Forecasting aggregate retail sales: The case of South Africa

Forecasting aggregate retail sales may improve portfolio investors׳ ability to predict movements in the stock prices of retail chains. This paper uses 26 (23 single and 3 combination) forecasting models to forecast South Africa׳s aggregate seasonal retail sales. We use data from 1970:01–2012:05, with 1987:01–2012:05 as the out-of-sample period. Unlike the previous literature on retail sales forecasting, we not only look at a wide array of linear and nonlinear models, but also generate multi-step-ahead forecasts using a real-time recursive estimation scheme over the out-of-sample period, to better mimic the practical scenario faced by economic agents making retailing decisions. In addition, we deviate from the uniform symmetric quadratic loss function typically used in forecast evaluation exercises, by considering loss functions that overweight the forecast error in booms and recessions. Focusing on the results of single models alone shows that their performances differ greatly across forecast horizons and for different weighting schemes, with no unique model performing the best across various scenarios. However, combination forecast models, especially the discounted mean-square forecast error method, which weighs current information more than past, not only produced better forecasts, but were also largely unaffected by business cycles and time horizons. This result, along with individual nonlinear models performing better than linear models, led us to conclude that theoretical research on retail sales should look at developing dynamic stochastic general equilibrium models that not only incorporate learning behavior, but also allow the behavioral parameters of the model to be state dependent, to account for regime-switching behavior across alternative states of the economy.