Case study Broker trade executions

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
triceratop
Posts: 5838
Joined: Tue Aug 04, 2015 8:20 pm
Location: la la land

Re: Case study Broker trade executions

Post by triceratop »

A data point that I'm not sure has all that much value:

I put a limit buy order of $63.84 for 5 VGIT through VBS this morning. After I placed my order, the online real time quote system showed the Size as being 1x5 with a Bid/Ask of $63.83/$63.88. Multiple trades executed at $63.83 (I do not know how many shares were transacted but the shares were not thinly traded this morning). After half an hour I moved my limit order up to $63.86 and the bid/ask spread almost immediately changed to $63.85/$63.89. I obtained an execution 5 minutes later, during which time shares were exchanging at $63.85 according to VBS' real time quotes. My order was executed at $63.86.

Might not mean much, but it seemed quite abnormal compared to my (all odd-lot) trades with VBR, VSS, VWO, VTI, and VCIT.
"To play the stock market is to play musical chairs under the chord progression of a bid-ask spread."
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

^ I am pretty sure that small odd lot limit orders get worse execution at many brokers. They just don't go into the pipeline. I would submit market orders for those kinds of orders myself. Indeed, I bought a few shares of AGG this morning with a market order and received a trade price of half-way between the 1 cent bid & ask prices.

This is why I will not submit limit orders for 105 or 224 or other lots. For those kinds of things, I will submit multiple orders: Perhaps a limit order for 100 or 200 or x hundred shares, then a remainder as a market order. In particular, I have been pleasantly surprised by small market orders for the least liquid of the ETFs I own: VSS and DGS. Even when the spread has been in the 10 cent range, I have received executions better than halfway towards the side that is beneficial to me (high when selling, low when buying).
Wiki This signature message sponsored by sscritic: Learn to fish.
pshonore
Posts: 8212
Joined: Sun Jun 28, 2009 2:21 pm

Re: Case study Broker trade executions

Post by pshonore »

I've put in orders to sell say 225.66 shares of XYZ. If the spread is 1 cent, I may put in a market order or maybe a few pennies over.

Its not unusual to see an execution for 100 shares, another for 80 and one for 45. The fraction of course gets liquidated pro-rata.

Same on the buy side. An order to buy 100 shares may have 1 execution or 3.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

I doubt anybody else is really looking at the intraday trades of bond ETFs today. The late December distribution was payable today, so if owners and brokers are reinvesting dividends, there could be some upward pressure on the prices. But looked what has happened in the last hour for AGG (and similar for BND):

Image
Wiki This signature message sponsored by sscritic: Learn to fish.
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

Passive oddlot orders are not protected quotes. A broker doing a poor job will trade through them.

An aggressive oddlot order (crossing the spread) will generally receive favorable executions from the broker taking the other side, since they are presumed to have little short-term predictive quality.
User avatar
whodidntante
Posts: 13114
Joined: Thu Jan 21, 2016 10:11 pm
Location: outside the echo chamber

Re: Case study Broker trade executions

Post by whodidntante »

Interesting. I have recently been using market orders for lots < 100 and I'm monitoring the executions with some interest. So far, so good.

I made the assumption that my broker would break an order > 100 if it benefited me. So I used a limit order if my goal is to buy 101 shares. I'll do it livesoft's way next time, assuming commissions are not a factor.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

@ftobin, thanks for chiming in. Would you mind giving some specific examples of what you just stated when you have time please? That might help clarify terms ilke "passive", "aggressive", "protected", and "trade through" which in this context are unfamiliar to me. Thanks!
Wiki This signature message sponsored by sscritic: Learn to fish.
*3!4!/5!
Posts: 1256
Joined: Fri Sep 02, 2016 1:47 pm

