Credit Quality and Credit Risk: A Rigorous Walkthrough

Abstract

This article provides an insight into the quality of credit portfolio of banks for the period from 2007–2008 to 2020–2021. The objective is to study the transition of public sector and private sector banks in India from standardised approach (SA) to internal rating-based (IRB) approach of credit risk. The overall structure of the article gives clear insight into why it is important for the banks to improve credit portfolio quality, considering default and recovery risk estimates, which in turn affect the returns of the bank. The study concludes that there is a need to improve credit quality of banks in India with a focus on small banks in public sector and large banks in private sector, as these banks are saddled with high defaults. Low rate of recovery in small banks in public sector and private sector adds to credit loss severity and high risk weights for the banks in IRB regime. However, the recent merger of the public sector banks, considering asset quality concerns, is still an unanswered question.

JEL Codes: G010, G210, G320, G330

Keywords

Non-performing asset Basel accords return on asset banks credit quality credit losses recovery rate

I. Introduction

Financial sector of a country is a powerful engine for economic growth. As a thumb rule, the financial sector has to have a growth rate of two-and-a-half times the general economy to meet the credit requirements (Chakrabarty, 2013). In India, banks, as an important fragment of financial sector, contribute 7.7 per cent to national gross domestic production (GDP). But as commonly quoted, Winning Comes with Pain, so is the case with Indian banking industry. Following Figure 1 clearly reveals the deterioration in borrower’s credit quality [presented through gross non-performing assets (GNPA) ratio of banks] over time in response to economic changes, as observed through GDP growth rate.

According to financial stability report (RBI, 2022), non-performing assets (NPA) ratio has declined to 5.9 per cent in India as of March 2022. However, it is expected to increase to 6.2 per cent and 8.3 per cent during medium and severe stress scenarios in the presence of worsening macroeconomic conditions. This brings out well in a nutshell that poor economic conditions, be it on account of sub-prime global financial crises (GFC) or the COVID-19 pandemic, normally leave a long-lasting impact on financial stability of the banking sector, in general, and an institution’s standing, in particular, as happened in Lehman Brothers in the year 2008. Way back in 1988, to effectively strengthen the regulation and to improve banks’ stability and soundness, the Basel Committee on Banking Supervision (BCBS) introduced capital charge measurement standards. First, the credit risk sought the regulatory attention through Basel I capital framework in the year 1988. Market risk and operational risk were introduced in the capital framework in the year 1996 and 1999, respectively. Basel committee issued ‘International Convergence of Capital Measurement and Capital Standards’ in the year 2004, revised and updated in 2006, known as Basel II. Basel II primarily focuses on: (a) Pillar I, minimum capital to risk-weighted assets (RWAs) of 8 per cent (9 per cent in India); (b) supervisory review and internal capital adequacy assessment process (ICAAP) and (c) pillar III, proper market disclosures and transparency of bank’s risk exposure to its stakeholders.

Further, in response to the 2008 subprime financial crisis, the Basel committee revised the guidelines and introduced Basel III in the year 2010 (revised and updated in June 2011). Basel III is the global regulatory standard on bank’s capital adequacy, stress testing and liquidity risk. To strengthen the regulation further, Basel Committee issued ‘Basel III: Post crises reforms¹’ in the year December 2017, which is considered to be more risk-sensitive in nature (BCBS, 2017). On the whole, through the various accords (Basel I, II and III) over time, the Basel Committee added risk sensitivity in assessment of risks (credit, market and operational) in the capital framework. However, the focus of this article is on approaches of credit risk capital charge estimation. In response to the Basel directives, Reserve Bank of India (RBI), in India, also issued the guidelines on ‘risk management system in banks’ in the year 1999 and came out with first ‘guidance note on credit risk’ in the year 2002. The guidance notes provide coverage on credit risk rating framework and approaches of credit risk capital charge computation, such as standardised approach (SA) and internal rating-based (IRB) approach of credit risk. These approaches are detailed in Table 1.

Table 1.

Standardised Approach vs Internal Rating-based (IRB) Approach.

Standardised Approach	Internal Rating-based Approach
Regulatory driven approach	Bank driven approach
Rating is given by RBI-approved external credit rating agencies	Rating is bank’s internal rating
Standardised risk weights are given by RBI	Risk weights are computed by the bank depending upon credit risk drivers, that is, PD, LGD, EAD and correlation, which are then inputted in IRB risk weight function (Basel Committee on Banking Supervision, 2005) for estimating capital charge using ‘k’ factor
Guidelines were given through ‘new capital adequacy framework’, revised and updated overtime	Guidelines were given on IRB approach on 22 December 2011
All scheduled commercial banks (except local area banks and regional rural banks) implemented the same from March 2009	No bank in India is reporting and disclosing their capital charge as per IRB approach. However, the banks are preparing themselves for migration to this approach

Source: Author’s own, based upon Basel and RBI guidelines issued overtime.

Note: EAD, exposure at default; LGD, loss given default; PD, probability of default; RBI, Reserve Bank of India.

