Research

Publications

In Search of the True Greenium (2026)

with Marc Eskildsen, Markus Ibert, and Lasse H. Pedersen

Forthcoming, Journal of Financial Economics
Code: Implied cost of capital estimates from around the world

Abstract:The greenium (the expected return of green securities relative to brown) is a central impact measure for ESG investors. Replicating the literature’s wide range of greenium estimates based on realized returns, we find that these are not robust to changing the greenness measure or time period. Instead, we propose a robust green score combined with forward-looking expected returns, yielding a more precisely estimated annual equity greenium of -25 basis points per standard deviation increase in greenness. The greenium is more negative in greener countries and over time. Finally, we provide greeniums for corporate bonds, weighted-average costs of capital, and sovereign bonds.

Machine Learning and the Implementable Efficient Frontier (2025)

with Bryan Kelly, Semyon Malamud, and Lasse H. Pedersen

Review of Financial Studies
Code

Abstract:We propose that investment strategies should be evaluated based on their net-of-trading-cost return for each level of risk, which we term the "implementable efficient frontier." While numerous studies use machine learning return forecasts to generate portfolios, their agnosticism toward trading costs leads to excessive reliance on fleeting small-scale characteristics, resulting in poor net returns. We develop a framework that produces a superior frontier by integrating trading-cost-aware portfolio optimization with machine learning. The superior net-of-cost performance is achieved by learning directly about portfolio weights using an economic objective. Further, our model gives rise to a new measure of "economic feature importance."

Is There a Replication Crisis in Finance? (2023)

with Bryan Kelly and Lasse H. Pedersen

Journal of Finance,
Winner of the Dimensional Fund Advisor Prize distinguished paper
Code: Learn how to work with global return and accounting data and how to use empirical Bayes to estimate a hierarchical model of global equity factor returns
Data: Stock returns, firm characteristics, and equity factors from 93 different countries

Abstract:Several papers argue that financial economics faces a replication crisis because the majority of studies cannot be replicated or are the result of multiple testing of too many factors. We develop and estimate a Bayesian model of factor replication, which leads to different conclusions. The majority of asset pricing factors: (1) can be replicated, (2) can be clustered into 13 themes, the majority of which are significant parts of the tangency portfolio, (3) work out of sample in a new large data set covering 93 countries, and (4) have evidence that is strengthened (not weakened) by the large number of observed factors.

Working Papers

A Market-Based Cost of Capital (2026)

with Niels Gormsen and Kilian Huber

Abstract:

A firm's cost of capital is a central object in economics. It is often defined based on the expected long-run returns on the firm's assets in financial markets, but these expected returns are unobserved and there exists no method that delivers unbiased estimates out of sample. We provide a method that uses machine learning to predict future cash flows of firms and then derives long-run expected returns from observed market prices and the present value identity. We show that this method works better than predicting returns directly if the model determining cash flows is more stable than that determining expected returns, a condition that holds empirically. The resulting estimates are unbiased out of sample and have greater predictive power than existing alternatives. We use them to construct a firm-level ``market-based cost of capital'' (MCC). Consistent with the standard prediction that firms maximize their market value by using expected returns as their cost of capital, we find that firms whose perceived cost of capital is closer to the MCC have higher market values.

Prediction Market Accuracy: Crowd Wisdom or Informed Minority? (2026)

with Roberto Gomez-Cram, Yunhan Guo, and Howard Kung, R&R at The Review of Financial Studies

Abstract:

Prediction markets produce remarkably accurate forecasts, but the source of this accuracy is poorly understood. Two explanations dominate: crowd wisdom and insider trading. Using the universe of Polymarket transactions, we show it is neither. Instead, accuracy comes from a minority of persistently skilled traders, around 3% of accounts. Unlike insiders, whose private edge is localized, these traders exhibit depth and breadth. They react to public news when it arrives, eliminate law-of-one-price violations, and trade against the crowd’s behavioral mistakes. The crowd, in turn, generates most of the volume but little of the information, and its losses fund the minority’s profits.

Financial Prediction Markets: A New Measure of Earnings Expectations (2026)

with Roberto Gomez-Cram, Yunhan Guo, and Howard Kung

Abstract:

We construct a high-frequency measure of earnings expectations using financial prediction markets. Unlike analyst forecasts, which are updated infrequently and prone to agency conflicts, prediction market prices reflect real-time, stake-backed beliefs. Each contract pays one dollar if realized earnings exceed the analyst consensus; we derive the conditions under which these prices represent subjective probabilities and introduce a methodology to convert them into implied expectations about earnings. Relative to analyst forecasts, we find that market-implied expectations are (i) more accurate; (ii) incrementally informative for earnings announcement returns; and (iii) significantly less biased, though they exhibit short-term overreaction in contrast to the underreaction typical of analysts. Our current sample covers the market's inception from September 15, 2025, through November 15, 2025; we will regularly update this analysis as new contracts resolve. We expect this measure to become increasingly informative as the market matures, liquidity deepens, and financial prediction markets become part of mainstream finance.

The Realized Return to Green Investing: A Global Replication (2026)

Marc Eskildsen, Markus Ibert, and Lasse H. Pedersen

Abstract:

We examine whether realized stock returns provide reliable evidence on the expected returns to green investing. The literature reports conflicting findings, with both significant green and brown outperformance. We conduct a comprehensive scientific replication that spans a broad set of greenness measures, portfolio constructions, risk models, and countries, covering the ex ante plausible empirical design space. Across these specifications, realized green-minus-brown returns are statistically insignificant once we take multiple testing into account. The distribution of published t-statistics differs significantly from the replicated distribution, consistent with publication bias. Finally, we show that realized returns are too noisy and samples too short to identify the greenium.

Complex Modern Portfolio Theory (2026)

with Oliver Hellum, Bryan Kelly, and Semyon Malamud

Abstract:

The literature has long wrestled with the practical usefulness of Modern Portfolio Theory (MPT), and extensive evidence shows its performance decreases rapidly with the number of assets (N ). We present several new and counterintuitive facts about MPT. Most importantly, the performance of MPT in fact increases with N once the number of assets exceeds the number of training observations (T ). This finding holds in a variety of settings: in conjunction with popular portfolio regularization methods, in a variety of asset universes, and when T is large or small.

The Power of the Common Task Framework (2025)

with Oliver Hellum, Bryan Kelly, and Lasse H. Pedersen

Abstract:

The “Common Task Framework” (CTF) is a collaborative and competitive process in which researchers solve a task using shared data, a predefined success metric, and a leaderboard. Using an economic model, we show that the CTF incentivizes effort, increases innovation, and curbs misrepresentation by reducing research costs and improving comparability. Historical examples from computer science underscore its effectiveness. To demonstrate its broader applicability, we propose a CTF for financial economics: a platform open to all researchers designed to identify the pricing kernel and systematically evaluate asset pricing models, from traditional factor-based approaches to modern machine learning techniques.

Subjective Risk and Return (2025)

Abstract:
Traditional asset pricing models like the CAPM explain realized returns worse than newer asset pricing models like Fama-French-5, but why? I show that traditional models explain subjective risk and return expectations well but also predict return-subtracting mispricing. Newer models, by contrast, explain subjective risk and return expectations poorly but predict return-enhancing mispricing. These results imply that the CAPM is a good model of risk but fails to explain realized returns because risk is correlated with mispricing. I explain this disconnect by a model in which all the CAPM assumptions hold, except that some investors have an optimism bias.

Page updated

Google Sites

Report abuse