Research

This paper turns more than 40,000 hours of financial-TV dialogue from 2012 to 2026 into speaker- and topic-resolved Q&As, providing high-frequency, cross-sectional observations on sentiment, uncertainty, and dispersion in expressed beliefs. TV-based uncertainty measures display greater topic specificity than their newspaper counterparts, co-move more closely with matched market benchmarks, and expand topic coverage. Using LLMs, we score forward-looking statements along two dimensions: direction and expressed self-uncertainty. The resulting topic-specific disagreement measures align systematically with disagreement-based benchmarks, but their relation with other uncertainty benchmarks is topic- and period-specific. The paper then studies disagreement in predictions about the future policy rate. Controlling for differences in macro projections, a mismatch in speakers' interpretations of the Fed's priorities is associated with a roughly six percentage point higher probability of a large rate-path disagreement, on par with a one-standard-deviation increase in inflation-projection disagreement. The result does not weaken when restricting the comparison to explicit conditional Fed statements and survives a battery of robustness tests. TV discourse is also rich in mental-state attributions. Fed-attribution language shows a tighter relation with market-based rate-uncertainty measures than both our own text-based uncertainty measure and standard text-based alternatives.

Post-GFC regulations are cited as a driving force behind banks' retreat from the mortgage market and the ensuing rise of nonbank lending. Whether banks facilitated this migration remains unexplored. Exploiting government lawsuits that affected the Federal Housing Administration (FHA) market but not the conventional market, and controlling for borrower-time and lender-borrower fixed effects, we find that affected banks increased their funding of nonbanks after settlement. We combine warehouse-credit data with mortgage originations and impute affected banks' indirect footprint. We find that both imputed FHA volume and the FHA share increase following settlement. The increase in the indirect FHA share is comparable in magnitude to the corresponding decrease in affected banks' direct FHA share. At the county level, greater warehouse funding from affected banks is associated with an increase (decrease) in nonbanks' share of FHA (conventional) originations.

In addition to stable core deposits, many U.S. banks also rely on non-core insured funding—especially brokered deposits (BDs)—which supervisors and analysts often associate with risk-taking and run-proneness. In sharp contrast to the "hot money" view, BD funding remains available and scales in system-wide stress episodes, including the 2008–09 crisis and the 2023 banking turmoil, and serves as a source of liquidity for vulnerable banks facing uninsured deposit outflows. However, banks with higher BD shares appear more fragile on outcomes—they experience sharper uninsured deposit outflows and weaker equity performance in shocks, especially during the 2023 banking turmoil. The adverse outcomes are concentrated among institutions that increased BD reliance on the eve of the turmoil, suggesting the outcomes reflect self-selection and negative market inference.

Financial crises are traditionally analyzed as purely economic phenomena. The political economy of financial booms and busts remains both under-emphasized and limited to isolated episodes. This paper examines the political economy of financial policy during ten of the most infamous financial booms and busts since the 18th century, and presents consistent evidence of pro-cyclical regulatory policies by governments. Financial booms, and risk-taking during these episodes, were often amplified by political regulatory stimuli, credit subsidies, and an increasing light-touch approach to financial supervision. The regulatory backlash that ensues from financial crises can only be understood in the context of the deep political ramifications of these crises. Post-crisis regulations do not always survive the following boom. The interplay between politics and financial policy over these cycles deserves further attention.