Scientific Beta

"Deep discussions, constructive criticism, and helpful suggestions." Appraising Scientific Beta's annual conference in Paris, that's how our Research Director Felix Goltz summed up his takeaways from the two-day event.

Going by the degree of engagement from asset owners and investment managers around key topics examined at the conference, it quickly became clear what is top of mind for many institutional investors right now, Felix observed.

Aside from machine learning – examined separately in this Spring Spotlight – attendees debated such issues as the value-add from ESG, variability in factor outcomes, improving benchmark diversification, and managing US concentration risk.

US Concentration

On the latter issue, Felix shared research showing that if investors think US and non-US stocks will deliver broadly similar returns over time, optimal Bayesian portfolios implied a 10% reallocation away from US equities in current conditions.

"And that's when confidence in that neutral view is low," he added. "Higher confidence in a neutral view leads to even larger reductions in US allocations."

"Of course, different investors will hold different views," Felix observed in relation to the variability of estimates. "(However), a systematic framework to reflect those views in portfolio construction helps avoid relying on ad-hoc timing decisions."

So if the Bayesian portfolio mechanism is an available fix to US concentration, where's the challenge for institutional investors? Felix observed that when it comes to benchmark choice, it is more a governance question than a technical one.

"Boards need to understand that staying with the standard benchmark is itself an active decision, not a neutral position," he said, while pointing also to the performance of investment teams building a culture of trust and staying power.

Reframing ESG

On the ESG debate, Senior Quantitative Researcher Giovanni Bruno presented research showing that adding 200+ ESG metrics to standard financial data did not improve portfolio performance out of sample.

"Our analysis delivers a clear message - despite the explosion of ESG data, more information does not mean better portfolios," Giovanni said.
The implication of this research for institutional investors was that ESG was more about managing trade-offs than about generating alpha.

On a related issue, Bruno's colleague Antoine Naly presented research showing that when it comes to identifying green innovators, patent data provides a significantly more reliable and verifiable measure than subjective opinion-based ratings.

Factor Design

A highlight of the conference's first day was the issue of how to design factor strategies to address absolute or benchmark-relative risk objectives.
In his presentation, Principal Research Analyst Ben Luyten emphasised that portfolio outcomes in equity factor investing don't just depend on factor choices.

"With the same set of equity factors, portfolio construction decisions generate very different risk profiles, whether aiming for diversification or remaining closer to a cap-weighted global benchmark," Ben said. "Providers need to be transparent about this, so investors know what they're buying."

Other Issues

Elsewhere over the two days, sessions explored how to harvest commodity factor premia through improved multi-asset diversification and, in what is an increasingly topical issue, how to address controversial weapons in equity portfolios.

Scientific Beta also demonstrated the benefits of its 130/30 single-factor indices and how they can be used to enhance equity portfolios.

Overall, the conference was about connecting, learning and exchanging ideas. The many discussions, both on the formal conference stage and in the informal networking, generated productive debate and innovative thinking.

Thanks to all the participants who attended the event and we look forward to engaging again next year.

This article was written as part of the Scientific Beta Spring 2026 Spotlight Edition. Access all articles here.