This is the 8th article in the blog series on Digital Transformation and Business Strategy for Financial Institutions, written by the Q-Lana Team. We have conducted our observations and analyses of prevalent trends, integrated with our experience, to shape the series’ structure and content. Our approach is grounded in common sense and a steadfast belief in evolutionary development. We kick off the series with insights into effectively navigating the realm of Digital Transformation. From there, we delve into compelling ideas for crafting a dynamic Business Strategy, tailored specifically for financial institutions. While our work so far focused on financial markets in lower and middle-income countries, it’s worth noting that the overarching concepts we present are universally applicable across all financial markets.
This article provides the background and details the factors used to quantify credit risk. We look at the Risk Adjusted Return on Capital (RAROC) which integrates these factors into a powerful formula that will allow us to develop and quantify the concepts of Risk Appetite and Relationship Pricing.
In this article, we take an initial look at Risk Management. The subject is broad and can fill its own blog series. Here we look at components that allow a financial institution to benefit from its skills and experience for the broader Business Strategy.
When implementing our customer-centric and risk-aware business strategy, we must consider concepts that facilitate its realization. We’ll delve into two crucial concepts: defining a specific Risk Appetite and developing a Relationship Pricing approach. Both rely on calculating the Risk-Adjusted Return on Capital (RAROC), which we’ll introduce first. These concepts streamline the approval and disbursement processes significantly. For instance, by defining a risk appetite for individual clients, we can preapprove risk exposures before reaching the decision authorities. Relationship pricing enhances transparency regarding the economic aspects of a relationship, empowering us to respond flexibly to customer demands and competitive offers aiming to attract our established client base.
Let’s begin with the fundamental approach to quantifying credit risk, utilizing three key variables: Probability of Default (PD), Loss Given Default (LGD), and Exposure at Default (EAD).
Probability of Default (PD): PD signifies the likelihood that a counterparty will default, typically within a one-year timeframe, expressed as a percentage. For instance, a 5% PD suggests a 5% chance that a specific loan will default within the next 12 months. In the context of a portfolio with 20 similar loans, each carrying a 5% PD, we anticipate one of those five loans to default within the next 12 months. The derivation of this 5% PD involves rating models introduced in a prior chapter. The specific methodology for rating calibration will be explained in a separate training video.
Loss Given Default (LGD): LGD quantifies the amount recovered by the lender from selling the underlying asset following a default. It essentially measures the bank’s potential loss in the event of a loan default. The assessment of LGD is influenced by factors such as the availability of collateral for a transaction. A more detailed discussion on evaluating LGD will be provided in a separate training session. In essence, it relies on the financial institution’s historical experience with loan defaults.
Exposure at Default (EAD): EAD refers to the expected outstanding loan amount at the time of default. This assessment considers the timing of loan defaults, meaning when during the loan’s life the default occurs. EAD also depends on the nature of the product. For example, a credit line is likely to be fully utilized at the time of default, while a term loan may have already seen some amortization before default.
These three factors enable us to calculate the Expected Loss (EL) of a loan exposure. As the word says, this is the amount that is expected to be lost. It can be understood as the cost of doing business. It is comparable to an insurance premium which in an ideal situation equals the actual losses a financial institution experiences over a period of time. The expected loss can be used in pricing as it is the charge a client needs to pay for the coverage of credit risk.
This calculation supports our pricing decisions and helps determine the necessary equity capital to support the loan exposure. The expression to remember here is the Unexpected Loss. This relationship opens the door to the RAROC many more calculations.
The expected loss is also important in the modeling of credit risk in which the information is used to derive the unexpected loss which is a basis for the calculation of the required equity coverage of credit risk exposure. Before we get to the unexpected loss, however, let’s have a look at a sample portfolio and conduct some calculations:
For our calculations let’s assume we have a bank with a portfolio of 100 similar loans:
- All these loans are linear amortizing four-year loans with an average loan amount of $10,000. This means, that these loans reduce their outstanding amount in a linear manner. For example, for simplicity, we assume that one year after the disbursement of the loan, 75% of the loan amount is still outstanding, after three years, 25% of the loan amount is outstanding.
