Actionable insights on equities, fixed-income, macros and personal finance Start 14-Days Free Trial
Actionable investing insights Get Free Trial

The Flash Crash: Fat Finger or Algo, Emkay Must Pay


At 9:50 AM on Friday morning, when the markets were expected to keep going up forever because we have new reform agendas, the markets went curiously the other way. Stunned traders watched the index fall over 10% – no, 15%. Down to 4888.2 from a close of nearly 5800 the previous day.


(Click for larger image)

The problem, says the NSE, was a fat-finger trade. Where Emkay Brokers hit 59 sell orders from a dealer terminal.

The market opened normally today and Nifty opened at 5,815. At 9.50.58, the Nifty circuit filter got triggered, upon which the cash market was closed automatically. The Nifty fall was apparently on account of the entry of 59 erroneous orders, which resulted in multiple trades for an aggregate value of over Rs. 650 crore. These orders were entered by a trading member, Emkay Global Financial Services, on behalf of an institutional client. These non-algo market orders have been entered for an erroneous quantity which resulted in executing trades at multiple price points across the entire order book, thereby causing the circuit filter to be triggered. These orders have been identified to a specific dealer terminal.

While the Exchange systems functioned normally without any glitch, the above abnormal trades caused market closure automatically, due to the index circuit filter getting triggered. The market was reopened by the Exchange with a pre-open phase at 10.00.22 and trading resumed at 10.05.00 and the market functioned normally. The Exchange is investigating the matter.

The operating phrase here is "non-algo" – meaning these were trades that were not algorithmic. How does NSE know? It asks algorithms to only send their orders with a code that identifies them as algos. And every algorithm must be approved by NSE.

But does everyone follow this rule? Does nobody attempt to execute orders through algorithms anyway, since you are allowed to programmatically send trades through an API? Excuse me while I laugh my head off.

Emkay’s official response is like this:

On October 5, 2012, while executing an order to transact a Nifty cash basket, in Nifty-50, a dealer committed a bona fide error. The error was in entering the value of the order as the quantity of Nifty-50 basket to be transacted. The order got transmitted to the NSE trading server as a single large Nifty basket order comprising of nifty 50 stocks. Immediately on realizing the error, we promptly got in touch with the NSE and kept in touch with them to co-ordinate all future course of action until the entire erroneous outstanding position was closed out.

All our remaining clients’ outstanding positions are intact and we will continue to service all obligations. We are hopeful that this obvious and apparent error would justify the annulment of these trades, and believe in the NSE’s professional management to see the merits for annulment, which is the practice worldwide. The NSE has disabled trading for us to facilitate investigations and primarily to prevent any further exposure until this settlement is completed.

(Emphasis mine)

Since the trades were on the Nifty index, which has 50 stocks, it’s very surprising that there were 59 orders that were "rogue" – a basket order should only send in 50 orders, or a multiple, even if it was a fat finger.

Note that many brokers – and I’m not naming names here – do run algorithms without letting the exchange know. It’s simple – there’s an API, and you can use it to create a trading software, but you can also use it to send orders algorithmically if you’re a broker. I’m not saying Emkay did that, and if it was an algorithm it would be a silly one to send such large market orders when there wasn’t enough depth.

However, this raises questions – was this a legal trade? Of course. An order of that size SHOULD bring the market down, fat finger or otherwise. The thing that helps is circuit breakers, and the Nifty circuit helped. But individual stock circuits (for those that have futures on them) are around 20%, which could be deemed as too high – instead, we could use dynamic circuits that open up automatically as more limit orders come in at the edge (such as in the case of Satyam when the stock should have fallen because the news really was bad).

Dynamic circuits can operate intelligently parallel to the market, ensuring that large market orders are rejected at the fringe, with a mandatory minute or two of a hold before they open up further. So you start with 10%, and if the limit is hit, all orders below the limit are rejected for two minutes, and then the limit is opened up another 5%, and then another 5% and so on. A stock that needs to fall due to fundamental factors will take about a half hour to fall 80%, if it really should – and the wait is indeed a good thing.

Emkay’s demand that the trades be annulled should be rejected outright. What if you bought a stock at a price, say, 10% below from Emkay’s trade, and then sold it intraday for a 10% profit when the market rebounded? It would be a travesty if they annulled one trade but not the other. And you had no idea you profited from Emkay’s fat finger – that news came only much later! Yes, such trades have been annulled in other exchanges (including twice in the BSE) but annulling such trades gives a horrible lesson that you can get away if your mistake is big enough. Emkay did not set up proper limits and they should suffer the consequences – even if the consequence is that they have to pay a hefty price.

(Note: If you and I do a trade like this that makes us bankrupt, you can be sure they won’t agree to annul such trades)

More importantly, we will see such situations more often – the next one might be triggered by an algorithm. It’s less useful to focus on "ban this" or "ban that". Create a big enough deterrent through fines and making brokers face their losses, and things will fall in line.


Like our content? Join Capitalmind Premium.

  • Equity, fixed income, macro and personal finance research
  • Model equity and fixed-income portfolios
  • Exclusive apps, tutorials, and member community
Subscribe Now Or start with a free-trial