Publications
First-Order and Higher-Order Inflation Expectations: Evidence about Households and Firms
with Christian König-Kersting, Robert Schmidt, Stefan Trautmann, and Franziska Heinicke, in: Journal of Economic Behavior & Organization, Volume 223, May 2025, 106988.
We study the relationship between first-order and higher order inflation expectations. Using two novel survey modules of German households and firms, we establish several new insights on the relation between different orders of beliefs. While our results are remarkably consistent for households and firm managers, they exhibit relevant differences to previous results documented in the literature. Most importantly, our results show that uncertainty aggregates when individuals think about the beliefs of others relative to their own beliefs. We derive implications for calibrating noisy information models with infinite regress and discuss potential sources for differences in results documented in the literature and their theoretical implications.
Expectation Formation Under Uninformative Signals
with Martin Weber, in: Management Science, Volume 71, Issue 6, June 2025, pp. 5123 - 5141.
How do individuals process non-diagnostic information? According to Bayes’ Theorem, signals which do not carry relevant information about the objective state of the world are treated as if no signal occurred. This paper provides experimental evidence that individuals update their expectations even after observing uninformative signals. Importantly, the direction in which they update depends on the valence of the signal. Prior beliefs become more optimistic after desirable uninformative signals and more pessimistic after undesirable uninformative signals. Our results provide novel insights why individuals form and entertain false beliefs in environments where potentially new information is easily accessible but costly to verify (e.g. online media).
Risk-Taking and Asymmetric Learning in Boom and Bust Markets
with Jan Müller-Dethard and Martin Weber, in: Review of Finance, Volume 27, Issue 5, September 2023, pp. 1743 - 1779.
An increasing number of studies depart from the rational expectations assumption to reconcile survey expectations with asset prices. While surveys are helpful to establish a link between subjective beliefs and investment decisions, precise inference about how investors depart from rational expectations can be challenging without relying on strong assumptions. In this paper, we provide direct experimental evidence of how systematic distortions in investors’ expectations affect their risk-taking across market cycles. As mechanism, we identify an asymmetry in how individuals update their expectations across boom and bust markets. The documented mechanism is consistent with survey data and provides important implications for recently proposed asset pricing models.
Experiments in Finance - From No to Maybe to Yes!
with Martin Weber, in: Handbook of Experimental Finance, E. Haruvy and S. Füllbrunn (eds.), Edward Edgar Publishing, 2022, pp. 17-26.
Since the inception of the behavioral finance paradigm, experiments in financial research have grown significantly in importance and prevalence. Yet, even after decades of research, experiments are also subject to specific criticism. In this article, we first review why we need experiments in financial research. Next, we discuss three approaches how to tackle future research questions with experimental methods. We argue that experiments are necessary for complete scientific validation and an important ingredient to aid in the development of improved (behavioral) financial theory, but as one method.
When Saving is Not Enough - Wealth Decumulation in Retirement
with Martin Weber, in: Journal of Pension Economics & Finance, Volume 21, Issue 3, May 2021, pp. 446 - 473.
We field a large online survey to study preferences and hypothetical product choices for phased withdrawal accounts and compare their demand to the demand of annuities. We find that most individuals prefer phased withdrawal accounts with dynamic withdrawal rates and equity-based asset allocation. Additionally, when offered the opportunity to exchange the phased withdrawal account with an annuity, most individuals decline to annuitize. Our results suggest that policymakers should consider offering combined solutions of phased withdrawals and annuities. Retirees who are averse to full annuitization could preserve some of their accumulated wealth while also acquiring protection against longevity risk.
Working Paper
Disconfirming Information and Overreaction in Expectations
with Jan Müller-Dethard and Martin Weber
We study an intrinsic property of Bayesian information processing that does not rely on individuals having rational absolute beliefs: two equally-diagnostic signals of opposite direction should cancel out. Using five preregistered experiments and financial market data on analyst forecasts and earnings announcements, we show that this is not always the case. In both settings, investors systematically overreact to new information that disconfirms prior signals. As mechanism, we identify that individuals update their expectations as if they must selectively allocate their cognitive resources. Overall, our study provides novel insights to the paradoxical co-existence of both over- and underreaction to new information.
