BRAZILIAN RETAIL SALES FORECAST: AN EVALUATION CONSIDERING DIFFERENT QUANTITATIVE TECHNIQUES

Abstract

The purpose of this paper is to evaluate the different possibilities to predict retail sales. Retail, which is one of the most important economic activities in the economy, has gone through important changes along the last fifteen years in Brazil economy. In fact reducing drastically the inflation rate, the Real plan completely changed the competition focus. Retail companies were forced to put much more attention on operational aspects. In other words operational losses could not longer be compensated in the financial market.  Obviously sales forecast is essential for efficiency and effectiveness in management especially in retail. This work, thus, compares alternative quantitative techniques in order to appraise the accuracy of these different methods. The conclusion is that the application of Box – Jenkins method is the method that produces the minimum residual square sum followed by the exponential smoothing technique with additive specification.