What's the Esade Target?

Esade Target publishes forecast deviations measured in absolut points. The reason for not using percentages, or relative figures, is because if an organization forecasts a rise of 1% in GDP and the actual figure is 2%, then the error based on the forecast is 50% (or 100% based on actual data). Using a percentage formula produces other problems, since it is not the same forecasting 2% if actual GDP growth is 4%, as forecasting 1% when the actual increase is 2%. Both cases would show the same percentage error, but the impact in the first case is two points, while the impact in the second case is just one point. In addition, percentage deviations offer little information to the manager or entrepreneur – and can be misleading. A deviation of 33% seems very large, but if the GDP growth forecast was 1% and actual growth is 1.33% – then the difference is small.
 
For these reasons, Esade publishes the MAE or Medium Absolut Error. If an organization forecasts 2% and the actual figure is 0.5%, then the deviation is 1.5 points. The same applies if the forecast is below the actual figure. If a forecast is 1.5% and actual figure is 3%, then the deviation is also 1.5 points. Information presented in this way is understandable, offers absolute values, and makes sense to managers and entrepreneurs. Deviations, both positive and negative, are measured by the number of points a forecasting organization tends to deviate.

Another advantage of working with direct differences in absolute value is that an evaluation of forecasters is independent of who is on the list. Esade is not interested so much in producing a ranking, as providing company managers with useful information for assessing the value of a forecast (that is, by how many points does the forecaster tend to deviate). This enables managers to calculate a range within which to place a given forecast.
 
By working with direct differences in absolute values between forecasts and reality, the calculation is unaffected by the degree of success or error of others, nor by the organizations included in the calculation. For example, if a forecaster registers a one-point deviation, it is useful to know if that organization is always high or low in its forecasts (or shows no bias). Including this information in the general Esade Target index would distort the analysis. For this reason, we provide an additional table with deviation calculation data for each forecaster.

Esade Target takes the forecast that each organization published for the year ahead, and compares it with the actual GDP and employment figures (using for jobs the officially published survey of the working population known as the Encuesta de Población Activa). The difference between forecast and reality is then calculated in absolute points, and the same operation is repeated for the previous two years to produce an average.
 
Longer and shorter time periods were considered (from one to five years). However, Esade opted for three years because a one-year horizon would give excessive weight to occasional error or accuracy, and lead to abrupt changes in the index; while a five-year horizon would be too harsh on organizations that make especially accurate one-off forecasts (because their standing would not improve much without a further three-year run of good forecasts).
 
A three-year horizon enables forecasting organizations to improve their showing relatively quickly, while also requiring enough of a track-record to enable robust conclusions to be made about an organization’s forecasting ability.