How to setup credit limits?

What do we want from robust corporate credit risk management?

In the last post we have written about the dangers of credit risk. The logical next step is to give recommendations on how to best construct a credit risk management system, that accounts for the issues outlined in the previous post.

 What do we want to achieve with credit risk management system?

 So what are the goals a robust credit risk system should achieve:

• Allows directing sales to economically most profitable counterparties.
• Calculates price markup needed for each of the counterparties.
• Enables fine tuning of the credit exposure on portfolio level.
• Can use financial statement data to form an internal credit rating of the counterparty, which can then be unified with external credit ratings.
• Is flexible and allows addition of new risk predictors or changes in operational situation

A simplified example of such a structure that achieves the aforementioned goals is given in the graph.

 Determining credit limits & credit markup

There are two major benefits of having a proper structural model for measuring credit risk. It allows for a robust and consistent evaluation/determination of credit limits and credit markup. 

With respect to credit markup we are often led to believe that we should only consider expected cost of credit risk. This however is deceiving, incorporating only expected cost of risk into markup will lead on average to a sufficient compensation most of the time but will expose the company to significant probability of default and credit tail risk. (See picture below)

 The real trick here is to determine both the expected and tail risk cost and determine how to allocate these costs to per customer basis. 

With respect to credit limits, at first glance it may seem that one cannot go wrong with using simple logic in setting up credit limits and that it all cancels out in a portfolio of many counter parties. There are two common pitfalls that companies tend to get wrong time and again:

1. Credit risk is highly codependent especially in the tails (in deep recession everybody is broke), so it’s not trivial how to set limits and markup that appropriately protect against risk but do not unnecessarily hinder sales.
2. Having a good insight about clients default risk does not directly translate to good use of that information and in most companies? the process of translating risk ?information to a limit is done in a haphazard way resulting in inconsistent limits and therefore causing unnecessary losses as well as missed profit opportunities. 

An example for how to conceptually tackle the problem of setting up credit limits within the methodological framework we presented is given in the following chart. The chart mostly deals with credit limits since that is the risk management part of the story but simultaneously we can calculate the credit markup any client and let the credit risk takers use this information when deciding upon margins.

Conclusion

Credit risk is a sneaky kind of animal. Most of the time you don’t see it and it may seem it is never coming back. When it reappears it is much bigger than the last time you remember it with far bigger consequences. Therefore credit risk really demands a disciplined comprehensive system to manage and should not be left on “I feel good about this client” basis.

by Ales Ahcan and Niko Jarnberg