Exchanges · Composition
NYSE, What the Listed Issue Count Holds
An exchange listing is not a list of companies, and the difference is not a technicality: every advance/decline figure published for the NYSE counts funds and preferred shares alongside businesses, which changes what the number can support.
What is on the list
Ask how many stocks rose on the NYSE today and the published answer counts rather more than stocks. The exchange lists operating companies, and alongside them closed-end funds, preferred shares, various trusts, exchange-traded products, warrants, units, and multiple share classes of the same business.
Each of those is an issue, and the advance/decline figures published for the exchange count issues. The arithmetic is correct. What is not correct is the sentence most people form from it, "most stocks fell today", because a substantial part of what moved was never a stock in the sense intended.
Why the rate-sensitive block is the real problem
Of everything on that list, closed-end funds and preferred shares matter most for reading breadth, because they share a driver that has nothing to do with equities.
Both behave substantially like fixed income. An interest-rate headline moves a large number of them the same way on the same session, and an advance/decline count faithfully records a broad-based move. Every breadth measure on this site then reports what looks like a decisive day across the market, the ratio, the line, the absolute breadth index, and no equity investor experienced one.
That is why practitioners have preferred a common-stock-only count since at least the 1990s, and why this reference recommends it wherever the data allows. It is the same data with the non-companies removed, and it answers the question people were actually asking.
Multiple share classes, counted twice
A smaller distortion, and a neat one, because it breaks the assumption that underpins every breadth measure: one company, one vote. A business with two listed share classes appears twice in the count, and on any ordinary session both move the same way, so it votes twice.
Depositary receipts add a variation on the same theme: a foreign company can be counted here and again on its home exchange, so a cross-border breadth comparison double-counts a population that is already the most correlated part of both lists. Neither effect is large enough to overturn a reading on its own. Both are reasons to treat a narrow breadth divergence, a few per cent one way, as being inside the noise of what the list is made of.
Listing is not trading
A second confusion is worth separating out. The listing venue says where a company is listed; it does not say where its shares change hands. A NYSE-listed company trades across many venues, including off-exchange execution, and the consolidated tape aggregates all of it.
Two consequences follow for anything volume-based. A single-venue volume figure is a fraction of the instrument’s activity and should be labelled as one, the short-volume research on this site uses consolidated data for exactly that reason. And an exchange’s own volume statistics describe its market share as much as they describe the market, which is a different quantity from the one most readers assume.
What to do when only the raw count is available
A common-stock-only series is the better input and it is not always on offer. When only the all-issues count is available, the reading is still usable with two adjustments, neither of which needs better data.
Discount broad readings on rate-driven sessions. If the day's news was about interest rates, expect the rate-sensitive block to have moved together and treat a decisive breadth figure with corresponding suspicion. Judge against the series' own history rather than a published threshold, since the distortion is roughly constant from day to day and largely divides out of a percentile computed on the same series. Neither restores what a clean count would have given, and both prevent the most expensive misreading.
Where it misleads
| Situation | What goes wrong |
|---|---|
| Issue count read as company count | Funds, preferreds, trusts and products are issues. The breadth figure is not a count of businesses. |
| Rate news | A large rate-sensitive block moves together, producing a broad-based reading no equity investor experienced. |
| Exchange breadth against a large-cap index | Two different lists of things. The mismatch alone manufactures apparent divergences. |
| Historical thresholds | The composition and the number of listed issues have changed substantially; old levels describe a different exchange. |
| Single-venue volume | Listing is not trading. An exchange's own volume describes its market share as much as the market. |
| Provider definition changes | A change in what counts as an issue produces a step in a long series that reads as a market event. |
Why an exchange page belongs in a volume reference
Because every count in the breadth section is taken over this list, and a measure is only as meaningful as the population it was computed across. The breadth pages each carry a version of this caution; this is where the underlying fact lives.
The practical instruction is short. Prefer a common-stock-only count. Compare breadth against an index built from the same list. Recompute any inherited threshold from the data you actually have. None of that is difficult, and skipping it is how a correct number produces a wrong conclusion.
Frequently asked questions
Is the NYSE issue count a count of companies?
No, and the gap is larger than most people expect. The list of issues traded includes closed-end funds, preferred shares, various trusts, exchange-traded products, warrants, units, and multiple share classes of the same business. Every one of them counts as an "issue" in the advance/decline figures published for the exchange, and none of the first four is an operating company reporting earnings.
Why does that matter for breadth?
Because breadth data is counted per exchange, over every issue listed. A large block of preferred shares and closed-end funds moving together on an interest-rate headline produces exactly the pattern breadth measures are designed to detect — a majority of issues moving one way — while no equity investor experienced anything of the kind. The count is accurate; the inference drawn from it is not.
What is a common-stock-only count?
The same advance/decline arithmetic restricted to operating companies, with funds, preferreds and other structures excluded. Analysts have preferred it since at least the 1990s for precisely the reason above, and where a data provider offers it, it is the better input for every breadth measure on this site. A raw exchange count and a common-stock count can diverge from each other, quite apart from either diverging from the index.
Does the same problem affect the index?
Much less, because a capitalisation-weighted index of selected companies never contained the funds and preferreds in the first place. That is the mismatch worth remembering: comparing an exchange-wide breadth count against a large-cap index is comparing two different lists of things. If the breadth data covers a list, the index it is read against should be built from that list.
How has the composition changed over time?
Substantially, which is why historical thresholds should be distrusted. The number of listed issues has risen and fallen with waves of listings, delistings and mergers, and the share of the list that is funds and exchange-traded products has grown markedly. A breadth threshold drawn from 1970s or 1990s data describes an exchange with a different composition, and it does not transfer without recomputation.
Do exchange-traded products double-count?
In a sense that matters for breadth, yes. A product holding a basket of listed shares is itself a listed issue, so its move is counted once in its own right while the issues it holds are counted individually as well. It is not double-counting in any accounting sense, and it does mean a single sector move can register more times in an exchange-wide count than there are underlying businesses involved.
Is the NYSE still where most trading happens?
Listing and trading are different things, and conflating them is a common error. A company listed on the NYSE trades across many venues, other exchanges and off-exchange execution, so the listing venue tells you where a company is listed and not where its shares changed hands. Consolidated volume figures cover all venues; a single-venue figure is a fraction and should be labelled as one.
What should I check before using exchange breadth data?
Three things, and they take a minute. Whether the count is all issues or common stock only. Whether it covers one venue or the consolidated tape. And whether the series has been consistent over the period you are studying, because providers change their definitions and a splice produces a step that looks like a market event and is not.