- Based on the assessment of the financial institution, it is likely that loans default halfway through their life, with an estimated remaining Exposure at Default of 50%.
- We’ve also made an assessment of the value of the collateral and derived that this is on average $3,000. This means that considering an exposure of $5,000 and the collateral value of $3,000, in the case of a default, an average loan loses about $2,000 or 40% of its outstanding amount.
Let’s use this data for some exercises and let’s answer the following six questions:
- What is the expected loss of an individual loan? Here we look at the loan amount of $10,000 and consider the Probability of Default, the Loss Given Default, and the Exposure at Default. The calculation is $10,000 times 5% for the PD times 50% for the EAD and times 40% for the LGD equaling $100. We expect that we lose $100 for each loan.
- What is the loss if one loan actually defaults? The difference between the first and the second question is that we are no longer looking at the probability of a default but at the actual default. If a loan of $10,000 of the original amount is defaulting, there is a $5,000 or 50% outstanding loan amount and the recovery is $3,000. So our loss, the LGD, is 40%, therefore, the actual loss is $2,000.
- What is the expected loss of the total portfolio? The Expected Loss of an individual loan was calculated at $100 and we have a total of 100 loans. Therefore, the Expected Loss of the total portfolio is $10,000. Here, let’s differentiate between the Expected Loss and the realized loss. While we calculate an expected loss of $100 for each loan, this number is most likely wrong in all cases. Most loans will repay in which case the actual loss is zero. Few loans will default, in which case the actual loss is $2000. Across the entire portfolio our expected loss is correct, broken down to individual cases, the expected loss is wrong in all cases
- How many loans are expected to default and what is the actual loss in that scenario? We are expecting a 5% default ratio and we have 100 loans in our portfolio this means, that we are expecting on average five loans per annum to default. How much are we losing in this case? The loss in each of these cases is $2,000 times five equals $10,000. There is no surprise, that the actual loss in case of five loan defaults is equal to the expected loss which we calculated in the questions before.
- What is the actual loss, if 4% of the loans default? We only have an expectation about losses. The Probability of Default was supposed to be 5%. That doesn’t mean that every year five loans default. We will experience years with fewer defaults and years in which there may be more than five defaulted loans. If indeed we experience a 4% default ratio, we will have an actual loss of 4 × $2,000 or $8,000. If the institution has done its calculation correctly, it will have provisioned $10,000. This results in a profit for the institution of $2,000. Please be aware that this example does not look at specific accounting rules such IFRS 9, but is just at the economic aspects. We have however the additional question, of whether those $2,000 really are a profit or should be considered as a reserve. In reality, bankers don’t look at the statistics, they consider themselves geniuses and think that they deserve a $ 2,000 bonus for their great work.
- What is the actual loss if 6% of the loans default? In the case of six loans defaulting, the actual loss is $12,000 or $2,000 more than our provisions. How are we covering this loss, especially since in the year before we distributed the excess reserves as a bonus? The only source we can use to cover those additional unexpected losses is equity.
This gets us right to the calculation of the Unexpected Loss. So far, we have only looked at one year. If we simulate let’s say 1000 years for example through a Monte Carlo simulation and assess how many loans in each of these years default, we will find a distribution similar to the one on the screen. In the most likely scenario, we will lose 5%, but there will be individual years in which we lose less or, in many cases, more than the 5% assumed. The distribution is skewed due to the limited upside but the significant downside. Meaning, that the best-case that could happen to us is that no loan defaults, five less than expected. But the downside, the worst-case could be that up to 100 loans default. In the distribution, seen on the screen, we experienced two years in which 13 loans defaulted.
Aggregating this expectation to a model we can see that the unexpected loss is calculated as the standard deviation from the mean of the defaults, using a specific confidence level. If the expression of confidence level is no longer present in your brain from your middle school statistics classes, don’t worry, we will have separate training for that. In brief terms, the confidence level defines the overall size of unexpected losses and by that in terms of banking the amount of equity capital an institution needs to have to prevent its own default. If the unexpected loss exceeds the confidence interval, the institution will experience a default on its own.