These approaches demand more capital for the bank with high exposure to low-quality borrowers, whose failure to repay the loans exposes the banks to credit risk. This failure in loan repayment arises because of unexpected changes in various idiosyncratic (borrower specific: Financials and management efficiency, business, collaterals) and systematic factors (Berger & DeYoung, 1997; Memmel et al., 2012; Rajan, 2018). That is why proper identification of credit risk, starting from proper appraisal (obligor and collateral assessment) to estimation of credit risk drivers, has become essential (Basel Committee on Banking Supervision, 2006). According to Regulatory Consistency Assessment Programme—India (Basel, 2015), credit risk in total RWAs of banks in India is 86 per cent; thus, maintenance of required capital for this risk for bank’s solvency is a major concern for the lending banker today.

Given this background, the objective of this article is to study the transition of public sector and private sector banks in India from SA to IRB approach of credit risk in order to emphasise on bank-driven risk assessment over a standardised framework. In line with Basel Committee on Banking Supervision (2006), the study of correlation between probability of default (PD) and loss given default (LGD) for public sector banks (PSBs) in pre- and post-merger scenarios has been attempted further. The risk weights of domestic systemically important banks (D-SIBs) and global systemically important banks (G-SIBs) have been presented next at a given level of PD and LGD. In the end, risk-adjusted performance of the banks through risk-adjusted return on capital (RAROC) is assessed.

This article is divided into five sections. After stating the importance of asset quality, credit risk management and introduction of various approaches of credit risk capital charge computation under Basel regime in first section, Section II presents the review of existing literature. Section III addresses the data and methodology used for this study. The results emerging from the data analysis are discussed in Section IV. Finally, Section V presents the conclusion.

II. Review of Existing Literature

Credit risk is risk of non-payment by the customer (Moradi & Mokhatab Rafiei, 2019; RBI, 2002). It is assessed through credit risk drivers like PD, exposure at default (EAD), LGD and correlation (Bonfim, 2009). Many methodologies by various authors can be found in the literature on these credit risk drivers. One way to estimate these drivers is by looking at the rating data of the customers of banks or through study of NPAs, which is publicly available in annual reports of the banks (Bawa et al., 2019; Mishra et al., 2021). Latter formed the basis of this study, considering the confidentiality of customer rating data.

NPA is defined as 90 days past due (RBI, 2001²). It arises due to unexpected changes in borrower’s credit quality on account of various borrower-specific and systemic factors (Bajaj, 2018). Bonfim (2009) reported that not only the borrower’s financial difficulties are central to borrower’s default but also the prevailing macroeconomic conditions (Salas & Saurina, 2002). Das and Ghosh (2007), in their study of state-owned banks, found that GDP growth rate and bank size influence the NPAs in banks in India. However, the influence of collateral on bad loans is well documented by Gabriel et al. (2006) and Bajaj et al. (2021). As, post the default event, the rate of recovery becomes important, which is defined as a fixed percentage of outstanding the bank managed to recover. Altman (2006) and Hu and Perraudin (2002) demonstrate risk of recovery in the presence of high defaults, and the influence of macroeconomic environment on rate of recovery is well documented by Altman et al. (2004), PECDC (2013), Frye (2000), Jarrow (2000) and Carey and Gordy (2001).

Way back in 1974, Merton observed an inverse relationship between PD and recovery rate (RR) on which IRB risk weight function (Basel Committee on Banking Supervision, 2005) is based. Keisman and Marshella (2009) view that ignoring the PD and LGD correlation may lead to underestimation of portfolio risk of banks (Merton, 1974). During recession period, NPAs increase and recovery reduces, which represents high recovery risk (or LGD) during downturn (Miu & Ozdemir, 2006). Not only the increase in default and recovery risk, but also the correlated defaults in concentrated portfolios expose the banks to high unexpected losses (UL) in the downturn period (Memmel et al., 2012). The extant of UL also depends upon degree of correlation (default correlation or asset correlation) in modelling portfolio credit risk (Kealhofer & Bohn, 2001; Moody, 2009). The assets correlation considered in estimating ‘k’ factor under IRB risk weight function points towards the linkage of PD with macroeconomic environment (Bandyopadhyay & Ganguly, 2011; Memmel et al., 2012).

These credit risk parameters, detailed above, influence the credit risk capital charges under Basel regime. Goodhart and Segoviano (2004) looked at a move from SA to IRB approach of credit risk for banks in the United States, Norway and Mexico from 1982 to 2003. The study shows that the banks with a portfolio of good quality (low PD) will have capital relief in IRB regime, as it is more risk-sensitive in nature (Jacobson et al., 2006). The study by Eva and Jaroslav (2012) found that the flexibility inherited in the advanced approaches of credit risk measurement is much more than the SA, and importance of financial and physical collateral as a mitigant in IRB approach provides sufficient coverage to banks against credit risk (Eva & Gabriela, 2016). In addition, usage of own methodology (i.e., IRB) for credit risk measurement can bring substantial savings of bank equity (Eva & Gabriela, 2016; Eva & Jaroslav, 2012). Oleiwi et al. (2019), with 3 years (2015–2017) data of eight different banks in Malaysia, emphasise the importance of bank’s risk management in boosting bank’s profitability. In the similar direction, Haubenstock (1998), Baer et al. (2011) and Huang et al. (2018) looked at RAROC model for risk-based strategic planning to calculate economic profit of a bank. It is considered as a better tool than return on assets (ROA) as it is based on risk of a loan or a portfolio (Bandyopadhyay & Saha, 2007; Padganeh, 2014). The authors view it as an improvement over traditional measures like ROA and return on equity (ROE), which are used to determine the value contribution of a loan or business unit.

