Building an Expert Judgement Credit Rating Tool for SME and Corporate Banking Customers

Building an Expert Judgement Credit Rating Tool for SME and Corporate Banking Customers

One of the key elements in improving the quality, consistency and efficiency of SME, and wider commercial and corporate banking, is the application of credit analysis and assessment tools. Sometimes called 'credit decisioning systems', these include a wide range of statistical and qualitative approaches. Increasingly these tools are also often demanded by regulators looking to implement improved capital management practices in their jurisdiction. In this case, we are going to take a deep dive into the expert judgement scorecard approach, in its many variations the most popular and practical approach for most commercial banks operating in most markets.

Risk Segmentation: The Foundation of SME Credit Rating

In this post we are exploring the idea of using risk segmentation to form the foundation of our SME credit risk model. Feel free to view the screencast below independently of the body text, as it might add some colour, or compress or substitute some valuable reading and comprehension time. Don’t forget to expand to full screen and watch in HD for best results!

You may recall from previous posts that Industry Sector can be considered the foot of the pyramid in terms of factors when rating a customer. Critically, industry sector analysis should also be at the root of our marketing and risk strategy for the SME market. However Industry Sector analysis, in itself, is not really nuanced enough to develop a truly sophisticated and differentiated approach to risk (and market segmentation).

For example, consider the case where you have an Industry Sector definition for, say something simple and easy to imagine like the ‘Food Service’ industry. Depending on the other parameters defining your SME banking business model, this could include tiny customers like street food or cart vendors (nano- or micro-enterprises), up to and including quite large restaurant or takeaway franchises or chains (perhaps classified as mid-cap or medium-sized enterprises). The risk characteristics of these customers will be very different, despite falling within the same industry sector.

We could just simply use size as an additional segmentation criteria (but what criteria? – turnover, capital, profit, staff or distribution channels), however what about other factors like longevity of the firm, or ownership structure, or some of the many other factors that may contribute to business risk. Rather than ignore them, or build a hugely complex multi-dimensional segmentation model, we have borrowed from Retail Banking marketing strategy somewhat to introduce the concept of ‘Commercial Profile’. The Commercial Profile can be used as an additional vector to be considered in risk segmentation, and will be covered in detail in future posts. It’s worth noting, however, that using Commercial Profiles is also a neat way to dovetail your risk strategy with your marketing strategy for the SME market.

So now we have two axes, Industry Sector and Commercial Profile, which we can assess discretely and combine to form a risk segmentation model, represented by the heat map below:

Remember from previous posts that Industry Sector and Commercial Profile are largely exogenous factors for a firm in the credit assessment process. They are not readily subject to manipulation by the firm itself, as opposed to business risk and financial criteria. For most practical purposes, firms can be allocated to segments based on these risk segmentation criteria, and not expected to shift markedly, at least for a short-medium term credit.

This is important, because it means we can prepare a risk segmentation model independently from the credit assessment process. Indeed, we may want to do this before even embarking on developing or changing our SME banking business model. And this process can be completed centrally, a collaboration between risk management and the SME Banking business, to ensure harmony between the risk and marketing strategies.

Note that the risk segmentation model is also ‘Through-the-Cycle’, and is only reviewed periodically or in the event of some significant problem or change in circumstances. This means we don’t have to complete a separate industry sector risk assessment (or commercial profile for that matter) for each and every customer or credit application. Believe it or not – we have seen some banks that do this! Not only is risk segmentation more efficient and reliable, but also imperative if you want to adopt any sort of portfolio approach to SME credit.

However, so long until next time. Please feel free to comment or email; feedback is gratefully received.

Screencast of Expert Judgement SME Rating Prototype

In the last post, we introduced the graphic below to illustrate the building blocks of an SME credit rating tool. Our approach builds an individual customer rating from four distinct components. These include 'exogenous' criteria such as the Industry Sector. These are criteria that the customer cannot really change (assuming they don't radically change their mission or business model). In which case, it may be necessary to reclassify them.

We supplement this with a 'Customer Profile' that captures some additional intangible characteristics of the firm. These are characteristics such as ownership structure, size, growth prospects, intellectual capital, or longevity for example, or even identifiable values and behaviours of owner/ managers (e.g. entrepreneurialism, risk appetite, ambition etc.).

The combination of these two factors creates a 'Segment' rating that is static and 'through-the-cycle' for any given firm. This supports active portfolio planning and management, and can dovetail nicely with a variety of approaches to market segmentation.

Towards the other end of the spectrum we have the 'endogenous' criteria. These are criteria that are specific to individual firms within the same segment. Management more or less has control (or should have) of its exposure to business specific risks and the management of its own finances.

The total of these building blocks gives us the total score and combined customer rating. Clear enough I hope, but we'll be getting into more detail in future posts, starting with defining and creating an Industry Sector rating. In the meantime, perhaps the screencast below will illustrate how these four elements hang together in the overall risk rating tool. However, don't hesitate to contact me if you have any queries or comments. I look forward to your feedback.

n.b. MAKE SURE YOU EXPAND TO FULL SCREEN AND WATCH IN HD FOR BEST EFFECT.

