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

4. Homogeneity in Bank Revenue Structure and Macroeconomic Volatility 
As discussed in the literature review above, financial consolidation and 
conglomeration may increase systemic risk potential because, although the extent of 
diversification can increase at individual banks, large banks tend to share increasingly 
similar characteristics in their business portfolios and asset structures. This section focuses 
on the degree of homogeneity in bank revenue structure across banks within countries and 
explores its implication on macroeconomic volatility. More specifically, we estimate the 
following simple cross-country regressions: 
j
j
j
j
j
MCF
SDNIIR
AVNIIR
MCV










'
(3) 
Where MCV
j
is a macroeconomic volatility variable for country j. The macroeconomic 
volatility variables are measured using the standard deviations of inflation, real GDP 
growth, the ratio of stock market capitalization to GDP, and real interest rate. AVNIIR is 
the time-average of mean non-interest income ratio across banks in a given country and 
SDNIIR is the time-average of the standard deviation of the non-interest income ratio 
across banks in that country. In other words, AVNIIR is constructed by computing the 
average NIIR across banks for each year, and then taking average across the time period 
(1992-2006). Hence, it denotes the degree of revenue diversification toward non-interest 

---

22 
income for each country. Likewise, SDNIIR is constructed by taking average of yearly 
standard deviations in non-interest income ratios across banks, and denotes the degree of 
heterogeneity in bank revenue structure within that country. We test the impact of the 
average level of non-interest income as well as the heterogeneity of non-interest income 
ratios on macroeconomic volatilities by including macroeconomic indicators as 
controlling variables. 
Table 6 reports OLS estimation results for regression (3). Note that the inflation 
rate is significantly positively associated with its own volatility as well as real interest rate 
volatility. The stock market capitalization to GDP ratio is also positively correlated with 
its own volatility. As for the commercial bank non-interest income ratio, the average level 
itself does not seem to be significantly associated with macroeconomic volatilities. Note 
however that, the standard deviation of non-interest income ratios across banks is 
significantly negatively associated with the standard deviation of the inflation rate whether 
or not other macroeconomic volatilities are included as additional control variables. Hence, 
we may conclude that increasingly homogeneous bank revenue structure tends to be 
associated with higher inflation volatility; potentially through higher indirect 
interdependence and herd behavior among banks. As banks share a common revenue 
structure, they may respond to shocks in an increasingly similar manner, thereby causing 
higher inflation volatility. 

---

23 
Ⅳ. Summary and Conclusion 
This paper investigated potential determinants and consequences of the changing 
revenue structure of commercial banks in the era of financial deregulation and 
conglomeration. Utilizing a firm-level dataset of 662 commercial banks from 29 OECD 
countries, we found several interesting empirical patterns, which can be summarized as 
follows: 
First, we investigated the factors that influence banks’ choice of the degree of non-
interest income diversification. We found that, in addition to bank specific factors such as 
total asset size, equity-asset ratio, and net interest margin, macroeconomic factors such as 
real GDP growth, inflation rate, and stock market capitalization are potentially important 
determinants of bank non-interest income shares. Commercial banks in countries with 
slow economic growth, a stable inflation environment, and well-developed capital markets 
tend to have higher non-interest income shares. 
Second, we investigated potential impacts of rising non-interest income shares on 
bank profitability and risks. We found that banks with higher non-interest income shares 
tend to exhibit a higher return on assets as well as higher equity-asset ratios. However, 
these banks also exhibit higher variances in return on assets. Consequently, diversification 
toward non-interest income does not seem to exert a significant impact on bank 
insolvency risk as measured by z-score. Namely, both the sum of earnings and capital, and 

---

24 
earnings variability seem to increase simultaneously without affecting net insolvency risk 
of banks. 
Finally, financial consolidation and conglomeration give rise to the emergence of a 
few large, complex banking organizations with increasingly similar business portfolios 
and asset structure, thereby raising systemic risk concerns. In this paper, we also studied 
macroeconomic implications of the increasing degree of homogeneity in revenue structure 
across banks within countries. We found that countries with more heterogeneous bank 
revenue structure tend to show lower inflation volatility, although there seem to be no 
significant impacts on real GDP growth or real interest rate volatilities. 
In sum, overall findings of the present paper suggest that expanding non-interest 
income businesses is not the panacea for commercial banks in the face of increased 
competition. Business diversification may not produce desired income diversification 
effects, and it does not necessarily imply a shift toward superior return-risk frontiers. 
Furthermore, at the macroeconomic level, excessive convergence in the revenue structure 
of banks may potentially undermine macroeconomic stability. Just as diversification may 
help stabilize earnings at the firm level, diversification across banks may be desirable at 
the macroeconomic level. 

---

25 
<