By Tahir Amin

Copyright brecorder

ISLAMABAD: A new econometric study has proposed a dynamic factor model (DFM) framework, claiming that it provides more accurate real-time tracking of Pakistan’s annual GDP growth than forecasts being made by the Asian Development Bank (ADB), the International Monetary Fund (IMF), and even provisional estimates released by the government.

In its report “An economic framework to now cast low-frequency data,” the ADB noted that the framework adapts a methodology originally designed for economies with quarterly GDP data, modifying it to suit Pakistan’s context, where GDP is measured annually. By integrating monthly, quarterly, and annual indicators, the model generates early estimates of economic activity that are updated as new information becomes available.

Covering the period from January 2000 to June 2023, the study applies the model to Pakistan’s GDP data from fiscal year 2005-06 to fiscal year 2022-23. It incorporates a wide set of variables, including industrial production, external trade, monetary aggregates, commodity prices, stock indices, and global leading indicators.

The main theoretical purpose of this paper is to develop an econometric framework for predicting and now casting low-frequency variables using readily available high-frequency variables. The empirical goal is to apply this econometric framework to track a specific low-frequency variable (annual GDP growth) to highlight its relevance and value addition. We apply the empirical framework to now cast annual GDP growth in Pakistan, a lower-middle-income country, using various indicators to generate a monthly measure of economic activity. Considering that high-frequency measures of economic activity are unavailable for a majority of countries, the proposed econometric framework has broader applications and interests a wider audience.

The best-performing specification, referred to as “Model E,” achieved 28 percent higher predictive accuracy on average compared to alternative models, according to the Diebold-Mariano statistical test. Importantly, its now casts consistently outperformed ADB’s Asian Development Outlook (ADO), IMF’s World Economic Outlook (WEO), and provisional estimates prepared by the National Accounts Committee, the paper added.

During fiscal year 2022-23, the model’s initial projection in July 2022 estimated GDP growth at 4.65 percent. However, as more data became available—including sharp declines in exports, remittances, and worsening inflation—the estimate steadily adjusted downward, settling at 0.32 percent by June 2023. This is closely aligned with the officially recorded contraction of -0.17 percent.

The study attributes the revisions to deteriorating economic conditions, including average inflation of 29.4 percent, a 19.1 percent fall in exports, and a 13.7 percent drop in remittances. It also found that incorporating global indicators, particularly relating to the United States, the United Kingdom, and China, significantly improved accuracy, underscoring Pakistan’s vulnerability as a small open economy.

Copyright Business Recorder, 2025