The Language of Uncertainty: Reading Türkiye's Economic Pulse Through Geopolitical Fog

Abstract

Geopolitical uncertainty is an increasing concern for investors, entrepreneurs, and researchers. Traditional methods of measuring uncertainty often fall short, whereas Artificial Intelligence (AI)-driven Natural Language Processing (NLP) offers more accurate and competitive options. In this study, we develop a Geopolitical Country-Specific Uncertainty (GCSU) index for Türkiye by applying the economic policy uncertainty index methodology. Since geopolitical uncertainty involves Knightian uncertainty—which cannot be quantified with probabilities in advance—our approach measures it by analysing the frequency of economy-, policy-, and uncertainty-related terms in Turkish-language newspapers. Unlike existing indexes for Türkiye, such as the Geopolitical Risk (GPR) and the Economic Policy Uncertainty (EPU), which rely on English-language sources, our localised method provides a more precise measure of country-specific uncertainty. Our findings show that an unexpected increase in geopolitical uncertainty leads to declines in employment, industrial production, trade, consumer confidence, and stock prices, while oil prices and inflation rise in response. However, these results may not be observed when uncertainty is measured using English sources (e.g., through the index from the GPR-Türkiye). These insights highlight the importance for policymakers to consider geopolitical uncertainty in their economic planning and suggest directions for further research into measuring uncertainty with local language sources.