Studies Relating to Mergers of Banks

In order to deal with the NPAs problem and to increase global competitiveness of Indian banks, many PSBs were merged in April 2020. This consolidation was a move to improve economies of scale and operating efficiency. However, Kaur and Kaur (2010) and Tanwar (2017) opined that stronger banks should not be merged with the weaker banks, as the weaker banks will have adverse effect upon the asset quality of the stronger banks. Given this correlation between PD and RR (Basel Committee on Banking Supervision, 2006), it becomes important to assess in pre- and post-merger scenario.

Upon reviewing previously conducted studies, it is clear that no detailed study has been conducted on all credit risk drivers and RWAs estimation for banks in India, to the best of our knowledge. Following are the contributions of our article:

First, the study of transition of all public sector and private sector banks in India from SA to IRB approach of credit risk is uniqueness of study.

Second, the correlation between default risk and recovery risk of PSBs in India in pre- and post-merger period makes the study more recent.

Third, the PD, LGD and risk weight of D-SIBs and G-SIBs provide a comparative picture.

Fourth, the risk-adjusted performance of Indian banks through RAROC makes the analysis investor-centric.

On the whole, the benefit of regime change from SA to IRB approach can be observed from the findings of this study.

III. Database, Methodology and Variables Definition

In order to comment on the credit quality of banks’ portfolios, we looked at following credit risk drivers: PD, LGD and EAD, assets correlations and maturity adjustment function (MAF) for each bank for the period from 2007–2008 to 2020–2021, in order to compute the ‘k’ factor and risk weight under IRB approach. These variables are explained in detail below.

PD

It is an estimate of default risk, and it is attached to rating. In the absence of customer-specific rating data, this article provides rupee-weighted average PDs for the bank, as it is a more conservative measure than frequency-based measure of PDs (Davis et al., 2004). To compute bank’s PD, the marginal PD is estimated first by using moving average method as given in the following equation (Bandyopadhyay, 2011):

Marginal PD = \frac{NPAs addition in a year}{3 years average gross advances} .

Next, long-run average (pooled) PD (LRPD), which is through the cycle (TTC) PD, is computed by weighted average of yearly marginal PD’s. Here, we have considered 5 years (T) of marginal PD for computation of pooled PD (Basel Committee on Banking Supervision, 2006). Thus, the LRPD for 2020–2021 is average of marginal PD from the year 2016–2017 to 2020–2021. This LRPD goes into estimation of ‘conditional PD’ under IRB risk weight function (Basel Committee on Banking Supervision, 2005) to assess conditional expected losses (EL). This conditional PD is nothing but inverse of standard normal distribution, that is, G (PD), and it is computed as:

N [\frac{1}{\sqrt{1 - R}} \times G (PD) + \sqrt{\frac{R}{1 - R}} \times G (0.999)] .

Here, N = standard normal distribution, R = assets correlation (explained later in the article).

EAD

The next credit risk driver is EAD. According to Basel Committee on Banking Supervision (2006), the gross exposure (outstanding) at the time of default is EAD. According to paragraph 475 of Basel text, the PD and EAD are correlated. Same is presented through Figure 2, which shows an inverse relationship between gross NPAs and credit growth in banks in India, which clearly reflects bank’s attempt to reduce exposure at risk (EAD) in years of high NPAs (PD) (Samantaraya, 2016).

LGD

LGD is the third credit risk driver, which is fraction of EAD that will not be recovered by the bank following a default event. In the annual reports of the banks, the data on aggregate recoveries (cash recovery) are available to compute bank’s level recovery rate (RR) from the year 2009–2010, only under the head ‘movement of NPAs’. It is computed as:

Loss given default (in per cent) = 1 - recovery rate.

Recovery rate (in per cent) = \frac{{Total amount of net recovery in a year}^{3}}{3 years of gross NPA ’ s}

Further, long run pooled recovery rate (LRPRR) is computed by taking 5 years average (T) of yearly RR. This article considers bank’s specific recovery in the absence of facility-wise information (Bandyopadhyay, 2011). However, paragraph 468 of the Basel II Framework (2006) requires banks to consider downturn LGD (D-LGD) in the estimation of credit risk capital charges. The D-LGD takes into account the economic downturn, the period with high slippages and low recoveries (RMA, 2005). We made an attempt to identify the downturn period, from 2009–2010 to 2020–2021, by taking the ratio of ‘recovery rate to slippages’. The slippage ratio³ is calculated by the following formula:

Slippage= \frac{Fresh NPA’s}{Standard advances}

Standard advances = Gross advances – gross NPA’s

From our analysis, we observed the year 2017–2018⁴ as the downturn period for the computation of D-LGD. This D-LGD percentage helps a bank in estimation of EL and conditional EL [i.e., UL at a given confidence level (CL)] in the IRB risk weight function.

Credit Loss Estimation

The next step is the estimation of losses, that is, expected and UL in the bank’s credit portfolio, considering the above-mentioned credit risk drivers.