Presentations at: AFA 2022, Paris December Finance 2021, EEA 2021, ESA 2021, MBEES 2021, Experimental Finance 2021, European Retail Investor Conference 2021, Society for Judgement and Decision Making 2020
The Portfolio Composition Effect
with Jan Müller-Dethard and Martin Weber (R&R at the Journal of Financial and Quantitative Analysis)
This study asks whether a simple, counting-based measure of performance, which is the fraction of winner stocks in a portfolio, affects people’s willingness to invest in the portfolio. We find experimental evidence that indicates that individuals allocate larger investments to portfolios with larger fractions of winner stocks, albeit alternative portfolios have realized identical overall portfolio returns and show identical expected risk-return characteristics. Building on our experimental findings, we empirically show that the proposed composition measure also matters for the demand of leading equity market index funds. A framework which combines category-based thinking and mental accounting can explain the effect.
Presentations at: Paris December Finance 2022, Experimental Finance 2022, German Foundation for Experimental Research 2022, MBEES 2022, AFA 2021, Society for Judgement and Decision Making 2019
A Choice-based Approach to the Measurement of Inflation Expectations
with Olga Goldfayn-Frank and Stefan Trautmann (R&R at the Journal of Monetary Economics)
In macroeconomic surveys, inflation expectations are commonly elicited via density forecasts in which respondents assign probabilities to pre-specified ranges in inflation. This question format is increasingly subejct to criticism. In this study, we propose a new method to elicit inflation expectations which is based on prior decision theoretic research. We demonstrate that it leads to well-defined expectations with central tendencies close to the corresponding point forecasts and to lower forecast uncertaitny than density forecasts. In contrast to currently employed methods, the approach is robust to differences in the state of the economy and thus allwows comparisons across time and across countries. Additionally, the method is not very time consuming and portable in the sense that it can be applied to difference macroeconomic measures.
Presentations at: European Central Bank, Deutsche Bundesbank, De Nederlandsche Bank, U.S. Securities and Exchange Commission, Experimental Finance Conference 2024, MBEES 2024
Temporal Resolution of Ambiguity: Preferences for Timing and Graduality
with Gleb Gertsman and Stefan Trautmann
We study agents’ preferences for ambiguity resolution in dynamic environments. In an extension of the original Ellsberg experiment, we first show the existence of preferences for late and gradual resolution of ambiguity. Most importantly, we document a strong interdependence between agents' attitude towards ambiguity and the preference for the timing and graduality of the resolution of ambiguity. We then compare our results to the predictions of popular recursive models of ambiguity aversion. Out of the considered models, only the smooth ambiguity model (Klibanoff, Marinacci, and Mukerji, 2005) if flexible enough to accomodate the observed preferences.
Presentations at: American Economic Association 2024, Economic Science Association 2023, Subjective Probability, Utility and Decision Making Conference 2023, Society for Judgement and Decision Making 2022
Information Partitioning, Learning, and Beliefs
with Lukas Mertes and Martin Weber
We experimentally study how information partitioning affects learning and beliefs. Holding the informational content constant, we show that observing small pieces of information at higher frequency (narrow brackets) causes beliefs to become overly sensitive to recent signals compared to observing larger pieces of information at lower frequency (broad brackets). As a result, partitioning infomation in narrow or broad brackets causally affects judgements. Observing information in narrow brackets also leads to less accurate beliefs and to worse recall than observing information in broad brackets. As mechanism, we provide direct evidence that partitioning information into narrower brackets shifts attention from the macro-level to the micro-level, which leads people to overweight recent signals when forming beliefs.
Presentations at: American Finance Association 2025, Financial Intermediation Research Society Meeting 2025, SAFE Household Finance Workshop 2025, Economic Science Association World Meeting 2025, Society for Judgement and Decision Making 2024, Experimental Finance 2024, DGF 2024