Forecasting ______________________________________________________
Many analytical frameworks and modelling techniques can be used to produce forecasts, incorporating different assumptions about how the economy may unfold over time. Unfortunately many standard models hide these underlying assumptions, creating a 'black box' that prevents users from sense checking and testing the robustness of their models. Volterra provide transparent and accountable analysis and supporting documentation. We fully discuss with our clients the methodology which has been employed and the assumptions chosen to underpin our analysis.
But forecasting is error-prone. Even if all the drivers of the business can be precisely identified and measured, which is itself no mean task, forecasts still fail. This is not just because random events can occur to knock a system off course. It is also because the interaction between different drivers can produce fluctuations which cannot be predicted. Indeed this was one of the insights of a mathematician called Vito Volterra, for whom we are named.
Thus the provision of forecasts needs to rest on two central propositions:
- Making the best use of all available data and statistical techniques
- Consideration of the likely range of outcomes and the potential for unforeseen outcomes
At Volterra Consulting we have developed an approach to forecasting which aims to achieve these. The key features of our approach are:
Open minded >>
A one-size-fits all approach to modelling is simply not appropriate. We take an innovative approach to economic modelling, based on a very wide experience in economic consulting. We also draw from diverse scientific backgrounds to bring new modelling techniques to business and economic modelling.
Our approach is also interactive. We are keen to work with our clients and use their expert knowledge of their business together with our statistical and modelling techniques to ensure we produce the best possible view of the future direction of the client's business.
Recognition of uncertainties >>
In the face of unavoidable uncertainty the correct approach is not to pretend that models are sophisticated enough to produce infallible predictions, but instead to quantify the uncertainty around predictions and analyse the potential effects on business.
An acknowledgement of uncertainty leads on naturally to a consideration of risks. We know that many things could happen in the future. It is, therefore, more sensible to move away from talking about what will happen and talk about how likely certain critical scenarios are to occur.
Advanced techniques >>
Following from all the above, it is natural that we need to be using the best possible tools to analyse likely future outcomes for our clients. This is where our expertise in statistics and modelling comes in, and is a big part of why we can add value to your business.
We recognise that our clients are the experts in their own business areas. Our strength is in working with our clients to ensure that the best information is used in the best possible way when forming a view about the future of the business. We do this by applying our statistical and modelling expertise and economics knowledge to the forecasting problem at hand, together with the insider business knowledge provided by the client.
Some examples of techniques we employ are:
- Econometric Modelling - using statistical techniques developed in the field of econometrics to identify the relationships between key variables
- Time Series Modelling - analysing the time path of key variables and identifying trends in variables and also identifying relationships between trending variables
- Evolutionary Modelling - dynamic modelling based on the technique of Tacit Knowledge Mapping
- Stochastic Simulation Modelling - using an agent-based approach, with a stochastic element included to reflect uncertainty of agent behaviour, to model the complex interactions of economic agents who make up markets through simulation
Although we use the best techniques available, it is not our ambition to bamboozle our clients with complicated mathematics. Our models are only as complicated as they need to be. We try to avoid the common problem in forecasting of delivering a 'black box', where the client is unsure of what methods have been used to get to results. Transparency in what we have produced and clear communication with our clients is fundamental to all of our work.
Forecast Auditing >>