EL are known as anticipated or average loss of a bank, which the bank has lost over 5 or more years (RBI, 1999). EL rate under IRB risk weight (Basel Committee on Banking Supervision, 2005) function can be calculated as:

Expected loss (EL) rate = PD \times D-LGD

For EL, the bank creates provision.

UL are on account of unknown risks, like the GFC or the current COVID-19 pandemic, which result in severe losses for the banks worldwide. For these losses, the bank needs to maintain economic capital or credit value at risk (CVaR), which is the required capital for the safety and soundness of the banking system (Basel Committee on Banking Supervision, 2006). Here, CVaR is nothing but multiple of loss volatility over and above the average loss (Ong, 2000). This multiple, that is, CL, is given by Basel committee as 99.9 per cent and it is inputted into conditional PD in IRB risk weight function (Basel Committee on Banking Supervision, 2005). The UL rate is computed as:

Unexpected loss rate = D - LGD \times conditional PD .

Correlation

Under IRB approach, the bank considers assets correlation (R), which reflects upon the size and default experience of the banks in the credit risk capital framework (Basel Committee on Banking Supervision 2005). The committee view that (a) ‘asset correlation decreases with increase in PD’, as higher the PD, the higher the idiosyncratic risk of a borrower, and (b) ‘asset correlations increase with firm size’, as large borrower/firm default is strongly linked to status of overall economy (Basel Committee on Banking Supervision, 2005, 2006). Basel committee has given the following formula for computation of assets correlations for wholesale (corporates, banks and sovereigns) exposure:

Asset correlation (R) = 0.12 \times \frac{(1 - e^{(- 5 0 × PD)})}{(1 - e^{- 50})} + 0.24 \times (1 - \frac{(1 - e^{(- 50 × PD)})}{(1 - e^{- 50})})

IRB Formula

In the end, ‘k’ factor is estimated, which is nothing but minimum capital per unit of exposure. It is estimated based upon above-mentioned credit risk drivers and maturity of 2.5 years. The MAF considered in the IRB risk weight formula is reflection of higher default risk for loans of longer maturity, as there is more room for the borrower’s credit quality deterioration (McCoy, 2008).

K = [LGD \times N [\frac{1}{\sqrt{1-R}} \times G (PD) + \sqrt{\frac{R}{1 - R}} \times G (0.999)] - PD \times LGD] \times (\frac{1 + (M - 2.5) \times b (PD)}{1 - 1.5 \times b (PD)})

\begin{matrix} Risk weighted assets = 11.11 per cent \times EAD \times k per cent \\ (11.1 1 per cent in case of India) \end{matrix}

CAR = \frac{k}{Risk weighted assets} .

RWAs Under Basel Approaches

After computing the RWAs under IRB approach, these RWAs are compared with RWAs under SA. The values for RWAs under SA has been collected from annual reports of the banks under study. For the estimation of RWAs under IRB approach, we have considered following scenarios for EAD:

EAD = 100 per cent of gross advances (i.e., regulatory definition),

EAD = 50 per cent of gross advances (mild scenario) and

EAD = 20 per cent of gross advances (represents portfolio with good credit quality)

These scenarios would help us in understanding the impact of high exposure risk on RWAs under IRB approach.

Impact of Mergers on Bank’s Credit Quality

According to Basel Committee on Banking Supervision (2006), there is an inverse relationship between PD and rate of recovery of banks (Figure 3). This study attempts to compare these two variables for PSBs in pre- and post-merger (merged from 1 April 2020⁵) period.

About Global Scenario: Credit Quality and Credit Risk

Next, the study presents and compares the PD, LGD and risk weight (in per cent) of D-SIBs and G-SIBs. The idea is to show how PD and LGD influence the risk weight of the banks in the capital framework worldwide. The data for the G-SIBs have been collected from their annual reports.

RAROC

In the end, we present the risk-adjusted profitability of the banks through RAROC. It is computed as:

RAROC: \frac{Spread + fee - operating expenses - provision for expected losses}{Economic capital for unexpected credit losses}

IV. Analysis and Interpretation

The analysis is divided into five parts: (a) Computation of credit risk drivers, ‘k’ factor and risk weights under IRB approach; (b) comparison of RWAs under basel approaches with IRB approach; (c) study of PD and LGD correlation in among PSBs in pre- and post-merger scenario; (d) PD, LGD and risk weight of D-SIBs and G-SIBs and (e) The RAROC of public sector and private sector banks is presented in the end.

(a) Computation of credit risk drivers, ‘k’ factor and risk weights under IRB approach

The influence of following credit risk drivers on ‘k’ factor was studied first.

PD

The long-run pooled PD of the public sector and private sector banks is presented through Figure 4. It is clearly observed from Figure 4(A) that PD is highest in the case of UCO bank and Punjab National Bank (PNB). Same is observed to be the lowest in the case of Indian bank. Among the private sector banks, it is perceived from Figure 4(B) that IDBI bank have the highest PD, and the same is observed lowest in the case of Kotak Mahindra Bank (KMB).

It is clearly observed from the descriptive statistics (Table 2) and sector-wise analysis (Table 3) that banks in public sector are burdened with high non-performing loans, as default percentages are high in comparison to their private counterparts. These high PD levels demonstrate an urgent need to address NPA issue in banks in public sector (Das, 2019).

Table 2.

