In the Swiss market for barrier reverse convertibles, more competition did not make retail products cheaper. It made them harder to compare. Issuers moved margin out of the simple products, where it could be competed away, and into the complex ones, where it could not. The clearest evidence is in a single gap. According to new research by Marc Chesney, Felix Fattinger, Jonathan Krakow and Simon Straumann in the Review of Finance (2026), products written on several underlying stocks carry annualized markups roughly 4 percentage points higher than single-stock versions and deliver realized returns about 2.1 percentage points lower. Both product types produced negative average excess returns. The multi-asset versions were substantially worse.
The usual reading of complexity treats it as innovation that investors pay for because they want tailored exposure. This paper supports a less comfortable reading. In this market, complexity works as a pricing technology. Its job is to break the comparison that competition depends on.
Where the Margin Went When Competition Arrived
The products studied are yield enhancement notes. An investor receives a fixed coupon and gets principal back at maturity unless the underlying asset falls through a barrier during the product's life. If the barrier is breached and the asset finishes below its starting level, the investor receives the depreciated stock, or its cash equivalent, instead of par. The investor has effectively sold a down-and-in put and is collecting the premium as a coupon. The issuer keeps the difference between what that option is worth and what it pays out.
"Singles" reference one stock. "Multis" reference a basket and pay according to whichever constituent performs worst.
As competition among issuers intensified, two things happened together. Markups on Singles fell, which is what textbook competition should do. Markups on Multis did not fall; they stayed high. Meanwhile the product mix shifted toward Multis. If competition were simply forcing everyone toward better terms, the complex products would have compressed as well. They did not. The more coherent explanation is that issuers relocated their margin to the segment where competition had no grip.
The mechanism is combinatorial. A single-stock note on a large Swiss bank or pharmaceutical name with a given maturity is a commodity. An investor, or more realistically an advisor, can line up five issuers' versions and pick the highest coupon. A worst-of note on three or four names can be assembled in so many combinations of underlyings, barrier levels and tenors that two identical products rarely exist at the same moment. Without an identical product, there is no clean comparison, and without that comparison price competition loses its force.
Why Worst-of Is the Opposite of Diversification
Intuition goes most wrong here, and the error is what makes the structure sellable.
To a retail buyer, a note on four blue-chip stocks sounds safer than a note on one. Four names suggest diversification. In a worst-of payoff, the logic runs the other way. The investor loses if any of the four breaches the barrier, so each added underlying is another way to lose. Lower correlation among the names makes it more likely that one of them breaks down. The investor is not diversified across the basket; they are exposed to its weakest member and are implicitly betting that the names move together.
That is why Multis can advertise higher coupons. The higher coupon is real compensation for a larger embedded short option position. Because the extra risk is hidden in the interaction between assets rather than in any single visible feature, buyers misjudge it. The authors' laboratory experiment shows how. Participants estimated barrier-breach probabilities accurately for single-asset products but underestimated them for multi-asset products by about 6.2 percentage points. They also valued the Multis more highly relative to fair value. The distortion grew with overconfidence: the participants most sure of their judgment were the most wrong about the joint downside.
Taken together, these findings show complexity being used in two ways. It blocks comparison between issuers, and it hides the size of the risk inside each product. The headline coupon carries the marketing while the correlation exposure carries the cost. The paper's abstract states the result in its strongest form: the more complex products are first-order stochastically dominated by the simpler ones. Across the return distribution, the simple product does at least as well. Buying the complex version does not trade return for safety or safety for return. It is simply a worse product.
A Market With No Twins
The most useful finding concerns what happens when comparison is possible. The authors identify "twin products": notes issued around the same time with identical underlyings and maturities. Where a Multi has a twin, the pattern changes. Terms improve and markups fall. Complexity alone does not protect margin. What protects it is the absence of a benchmark.
Twins are scarce in the complex segment. Only 29% of Multis had a preceding twin product, compared with 68% of Singles. That difference is the market structure in one statistic. Two-thirds of simple products face a direct competitor, while more than two-thirds of complex ones are effectively alone. An issuer designing a Multi can avoid a twin by choosing an unusual basket. An issuer designing a Single on a heavily traded name cannot.
The conclusion is that comparability has measurable economic value. It is the variable through which competition reaches the investor's pocket. Disclosure regimes and product warnings that focus on describing features may be aimed at the wrong variable. A buyer can understand every term of a worst-of note and still overpay, because the problem is not the clarity of the description. The problem is that there is nothing to compare the price against.
Who Ends Up Holding the Complex Version
Overpayment is not evenly spread. Investors holding competitively priced simple products tend to be wealthier and more diversified, and they earn better portfolio-level returns, Sharpe ratios and alpha. Less sophisticated investors are more likely to hold the complex products.
The result is segmentation, and it explains why the arrangement lasts. Competition does work, for the clients able to use it. Sophisticated buyers gravitate to comparable Singles and receive the compressed markups. Less sophisticated buyers are steered, or steer themselves, toward Multis with attractive coupons, and issuers keep their rents there. Market-wide averages can show competition improving terms while a large share of retail flow sees no improvement at all.
For an advisor, the implication is direct. If a client's structured product shelf leans toward worst-of baskets with no obvious twin, the client is likely in the segment that funds issuer margin. A high coupon is not evidence against this. Under this mechanism, a high coupon is often the evidence for it.
What the Evidence Does Not Settle
The case is strong but has gaps, and several weaken the "deliberate design" framing more than the investor-outcome finding.
First, intent is inferred rather than observed. The authors present their evidence as consistent with strategic complexity, which is the right level of claim. An alternative explanation is that issuers simply follow demand. Retail clients chase higher coupons, worst-of structures deliver them, and the shift toward Multis follows yield-seeking behaviour rather than an effort to avoid comparison. The persistence of Multi markups as Single markups fall is harder to explain this way, but it does not rule demand out.
Second, the markup and return figures do not match. A 4 percentage point markup gap alongside a 2.1 percentage point realized return gap means roughly half the measured extra margin did not appear as lost performance over the sample. Several explanations are possible. Fair-value models for multi-asset barrier options depend on correlation assumptions that are themselves uncertain. Issuers bear real hedging costs on correlation that they cannot fully offload. Or the sample period's realized breaches were milder than the products priced. Each reading changes how much of the markup is rent and how much is the cost of manufacturing a hard-to-hedge risk.
Third, the behavioural mechanism comes from a laboratory. A 6.2 percentage point misjudgment by experiment participants is suggestive. It does not show that actual Swiss retail buyers, many of whom buy through advisors, make the same error at the same size.
Fourth, the evidence covers a single national market and a single product family. Switzerland's structured product market is deep and well documented, which is why it is useful for study. Whether the same dynamic applies to US-listed buffered notes or autocallables sold through brokerage platforms is a separate empirical question, and this paper does not answer it.
None of this rescues the complex product for the buyer. The stochastic dominance result does not depend on why issuers built Multis. Whether the extra margin comes from strategy or demand, the investor holding the worst-of version was worse off.
The Twin Test
The decision rule follows from the mechanism. Before buying a complex structured product, check whether a twin exists: the same underlyings, the same maturity, from a different issuer, around the same date. If one does, compare terms and take the better one. If none does, treat the missing benchmark as a cost in itself, because the evidence shows that is where markups survive.
The thesis would weaken under one observable condition. If markups on multi-asset products begin to compress as issuer competition intensifies, matching what already happened to single-asset products, then complexity was a temporary refuge and not a durable defense of margin. Until that happens, the comparison holds: in this market, the simpler product was the better one, and the complex version mainly protected the issuer's margin.
