Felix Reichenbach
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Felix Reichenbach has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Martin Walther
Technische Universität Berlin
6 shared publications - Alexander Karlberg
European Organization for Nuclear Research
2 shared publications - Andrea Banfi
University of Sussex
2 shared publications - Barbara Jäger
University of Tübingen
2 shared publications - Giulia Zanderighi
Max-Planck-Institut für Physik
2 shared publications - Silvia Ferrario Ravasio
Theoretical Physics Department, CERN
2 shared publications
Latest Publications
- Open Access
- 15 October 2026
- Published by Elsevier in SSRN Electronic Journal
Multi-Asset Portfolio Selection with Machine Learning: Long-Run International Evidence
This study examines whether machine learning improves asset-class return forecasts, whether crosscountry information pooling further improves accuracy, and whether these gains also lead to better portfolio decisions. Using annual data for 16 countries covering equities, long-term bonds, bills, and housing from 1870 to 2020, we combine historical returns with macroeconomic and financial variables in a recursive forecasting framework. Pooled machine-learning forecasts significantly outperform both historical-mean forecasts and country-specific ML. These gains translate into higher Sharpe ratios for both international investors allocating across countries and domestic investors investing only within their home markets. Pooled ML also outperforms simple allocation rules such as equal weighting. The passive benchmark based on capital-weighted aggregate asset returns remains competitive in the baseline analysis, but pooled ML performs better when the universe is restricted to liquid asset classes or portfolio limits are relaxed. The results extend the machine-learning literature on long-run multi-asset allocation and highlight the value of pooling information across markets.
- 15 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessMulti-Asset Portfolio Selection with Machine Learning: Long-Run International Evidence
This study examines whether machine learning improves asset-class return forecasts, whether crosscountry information pooling further improves accuracy, and whether these gains also lead to better portfolio decisions. Using annual data for 16 countries covering equities, long-term bonds, bills, and housing from 1870 to 2020, we combine historical returns with macroeconomic and financial variables in a recursive forecasting framework. Pooled machine-learning forecasts significantly outperform both historical-mean forecasts and country-specific ML. These gains translate into higher Sharpe ratios for both international investors allocating across countries and domestic investors investing only within their home markets. Pooled ML also outperforms simple allocation rules such as equal weighting. The passive benchmark based on capital-weighted aggregate asset returns remains competitive in the baseline analysis, but pooled ML performs better when the universe is restricted to liquid asset classes or portfolio limits are relaxed. The results extend the machine-learning literature on long-run multi-asset allocation and highlight the value of pooling information across markets.
- Open Access
- 2 October 2026
- Published by Elsevier in SSRN Electronic Journal
When Can Prediction Market Prices Be Trusted? Price Informativeness, Calibration, and Bias on Polymarket
Prediction markets provide real-time probability estimates that are increasingly used by managers, analysts, journalists, and investors. Using 1.5 billion trades across 779,000 Polymarket markets, we examine when these prices can be trusted as probability estimates. Overall, prices are highly informative and improve with market maturity, with higher trading activity associated with better forecasts, particularly at early stages. However, accuracy varies across categories: weather markets are most precise, whereas mentions and geopolitics markets are least precise. We also find no general market-level longshot bias. Instead, low-probability outcomes are initially underpriced before reversing late in the lifecycle. Default outcomes are persistently overpriced. In sports markets, Polymarket matches professional bookmakers' accuracy. However, live trading shows that rapid information arrival can generate temporary distortions even as overall forecast accuracy improves. Prediction market prices therefore offer valuable probability estimates, but their interpretation should account for when, where, and under what trading conditions prices are observed.
- 2 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessWhen Can Prediction Market Prices Be Trusted? Price Informativeness, Calibration, and Bias on Polymarket
Prediction markets provide real-time probability estimates that are increasingly used by managers, analysts, journalists, and investors. Using 1.5 billion trades across 779,000 Polymarket markets, we examine when these prices can be trusted as probability estimates. Overall, prices are highly informative and improve with market maturity, with higher trading activity associated with better forecasts, particularly at early stages. However, accuracy varies across categories: weather markets are most precise, whereas mentions and geopolitics markets are least precise. We also find no general market-level longshot bias. Instead, low-probability outcomes are initially underpriced before reversing late in the lifecycle. Default outcomes are persistently overpriced. In sports markets, Polymarket matches professional bookmakers' accuracy. However, live trading shows that rapid information arrival can generate temporary distortions even as overall forecast accuracy improves. Prediction market prices therefore offer valuable probability estimates, but their interpretation should account for when, where, and under what trading conditions prices are observed.
