Payment For Order Flow Negatively Affects Retail #Stock #Traders

Item #1: Payment for order flow is bad for stocks with large spreads for, instead of encouraging competition and tighter spreads, it allows the execution firm to get the juicy flow and provides no incentive for other firms to narrow the spread. Also, there is often hidden liquidity inside the spread and internalized orders never benefit from it.

Item #2:

This is Google’s cache of

Who Executes Retail Trades? A Look at Market Share and Payment for Order Flow
1 Reply
The trading strategies posted on this blog have so far focused on identifying resting limit orders that are likely to have come from a human. There are all kinds of human traders, but probably the most coveted counterparties are retail traders. So coveted, in fact, that wholesalers will pay to interact with them and exchanges will pay to get them into their order books.

Public data that shows where brokers choose to direct customer orders is scarce. The most readily available source is probably SEC 606 reports. 606 reports are generally in a non-standardized format, lack important details, and appear even at a glance to have significant data quality issues. Nonetheless we will try to glean from them what we can. I’ve undertaken the tedious task of converting the unformatted reports, by hand, into something machine readable. I hope the results make the effort worthwhile.

SEC 606 reports generally disclose the following:

The proportional share of destination Broker-Dealers (i.e. wholesalers, ELPs, dark pools, exchanges) for the entirety of a broker’s orders.
The proportional share of a broker’s orders that are market, limit, or “other” orders.
The proportions of orders in each of these 3 categories that were directed to each destination Broker-Dealer.
The average payment per share received by the broker from each destination Broker-Dealer.
Unfortunately, I am not inspired with confidence in the reports’ accuracy or consistency in how they define the various categories. There are data fields that are often missing outright. And there could be additional errors since some of the data entry and cleaning was done manually. Given these limitations, I try to manipulate the data as little as possible to avoid propagating distortions.

First, we need a list of the largest online retail brokerages. I figured that most large brokers would have the money, brand, or shared business interests to be advertised on Nasdaq’s website. I picked all of the “Online Broker Partners,” except LOYAL3 because I couldn’t immediately find a 606 report from them. The aggregated 606 reports from Q4 2014 are here, with sourced notes about volume, AUM, regulatory net capital (‘Net Cap’), revenue per trade, and whatever else struck me from public filings. [1]

The attached R script uses the more regularized data here. See [2] for a rough overview of some of its assumptions.

Market Share and Payments
After some formatting, the script generates the following market share breakdown for destinations receiving a high share of marketable orders:

Order Destination Market Share (%) Avg Payment ($/shr) Marketability Index (Between 0 and 1, lower numbers more marketable)
UBS 29.7 0.0015 0.28
Citadel 23.2 0.0009 0.35
KCG 18.7 0.0011 0.38
G1 Execution Services 14.8 0.0015 0.24
IB ATS 4.1 NA 0.12
Two Sigma 3.8 0.0018 0.22
Goldman Sachs 3.6 0.0010 0.36
BatsY 1.1 NA 0.14
National Financial Services 0.7 NA 0.26
NITE 0.2 NA 0.34
AUTO 0.1 NA 0.27
AMEX 0.0 NA 0.00

And for destinations receiving orders that are less marketable:

Order Destination Market Share (%) Avg Payment ($/shr) Marketability Index (Between 0 and 1, lower numbers more marketable)
Nyse 28.2 NA 0.75
DirectEdge 27.7 0.0032 0.89
Nasdaq 20.6 0.0029 0.71
Arca 13.6 NA 0.77
Lava 8.3 0.0031 0.95
Nyse Mkt 1.2 NA 0.88
Bats 0.3 0.0020 1.00
EDGA 0.1 NA 0.74
See [3] for an overall table.

Non-Marketable Orders
One of the very first things we see is the higher payments brokers tend to receive for non-marketable orders. This shouldn’t be too surprising, since a broker has an incentive to post non-marketable limit orders on exchanges with high liquidity-provider rebates, which they receive upon execution. Could brokers get their clients a higher fill rate by posting these orders on exchanges with no rebate? Maybe. It’s well known that orders executed on inverted exchanges tend to perform better after trading, by roughly the amount of the rebate difference (we saw this in an earlier post). Retail orders may be a bit different though; we shouldn’t ignore the possibility that resting retail orders are almost destined to lose a penny upon trading. Should a client mind if their broker managed to capture a third of that through a rebate instead of feeding all of it to whoever else is trading? Maybe clients should get a commission discount for sending a resting order instead of a marketable one? I can see why brokers might shy away from that kind of pricing transparency. Some brokers already offer commission-free trading; going any further down this road would mean clients actually getting paid to trade. At a certain point, brokers might invite uncomfortable comparisons.