Descriptive Statistics of Indicators of Credit Quality of Banks in India (in per cent).

	Public Sector Banks		Private Sector Banks
	Probability of Default	Loss Given Default	Probability of Default	Loss Given Default
Minimum	3.31	85.27	1.66	31.48
Maximum	6.82	93.24	9.50	93.32
Average	5.05	89.72	3.78	74.51
Standard deviation	1.19	2.35	1.77	17.93

Table 3.

Estimates of Probability of Default (PD), Loss Given Default (LGD), k Factor and Risk Weights Under Internal Rating-based (IRB) Approach: Sector-wise and Size-wise Analysis (in per cent).

	Probability of Default	Loss Given Default	k Factor	Risk Weight
Public sector banks	5.05	89.72	25.88	287.53
Large-size banks	4.96	88.53	25.71	285.65
Mid-size banks	4.66	89.83	25.22	280.16
Small banks	5.53	90.80	26.71	296.79
Private sector banks	3.78	74.51	23.43	260.26
Large-size banks	4.54	67.94	24.53	272.54
Mid-size banks	3.59	73.33	23.22	257.95
Small banks	3.22	82.27	22.53	250.29
Old private sector banks	3.43	78.42	22.93	254.71
New private sector banks	4.17	70.21	23.98	266.36

The size-wise analysis in Table 3, indicates that large and mid-size banks in public sector have low PD of 4.96 and 4.66 per cent, respectively, as they have strong risk management policies in place and spend sufficiently on loan appraisal and monitoring. Studies by Salas and Saurina (2002), Ranjan and Dhal (2003) and Hu et al. (2004) also observed negative relationship between the bank’s size and bad loans. Same is reported by Bhardhan and Mukherjee (2016), as large banks may choose to diversify their loan portfolio. In contrast, default probabilities are high for large banks in private sector, particularly because of high NPAs in IDBI and Yes Bank.

LGD

LGD of all public sector and private sector banks in India is presented through Figure 5. It is clearly observed from Figure 5(A) that LGD is observed highest in the case of State Bank of India among public sector, as they have more exposure in big-ticket loans given to large corporate and Infrastructure⁶ projects, which have turned bad in recent years. In contrast, LGD is observed lowest in the case of PNB. On the other hand, in private sector (Figure 5(B)), Bandhan bank has the highest LGD, and same is observed lowest in the case of Yes Bank and IndusInd Bank.

RBI has suggested LGD of 65 per cent and 75 per cent for senior unsecured claim and subordinate claim, respectively, for banks in India (RBI, 2011). We observe LGD of 89.72 per cent and 74.51 per cent for public sector and private sector banks, respectively. These estimates for PSBs are slightly higher than estimates suggested by RBI for banks in India. This clearly demands improvement in recovery strategies for banks in this sector. As, LGD percentage in PSBs varies in the range of 85.27–93.24 per cent (Table 2). Sadly, on average, PSBs are able to recover only around 11 per cent of what they have lent. One reason for low recovery could be longer time spent in recovering the money because of traditional legal measures (SARFAESI and DRT)⁷ prevailing in the country. That is why rapid and timely recovery is a crucial condition for improved recovery (Inwon, 2002). In contrast, RR is around 25 per cent in the case of private sector banks in India, and the LGD percentage varies in the range of 31.48–93.32 per cent. On the whole, the analysis shows better recovery in the case of private sector banks in comparison to their counterpart PSBs because they have better recovery mechanisms in place. Another reason could be that their NPAs are low, so their recoveries are better (Basel Committee on Banking Supervision, 2006).

The size-wise analysis in Table 3 indicates that LGD has been observed low in the case of large-sized banks in public sector and private sector, as they have better recovery strategies in place for quick and timely recovery. On the whole, it is very important for banks to improve their RRs to reduce their credit portfolio losses and capital burden under IRB regime. Through the systematic country diagnostic of India, the World Bank (2018) pointed the negative implications of the failures on the part of the banks to recover bad debts, which weakens the market discipline and gives signal to debtors that loans can be defaulted with impunity.

Assets Correlation

Next, the impact of assets correlations on IRB capital charges needs a special mention, as this parameter is linked with bank’s PD and size. Figures 6(A) and (B) reveal decreasing assets correlation with increasing PD for banks in India (Basel Committee on Banking Supervision, 2005). Large banks have high assets correlation, which clearly shows that the borrowers’ default in large-sized banks, be it State Bank of India (SBI), HDFC bank, etc., strongly depends upon systematic factors (Lopez, 2004).

MAF

MAF is another important input considered in the IRB risk weight function. Figures 7(A) and (B) clearly show higher MAF for banks with lower long-run pooled PD. As the banks with lower PD have more room for down-gradation in loans of longer maturity, another reason of decline in default risk with increase in term of loan could be that longer maturity allows for better monitoring by banks (Bajaj et al., 2021; Mishra & Dhal, 2009).

‘k’ Factor and Risk Weight Percentage Under IRB Approach

Further perusal of Table 3 presents high capital requirements for PSBs in comparison to their counterpart private sector banks. From the analysis, we observed that ‘k’ factor and risk weight is around 25.88 per cent and 287.53 per cent, respectively, for PSBs. Same is observed to be 23.43 per cent and 260.26 per cent for private sector banks in India. It is clearly evident that PSBs are exposed to high credit risk because of poor credit quality of their loans (Bajaj & Krishnakumar, 2023); thus, they need to maintain more capital in comparison to their counterpart private sector banks in IRB regime.

