Amy Kwan, Author and Researcher, Capital Markets Cooperative Research Centre
A recent report from the Capital Markets Cooperative Research Centre has found that US dark venues are able to attract trades from lit venues because regulations, specifically tick sizes, encourage traders under certain circumstances to shift trades into the dark, affecting liquidity on lit exchanges.
This has led to a call for reforms, ranging from tick size changes to a Trade At Rule. Others believe that markets benefit from a multiplicity of venues.
Amy Kwan, author of the report, explained her findings: "Our focus is on a particular trading rule, which is this minimum pricing increment."
That's because on US exchanges, the minimum pricing increment is a cent. If there is additional trading interest there are going to be long queues at the best bid and ask prices unless someone crosses to trade.
"With these long queues, what some traders can do is to bypass these queues and trade immediately in the dark pool. So we are saying when the minimum pricing increment is a constraint on trading in the lit exchanges, people will bypass these queues by trading in dark pools," she said.
That means a lot of volume going to dark pools, which leads to increasing growth and market share.
Kwan said the result is that dark pools would have a competitive advantage when the price is above a dollar, an advantage they lose when the price goes below a dollar, at which point their market share drops correspondingly.
BENEFITS AND DRAWBACKS
Kwan added that traders using dark pools in this way took liquidity away from lit exchanges, and the phenomenon was also seen when higher priced stocks traded.
Andrew Upward, a market structure analyst with Rosenblatt Securities, said that in the US markets this phenomenon did happen but he did not believe "that it happens in the same way as some critics of the way that exchanges operate might describe a violation of price-time priority".
Andrew Upward, Market Structure Analyst, Rosenblatt Securities
Upward explained: "Particularly in the US, but in many other jurisdictions globally, we have a fragmented or competitive market structure where we have many different market venues, many of which have had protected quotes and operated lit markets and many of which are dark pools or other kinds of off-exchange platforms.
"When you have multiple market centres, even though each of those market centres might operate on the basis of price-time priority, when you take them collectively there is no way for one market centre to say 'well hold up you cannot trade with us because our quote arrived after the quote that arrived at our competitor's platform', which would circumvent price-time priority in the broadest sense."
In other words, because price-time priority is the convention on most individual venues, but is not enforced across venues, dark pools can jump ahead of lit exchanges and they can jump ahead of one another in that kind of broad price-time priority and concept.
Upward said that for him the debate was centred on whether the benefits of fragmented markets outweigh the detrimental aspects of the US market model, while conceding that 'the lack of price-time priority across markets is a part of the detrimental aspect'.
But there are benefits to having multiple venues, though traders will need to accept the reality of a fragmented marketplace as well, he said.
"Given the fact that if I post a bid or offer on one of those market centres and I am the first one to bid that price, or offer at that price, I might not fill," he said. "Someone who comes in after me could get that fill because the order is routed to that marketplace, and not the marketplace where I was bidding or offering stock. I think that is just something that today's trader has to kind of cope with."
For Adam Sussman, head of market structure and liquidity partnerships at Liquidnet, tick sizes are an area where US regulators could make positive changes.
"I think the SEC is realising that one- tick-fits-all is not optimal. The challenge of the tick pilot in the US is that there are many other components to it which create three different Petri dishes," he said. "The overall goal therefore of trying to create more flexibility at the tick level is being lost in this controversy surrounding some of the other components of the tick pilot."
Faced with competition from dark venues, the Tokyo Stock Exchange at the beginning of this year implemented a reduction in tick sizes for 80 categories of stocks, going from 1 yen increments to 0.5 yen. Sussman believed Japan showed the way for the US.
He commented that the TSE realised that the tick in the most liquid of its stocks was probably too wide and it narrowed the size.
"Thus far data has shown that narrower tick sizes have de-fragmented the Japanese market and brought more liquidity back to the exchange. So if you wanted to trade within the exchange spread you went to Proprietary Trading Systems (PTS) but now the exchange and the PTSs are on a level playing field."
In the US, said Sussman, tick size was a penny over every stock regardless of liquidity, but he was not keen on a move to a nickel increment.
"I would have rather (see) a two penny tick or a three penny tick, something a little bit more gradual. Increasing the tick size by five times the amount seems pretty extreme," he said.
But for Sussman there was also the issue of what purpose a dark pool should serve, namely that it should further price or size discovery in a way the exchange could not. He further highlighted Canada and Australia as jurisdictions where legislation means executing away from the exchange is only possible if the value found on the dark venues was above and beyond what could be found on the exchanges, in terms of size or price discovery.
Adam Sussman, Head of Market Structure and Liquidity Partnerships, Liquidnet
"Our members are executing at a price or a size that would not be available to them at the exchange without market impact and therefore without hurting the performance of long term investors such as pension funds," he said.