Re: Case study Broker trade executions

Post by *3!4!/5! »

livesoft wrote:@ftobin, thanks for chiming in. Would you mind giving some specific examples of what you just stated when you have time please? That might help clarify terms ilke "passive", "aggressive", "protected", and "trade through" which in this context are unfamiliar to me. Thanks!
It means I'm sticking with mutual funds. If I'm looking around the poker table and I can't see the sucker ...
matto
Posts: 163
Joined: Fri Jan 03, 2014 2:51 am

Re: Case study Broker trade executions

Post by matto »

livesoft wrote:@ftobin, thanks for chiming in. Would you mind giving some specific examples of what you just stated when you have time please? That might help clarify terms ilke "passive", "aggressive", "protected", and "trade through" which in this context are unfamiliar to me. Thanks!
Passive = limit order on your side of the book (so someone has to hit *you* for the order to fill)

Aggressive = market order or a limit order into the other side of the book (so you hit someone else's passive order to get a fill)

Because of Reg NMS, quotes on exchanges are protected. What this means is that triceratop's situation is supposed to be prohibited. That is, the current best bid/offer (NBBO) on any exchange are protected, all other market participants are prohibited from making a trade at a worse price than the NBBO. The order is called 'protected' since it must be filled before worse prices can be filled.

So specifically, if the b/a is 63.83/63.88, and triceratop posts a bid at 63.84, once that bid is 'processed', no buy orders at 63.83 can be hit unless the 63.84 bid is hit first.

But because the rules (Reg NMS) are terrible, there are a million loopholes. What ftobin is stating is that odd lots (not multiples of 100 shares) are not protected. So this is why triceratop *could* have seen orders in the market seemingly go through his/her order. This is what ftobin means by 'odd lots are not protected'.

For what it's worth, I work in finance and even I didn't know this loophole. There are so many it is hard to keep track.

The other thing ftobin states is that brokers should want to fill odd lots since they are mostly noise. This is because any *real* trader will want to execute in round lots, so this implies that people who are willing to buy 5 shares aren't informed and therefore their orders are mostly noise. Market makers love noise since it tends to be bidirectional and annihilate. What they are afraid of is that the order for 100 shares is part of a mutual fund selling 100,000,000 shares of stock. This is what ftobin means by a good broker will want to fill odd lots rather than let them trade through.

Going one step further, every order you place at TD Ameritrade, VBS, Fidelity, etc doesn't ever get routed to an exchange. It gets routed to Citadel/KCG/Virtu/Two Sigma who then decide if they want to internalize that order. Retail order flow is very bidirectional so this is free money. This is why in the above example, a 'good' broker would often fill you internally at 63.84 once the broker sees a trade in the market at 63.83. They will fill you because 1) you are retail, so noise. 2) odd lot, so doubly noisy. 3) you are willing to trade at a price 1 cent higher than the market (almost like Citadel buys it from the market at 63.83 and then sells it to you at 63.84. If I recall from my Series exam, brokers are allowed to trade in front of you as long as they fill you within 100ms or something afterwards.).

If this gives you a headache, you are not alone. The US exchange system is needlessly complex. Would be better IMO to turn it into a single regulated exchange almost like a public utility. Exchanges like IEX just add to the complexity.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

Now that's the kind of post that really belongs in this thread. Thanks!
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
triceratop
Posts: 5838
Joined: Tue Aug 04, 2015 8:20 pm
Location: la la land

Re: Case study Broker trade executions

Post by triceratop »

^ Thank you very much! I inferred much of the language from context but that explanation is confirming and wonderfully done.
"To play the stock market is to play musical chairs under the chord progression of a bid-ask spread."
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

matto wrote:Going one step further, every order you place at TD Ameritrade, VBS, Fidelity, etc doesn't ever get routed to an exchange. It gets routed to Citadel/KCG/Virtu/Two Sigma who then decide if they want to internalize that order.
It is true that every order does not get to an exchange, but that does not mean that some orders do not get to an exchange. I often see my orders in the "book" shown by real-time level II quotes. I understand many of the reasons when orders are not shown and I understand what it takes to have an order shown.
Wiki This signature message sponsored by sscritic: Learn to fish.
matto
Posts: 163
Joined: Fri Jan 03, 2014 2:51 am

