Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
Parliamentary Joint Committee on Corporations and Financial Services
Oversight of the Australian Securities and Investments Commission

COMERTON-FORDE, Professor Carole Anne, Australian National University


CHAIR: Welcome and thanks for making the time to come and speak to us.

Prof. Comerton-Forde : By way of background, I do research in market structures and have been doing that for about 15 years. My research focuses mainly on the US and the Australian markets, but I am also pretty familiar with the European and Asian markets so I can give you a flavour of those also. I have recently been an economic consultant for ASIC, working with them on market structure issues and in the development of the market integrity rules for the launch of competition in the Australian equity markets. But I am speaking to you today in my capacity as a professor of finance and not representing ASIC. Obviously, the views I express are mine and not the views of ASIC.

Ian Holland heard me speak at the ASIC summer school earlier in the year and asked me to speak to you on similar topics to those I covered at the summer school. I think you have a copy of the slides that I used at the summer school, and probably those give you the flavour of the richness and complexity of the issues involved in market structure and competition. I am not going to speak to those slides because there is a lot of information there. Rather, I want to pick up on three key issues that I talked about. They are, firstly, giving you a big picture sense of what is happening globally in equity markets and how we might learn from that for what is going to happen in Australia and how we should regulate in Australia. Then I want to talk about two fairly hot and contentious issues—dark pools and high-frequency trading—because I think they are issues that are often poorly understood and it is useful to get further background on those.

At the outset I want to say that competition has had an extremely positive effect on equity markets around the world. I am going to talk mainly about some of the problematic areas. You may get a sense that it sounds very negative, but that is not the impression I want to convey; I think competition in equity markets is very good and very important for the development of the markets, but there are risks and issues that need to be managed. Those are the things I want to focus on today.

Firstly, the global trends. The Australian market has a long history of being an innovator in financial markets. Australia was one of the first markets to automate and one of the first markets in the world to demutualise. Despite that history of innovation, on the issues that I am talking about today the Australian market is lagging; we are probably three to five years behind the European and the US markets.

In some senses in this case that is a good situation to be in, because we can observe what has happened there, identify any issues and try to maximise the benefits and minimise the downsides. Equity markets are constantly evolving. We saw in the late 1990s, early 2000s exchanges go public and demutualise. That really had a significant effect on the way markets operate and the way exchanges do business. It has really set the scene for the changing landscape in markets around the world. Change in the US and Europe has been further accelerated by regulatory change, so the regulators in those two regions really pushed for competition between markets as a way of developing those markets, and that is coming much later here than it has in those markets. The approaches taken for those two different areas were quite different, and I do not want to go into the details of the differences, but they drove different outcomes. In a big picture sense the outcomes have been very similar across the US and Europe, and they are the things that I want to focus on.

Promoting competition means that there has been a proliferation of new trading venues in the US and Europe. In the US there are now over 50 different places where you can trade your stock and in Europe over 100 different places. That creates issues in terms of the fragmentation of the market and being able to access liquidity. New entrants have come into the market in both places and, typically, there have been one or two entrants who have been very, very successful. In the US that was BATS and Direct Edge—probably not names that you know, but together they represent 25 per cent of US equities trading, which is something not known even by a lot of investors in the US. Triex has been the most successful participant in Europe in some markets; they have been most successful in London, where they have captured close to 50 per cent of the market. They are very successful businesses.

What is the impact of that regulatory change and what impact has that had on the businesses in the incumbent markets? I draw your attention to the three figures I distributed today. One shows the changes in market share for exchange businesses in New York, France, Germany and Toronto. It is just a chart with a few lines on it. The zero on the horizontal axis is when competition commenced in those countries. You can see the incumbent businesses very quickly lost huge amounts of market share—within two years those markets had lost between 25 and 50 per cent of their business, which clearly changes the dynamic in the market. I think this picture gives us a strong indication of what might happen in Australia when Triex launches later this year. I will talk a bit more about ASX and how they are prepared for that, so we may not expect such an extreme change, but I do think the impact will be very significant and I do not think everyone appreciates that yet.

So why have the new entrants in the various markets been very successful? There are probably a few dimensions on which they are competing that have driven this outcome, the most important of which are price, ownership and technology, all of which create issues that we need to think about. Trading venues have competed very aggressively on price, so we have seen very big reductions in the cost of trading in markets where there is competition, and that is clearly a very good thing for the participants in the market. Often these fee savings are not passed on to the end investor—it is the brokerage community that gets the benefit of that—but hopefully, with competition and brokerage, you will see that being passed on to the end client as well, through their commissions.

