Basel I to Basel III: Impact of Credit Risk and Interest Rate Risk of Banks in India

2.1. Basel I

In 1988, the Committee refined its structure to address market risks apart from credit risk. A vital part of which was that banks were permitted to use an internal model for measuring their market risk capital prerequisite which were subject to quantitative and qualitative guidelines (BCBS, 2015a). However, this led to an increase in the notional reduction of the weight factors of assets, thus concealing the real risk exposures of credit institutions and leaving the true picture of the institution hidden behind the cloak of manipulations (Vousinas, 2015).

Summary points of Basel I:

Introduced the concept of risk weighted assets (RWA).

2.2. Basel II

The foundation of Basel II was laid down by the deficiencies of Basel I, especially with regard to the actual and applied credit risk weights. Basel II enforced three pillars, namely, capital requirements, supervisory review process and market discipline, which were closely interlinked, inter-related and mutually reinforcing in safeguarding against the operational risk of credit institutions.

Summary points:

Ratio of core capital to total RWA remained at, at least 4 per cent.

Indeed, even before the breakdown of Lehman brothers in 2008, the requirement for further fortifying Basel II became significant. An excessive amount of leverage and insufficient cushion of liquidity combined with poor governance and risk management prompted the severe impacts of worldwide recession and financial crisis (BCBS, 2015b). Also, there were conflicts of interest in the ratings of counterparty risk assessing credit rating agencies. Creative methods of estimating the credit risk and complex financial products like derivatives aggravated the situation. The capital requirement was not considered pro-cyclical. All these deficiencies set the path of Basel III accords.

2.3. Basel III

In 2010, Basel III (BCBS, 2010) was accepted and was agreed upon to be implemented in a phased manner starting from 2013 over 6 years till 2019.

The proposed rules (as graphically represented in Figure 1):

Tier 1 capital ratio to be increased from 4 per cent to 6 per cent.

OPEN IN VIEWER

Figure 1.

Source: Authors’ compilation.

The rest of the article is organised as follows: Section 3 provides an extensive survey of the literature on the implications of Basel norms, that is, gross non-performing assets (GNPAs) and net interest margins (NIM; also called credit risk and interest risk, respectively), while the empirical methodology and description of the variables used in the empirical analysis are outlined in Section 4. Section 5 provides an empirical analysis and discussion of the results is in Section 6. Section 7 presents the conclusion.

3.1. Gross Non-performing Assets (GNPAs)

The determinants of GNPAs can be classified into three segments. The first set of studies focus on macroeconomic variables (Dash & Kabra, 2010; Espinoza & Prasad, 2010), the second set focus on both macroeconomic and microeconomic variables to identify the determinants of GNPAs (Messai & Jouni, 2013). The third set of studies focus primarily on bank-specific factors (Rajaraman & Vasishtha, 2002). Since these studies are based on different countries, the legal framework, government policies and other country-specific variables are different for each country; a general conclusion cannot be reached. Barr and Siems (1994) and Hou and Dickinson (2007) stated that the issue of NPAs has received attention lately as increasing NPAs had led to banking crisis.

This is the fundamental thought process for the endeavour in this present research effort. Poghosyan and Cihak (2011) are of the opinion that bank distress could be significantly determined by the asset quality and earning profiles of banks, along with leverage.

3.2. Interest

The direct implication of Basel norms on interest is that in some cases its implementation may lead to a reduction in the credit supply provided by banks (Vousinas, 2015). When banks will have less money available to be disbursed and the demand for money is the same, the interest rates will increase due to inelastic supply (Miles, Yang, & Marcheggiano, 2013; Yan, Hall, & Turner, 2012). Amidu (2014) observed that regulatory initiative which restricts banking activities influences banks’ decisions to supply loans in sub-Saharan Africa.

Like the GNPAs, the literature on determinants of interest can likewise be characterised into those of bank-specific variables, industry-specific variables and macroeconomic elements. Some studies concentrate on one classification of elements while others consider two or all the three classifications of elements. By and large, many factors have been mentioned in the past studies on the determination of interest spreads and margins.

In 1998, Demirguc-Kunt and Huizinga in their benchmark research had analysed interest spreads of a sample of 80 nations. They found that few elements including bank characteristics, macroeconomic variables amongst others assume an essential part in the determination of interest. Gambacorta & Mistrulli (2004) considered both microeconomic and macroeconomic factors in an investigation of Italian banks. Grenade (2007) gauged the determinants of banks, interest spreads in the Eastern Caribbean Currency Union. In other studies of Chirwa and Mlachila (2004); Naceur and Goaied (2008); Entrop, Memmel, Ruprecht and Wilkens (2015) and Siddiqui (2012), among others, interest rate spreads are hypothesised to be a function of bank-specific and industry-specific variables, as well as macroeconomic factors. In any case, there is still a lack of empirical studies on the determination of interest rate spreads with respect to India in the context of Basel norms.