(b) Comparison of RWAs under Basel approaches

Next, the transition of public sector and private sector banks in India from SA of credit risk to IRB approach of credit risk is presented through Figure 8. Figures 8(A) and (B) present that RWA under IRB approach would be lower than SA when only 20 per cent of the exposure of public sector and private sector banks is at risk. In contrast, RWA would be high for the bank with high exposure risk (i.e., EAD), default risk (i.e., PD) and recovery risk (i.e., LGD). These results are in line with study by Goodhart and Segoviano (2004) that shifting to IRB would provide benefit to only those banks, which are with improved portfolio quality. Thus, this analysis demands regulatory focus on PSBs in the country.

The correlation values between PD and LGD in pre- and post-merger scenarios are presented next through Table 4. In line with Basel Committee on Banking Supervision (2006), strong positive correlation between PD and LGD has been observed for PNB, Canara Bank and Union Bank of India post-merger. It is interesting to observe decline in correlation values for SBI and BoB post-merger, with their better monitoring and strong risk management to deal with NPA menace.

(d) PD, LGD and risk weight of D-SIBs and G-SIBs

Table 4.

Probability of Default (PD)–Loss Given Default (LGD) Correlation in Pre- and Post-merger Era.

	Pre-Merger	Post-Merger
State Bank of India	0.796	0.524
Punjab National Bank	0.992	0.994
Oriental Bank of Commerce	0.993
United Bank of India	0.757
Bank of Baroda	0.974	0.896
Canara Bank	0.983	0.986
Syndicate Bank	0.968	0.986
Union Bank of India	0.981	0.973
Andhra Bank	0.992
Corporation Bank	0.989
Indian Bank	0.578	0.625
Allahabad Bank	0.996	0.625

Next, we present the data of D-SIBs in India with G-SIBs to show the reflection of default risk and recovery risk estimates on risk weights of banks under IRB regime. The PD percentage of G-SIBs up to 10 per cent is presented in Table 5. It can easily be observed from Table 5 that higher the PD and LGD, higher the risk weights of the banks.

(e) RAROC of public sector and private sector banks

Table 5.

Probability of Default (PD), Loss Given Default (LGD), Risk Weights of Domestic Systemically Important Banks (D-SIBs) and Global Systemically Important Banks (G-SIBs).

	Exposure-weighted Average
G-SIBs⁸	PD	LGD	Risk Weight
JP Morgan Chase	3.65	28.48	85.98
Citigroup	7.28	32.21	87.8
HSBC	4.2	36.7	113.0
Bank of America	4.5	36.8	120.6
Bank of China	3.71	41	125
Barclays	5.2	31.4	114.3
Deutsche Bank	4.9	30.2	92.3
Goldman Sachs*	7.7	37.8	156.5
Mitsubishi UFJ FG	5.94	22.08	77.4
State Bank of India	3.39	93.24	254.7
HDFC Bank	1.93	61.98	218.0
ICICI Bank	3.99	86.82	267.3

Source: Author’s own analysis for D-SIBs and data collected from respective G-SIBs, basel III disclosures.

Note: *Between 5 and 20 per cent.

In the end, the risk-adjusted performance of the banks was assessed through RAROC. We observed low RAROC for banks with poor asset quality as reflected through Figures 9(A) and (B). Only Indian bank in public sector is performing well on risk-adjusted basis. This clearly reflects the cost of poor credit quality on the performance of PSBs in India. However, many private sector banks have decent risk-adjusted performance. HDFC bank has the highest RAROC followed by KMB.

V. Conclusion

This study makes an effort to provide insights on default risk, exposure risk, recovery risk and concentration risk estimates on credit RWAs of public sector and private sector banks in India. The idea is to highlight importance of improved credit quality on performance of banks. The study captures the transition of banks from SA to IRB approach of credit risk. The analysis shows that RWAs under IRB approach will be lower than SA only for those banks who have less exposure at risk, have good quality of loan portfolio, timely monitoring and quick recovery strategies in place. High PD in PSBs is a key prerequisite for them to improve their appraisal standards to reduce subsequent capital burden under IRB regime. Big banks like, SBI, ICICI, IDBI, etc., have more money stuck in big projects, which is resulting in high LGD, thus high ‘k’ factor and the respective risk weights in IRB regime. As severity of the loss through high LGD in banks has reflection on IRB risk weights of banks under study.

The small banks in public sector need special regulatory attention, as they have high PD and LGD, thus high UL, ‘k’ factor and risk weight. Credit risk drivers and correlation estimates presented in this article help us draw a perspective behind mergers of banks in India. Although the impact of mergers on bank’s assets quality is still an unanswered question? Still, high PD and LGD correlation for Canara Bank, PNB and Union Bank of India raise necessary concerns. RBI (2020) even has pointed towards increase in RWAs of PSBs post-mergers. The GSIBs and DSIBs statistics give us a clear lookout and a concluding statement on the risk weights and PD–LGD dependency. On the whole, the higher the PD and LGD estimates, the higher the risk weights of the banks in India and Internationally. In the end, the study concludes that lower the credit quality of loans in banks, the higher the credit risk and lower the risk-adjusted returns of banks in India.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Notes

ORCID iD

Richa Verma Bajaj

References

Altman

E. I.

(2006). Default recovery rates and LGD in credit risk modeling and practice: An updated review of the literature and empirical evidence. New York University, Stern School of Business.

