II. Literature Review
2.2. The Relationship and the Determinants of NIM
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
economic implications are kept. As mentioned in section 1, this approach displays econometric benefits of allowing for the presence of composed errors and of being immune from random shocks. We therefore select this method to deduce one of our dependent variable, i.e., the Lerner index, which proxies for market power.
2.2. The Relationship and the Determinants of NIM
Ho and Saunders (1981) is the pioneer discussing the determinants of banks’ net interest margins (NIM). They extend and integrate the hedging and expected utility approaches to examine the relevant factors on bank’s NIM, where a bank is viewed as a dealer in the credit market, acting as an intermediary between fund demanders and suppliers. It is shown that bank margins rely on four factors, i.e., the degree of managerial risk aversion, the size of transactions undertaken by the bank, market structure, and the variance of interest rates. According to their empirical results on the U.S. commercial banks, the degree of competition within the industry and interest rate risk exposure are found to be critical variables. Afterwards, their initiative model has been extensively employed by other researchers. Allen (1988) extends the
Ho-Saunders model to factor in loan heterogeneity and cross-elasticities of demand between bank products, and propose the proposition that diversification benefits, stemmed from the interdependence of demands across bank services and products, may give rise to the reduction in pure interest spreads. McShane and Sharpe (1985) proxy the interest rate risk by statistics from money market as extension, instead of the interest rate on traditional activities.
Studies on the U.S. commercial banks are supplemented by Angbazo (1997) and Rogers and Sinkey Jr. (1999) as well. The former tests for the maturity-mismatching hypothesis,7 regulatory tax-avoidance hypothesis,8 and underinvestment rationale,9
7 Flannery and James (1984) proffers maturity-mismatching hypothesis that interest rate risk exposure is negatively associated with the average maturity of assets.
8 Regulatory tax-avoidance hypothesis suggests that as deposit financing becomes competitive and 8
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
using dataset of 286 U.S. commercial banks. The empirical results show that bank interest margins reflect default and interest rate risk premia, and risk effects demonstrate heterogeneity according to bank size. Moreover, off-balance sheet activities are found to be positively correlated with profitability for the U.S.
commercial banks, and the involvement in OBS activities raises interest rate risk and liquidity risk exposure, consistent with the moral hazard hypothesis.10 The author also introduces year dummy to discuss the time effect in accordance with the actual credit market conditions at that time. Rogers and Sinkey Jr. (1999) focus on the analysis of non-traditional activities for 11,534 observations of U.S. commercial banks. They reach the conclusion that banks with heavy engagement in
non-traditional activities tend to be larger, with lower NIMs, and accommodate fewer core deposits. Their empirical results suggest more engagement in non-traditional activities in order to reduce the idiosyncratic risk, after controlling for bank size. They claim that non-traditional activities can serve as a compensating avenue for the
shrinking profits from traditional ones and banks can also benefit from more sources of funding, market values strengthening, and operation safety.
Maudos and Solís (2009) analyze key factors on the NIM of Mexican banks by including operating cost, diversification, and specialization, the first study on this region. Market power and average operating costs are found to play critical roles on the fluctuations of net interest income, and the effect increases with time. Although the non-interest income is negatively associated with net interest income during the
costly (due to reserves), OBS activities supply avenues for banks to finance loans less costly and meanwhile get rid of the opportunity cost from holding required reserves, which is non-interest bearing.
(See Pennacchi, 1988)
9 Underinvestment rationale suggests that OBS activities benefit banks from reducing the frequency of abandoning worthy investments on uncollateralized loans, shunning wealth transfer from shareholders to debt claimers. (See James, 1988)
10 Moral hazard hypothesis postulates that banks are apt to produce higher asset risk through increasing involvement in OBS activities to raise the subsidy value of deposit insurance when the premium fails to instantly factor in the enhanced risk derived from new activities. (See Pyle, 1985)
9
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
sample period, the impact is low since the sample period, 1993-2005, is a bit far from now, when non-traditional activities are not the dominant activities in Mexico.
Nguyen (2012) explores the relationship between net interest margin and noninterest income on 28 financially liberalized countries in the context of
simultaneous equations. He considers the variable of BANKHI, defined by the ratio of each bank’s total deposits to the sum of the total deposits of the entire banking sector of a country, as the measure of market concentration and discovers no consistent benefits of diversification.
