The agricultural sector is the backbone of the economy and plays a vital role in the economic development of Pakistan. It contributes about 24 per cent towards the GDP of the country. The agricultural sector mainly depends upon the major crops grown in Pakistan of which, two of these crops (cotton and rice) are major foreign exchange earners. Wheat and rice are major food crops and wheat is one of the major crops grown in all the four provinces of the country, cultivated on 8.18 million hectares and accounting for 37.19 per cent of the total cropped area. As wheat is the principal staple food, its magnitude and size, particularly shortfalls in production, deeply influence the economy of the country.

Many agencies are involved in this work in Pakistan. The Pakistan Agricultural Research Council (PARC) is one of these agencies playing a leading role in agricultural research. The Biometrics Programme is one of three components of the Social Sciences Institute, placed at NARC, which is the federal component of PARC. This programme has undertaken considerable work on forecasting the area and yield of important crops, including rice.

Recently, the Biometrics Programme has developed a forecast model for wheat crop production. The forecast model is based on time series data of 28 years (1975-1976 to 2002-2003) collected from secondary sources. The performance of the forecast model as tested for area, yield and production of wheat was significant. By using this model policy makers can predict the yield of a particular crop or forecast area which will be sown in the next crop season.

The technique used for crop production forecasting is Ordinary Least Square (OLS) of regression analysis. Five equations (one for Pakistan and one for each of the four provinces) are used for wheat area and one equation to forecast the wheat yield in Pakistan. Production is estimated by multiplying area and yield forecasts.

The production of wheat depends upon a number of inputs/factors like fertilizer, water availability, rainfall, temperature and so on. The variables which have a linear relationship are kept for this model. The variables selected for the wheat area model are lagged area under wheat by one year and wheat procurement prices. The variables used in the wheat yield model are fertilizer consumption for wheat (kg/ha) and total water availability at farm level.

This model enables the policy makers to estimate the yield of wheat before harvest. The planners, with the help of this model, can also estimate the area of wheat that will be sown during the next crop season. The accurate forecasting of wheat area and production may support the policy makers and planners for making policy decisions regarding supply, demand and imports of wheat in the country.

The wheat area forecast model was tested for three years (2000-2003). It was observed that the estimated wheat area of Pakistan was 7 per cent and 4 per cent higher during 2001-2002 and 2002-2003 respectively, than the official target area of wheat.

The performance of the model was also tested for the yield of the wheat crop. It was found that the estimated wheat yield of Pakistan was 2 per cent and 6 per cent lower during the years 2001-2002 and 2002-2003 respectively, compared to the official estimated yield.

Written by Muhammad Asif Masood and Anver Javed, Pakistan Agricultural Research Council

References:
Ahmed, N., Bengal, K., and Iqbal, F., 1983. Long-run Demand and Supply of Major Agricultural Products. Appl. Econ. Centre, University of Karachi.
Azhar, B.A., Chaudhry, G.M., and Shafique, M.A., 1973. A Model for Forecasting Wheat Production in the Punjab, Q.J. PIDE XIV (1):407-414.
Granger, C.W.J., 1980. Forecasting in Business and Economics, Academic Press, Inc.
Khan, M.I. and Khan, N.A., 1988. Rice Forecast Model of Pakistan, Gomal Univ. Journal Res. 8(1and 2): 95-105.