New research demonstrates how neural networks and machine learning overcome two of the biggest barriers in econometric research: optimization of non-smooth functions and analysis of data that can never be jointly observed.

This newsletter curates the ten most-read pieces offering a lens through which to discern both the preoccupations that defined this year’s economic discourse and the fault lines likely to structure policy debate in the year ahead.

Two breakthrough papers demonstrate how advanced computational methods are making previously intractable macroeconomic models feasible to estimate. One approach integrates local linear solutions into filtering procedures for computational gains of several orders of magnitude, while another uses neural networks to estimate complex heterogeneous agent models.

This issue examines how economists can reclaim a vital role in policymaking by rebuilding public trust, engaging more deeply with real-world challenges, and communicating more transparently. Leading voices offer recommendations on how the economics profession  can tackle the 21st-century challenges.

Two new papers provide guidelines for using LLMs in empirical research, ensuring rigorous and reliable outcomes while the other leverages LLMs to simulate complex social interactions.

Two recent papers explore how cognitive limitations like bounded attention and memory affect people’s perceptions of risk and uncertainty as well as saving behavior and interpersonal trust.

This newsletter highlights the issues and research that attracted the greatest interest among our readership over the past twelve months.

Two recent papers explore the applications of deep learning in economics. One paper introduces practical methods for extracting structured information from unstructured data using deep neural networks, while the other demonstrates how deep learning can help solve high-dimensional dynamic equilibrium models.

Two new research papers that emphasize how field experiments not only address critical challenges in causal inference but also enhance the generalizability of findings, showcasing their enduring relevance and potential for future research in economics.

Recent research highlights issues with the forecasting accuracy of HANK models as well as issues with indeterminacy.