I don’t really know how best to make payments for order flow transparent. Maybe in the distant future there will be some sort of auction process where even non-marketable retail orders can receive price improvement. In the meantime we have to ask whether a broker seeking to pocket a large rebate might sacrifice their customer’s desire to get filled. I find it a bit unlikely since the broker won’t get paid at all if the customer isn’t executed. Ultimately, even with the seemingly perverse system of payment for order flow, brokers and clients probably have similar interests when it comes to resting orders.

We can see that resting orders are generally directed to major exchanges. Nyse, EdgeX, Nasdaq, and Arca make up about 90% of their volume. These exchanges are highly active and I’d expect that they get customers filled without increasing adverse selection by more than their roughly 0.30 cent rebates. I was a bit surprised to see so much volume directed to Nyse, because some exchanges, particularly EdgeX, offer much higher rebates for retail orders (see “Retail Order Tier“). Interactive Brokers was the only broker directing signficant volume to Nyse and their fee schedule indicates that at least some of their clients receive part of the liquidity-provider rebate. Maybe because they don’t receive some of it, Interactive Brokers cares less about squeezing the maximum rebate out of client orders (that, or their revenues from rebates are more complex due to the partial pass-through and presence of exchange tiers).

I was more surprised to see Lava getting about 8% market share. I don’t have much experience with Lava, and it would depend a lot on the specific stocks in question, but Lava’s volume during this time would have been an order of magnitude lower than these other exchanges. That could mean orders resting on it did receive noticeably lower fill rates than expected. Hopefully, brokers directed orders to that exchange only when it showed significant activity on the relevant stock. In any case, if there was an issue with Lava, there isn’t one any longer; Lava has been shut down.

It’s also interesting that almost all of Schwab’s orders are directed to UBS. For the other brokers, UBS appears to be mainly a destination for marketable orders. But Schwab had a contract with UBS during this period and appears to have sent UBS passive orders as well. Could some of these orders have found their way to Nasdaq, marked with UBS’s MPID? It seems like a definite possibility, given that we saw that orders marked UBSS performed as expected for retail orders (poorly), were larger than most non-retail orders, and numbered about 10,000 executions per day (Schwab as a whole executes hundreds of thousands of trades per day). I wonder if UBS might have kept the rebates from these orders, which were probably worth about 10-15k/day.

Marketable Orders and Conflict Of Interest
This brings us to the question of whether Schwab has extracted maximal value out of their customers’ orders. They were only getting about 0.10 cents per share from UBS during this period. Some brokers managed to get more than that from marketable orders alone – for instance, TD Ameritrade, which apparently received 0.18-0.19 cents/share from each of Citadel, G1, Two Sigma, and UBS itself. Adding some non-marketable orders into the mix like Schwab presumably has should only make that rebate higher, suggesting that maybe Schwab did not negotiate as hard as Ameritrade with UBS. Or, perhaps, that Schwab demanded better execution prices and faster fills than Ameritrade. If indeed Ameritrade’s higher received payments came at the expense of customers’ executions, it raises an interesting question about whether that expense can be argued as being baked in to their commissions, or if disgruntled customers could claim that Ameritrade is violating the spirit of best execution requirements. Although it sounds similar to a broker pocketing the rebate on non-marketable orders, this would not be identical to what we discussed above; the executed prices of retail marketable orders are to some extent chosen by their counterparty. Executing clients at a worse price somehow feels different to me than getting them a slightly lower fill rate, but still at their chosen price. It seems to me that payment for marketable order flow may entail greater legal risk for brokers when they knowingly accept higher payments than their competitors (but I’m far from an expert on such things).

Brokers’ businesses can be complex though, as seen in the (above-linked) article by John McCrank. Evidently, Fidelity’s retail orders may first interact with institutional flow (possibly Fidelity clients as well), reducing the economic value of any orders left over. I don’t know if and how any orders internalized by Fidelity itself would appear on their report. One potential sign in Fidelity’s 606 report is that a bit more than 1% of their flow is executed at a sister company, National Financial Services. It’s conceivable Schwab is doing something similar and the real reason they receive smaller payments is that they send UBS less attractive order flow. Or it’s conceivable that Ameritrade charges more for order flow and gets worse prices for their customers.

Remember, in contrast to resting orders, brokers are almost automatically guaranteed a commission for every marketable order sent to them. That means their incentives are not completely aligned with the interests of their clients.