Altman

, Resti

, & Sironi

(2004). Default recovery rates in credit risk modelling: A review of the literature and empirical evidence. Economic Notes , 33(2), 183–208.

Baer

, Mehta

, & Samandari

(2011). The use of economic capital in performance management for banks: A perspective (McKinsey Working Paper on Risk, Number 24, pp. 1–20). McKinsey & Company.

Bajaj

R. V.

(2018). Credit risk capital estimation under IRB approach for banks in India. Journal of Quantitative Economics , 16(2), 475–500.

Bajaj

R. V.

, & Krishnakumar

(2023). Overhang of NPAs problems in banks in India and SAARC countries: A survey based approach. Finance India , 37(2).

Bajaj

R. V.

, Sanati

, & Lodha

(2021). Impact assessment study of NPAs and rate of recovery: Are private sector banks in India better off? Global Business Review , 25(3) 724–749. https://doi.org/10.1177/0972150920980305

Bandyopadhyay

(2011). Internal assessment of credit concentration risk capital: A portfolio analysis of Indian public sector bank (MPRA Paper 28672). University Library of Munich.

Bandyopadhyay

, & Ganguly

(2011). Empirical estimation of default and asset correlation of large corporates and banks in India. Journal of Risk Finance , 14(1), 87–99.

Bandyopadhyay

, & Saha

(2007). RAROC & EVA: The new drivers of business growth in Indian banks . MPRA.

10.

Basel Committee on Banking Supervision. (2005). An explanatory note on the Basel II IRB risk weight functions .

11.

Basel Committee on Banking Supervision. (2006). Basel II: International convergence of capital measurement and capital standards: A revised framework (Comprehensive Version) .

12.

Basel. (2015). Regulatory Consistency Assessment Programme (RCAP)—assessment of Basel III risk-based capital regulations—India. https://www.bis.org/bcbs/publ/d320.pdf

13.

Bawa

J. K.

, Goyal

, Mitra

S. K.

, & Basu

(2019). An analysis of NPAs of Indian banks: Using a comprehensive framework of 31 financial ratios. IIMB Management Review , 31(1), 51–62.

14.

BCBS. (2017). Basel III—finalizing post-crises reforms .

15.

Berger

A. N.

, & DeYoung

(1997). Problem loans and cost efficiency in commercial banks. Journal of Banking and Finance , 21(6), 849–870.

16.

Bhardhan

, & Mukherjee

(2016). Bank-specific determinants of nonperforming assets of Indian banks. IEEP , 13, 483–498.

17.

Bonfim

(2009). Credit risk drivers: Evaluating the contribution of firm level information and of macroeconomic dynamics. Journal of Banking and Finance , 33, 281–299.

18.

Carey

, & Gordy

(2001). Systematic risk in recoveries on defaulted debt. In Working paper presented at the 2001 financial management association meetings, Toronto .

19.

Chakrabarty

K. C.

(2013). Two decades of credit management in Indian banks: Looking back and moving ahead . Reserve Bank of India.

20.

Das

, & Ghosh

(2007). Determinants of credit risk in Indian state-owned banks: An empirical investigation. Economic Issues , 12(2), 27–46.

21.

Das

(2019). Indian banking sector: Current status and the way forward (Speech by Mr Shaktikanta Das, Governor, Reserve Bank of India, delivered at the NIBM, 15th Annual Convocation of Post Graduate Diploma in Management, Pune, on June 8, 2019).

22.

Davis

P. O.

, & Williams

(2004). Credit risk measurement: Avoiding unintended results. Part 2: Weighting on defaults—knowing your institution’s default metrics. The RMA Journal , 82–85.

23.

Eva

, & Gabriela

(2016). Comparison of different methods of credit risk management of the commercial bank to accelerate lending activities for SME segment. European Research Studies Journal , 19(4), 17–26.

24.

Eva

, & Jaroslav

(2012). Assessment of credit risk approaches in relation with competitiveness increase of the banking sector. Journal of Competitiveness , 4(2), 69–84.

25.

Frye

(2000). Collateral damage. Risk , 13(4), 91–94.

26.

Gabriel

, Salas

, & Saurina

(2006). Determinants of collateral. Journal of Financial Economics , 81, 255–281.

27.

Goodhart

, & Segoviano

M. A.

(2004). Basel and procyclicality: A comparison of the standardised and IRB approaches to an improved credit risk method (Discussion Paper 524). Financial Markets Group, London School of Economics and Political Science.

28.

Haubenstock

M. J.

(1998). Linking RAROC to strategic planning. The Journal of Lending, and Credit Risk Management , 26–31.

29.

J. L.

, Li

, & Chiu

Y. H.

(2004). Ownership and nonperforming loans: Evidence from Taiwan’s banks. The Developing Economies , 42(3), 405–420.

30.

Y. T.

, & Perraudin

(2002). The dependence of recovery rates and defaults (Working Paper). Birkbeck College.