Liquidnet is an off-exchange institutional trading venue that provides value above and beyond what can be found on an exchange, explained Sussman. It was created to meet the needs of buy side firms to match large block size orders with other buy side institutions, mostly at the midpoint and in sizes that are 300 times the amount that is typically crossed at exchanges.
"It is highly different from what is happening on the exchange," said Sussman."Typically those orders are being negotiated and that has nothing to do with time priority or what is really happening on the exchange - there are two big institutional players looking to get a deal done and they need that kind of anonymity and negotiation in order to facilitate a large in size block trade. You will not see those kind of orders being crossed at the exchange."
One idea advocated by Kwan is the Trade At Rule, which the CMCRC study supports - particularly with respect to minimum price improvement.
"What we propose is that you can only queue jump if you provide minimum price improvement, which currently for many regulators just means at the midpoint of the bid and ask prices," Kwan said. "Instead of queue jumping and trading at the same price you have to be penalised to some extent now, which will decrease the amount of dark trading."
Australia and Canada are two jurisdictions being keenly eyed for having instituted such measures.
"There have been a couple of studies that looked at the introduction of this Trade At Rule of the minimum price improvement rule in Canada and in Australia and there was a drop-off in dark pool volumes because they are becoming less competitive."
Dave Lauer, president of KOR Group and a specialist in market structure, is an advocate of the Trade At Rule and believes the US should follow the Canadian and Australian lead.
Speaking on the sidelines of the Trading Show in New York, Lauer said: "In Australia and in Canada you have a Trade At Rule, which says that if you want to execute off exchange -- the principle is you are damaging the lit markets, you are damaging the price discovery process by not posting a quote on a lit market,
"In order to compensate for that damage you have to provide significant price improvement. And to me, philosophically, that makes a lot of sense."
Dave Lauer, President, KOR Group
Lauer outlined the model by which lit markets were being "free-ridden" and some of their liquidity siphoned into dark pools.
"If you have a quote posted on the lit market which comes in at a time T0 and you have a quote, an order, on a dark market that comes in at a time T1 and someone tries to check hit the dark market first because of lower fees or broker routing conflicts, the order that came in later than the one on the lit market will get filled because someone came into the dark market for it."
So lit markets are giving away free rides because trades in the dark market must take place within the public NBBO, which is damaging to market marking.
"It disincentivises market makers from posting lit quotes if they know that they are just going on to get queue jumped by the dark market and those dark markets are not adding anything -- there is no significant price improvement, it is not block trade," he said.
That's backed by the figures, which show that in the US, the average dark trade size is about 200 shares for the top 20 ATSs (it's under 200 shares if you exclude IEX).
The SEC filed a proposal for a Trade At Rule pilot scheme at the end of August that will see tick sizes widened for some stocks with smaller capitalisation. The pilot will run for 12 months, but no starting date has been settled.
Kwan welcomed the idea: "I definitely support any move towards more rigorous regulation of these dark pools and I think it is a good step that the SEC is doing in trying to study this in a little more detail," she said. "It has been obvious in recent years that dark pools are taking a huge chunk of market share away from the lit exchanges which can have detrimental effects for price discovery. So, any move to try to study this just a little bit further, making controlled experiments to do so, is definitely a good move."
Regression discontinuity plots: Mktshare against
This figure plots Mktshare against PriceBin for each trading venue type: exchange (EXCH), dark electronic communication networks (DARK-ECN), block crossing networks (BLOCK), ping destinations (PING), retail market makers (INTERNALIZE), and others (OTHER). For each penny price bin from $0.80 to $1.20, we find the Mktshare for each stock, with Mktshare calculated as the number of shares traded for the trading venue type as a percentage of the total number of shares traded in the price bin. These plots show the mean Mktshare for each PriceBin across all stocks. The vertical line in each plot indicates the $1.00 price level.
Valerie Bogard, Analyst, TABB Group
Valerie Bogard, analyst, TABB Group, warned that the issue was not clear cut, and that any actions would have consequences, some of which might not be foreseen.
"The difficulty with talking about regulatory advantage between off exchange and on exchange is that they each have their own different sets of advantages and disadvantages. Somewhat intentionally ATSs are meant to be breeding grounds for innovation and creativity," she said. "Some of that can kind of seep on to exchanges and help them in creating competition. It is somewhat difficult to look at advantage in isolation instead of taking that on as one part of the whole, the wider arena of differences."
The relationship between on and off exchange is complex, and if there was a simple solution to bring more volume to the lit markets, regulators would have already implemented it, she added.
"When you start exploring solutions or multiple solutions you have to be careful about what the unintended consequences are -- who benefits, who loses, it's not as simple as just mandating, or regulating, just one component."