[1] One thing that struck me in particular was the very large loss that Interactive Brokers experienced in their market making division:

During the quarter ended September 30, 2014, income before income taxes in our market making segment decreased to a loss of $111.8 million from income of $87.5 million in the year-ago quarter. This reflects a $198.9 million decrease in trading gains from the year-ago quarter. Removing the effects of currency translation, the market making segment produced $21.2 million pretax income in this quarter, compared to $41.3 million for the same period last year. Currency translation losses were $133.0 million this quarter, compared to a $46.2 million gain in the year-ago quarter. The decrease in market making profits, excluding translation effects, can be attributed to the ongoing competitive environment, lower volatility as measured by the CBOE Volatility Index, or VIX® and a trading error that resulted in a loss of approximately $16 million.

You might expect that kind of loss would prompt a review of risk management. If there was a review, I’m not sure how effective it was; just a few months later Interactive Brokers possibly lost $120M when the Swiss Franc peg was broken. I guess retail market makers can lose money just as easily as HFT market makers can.

If you’re interested, there was slightly more detail given in their earnings call:

Rich Repetto – Sandler O’Neill
I guess the first question, Thomas, could you just give us a little bit more details on the $16 million trading loss, because that certainly weighed on the results for this quarter, the Market Making results.

Thomas Peterffy – Chairman and CEO
Well, it was in an area of our software that is used from time-to-time and it’s kind of technical to really go into it in much detail. At this time the situation that it was applied to had a strange twist in it and the software misunderstood the situation and did their own thing.

Rich Repetto – Sandler O’Neill
So, could something like this — is this like – it could ever happen again or do you think it’s prevented now?

Thomas Peterffy – Chairman and CEO
Well, this same thing will not happen again. But software, from time-to-time, errors can always happen in a software, as you know. Apple comes out with a — or anybody comes out with a new chip, sooner or later you find the bug in the chip.

And in Q4:

[W]e took several steps to make our currency strategy more transparent which we believe will bring more clarity to understanding the results of our operating businesses. First, we transferred nearly all of the currency spot positions held as part of our global composition from the market making unit to the parent holding company IBG LLC.

[2] 606 reports are split by ticker tape. To aggregate the tapes I weighted the results by the Q4 2014 overall market volume statistics, which have about 53% of volume on Tape A, 19% on Tape B, and 27% on Tape C. Retail participants may trade a different mix than the global market average (e.g. they may trade more on the ETF-heavy Tape B), but I doubt the difference would affect the overall results much.

Exchange rebates and payments for order flow are not always disclosed. And sometimes when they are, there is an accompanying note that the disclosed payments are an upper bound. I assumed that the upper bounds were representative of the actual payments.

To evaluate the marketability of the orders a broker sends to a given destination, I used a very simple ratio: MarketabilityIndex = PercentOfLimitOrdersSentToDestination / ( PercentOfLimitOrdersSentToDestination + PercentOfMarketOrdersSentToDestination ). There are other possible measures, but I thought this was best given the data quality (sometimes proportions don’t even sum to near 100%). A destination was marked as accepting mainly marketable orders if this ratio, after weighting, was less than 0.4. Destinations were marked as mainly non-marketable if this ratio was greater than 0.7. The only destinations that didn’t fall into these two categories had less than 1% market share combined; these were: FBCO, Citi (which may be a blend of destinations, including Lava), and “Other.”

For statistics on overall market share, I approximated each broker’s overall market share as being proportional to their net capital, which is the only number that was available and reliable for every broker. Net capital may somewhat measure the risk of the trading that goes through a broker, which may be representative of the overall level of activity for their customers. TD Ameritrade and Fidelity have similar levels of both net capital and daily volume, even though net capital is usually posted well in excess of the regulatory minimum. Interactive Brokers has less AUM but a greater level of net capital than these two, which makes sense given its reputation for highly active clients.


Broker Order Destination Share of Broker (%) Marketability Index (Between 0 and 1, lower numbers more marketable) Avg Payment ($/shr) Broker Portion of Net Cap (%)
Etrade G1 Execution Services 50.3 0.30 0.0010 8.9
Etrade DirectEdge 20.8 1.00 0.0031 8.9
Etrade Citadel 9.2 0.16 0.0010 8.9
Etrade Lava 8.4 1.00 0.0027 8.9
Etrade KCG 5.5 0.14 0.0010 8.9
Etrade Nasdaq 3.2 1.00 0.0028 8.9
Fidelity KCG 33.1 0.38 NA 22.8
Fidelity Citadel 28.8 0.34 0.0002 22.8
Fidelity DirectEdge 16.9 1.00 0.0031 22.8
Fidelity Goldman Sachs 7.4 0.30 NA 22.8
Fidelity UBS 3.5 0.38 NA 22.8
Fidelity National Financial Services 1.4 0.26 NA 22.8
Fidelity Bats 0.5 1.00 0.0020 22.8
InteractiveBrokers Nyse 35.5 0.75 NA 35.5
InteractiveBrokers Nasdaq 23.8 0.61 NA 35.5
InteractiveBrokers Arca 17.2 0.77 NA 35.5
InteractiveBrokers IB ATS 5.5 0.12 NA 35.5
InteractiveBrokers Lava 4.4 0.92 NA 35.5
InteractiveBrokers Nyse Mkt 1.5 0.88 NA 35.5
InteractiveBrokers BatsY 1.2 0.14 NA 35.5
InteractiveBrokers DirectEdge 1.0 0.73 NA 35.5
InteractiveBrokers UBS 0.9 0.16 NA 35.5
OMSL KCG 42.7 0.40 NA 0.0
OMSL Lava 32.8 1.00 NA 0.0
OMSL Citadel 21.4 0.34 NA 0.0
Schwab UBS 94.2 0.49 0.0010 12.8
Schwab Citadel 2.9 0.68 0.0009 12.8
Schwab KCG 1.3 0.58 0.0012 12.8
Schwab Citi 1.1 0.58 0.0010 12.8
Schwab Goldman Sachs 0.4 0.46 0.0010 12.8
Scottrade KCG 26.3 0.18 0.0010 3.1
Scottrade Citadel 20.7 0.16 0.0010 3.1
Scottrade Nasdaq 14.9 1.00 0.0031 3.1
Scottrade DirectEdge 14.2 1.00 0.0029 3.1
Scottrade Citi 10.0 0.15 0.0010 3.1
Scottrade G1 Execution Services 5.3 0.16 0.0010 3.1
Scottrade Two Sigma 3.3 0.15 0.0010 3.1
TDAmeritrade DirectEdge 34.9 1.00 0.0034 16.4
TDAmeritrade Citadel 17.1 0.24 0.0018 16.4
TDAmeritrade G1 Execution Services 15.5 0.22 0.0018 16.4
TDAmeritrade Two Sigma 10.5 0.23 0.0019 16.4
TDAmeritrade Lava 8.5 1.00 0.0033 16.4
TDAmeritrade UBS 7.5 0.21 0.0018 16.4
TradeStation DirectEdge 26.8 1.00 NA 0.5
TradeStation NITE 23.0 0.34 NA 0.5
TradeStation BatsY 19.4 0.17 NA 0.5
TradeStation FBCO 14.5 0.50 NA 0.5
TradeStation EDGA 9.7 0.74 NA 0.5
TradeStation AUTO 6.2 0.27 NA 0.5
TradeStation Citadel 2.6 0.43 NA 0.5
TradeStation AMEX 2.3 0.00 NA 0.5
TradeStation Other 2.2 0.49 NA 0.5
TradeStation KCG 0.7 0.63 NA 0.5
Share this:

This entry was posted in Uncategorized on March 10, 2015.
Post navigation← Should IEX Become a Full Exchange? Incentives for Exchanges to Compete by Increasing DelaysManipulation Using Compliance Tools: Equities Analysis Inspired By Allston Case →
One thought on “Who Executes Retail Trades? A Look at Market Share and Payment for Order Flow”

March 18, 2015 at 8:02 pm
This is pretty horrible — the big retail brokers are routing mostly to dark pools and only select exchange/exchanges presumably because they can get paid more to route there. This can’t be right according to “best execution requirements”.

Related Party Transactions In ETH?
If ETH Isn’t a Security Then Nothing Is
Some Questions for Crypto-Exchanges
Do Bitcoin Markets have a Conversion Tort Problem?
A Curious Feature of IEX Auctions
Liquidity Disincentives and IEX
The Shape of Supply and Demand Curves in Rapidly Clearing Markets
CHX and Four Types of Speed Bump
High-speed Trading Networks and Societal Value
Price Impact in Efficient Markets
Pershing Square and Information Leakage on IEX
Possible Compromises for IEX
Can We Tell Who Trades on Which Dark Pools?
Could HFTs Benefit from a Cancellation Tax?
IEX Peg Orders: Last Look for Equity Markets?
A Close Look at the Treasury Flash Rally Report
IEX, Ideology, and the Role of an Exchange
Plain, Old Fraud in the Twitter-Hack Flash-Crash?
Are Data Centers that Host Exchanges Utilities?
Market Impact, Informational Efficiency, and the Value of Liquidity
Blind Analysis, Inefficient Markets, and UK Polling Accuracy
The Flash Crash, Systems Failures, and Mission-Critical Engineering
Market Data Patterns, Order Anticipation, and An Example Trading Strategy
Sarao Spoofing Allegations and Queue Position
Spoofing Corrupts Markets: A Reply to John Arnold
A Few Notes about HFT as a Business and the Virtu IPO
Manipulation Using Compliance Tools: Equities Analysis Inspired By Allston Case
Who Executes Retail Trades? A Look at Market Share and Payment for Order Flow
Should IEX Become a Full Exchange? Incentives for Exchanges to Compete by Increasing Delays
Ending Market Abuse: Limits on Anonymous Trading
Protecting Client Interests: Anonymity in US Equities
Order Size in the HFT Era: Identifying Trader Type Pt. 3
Creating an HFT Strategy: Identifying Trader Type Pt. 2
Finding Low-Alpha Orders: Identifying Trader Type Pt. 1

