The Indicators’ Inadequacy and the Predictions’ Accuracy

In this article, we proposed the introduction in literature of a new source of uncertainty in modeling and forecasting: the indicators’ inadequacy. Even if it was observed, a specific nominalization in the context of forecasting procedure has not been done yet. The inadequacy of indicators as a supplementary source of uncertainty generates a lower degree of accuracy in forecasting. This assumption was proved using empirical data related to the prediction of unemployment rate in Romania on the horizon 2011-2013. Four strategies of modeling and predicting the unemployment rate were proposed, observing two types of indicators’ inadequacy: the use of transformed variables in order to get stationary data set (the difference between the unemployment rates registered in two successive periods was used instead of the unemployment rate) and the utilization of macro-regional unemployment rates whose predictions are aggregated in order to forecast the overall unemployment rate in Romania. The results put in evidence that the predictions of the total unemployment rate using moving average models of order 2 are the most accurate, being followed by the forecasts based on the predictions of active civil population and number of unemployed people. The strategies based on the aggregation of the predictions for the four macro-regional unemployment rates imply a higher inadequacy and consequently a lower degree of forecasts’ accuracy