Autoregressive LLMs generate text by sampling from estimated probability distributions over the next token, conditional on prior context. We use these probabilities to construct an entropy-based ...

Organized interests are thought to influence policy, but whether and when interest group money overrides public opinion remains poorly understood. We investigate how gun interest group money and ...

However, a party can sustain extremism only if the other side is extreme, too. A small moderation of one side’s voting electorate can trigger a discontinuous collapse of candidate extremism on both ...

We construct a novel dataset linking academic publication records to U.S. Census employer–employee data to track 42,000 AI researchers over two decades. We document systematic changes in the ...

We investigate how fertility relates to work from home (WFH) in the post-pandemic era, drawing on original data from our Global Survey of Working Arrangements and U.S. Survey of Working Arrangements ...

Research Associate Kosali Simon, the Herman B. Wells Professor and Distinguished Professor at the Paul H. O'Neill School of Public and Environmental at Indiana University, will direct the NBER Program ...

We use high-frequency retail microdata to measure the short-run impact of the 2025 U.S. tariffs on consumer prices. By matching daily prices from major U.S. retailers to product-level tariff rates and ...

We examine two approaches to improving urban school systems: changing who gets to go to existing schools (reallocation) and restructuring school portfolios through closures and reconstitution ...

Using rich Swedish administrative data, we apply causal machine learning methods to study how earnings losses after job displacement vary with observable characteristics that may be relevant for ...

Do firms respond similarly to corporate tax incentives across countries? We provide globally comparable estimates of the corporate elasticity of taxable income using administrative tax return data ...

Useful social science theories predict behavior across settings. However, applying a theory to make predictions in new settings is challenging: rarely can it be done without ad hoc modifications to ...

This paper shows that the pace of technology creation is a key driver of the skill premium. It develops a model in which skilled workers have a comparative advantage in learning new technologies. As ...