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

2. Determinants of Non-Interest Income Ratio at the Bank Level 
Our first objective is to identify factors that may influence the choice of non-
interest income share on the bank level. The degree of diversification towards non-interest 
income can be affected not only by bank specific factors but also by macroeconomic 
factors. Indeed, diversification efforts of individual banks may be significantly constrained 
by macroeconomic factors. The degree of competition in the banking industry and the 
incentives to expand non-traditional business activities may also change with the 
development of the underlying economy and capital markets. 

---

15 
The following regression equation is estimated to explore potential determinants 
of non-interest income share at the bank level: 
t
i
i
t
t
i
t
i
t
i
CT
YR
MCF
BSF
NIIR
,
1
,
1
,
,
'
'














(1) 
Where, NIIR denotes the non-interest income ratio of bank i at year t. BSF and MCF are 
vectors of bank specific variables and macroeconomic variables respectively. As bank 
specific factors we consider log asset size, equity capital to asset ratio, impaired loan ratio, 
ROA, loan ratio, net interest margin and cost-income ratio. As macroeconomic factors, we 
include GNI per capita in US dollars (under the Atlas method), real GDP growth rate, real 
interest rate, inflation rate and stock market capitalization relative to nominal GDP. YR 
and CT are year and country dummy variables respectively. Above regressions are 
estimated using the pooled OLS method with one year lagged explanatory variables to 
avoid potential endogeneity problems. 
Regression results are summarized in Table 2. The first column shows the list of 
explanatory variables. The second column reports estimation results for the regression 
specification including bank specific factors only, and the third column reports results on 
the specification including both bank specific and macroeconomic variables. Consider 
first the bank specific factors. Note that relatively large, well-capitalized, and more 
profitable banks exhibit a higher non-interest income ratio. Note also that banks with 
relatively low net interest margins and low loan to asset ratios tend to show a higher non-

---

16 
interest income ratio. These findings are consistent with US experience reported in Rogers 
and Sinkey (1999), De Young and Hunter (2003), and De Young et al. (2004). 
It is also noteworthy that risk taking banks with high impaired loan ratios and less 
cost-efficient banks with high cost-income ratios tend to diversify their revenue sources 
more aggressively by increasing their non-interest income shares. It is plausible that banks 
need to invest in more staff and technology, thus incur higher costs to conduct non-interest 
income businesses. Note that Davis and Tuori (2002) also report a similar result for 
European banks. 
As for macroeconomic factors, the economy size as measured by per capita GNI is 
shown to be insignificant, while fast growing countries with a high GDP growth rate 
exhibit a lower non-interest income ratio. This result indicates that banks tend to diversify 
toward non-interest income as the economy slows down in growth and competition in loan 
market becomes more intense. Note also that the coefficient of the inflation rate is 
significantly negative and the coefficient of stock market capitalization to GDP ratio is 
significantly positive, which suggests that a low inflationary environment and a high level 
capital market development facilitate non-interest income expansion of commercial banks. 
The evidence in this section also implies that the efforts of individual banks and the 
government policies directed toward bank revenue diversification may not materialize if 
not supported by an adequate macroeconomic environment. 

---

17