To understand why trade-at is a bad idea, one need look no further than its proponents: the for-profit exchanges and their largest customer base – high frequency traders. The exchanges, which before demutualization were once considered a bastion of market structure integrity, today answer to shareholders, just like the broker-dealer community.
In the past decade, exchanges have seen their market share deteriorate due to competition and innovation by broker-dealers, fostered by the SEC’s adoption of Reg NMS in 2005. When broker-dealers internalize their customers’ orders, the for-profit exchanges lose out on exchange fees. Exchanges’ best customers today are not retail or institutional investors; they are the high frequency shops that pay for co-located servers and feed the exchanges’ maker-taker pricing model. And those high frequency traders would like nothing more than to see more & more displayed order flow for them to prey on.
Who is against a trade-at rule? Just about everyone else: both the buy-side and the sell-side, and most retail investors in the know. Within the past few weeks, the largest trade association for mutual funds (ICI) and the largest trade association for the broker-dealer community (SIFMA) both filed comment letters opposing a trade-at rule, as it would peel back several of the benefits investors have come to expect in the past decade.
In large part due to broker-dealer internalization practices, it’s never been a better time to be a retail order. Liquidity is prevalent, spreads are narrow, executions are fast, transaction costs are low and opportunities for price and size improvement abound. Internalization is the primary reason behind the $8 trade and empowerment of the individual investor. And a trade-at rule would do away with all this.
Let’s compare retail execution data from 2006 (pre-Reg NMS implementation) vs. 2010 (post-Reg NMS) for the largest retail market makers (Citi’s affiliate ATD, Knight, UBS and Citadel) for all stocks and all “marketable” orders (market and marketable limit orders up to 10,000 shares).
Market makers executed roughly three times the volume in 2010 vs. 2006 because order sending firms like TD Ameritrade, Scottrade and E*Trade have embraced the reduced cost of execution via a market maker and the price improvement of their client orders.
The speed of execution has declined approximately 50 percent, which indicates both an increase in available liquidity at the inside market (NBBO) and a commitment of more capital by the market makers. Average spread size has also declined substantially, by approximately 23 percent. The retail investor today is receiving a better price in a reduced spread market, as evidenced by SEC Rule 605 best execution statistics. All good things.
The percentage of orders price improved has increased by 74 percent and the actual number of shares price improved has quadrupled. Comparatively, retail investors received nearly $238 million of price improvement in 2010 vs. $126 million in 2006. These are real dollars that go directly into the pockets of retail investors. Additionally, fewer orders are being filled outside the quote, meaning that more retail investors are “getting what they see on the screen” when their orders are executed by market makers.
And they’re getting all this much cheaper than they would if they were forced to execute all on exchanges and pay ever-increasing exchange and regulatory fees. A trade-at rule would eliminate this almost entirely as it would require by regulation a minimum price improvement that would, practically speaking, not make any economic sense to those market makers who currently provide such improvement.
When expanded to include institutional execution data, research further reveals a greater percentage of price improvement occurring in trades that are reported to the ADF (which includes virtually any liquidity that cannot be ascribed to an exchange, such as dark pool liquidity). Across large, mid- and small capitalization stocks, roughly 37 percent of the executions reported to the ADF receive some degree of price improvement (45 percent for small cap names). In the case of mid- and small cap stocks, the percentage of ADF-reported executions occurring at the midpoint of the NBBO exceeds that of all other exchanges and ECNs.
Some have argued that internalization and price improvement practices should be done away with altogether because, despite the impressive aggregate amount of improvement, the per-share amount is not large enough. Isn’t that like biting off one’s nose to spite one’s face?
Would retail investors rather have some price/size improvement or none (which is exactly what a trade-at rule would do in practice)? Why would we encourage a market structure that sacrifices the good for the perfect, recognizing that a “perfect” structure is unobtainable if we force market makers out of the game?
Adopting a trade-at rule would be step one toward imposing a central limit order book (or CLOB), where all orders would have to be publicly displayed and none could be price improved. Any investor of any substantial size will tell you that publicly displaying all of its trading interests would be like jumping into a tank of piranhas (only here they’re called high frequency traders). A trade-at rule would cause too much information leakage into the market and could actually have a deleterious, not additive, effect on liquidity.
