Predatory trading

By James Hymas | June 9, 2010 | Last updated on June 9, 2010
8 min read

It happens to every trader: you place a bid for a thousand shares of thinly traded stock at a nickel above the best bid showing, and before you can blink, there’s another bid a penny in front of you. While desirable from a capital markets perspective – the bid/offer spread is, after all, declining, which is good news for the sellers – it’s annoying for the trader and very time-consuming to address.

The obvious solution is to employ trading algorithms – computer programs, sometimes very simple ones such as Excel spreadsheets – that place orders based on user-defined rules. After all, that’s how the institutional participant was able to react so quickly.

Unfortunately, according to the Investment Industry Regulatory Organization of Canada (IIROC), brokerage firms are responsible not just for the actual orders placed by clients, but also for the supervision of the algorithms themselves. This has discouraged retail access to such services in Canada compared to the U.S. and other jurisdictions.

There is, however, some hope that retail investors will be able to level the playing field to compete more effectively in illiquid markets with the institutional investors and their access to algorithmic trading.

How so?

Well, because pegged orders may become more commonplace.

In a recent consultation paper, the Canadian Securities Administrators and IIROC had a number of questions for marketplace participants concerning some of the effects of recent competition-driven innovation. The most recent changes, such as dark pools, have been designed for – and are almost exclusively pertinent to – institutional traders. But the section of the paper dealing with pegged orders should be of immense interest to retail clients and their advisors.

The document provides a definition: A primary peg order is a visible order that is automatically priced (and then subsequently re-priced as necessary) to equal either the best bid, in the case of a buy, or the best offer in the case of a sell.

Brokers who have spent entire days changing orders to stay on the bid – or dealing with clients seeking such changes – will immediately recognize the potential utility of such orders. In relatively illiquid markets – such as Canadian preferred shares – it would be very useful to place such an order.

For example, suppose that a certain issue is quoted at 20.00-50; and that an investor wishes to purchase the issue without – if at all possible – paying the full spread on the transaction.

Instead of placing a limit order for 20.05 (and in all probability seeing the bid move to 20.06), he might wish to place a pegged order with a limit of 20.20. That order will be visible at the bid level of the National Best Bid and Offer book and move instantaneously with upward or downward changes in this bid, to a maximum of the 20.20 preset limit. The improvements in terms of required effort and latency (the time between a change being triggered and the change being made) are immense.

Two major arguments have been proposed to enforce a ban on pegged orders: the violation of price-time priority and the increase in messaging traffic between marketplaces.

Price-time priority

Jeffrey MacIntosh of the University of Toronto and a director of the Canadian National Stock Exchange (CNSX), which owns and operates the Pure Trading exchange, takes exception to the queue-jumping inherent with pegged orders in a fragmented marketplace.

If a trader places a pegged order with Exchange A, it will immediately move to reflect a better bid on Exchange B – and may well be executed prior to execution of the order that triggered the repricing. He claims this violation of price-time priority will result in a material reduction in the attractiveness of limit orders, to the ultimate detriment of the capital markets, an assertion that was repeated by the CNSX in its response to the consultation paper. More generally, this concern may be referred to as “free riding.”

This conclusion flies in the face of results found by Brown and Holden in 2005, who believe limit orders are already unattractive because they do not automatically reprice in response to changing market conditions and are therefore subject to being picked off by arbitrageurs when these changes occur (particularly, for example, during market crashes).

Their model shows that pegged limit orders increase the profits made by those who place these orders, largely at the expense of the dealers. Additionally, when dealers supply little liquidity, market order losses are decreased. This decrease in the frictional costs for the buy-side should be considered beneficial for the market.

Additionally, it should be noted that price-time priority is already routinely violated due to internal crosses processed by individual brokers and due to market makers electing to participate in the interaction between an order eligible for a minimum guaranteed fill and a limit order (even if they have not posted an order that is on the market at all).

To some degree, the price-time priority issue is a red herring: there can be no question of unfairness if pegged orders are available to all those who wish to use them.

The only valid objection to pegged orders on such grounds must be based on harm done to the capital markets due to their potential for regular limit orders to be deprecated.

However, despite the years of experience with pegged orders on the NYSE-Euronext, the NASDAQ Baltic and Nordic exchanges, most American ECNs and NASDAQ Supermontage, evidence of harm to the capital markets remains a matter of conjecture and fear-mongering – much like the harm allegedly due to short-sales.

Messaging traffic

The Toronto Stock Exchange – which currently does not offer pegged orders – was able to come up with a valid objection to their introduction.

