Forecasting agricultural exports and imports in South Africa

Abstract

The implementation of wide-ranging policy reforms, including trade and exchange rate policies, is improving the efficiency of the South African economy and its reintegration into the global economy with rapid export expansion. Agricultural exports in the Southern African Customs Union increased from R8.14 billion in 1995 to R23.0 billion in 2003, whilst agricultural imports rose from R6.83 billion to R13.84 billion during the same period. This article uses alternative approaches to forecasting agricultural exports and imports in South Africa. The models used include: exponential smoothing, autoregressive integrated moving average (ARIMA), vector autoregression (VAR), Engle–Granger (EG) single-equation and vector error-correction models (VECM). We found that the ARIMA and EG methods outperform the VAR and VECM according to Theil's U-statistic. The VAR outperforms the VECM in forecasting agricultural exports in South Africa. The combined forecasts have a lower variance compared to individual forecasts, thereby, reducing the risks of making wrong decisions based on the forecasts. The article provides empirical evidence that is beneficial to policymakers and business leaders in South Africa as they strive to reduce poverty and inequality and increase economic growth.

Acknowledgements

The comments and suggestions by Dr P. J. Nomathemba Seme are very much appreciated. Thanks to Siphokazi Kargbo for her encouragement to pursue this work. My sincere thanks also go to various staff members at the South African Mission to the IMF/World Bank in Washington, DC; the National Department of Agriculture, Statistics South Africa and the South African Reserve Bank for providing some of the data and materials used in this article. The views expressed in this article do not reflect the official position of American Express Company or its affiliates. The author accepts sole responsibility for any errors and omissions.

Notes

¹ SACU is made up of South Africa, Botswana, Lesotho, Swaziland and Namibia. South Africa has the dominant economy and the union's largest agricultural exporter and importer (Sigwele, 2003; Kargbo, 2006).

² The total trade weights (in parenthesis) for the trading partners used in computing the real multilateral exchange rate index are: USA (14.44), UK (14.09), France (4.98), Japan (9.9), Belgium (3.54), Switzerland (4.99), Germany (16.91), Italy (5.07), South Korea (2.50), the Netherlands (3.90), Australia (1.59) and Zimbabwe (2.27). See Walters and de Beers (1999) for calculation of the trade weights and Edwards (1989) or Kargbo (2000, 2006) for the formula used in calculating the multilateral real exchange rate (MRER). The wholesale price index (WPI) was used for foreign prices of trade partners and South Africa's GDP deflator for domestic prices.

³ARIMA(1,1,1)-GD is an ARIMA model with the dummy variable GOVD included in the estimation. H_Winters = Holt–Winters and Brown's DES = Brown's methods.