Internal ratings based (IRB) approach was created with the expectations that it will accurately align capital requirements with the intrinsic amount of credit risk and, consequently, will produce lower overall capital requirements. IRB is also complex to understand and implement, but it is undoubtedly more sensitive and effective than previous standardized approach. Thus it encourages banks to improve their internal risk management.
1. Motivation to banks for moving to IRB approach 1. As the picture below summarizes, by moving to IRB approach, banks will be able to quantify risk accurately and hence they may require only low capital charge. For example, capital requirements should drop substantially at a bank with a prime business portfolio that is well collateralized. On the other hand, a bank with a high-risk portfolio will likely face higher capital requirements and, consequently, limits on its business potential. Thus, over time, it presents banks with the opportunity to gain competitive advantage by allocating capital to those processes, segments, and markets that demonstrate a strong risk/return ratio
2. Bank will also be able to do risk based pricing which means borrowers whose risk is high will be charged a high interest rate, thus using pricing they will be able to compensate the capital required for meeting unexpected losses. 1. IRB Components - Expected & Unexpected Losses The concept of expected and unexpected losses plays an important role in the economic foundation of the IRB. Whereas a bank cannot predict in advance what losses it will suffer over a given period, it can forecast the average level of credit losses. Expected losses (EL) are those within the average level of reasonably foreseeable credit losses. Unexpected losses (UL) relate to potential volatility in the expected losses. These expected and unexpected losses are mainly calculated in order to cover them by capital provisioning and write-offs which are regarded as cost of a banking business.
Haven’t found the relevant content? Hire a subject expert to help you with CreditMetrics IRB Model
1. JP Morgan’s CreditMetrics IRB Model CreditMetrics is the popular methodology developed by J.P. Morgan and it utilizes historical default rates to model default probabilities using a credit migration or a transition matrix. Transition matrices measure the probability of a change in a given credit rating for each debt security over the credit horizon. Consider the example of a five-year fixed-rate BBB rated loan of $100 million made at 6 % interest. Based on historical data on publicly-traded bonds, the probability that a BBB borrower will stay at BBB over the next year is estimated at 86.93 percent. There is also some probability that the borrower will be upgraded (e.g., to A) or will be downgraded (e.g., to CCC or even to default, D). Table below shows the possible probabilities of transition.
The effect of rating upgrades and downgrades is to impact the required credit risk spreads or premiums on the loan's remaining cash flows. BBB rated Loan’s current price is given by P=6+ 6(1+r1+s1)+6(1+r2+s2)2+6(1+r3+s3)3+ 106(1+r4+s4)4 Where ri are the forward risk-free rates on BBB rated Zero-coupon treasury bonds and si is the annual credit spread on (zero coupon) loans of a BBB rated class. Thus using the above formula and appropriate risk adjusted rate, Loan cash flows are discounted to obtain the respective current loan values.
The distribution of loan values on the one year credit horizon date can be drawn using the transition probabilities and the loan valuations. The mean of the value distribution shown is $107.09 million. If the loan had retained its BBB rating, then the loan’s value would have been $107.55 million at the end of the credit horizon. Thus, the expected losses on this loan are $460,000 (=$107.55 minus $107.09million).
However, at 99% confidence level (1% chance that loan values will be lower than $100.12), unexpected loss is found to be $6.97 which is also nothing but the economic capital. Similarly, portfolio loss distributions are then obtained by calculating individual asset values for each possible joint migration scenario. 1. Data Requirements – IRB approach One thing which is common in all IRB models is robust information system to store and analyze long-term time series databases on the credit performance. Hence a bank looking to move forward with IRB approach has to have an integrated risk management system which aid the banks in better data collection, support high quality data and provide scope for detailed technical analysis. Hence, banks aiming at maintaining lower capital by adopting the advanced approaches would also have to be prepared to meet the higher information needs.
