COMPARATIVE FORECASTING OF PAKISTAN’S REAL GDP GROWTH: AN EMPIRICAL ASSESSMENT OF ALTERNATIVE FORECASTING APPROACHES

Abstract

Gross domestic product (GDP) growth is an important indicator of macroeconomic performance and provides essential information for economic planning, resource allocation, and policy formulation. Accurate forecasting of economic growth is therefore crucial, particularly for developing economies such as Pakistan, where economic conditions can be influenced by substantial structural and macroeconomic changes. This study comparatively evaluates alternative forecasting approaches for Pakistan’s real GDP growth and generates forecasts for the period 2023–2032. Annual real GDP growth data covering 1960–2022 are used to develop and evaluate five forecasting approaches: Autoregressive Integrated Moving Average (ARIMA), double exponential smoothing, multilayer perceptron (MLP), neural network autoregressive (NNAR), and a hybrid forecasting model. The models are assessed comparatively to identify their ability to capture the underlying temporal patterns in Pakistan’s economic growth and provide reliable long-term forecasts. The empirical results indicate that the multilayer perceptron (MLP) model provides the most accurate forecasting performance among the evaluated approaches. The findings demonstrate the potential of nonlinear neural-network-based methods to capture complex patterns in Pakistan’s historical GDP growth that may not be adequately represented by conventional statistical and exponential smoothing techniques. The resulting forecasts for 2023–2032 provide an empirical basis for assessing the prospective trajectory of Pakistan’s economic growth. Overall, the study contributes to the forecasting literature by providing a comparative assessment of conventional, exponential smoothing, neural network, and hybrid approaches using a long historical series of Pakistan’s real GDP growth. The findings may assist researchers, economic planners, and policymakers in selecting appropriate forecasting techniques and developing evidence-based strategies for medium- to long-term economic planning.