|
Matthias Knab, Opalesque: With Jazz managing over EUR390 million, Vincent Weber discusses the judgment behind portfolio construction, what he underestimated about AI, and when earning risk premia calls for buying protection.
Trading strategies are coming to market faster. Vincent Weber sees the change across the quantitative investment strategy (QIS) providers Resonanz Capital works with, and it has widened the choices available to his team.
"A change in market structure used to mean waiting months for a related strategy to become available," he says. "Now, increasingly, we are talking about weeks, or sometimes even days. There is much more we can work with. But having more strategies available does not tell you which ones belong together."
Speaking to Opalesque eighteen months after our previous conversation, the CEO and Head of Research of Frankfurt-based Resonanz Capital attributes much of that acceleration to years of infrastructure investment across the provider base. AI is adding to it. That competition puts pressure on some opportunities, even as the larger universe opens new possibilities for portfolio construction.
The distinction matters to Weber: more strategies mean more candidates to assess. Their availability says little about the quality of their returns or their usefulness in a particular portfolio.
Choosing what earns a place
Resonanz tracks more than 4,000 strategies. The task is to select a relatively small portfolio from that universe, and judge each candidate by what it adds to the whole.
"You can show me a very attractive strategy, and I may still have no use for it," Weber says. "If it gives me more of something I already have, what have I gained? I want to know what changes in the portfolio when I add it."
Most professional investors have encountered portfolio optimisation. Less familiar is the difficulty of producing a relatively sparse portfolio when the candidate set runs into thousands. Deciding how to formulate that problem involves choices that the mathematics cannot make on the manager's behalf.
"There is more art in this than people might expect," he says. "There is no single right solution. You have to decide what you are optimising for, which assumptions you trust and how much complexity you are prepared to carry. Those decisions shape the answer."
He returns to an analogy from the previous interview. "I called it being a DJ for alpha, and I still think that fits," he says. "You have to understand each strategy, but you are also listening to what happens when you put them together."
The portfolio has to work in practice
The design also has to account for the conditions under which the portfolio will operate. Weber puts funding costs, liquidity and regulatory constraints alongside transaction costs as considerations that affect the allocation itself.
"A portfolio is a real-world thing," he says. "You are not just optimising a vector of positions. You have to fund it. You have to be able to trade it. And it has to work within the rules of the structure you are using."
Those requirements belong in the construction process from the outset. A different funding arrangement or liquidity requirement can change which combination is worth implementing.
Jazz applies the approach within a daily liquid UCITS structure for professional investors, combining open-architecture QIS with specialist external managers. Launched in 2025 with more than €200 million, the Resonanz Jazz Multi-Strategy Fund now manages over €390 million.
The portfolio comprises more than 40 distinct return engines spanning systematic macro, quantitative equities, volatility and discretionary sleeves, with individual risk budgets and strategy-level stop-loss discipline.
The firm also builds bespoke index solutions. The underlying strategy universe may be shared, but the requirements of each mandate can lead to a different portfolio.
What AI changes
In his June essay, "I Was Wrong About AI. Twice.", Weber described two successive underestimations of the technology.
"First I saw a better way to work with written information and filed it under better search," he says. "Then the models became good coders, and work that had taken months began taking days. I saw the efficiency gain. I underestimated how much the work itself would change."
In systematic research, he sees the distance between reading a paper and testing a first implementation shrinking. Researchers remain responsible for judging which hypotheses deserve attention and whether a result survives scrutiny.
That raises a question for portfolio managers too: why should selecting and combining strategies be insulated from the same change? Weber's argument rests on the work surrounding the model: historical data, realistic simulation, implementation and the design of the investment process. Access to a capable model does not settle those decisions. Portfolio construction will evolve with the tools.
Earning the premium and managing the tail
For Weber, the core portfolio will generally favour strategies with positive expected risk premia: returns earned for accepting particular risks. His analogy is writing insurance.
"You receive a premium because you are willing to bear a risk," he says. "There has to be a reason to expect that compensation. Then you have to understand the potential claims, especially when you put several of those exposures together."
That is why the tail of the combined return distribution receives particular attention. The decision to buy additional protection depends on the risks already in the core.
"If we have a lot of carry exposure, we will often buy outright convexity to mitigate the tail risk," Weber says. "With a more diversified mix, including trend or quality equities, we may decide against that additional protection. We can be tactical here because the need changes with the portfolio."
Convex protection is intended to provide an increasingly strong offset as adverse moves become larger. Its cost forms part of the economics of the whole portfolio. The presence of trend or quality equities does not automatically remove the need for it; the decision follows the aggregate exposures.
The same question therefore applies to a return-seeking strategy and a hedge: what does it contribute alongside everything else?
"We generally want to be earning risk premia," Weber says. "But we also want to be deliberate about the tail risks we retain. What we choose to protect depends on what we choose to own."
|
RSS
