Both markets end the day with an auction. New York has run one for twenty years without incident. India's is three weeks old and has already produced a manipulation case. The difference is not the auction — it is four design decisions around it.
Between the moment the cash market is switched off and the moment same-day index options stop trading. The closing price is decided inside it, and nobody can trade against it.
This needs saying first, because most of the anger currently pointed at CAS is pointed at the wrong thing.
A closing auction is a good mechanism. Instead of letting the last trade of the day set the official price — which is trivially cheap to nudge — it pools every closing order and matches them at one price: the price at which the greatest number of shares can change hands. New York has closed this way for two decades. It is where every index fund and ETF in America does its business, and it is routinely the single deepest liquidity event of the American day.
India adopted the same mechanism on 3 August 2026, for F&O-eligible stocks on the NSE and BSE. The matching logic is identical. The equilibrium price is calculated the same way.
Everything that follows is about that gap, and about three smaller design decisions that turned a sensible reform into something that could be gamed inside ten days.
Two order types exist only for the closing auction. You submit them during the day, they sit idle, and they execute only in the auction itself.
"Get me in at the close, whatever the closing price turns out to be." No price condition. Guaranteed to fill — you just do not know at what.
"Get me in at the close, but only at ₹500 or better." A price condition is attached. You control the price, so you may not get filled at all.
An index fund obliged to buy ₹500 crore at the official close cannot work that into the order book without moving it. One MOC into the auction crosses it in a single print.
Same trade-off as an ordinary market versus limit order. The only difference is timing — they are parked for the auction rather than working the book through the day.
India has both of these already, without the labels. Order Entry Session I accepts market and limit orders; Session II is limit-only. The order types are identical to New York's. The discipline around them is not, and that turns out to be the whole story.
Read these side by side. The divergence is not in how the closing price is calculated — both use the same maximum-executable-volume logic. It is in what else is happening while the auction runs.
Nasdaq runs the same shape as the NYSE: MOC to 3:55, LOC to 3:58, imbalance published from 3:50 — every 10 seconds, then every second after 3:55. Non-F&O stocks in India are untouched by CAS: they still close at 3:30 on the old VWAP method.
| NYSE | India (CAS) | |
|---|---|---|
| Continuous market during the auction | Runs in parallel to 4:00 | Frozen from 3:15 |
| Can you act on what you see? | Yes — arbitrage it in the live book | No — only by bidding into the same auction |
| Can you add to an imbalance late? | No — contra side only after 3:50 | Yes — freely, through 3:25 |
| Can you cancel what you added? | Effectively no after cutoff | Yes, until the entry window shuts |
| Imbalance transparency | Every 1 sec from 3:50 | Price, quantity, imbalance and index — a richer feed |
| Derivatives vs cash close | Aligned at 4:00 | Cash 3:35, derivatives 3:40 |
| Participation depth | Largest liquidity event of the day | Thin — one actor can dominate a side |
| Human oversight | DMM assigned per stock | Fully automated |
Note the fifth row carefully — it runs against the popular complaint. India is not the opaque one.
Four differences. The second one is doing most of the damage.
On the NYSE the auction price has to answer to a live continuous market trading beside it. Push the auction away from fair value and the regular book takes the other side. In India, from 3:15 the auction is the only market there is. A distorted print has no competing price to correct it.
This is the single biggest gap. New York's rule is blunt: past the cutoff you may only enter orders that reduce the imbalance, and Closing Offset orders can never add to one. That makes the place-heavy-then-cancel play structurally impossible. CAS permits exactly that play — enter size, move the indicative price, cancel before matching.
India is not the opaque one here. The NSE broadcasts the indicative equilibrium price, tradable quantity, cumulative buy and sell quantities, imbalance and an indicative index value — a richer feed than New York's. The difference is what you can do about it. The NYSE publishes while continuous trading is still live, so a distortion can be arbitraged in the regular book. In India the cash market is dead: you can watch ₹66 crore land in real time and your only recourse is to bid into the same auction being pushed. Thin participation finishes the job — one actor was 99.9% of a side.
India freezes cash at 3:15, fixes the close between 3:30 and 3:35, and lets 0DTE index options run to 3:40. The settlement input is decided in a window nobody can trade against — and option writers cannot hedge through it. That is what the delink 0DTE campaign is aimed at, and it is the ask most likely to succeed.
The debate keeps merging these, and it weakens the argument badly. They have different causes, different victims and different fixes.
A genuine structural invitation, and it was accepted inside ten days of launch. Whoever was on the wrong side of those two prints on 13 August paid for it. Fixable with one rule imported directly from New York: past a cutoff, contra-side orders only.
This is what generates the daily grievance. The Dixon 15000 CE running from 10 paise to 90 during the auction. Straddle premiums down roughly 40% week-on-week. Theta collapsing from ₹120 to ₹14.95 in seconds. None of it requires anyone to have manipulated anything.
SEBI's interim order of 19 August sets out two separate plays on the same day — 13 August, a weekly Sensex expiry. Both are textbook demonstrations of gap ii.
JPMorgan-linked, holding long calls. In the first seconds of the session, placed 32 buy orders worth roughly ₹66 crore across all 30 Sensex constituents at maximum permitted prices — 99.9% of all buy-order value in those two seconds. Repeated twice more at 96% and 85%, then cancelled once the price had moved.
Holding long puts. Between 3:21 and 3:25, placed sell orders worth ₹145 crore across eight stocks priced 1.5–3% below their reference prices, holding the index down for five minutes — then cancelled almost all of them within three seconds.
Both entities barred. The index was moved +362 points one way and −233 points the other on the same afternoon — neither move reflecting anything a buyer and a seller actually agreed on.
In fairness to the regulator: this was detected and acted on within ten days, and the enforcement worked. The design is what let it happen in the first place — and enforcement after the fact does not refund the option sellers who were on the other side of those two prints.
The credible asks are narrow, and every one of them has a working precedent in New York. "Roll back CAS" is the weakest of them, because the auction itself is sound.
CAS has been live for roughly three weeks. The SEBI chief has said publicly that the system is transparent and that he is open to suggestions on tweaks. A campaign under #CASkiAwaaz is collating trader submissions to the regulator, and the community's ask has already narrowed from "roll it back" to the more winnable "delink 0DTE" — which is the right move.
One thing worth saying out loud, because it is true and because someone will otherwise say it back to you: a good share of the loudest loss posts of the last three weeks were naked short out-of-the-money options held into 15:10 on expiry day. The system change did not create that position. It made an existing bad habit far more expensive, very quickly. If your plan depended on premium decaying quietly through the last half hour, that plan needs rewriting regardless of what SEBI decides.
Timings for CAS are taken from the NSE's own Closing Auction Session documentation and broker explainers of the SEBI framework effective 3 August 2026. NYSE and Nasdaq timings are from the exchanges' published auction fact sheets. The 13 August figures are as reported from SEBI's interim order of 19 August.
One link has deliberately not been asserted here: the precise chain by which constituent auction closes roll up into the Nifty and Sensex settlement values. That computation step should be confirmed against the NSE and BSE circulars directly before anyone leans on it in an argument.
Join the discussion
Questions, counterpoints and alternative readings of the auction design are all welcome. Every comment is reviewed by Abhijeet before it appears on the page.