- Open Access
- 17 September 2026
- Published by Springer Nature in Electronic Markets
- Vol. 36 (1) , 77
Leveraging crowd wisdom: Can critical online communities predict post-campaign failure in crowdfunding?
Failure to deliver on earlier promises undermines trust in creators and platforms, threatening trust-based digital business models such as crowdfunding. Drawing on research on information asymmetry, collective wisdom, social media discourse, and platform design, we investigate the ability of online communities to predict post-campaign failure. Our results show that campaigns discussed in a critical community on Reddit before successful funding are subsequently more than twice as likely to fail to deliver. Additionally, these campaigns are more prone to delivery delays and quality issues. Disaggregating the crowd’s evaluation into different types of criticism, we find that campaign-related criticism, such as doubts about the campaign’s credibility or feasibility, has the highest predictive power. This suggests that pre-funding criticism on social media can serve as an early indicator of post-campaign failure. Our study contributes to research on crowdfunding and online information environments by demonstrating that criticism in external online communities can provide useful early warning signals. In doing so, we extend prior work that has mainly focused on platform-internal information and funding success rather than on post-campaign outcomes. Our findings also have practical implications for platforms and backers, as they suggest that external community signals can complement existing due diligence and early risk detection in crowdfunding.
- 17 September 2026
- Published by Springer Nature in Electronic Markets
- Vol. 36 (1) , 77
Open AccessLeveraging crowd wisdom: Can critical online communities predict post-campaign failure in crowdfunding?
Failure to deliver on earlier promises undermines trust in creators and platforms, threatening trust-based digital business models such as crowdfunding. Drawing on research on information asymmetry, collective wisdom, social media discourse, and platform design, we investigate the ability of online communities to predict post-campaign failure. Our results show that campaigns discussed in a critical community on Reddit before successful funding are subsequently more than twice as likely to fail to deliver. Additionally, these campaigns are more prone to delivery delays and quality issues. Disaggregating the crowd’s evaluation into different types of criticism, we find that campaign-related criticism, such as doubts about the campaign’s credibility or feasibility, has the highest predictive power. This suggests that pre-funding criticism on social media can serve as an early indicator of post-campaign failure. Our study contributes to research on crowdfunding and online information environments by demonstrating that criticism in external online communities can provide useful early warning signals. In doing so, we extend prior work that has mainly focused on platform-internal information and funding success rather than on post-campaign outcomes. Our findings also have practical implications for platforms and backers, as they suggest that external community signals can complement existing due diligence and early risk detection in crowdfunding.
- Open Access
- 23 April 2026
- Published by Elsevier in Journal of Economic Behavior & Organization
- Vol. 246
Decision models and online betting behavior: Comparing mean-variance, expected utility, and prospect theory
- 23 April 2026
- Published by Elsevier in Journal of Economic Behavior & Organization
- Vol. 246
Open AccessDecision models and online betting behavior: Comparing mean-variance, expected utility, and prospect theory
- Open Access
- 30 May 2025
- Published by Elsevier in Finance Research Letters
- Vol. 83
Skin in the game: The returns of digital assets from computer games
- 30 May 2025
- Published by Elsevier in Finance Research Letters
- Vol. 83
Open AccessSkin in the game: The returns of digital assets from computer games
- Open Access
- 1 January 2025
- Published by Elsevier in SSRN Electronic Journal
Has investor behavior changed due to the Internet? Examining the impact of short-term and high-frequency stock charts
Since the advent of the Internet, the investment landscape has changed substantially. We consider the growth of stock ownership and Internet usage in recent decades and summarize the major steps in the evolution of (online) trading. Based on these developments, we hypothesize that the increased prevalence of short-term and high-frequency charts alters investors' perceptions of past returns. We then empirically investigate the impact of different timeframes and frequencies depicted in stock charts on the behavior of retail investors and specifically the cumulative prospect theory (CPT) anomaly established by Barberis et al. (2016). However, we find little evidence that stock charts are a major driver of the anomaly. Instead, longer timeframes result in higher returns in most periods and these increase strongly in the 1990s. An alternative explanation for this increase could be the emergence of electronic retail trading, which brought with it a wave of new investors. Furthermore, we provide guidance to researchers and practitioners on which CPT specifications should be used in the context of asset pricing.
