Credit risk refers to the risk of default on part of the borrower if he or she fails to make the required and scheduled payment on a specific debt. Credit risk arises from the uncertainty regarding whether the financial instrument cardholders or borrowers could fulfil the required obligations. In the recent past, there has been a massive global financial crisis which created turmoil in the banking industry. With an immediate effect, it had engulfed the banking sector and had affected the entire global financial stability (O’Riordan, 2014).The global financial crisis begun in mid-2007 with the credit crunch, it was this period of time when a loss of confidence by US investors in the value of sub-prime mortgages caused a liquidity crisis. Eventually, the US Federal Bank had to insert a large amount of capital into the existing financial markets.
By September 2008, the crisis had deteriorated as stock markets around the globe crashed and became volatile. Soon the whole world was doomed. As Keating. D (2008) stated, “When the US sneezes, the world catches a cold”. On 6th Feb 2013, the chairman of FSA Adair Tuner delivered a speech regarding the crisis ” The financial crisis of 2007 to 2008 occurred because we failed to constrain the financial system based creation of private credit and money”. In the contemporary time, the increasing rate of global financial crisis and the credit crunch put the credit risk management strategy into the regulatory spotlight. After the crisis, a substantial amount of research has been devoted to the field of credit risk evaluation and credit risk management. “Improving risk management tools and practices became the top priority for a financial institution”, Yurdakul, F (2014).
The Dodd-Frank law was promulgated in the US during 2010 with significant changes to financial regulation. It also created a credit rating at the securities and exchange committee (SCE) to regulate agencies like Moody’s and Standard Poor’s.
According to the Basel Committee on Banking Supervision (2001), the most common cause for the banks to suffer insolvency and bankruptcy is the credit risk. Credit risk management is considered as a practice, which has to be performed by any financial institution with the aim to mitigate the acquired losses by taking the adequacy and availability of the capital and financial resources into account for a given period.
There are credit risk models in support of this. It can broadly be classified into two models: proprietary or internal models and the vendor marketed models. These models have general application nature and are very easy for the application of the understanding regarding the concepts of risk management (Altman, 2006).
Credit Risk Measurement and Modelling
Financial Institutes have gradually developed in recent years from the subjective system (Credit scoring) and, are now more focused on the objective-based system(Structural and reduced form) to credit risk models.
Credit modelling tools such as credit scoring model is used to evaluate and analyse the creditworthiness of an enterprise. Altman (1968) first proposed a scoring model to calculate a firm’s default probability by using accounting based variables. Today, in the market, there are over 50 generic credit scoring systems that contain over 100 different credit scoring models Hanic. A et al (2013). However, Xuan Z (2017) argues that the structural model provides an alternative and potential superior source of information compared to that of the credit scoring system. He had encountered several problems when modelling probability of default based on credit scoring models like the historical model. Data used in such credit scoring model didn’t necessarily contain enough information to estimate future performance. Hillegeist S et al (2004) have a similar point of view as their result on assessing the probability of bankruptcy demonstrate that the market- based Black-Scholes model (BMS-Prob) provides significantly more information about the probability of bankruptcy than traditional accounting- based measures.
Traditional approaches to credit risk measurement have an aim to define the Probability of Default (PD) based on historical accounting data. However, the traditional approaches pay no attention to downgrades and upgrades in counter party creditworthiness which is a comparative consideration of market models and completely ignoring the company equity structure, information on the equity market. There are three traditional models used to estimate PD: (1) expert systems, including artificial neural networks; (2) rating systems and (3) credit scoring models.
The 4 “Cs” is considered as an important framework associated with the expert system and subjective analysis of credit risk measurement. 4 “Cs” model indicates to the fact that, at the time of measurement and assessment process, bankers are used to procuring information from the borrowers based on their characteristics features, such as, the character of the borrower (reputation), capacity (volatile of earning), collateral and capital (leverage). The beneficial aspect, associated with the 4 “Cs” is that, by analysing that information of the borrowers, bankers can easily make to a subjective judgment of whether to grant credit or not, when a borrower demand for the corporate loan from any financial institute.
Altman and Saunders (1998) presented a traditional overview of credit risk measurement. They summarised the key development on credit risk measurement over its past 20 years. Similarly, Carling et al. (2010) explained the recent development of credit risk modelling in the past 10 years. When reviewing these two articles, we observe some important changes and development in credit risk measurement and modelling. Carling K et, al prove the factor models have increased significantly in the number of applied studies. As a suggestion of internal rating based (IRB) approach in Basel II factor model.
Credit risk modelling with respect to large corporations, small and medium-sized enterprises, retail borrowers, sovereign and sub-sovereign
Credit Risk Modelling is aimed at building of better understanding of credit risk of the banking systems. External factors generally do not react fast enough to the changes made in the economic environment of the financial institution. So the banks are recommended to use their internal rating system. There are few statistical tools which can be used as well.
