Summary: |
Using individual shopping centre data from Portugal, the paper compares the performance of traditional econometric techniques (i.e non-pooled estimators) and panel data techniques (pooled estimators) as forecasting frameworks. The data provided by the main Portuguese retail operator was not long enough, in some cases, to be used with traditional econometric techniques. For this reason, pooled estimators are suggested as an alternative. The paper discusses the advantages and disadvantages of this approach, compared with traditional econometric techniques (TET). It shows that pooled estimators can be a good alternative to traditional econometric techniques when data is scarce. Moreover, pooled estimators provided, in almost all the cases, more accurate forecasts than TET. |