Re: Case study Broker trade executions

Post by matto »

livesoft wrote:
matto wrote:Going one step further, every order you place at TD Ameritrade, VBS, Fidelity, etc doesn't ever get routed to an exchange. It gets routed to Citadel/KCG/Virtu/Two Sigma who then decide if they want to internalize that order.
It is true that every order does not get to an exchange, but that does not mean that some orders do not get to an exchange. I often see my orders in the "book" shown by real-time level II quotes. I understand many of the reasons when orders are not shown and I understand what it takes to have an order shown.
I should have been more clear. No order is ever *initially* routed to an exchange. They go to the HFT firms who decide if they want to internalize it, and if not, they can forward the order to the exchange. You are correct that some orders make it to the exchange.

I will further add that some exchanges have larger rebates than others. Because the market is pretty efficient, you do more or less the same no matter where you route your order. That is, routing to the low rebate exchanges will more likely get a fill, and routing to the high rebate exchanges will less likely get a fill, but the extra rebate makes up for this.

However, since the HFT firms keep the rebate, they'll always route you to the exchange that's least likely to get a fill.

Interactive Brokers on the other hand will pass through the rebate to you. It is the only broker that does this, to the best of my knowledge.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

A lot of what you wrote is also described in "Flash Boys" by Michael Lewis and the earlier book by Scott Patteron "Dark Pools" which went into more depth.

Any estimate of how much Schwab makes from all this routing first to places like Citadel?
Wiki This signature message sponsored by sscritic: Learn to fish.
matto
Posts: 163
Joined: Fri Jan 03, 2014 2:51 am

Re: Case study Broker trade executions

Post by matto »

livesoft wrote:A lot of what you wrote is also described in "Flash Boys" by Michael Lewis and the earlier book by Scott Patteron "Dark Pools" which went into more depth.

Any estimate of how much Schwab makes from all this routing first to places like Citadel?
Just an FYI, Flash Boys is an entertaining read, but it's almost criminally embellished. It raises few of the substantive problems and raises a lot of trivial ones.

People like Katsuyama (he worked on an equities desk) were extracting far more money than modern HFT firms.

I would estimate the total order flow payments across all exchanges are < $1B. So maybe $100-200M per brokerage.
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

[update: matto posted while I was writing, and his post looks better than mine.]
livesoft wrote:@ftobin, thanks for chiming in. Would you mind giving some specific examples of what you just stated when you have time please? That might help clarify terms ilke "passive", "aggressive", "protected", and "trade through" which in this context are unfamiliar to me. Thanks!
Sure -- my original text, with translations/examples interspersed:
Passive oddlot orders are not protected quotes. A broker doing a poor job will trade through them.
First, a definition: NBBO: national best bid and offer, which indicates the highest-priced buy non-oddlot order (>=100 shares) on an exchange, as well as the lowest-priced sell non-oddlot order. Let's say the MSFT NBBO is 10.00x10.05. The NBB (national best bid) is 10.00, NBO (national best offer) is 10.05.

Now that that's out of the way, I'll presume you are buying in all my examples, to simplify the language. I'll also assume just one exchange at the start.

Generally speaking, a passive order is one that, when placed, is not expected to be immediately crossed on the exchange's book. So placing a buy in MSFT up to and including 10.04 would generally be considered passive. An aggressive order would be one that is priced to immediately execute at least partially (>= 10.05).

If you place buy priced at 10.02 for >= 100 shares, it will likely become the new NBB, so the spread becomes 10.02x10.05. However, if you place an oddlot at the same price, this is considered *not part of the NBBO*, and the NBBO remains 10.00x10.05.

Now, RegNMS rules state that nobody can trade outside the NBBO; that means, in our example, no trades can execute outside of [10.00, 10.05].