Altman (1968) analysis corporate bankruptcy prediction models which depend on accounting based measure as variables. Recent accounting- based models use a composite measure that statistically combines several accounting variables such as Altman (1968) Z-score. The Z-score model is widely used in corporate borrowers indicate the probability of a company entering bankruptcy within next 2 years. The higher score, the lower the probability of firm bankruptcy. However, the historical data does not hold sufficient information to estimate the risk. Due to this, there has been some development of new models , e.g. RORAC model.
Small and Medium-sized enterprises:
According to OECD(2017), SMEs plays a key role in national economies around the world, generating employment, income and drives of innovation and growth. During the global financial crisis 2008-09, the bank credit constraints experienced by SMEs in many countries have further highlighted the vulnerability of the SME sector to the changing conditions in lending at the banking sector. When seeking bank credit, SMEs continue to face problems financing condition and higher interest rates compared to large businesses.
Every bank has their own model in measuring the company’s performance, even if they possessed the best model, there’s a possibility of default occurring. For example Altman and Sabato, (2007), have developed a model using a complete set of financial ratios considering profitability, leverage and liquidity ratio to find out the company with prudent creditworthiness. These categories of the ratio are enough in predicting the likelihood of SME experiencing financial distress.
Several researchers have applied structural model approach, which was pioneered by Merton (1974), e.g. (KMV, Risk Metrics). The model uses the modern principle of option pricing in corporate debt valuation (Black and Scholes, 1973). Under the structural models, a default occurs when its assets reach a sufficiently low level compared to its liabilities. Since the work of Black, Scholes and Merton was introduced, many other researchers have proposed extensions to the original Merton framework by removing one or more of the unrealistic assumptions. Since the 1970s, many structural models have been developed to forecast the credit risk. The structural model approach was further implemented in Moody’s KMV model. This is also called KMV-Merton model. Altman.E et al (2004) explained the three extending models of structure in their study: Black and Cox (1976), Geske (1977) and Vasicek (1984). Black and Cox (1976) introduced the possibility of more complex capital structures with subordinated debt; Geske (1977) proposes interest payment debt; Vasicek (1984) introduce the difference between short and long-term liabilities, which now represents a unique feature of the KMV (Kealhofer, McQuown and Vasicek) model. The extended structural models represent important improvements for Merton’s original framework as they are more realistic and able to better align with that of the market data (e.g., CDS spreads) Yu Wang (2009). While Bharath et al. (2004) assess the accuracy and contributions of KMV-Merton model it was found the KMV-Merton model does not generate appropriate statistic for the probability of default and there’s a difficulty of implementation for its derivative.
Chen. X et al. (2010) conducted a study to monitor the measurement of the credit risk of SMEs in China. However, it was a challenge to measure SMEs using the modern approaches like J.P Morgan’s Credit Metrics, McKinsey’s Credit Portfolio View and Credit Risk+ model due to insufficient financial data of the enterprise. They preferred KMV model to be most suitable for SMEs companies because it didn’t rely on the historical probability credit rate but mostly depends on the enterprise’s asset.
Luppi’s et al. (2007) analysed logit model to model non- profit SMEs and found that traditional accounting-based credit scoring model held less explanatory power in non-profit firms than that in for firms. Also, Diestsh and Petey (2002) developed a credit risk model which focused on one-factor probit model to analyse the probability of default for SMEs. Their study proved that probability of default has a positive correlation with the company’s asset.
Altman and Sabato (2007) apply a logit on SMEs bankruptcies in the USA. They claim that practitioners should differentiate between corporate and SME modelling credit default as their SME specific model could outperform the corporate model on SME bankruptcies.
Credit cards came into existence in 1960s, giving consumers full hand access in purchasing from hair pins to hardware tools. Consumer credit is and extensive monetary term. It has an impact on consumers who previously did not accept the ramifications of utilisation of such a term. It helps in taking car loans, house loans etc.
The importance of the credit risk of a retail borrower was increased by the changes in banking regulations incorporated in the New Basel II Capital Accord. This accord allows banks to apply the internal operations based model to help determine how much capital to put aside against any possible losses in their loan portfolio. For the retail portfolio, credit scoring
system could thus be envisaged to be one of such internal based models.
Banks cannot estimate small retail loans by using the models that estimate for the large or SME loans. Retail credit facilities loans to individual or first-time borrowers. Over the decades, consumer lending has become increasingly sophisticated and sensitive, as lenders have moved from a traditional interview-based decision-making to data-driven models to quantify credit risk. The trend in retail credit decision making is strongly toward increased reliance on statistical data-based models of credit risk measurement.
The advanced IRB approach of Basel II accord requires banks to estimate the probability of default (PD), loss given default (LGD) and exposure at default (EAD) for their retail lending Tripe D (2005). These parameters are related to the corporate models of credit risk. Due to the differences in repayment behaviour, these parameters cannot be used directly concerning the retail borrower portfolios.
When lending to an individual, banks need to take some risk factors like loss of income, jobs etc. under consideration (Thomas, 2003). Traditional models like neutral network, internal rating systems and credit scoring models are used to estimate the probability of defaults. Through Hasan K (2016), it is empirically confirmed that credit scoring system is the most appropriate for estimating a probability of default on retail loans applicants. The idea of employing credit scoring was first introduced by (Narain, 1992) and then developed further by (Thomas et al. 1999 and 2002).
