Celebrating Success – Professor Darren Duxbury

Congratulations to Professor Darren Duxbury, and his co-authors, whose paper Domain-dependent diversification: The influence of gain–loss domain on correlation choice has been published in the Journal of Economic Behavior & Organization.

The paper is the product of an international research collaboration with colleagues at Radboud University, Netherlands.  The collaboration developed when I hosted/supervised a visiting international PhD student (i.e. occasional PGR student), Charlotte Borsboom (Radboud University), for three months (September-December, 2019) funded under the Erasmus+ mobility scheme. The visit resulted in a research project joint with Charlotte and her Radboud supervisor, Professor Stefan Zeisberger, and joined by another PhD student at Radboud University, Alex Nieber.

Abstract: Despite compelling evidence of widespread gain–loss-domain-dependent behavior, research on domain-dependent diversification is scarce. We recruited 251 experienced US retail investors to participate in a controlled experiment with the task to select portfolios that differ in asset correlation and, hence, diversification benefits in both the gain and the loss domain. We find evidence of domain-dependent diversification, both unconditional and conditional on benchmark portfolio preferences. Consistent with a loss-attention hypothesis, diversification errors are not observed in the loss domain but are clearly present in the gain domain (with much lower diversification relative to the benchmark).

The paper is published open access and available at https://doi.org/10.1016/j.jebo.2024.106681

Layperson summary: Diversification is a fundamental risk-reducing strategy in many aspects of life, with particular importance in the world of finance where superior risk-adjusted investment performance can be achieved via portfolio diversification.  It remains a puzzle, therefore, why households generally hold under-diversified portfolios.  It has been shown that

investors make mistakes when aggregating information and neglect correlation of asset returns – both fundamental to portfolio diversification.  But when and why this might occur has not been well-understood.  To better understand the failure to diversify investment risk, we examine how domain dependence (that is, whether asset returns are expected to be positive or negative – i.e., gains or losses) might impact diversification decisions.  Across a wide range of decisions, losses, relative to gains, are known to promote increased cognitive effort and attention, prompting less reliance on automatic decision processes and promoting enhanced logical reasoning.  We might expect, therefore, more portfolio diversification in bear markets where investment returns are generally negative than bull markets characterised by positive returns.  To test this, we recruited experienced US retail investors to participate in a controlled experiment with the task to select portfolios that differ in asset correlation and, hence, diversification benefits, in both the gain and the loss domain.

We find evidence of domain-dependent diversification, both unconditional and conditional on benchmark portfolio preferences. Consistent with our loss-attention hypothesis, whereby investors pay more attention to asset return correlations and are more likely to correctly aggregate outcomes in the face of losses than gains, diversification errors are not observed in the loss domain but are clearly present in the gain domain (with much lower diversification relative to the benchmark).  Our findings, therefore, help to better understand the puzzle of poor portfolio diversification in financial markets.