4.1. Data

The study of 36 entities (banks) covers the period 2002–2015 which is further segregated into 2002–2008 and 2009–2015 to account for the impact (if any) of financial crisis as also stated by Macroeconomic Assessment Group (MAG, 2010). ²

4.2. Selection of Variables

Dependent variable is GNPAs in the first model and NIM in the second model. The possible explanatory, bank-specific determinants are selected keeping in mind the frequently used ratios of RBI. The CAMELS ³ model for performance measurement is selected as it is an important supervisory tool that defines the Indian bank performance in an objective manner (Dhar & Bakshi, 2015). CAMELS has been used in the study by Prasuna (2004), Siva and Natarajan (2011) and Chaudhry and Singh (2012), among others. Subsequently, the independent variables are selected from the extensive literature reviewed.

4.3. Model 1: GNPAs

In this study, an existing well-established model as used by Demirgüç-Kunt and Huizinga (1999), Mendes and Abreu (2003) and Athanasoglou, Brissimis and Delis (2008) is used to study the relationship and impact of bank-specific variables with the GNPAs. The model is presented ahead:

G N P A i t = α + ∑ i = 1 k β k X k i t + ε i t ,

(1)

where GNPA_it = GNPA ratio, that is, ratio of GNPA to gross advance for i-th bank at t time.

α  = Constant.

β  = coefficients of determinants of GNPA ratio.

X_k = k-th regressor.

K  = Number of regressors or independent variable.

u_it  = error term

G N P A i t = α + β 1 C D R i t + β 2 S E N i t + β 3 P P E i t + β 4 C A R i t + β 5 N I M i t + β 6 R O A i t + β 7 I D R i t + u i t ,

(2)

where α = constant,

β₁ to β₉ = coefficients of determinants of GNPA ratio,

i (index of banks) = 1, 2 […] 36 and

t (time-interval) = 1, 2 […] 14.

I n t e r e s t i t = α + ∑ i = 1 k β k X k i t + ε i t ,

(3)

where Interest_it = NIM ratio, that is, ratio of net interest income of the bank divided by total assets. Other variables are same as the previous model.

N I M i t = α + β 1 C D R i t + β 2 S E N i t + β 3 P P E i t + β 4 C A R i t + β 5 N I M i t + β 6 R O A i t + β 7 I D R i t + β 8 G D P i t + u i t ,

(4)

where α = constant,

β₁ to β₉ = coefficients of determinants of GNPA ratio,

i (index of banks) = 1, 2 […] 36,

t (time-interval) =1, 2 […] 14 and u_it = error term.

5.1. Model 1: GNPAs (Credit Risk)

5.2. Model 2: NIM

6.1. Credit Risk for Public Sector Banks (PSBs) ⁴

As it can be observed from the table (in Annexure 3), the PPE and ROA conforms to the results in the findings of the literature for all the time frames. The CDR and NIM exhibit consistency with the results present in the literature only after the crisis. This is perhaps because before the crisis economy faced a boom and the return avenues were much more than the risk involved, therefore a higher CDR resulted in higher profits for the banks rather than GNPAs.

The lending to sensitive sector exhibits an impressive negative trend contradictory to the findings of the available literature. It clearly indicates that the PSBs are managing their SEN lending well and are thereby reducing their GNPAs. CAR exhibited consistency before the crisis, however, after the commencement of implementation of Basel III’s stricter capital measures the variable indicates a positive relationship with GNPAs similar to the result of the entire banking sector post-crisis. A reason for this can be that as the compliance for tighter capital is increasing the banks in order to maintain their previous profits invest in risky avenues for higher profits. IDR is observed to indicate a positive relation unlike the negative trend expected. The reason is similar to CAR.

6.2. Credit Risk for Private Sector Banks ⁵

The PPE (being consistent with the existing studies) exhibits a negative trend. The CDR and lending to sensitive sector indicates a contradiction with the findings in the literature (Jesus & Gabriel, 2006; Keeton, 1999). Both these variables were expected to be positive, however, under all time frames a negative relationship has been observed. This indicates the degree of efficiency, caution and high recovery rates at the private sector banks. A higher CAR and IDR are expected to reduce GNPAs, however, it is actually working in the reverse. The fear of less money for loan disbursement due to stricter and tighter capital norms has induced banks to invest in riskier proposals. NIM and ROA have exhibited consistency with existing studies after the crisis.

6.2.3. Interest Risk (NIM) for Public Sector Banks (PSBs)

Interest margin presents a negative relation with PPE and a positive relation with CAR, NPA and ROA as suggested by the existing studies. ⁶ CDR and SEN lending that ought to have positive relation with NIM comply only in the post-crisis phase (and contradicts in the other two time frames). This is perhaps because pre-crisis banks had excess money available to disburse as loans to sensitive sectors. IDR although consistent with the literature findings in the pre-crisis phase exhibited a positive relation; however, exhibited a negative relation in the post-crisis phase. GDP was expected to be negatively related to NIM, however, it indicated a positive relation in the pre-crisis phase. This could be due to the high return earning projects available before the crisis in the period of boom. GDP indicated a negative relation, consistent to the findings in the literature only in the post-crisis phase.

6.2.4. Interest Risk (NIM) for Private Sector Banks

CDR, PPE, SEN, CAR and ROA reinstated the existing findings in the literature. ⁷ IDR exhibited a contrasting result in the different time frames. According to the literature findings, IDR has a positive relation with interest margins. GDP exhibited a negative trend post commencement of implementation of Basel III which indicates that as capital requirements become tighter, less funds are available for disbursement which in turn increases the interest rates. The findings of Wong et al. (2010) and Gambacorta (2011) are reported to be similar.