- 1 January 2025
- Published by Elsevier in SSRN Electronic Journal
Open AccessHas investor behavior changed due to the Internet? Examining the impact of short-term and high-frequency stock charts
Since the advent of the Internet, the investment landscape has changed substantially. We consider the growth of stock ownership and Internet usage in recent decades and summarize the major steps in the evolution of (online) trading. Based on these developments, we hypothesize that the increased prevalence of short-term and high-frequency charts alters investors' perceptions of past returns. We then empirically investigate the impact of different timeframes and frequencies depicted in stock charts on the behavior of retail investors and specifically the cumulative prospect theory (CPT) anomaly established by Barberis et al. (2016). However, we find little evidence that stock charts are a major driver of the anomaly. Instead, longer timeframes result in higher returns in most periods and these increase strongly in the 1990s. An alternative explanation for this increase could be the emergence of electronic retail trading, which brought with it a wave of new investors. Furthermore, we provide guidance to researchers and practitioners on which CPT specifications should be used in the context of asset pricing.
- Open Access
- 13 November 2024
- Published by Elsevier in Finance Research Letters
- Vol. 71
Which sustainable development goals favor crowdfunding success?
- 13 November 2024
- Published by Elsevier in Finance Research Letters
- Vol. 71
Open AccessWhich sustainable development goals favor crowdfunding success?
- Open Access
- 9 July 2024
- Published by Elsevier in Journal of Financial Markets
- Vol. 71
Robinhood, Reddit, and the news: The impact of traditional and social media on retail investor trading
- 9 July 2024
- Published by Elsevier in Journal of Financial Markets
- Vol. 71
Open AccessRobinhood, Reddit, and the news: The impact of traditional and social media on retail investor trading
- Open Access
- 12 June 2024
- Published by Springer Nature in Digital Finance
- Vol. 6 (3) , 501-542
Improving credit risk assessment in P2P lending with explainable machine learning survival analysis
Recent research using explainable machine learning survival analysis demonstrated its ability to identify new risk factors in the medical field. In this study, we adapted this methodology to credit risk assessment. We used a comprehensive dataset from the Estonian P2P lending platform Bondora, consisting of over 350,000 loans and 112 features with a loan volume of 915 million euros. First, we applied classical (linear) and machine learning (extreme gradient-boosted) Cox models to estimate the risk of these loans and then risk-rated them using risk stratification. For each rating category we calculated default rates, rates of return, and plotted Kaplan–Meier curves. These performance criteria revealed that the boosted Cox model outperformed both the classical Cox model and the platform’s rating. For instance, the boosted model’s highest rating category had an annual excess return of 18% and a lower default rate compared to the platform’s best rating. Second, we explained the machine learning model’s output using Shapley Additive Explanations. This analysis revealed novel nonlinear relationships (e.g., higher risk for borrowers over age 55) and interaction effects (e.g., between age and housing situation) that provide promising avenues for future research. The machine-learning model also found feature contributions aligning with existing research, such as lower default risk associated with older borrowers, females, individuals with mortgages, or those with higher education. Overall, our results reveal that explainable machine learning survival analysis excels at risk rating, profit scoring, and risk factor analysis, facilitating more precise and transparent credit risk assessments.
- 12 June 2024
- Published by Springer Nature in Digital Finance
- Vol. 6 (3) , 501-542
Open AccessImproving credit risk assessment in P2P lending with explainable machine learning survival analysis
Recent research using explainable machine learning survival analysis demonstrated its ability to identify new risk factors in the medical field. In this study, we adapted this methodology to credit risk assessment. We used a comprehensive dataset from the Estonian P2P lending platform Bondora, consisting of over 350,000 loans and 112 features with a loan volume of 915 million euros. First, we applied classical (linear) and machine learning (extreme gradient-boosted) Cox models to estimate the risk of these loans and then risk-rated them using risk stratification. For each rating category we calculated default rates, rates of return, and plotted Kaplan–Meier curves. These performance criteria revealed that the boosted Cox model outperformed both the classical Cox model and the platform’s rating. For instance, the boosted model’s highest rating category had an annual excess return of 18% and a lower default rate compared to the platform’s best rating. Second, we explained the machine learning model’s output using Shapley Additive Explanations. This analysis revealed novel nonlinear relationships (e.g., higher risk for borrowers over age 55) and interaction effects (e.g., between age and housing situation) that provide promising avenues for future research. The machine-learning model also found feature contributions aligning with existing research, such as lower default risk associated with older borrowers, females, individuals with mortgages, or those with higher education. Overall, our results reveal that explainable machine learning survival analysis excels at risk rating, profit scoring, and risk factor analysis, facilitating more precise and transparent credit risk assessments.
- Open Access
- 6 February 2024
- Published by Taylor & Francis in Journal of Behavioral Finance
- Vol. 26 (3) , 317-334
Attention Allocation of Investors on Social Media: The Role of Prospect Theory
- 6 February 2024
- Published by Taylor & Francis in Journal of Behavioral Finance
- Vol. 26 (3) , 317-334
Open AccessAttention Allocation of Investors on Social Media: The Role of Prospect Theory