Economic modelling and forecasting

– Data-driven business forecasting

There is a substantial number of methods and approaches that are used for business planning purposes. You could for example use a bottom up approach, focusing on the individual elements and then combining them up. Or you can use a top-down approach – starting from the overall picture and going down towards your position. You can also use an expert approach – relying on your team’s experience in the respective field, or perhaps do a combination of all the above. While they have historically proven to be often quite effective, there are several important problems with their use:

= Not all parameters can be reliably estimated based on the suggested methods, especially in dynamic environments

= Uncertainty is often difficult to incorporate with business forecasting and point forecasts are usually used

= The element you are forecasting may have quite complex relations with other parameters that require advanced models

If the traditional planning methods work for you, then you probably don’t need to go a step further. If however you are reaching the end of their capabilities, data driven analytical tools for forecasting are a logical step forward. They bring much richer mathematical and statistical methods that can be used to plan a complex, dynamic and risk centric models, which can achieve substantially more than the traditional methods.

– Macroeconomic modelling

Understanding the state of the economy and the probable path it will take in the future is crucial for appropriate planning, risk management and decision making. Developments in aggregate demand, prices and the labour market will affect revenues and have implications for costs and personnel in a company. In order to make sense of current macroeconomic data and produce reliable forecasts one needs to go beyond purely descriptive analysis and rely on a solid modelling framework. This will help to establish the sign and magnitude of the relationships between macroeconomic variables, ensure consistency in projected developments and enable the creation of different scenarios to capture uncertainty about the future. Depending on the task at hand – analysis of past developments, forecasting or simulation – different macroeconomic modelling approaches will be called for. In any case, the use of an appropriate model will provide a sound basis for the analysis and boost the consistency of the obtained conclusions.

– Risk modelling and ECL calculation

Financial institutions as well as non-financial institutions with significant exposures to risk assets need to do risk modelling for their portfolios. They also need to do ECL calculations on their exposures – as required by the financial institutions regulatory framework, as well as by accounting standards like IFRS 9. However, the issue is not only regulatory as risk assessment is vital for the operational effectiveness and overall risk performance of any financial institution. That is why proper risk modelling is important for operational purposes as well. There are many statistical methods employed in risk modelling – starting from the basic 12 month PD calculation and calculating ECL from it to using more complex models for risk estimations such as decision trees, scorecards and other advanced ML algorithms. In any case proper risk modelling is essential, especially in more turbulent economic environments like the one we have now. Analytical tools are a great way to model risk and provide a much deeper understanding of risk in your institution.