Overview of an Expert Judgement Model for SME & Commercial Customers

While there is no magic formula to developing an expert judgement credit rating tool for SME and commercial customers, there are some sound principles that most models adhere to. In the presentation below we take an overview of the GBRW approach, which attempts to distill the best approaches we have seen used successfully in banks around the world, and to reassemble them in an orderly and logical bundle. For the purposes of illustration, we will be using the GBRW model as a working prototype to explore some of the key concepts underpinning expert judgement rating in future posts.

However, every bank must take care to adapt and refine their own model to suit their own circumstances, considering the local business environment, the target market, and the internal characteristics of the bank itself. Remember, a key element of competitive advantage in SME banking is the ability to understand the risks of the target market better than competitors, to define a segmentation strategy and market positioning that plays to the distinct strengths and weaknesses of the bank, and to clearly differentiate their Customer Value Proposition. An expert judgement rating tool is not just a risk management tool. It is a marketing tool as well.

In the next post, we will be taking a walk through of a working prototype expert judgement model, including a screencast so we can really get to grips with how the different components fit together. This will form the foundation for exploring each element in greater detail during the rest of the series.

Barriers to Developing an SME Credit Risk Expert Judgement Model

customer site visit

You're an SME or Commercial Banking Relationship Manager driving home from a team building event which finished early. Your spouse is still at work, and your kids not back from school yet. So rather than sneaking home early, you decide to get a head start on things. You pull over into a lay-by, fire up your tablet computer, login to your bank's CRM app, and check on your work flow.

You notice that Johnson’s Widgets, a long-standing customer of yours, is due a ‘Business Risk Review’ and, as it happens, they are just around the corner. You place a call to their Financial Director and luckily he’s in. Actually, he’s been meaning to call you because he’s interested in applying for an extension of their trade finance credit line in advance of a big order. Before you set off, you notice that the most recent set of accounts has been updated, and the app has neatly calculated the relevant ratios and trends. Interestingly you note that a significant improvement in debtor collection has resulted in a boost in the ‘Financial Assessment’ score.

When you get to the factory headquarters, rather than sitting in the FD’s office, you both take a stroll around the facilities. Along the way, you use your tablet to dynamically update the ‘Business Risk Review’ based on the structured questionnaire which underpins it. You note that while business is growing, Johnson’s is becoming increasingly dependent on a single customer, which is reflected in the ‘Business Risk’ score. The FD counters that they are developing several new products which should expand their range of potential customers.

Back at the office, the FD enquires about the potential increase in the trade finance credit line. Without leaving, you immediately discuss and add some basic product parameters and pricing, and test them against the model. With a little tinkering, you find a solution which meets the customer’s needs and is comfortably within the risk/ reward tolerances of the bank. You give the FD a provisional, positive response and promise to follow-up with a formal application tomorrow.

The FD is delighted, and you sleep soundly having not only completed some risk management administration, but also closing another deal which brings you closer to your targets.


While there are no doubt some banks in emerging markets which operate at this level, there are far too many which do not. Why not? I often hear complaints in emerging markets about relationships between bankers and SMEs which seemed to be based on obfuscation and avoidance. That is not an excuse for avoiding change, rather it is a compelling reason for banks to proactively adopt more objective and transparent approaches (both internally and externally) to credit assessment.

We all have our personal theories on the barriers to achieving a seamless relationship between risk and relationship management, especially in relation to credit assessment and rating. Here are some of mine:

» Some Relationship Managers resist implementation of changes in credit risk assessment methodologies because of the potential loss of personal influence in approving and pricing credit, or the status of their customer set.

» Some Relationship Managers worry about the competitive impact of a credit risk model, the impact on favoured customers, and the potential upheaval in the pecking order of customers.

» Some banks may lack the confidence and perhaps some of the in-house skills required to take an independent approach and are rather dependent on vendors.

» Some banks seem to be driven by solutions from technology vendors which often drive up costs and complexity massively, thus killing any project in its infancy.

» Some ratings agencies are pushing statistical solutions based on data and models which are not fully understood or trusted by banks in emerging markets in particular.

» Many banks have poor historical data, which can be used as a ‘straw man’ by opponents of change. Sometimes the inadequate financial reporting environment is also used.

» Some banks do not seem to have sufficient trust in the basic analytical skills of their staff, and are biased towards a pure statistical solution at the expense of a blended heuristic approach.

This post aims to be a launch pad for a longer running series on the practical implementation of credit rating models, particularly expert judgement solutions for MSME customers. However, before we do so I think it useful to flush out as many challenges and issues as possible, so comments are greatly appreciated. What do you think are the most compelling reasons for the lack of progress in what seems to be a very surmountable challenge?