In summary, we now have learned how we can derive the amount of capital that an institution needs to provide to cover the credit risks it is assuming in its loan portfolio. This serves us as a quantification of the credit risk. We can take this concept and apply specific formulas that have been developed to calculate the amount of capital a financial institution needs to provide for specific loans. We cover the background to those formulas in a separate training, for now, let’s just use an example.
If we apply the formulas similar to those that are derived from regulations such as the Basel risk framework, our $10,000 Loan which we used in our example requires a total amount of equity capital of $1066.49. We have used the 5% probability of default and the 40% loss given default from our example. Our loan has a 4-year maturity. We further used a confidence level of 99.9%, meaning, we like to have enough capital that our institution in theory only defaults once every 1000 years. As a result, we calculate an Unexpected Loss of 966.46 and an Expected Loss of $100. The unexpected loss can also be interpreted as economic capital.
The Unexpected Loss or Economic Capital is a factor that we will use later in the Risk Appetite concept. Given that we now have an indicator of the risk related to a specific loan in the form of an equity contribution, we are in the position to calculate a Risk-Adjusted Return on Capital or RAROC. Conceptually, the formula for RAROC looks like this:
We look at all the revenues that are generated from the loan, this could come from Interest Income and Fee Income. We put all expenses against that. They usually consist of Operating Expenses, Expected Loss, and Funding Expenses, i.e. the cost of borrowed funds to finance the loan. Further, we need to credit the Capital Benefit back. Capital Benefit refers to the fact that our loan is not fully financed with borrowed funds. Parts of the funding, namely the amount equal to the Unexpected Loss or economic capital come from equity. As we want to calculate a return on equity, we do not charge funding costs for that part.
Let’s look at some numbers:
Let’s assume that our Loan carries an interest rate of 10%. In addition, we have a 1% Disbursement Fee which we distribute over 4 years. We assume an interest income of $1,000 as the loan amortization happens at the end of the first year + $25 for a quarter of the fee. We assume operating expenses of $300 per year based on input from the finance/operations team and Funding Cost of 5% or $500 as this is the amount against which we can borrow money. We also calculate a capital benefit of $48.32 which is equivalent to multiplying the $966.49 of unexpected loss with the 5% of the cost of funding. Our total revenues $1,073.32 against total expenses of 900. This leaves a positive income of 173.32 dividing this income by the total unexpected loss of $966.49 we can calculate a RAROC of 17.93%. This method is most transparent to document the actual return an investment can achieve. In this case of course the investment is the bank’s decision to provide a loan to a customer.
Let’s have a look at a few scenarios to play around with the calculation so far:
What happens if we increase the size of the loan but leave all other terms and conditions unchanged? If we double the loan amount from $10,000 to $20,000 our RAROC will increase. The reason for this increase is the assumption that our operating expenses remain stable. This is a fair assessment, considering that the bank most likely has the same operational effort to provide a $20,000 loan versus a $10,000 loan.
We can also see the effects of an increase in the risk of an exposure. This is expressed in the Probability of Default. Our initial loan had a 5% probability of default resulting in a RAROC of 17.9%. If we increase the probability of default to 6%, the RAROC goes down by 2.6% to 15.3%. You can see the significant effect a change in risk profile has on profitability. The reason for this is the increase in the unexpected loss, up from 966.49 to 1021.75. At the same time, the expected loss grew from 100 to 120. The increase in economic capital also led to a small increase in the capital benefit, which does not significantly influence the end result. An interesting side aspect here is the fact that the profitability of a loan can vary for the institution if over time, as the risk profile of a client changes.
Based on these three simple examples, you can already see the potential that lies within this formula to steer the overall financial institution. It is really powerful and has many use cases. We are looking at two of them.
Risk Appetite Concept
As a background: Financial institutions tend to optimize processes and workflows to improve efficiency. When it comes to assessing the risk of a counterparty or a transaction, such improvements reach limits, when the thoroughness of an assessment competes with speed. If we assume that we have to assess the credit risk for a specific transaction, and there are many reasons to do that, then our attempt to optimize turnaround time reaches its limits.