Credit scoring models have greatly facilitated over the recent years. The credit agencies take different key accounting ratios of the borrowers and compare it with the market. Zhang (2016) has identified that JP Morgan’s credit metrics as the most exceptional model since 1990s. This procedure is carried out using the credit rating while estimating the expected probability of default of at least 5 years of data and migration of the credit. However, this model is not suitable for analysing the future performance as it uses the historical data of a company. For example, a company which is newly developed in the market will not have enough data to estimate the score. After the introduction of standards “Basel II”, credit scoring models have become mandatory for the prediction of bankruptcy debtors become widespread. Simplified models are: Moody’s RiskCalc, KMV approach and Altman’s ZETA credit scoring.
Tam K (1991) employs a neural network approach to evaluate the credit risk of individual loans prediction; however, he argues that the results he obtained were difficult to interpret and lacks transparency. Handzic et al. (2003) support his argument by stating that the parameters used in this model are not fixed and takes long to find perfect parameters of any neural network model. Nowadays, banks have moved away from the neutral network-based model.
Almost all Fls use credit scoring model to estimate the PD of consumer loans. Fair and Isaac corporation developed credit score model like FICO score in the late 1950s. Since then there have been several redevelopments of the model, e.g., FICO Score XD 2. FICO model has become widely used model to measure individual’s creditworthiness. Credit bureaus like Equifax and Experian apply this model to assess mortgage score, personal loans and many more.
Sovereign and sub-sovereign:
Sovereign risk comes in many forms. However, anyone who faces sovereign risk is exposed to a foreign country in some way. In that case, foreign exchange traders and investors face the risk that a foreign central bank will change its monetary policy could affect the currency trades for the traders as a whole. This is not always beneficial for foreign traders. Sovereign risk also has a great impact on personal investors. It is obvious that there is always a risk to owning a financial security if the issuer resides in a foreign country.
Future of Credit risk modelling
Financial firms are under pressure to expand their business while containing risk and obeying with worldwide regulations. Though banks have spent a substantial amount of time to meet strict regulatory requirements, there are still many defaults in the current risk modelling methods. They are always on a lookout for new methods of measuring the risk assessment of credits. These can be done by including rapid pace of new product innovation, diversification of financial institutions, and a step up the rate of credit intermediations. Banks are more inclined towards developing their own as this will help the banks to understand the portfolio of customers in a better manner and with the help of these, the DP and LDP can reduce drastically.
It is relevant to say that technological progress in various fields of marketing where large data can be computed and analysed in a wink has given the banks greater scope to apply a different variety of credit models to their management of credit risk. Technology motivates the bank to develop new models every day. Technology is getting cheaper day by day as well as it is initiating large amount of work in less time. Banks are getting dependent on bulk technologies for managing their risk management models (Chance and Brooks, 2015).
Transparency: Nowadays, the bank has also introduced QR code and special security pins to its transactions to give its clients transparency. It also keeps them informed of the status of every request and transactions made.
Risk identification: Banks have also progressed from the traditional credit application process to digitization of the entire method. These are supported by portfolio experiences such as scores built on customer profiles, a behavioral score of the customer and other factors like risk appetite to control risk (Bielecki and Rutkowski, 2013).
Data analytics: This helps the bank having a clear record of each and every SME account, their loan details, their corporate details etc. This helps them to identify clients who are suffering from the situations where they are required to undertake repayments.
According to a recent press release (11 April 2018) on global banking & Financial review, the modern analytics “Cloudera” has announced a partnership with Simudyne, a leading simulation software company. They jointly offer banks various advanced simulations solutions for risk management. In the article the chief executive officer at Cloudera, Tom Reilly mentioned, “With Cloudera Enterprise and our Shared Data Experience (SDX), we provide the identity management, security capabilities, policy enforcement and data governance that banks need for regulatory compliance”. Combined with Simudyne, this modern approach to gaining insights and understanding risk helps financial institutions make better predictions and business decisions as economic scenarios emerge. With Cloudera’s modern platform, financial firms can manage credit risk, enable scalable stress testing tools, drive better customer insight and promote economic stability.
In the press, Jes Staley chief executive officer at Barclays Bank says “Simudyne is ground breaking technology currently being leveraged across Barclays and it enables us to model multiple scenarios on substantial data sets so that we can understand our risks, exposures and options”.
An invention of frontier using cognitive technology will protect banks to strengthen credit risk management. IBM is in process to fully explore the many possibilities for deployment to enhance banks’ credit risk management strategies
In the past few decades, the importance of a healthy banking system has increased in larger rate. After the financial crisis, we have seen significant improvements on credit risk modelling. Today, banks prefer using various modern models instead of traditional banking system. The above discussion provides an overview associated with the credit risk practice which is covered in the operations along with the action of governance of every organisation. However, there is no ‘one size fits all’ way of implementing credit risk scoring tools and techniques. Instead of doing this, the process must be well suited and furnished to fit the size, complexity, market value, industry competition and environmental uncertainty which are generally faced by the banks while crediting.