RISE Finance Lab

Abstract: We utilize high-frequency data from two major Australian online sports betting platforms to investigate how bettors react to past outcomes and temporary income windfalls. Bettors stake significantly more after wins and are more likely to return after both wins and losses. These patterns are strongest among individuals flagged as at-risk of gambling harm. Exploiting the timing of annual tax refunds, we find that at-risk bettors relatively increase their stake size by 13% following the refund. Voluntary harm-minimization tools do not significantly dampen this response. Our findings highlight how liquidity amplifies feedback-driven risk-taking and underscore the limitations of self-regulation in managing risky gambling behavior.

Abstract: The paper develops a factor-based framework to measure the exposure of US banks to physical climate risk. Using monthly temperature data, the authors estimate that the US banking sector’s systemic physical climate risk peaked at approximately $500 billion, equivalent to roughly one-quarter of the sector’s reported equity, by the late 2010s. Exposed banks also reduce lending to small businesses, especially in counties where they lack branches. The framework offers a scalable, disclosure-free measure of physical climate risk, with direct implications for supervision and stress testing.

Abstract: Exploiting the feature that central bank governors in many economies serve fixed terms with predetermined turnover dates, while a successor’s identity and policy stance remain unknown until shortly before appointment, this paper studies how turnovers affect global bank credit across 43 countries. Cross-border lending rises 11.8% before a turnover, while total lending is unchanged, indicating a reallocation; lending reverses once the new governor takes office. Policy rates are stable while economic policy uncertainty and interest rate volatility rise. Effects are larger at central banks with supervisory authority, isolating governor turnovers as a distinct, plausibly exogenous source of economic uncertainty.

Abstract: We examine shareholder responses to major environmental and social incidents through director elections. Our findings show that directors at firms experiencing such incidents receive 9.8% more negative votes. Shareholder dissent reduces the likelihood of future ES incidents when boards implement governance responses such as adopting ES-linked pay or enforcing CEO turnover. The results also suggest that shareholders substitute voting against directors with support for ES-related proposals. Overall, our study underscores that the effectiveness of shareholder voting is conditional on complementary board governance responses, with the two forces jointly driving improvements in future ES performance.

Date: Journal of Financial Economics, Vol. 164 (2025)

Abstract: We study the impact of disclosure and inattention on the decision to retain fee-based financial advice using a two-tiered natural regulatory experiment. Increased salience in fee disclosure raises the drop rate for advice, implying improved attention — particularly for relatively sophisticated investors. However, a novel auto-drop requirement for inattentive investors generates far more drops, implying limited attention despite salient disclosure — particularly for the unsophisticated. Contrary to studies of commission-based advice, we find that investors benefit from fee-based advice. Benefits are higher for less sophisticated investors, who tend to be detrimentally auto-dropped. Drops triggered by salient disclosure tend to be beneficial.

Date: Review of Finance, Forthcoming

Abstract: In Australian real estate markets, about a third of properties are sold at auction. We show that properties that fail auctions sell later for a 2.6% discount. This effect increases for properties failing multiple auctions and when no bids are made. Consistent with a causal channel, the effect holds when auction failure is instrumented by the tendency of owners to anchor on nearby better properties (and thus set reserve prices too high). Prices cluster just below salient round numbers, and the discount fades over time, inconsistent with our effects reflecting unobserved property characteristics. We test for several mechanisms and conclude that most of the pricing discounts reflect stigma, which reduces potential buyers’ willingness to pay.