findings. Section 4 summarizes and concludes. 

---

4 
Ⅱ. Literature Review 
The structure of the international financial industry has changed dramatically over 
the past twenty years and these changes have been documented extensively in the 
literature. One notable feature of this change is in the composition of financial institutions’ 
businesses and product lines, namely, a movement towards financial conglomeration. This 
shift stems from a more competitive market environment in which financial institutions 
are actively seeking strategies to cut costs and enhance revenue. 
In order to better understand the effects of conglomeration, we must first 
understand factors driving the recent surge in banks’ shares of non-interest revenue. 
Widespread deregulations accompanied by technological development have reshaped the 
profit structure of banks. There now exist more efficient means of production of financial 
information and better techniques for assessing and pricing risks, which have shifted firms’ 
financial demands from traditional bank loans to non-bank funding sources such as 
corporate bonds and commercial papers. At the same time, there have been changes in the 
structure of financial savings with household portfolios moving away from bank deposits 
towards more diverse non-bank and capital market related products (Davis and Tuori, 
2000). Under the increased competition and with the increased demand for diverse 
financial products, commercial banks have been increasingly diversifying their financial 
services as a means of realizing scope economies. 

---

5 
On a bank level, there are several factors that have led to banks’ decisions to 
diversify their revenue structure. Rogers and Sinkey (1999) find that bank size is 
positively correlated with non-interest income, while net-interest margins and core 
deposits are both negatively correlated with non-interest income. Banks with fewer core 
deposits and lower net-interest margins earn less revenue from traditional interest income 
sources and so they must transition into non-traditional banking in order to remain 
profitable. 
De Young and Hunter (2003) and De Young et al. (2004) laid forth two banking 
strategies that help to explain the role that bank size plays in conglomeration. Relatively 
large banks make use of economies of scale in order to dominate the production of 
consumer loans. Despite their low unit costs, the market for these products is very 
competitive, and in order to make this model profitable, large banks must supplement their 
revenue stream with non-interest income. Smaller banks are able to earn higher interest 
margins, despite their relatively high unit costs. This is because they can charge higher 
interest rates due to higher switching costs on the part of borrowers. They can also pay 
lower interest rates because of their relatively loyal customer bases. Therefore, for smaller 
banks, non-interest income is less important. 
The diversification of bank revenue sources has also led to empirical research on 
the question of how non-traditional bank activity affects profitability and risk. The 
evidence seems to be mixed and varies from study to study. Saunders and Walter (1994) 
review 18 of the related studies that examine whether non-bank activities reduce bank risk. 

---

6 
They conclude that the evidence is mixed, with differing results across geographical 
regions. In the US, most studies agree that the diversification benefits of non-interest 
income are somewhat limited and that it often serves to increase bank risk. In Europe 
results have been more in line with the conventional wisdom that non-interest income 
stabilizes profit and thus offers diversification benefits. 
In analyzing the risk associated with non-interest income the empirical literature 
has taken two broad approaches: counterfactual simulation studies and studies that use 
data from banks that are actually involved in diversification. The counterfactual studies 
simulate mergers between banking and non-banking institutions and then analyze the 
riskiness of the simulated financial institution. These studies seem to support the 
hypothesis that small scale non-banking activity can reduce risk and increase profits. Wall 
et al. (1993) find that for 1981-1989 diversifications into insurance, mutual funds, 
securities brokerage, or real estate led to higher returns and lower risk. Boyd et al. (1993) 
use data for 1971-1987 and find that a bank holding company can reduce risk by merging 
with insurance firms, but will increase risk by merging with securities or real estate firms. 
Recently, Lown et al. (2000) find that diversification into life insurance has the biggest 
potential to reduce risk. 
Recent studies that directly investigate banking data to analyze profit and risk 
show mixed evidence on the impact of non-interest income. Using US data, Stiroh (2004a) 
finds that, in the aggregate, non-interest income is quite volatile and increasingly 
correlated with interest income, suggesting that there are few diversification benefits 

---

7 
associated with non-interest income. At the bank level, he finds that a greater reliance on 
non-interest income, especially trading income, tends to decrease risk-adjusted profits and 
increase risk. Stiroh (2004b) finds similar results for smaller community banks. De Young 
and Rice (2004) find that well-managed banks expand more slowly into non-interest 
activity and that greater levels of non-interest income are associated with poorer risk-
return tradeoffs. 
De Young and Roland (2001), using 472 US commercial bank data between 1988 
and 1995, find that non-traditional activities of banks are associated with both higher 
revenue volatility and higher total leverage. They suggest three observations as potential 
explanations for why fee-based banking is not necessarily more stable than traditional 
banking. First, there are considerable costs associated with changing lenders; so traditional 
loans are usually quite stable. These costs are not associated with abandoning fee-based 
services. Second, increasing fee-based services often requires substantial fixed costs. 
Banks must invest in new technology and staff. These costs serve to increase the 
operational leverage of banks. Third, fee-based services do not have significant regulatory 
capital requirements, suggesting that they increase financial leverage and thereby risk. 
Since, many of the characteristics that lead to conglomeration are country specific, 
the levels of non-banking activity may also differ across borders (Davis and Tuori, 2002 
and Smith et al., 2003). Also, since cost-efficiency frontiers differ across borders, 
efficiency changes are likely to differ depending upon geography. Fecher and Pestieau 
(1993) find that measures of efficiency are correlated with levels of regulation and 

---

8 
competition in OECD countries. It follows that empirical analyses of non-interest income 
in Europe draw different conclusions than the work conducted on banks in the US.
Smith et al. (2003) use a portfolio view of banking to analyze 2,655 banks within 
15 European countries from 1994-1998. They find that, although non-interest income is 
positively associated with revenues, the positive effect is offset by the higher operating 
costs that accompany fee-based services. Furthermore, they find that, while volatile, the 
income from these services serves to stabilize profits. They also draw a different 
conclusion than Stiroh (2004a) about the relationship between non-interest income and 
interest income, finding that the two are negatively correlated suggesting potential 
diversification benefits. This difference could stem from the fact that fee-based banking 
has played a minor role in Europe than in the US (Davis and Tuori 2002). Valverde and 
Fernandez (2007), using European bank data, also find that revenue and market power 
increase as output becomes more diversified towards non-traditional activities in banking. 
As for the system-wide risk, there is concern that as the industry moves toward 
conglomeration, systemic risk potential may increase. First, increased concentration of the 
financial industry around a small number of large banking conglomerates that are too big 
to fail may significantly increase systemic risk potential. The emergence of a few large 
banking institutions may also undermine the effectiveness of financial supervision and 
market monitoring. Increasing complexity of financial conglomerates would make it more 
difficult for regulators and market participants to comprehend risks and take early 
corrective actions. As a consequence, excessive risk taking and moral hazards on the part 

---

9 
of large banks may lead to higher systemic risk potential. 
Increasing degree of interdependence among the large and complex banking 
conglomerates is another source of systemic risk potential. The Group of 10 Ferguson 
report (2001) indicates that areas of direct interdependency include mutual credit risk 
exposures through inter-bank loans, on and off-balance sheet activities such as financial 
derivatives, and payment and settlement relationships. The systemic risk potential may 
also increase if large conglomerates are simultaneously and similarly exposed to adverse 
shocks due to homogeneity in asset and revenue structures. While banking conglomerates 
are able to diversify within each group, they are getting more homogeneous as business 
areas become increasingly similar. Resulting indirect interdependencies among large 
conglomerates may raise systemic risk potential as well. 
Finally, financial conglomeration may aggravate the problem of systemic risk as 
banks expand their involvement in high risk activities that are closely tied to non-bank 
financial firms and capital markets. As a result, banking institutions would be more 
vulnerable to contagion risks from non-bank and non-financial sectors as well as capital 
markets. Examples include banks’ active securitization of mortgage and credit card loans. 
The use of identical brand names for affiliated non-bank subsidiaries may also erode the 
firewall within a conglomerate and increase pressure for banks to bail out affiliated non-
bank subsidiaries. The shift of financial savings from bank deposits to affiliated non-bank 
financial subsidiaries also implies a de facto extension of the public safety net. 
The body of empirical literature on systemic risk is somewhat limited. G10 report 

---

10 
(2001) suggests that interdependencies among large and complex banking organizations 
have increased over the last decade in the US and Japan and began to increase in Europe. 
De Nicolo and Kwast (2002) investigate the systemic risk potential presented in the US 
banking industry over the period of 1988-99 based upon correlation measures of stock 
returns of large and complex banking organizations. They find a positive consolidation 
elasticity of stock return correlations and interpret the evidence as suggesting that 
systemic risk potential increases with consolidation in the banking industry. As for the 
cross-country studies, empirical evidence seems to be mixed. Beck et al. (2003), using a 
logit model, find that banking crises are less likely in countries with a more concentrated 
banking system. On the other hand, De Nicolo et al. (2003) report that the aggregate z-
score index obtained from the top 5 banks in each country is significantly negatively 
associated with the degree of bank concentration. Namely, bank consolidation is positively 
associated with the systemic risk potential. 
In Korea, some recent studies examine the consequences of financial 
conglomeration on profitability and risks of commercial banks. As the conglomeration 
trend is a relatively recent phenomenon, the empirical evidence is still mixed. Kang (2001) 
investigates the risk impact of non-interest income using z-score and return on assets’ 
variability for commercial banks. He finds that the expansion of non-interest businesses 
such as securities trading contributes to the reduction of banks’ earnings risk. Bae (2006) 
also finds that non-interest income contributes to higher profitability as well as lower 
profit variability in Korea, and interprets this positive evidence as resulting from the fact 

---

11 
that Korean banks focus on relatively stable fee income and credit card businesses rather 
than securities, foreign exchange and derivative activities. 
Hahm and Kim (2006) perform a more general study of the z-score index and its 
risk components for financial conglomerates in Korea. In their panel regressions, they find 
that, during the post-crisis period of 2001-03, asset size is significantly positively related to 
profitability and significantly negatively related to risk as measured by the standard 
deviation of ROA. However, they find no significant conglomeration effects on profitability 
and risk after controlling for the size effect. Oh et al. (2007) investigate the systematic and 
idiosyncratic components of risks as measured by stock prices of Korean banks and find 
that financial conglomeration tends to stabilize financial risks. 
As for the systemic risk potential, Hahm and Hong (2003) provide a diagnostic 
analysis of various channels through which financial consolidation and conglomeration 
influence financial stability. They find that both direct and indirect interdependencies 
among large banks have increased substantially after the crisis in Korea, which may have 
also raised the potential for systemic risk. Using bank panel data from 1999 to 2004, Kang 
(2007) also finds that herd behavior among Korean commercial banks has increased after 
the crisis. He also reports evidence that increases in asset size and expansion of retail 
businesses have negatively affected bank stability, while business diversification and 
governance transparency have exerted positive effects on bank stability. 

---

12 
Ⅲ. Methodology and Empirical Results