October 4, 2010
Good morning. I thank the Wholesale Markets Brokers’ Association for inviting me to speak at the “Swap Execution Facility Conference.” Swap execution facilities – or SEFs – comprise a new category of trading platforms that was established in the Dodd-Frank Wall Street Reform and Consumer Protection Act. In so doing, Congress sought to bring greater transparency, more efficient markets and better pricing for end users and to lower risk in the financial system.
As we work to develop requirements for SEFs – with broad input from the public – it is important that we keep in mind the context of the unusual market events that took place on May 6 of this year. Those events created significant uncertainty for American investors.
As outlined in the joint staff report released this past Friday by the Commodity Futures Trading Commission (CFTC) and Securities and Exchange Commission (SEC), there were three chapters of the May 6 market events:
I will focus primarily on the second point – the liquidity crisis in the broad market indexes. In discussing SEFs, though, I will draw upon some of the important lessons related to the second liquidity crisis – in individual stocks– as well.
The equity index futures market is interlinked with the markets for exchange traded funds (ETFs) and the underlying markets for individual securities. There is far more liquidity in the futures market than the ETF markets, as evidenced by the depth of the respective order books. On May 6, the E-Mini’s order book started out with about 20 times more orders, measured by total notional amount, than there were in the S&P 500 SPDR ETF.
The morning and early afternoon uncertainty in markets had drained liquidity across the markets, most dramatically in the E-Mini, whose buy side order book by 2:30 p.m. was only 45 percent what it was earlier in the day.
What happed next in this fragile market is important: a large fundamental trader came into the E-Mini market to hedge about $4.1 billion of equity market exposure by selling 75,000 futures contracts. Like most traders, this large trader needed to execute its transaction through an executing broker, as it did not have direct access to the trading platform. The trader and its executing broker had a series of choices to make when determining how to execute the sell order. In this case, they chose to use an automated execution algorithm rather than having traders manage the position. Another key choice it appears they made was to put the entire order into the execution algorithm rather than breaking down the trade. And lastly, they chose to use an algorithm that did not take price or time into consideration. It was just based upon a limit of 9 percent of the trading volume calculated over the previous minute. In essence, once the $4.1 billion order was entered into the algorithm, it was as if it was on auto-pilot in an otherwise fragile market.
During the next 13 minutes, the large trader sold about 35,000 contracts – with a notional value of about $1.9 billion – and the market went down another 6.4 percent. During the same time – as the report shows – a diverse group of other fundamental sellers sold a net of about 45,000 contracts for a total of 80,000 contracts sold. Fundamental buyers, however, bought a net of only about 50,000 contracts. Fundamental buyers and sellers are entities that the report identifies as holding their positions overnight, rather than buying and selling during a single day. Though market makers, including high-frequency traders, stepped into buy some of these contracts, they sought, often in seconds, to flatten their positions. It took time and significant price concessions for opportunistic buyers – those willing to hold position for more than a few seconds but not necessarily overnight – to purchase the net difference of 30,000 contracts of the sell orders that had not been purchased by market makers. The buy-side order book of the E-Mini was rapidly exhausted.
Just after 2:45, trading on the E-Mini was paused for five seconds when the Chicago Mercantile Exchange’s Stop Logic Functionality was triggered. At that time, its buy-side order book was down to only 1 percent of the size it was earlier in the day. During the pause, the E-Mini’s order book began to replenish. The fundamental buyers started to be more balanced with the sellers, and the markets began to rebound.
So, what are some of the lessons from the first liquidity crisis of May 6?
One key lesson is that, under stressed market conditions, the interaction between the automated execution of a large sell order and trading algorithms can quickly erode liquidity and result in disorderly markets, especially if algorithms use volume as a proxy for liquidity. The events of May 6 demonstrate that, in volatile markets, high trading volume is not necessarily a reliable indicator of market liquidity.
Just as there’s a difference in tennis or ping pong between the rally before the point and the point itself, in markets, there’s a difference between a position going back and forth between market makers and a position actually being bought by a fundamental buyer who will hold it overnight. Much of the volume on May 6 was just positions being moved back and forth over a matter of seconds between high-frequency traders and other market makers. This is what our economists refer to as “hot potato volume.” For the large trader’s order to actually be absorbed by the market, it had to find fundamental or opportunistic buyers who were willing to hold the position at least for more than a few seconds.
