We imported the auction, but we left out the brakes
India’s Closing Auction Session (CAS) was introduced to improve price discovery, but recent sharp swings have exposed gaps in its safeguards. Jimeet Modi argues that allowing order cancellations until the auction ends can leave markets vulnerable ...
By ET CONTRIBUTORS | Updated:
Agencies
Every trading day ends with a single number: the closing price. It looks innocuous. It is anything but.
That number determines mutual fund NAVs, influences the Nifty and Sensex, and settles expiry-day derivatives. If the closing price is distorted, the consequences do not end with traders. They travel quietly through the financial system.
On August 3, India changed how that number is determined. Instead of calculating the close from the average price during the last half-hour of trading(VWAP), exchanges introduced a short Closing Auction Session (CAS), in which buy and sell orders are aggregated, and a single equilibrium price is discovered.
The concept is hardly radical. London, New York and Hong Kong have used closing auctions for years. But there is an important lesson in their experience: the auction is only as robust as the safeguards surrounding it.
India's experience has already underlined that point.
On August 13, the ninth trading day of CAS and a Sensex weekly expiry, the auction's indicative price spiked within seconds, followed by two more sharp moves. On August 19, SEBI passed an interim order finding prima facie manipulation and impounding the alleged gains.
The order describes two parties. One allegedly placed an aggressive order that pushed the indicative price higher, in a manner linked to its expiry-day options positions. Another allegedly placed large sell orders across eight Sensex constituents at prices substantially below the prevailing market, potentially pulling the indicative price lower, and cancelled them moments before the auction ended.
The crucial issue is not merely what happened. It is what the market's design allowed to happen.
India's auction currently permits limit orders to be cancelled right up to the end of the session. Mature markets have deliberately built friction into that final window. On the NYSE, closing orders are subject to restrictions as the auction approaches; in the final minutes, cancellations are increasingly constrained. Nasdaq has similar protections, while Hong Kong operates its own order-freeze mechanism.
The principle is simple: once the market gets close to the final price, participants should not be able to freely change the orders that determine it.
India has imported the auction, but not yet all of the brakes that make it resilient.
The episode on August 27 showed why that matters.
The Sensex traded comfortably above 77,100 through most of the day. Yet during the final 15-minute auction window, its indicative price plunged to around 74,983—nearly 3% below the level at which continuous trading had left the market—before recovering to close at 76,933.
For derivatives traders, that was not an academic movement.
A 77,000 call option worth Rs 327 at 3:15 pm became worthless after the auction. Stop-losses around the close could be executed at prices that disappeared minutes later. Index funds, whose mandate is to transact around the official closing price, had no choice but to absorb the difference.
An investor could therefore read the market correctly for the entire day and still have the final fifteen minutes rewrite the outcome.
The information was visible. The exchange publishes the indicative price, quantities and order imbalance. Anyone watching could see the Sensex falling nearly 2,000 points in real time.
But there was a crucial piece of information nobody could know: would the orders causing the imbalance still exist when the auction ended?
That is the distinction between genuine price discovery and a liquidity vacuum.
The 3% price band applicable to individual stocks may constrain the movement of each constituent, but it does not necessarily prevent a coordinated movement in an index made up of 30 stocks. An index-level settlement can therefore experience a much larger distortion than the apparent safeguards at the individual-stock level suggest.
None of this means the closing auction was a mistake.
Closing auctions are an important piece of modern market infrastructure. They concentrate liquidity, reduce the impact of fragmented end-of-day trading and provide a transparent mechanism for establishing an official closing price.
But they work best when the ecosystem around them is sufficiently deep—and when the rules prevent participants from exploiting the mechanics of the auction itself.
Hong Kong's experience is instructive. It introduced a closing auction in 2008, suspended it after concerns surrounding sharp movements in a heavyweight stock, and eventually reintroduced the mechanism in 2016 with additional safeguards.
Academic research has also identified greater vulnerability to closing-price manipulation around derivatives expiry and when large orders arrive near the close.
That makes the lesson for India particularly relevant: the surveillance system can detect manipulation, but good market design should make manipulation difficult in the first place.
To Sebi's credit, surveillance appears to have identified the August 13 episode quickly, and regulatory action followed before the next expiry. The watching worked.
The building needs work.
I am therefore not arguing for another layer of regulation. Indian markets have already absorbed a series of changes since October 2024—from fewer weekly expiries and larger lot sizes to higher transaction costs and now a new closing mechanism.
Constantly adding restrictions can have an unintended consequence: participation falls, liquidity becomes thinner and the cost of capital rises.
The answer is not to abandon CAS. It is to finish building it properly.
Freeze cancellations during the critical final window.
Introduce order types that can reduce an imbalance but cannot deliberately widen it.
Deepen market-making participation around the auction, because a mechanism designed for deep institutional liquidity cannot be expected to behave identically in a market where that liquidity is still developing.
And reassess whether stock-level price bands are sufficient protection for an index-level settlement price.
None of these measures should cost the ordinary investor a rupee.
But they could protect millions of investors who will never place an order after 3 pm, because the closing price ultimately flows into their mutual fund NAV, index fund execution and derivative settlement.
That is why a distorted close is not merely a trader's problem.
It is an investor problem.
It is a market-structure problem.
And ultimately, it is a confidence problem.
CAS is not the problem; deploying CAS without the full supporting microstructure is the problem.
We imported the auction. Now we need to build the brakes.
