Marina Diakonova

We introduce a new framework for measuring geopolitical risk (GPR) by tracing it back to its geographical origin [1]. Our approach builds on the seminal work of Caldara and Iacoviello, whose GPR index measures the share of articles in the Anglo-Saxon press that mention elements of geopolitical tension. In contrast, our framework combines two key innovations: (a) using national press narratives and (b) identifying articles that focus on specific regions. This allows us to construct bilateral indicators that capture how each country perceives the risks associated with different parts of the world.

These bilateral GPR indices have several important applications. First, they allow the decomposition of national risk into distinct, non-overlapping components (achieved through a combinatorial approach). Second, they enable us to track how the perceived risk associated with a particular region evolves over time. For example, in the case of Russia, we observe that surges in perceived risk correlate closely with geographical proximity. Finally—and most importantly—we show that the macroeconomic effects of GPR shocks depend not only on the recipient country but also on the origin of the shock, both in magnitude and even in sign. While most shocks depress GDP, some generate positive effects—for instance, Middle East shocks that boost U.S. output through oil and safe-haven channels.

This work has been presented at multiple research and policy fora and has generated a number of follow-up projects, including the development of the Global Bilateral GPR Database. For a non-technical overview, articles in SUERF or e-axis. 

[1] Rethinking GPR: the Sources of Geopolitical Risk, I Alonso-Alvarez, M Diakonova, JJ Pérez, Banco de España Working Paper, n. 2522 (2025)

[2] Geopolitical Risk: a Database of General and Bilateral Measures, I Alonso-Alvarez, E Bukina, M Diakonova, N Khitarishvili, JJ Pérez, P Piqueras. Draft available upon request

We develop a new way to measure elite political polarization and legislative gridlock across European countries. While most existing studies focus on the U.S. or use infrequent survey data, we build monthly, press-based indicators for France, Germany, Spain, and Italy. Using dictionary-based text analysis of national newspapers we constructs two distinct indices: a Political Polarization Index (capturing ideological division) and a Legislative Gridlock Index (capturing institutional deadlock).

Our results show that political polarization in Europe has been on a rise since the Global Financial Crisis, though its institutional consequences differ by country. In France and Germany, polarization is tightly linked to legislative paralysis, often during periods of social unrest or migration-related debates. In Spain, polarization has surged since 2016, reflecting electoral fragmentation and the Catalan crisis. Italy stands out: despite increasing ideological division, reforms to the electoral system have limited the extent of gridlock.

By explicitly distinguishing between polarization and its institutional consequences we provide a high-frequency, comparable tool for studying the interaction between politics and the economy in Europe. We are currently extending this work to include more countries and capture further narrative nuances with more sophisticated NLP techniques.

[3] Political polarization in Europe, M Diakonova, C Ghirelli, JJ Pérez. Banco de España Working Paper, n. 2533 (2025) → Media impact: SUERF Policy Brief; Real Instituto Elcano; Business Insider, El Confidencial, El Español, El Comercio, La Razón, La Vanguardia. 

Figures shown as they appear in the SUERF Policy Brief.

Here we revisit how economists measure uncertainty, a concept that has become central to understanding economic cycles and policy transmission. Our starting point are the indices developed by Andrés Azqueta-Gavaldón, who uses LDA to find the topics that make up the discussion around economic uncertainty. We recompute these topics for the European countries and group them into main non-overlapping themes —politics, policy, markets, society, and others. We then compare them with established uncertainty proxies such as the Economic Policy Uncertainty (EPU) index and the ECB’s Country-Level Index of Financial Stress (CLIFS). The results reveal clear differences between how the press and markets perceive uncertainty. Political and policy-related topics dominate the EPU, while market and corporate themes explain CLIFS. Yet during major crises—such as 9/11, the Global Financial Crisis, or the Russia–Ukraine war—both indices spike together, reflecting systemic shocks that bridge political and financial channels. By decomposing uncertainty in this way, the paper provides a framework for understanding how different forms of uncertainty interact and propagate across countries.

[4] Sources of economic policy uncertainty in the euro area: a ready-to-use database, A Azqueta-Gavaldón, M Diakonova, C Ghirelli, JJ Pérez, Banco de España Occasional Paper, n. 2315 (2023)

[5] The Anatomy of Economic Uncertainty: Deconstructing Proxy Measures through News Narratives, A Azqueta-Gavaldón, M Diakonova, C Ghirelli, JJ Pérez, GT Larrauri. Draft available upon request