“Marketplace-visible pegged orders have a negative impact on market structure because they result in significant messaging increases that place unnecessary strain on marketplace and regulatory infrastructure,” it said.

Omega ATS confirmed this view: “The message-to-execution ratio of a visible pegged order is significantly higher for visible pegged order types than any other order type.”

However, neither of these groups provides any quantification of this effect. Given that the maximum exchange fee charged by the TMX is currently $0.0037 per share (for a liquidity-taking order in a high-priced equity placed by a low-volume participant), the immediate question is: assuming there is a charge at all for pegged orders (liquidity-providing by definition), what’s the fair price?

As Omega ATS suggested, traders should simply be shown a price and make up their own minds.

A question of fairness

The request for comments pointed out “the strategy that underpins market-pegged orders has been used by traders in Canada for some time” and many commenters agreed.

It has been present initially through frequent manual repricing of limit orders and, more often nowadays, through algorithmic trading software.

Given the emphasis on fairness in the regulators’ questions, it’s surprising none of the big players – BMO Nesbitt Burns, TD Securities, CIBC World Markets, RBC Dominion Securities, and National Bank Financial – have given any consideration to the impact of the issue on their retail client base.

Retail clients do not have access to algorithmic trading and this lack of access places them at a distinct disadvantage to institutional traders when competing for fills – particularly in illiquid markets. Pegged orders have the potential to redress this imbalance and should be made available – or, at the very least, not subject to a regulatory ban – by exchanges to their participants and by the participants to all of their clients.

It is wise to remember the example of Iceberg orders, which have been offered to participants by the TSX for quite some time. There are no regulatory obstacles to making this order type available to retail; the sole impediment is the lack of execution-based competition in Canada.

Any attempt by the regulatory authorities to address fairness with respect to order types should include a provision that all order types available to a retail brokerage through its membership in various exchanges should be disclosed to its clients, with its decision made clear via a “comply or explain” notice posted on its Web site.

Not a panacea

Despite the attractiveness of pegged orders to retail, it must be understood that they should not be regarded as a panacea.

For example, consider a preferred share quoted at 20.00-50, 11×1. It is a relatively simple matter for a hostile program-trader to determine that of the 11 board lots quoted at the bid, only one is a regular limit order while ten lots are pegged.

A predatory trader seeking to sell one thousand shares could potentially place a new bid at 20.49 and, once the pegged orders drift up to the new bid, hit these bids at a substantially better price than he would have received in the absence of technique.

This operation could be construed as abusive and subject the trader to regulatory sanctions – which shows that pegged orders are not the answer to all trading problems. If users of such orders need regulation in order to protect them from the consequences of their actions, it should be flagged that pegged orders contain a weakness that must be considered carefully before they are placed.

Omega ATS remarked on this potential predation. It recommended that, in order to address the free-riding concerns, thought by some to be so important, this type of predatory trading should be explicitly allowed.

Equilibrium will then be reached between the number of regular limit orders and pegged orders in the marketplace, which addresses the question of depreciation of regular limit orders.

It is recognized that as the proportion of regular limit orders declines, markets may well become more volatile on an intra-day basis, as a large proportion of pegs are moved by a small proportion of firm limit orders.

This should be counted as a feature of the proposal, not a bug – markets should be designed to serve intelligent investors with a long-term view.

Such investors will care about intra-day volatility only one day in every thousand, if that, and when prices do align to trigger a trade, the volatility will give them a wider variety of potential entry and exit prices for their decisions. Volatility is a value investor’s best friend.

Allowing competent traders to thrive at the expense of the incompetent is something of a novel idea in Canada, but one that is long overdue.

The time has come

Pegged orders, when used wisely, will allow retail clients to compete more effectively with institutional traders for fills, particularly in illiquid markets.

Some academics back this: there is support for the idea that they contribute to the efficiency of capital markets by reducing dealer profits and improve the overall returns to longer-term investors as a class.

Despite alarmist talk from some opposed to pegged orders, there is no evidence that capital markets in jurisdictions in which this type of order is allowed have been harmed. And while there may be additional costs to the exchanges of making such orders available, it seems unlikely, based on the international experience, that these costs will be crippling or irrecoverable.

Exchanges should be allowed to offer this type of order if they feel it will improve their competitive position. Brokerages with a retail client base should be encouraged to pass through the availability of the order type through a “comply or explain” rule applying to all order types that they could potentially allow.

  • James Hymas is president of Hymas Investment Management Inc.

    James Hymas