Stock exchanges have an interest in promoting fair and efficient pre-trade activity. So why don’t they offer tools that will do the job?
The electronic model has now been fully accepted by most exchanges around the world. In the competition to offer liquid and efficient markets, the electronic central limit order book has won.
Or has it? The reality is that most stock exchange trading systems are not really designed to promote liquidity. These systems are passive recipients of liquidity. They do not participate at all at the pre-trade phase - sometimes referred to as the “upstairs market” - the phase where liquidity originates. While the standard central limit order book is very good at matching orders from existing liquidity sources, it is quite weak at unearthing potential suppliers of liquidity beyond those who happen to be looking at the market right now.
Illiquidity is always with us
All exchange traded instruments are in some degree illiquid; any instrument is only liquid up to a certain size. Beyond that size, it can be hard to find counterparties. For minor, small cap stocks, that size is quite small – sometimes the market in those stocks has no bids or offers at all. At the other extreme, even with substantial market depth, large institutional block trades are seldom filled immediately based on existing market supply and demand.
If the fundamental problem is lack of interest, there is no solution to that. We could ask brokers to work harder but in today’s low commission environment, there is a limit to what you can expect from them. If the number of people interested in trading a stock right now is very small, then there is very little that anybody can do to drum up interest in a short space of time. Over time, it can be done – as happens when sponsors promote an IPO – but when there are insufficient potential counterparties who are currently active, it is just about impossible to execute with the degree of immediacy that participants expect in the secondary market. The main forum offered by most exchanges, the central limit order book, is quite ineffective in these circumstances. It advertises only to those people who happen to be watching that stock at that time. This works well much of the time but when that liquidity limit has been reached, the central limit order book offers no solutions.
And so into the breach have stepped the “liquidity providers”. The likes of ITG and LiquidNet and, indeed, the traditional broker who knows how to “work” an order all purport to resolve the illiquidity problem. To various degrees they are effective, but most institutions will argue that executing a block trade remains a challenge. And at the small-cap end of the market, the situation is worse: these crossing networks do not have adequate reach to be of any value to small cap investors. The economics of the small cap market provide poor incentives for liquidity providers to step in.
Pre-trade: the neglected phase of the transaction chain
It is surprising how little attention exchanges have paid to this problem – the problem of service levels at the pre-trade phase. Upstairs trading is more prevalent at all-electronic exchanges than it is at those with a trading floor (Euronext, the archetypal electronic exchange, recently reported 67% of its block trades as being executed at the Paris Bourse, far higher than the New York Stock Exchange[1]). Yet very little of the electronic stock exchange’s technology investment goes into this segment.
It is true that, from time to time, stock exchanges have tried to participate in pre-trade liquidity formation (Optimark and joint implementations of POSIT come to mind) but they have not addressed one of the key problems: that the population of potential liquidity providers is much larger than the firms who are members of – or have direct access to - the exchange.
What is needed is a network model that allows bilateral or even multilateral negotiation between parties, with the degrees of anonymity that are appropriate at each stage of negotiation. This network should use every standard widely-adopted communications device and protocol that is now available: FIX-compliant RFQ and associated transaction types, instant messaging, Voice-over-IP, the varieties of finance-oriented XMLs, wireless access, e-mail alerts, and so on. Moreover, it needs to be structured in a way that allows all stakeholders and potential suppliers of liquidity to have access to it through the most important standard channel of all: a thin client web interface that is relatively quick and easy to install.
It should aim to maximize the potential for a trade to occur by targeting sources of liquidity that subsist in an existing marketplace (even those who are not formal members of the exchange). In short: wherever and whatever potential interest in a trade exists, the network can try to uncover that interest. The most logical “owner” of this service is the exchange itself. Think of it as an exchange-supervised Pre-Trade Network. Implemented correctly, it will exploit the network effect in a way that was never possible in those pre-internet days when the order traffic was policed by the broker or a market maker or a “jobber”.
Deploying this kind of technology over an existing exchange infrastructure is not hard. In fact, exchanges already possess the perfect host for this infrastructure, namely the network that supports its current trading systems. The introduction of a negotiation model into this infrastructure should be neither costly nor risky, for this reason:
A negotiation system can actually be implemented as a discreet application that stands separate from order management and order-matching. Instead of being part of the core trading system, it interfaces with it and with other aspects of exchange operations. The negotiation model for hard-to-execute trades creates a forum for the parties to come together and the negotiation to commence. Once a successful match has been found, the trade can be reported to the exchange using the same interfaces that are available for reporting negotiated trades today. None of this involves tampering with the infrastructure. It is therefore low-cost and low-risk.
The Exchange can do it differently
It may appear that this is just another version of such pre-trade communication systems as Autex or the offerings of Bloomberg and Reuters. While it can perform a similar role to these products, there are some important advantages in the way an exchange could deploy a Pre-trade Network. Here are a few:
- The exchange is the natural place for this kind of network because it is already a focus of liquidity – that is its reason for being in business – and it offers a high quality infrastructure in terms of connectivity, security, reliability, auditing, performance, clearing connections, central counterparty and service levels across the board.