Item #3: wocram on Apr 1, 2017 [-] Literally all retail brokers are getting paid by citadel for their orders.

OTCBB: VEND real-time intraday trading flow – 11/26/2013

VEND is one of the more newer stock promotions on the block. If you’re new to this game and need to find out which brokers you should get to participate in the action, see my blog post about brokers.

Today I decided to take some raw footage of the intraday trading since we in the TimAlerts subscription chatroom service were expecting somewhat of a sell off today. VEND ended the day down 13%, not bad. We expect more downside to come yet. Some are even thinking that VEND might get halted like LIFS just did. See Michael Goodes blog post about LIFS:


If you are interested in joining this live trading service now is a perfect time to get in!
Timothy Sykes if offering 50%+ off annual plans: CLICK HERE.  (This is a limited time offer)

$NAMG & stock brokers required to bet against fraudulent pump N’ dumps

I was browsing the NAMG iHub message board, rather surprised there are no posts today after seeing the stock breakout 13% in a single day. I was long NAMG from Tuesday of this week at a price of $1.01. I sold half of my position yesterday at $1.08 and the last half today at the open, also $1.08. Why did I sell right at the open today? Because I could not watch my level 2 and charts today, and therefore instead of exposing myself to blind risk, I decided to play it safe and just close the position for a small profit. Was I disappointed that I didn’t hold all the way to $1.21 (today’s closing price); yeah sure I am. But it’s no big deal; there’s always another trade to be had.




Here is my post on the iHub message board the day that NAMG started to test the $1.05 level. (Thursday, 10/24/13 03:00:36 PM)


Boy did I call this one right! I’ve been telling several people in the TIMAlerts chat room that it was time to start covering and/or boxing their short positions on NAMG.



After seeing no real activity on the iHub NAMG message board today, I checked the Yahoo Finance NAMG message board. There is some activity here. I see one guy who asks this question which I will answer for him here.

I believe this is a pump & dump stock promotion, as many here do. Does anyone know a broker that will allow shorting (short sale) of this stock and what would be the minimum investment and margin amount required? Potential returns are massive.


Question #1 Does anyone know a broker that will allow shorting (short sale) of this stock?

Answer #1 Interactive Brokers and SureTrader will allow one to sell short NAMG and plenty other PINK/OTCBB stocks.

– Interactive Brokers requires a minimum amount of $10,000 USD to open a margin account.

– SureTrader only requires $500 USD to open an account!

(If you want referrals to either broker, send me an email request: dazetrader at gmail dot com)


Question #2 What would be the minimum investment required?

Answer #2 The minimum investment would be 1 share of stock, which is literally peanuts, you can invest as little or as much as you want!


Question #3 What would be the margin amount required?

Answer #3 Both of the above mentioned brokers adhere to the $2.50 rule. Which means if a stock trades below $2.50 per share, you are required to put up $250 margin for every 100 shares that you want to sell short. For example, say you have a XYZ stock trading at $1.00 and you want to sell short 1000 shares. The amount of cash to short the stock would be $1.00 x 1000 = $1,000, but the margin requirement is actually $2.50 x 1000 = $2,500. So to hold this XYZ stock short, you will be required to maintain at least $2,500 in your trading account.

I have been subscribing to TIMAlerts subscription service for just over 1 year now. The best part about this service is the live chat. You can be sure to hear about the very latest and up & coming pump n’ dumps that are coming to the table and are currently in-play. You will also greatly reduce your learning curve in the OTCBB/PINK market place.

If you are interested in subscribing, click on this link! Or send me an email and I can get you connected as well.  dazetrader at gmail dot com

%d bloggers like this:

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.