Risk Management in Financial Modeling

Financial models were widely used by corporations, even in 2008. However, the severity of the 2008 crash forced financial institutions to rethink their approach towards modeling. Many assumptions which are inbuilt in a financial model were being changed to imbibe the lessons learned in the great recession. One such lesson learned was about risk management. After the collapse of 2008 and all the financial losses that it caused, risk modeling ended up becoming an essential part of every financial model.

In this article, we will understand the details about why risk management is important from a financial modeler’s point of view:

How Financial Modeller’s Define Risks?

In common parlance, the risk is defined as the possibility of a loss or injury of any kind. However, in terms of finance, the risk is defined as the deviation from the mean. Hence, risk management is all about understanding the relationships between the possible outcomes of an event. The whole exercise focuses on the identification of the probability that a certain negative event will occur in the future. Once the probability is determined, the impact of the event is also considered to arrive at a decision.

Simply put, financial modelers define risk as a probability that an event will occur and multiply it with the impact of the event.

Risk = Probability * Monetary Impact

In most cases, probability and impact are inversely related. This means that incidents which have a high possibility of happening generally have low impact and can be easily mitigated. These are events which most companies will have mitigation plans ready for. From a financial modeler’s point of view, it is easy to identify and simulate these events to enable a response which would cause the least financial damage.

On the other hand, there are some events which have a very small probability of happening. However, if these events do occur, the loss would be enormous and would even jeopardize the very existence of a business. After the events of 2008, these events came to be known as a “black swan” event based on the book “Black Swan” by Nicholas Naseem Taleb. Prior to 2008, these risks were seldom included in financial models. However, post-2008, almost every company considers some black swan events and creates basic plans to mitigate these events if they do occur.

Common Approaches to Risk Modelling

Risks often mean uncertainties. This means that risks represent different situations which the business may have never seen before. As a result, different approaches have to be taken in order to model different types of risks. Some of these approaches have been written below.

1. Statistical Risk Modelling: Statistical risk modeling is the type of approach which can be used even when the underlying causes of the risk are not known. For instance, companies do know the exact causes which result in a rise in commodity prices. The reality is that a rise or fall in commodity prices is caused by many factors which may be too complex to express in the form of a cause-effect relationship.

   However, if the company conducts statistical analysis, they may be able to find correlations between rising commodity prices and other variables. These correlations can act as close substitutes since the exact causal relationships are unknown. The indicative variables can then be considered to be a leading indicator in risk modeling. Statistical risk modeling is the backbone of scenario analysis. This is because a scenario is nothing but a group of inputs. Statistical risk modeling is crucial to find out the combinations in which these inputs are found.

2. Mathematical Risk Modelling: There are many instances when financial modeler is fully aware of the cause-effect relationship, which creates the risk. In such cases, elaborate statistical modeling is unnecessary. In such cases, the financial modelers are expected to create a smaller sub-model, wherein the particular risk in question can be modeled. The output of that sub-model must then be filled into the main financial model as an input. The problem is that here it is the modeler’s job to create formulas which mimic reality. Hence, the models will only be as good as the financial modeler creating it!

3. Computational Risk Modelling: There is a relatively new field of study called computational risk modeling, which harnesses the power of computers to create millions of scenarios in nanoseconds and provides information about the various inputs. Let’s understand this with the help of an example.

   Let us assume that a pizza needs to be baked wherein there are 20 different ingredients used. Now, if the baker wants to find out what each ingredient does, the ideal way to do so would be to leave out one ingredient at a time and record the effect on the taste of the pizza.

   Hence, a total of 20 cakes would have to be baked just to obtain some basic information. Similarly, the baker can then change any two ingredients at a time and run an experiment.

   This time a total of 190 cakes would have to be baked! Similarly, 1100 cakes will have to be baked if four ingredients are varied at the same time.

   As we can see, varying inputs will provide us information. However, it proves to be time consuming and expensive which makes it infeasible. This is where computational models come into the picture. They allow the use of computers to create millions of scenario, collect data points from each scenario, analyze the same using big data techniques, and provide actionable information to the user.

   Since financial modelers already have the tools required for analyzing large volumes of data, this method is preferred by them.