To begin, it is worth understanding the need for closing auction session (CAS) in the first place. The closing price of the day is the most important. It is used for several operational purposes, including settling F&O contracts, Index calculations, NAVs of mutual funds and ETFs, driving mid-frequency algorithms and margin calculations. Given its importance, some processes, like CAS, to ensure its validity and wide acceptance are indeed in order.
Further, passively managed funds such as Indexed funds depend on this closing price for their transactions and need to trade as close to the official closing price as possible in order to minimise differences in performance from the index they track – tracking error, in Industry parlance. An index is, by design, calculated using the closing price of its constituents. Hence, any new investors in the fund assume they will buy all the constituents of the index the fund tracks at their close. If the fund cannot replicate this and buys at prices different from the actual close, it’s performance will deviate from the index performance, killing the raison d’etre for the fund itself.
Prior to the CAS, the closing price was determined as a volume weighted average price in the last 30 minutes – the VWAP price. To minimise tracking error and execute as close to the VWAP price as possible, index funds had to employ sophisticated execution algorithms that attempted to mirror the VWAP price. By definition, this requires some kind of predictive volume slicing, which deviates from actual volumes traded, resulting in tracking errors that were sometimes large, particularly when large orders at the end of the day allowed determined manipulators to force a specific close, “painting the close” in industry parlance. CAS allows pooling of orders from 3:15 pm to 3:30 pm into a price clearing auction which is determined as the price that allows the maximum executable volume with some safeguards and fallbacks. This equilibrium price is then set as the closing price and all trades during CAS are priced at this closing price, thus making it possible to execute at exactly the close and thereby minimising tracking error.
While daily flows in specific funds may not be very large, the problem becomes excessively acute when an index rebalances, and an indexed fund is required to exit and enter large blocks of stock at the close. CAS scores over VWAP algorithms significantly in this situation.
Given the steep increase in assets under management of passively managed funds in India, CAS is indeed a step in the right direction and seemingly aligns Indian markets with global standards where such closing auctions are commonplace.
The devil, as usual, is in the detail. The cutover from VWAP to the CAS last week was opened up for all FnO stocks which make up the bulk of the market capitalisation – and it was chaotic to say the least. On the first day of its introduction, the average NIFTY stock got marked up enormously in the CAS, causing a 200 point surge in the NIFTY relative to its corresponding value at 3:15 pm. This was repeated on Tuesday with a 150 point surge. Worse, closing prices on the NIFTY and BSE were very different leading to vastly different daily returns for two almost identical indices. Since futures and options continue to trade during the CAS when the underlying does not, these essentially lose their pricing reference and open unhedged FnO positions became essentially directional bets with associated PnL and margin impact. This constitutes only the obvious damage. Given the well-known anchoring effect from Behavioural Finance, the opening price the next day is assessed by traders as the previous day’s closing plus price discovery adjustments, prices for the next several days reflected this anomaly.
Why did this occur? Faced with completely unfamiliar dynamics around the closing that CAS introduced, most traders took a wait and see approach to assess the new feature, akin to buyers of new products delaying their purchase until early adopters report a positive experience. This abstinence from CAS participation particularly from retail resulted in exactly what CAS was designed to avoid - large orders painting the close.
This should not have come as a surprise. Every action has second order effects and thinking these through must become a part of any important implementation. This author has written earlier about the unwanted spillover impact of the removal of mini NIFTY futures on the options market. In this instance, no great foresight was required in anticipating reduced participation on the first few days of CAS. Again, no crystal ball gazing was necessary to gauge the impact of “incorrect” closing prices on FnO markets. As this author wrote on this platform last year when CAS was first contemplated, “prudence demands that this (CAS) be tried first on securities where there is no impact on the much larger derivatives market”. To minimise disruptions, experiments should start small before becoming big-bang implementations.
While CAS related volatility now appears to have settled down, with official closing prices reflecting 3:15 pm prices more closely in the last few days, the larger learnings from its implementation must not be lost, regardless of SEBI officialdom dismissing this incident as “teething trouble”. Each untoward incident like this only damages markets, and confidence in the long run, introducing needless risk premiums. SEBI will do well to consider the second order impact of its actions each time it tinkers with markets. At the very least, the SEBI’s experiments should be sandboxed or exploratory and simulated using a digital twin before becoming such big-bang interventions.
.png)