We have also seen innovation in the way in which exchanges charge for their services, so the development of what is known as maker-taker pricing, which has been contentious in Australia, at least, where you charge different fees for people who provide liquidity of the market versus those that take liquidity from the market. Those models have been put in place to incentivise people to move to the new markets and conduct their business there. Typically these new entrants have been owned by either investment banks or by very large trading firms, so they clearly have an incentive to drive down fees and to send order flow to the markets which they own.

Technology has also been a very important dimension of the change in the markets. The new entrants have come into the market with superior and much faster technology and have innovated significantly in the way they offer trading services to the market. Generally this has been good, but there have been some risks along the way. There has been a massive change in the types of orders that are available to investors in the market: hidden orders, pegged orders, flash orders. They have generally been good, but there have been some instances, such as with the flash orders in the US, which have been problematic. It is important that the regulator is on top of those sorts of things and can step in and make changes quickly if needed.

Unlike the incumbent markets in other countries, ASX is much better prepared for these sorts of changes. Rather than waiting for competition, they have innovated with the threat of competition. In the last 18 months ASX have cut their fees and introduced new trading technology and new trading platforms, offering new services to the market. So they are probably better positioned than other markets.

The two other major trends that I wanted to talk about were dark pools and high-frequency trading. Before I do that, I think it is useful to get a sense of how markets operate traditionally and how that is changing, and what the characteristics of the market are that we should be looking for in regulating. Historically, equity market trading, as you know, took place on the floor. Trading now takes place on trading screens and the most common form of trading is an electronic limit order book, which is what the Australian Stock Exchange uses. That is a very transparent process in which orders get sent to the market, and people who observe that trading process can see the demand for and supply of stock. That process is very important in enabling people to identify liquidity and to know when is a good time to trade, and also for ensuring that there is an efficient price discovery process. That is also very important for ensuring investors get a good deal and ensuring that prices are set in a fair way when we are trying to price assets outside the financial markets. That was the way trading took place historically. Because it was all centralised in a single pool, everyone was coming to the same location, so participation in the markets involved lots of different types of traders. You had retail traders, you had institutional traders, you had speculators—all sorts of people participating in and ensuring that the market was operating well and that prices were efficient.

The two most important qualities that I think we need to keep in mind, as we look at how markets are evolving, are that liquidity and the price discovery process. You can get a sense from the picture where you see the decline in traditional markets and the proliferation in the number of venues—50 venues in the US, for example. It is now becoming much more difficult to find liquidity and trade across many markets. That problem can be solved through the consolidation of information. All information can be brought back together and reported to the market and people can still trade on that. But where things perhaps start to become a bit more difficult is when you have non-transparent trading—trading where you do not observe the trading process. That is what dark pools are essentially.

I guess 'dark pool' is an unfortunate term because it has sinister undertones, but really it is not. It is not a new concept; it is just a new term. Dark pools are essentially trading venues where there is no pretrade transparency. People place their orders to buy and sell stock in the market, but no-one sees those. The orders interact with one another electronically and when there is a match they get reported to the market. After the trading takes place it becomes transparent to everyone. Thinking about the price discovery process, you can see already there is a loss of information.

Dark pools, or non-transparent trading, have had considerable attention in the media, both here and overseas. Concerns have been raised about the potential negative impact of that. In the US, dark pool trading, or non-transparent trading, is now a very significant part of the market. It has grown from about 17 per cent in December 2008 to about 30 percent in December 2010. It is a substantial part of the market and it has raised concerns about both price discovery and liquidity.

When people talk about dark pools they often lump them into one category. I think it is useful to break down that category to get a better sense of what they offer to the market and think about how they should be regulated. You could categorise them in a range of different ways. Very basically, you can think of them as those that offer services to large block traders and those that do not. Those that offer services to large institutional traders have existed in some form for a long while and are a very important part of the market. The reason for that is that it provides opportunities for investors trading large amounts of stock to do so without having a price impact. If I am a large institution and I want to buy a very large amount of stock, I do not want to move the price simply because I am a large buyer or seller. The dark pool is an effective way for me to be able to do that.

This type of trading has always been an important part of the Australian market. For trades over $1 million these dark facilities have always been available. Historically they have been over the phone, now there are automated systems where orders interact with one another anonymously and electronically. When a sufficiently large order is found, the two parties electronically negotiate the price at which they want to trade. Typically they will trade with reference to the price in the underlying market—in the Australian case, the ASX.

