Who Ever Said Child Order PI?

In the first two blog posts in this series, we discussed the performance of IEX D-Peg, looking at how trades perform after and at the moment of the trade. In this post, we look at pre-trade execution quality, rounding out our discussion of how D-Peg is designed to enable performance throughout the life of an order — from the moment it is placed to after its final execution.

So what do we mean by “pre-trade” execution quality? Typically, most investors expect that when they put an order into the market, they will trade at their limit price or worse: either their order will trade at their limit when the market moves into it, or the algorithm has to chase the market price as it moves away from the order’s original limit.

But it doesn’t have to be that way every time. D-Peg can offer price improvement not just compared to the midpoint when it executes, but also compared to the NBBO midpoint[1] when the order was sent (i.e., the “arrival price”).

Trading at prices better than arrival price (a metric known as “child order price improvement”) is unique to order types that are either pegged or can otherwise reprice, as compared to a limit order which will execute at their limit price when the market moves into them. Simply put, a displayed limit order to buy at $10.05 is not going to trade at a price lower than $10.05[2]. Pegged orders, on the other hand, shift their prices as the market changes. And D-Peg is particularly well-designed to capitalize on the opportunity to get a better price than where it started because of its ranking in IEX’s limit order book[3], the Speed Bump[4], and our proprietary IEX Signal (i.e., Crumbling Quote Indicator or CQI).

Better than where you started

To calculate the value of price improvement for D-Peg orders, we compared the NBBO midpoint at the time of trade[5] to the NBBO midpoint when the order was placed (in other words, the price the D-Peg could have executed on entry) [6]. Basically, we wanted to see how the midpoint changed from when the order was placed to when it executed.

We found that in October 2019, 14% of all D-Peg shares, or 22% of all notional value, traded at an NBBO that was better than when they entered the market. Fourteen percent of shares might seem immaterial, but it’s a phenomenal outcome compared to a limit order that cannot be price improved upon. For the 14% of D-Peg volume that received price improvement, their average PI was 3.23 cents per share. We’re talking cents per share, not mils. Even if you average the price improvement across all D-Peg shares traded on IEX (not just those that received price improvement) D-Peg trades received an average of 45 mils of PI per share!


Three reasons:

  • First, D-Peg is designed to avoid leaking information. Not only because it is a non-displayed order type, but as we touched on this in Part II, D-Peg cedes priority to displayed orders at the near touch, it leaks less information to market observers.
  • Second, D-Peg orders don’t hold up the quote in the way a Midpoint Peg can, because they only step up as far as necessary to meet a counterparty. This makes it easier to take in-the-money ticks if another market participant is trying to clear the quote.
  • Third, the IEX Signal. This functionality allows resting orders to take in-the-money ticks when the IEX Signal predicts that the price is about to change in their favor. Put simply, D-Peg hangs back (1 MPV outside of the NBBO) when the IEX Signal predicts that the price is changing in its favor, which helps it to avoid getting run over by a falling or rising price. It then returns to operating normally when the price stabilizes and can trade at the better price.

Demand a higher standard

Pre-trade child order arrival price improvement is an outcome you may rarely even think about — it’s certainly not possible with displayed limit orders in today’s marketplace. Of course, there are times when it is a disadvantage to be hanging back, but in many situations, it’s worth it to have the opportunity to see these kinds of results.

Buying lower or selling higher on 14% of volume, or 22% of notional value? That’s a benefit that flows all the way from the broker to the end investor.

Thanks so much for reading through all three of these posts. We hope this has provided new insight into the performance of D-Peg. Please reach out anytime at eric.stockland@iextrading.com if you have any questions or would like to discuss this data and how D-Peg can fit into your trading strategies.

[1] For this analysis we only consider DPEG orders that were marketable to the NBBO mid on entry. i.e., a D-Peg order to buy with a limit price <= the NBBO midpoint would be excluded.

[2] There may be exception to this in a crossed market, but under regular market conditions on IEX a displayed limit order cannot trade better than its displayed and working price.



[5] Note that we are not comparing the D-Peg execution price to arrival NBBO mid, rather we are comparing the NBBO mid at time of execution to arrival NBBO mid. This is done to avoid double counting the PI retention feature of D-Peg with arrival price improvement.

[6] In calculating arrival price improvement, we consider only D-Peg orders that traded at prices better than the arrival NBBO mid. However, the data set used when calculating an average PI includes all D-Peg shares traded on IEX. We assert that a displayed limit order on large, legacy exchanges and a D-Peg on IEX are equally likely to have the NBBO move away from their entry price.