Conference Agenda

Timely assessment of economic activity is crucial for effective policymaking at the national, regional, and global levels. However, many economies still do not publish GDP data at a quarterly basis, creating persistent information gaps. In 2025, 34% of economies publish only annual GDP statistics. This lack of higher‑frequency and timely data is particularly restrictive for emerging market and developing economies, where economic volatility and spillover risks are often highest. The problem is more severe for historical data: only 42% of economies have quarterly GDP estimates for a period longer than 20 years. To address these gaps, this paper develops a model that estimates missing quarterly GDP series by leveraging global and regional economic interconnections. The method transforms sparse annual data into quarterly estimates by exploiting higher-frequency information from the rest of the world, enabling real-time policymaking in both data-scarce economies and in global-level discussions. Moreover, this method ensures internally consistent estimates of regional and global economic activity, allowing both top‑down and bottom‑up scenario analyses.

We use web-search data to construct indices that proxy the derived demand for metals - specifically cobalt, lithium, and nickel - which are key inputs in technologies driving the energy and digital transition. These indices are incorporated into monthly Structural Vector Autoregressive (SVAR) models of the global markets for these metals. Identification of structural shocks relies on a combination of zero, static and dynamic sign restrictions, allowing us to disentangle supply shocks from multiple demand-side shocks that drive the real price of metals. In particular, we isolate a demand component linked to the technological uptake of metals in the energy and digital transition. Our framework enables a quantitative assessment of the relative contribution of each structural driver to price dynamics and highlights the growing macroeconomic relevance of technology-linked metal demand.

This paper develops a new approach to characterize regional business cycle dynamics using high-dimensional state-level data. I employ a three-way tensor decomposition subject to orthogonal time components and non-negative factor loadings to estimate the latent regional cycles. The constraints act as inductive biases that resolve scale, sign, and rotation indeterminacy and produce what is known in the machine learning literature as a “parts-based” representation of the data. A model with four factors captures over 80% of the variation in the data while reducing its dimensionality by more than 90%. The estimated regional cycles correspond to distinct economic forces and align closely with well documented regional specializations. These patterns emerge directly from the data without the use of covariates or geographic clusters.