This is obviously a very important and useful part of the market. Where things become a bit more concerning is due to the fact that the development of technology now means that dark pools are available to everyone. This creates more risks, both because there is no longer a strong motivation for why that sort of trading is allowed and because there is no longer a lever for ensuring that sufficient transparency and liquidity are going to the main market. Potentially you get too much order flow going into dark pools and this is where the concerns arise in the US. In the US, 30 per cent of trading is done in the dark and 18 per cent is done in internalisation engines—these are engines that match small sized trades. In the US, the SEC reported that close to 100 per cent of retail orders are not sent to public markets; they are executed in internalisation engines, matched between the broker's own account or the broker matching orders with other clients and not sending them to the public market. The risk with that is that there is no longer liquidity being displayed to the markets and prices in the market are no longer efficient.

Should we be concerned about this sort of trading? I think the answer is clearly yes. The internalisers—the parties that are internalising the trades—argue that they are offering value to their clients by offering better price improvement. Let us think of a very simple example—say, you want to buy a stock and the stock is currently priced in the market at $10 to buy and $10.05 to sell. If you wanted to buy the stock in the market, you would need to pay $10.05. An internalisation engine would give you the opportunity to trade within those prices, so rather than paying $10.05 you would pay somewhere between $10 and $10.05. Potentially, that is very valuable. You could get a better price than what is available in the public market.

What happens in reality when there is no regulation around that process? In the US, a very large amount of that internalisation takes place at tiny price improvements—maybe one one-hundreth of a cent per share. If I am trading 100 shares, I might get a cent price improvement on my whole parcel of stock. You might say: 'Any price improvement is a good price improvement. I'm happy to have this cent rather than it being in someone else's pocket.' But what that does not capture is the cost to the market for the guy who was offering his order to sell at $10.05 who is no longer able to trade. He no longer has any incentive to display his orders to the market. What is going to happen is that prices will become less efficient. Rather than that spread being $10 and $10.05, it might go out wider to $9.90 and $10.10, in an extreme case. If we take out a large amount of volume from the market in terms of internalisation, we disincentivise people from providing their liquidity to the market and potentially create significantly increased costs and lower liquidity in the market. It is important to think about and monitor that internalisation process and manage that.

The other figure I provided is figure 2. This is a submission to the SEC by one of the US brokerage firms that explains this process. A retail order from an investor in the US goes to their broker dealer. Rather than sending it to the market, they will try to internalise that order themselves. They will obviously pick the best orders—the orders that they think are the least informed—and then if they decide they do not want to execute it, they send it on to another broker. You can see they have a number of layers where it takes a while for it to get to the public market. In terms of time it is a very short elapsed time—maybe only a second—but in terms of the orders that are getting to the public market, they are now very informed markets and the risk of trading on an exchange are much greater, discouraging people from displaying their liquidity to the market.

In Australia, brokers have always been able to internalise but there have been controls in place to minimise the level of internalisation. In recent times and as we move to a more competitive environment, brokers are building more sophisticated internalisation engines. We would expect in the absence of any regulation for more order flow in Australia to be internalised. That is something important to consider, watch and get right. In its consultation with the market at the end of last year, ASIC proposed a $20,000 threshold. Orders could not be internalised unless the order was at least $20,000. That was not welcomed by the market. So in subsequent feedback on the consultation process, ASIC indicated that it will consult further with the market to find an alternative mechanism to keep in check the rate of internalisation and ensure that price discovery and liquidity are protected in the displayed markets.

Finally, moving to high-frequency trading—again, another term which creates ill-feeling. High-frequency trading, or HFTs, have received significant media attention both here and overseas. Much of that attention has been negative and much of it unwarranted, in my view. High-frequency trading is not a single category of trading. We need to think about the differences in the way these types of traders participate in the market. They provide different functions and trade in different ways. The thing they have in common is that they all trade at very high speed and require managing their risks on very short time horizons. They will be sending lots and lots of orders to the market and increasing the technology requirements and capacity requirements for everyone in the market—participants, market operators and regulators. In the US and in Europe these have become very dominant players. In the US, they are estimated to represent about 50 to 60 per cent of the market. In the UK, they are estimated to be about 35 per cent. Early last year, the ASX estimated between three to four per cent of the Australian market was high-frequency trading. It is not possible to quantify with certainty but estimates these days are around the 15 to 20 per cent mark. The introduction of Chi-X and the introduction of ASX's PureMatch will increase the level of high-frequency trading in the market. We saw at the end of last month, Getco, one of the biggest high-frequency trading firms in the world, become a participant of the ASX. We would expect this trend to continue in Australia.

