Working Papers
Working Papers
The Intersection of Expected Returns [SSRN]
Abstract:
A relatively small number of stocks plays a disproportionately large role in explaining the premia of 164 cross-sectional asset pricing anomalies. For instance, excluding the top 10% of stocks that are shared across the most anomaly portfolios for a given month reduces the average anomaly's return and alpha by approximately 40%. These stocks can be identified ex ante and used to form long-short portfolios that generate abnormal returns more than three times larger than that of the average anomaly portfolio. I provide evidence that the returns of these stocks reflect mispricing due to biased investor expectations. Overall, my analysis suggests that (i) anomalies are largely not independent of one another and (ii) that they likely reflect theoretically arbitrageable mispricing, rather than exposures to latent systematic factors.
Presentations: Eastern Finance Association Annual Meeting, Financial Management Association Annual Meeting, American Finance Association Annual Meeting (poster session), Chicago Quantitative Alliance Spring Conference (scheduled), University of Arizona
Accolades: SSRN Top 10 Papers Lists (Behavioral & Experimental Finance eJournal, Capital Markets: Market Efficiency eJournal, Capital Markets: Asset Pricing & Valuation eJournal, Financial Economics Network Subject Matter eJournals, Financial Economics Network eJournal), AFA Travel Grant⁽$⁾
⁽$⁾ Monetary award
Political Polarization and Stock Market Expectations [SSRN]
with Marco Angrisani, Richard Sias, and Harry Turtle
Abstract:
Political polarization explains much of the heterogeneity in how individuals' perceptions of the distribution of stock returns evolves over time: individuals are more optimistic when "their" party occupies the White House and become more pessimistic when the "other" party takes office. Polarization influences perceptions of near-term returns, long-term returns, and tail returns. Polarization has explanatory power beyond party or ideological affiliation alone. The effect is particularly pronounced for Republicans and has grown over time; by 2025, a one-standard deviation shift in polarization has a greater effect size than a five-standard deviation shift in income.
Presentations: University of Arizona, University of Toledo
Accolades: SSRN Top 10 Papers Lists (Behavioral & Experimental Finance eJournal), SSRN Editor's Choice (Behavioral & Experimental Finance eJournal)
Target Date Fund Glide Paths: Do Active Bets Enhance Retirement Outcomes? [SSRN - link coming soon]
with David Brown and Shaun Davies
Abstract:
Target Date Fund providers claim to provide value for investors through glide path design, implementation, and tactical asset allocations. We study time series variation in glide paths arising from these discretionary adjustments—what we call Glide Path Activeness. We document significant heterogeneity in Glide Path Activeness in the cross section (between TDFs, TDF series', and TDF providers) and in the time series. We provide evidence that Glide Path Activeness does not add value for investors—in the worst case, we estimate an annualized cost of 1.38% on a risk-adjusted basis. We find that this underperformance is due primarily to poor timing of asset classes and return chasing rather than the selection of more expensive underlying funds or larger management fees. Our results have implications for investors, plan sponsors, and policy makers.
Presentations: Investment Company Institute, University of Kentucky*, Vanguard*, Arizona State University*, Australian National University*, Baylor University*, Claremont McKenna College*, University of Arizona*, University of Hawaii at Manoa*, University of Melbourne*, University of New South Wales*, University of Sydney*, University of Tennessee*, Villanova University*
* Presented by coauthor
Accolades: Investment Company Institute Paper Acceptance Award⁽$⁾
⁽$⁾ Monetary award
Coverage: Rational Reminder Podcast
Political Polarization and Partisan Consumption Cycles [SSRN]
with Nathan Fernig and Richard Sias
Abstract:
Partisanship and political polarization have a substantial impact on discretionary consumption: counties aligned with the party of the sitting president increase consumption significantly more than unaligned counties. The effect size of political alignment is comparable to that of per capita income and population growth and has grown over time as Democrats and Republicans have become increasingly polarized. We show that the consumption difference between politically aligned and unaligned counties persists throughout the presidential term, consistent with a channel through which expectations influence consumer behavior and propagate through local economic activity. These findings highlight how political affiliation and polarization influence economic behavior at the local level, with broader implications for macroeconomic dynamics and both theoretical and empirical finance.
Presentations: Chicago Quantitative Alliance Spring Conference (scheduled)*, University of Arizona
* Presented by coauthor
Do Characteristics Proxy for Latent Factors? Evidence From Characteristic Spreads
with Chenyanzi Yu
Abstract:
Spreads in characteristics between the long and short legs of anomaly portfolios (e.g., the spread in price-to-earnings ratios in the Value anomaly) do not consistently predict future anomaly returns. This finding implies that, on average, individual characteristics are weak proxies for latent factors. A key contribution of this paper is to formally establish a general mathematical link between proxy quality and return predictability.
Presentations: Financial Management Association (early ideas session), University of Arizona
Losing the Trees for the Forest: Strong Factors, Weak Factors, and Statistical Estimates of the Pricing Kernel
Abstract:
The detectability of latent factors is closely tied to their signal-to-noise ratios. By first extracting dominant factors—such as the market factor—both principal component analysis and autoencoder neural networks more effectively recover weaker latent structures. Empirically, statistical estimates of the pricing kernel differ substantially when derived from raw data versus from data residualized on the market factor.
Presentations: University of Arizona