A CLOB would be like replacing Restaurant Row with a government breadline. It’s just plain un-American. Innovation and investor choice would disappear. A trade-at rule would set us down that path.
Equity market structure is a mosaic, with different participants each providing valuable contributions to the overall flow. Trade-at seeks to replace this with a “one size fits all” approach, replacing a rainbow with a single swath of color. The only beneficiaries would be the for-profit exchanges and high frequency traders…and it would come at the detriment, once again, of the retail investor.
Comments | Post a Comment
10 Comments to "Trade-at Rule: A Breadline on Restaurant Row?":
crammond1964
13 April 2011
since 2006 markets = exchanges have allowed delayed publications quote stuffing and block trades ; which has not increased liquidity or volume . HFT are not the only guilty party but the internal crossing which has destroyed the real liquidity and volume . The exchanges have to allow ALL orders to enter through the same gate whether large or small or we do not progress ; in fact since 2006 we have gone backwards . Perhaps if the exchanges were regulated properly and not just for profit ; as they were pre 2003 we might see some light !
Comments (252)
Anonymous
13 April 2011
if it strikes a death knell to internalization then it could only be a good thing. to paraphrase the author, one need look no further than its detractors: the investment banks. the banks are a bastion of anti-competitive behavior. internalization is one example of that and so is their suite of too big too fail subsidies. it is about time that regulatory agencies remove conflict of interest in the marketplace so that a firm may act as agent or act as principal.
Anonymous
13 April 2011
Although well written, Mr. Masone's editorial seems to miss the forest for the trees. Non-exchange proponents of trade-at understand that a de-centralized market such as what Citi is advocating is ultimately bad for investors. If all orders were forced to a central location spreads would undoubtedly have decreased far more than that experienced in the past.
The 605 stats on so-called 'price improvement' that Mr. Masone quotes indicate that there is a lot more latent liquidity in the market that is being held off-exchange by internalizers and dark pool operators like Citi. Think about it--whose order is improving the price? If those orders were residing at a public exchange there would be tighter spreads and less need for price improvement.
To say that smart retail investors understand this is just wrong. How many Citi clients understand when their order is internalized and how many have the ability to determine what market conditions were in effect at the time? And how many can tell whether they got "best execution?" The answer of course, is none.
Finally, I wonder if this editorial is a policy statement of Citi or just of their Equities Division? I think if it gains wider distribution it could prove embarassing for the company.
Anonymous
13 April 2011
1) Please explain how price improvement of $0.0001 benefits the retail investor or anyone other than automated market makers and HFT. 2) Show the statistics that quantify how may times an a broker dealer or exchange via HFT/dark pool doesn't execute a limit order (retail or institutional) against a market order, but instead improves the market order by 1/100 of a penny. 3) Show me how many times a broker dealer or an exchange first routes a customer mkt order to a dark pool soaked with HFT with fill rates in line with the % of water you'll find in the Sahara desert alert the same HFT that are posting on the other side of the spread who then move or cancel their quote before an exchange or broker dealer can access the original displayed bid/offer. 4) etc. etc. etc. Isn't it all about the incentives that exist due to the make or take model which arguably allow for information leakage to move the market before a customers order flow gets executed? Isn't the first look at liquidity worth paying a rebate for if your only paying when you chose to trade (IOIs and reverse IOIs, - whatever you all decide to call them)? Is this what we want to protect and the behavior we want to provide incentives for? In theory doesn't the Trade-At Rule should eliminate the very routing practices that exist due to the make or take model which cause routing to fragmentation and hurt real investors (retail or institutional)? If not please explain. Considering the exemptions for trading blocks, (the majority of dark pools don't trade blocks) Isn't the only real problem with the Trade @ Rule the fact that it can't be addresses independently of the make or take model considering the access fees for various displayed venues differ and therefore the net pricing in the displayed markets at the the same displayed price points? What am I missing? Please do tell.
sarnuk
14 April 2011
Anonymous comments #3 and #4. Outstanding. Where is the Like Button here?