The academic evidence on high-frequency trading is predominantly very positive. They add a lot of value to the market in regard to their liquidity, tightening the spreads and reducing trading costs, and there is no evidence that they create an adverse market in terms of volatility. Despite those very clear benefits, there are risks. The main thing that needs to be thought about is the impact they have on the market in terms of the increased technology costs which need to be borne by everyone. Thought needs to be given about how to manage that and how to ensure that those costs are borne by those that create the costs rather than by the market as a whole.

The other thing that high-frequency trading does is to create a more complex environment for the surveillance. Again, the regulators need additional tools, technology and understanding of the market in order to monitor the market effectively. In general, the trends with high-frequency trading, dark pool trading and competition increase the need for better understanding of the markets and better access to data. It is important as we go forward that ASIC are able to obtain that information and that data to ensure they keep abreast of the market and abreast of developments in the market, and that they are able to be proactive and change quickly in response to the market dynamics.

That is all I have to say, but I am happy to take your questions on any market structure issues or concerns you might have.

CHAIRMAN: Thank you very much for that comprehensive and informative presentation. You raised the issue of electronic price discovery and liquidity, particularly in large trades, and that the price is electronically negotiated. Could you explain further how that works?

Prof. Comerton-Forde : Different trading venues will manage that in different ways, but a common mechanism are pools that are only open to institutional investors. They will send their orders to the market—a loose term, not in a legal sense but in the sense that there is a pool of liquidity—and where there are two parties that are wanting to trade a sufficiently large size, they both get a message saying, 'There is a party wanting to trade on the other side,' and then they negotiate. It is almost like sending an email anonymously to someone. Typically, these types of venues will trade somewhere between the two prices in the displayed market. These venues are not provide any price discovery to the market; they are about providing volume discovery—allowing large institutions to find one another without moving the price, which is very important.

CHAIRMAN: Do you have a regulatory view on how dark pools operate in different markets? Is there a particular course that seems to be the right course of regulation?

Prof. Comerton-Forde : One thing that is very important is that they are regulated in their own right. The current Australian regulatory environment does not provide for a dark pool class of markets. They are currently operating under the rules of the ASX. It is largely brokers operating a venue where they match orders and then they report them to the ASX.

Prof. Comerton-Forde : One thing that is very important is that they are regulated in their own right. So the current Australian regulatory environment does not provide for a dark-pool class of markets. They are operating currently under the rules of the ASX. So it is brokers, largely, operating a venue where they match orders and then report them to the ASX. In the US and Europe they have specific regulatory frameworks. In the US it is regulation ATS and in Europe it is multilateral trading facilities. It is a market designation where they operate under different rules from exchanges because they are only offering part of the services that an exchange offers. I think that is an important thing to have in place to ensure that you can appropriately regulate those entities.

Senator BOYCE: Is that almost a regulation that got left behind when the regulation of the ASX moved to ASIC?

Prof. Comerton-Forde : My understanding—and I am not an expert on this—is that to change this it requires Treasury to take action. It is not within ASIC's domain.

Senator BOYCE: I was just thinking that we probably should have put that in the bundle that was moved to ASIC's control.

Prof. Comerton-Forde : Yes. ASIC regulates the brokers who are operating these venues, but they are not regulating the venue per se. So the brokers have an obligation to the market and to their clients and all that sort of stuff, but they are not regulated as a trading venue. And they probably should be.

CHAIRMAN: I made a note when you were talking about internalisation processes and the requirement to monitor those. Could you just explain how that might actually look and how you would monitor that process. Would you require that the markets themselves disclose that information that is done under regulation, or is it really a case of exploratory monitoring after the fact? How might that work?

Prof. Comerton-Forde : There are a couple of issues. One is regulating the conditions under which you allow internalisation. My view is that, in order to protect the displayed markets and ensure you have efficient price discovery, the internalisation needs to offer something extra. They cannot just be trading at the same price as the displayed market. So they need to offer something better, and not one one-hundredth of a cent better; they need to offer something meaningfully better. So that is one issue. The other issue, and I guess this is more what you are asking about, is: how do you manage and regulate that process of ensuring that the client's orders are being dealt with fairly? Is that your question?


Prof. Comerton-Forde : It would be part of the best execution obligation of the broker, and you would have to manage and monitor that after the fact.

CHAIRMAN: So it would use the same technology that is currently used to monitor the market? Is it as simple as that? Is it just a matter of applying the same technology, software and so forth as is there now?