Maudos and de Guevara (2004) examines the factors explaining the interest margins in the principal European banking sectors, using a sample of 1,826 banks for the period 1993-2000. Differing from the previous relevant literatures, banks’
operating costs and direct measures of degrees of competition, such as the Lerner index and HHI, are taken into account. The outcomes show that the increase in the degree of concentration, corresponding to a less competitive market, causes upward pressure on interest margins. The authors empirically demonstrate the robustness of the Lerner index over the HHI. Specifically, the Lerner index displays a highly significantly positive relationship with the dependent variable. However, the benefit derived from the increase in market power would be counteracted by the fall in operating costs, credit risk, and interest rate risk. These three factors are significantly and positively correlated with NIM.
Carbó and Rodríguez (2007) compile a sample of 19,322 observations for European banks, which are operating under more flexible regulatory environments, to investigate the relationship between interest margins and specialization using a multi-output framework. Their results strongly support the loss-leader behavior among these European banks, which motivates banks to attract fund suppliers through higher returns and correspondingly charge relatively underestimated rate on fund
10
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
demanders.11 Under such a circumstance, NIM does not increase, and inducing banks towards providing more non-traditional activities that serve as compensation. Taking the HHI as the proxy for market power, their results confirm that market power increases as the degree of product diversification towards non-traditional activities increases. They further use the ratio of other earning assets to total assets as proxy for income diversification, and discover that more diversified banks have higher interest margins in traditional activities than specialized banks. Thus, the engagement in non-traditional activities benefits banks through increasing their market power and pricing ability. Similarly, Lepetit et al. (2008) collect the data on 602 European banks and attempt to investigate the consequence of the engagement in fee-based activities on bank margins and loan pricing. They test both the loss-leader hypothesis and credit risk underpricing hypothesis.12 In contrast to Carbo and Rodríguez (2007), they attain negative association between commissions and fees income and NIM. The evidence casts doubt on whether banks set lower lending rates, implying the underpricing of credit risk, to attract new customers in an attempt to establishing long-term customer relationship leading to the generation of fees and commissions income afterwards.
Their empirical evidence thus brings up operational risk concerns to authorities thanks to cross-selling strategies.
As for Asian financial markets, Berger, Hasan and Zhou (2010) investigate the benefits of specialization versus diversification for Chinese banks. They provide ample regressions to explore the effects of diversification from four dimensions: loans, deposits, assets, and geography. It is shown that diversification is insignificantly
associated with bank performance, but foreign ownership and conglomerate affiliation may mitigate the negative effects brought by diversification.
11 The loss-leader hypothesis asserts that banks attract new customers and establish long-term relationships by setting lower interest margins or charging lower lending rates.
12 Banks with more engagement in non-traditional activities generally display credit risk underpricing.
11
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Heffernan and Fu (2010) employ system GMM model to identify the
determinants of Chinese bank performance and obtain similar empirical results to Berger, Hasan and Zhou (2010), i.e., foreign ownership boosts bank performance, while non-traditional activities produce negative and significant influence on the NIM of Chinese banks. Lin et al. (2012) consider the dealer model, which views banks as risk-averse dealers, and identify two regimes of high and low degrees of
diversification. Using the endogenous regime switching model, they examine the determinants of interest margins and diversification effects for banks in 9 Asian countries.13 Their empirical evidence validates that banks benefit from diversifying their outputs that can reduce the sensitivity to risk factor fluctuations and the impact of idiosyncratic risk on margins.
13 The sample nations include China, India, Indonesia, Japan, the Philippines, Singapore, South Korea, Taiwan, and Thailand.
12
‧
According to Lerner (1934), the Lerner Index (LI) assesses pricing ability by measuring the mark-up of output price over marginal cost. Its formula is defined as:
𝐿𝐼𝑖𝑡 = (𝑃𝑖𝑡− 𝑀𝐶𝑖𝑡)/𝑃𝑖𝑡⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ (1) where 𝑃𝑖𝑡 denotes the output price of bank i at time t, and 𝑀𝐶𝑖𝑡 denotes the
marginal cost of its output, which is derived by taking the partial derivative of the cost function with respect to the single output. This requires estimating the translog cost function, specified as a function of a single output (Q) and three inputs (𝑊𝑘, 𝑘 = 1,2,3): of the half-normal random variable u, serving as the measure of cost inefficiency.
Variables α, β, η, γ, ω, φ, 𝜎𝑣12 , 𝑎𝑛𝑑 𝜎𝑢12 are unknown corresponding parameters to be estimated. As required by the microeconomic theory, restrictions such as symmetry and homogeneity of degree one in input prices are imposed to (2) prior to estimation.
The implied marginal cost function can be easily deduced by the following formula:
13