At the same time, we are conscious that we are preparing the Financial Institution for a fully upgraded Business Strategy. Such a strategy should be complemented by the definition of a risk appetite. A financial institution’s risk appetite is typically defined through a combination of quantitative and qualitative factors and is influenced by risk tolerance, business strategy, regulatory environment, and the overall risk management framework. If designed and executed well, the risk appetite will help to establish a clear boundary within which the financial institution could operate, without exposing itself to excessive or unacceptable levels of risk. Developing the risk appetite is an exercise that is best done with an outside advisor or consultant such as those provided by Q-Lana.
There are several components that are combined to develop the risk appetite:
Let’s begin with risk capacity, which is the maximum allowable risk an institution can assume without breaching regulatory capital, liquidity, operational constraints, or external obligations like covenants. It serves as a fundamental constraint on an institution’s risk-taking.
Actual risk appetite represents the level and types of risk an institution is willing to embrace within the confines of its risk capacity. This alignment is crucial for achieving strategic objectives and business plans.
Risk limits are quantifiable metrics based on future assumptions that allocate an institution’s overall risk appetite to business lines, legal entities, specific risk categories, and concentrations.
Several factors influence overall risk appetite:
- Regulatory and Compliance Requirements: The risk appetite must align with regulatory mandates to ensure compliance with laws and regulations.
- Business Strategy: Risk appetite mirrors an institution’s strategic goals. Growth-oriented strategies may accept higher risk compared to stability-focused ones.
- Stakeholder Expectations: Shareholders, customers, regulators, and the broader market can influence risk appetite.
- Shaping Risk Culture: The risk appetite statement sets expectations for risk awareness, communication, and behavior across the organization.
To quantify risk appetite, Economic Capital and unexpected loss are common metrics. An institution assesses its current Economic Capital requirements and risk profile and compares them to the target profile based on strategic considerations and other input factors. This guides adjustments in risk allocation over time, albeit at a slower pace than in hedge funds or similar entities.
Once the new target risk appetite is defined, it can be allocated to sectors, products, regions, and even individual counterparties. This specific allocation facilitates quicker credit decisions and necessitates regular updates of borrower risk profiles, along with an escalation mechanism for exceptions.
This approach not only enhances the institution’s business strategy but also improves its risk assessment and allocation processes, resulting in more effective risk management.
Relationship Pricing:
Now, let’s explore the second concept: relationship pricing. Traditionally, financial institutions relied on fixed pricing tables to determine interest rates and fees for their financial products. Unfortunately, this approach often overlooked the broader relationship aspect and resulted in missed business opportunities. In a customer-centric banking approach, pricing financial products revolves around the entire customer relationship.
To achieve this, financial institutions need a pricing model that can calculate both actual and projected revenues for a customer relationship, whether considered on its own or in conjunction with related clients. For instance, if a company is a client, the pricing of its financial products could be adjusted favorably if the company’s employees also maintain a relationship with the institution.
The ideal pricing concept for this purpose is Risk-Adjusted Return on Capital (RAROC), which we’ve discussed previously. It factors in overall product revenues, direct costs, the total risk capital utilized in the client relationship, and target returns. The method used to calculate overall risk mirrors the previously defined risk appetite. By having such a real-time system in place, institutions can dynamically adjust pricing for specific clients, ensuring mutually beneficial relationships.
Let’s illustrate this with an example using the loan we discussed earlier, which achieved a 17.93% RAROC. Suppose we add a current account for the borrower and their spouse, charging $10 per month per account and estimating a $5 monthly cost per account. Since current accounts generally don’t introduce credit risk, there’s no need to adjust the risk capital. The RAROC increases to 30.35%. This means that if the client requests it, we could reduce the loan’s interest rate by 1% and still maintain higher profitability compared to not having the current accounts.
This calculation demonstrates the potential of this approach when you have the necessary data and tools at your disposal. This wraps up our in-depth exploration of some key concepts within our business strategy.
This concludes our excursion into the explanations of RAROC, Risk Appetite, and Relationship Pricing. In the next chapter, we will delve into the tools and partnerships that facilitate the implementation of this strategy. Below, you can also find the content of the Blog as a video. Due to its length, we have split the recording into two parts.
Part 1:
Part 2:
