This study examines how fintech lending platforms affect local labor market outcomes among small businesses. Using county-level data on Lending Club small-business loan activity from 2008 to 2020, we find that greater platform lending is associated with lower job creation, reduced employment, and depressed worker earnings among small businesses. To address endogeneity concerns, we exploit the staggered regulatory approval of Lending Club across U.S. states as a quasi-natural experiment and implement a stacked difference-in-differences design with entropy balancing. The results show that expanded access to Lending Club credit leads to significant declines in small-business labor market outcomes. Mechanism tests indicate that these adverse effects are not driven by borrower risk selection. Instead, they operate mainly through debt-trap behavior, whereby borrowers take on platform debt that their businesses struggle to service, and strategic debt restructuring, whereby borrowers use unsecured platform loans to refinance existing business debt rather than fund productive investment. Additional analyses show that the effects are robust to controls for Prosper approval and local house prices, are concentrated in the retail-investor-dominated period and persist in counties with lower sole-proprietorship density. Overall, the findings highlight a real-economy cost of fintech lending platforms.
Quarterly-updated asset growth (AGQ) positively predicts subsequent stock returns in China. This positive premium contrasts sharply with the well-documented negative relation between annually-updated asset growth and returns in other markets, while annually-updated asset growth does not predict average returns in China. The AGQ premium is robust across measurement horizons and is not driven by a particular fiscal quarter. Higher AGQ predicts stronger future profitability, while returns on AGQ-sorted portfolios forecast future GDP growth, relating the premium to firm fundamentals and macroeconomic investment opportunities. Moreover, successive AGQ updates revealed by financial reports predict returns beyond the previously available AGQ. These findings show that the same accounting measure could produce different return patterns when updated at different frequencies.
We show that credit rating reports contain valuable information beyond the letter rating itself. Using 3352 Moody's rating action reports, we analyze clarity, length, numerical detail, uncertainty, and uniqueness and find that the stock market reacts to rating content and that this content predicts future rating actions. Analyst fixed effects account for about 20% of the variation in report style, indicating that individual analysts matter but do not dominate. Following the Dodd-Frank Act, reports became longer and clearer, especially when ratings diverged from expectations, consistent with heightened litigation risk. Our findings indicate that the words behind the letter are associated with stock market reactions and change systematically with regulation.
This study examines whether information from Bitcoin perpetual futures improves forecasts of Bitcoin spot volatility. Using high-frequency intraday data, we construct two types of predictors from perpetual futures: funding-rate-based measures and realized volatility measures. These predictors are incorporated into a Heterogeneous Autoregressive model to forecast daily, weekly, and monthly spot volatility. Both in-sample and out-of-sample results show significant improvements in volatility forecasts, particularly from funding rates and at longer forecast horizons. This is consistent with funding rates capturing imbalanced leveraged positions whose subsequent unwinding and the associated price fluctuations unfold over extended periods. Our results underscore the distinct informational value of perpetual futures for forecasting Bitcoin volatility.
Contrary to the conventional view that unionized firms are associated with less disclosure, we find that U.S. firms with organized labor provide earnings with better quality. We establish the causal relationship using a regression discontinuity analysis that exploits the random assignment around close-call union elections and through a difference-in-differences approach, utilizing the staggered enactment of right-to-work laws. The influence of organized labor is significantly stronger in firms with greater capacity to wield influence, enhanced incentives to monitor, and in contexts where alternative information sources are limited. The impact of organized labor on increasing labor-related costs is also significantly reduced in firms with better earnings quality. Overall, our research indicates that managers face a tradeoff between shielding shareholders from union-driven rent-seeking and building a trusting relationship with employees by accommodating employees' demand for useful information.