(The author is Founder and CEO, SAMCO Group)
That number determines mutual fund NAVs, influences the Nifty and Sensex, and settles expiry-day derivatives. If the closing price is distorted, the consequences do not end with traders. They travel quietly through the financial system.
On August 3, India changed how that number is determined. Instead of calculating the close from the average price during the last half-hour of trading(VWAP), exchanges introduced a short Closing Auction Session (CAS), in which buy and sell orders are aggregated, and a single equilibrium price is discovered.
The concept is hardly radical. London, New York and Hong Kong have used closing auctions for years. But there is an important lesson in their experience: the auction is only as robust as the safeguards surrounding it.
India's experience has already underlined that point.
On August 13, the ninth trading day of CAS and a Sensex weekly expiry, the auction's indicative price spiked within seconds, followed by two more sharp moves. On August 19, SEBI passed an interim order finding prima facie manipulation and impounding the alleged gains.
ADVERTISEMENT
The order describes two parties. One allegedly placed an aggressive order that pushed the indicative price higher, in a manner linked to its expiry-day options positions. Another allegedly placed large sell orders across eight Sensex constituents at prices substantially below the prevailing market, potentially pulling the indicative price lower, and cancelled them moments before the auction ended.
The crucial issue is not merely what happened. It is what the market's design allowed to happen.
India's auction currently permits limit orders to be cancelled right up to the end of the session. Mature markets have deliberately built friction into that final window. On the NYSE, closing orders are subject to restrictions as the auction approaches; in the final minutes, cancellations are increasingly constrained. Nasdaq has similar protections, while Hong Kong operates its own order-freeze mechanism.
The principle is simple: once the market gets close to the final price, participants should not be able to freely change the orders that determine it.
ADVERTISEMENT
India has imported the auction, but not yet all of the brakes that make it resilient.
The episode on August 27 showed why that matters.
ADVERTISEMENT
The Sensex traded comfortably above 77,100 through most of the day. Yet during the final 15-minute auction window, its indicative price plunged to around 74,983—nearly 3% below the level at which continuous trading had left the market—before recovering to close at 76,933.
For derivatives traders, that was not an academic movement.
A 77,000 call option worth Rs 327 at 3:15 pm became worthless after the auction. Stop-losses around the close could be executed at prices that disappeared minutes later. Index funds, whose mandate is to transact around the official closing price, had no choice but to absorb the difference.
An investor could therefore read the market correctly for the entire day and still have the final fifteen minutes rewrite the outcome.
The information was visible. The exchange publishes the indicative price, quantities and order imbalance. Anyone watching could see the Sensex falling nearly 2,000 points in real time.
But there was a crucial piece of information nobody could know: would the orders causing the imbalance still exist when the auction ended?
That is the distinction between genuine price discovery and a liquidity vacuum.
The 3% price band applicable to individual stocks may constrain the movement of each constituent, but it does not necessarily prevent a coordinated movement in an index made up of 30 stocks. An index-level settlement can therefore experience a much larger distortion than the apparent safeguards at the individual-stock level suggest.
None of this means the closing auction was a mistake.
Closing auctions are an important piece of modern market infrastructure. They concentrate liquidity, reduce the impact of fragmented end-of-day trading and provide a transparent mechanism for establishing an official closing price.
But they work best when the ecosystem around them is sufficiently deep—and when the rules prevent participants from exploiting the mechanics of the auction itself.
Hong Kong's experience is instructive. It introduced a closing auction in 2008, suspended it after concerns surrounding sharp movements in a heavyweight stock, and eventually reintroduced the mechanism in 2016 with additional safeguards.
Academic research has also identified greater vulnerability to closing-price manipulation around derivatives expiry and when large orders arrive near the close.
That makes the lesson for India particularly relevant: the surveillance system can detect manipulation, but good market design should make manipulation difficult in the first place.
To Sebi's credit, surveillance appears to have identified the August 13 episode quickly, and regulatory action followed before the next expiry. The watching worked.
The building needs work.
I am therefore not arguing for another layer of regulation. Indian markets have already absorbed a series of changes since October 2024—from fewer weekly expiries and larger lot sizes to higher transaction costs and now a new closing mechanism.
Constantly adding restrictions can have an unintended consequence: participation falls, liquidity becomes thinner and the cost of capital rises.
The answer is not to abandon CAS. It is to finish building it properly.
Freeze cancellations during the critical final window.
Introduce order types that can reduce an imbalance but cannot deliberately widen it.
Deepen market-making participation around the auction, because a mechanism designed for deep institutional liquidity cannot be expected to behave identically in a market where that liquidity is still developing.
And reassess whether stock-level price bands are sufficient protection for an index-level settlement price.
None of these measures should cost the ordinary investor a rupee.
But they could protect millions of investors who will never place an order after 3 pm, because the closing price ultimately flows into their mutual fund NAV, index fund execution and derivative settlement.
That is why a distorted close is not merely a trader's problem.
It is an investor problem.
It is a market-structure problem.
And ultimately, it is a confidence problem.
CAS is not the problem; deploying CAS without the full supporting microstructure is the problem.
We imported the auction. Now we need to build the brakes.
(The author is Founder and CEO, SAMCO Group)
(Disclaimer: The opinions expressed in this column are that of the writer. The facts and opinions expressed here do not reflect the views of www.economictimes.com.)
Too busy to trade in stocks? Invest in Mutual Funds with ET Money!
Bookmark or read stories offline -
Download ET Markets APP
Download ET Markets APP
ADVERTISEMENT