Prof. Comerton-Forde : Yes and no. You are not seeing the orders in the public markets, so you cannot monitor the order flow. You can only monitor after the trade gets reported. ASIC have put in place monthly reporting from dark crossing engines in the Australian market. So they need to report aggregate statistics. But that is at quite an aggregate level, not at an individual order level. So it would have to be a more manual process to manage that.

CHAIRMAN: I think this will be a good question for ASIC!

Senator BOYCE: They do not do the work on dark pools, I know.

Senator STEPHENS: Thank you very much—that was a great presentation, and it really raised lots of issues for us. Can I go to the third key point that you were making about high-frequency trader participation and that point that you made about technology costs. What kind of advice would you have to the committee about how you can ensure that the costs are being borne by the people who are creating those cost pressures as opposed to the whole of the market?

Prof. Comerton-Forde : I think one of the differences between high-frequency traders and other traders is the rate with which they send orders to the market and also the rate with which they cancel their orders. So they might send a bunch of orders to the market and, when prices move, they need to adjust their orders, so they will send another set of cancellations and new orders to the market. So that is where the growth in message traffic comes from. I think one sensible proposal is to charge a fee when the ratio of orders being placed to cancelled orders becomes excessive. The only people that would be subject to that charge would be those using excessive message traffic and cancellation of orders. I know other markets have talked about the possibility of having a flat transaction tax or having minimum times in which orders need to be put in place in the market. I think both of those are bad ideas and will have a negative impact on people's willingness to provide liquidity. Looking at the extremes, which is where the costs become worst, is a better place to try to start managing the issue.

Senator STEPHENS: What kind of a fee would you suggest would be appropriate?

Prof. Comerton-Forde : In terms of dollar value?

Senator STEPHENS: Yes.

Prof. Comerton-Forde : I do not think I could come up with anything off the top of my head.

Senator STEPHENS: Or a proportion?

Prof. Comerton-Forde : I think it would be a matter of looking at normal patterns of behaviour in terms of the ratio of order placement and order cancellation and finding where things become extreme. I have not done that analysis. I am not really in a position to answer. One thing to bear in mind is that in the US markets some of the individual exchanges would like to impose a fee like that but they cannot because if they do they lose business to their competitors. I think these types of issues need to be dealt with at a whole-of-market level not at an individual operational level. Clearly there are reasons for an individual operator not to try to address this, because it has an adverse effect on order flow coming to their markets

Senator STEPHENS: You talked about the surveillance challenge increasing. What exists now to monitor that significant change that has already happened and the potential for those cancellations—the huge transaction load that is happening in this space?

Prof. Comerton-Forde : As part of the transfer of supervision ASIC adopted SMARTS technology. That is a surveillance tool that essentially takes every individual message that comes into the market and they have the capacity to replay the market, analyse that data and investigate it. So they already have tools to do that. It becomes more complex though, as I said, as you have multiple markets. So when we have the introduction of Tri-X and the introduction of PureMatch there are technology issues in bringing those things together. Once those sources of data are brought together it is the same process, but obviously the volume of data becomes much larger.

Senator BOYCE: Professor, I think I may have misunderstood you at the beginning. You said the ASX had been quite innovative but that it was 3 to 5 years behind the US and Europe. Can you explain what you meant.

Prof. Comerton-Forde : I mean more in the sense that the US and Europe had major regulatory reform to drive competition. Australia has not had that.

Senator BOYCE: So it was about where we are in the competition spectrum.

Prof. Comerton-Forde : That is right. Because of the competition, markets have developed and the ASX has not had that pressure.

Senator STEPHENS: My question follows on very neatly from that. We have got the ASX and we are about to see Tri-X. Do you expect other players to come into the market in Australia?

Prof. Comerton-Forde : That is an interesting question and I do not know the answer. My expectation would be that the Australian market cannot sustain more than three or four markets. I guess it will be up to the brokerage community to decide whether or not they are going to support other new entrants. It is a small market, so you cannot have multiple competitors. Having said that, the Canadian market, which I guess is fairly similar in size and scope, has about seven or eight different trading venues—although three or four of them have most of the business. We could get to that sort of position as well.

CHAIRMAN: Professor, can I thank you very much for your presentation. The information you have provided, including the document you have given us, I will take as tabled evidence for the committee and incorporate that. Again, thank you very much for your very informative presentation.

That concludes this hearing. I thank everybody for their participation. The committee has resolved that questions taken on notice should be returned by Friday, 22 July.

Committee adjourned at 18:19