Comments (160)
kkhetan
15 April 2011
Let's put aside the rhetoric and work through an example (I'm speaking as a private investor): The market for XYZ is 35.10 to 35.20. The internalizer (and there are 200 of them now per Commissioner Mary Schapiro) gets a market order to simultaneously buy and sell XYZ. The mom and pop investor get filled at 35.1001 and 35.1999. So 99.9% of the spread is captured by the internalizer. The real damage is the reluctance for people (including mom and pop) to leave limit orders because they never get filled unless the market goes against them! I sense the hollowing out of order flow by internalizers, leading to enervating the liquidity providers, had a big role in the May 6th crash. There is room for all to survive. The SEC is clearly overworked. But I suspect they will act soon. Maybe decide that an improvement of .0001 is not material and, for argument sake say 25% of the spread must be improved (so in the example mom and pop get filled at 35.125 and 35.175) leaves a decent improvement for mom and pop, AND money, but not as much, for the internalizer. This is an interesting space and I'm sure will see resolution sooner than later.
Comments (12)
prgtrdr
15 April 2011
@kkhetan:
Your example isn't quite right. It is highly unlikely that two opposing market orders are entered at exactly the same time that would satisfy each other's interest. This is especially true for retail firms that tend to do a lot of internalization like UBS, Morgan Stanley, etc., where a lot of their order flow is research-driven. When a Morgan Stanley analyst says "I think Pfizer is overpriced" their retail clients will generate mostly one-way order flow to sell. You don't want to be standing on the tracks in front of that freight train.
The reality is that the internalizer (or market maker or exchange specialist) acts as a liquidity provider for the each market order until the contra interest arrives. During that period he is at risk of the market moving, thus he is entitled to earn the spread (minus slippage).
The point I'm making here is not that specialists aren't entitled to make money--they are. The issue is where in the food chain they reside and whether that process impedes price discovery.
Comments (7)
kkhetan
15 April 2011
@prgtrdr: i suspect the real debate is what price should the internalizer have to pay to be able to jump the queue and get in front of the NBBO. .0001 seems too low. I suggested splitting the difference between the mid and the bid (or offer as the case maybe). earning half the spread for providing 'liqudiity' seemed a reasonable opening gambit to me. Is there a better suggestion out there?
Comments (12)
Anonymous
16 April 2011
....... im afraid jumping the queue will be illegal under the new regulations ! The market has become so use to abusing the system . dark pools and internalisation and delayed publications have to be withdrawn if we are to have an honest market ; we are all here to make money but please within the rules .
overbet
18 April 2011
The internalizers entire argument for internalizing and sub pennying is a fallacy. They proudly boast that spreads for most stocks have narrowed and they give their customers better service, price improvement and all the other we know whats better for you than you do mumbo jumbo garbage. The fact is these spreads, meaning the bid and the ask may in fact have narrowed when you just look at them on the surface. Broker dealers are happy to present you with this "evidence" but the truth is the execution cost has increased. If I need to execute for a 100 shares or less great liquidity is there if I pay the spread. If I need to execute more than 100 shares I am going to move the price on anything over the first 100 on most of the stocks they brag about spreads narrowing on. So in the past if I sold 1000 shares of stock and the spread was $30 bid by $30.10 offer I could have sold most of it at $30 because that was a solid bid not a fluffy bid supported by air liquidity. So I get out of everything at $30. Now we have the same trade but the spread is $30 bid by $30.05 offer wow that spread is much tighter shouts the broker dealers. But I try to sell my 1000 shares at $30 and I get filled on 100 shares and now I am the offer of 900 at $30 and the bid drops out the algos freak and start sub penny selling in front of me. now I hit the new current bid of $29.95 and I get another 100 shares and this cycle continues until I have exited my position with an average exit price of $29.80. BUT THE SPREADS ARE NARROW!!!! great thanks for the favor mr broker dealer with friends like you......................
Comments (6)