- As an exchange product, owned by and licensed to whomever the exchange authorizes, the service could be provided for free (with fees only charged if a match is recorded), unlike the for-profit products of the major financial information vendors and intermediaries. Indeed it could be made available directly to investors, via a trading participant – a device that would allow the exchanges to really start addressing the problem of illiquidity in small and medium cap stocks. The marginal cost for an exchange to do this is lower than any other entity in the market place because of the pre-existing infrastructure.
- Because the exchange owns the network, it can assign market makers to the negotiation process operating under exchange rules that can enhance investor confidence; this can be complementary to whatever market making services are available in the regular market.
- It would be deployed specifically with compliance in mind i.e. both regulators and participants would actively look to the Pre-Trade Network as the means for monitoring pre-trade compliance because it has a natural built-in audit trail (since all communication would be electronic) thus addressing all the regulatory concerns about pre-trade communication.
- It can be set up to interact with the central liquid markets in ways that improve execution and liquidity formation. For example, call auctions can be triggered when interest rises to a certain level and, as mentioned above, best execution rules can be enforced.
- While it could be configured to interface with external products like Autex, the ultimate benefit would be for the exchange to bring “upstairs” business back onto the exchange. A trade negotiated through this exchange-sponsored network is far more likely (indeed can be mandated) to be reported to the exchange.
- It is adaptable, unlike a conventional auto-matching system, to a very wide variety of instruments and products, including baskets, strategies, stock lending and even non-fungible products or services, because a negotiation model has much greater scope for handling unstructured data than does an auto-matching system. It can thus enhance the ability of stock exchanges to attract OTC business (something which, to their credit, several derivatives exchanges have begun to address, notably LIFFE, CME and Eurex).
The term is clichéd but this network, implemented with the exchange’s conventional auto-matching technology really does become a “virtual trading floor” in a way that has hitherto not been true of exchange technology implementations. It preserves the best qualities of the old trading floor and dispenses with all the drawbacks.
There’s always a catch
Many good initiatives in the world of trading fail to be realized, or at least take longer to be adopted than they should. There are normally two reasons for this: (a) the tyranny of the status quo, i.e. vested interests who stand to lose out if the initiative is implemented and (b) the wariness of regulators who are slow to approve change until they have undertaken exhaustive and exhausting analyses of the pros and cons.
This concept does not suffer from either of these problems. Theoretically, voice-based brokers or dealers who already offer a quote service may boycott the service in the belief that it diverts market interest away from their desks but it is hard to see why they would want to do this. The Pre-Trade Network merely offers them another sales counter in the marketplace. The buy-side similarly would not have any incentive to object to the presence of such a platform – again, they only stand to benefit from another venue for improving execution.
As mentioned above, exchanges stand to benefit from the increased price formation capacity that the system can provide. Concerns about upstairs crosses and the removal of liquidity from the central order book can be addressed by the adoption of appropriate rules, as with any market. In other words the “threat” of upstairs trading is already there – the Pre-Trade Network actually provides a vehicle for better management of that activity.
Regulators are of course unpredictable and even with the best planned deployment, anything that affects market activity and involves traders negotiating prices will inevitably attract their attention. The important issue will be to provide the appropriate degree of transparency and auditability that the regulators require. If implemented as a tool that helps to formalize inter-dealer and inter-investor communication, regulators ought to welcome it. Indeed, the Pre-Trade Network could be deployed in a way that brings regulators and participants together – the formal platform for ensuring compliance with pre-trade regulations.
If this is such a good idea…..
Why isn’t everybody doing it? This is an important question, not least because the technology required to implement this is not terribly radical or ground-breaking. It can all be built from components that already exist.
We think the reason this has not already happened is simple: the idea that the exchanges can have a role in pre-trade activity has not yet been fully absorbed. As recently as 10 years ago, the exchanges were the owners of the proprietary technology that linked participants together but, in those days, they lacked the business motivation and ability to conceive of and implement a notion like this. The exchanges focused on their core business – order execution for liquid markets.
On the other hand, a local implementation of the Pre-Trade Network by a third-party supplier would have faced insurmountable difficulties in trying to convince a critical mass of users to adopt a platform like this, install it, test it and launch it.
Participants were simply not familiar with the notion of a networked user community. Even if the exchange had built it, its users would not have known what to do with it. For the vast majority of users, there was no internet and no FIX protocol. Every network was proprietary.
Then along came the internet and everything changed. Direct participation by all stakeholders became possible and networked communities began to spring up everywhere.
And so now a broad, inclusive, exchange-owned Pre-Trade Network can be implemented cheaply and at low risk. The exchanges should get into this business.
[1] Source: Bessembinder, H., Venkataraman, K., 2003. ‘Does an electronic stock exchange need an upstairs market?’, Journal of Financial Economics, 12-13 (http://jfe.rochester.edu/03109.pdf)