31.

Huang

J. B.

, Li

Z. F.

, & Ding

(2018) Research on fund performance measure with CVaR. Management Review , 30, 20–32.

32.

Inwon

(2002). Collateral in loan classification and provisioning (Working Paper WP|02|122). International Monetary Fund.

33.

Jacobson

, Lindé

, & Roszbach

(2006). Internal ratings systems, implied credit risk and the consistency of banks’ risk classification policies. Journal of Banking & Finance , 30(7), 1899–1926.

34.

Jarrow

R. A.

(2000). Estimating recovery rates and (pseudo) default probabilities implicit in debt and equity prices (Working Paper). Cornell University.

35.

Kaur

, & Kaur

(2010). Impact of mergers on the cost efficiency of Indian commercial banks. Eurasian Journal of Business and Economics , 3(5), 27–50.

36.

Kealhofer

, & Bohn

(2001). Portfolio management of default risk . Moody’s KMV Technical Document.

37.

Keisman

, & Marshella

(2009). Recoveries on defaulted debt in an era of black swans . Moody’s Global Corporate Finance, Moody’s Investments Service.

38.

Lopez

J. A.

(2004). The empirical relationship between average asset correlation, firm probability of default, and asset size. Journal of Financial Intermediation , 13(2), 265–283.

39.

McCoy

(2008). Maturity adjustments under asymptotic single risk factor models: A comparative analysis. Journal of Credit Risk , 4(3), 103–136.

40.

Memmel

, Gündüz

, & Raupach

(2012). The common drivers of default risk. Journal of Financial Stability , 16, 232–247.

41.

Merton

R. C.

(1974). On the pricing of corporate debt: The risk structure of interest rates. Journal of Finance , 2, 449–471.

42.

Mishra

A. K.

, Jain

, Abid

, & RL

(2021). Macro-economic determinants of non-performing assets in the Indian banking system: A panel data analysis. International Journal of Finance & Economics , 26(3), 3819–3834.

43.

Mishra

, & Dhal

(2009). Pro-cyclical management of banks’ non-performing loans by the Indian public sector banks (DBOD No. BP. BC. 6/21.01.002/2009-10, RBI/2009-2010/37). RBI. www.rbi.org.in

44.

Miu

, & Ozdemir

(2006). Basel requirements of downturn loss given default: Modeling and estimating probability of default and loss given default correlations. Journal of Credit Risk , 2(2), 43–68.

45.

Moody’s

KMV.

(2009). The relationship between average asset correlation and default probability. https://www.moodysanalytics.com/-/media/whitepaper/before-2011/07-22-09-the-relationship-between-average-asset-correlation-and-default-probability.pdf

46.

Moradi

, & Mokhatab Rafiei

(2019). A dynamic credit risk assessment model with data mining techniques: Evidence from Iranian banks. Financial Innovation , 5(1), 1–27.

47.

Oleiwi

A. T.

, Ali

, Jassim

S. H.

, Nadhim

M. H.

, Gueme

G. M.

, & Bujang

(2019). The relationship between credit risk management practices and profitability in Malaysian commercial banks. International Journal of Engineering and Advanced Technology (IJEAT) , 8(5C), 53–59.

48.

Ong

M. K.

(2000). Internal credit risk models—capital allocation and performance measurement . Risks Books.

49.

Padganeh

(2014). Risk-adjusted return on capital (RAROC) (Working Paper No. 1499114). GARP.

50.

PECDC. (2013). Default rate study .

51.

Rajan

R. G.

(2018). Note to parliamentary estimates committee on bank NPA (September 6th 2018, prepared at the request of the Chairman of the Parliament Estimates Committee, Dr. Murli Manohar Joshi, MP). https://www.thehindubusinessline.com/money-and-banking/article24924543.ece/binary/Raghuram%20Rajan%20Parliamentary%20note%20on%20NPAs

52.

Ranjan

, & Dhal

S. C.

(2003). Non-performing loans and terms of credit of public sector banks in India: An empirical assessment. RBI Occasional Papers , 24(3), 81–121.

53.

RBI. (1999). Guidelines on risk management systems in banks .

54.

RBI. (2001). Monetary and credit policy measure , 2001–2002.

55.

RBI. (2002). Guidance note on credit risk management .

56.

RBI. (2011). Implementation of the internal rating based (IRB) approaches for calculation of capital charge for credit risk .

57.

RBI. (2020). Financial stability report , issue no 21.

58.

RMA. (2005). Downturn LGDs for Basel II (RMA Capital Working Group).

59.

Salas

, & Saurina

(2002). Credit risk in two institutional regimes: Spanish commercial and savings banks. Journal of Financial Services Research , 22(3), 203–224.

60.

Samantaraya

(2016). Procyclical credit growth and bank NPAs in India. Economic & Political Weekly , 51(12), 112–119.

61.

Tanwar

D. N.

(2017). Performance analysis of Indian banks using camel approach. International Journal of Commerce and Management Research , 3(1), 59–67.

62.

World Bank. (2018). India—systematic country diagnostic: Realizing the promise of prosperity (English) . World Bank Group. http://documents.worldbank.org/curated/en/629571528745663168/India-Systematic-country-diagnostic-realizing-the-promise-of-prosperity