Dynamic Relationship Between Rainfall and Wheat Production in Balochistan: A Time Series Analysis Using the VAR Model

Abstract

This paper investigates the dynamic relationship between rainfall and wheat production in Balochistan using a Vector Autoregression (VAR) framework with annual data from 2000–2025. Stationarity tests reveal that rainfall is stationary, while wheat production is non‑stationary and requires differencing. Lag order selection identifies two lags as optimal for capturing short‑term dynamics. VAR estimation shows that rainfall significantly influences wheat production at the second lag, whereas wheat production does not affect rainfall. Granger causality confirms unidirectional causality from rainfall to wheat production. Impulse response functions and forecast error variance decomposition further demonstrate that rainfall shocks exert strong and persistent effects on wheat output, while wheat production shocks have negligible influence on rainfall. Forecasting analysis highlights the sensitivity of wheat production to rainfall variability, with negative shocks leading to sharp declines followed by recovery. These findings underscore the dominant role of climatic variability in shaping agricultural productivity in Balochistan.

Keywords: Rainfall variability, Wheat production, Vector Autoregression (VAR)