We have asked the CFTC-SEC Joint Advisory Committee on Emerging Regulatory Issues to consider the report and make recommendations to both agencies. Some of the ideas that I look forward to hearing about from this expert panel, as well as from market participants and the public, include:
1. Requiring executing brokers to have an obligation to enter and exit in an orderly manner. On May 6 – as is often the case – the large customer did not execute the trade itself, but used an executing broker. Should executing brokers have to adopt certain trading practices when executing a large order by use of an algorithm, such as price or volume limits? Exchanges have maximum order size limitations and price banding that prevent entry into the trading engine of an order that exceeds a predefined maximum quantity or a price difference from the current market – should executing brokers have similar limitations? Should they have an obligation to monitor and make non-disruptive trading judgments? Might customers also have an obligation for orderly execution, and if so, under what circumstances?
2. Increasing visibility into the full order book, either in aggregate or in detail. Currently, for most products, traders in the futures markets can see up to the tenth offer or bid in an order book. One of the problems on May 6 was that a large sell order overwhelmed existing liquidity in the E-Mini order book. Might fuller visibility of the order book lessen the chance of market disruptions resulting from such large buy or sell orders in the future? If so, what is the best way to accomplish a goal of increased visibility into the full order book?
3. Potential revisions to market pauses – either for single exchanges, such as stop loss functionality logic, or for cross market circuit breakers. On May 6, the E-Mini stopped trading for five seconds. This was a critical moment in the markets. The pause gave the order book time to replenish. Would it have been better if the pause came earlier? If so, what conditions should trigger a pause? Should cross market circuit breakers be adjusted from their current 10 percent limit?
The Dodd-Frank Act
These events inform the discussion on the transparency and market structure provisions of the Dodd-Frank Act. I recognize that many in the audience may be interested in what we’re doing over on 21st Street to implement that essential legislation. The bill requires the CFTC and SEC to write rules generally within 360 days. For those of you keeping track, we have 285 days left. On Friday, the Commission voted on three rules to implement the Dodd-Frank Act. I thank my fellow Commissioners and CFTC staff for all of their hard work on these rules. We anticipate putting out proposed rules SEFs and real time reporting in November or December.
One of the central reforms of the Dodd-Frank Act is the trading requirement for clearable swaps. The Act mandates such swaps to trade on exchanges or SEFs if they are made “available for trading.” This brings significant benefits to users of derivatives and to the American public. The more pre-trade transparency there is, the more likely there will be additional competition in the markets and tighter pricing for end-users. The statute says that one of the goals of swap execution facilities is “to promote pre-trade transparency in the swaps market,” though it appropriately authorizes the CFTC to write rules to facilitate block trades.
We now estimate that approximately 30-40 new entities will register as SEFs or designated contract markets (DCMs). Last month, I said in a speech that we estimated the number of new entities registering as SEFs or DCMs would be 20-30. Since then, we have heard from parties suggesting that the earlier estimate might have been low. A look around this room may suggest that we may still be low.
Some have asked: what is a swap execution facility? I think the statute is clear. It’s “a trading system or platform in which multiple participants have the ability to execute or trade swaps by accepting bids and offers made by multiple participants.” It is not a facility where many buyers have access to only one dealer.
Real Time Reporting
In addition to trading, Congress also has been very specific that market participants and end-users will benefit from real time reporting. Such post-trade transparency must be achieved “as soon as technologically practicable” after a swap is executed to enhance price discovery, other than a possible delay for the reporting of block trades.
The Dodd-Frank Act brings post trade transparency to the swaps market by requiring real time reporting for both cleared swaps and bilateral swaps. As we write the rules on real time reporting, the statute mandates that we “ensure that such information does not identify the participants.” Thus we are required to both promote price discovery and protect market participant confidentiality.
Further, transparency – both through trading and through real time reporting – enhances clearinghouses’ ability to appropriately manage their risk.
As we hear from the public regarding SEFs, we are interested in two sets of cross-market issues that arise out of the goal of promoting competition amongst and between trading platforms. The first set of issues, which is informed by the events of May 6 and particularly the liquidity crisis in individual securities that day, relates to the functioning of a transparent market where there is likely to be multiple trading venues. How do we best accomplish the benefits of competitive markets while avoiding some of the shortcomings of a fragmented market structure? How do we ensure that end-users most benefit from pre-trade transparency across multiple SEFs? What obligations are appropriate to place upon swap dealers and introducing brokers with regard to using SEFs in best achieving their clients’ goals for price and execution? How do we ensure that SEFs have impartial access to all market participants?
The second set of cross-market issues arises from the challenges of trading platforms fulfilling their self-regulatory functions in a multi-platform environment. Under current law, DCMs are permitted to contract out some of their compliance function while maintaining their responsibilities. How should we approach this with regard to SEFs? Furthermore, how best should the CFTC address cross-market compliance functions?
The CFTC faces challenges in the months ahead, but we are prepared and geared up to meet those challenges. I look forward to continued dialogue with the public and interested parties as we work to implement the Dodd-Frank Act.
Thank you, and I’d be happy to take questions.
Last Updated: January 18, 2011