Lea en español.
On January 28, 2025, the School of Economics and Data Hub at San Francisco University of Quito (USFQ) unveiled the U-Index. This new tool seeks to convey broad-based public perceptions and reflects “the intensity of economic uncertainty at a specific moment,” with an emphasis on economic policy.
Led by economics professors Pedro Romero and Rodrigo López, the U-Index analyzes 3.2 million news articles from 2001 to the present, providing a comprehensive view of how uncertainty perceptions have evolved over time. The project also benefited from the collaboration of computer-science students and the university’s newspaper archive.
Rodrigo López, who holds a master’s degree in economics and econometrics from the University of York in England, explained in an interview with the Impunity Observer: “In US markets, there is an abundance of data helping economic agents make decisions.” However, in Ecuador—a dollarized economy with fixed interest rates—these numbers do not always reflect the reality on the ground.
Economic indicators such as the Emerging Markets Bond Index (EMBI) measure financial risk on a global scale. López, nevertheless, noted that these indicators have limitations when applied locally.
Building the Uncertainty Index
Pedro Romero, who holds a PhD in economics from George Mason University, highlighted the large amount of data used to build the U-Index. The team aggregated 2,374,543 articles from printed newspapers—such as El Universo, El Comercio, and El Telégrafo—alongside 247,325 articles from digital media outlets like Diario Expreso and Primicias. These articles, focused on economic policy and related topics, were then subjected to cutting-edge analysis.
From print media, the U-Index team took 761,888 images, containing various articles within. To extract meaningful information, they applied Optical Character Recognition (OCR) technology, which converts scanned documents into digital text, and Detectron 2, a machine learning tool that identifies key data within images. Roboflow and Data Augmentation allowed them to categorize images based on keywords such as economy, politics, uncertainty, and related terms.
Despite the data’s complexity and enormous breadth, the processing proved successful, with an average time of 20 seconds per image. That means the entire data processing took approximately four weeks. The reliability of the results was validated by 160 students, who reviewed a random sample of 5,345 articles. This review process ensured an accuracy rate of 91 percent. Digital articles were processed with less difficulty, given their easier accessibility and lack of need for conversion.
What the U-Index Offers
The U-Index offers dynamic graphs that allow users to analyze trends over different time frames—ranging from six months to a decade. In addition, a “context” feature correlates the uncertainty indicator with significant historical events in Ecuador, providing a deeper understanding of how political, social, and economic conditions influence public perceptions.
The U-Index is based on the frequency of specific keywords, and it includes a confidence interval to ensure greater confidence in its results.
López asserted: “Uncertainty is a dynamic process, shaped by the interaction between available economic data and the opinions of specialized agents.” He then highlighted that the U-Index has the ability to track how uncertainty fluctuates around pivotal events.
López referred to two notable examples:
- Peaks in uncertainty often coincide with presidential elections. However, the U-Index does not establish causality; it simply indicates a correlation.
- When a Constitutional Assembly was developing Ecuador’s current constitution, uncertainty levels grew steadily. Once the new constitution was officially enacted in 2008, uncertainty declined. Regardless of framing a favorable or unfavorable business environment, economic agents gained clarity about what to expect moving forward.
The Future of U-Index
During a Q&A session, a common question was whether social-media content could be incorporated into the U-Index. While social-media platforms offer vast amounts of real-time data, experts warned that including them could undermine the integrity of the index. As a result, the team suggested a separate index specifically focused on social-media trends.
Despite this challenge, the U-Index creators seek to offer a valuable complement to global economic indicators like the EMBI. The U-Index is uniquely designed to reflect Ecuador’s economic realities, offering a tool that is both specific and dynamic.
Francisco Zalles, an Ecuadorian economist and professor at Hespérides University in Spain, has argued: “the U-Index is a valuable data tool that will raise the quality of economic research in the country. For instance, analysts can identify how social factors such as unemployment and violence relate to uncertainty.”
Zalles adds another potential application of the U-Index: “In Ecuador, we can frame the economy through fiscal policy and regulations. The U-Index could help us to measure the efficiency of policies and become the base for economic policymaking.”