1. Credit Risk Management (CRM) System – Data Flow Diagram Integrated Credit Risk Management (CRM) System – DFD diagram Hence, based on the understanding and analysis so far, one point that can be stressed on for all the Indian banks is to importance on developing a completely integrated credit risk management system. Above suggested data flow diagram is one such model. As the data gets collected from different sources – Customers, Rating agencies External data sources, the centralized data warehousing system amasses the data for further processing through different CRM modules. Finally the required measures can be presented through the dashboards for the risk management group to monitor and control the risk.
For the design and development of data base system, dimensional modeling is used because of its obvious benefits like easier interpretability and understandability, efficient performance of complex queries and extensibility to accommodate unexpected new data. 1. Data understanding and Collection As discussed in the previous stages, we required Data from the Corporate (Commercial) Banking sector of a bank as an input for our DW model. Along with this, as a second input to the model, we also required rating data from an external rating agency (CRISIL, CARE etc.).
For this purpose we collected data from a reputed nationalised bank of India and scaled it down, in dimensions, to include only the relevant columns and in size, to include only a limited no. of rows to enable a satisfactory processing speed. Banking Data For the purpose of designing and implementing this model, we selected a list of 10 big to moderate size customers of the bank with varying degree of exposure (in loan size), varying ownerships (combination of pvt. and public co.s) and varying degree of risks.
After this step, we included a range of loan products with various features (secure/unsecure) to which these customers were exposed. The data contained several properties of loan products as Product codes, Description, Product wise limits, Currency, Collateral Required, Interest Rate etc. which were relevant to our model designed. We also collected the industry codes and descriptions to be used in the model.
We also collected details for more than 20 loan accounts opened by the selected list of customers under the above mentioned industries and from the selected range of products. We captured the following details for these accounts: Account No., Product Code, Balance outstanding, Term of the loan etc. Rating Data In rating data, we had a table of ratings which were awarded by the agencies and the probabilities associated with the ratings (as explained in the previous section). Based on the rating data, we created a rating fact table which contained derived multipliers to be used as an input for the model. These multipliers were used to calculate the expected and unexpected portion of the loss in the risk assessment model.
Note: The data that we collected was modified (names and codes changed) and scaled down to maintain confidentiality. 1. Dimension modeling 1. Business process Business process or measurement event that is captured for modeling is each of the new transactions that occur in the branch offices with new or existing clients. Also a separate start schema is created to store the information about the rating. It holds measures like risk free rate, default spread for different ratings. It is basically generated or updated yearly based on historical data captured form external data source.
1. Granularity Granularity of this dimension design is the account number for each type of product and for each of the customers. Assumption here is each of the customers can hold multiple products and each of the products will be recorded and tracked in unique account number. 1. Dimensions Dimensions are designed such that they take on a unique value at the declared grain of the fact table. Dimensions used are Product, Account, Customer, Industry and Date.
1. Facts – Measures 1. Extraction, Transformation, and Loading Extraction, Transformation, and Loading (ETL) processes are responsible for the operations taking place in the back stage of data warehouse architecture. In a high level description of an ETL process, first, the data are extracted from the source data stores that can be On-Line Transaction Processing (OLTP) or legacy systems, files under any format, web pages, various kinds of documents (e.g., spreadsheets and text documents) or even data coming in a streaming fashion.
Typically, only the data that are different from the previous execution of an ETL process (newly inserted, updated, and deleted information) should be extracted from the sources. After this phase, the extracted data are propagated to a special-purpose area of the warehouse, called the Data Staging Area (DSA), where their transformation, homogenization, and cleansing take place. The most frequently used transformations include filters and checks to ensure that the data propagated to the warehouse respect business rules and integrity constraint, as well as schema transformations that ensure that data fit the target data warehouse schema.
Finally, the data are loaded to the central data warehouse (DW) and all its counterparts (e.g., data marts and views). In a traditional data warehouse setting, the ETL process periodically refreshes the data warehouse during idle or low-load, periods of its operation (e.g., every night) and has a specific time-window to complete. Nowadays, business necessities and demands require near real-time data warehouse refreshment and significant attention is drawn to this kind of technological advancement.
Haven’t found the relevant content? Hire a subject expert to help you with CreditMetrics IRB Model