Hedging Criteria: An Experimental Study on Minimum Variance vs. Lower Partial Moment Approaches

How to cite

Lee, Siun; Feldman, Paul. 2025. "Hedging Criteria: An Experimental Study on Minimum Variance vs. Lower Partial Moment Approaches." AEA RCT Registry. AEARCTR-0014675. https://doi.org/10.1257/rct.14675-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Siun Lee* , Paul Feldman

* Primary investigator

Registry ID
AEARCTR-0014675
DOI
https://doi.org/10.1257/rct.14675-1.0
First registered
Last updated
Status
Completed
Trial dates
2024-09-02 – 2024-12-31
Countries
United States of America (Texas)
Registry record
http://www.socialscienceregistry.org/trials/14675

Abstract

This study investigates whether investors’ hedging behaviors align more closely with the traditional minimum variance (MV) approach or the second-order lower partial moment ($LPM_2$) criterion. Using a scenario-based design that differentiates MV and $LPM_2$ hedge ratios, participants allocated assets between two correlated lotteries representing spot and futures prices. The results reveal a divergence in hedge ratios, with some participants favoring $LPM_2$ due to its focus on downside risk rather than treating upside and downside risk equally, as MV does. This suggests that MV may underestimate risk aversion in scenarios involving asymmetric risk profiles. Notably, $LPM_2$ hedge ratios likely align more closely with participant decisions, particularly in situations where managing downside risk is prioritized over variance minimization. These findings bridge the gap between normative hedging theories and observed behaviors, offering insights for refining risk models and enhancing portfolio management strategies.

Experimental design

Participants decide asset allocation between two correlated lotteries with 45 rounds. This design aims to compare the predictive power of two theoretical models: the minimum variance (MV) approach and the Lower Partial Moment (LPM_2) approach.
The experiment consists of two lotteries, L1 and L2, each with two possible outcomes. These lotteries are designed to reflect the relationship between spot and futures prices in financial markets. The outcomes represent the changes in spot and futures prices.

A key feature of this design is the perfect negative correlation (-1) between the two lotteries. This means with probability p, L1 yields a positive outcome while L2 yields a negative outcome. By the same token, with probability (1-p), L1 yields a negative outcome while L2 yields a positive outcome. This correlation structure is crucial because it accurately reflects the high correlation typically observed between spot and futures prices in financial markets. Also, it ensures a positive hedge ratio, which is consistent with real-world hedging strategies.

Design details

Randomization method
randomization done in office by a computer
Randomization unit
randomization done in office by a computer
Clusters
5 experimental scenarios
Observations
180 participants
Arms
180 participants
Primary outcomes
Subjects are presented with these lottery scenarios sequentially at a random order. For each scenario, subjects make their allocation decision between L1 and L2. We compare this subject's choice of alpha with calculated hedge ratios from the MV and LPM2 models to assess which theory better predicts the subjects' choices. This ratio is of particular interest as it quantifies the extent of hedging behavior. Note that alpha= 1 implies no hedging (all assets allocated to L1). As alpha decreases, the hedge ratio increases, more assets are allocated to L2 (increased hedging) indicating more aggressive hedging.

Cited by

No citation data available for this registration yet.

Keywords