Worst-Of Baskets: The Dispersion Risk Hiding in Multi-Index Structured Notes
Picture two structured notes from the same issuer. Same five-year term, same contingent coupon, same downside barrier. The first is linked to the S&P 500. The second is linked to the worst of the S&P 500, the Russell 2000, and the Nasdaq-100. On a fact sheet they look like cousins. They behave like different animals.
The difference is one phrase, "worst-of," and it carries more weight than anything else on the page. The first note rises and falls with one broad index. The second pays out based only on whichever of its three indices does worst, and the other two might as well not be there.
That distinction is everywhere in the market. Across the notes SQX has classified, roughly half carry more than one underlier, and when a note has more than one, it is structured as worst-of about 94 percent of the time. The baskets themselves are remarkably consistent: the single most common combination, on nearly 2,000 notes, is the S&P 500, the Russell 2000, and the Nasdaq-100 together. Large-cap, small-cap, and tech, bundled into one note over and over.
The risk that bundle creates has a name on the desk: dispersion. It is worth understanding, because adding indices to a worst-of note adds risk rather than spreading it, and the size of that risk comes down to how far the indices can drift apart.
What worst-of means, and why issuers use it
On every observation date, a worst-of note looks at each underlier and acts on the weakest one. The laggard governs everything that matters: whether the note auto-calls, whether the contingent coupon pays, whether the downside barrier has been breached. The two indices that held up are irrelevant to the outcome. Only the loser is read.
That is one of three ways a note can reference its underliers. A single-underlier note tracks one thing. A worst-of note tracks the weakest of several. A basket-average note tracks the blended performance of the group, which actually does spread risk the way investors expect. The SQX data shows how lopsided the usage is. Among multi-underlier notes where the scope is recorded, worst-of accounts for the overwhelming majority, basket-average appears on only a few hundred notes, and best-of is close to nonexistent.
Issuers favor worst-of for a straightforward reason. A worst-of note pays a higher headline coupon than an otherwise identical single-index note, because the investor is taking on more downside to earn it. The extra coupon is the price the issuer pays for the extra risk the investor accepts. The yield is the compensation, and reading the structure tells you what you sold to get it.
More indices, more risk, and dispersion sets the size
The instinct is to read three big indices as diversification. Under a worst-of payoff, the instinct runs backward. Each index you add to the basket is one more candidate to be the laggard, one more way the note can miss a coupon or break a barrier. The basket does not pool the outcomes. It selects the worst of them.
A single observation makes the point. Suppose on a coupon date the S&P is down 5 percent, the Russell is down 12 percent, and the Nasdaq is up 3 percent. A note linked to the S&P alone is comfortably fine. A worst-of note across all three reads the Russell at down 12 percent and acts on that figure alone. The healthy index and the flat index do nothing for the holder.
This is where dispersion does its work. Dispersion is the degree to which the underliers drift apart from each other. When the indices in a basket are tightly correlated and move together, the laggard is never far behind the leaders, and the worst-of feature costs the holder relatively little. When the indices spread out, the gap between best and worst widens, and a worst-of holder lives at the bottom of that gap. Lower correlation, which helps an ordinary diversified portfolio, hurts a worst-of holder. The more the names can diverge, the more the worst-of feature can cost.
The baskets in the market sharpen this further. The S&P, the Russell, and the Nasdaq are positively correlated, which makes a worst-of note on the three feel safe in a calm market, since the indices tend to move in step. The danger is what correlation does under stress. In a sharp selloff, correlations across equity indices tend to climb toward one, so the indices fall together, and one of them still falls furthest. The basket feels diversified right up to the moment it matters, then behaves like a bet on the weakest member. A holder who read "three major indices" as safety finds out otherwise at the worst time.
The fields that make the risk visible
None of this can be assessed from a payoff name. To see the dispersion risk in a note, you need two things in your data, and most reference feeds carry neither cleanly.
The first is the observation scope as a real, controlled field: single, worst-of, or basket-average, recorded explicitly rather than implied. A note's entire risk character turns on this value, and summary data often omits it or buries it in prose. The second is every underlier resolved to an identifier you can actually use. A basket listed as the free-text strings "S&P 500," "Russell 2000," and "Nasdaq-100" joins to nothing. The same three indices, each resolved to a real ISIN and validated, let you pull return histories, compute the pairwise correlations, and estimate how far the basket can disperse. The difference between those two states is the difference between guessing at the risk and measuring it.
That gap shows up at several desks at once. An advisor at the point of sale should be able to explain why a worst-of note pays more than a single-index note, and ideally see how correlated the basket is before recommending it. A risk team modeling the note needs the joint behavior of the underliers, not three separate single-name views, which starts with resolved identifiers. A portfolio manager holding several worst-of notes often holds overlapping baskets, since the market reuses the same handful of indices, so the same Russell 2000 exposure can sit inside a dozen different notes at once. A position-level view misses that concentration entirely unless the underliers are resolved and joinable across the book. Every one of these needs the scope and the resolved basket as data, in hand before the market moves, not reconstructed from prospectuses after it does.
What the structure is really telling you
On a multi-underlier note, the worst-of feature is the first thing to read, ahead of the coupon. The coupon tells you what you might earn. The worst-of feature, and the dispersion it exposes you to, tells you what you are risking to earn it. More indices means more risk, correlation sets how much, and the direction is the opposite of the diversification the basket appears to offer.
The two fields that make that risk legible, the observation scope and the fully resolved underlier list, are exactly the fields that belong in reference data as structured, queryable values rather than sentences in a filing. This is the core of what SQX's structured note reference data delivers. Observation scope is a controlled field, classified as single, worst-of, or basket-average from the source filing rather than left as prose. Every underlier on every note resolves to its own ISIN, cross-validated against an authoritative source, so a basket of three indices becomes three joinable securities instead of three text strings. The two pieces arrive together, keyed to the same instrument, across the full universe of notes SQX covers.
What that combination unlocks is the analysis the term sheet alone cannot give you. Because the scope is a field, you can pull every worst-of note in a book in a single query, or screen new issuance for worst-of structures on three or more underliers before anyone buys one. Because the underliers are resolved, you can take any one of those baskets, pull the return history for each constituent, and compute the pairwise correlations that tell you how far the basket can disperse and how badly the worst-of feature can bite. The same resolution lets a portfolio manager roll exposure up across notes, so the Russell 2000 sitting inside a dozen separate worst-of notes shows up as one concentrated position rather than twelve hidden ones. None of that is possible when the basket lives as free text in a PDF, and all of it falls out naturally when the scope and the underliers are data.
That is the difference SQX's structured note reference data is built to provide: the worst-of feature and its basket captured as fields you can query, join, and model, turning "the worst of three major indices" from a line on a term sheet into a risk you can see and size before it moves against you.
To learn more about our structured note reference data, contact us today!
The figures in this article reflect the subset of the structured note universe SQX has classified to date, not the entire outstanding market.
Want more SQX updates? Follow SQX on LinkedIn and subscribe to our
LinkedIn
newsletter to stay ahead of the curve.
Latest News