If I have access to an exchange's direct feeds, I can see the oddlot order resting at 10.02, so I know I could sell to you at 10.02. If I don't, I only know 10.00x10.05, and if I sent an order to another exchange, I could "trade through" your order and get executed at 10.00. In contrast, if I send a sell order to the first exchange priced at 10.00, I'd cross your order at 10.02 first, then execute the rest at 10.00. (I think first exchange *could* technically execute me at 10.00 as well but in practice they won't).

A good broker has direct feeds to all exchanges and looks for oddlots that are "inside" the NBBO before routing dumbly to any ole' place.
An aggressive oddlot order (crossing the spread) will generally receive favorable executions from the broker taking the other side, since they are presumed to have little short-term predictive quality.
We covered what an aggressive order is above. "Smart" traders (looking to get in/out on small movements of stock) generally aren't trading <100 shares at a clip. It's expensive (fees & commissions), and it gives away information. So the inverse states that if an order is <100 shares, it's generally from someone who isn't trying to time the market in a short-term manner. A market maker (the person you send your order to) will often sell from their inventory to people buying oddlots, since it'd be both profitable and helping their quality metrics (a whole other topic).
matto
Posts: 163
Joined: Fri Jan 03, 2014 2:51 am

Re: Case study Broker trade executions

Post by matto »

To add: I met one of the founders of Nasdaq and he said he tried contacting Lewis in order to help proof read or otherwise advise Flash Boys, but never got a response. He was pretty upset with the final product, and I tend to agree.

Just so nobody accuses me of picking on equities, the following are much worse:

1. Municipal bonds
2. Forex
3. Corporate bonds

Forex in particular is hilarious. If your counterparty is important (like a major bank) and doesn't like a trade, they can request to roll it back. If you refuse, they can get you kicked off the exchange.

All 3 are improving, and I hope HFT continues to enter the bond market and brings down spreads. Forex spreads are already quite low.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

@ftobin, thanks. So can I then say that the idea to break up an oddlot >100 shares order (103, 246, 318, e.g.) into at least two orders of a round-lot and a separate sub-100 share oddlot makes some kind of sense?
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

matto wrote:To add: I met one of the founders of Nasdaq and he said he tried contacting Lewis in order to help proof read or otherwise advise Flash Boys, but never got a response. He was pretty upset with the final product, and I tend to agree.
I'm not surprised, but any comments about "Dark Pools" which has some history of Nasdaq so your conversation was probably with someone mentioned front and forward in Patterson's book?
Wiki This signature message sponsored by sscritic: Learn to fish.
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

matto wrote:But because the rules (Reg NMS) are terrible, there are a million loopholes.
I agree with almost everything you wrote, but I'd disagree here. The rules aren't perfect, but I worked pre- and post-RegNMS, and I think it added a lot of sanity to the market. Personally, I think not allowing people to trade outside the market is a good thing. People who believe otherwise are free to send not-held orders.
matto
Posts: 163
Joined: Fri Jan 03, 2014 2:51 am

Re: Case study Broker trade executions

Post by matto »

livesoft wrote:@ftobin, thanks. So can I then say that the idea to break up an oddlot >100 shares order (103, 246, 318, e.g.) into at least two orders of a round-lot and a separate sub-100 share oddlot makes some kind of sense?
I believe mixed lots (> 100 shares) are broken down into round and odd lots. But there's probably another loophole somewhere.
livesoft wrote:
matto wrote:To add: I met one of the founders of Nasdaq and he said he tried contacting Lewis in order to help proof read or otherwise advise Flash Boys, but never got a response. He was pretty upset with the final product, and I tend to agree.
I'm not surprised, but any comments about "Dark Pools" which has some history of Nasdaq so your conversation was probably with someone mentioned front and forward in Patterson's book?
Sorry I don't remember hearing anything about that book in particular. I've heard "Flash Boys Not So Fast" (a rebuttal) is pretty good, but I believe it is a bit too friendly to the internalizing firms. Like I mentioned earlier, retail flow is bidirectional so Citadel can fill you at *better* than the market because you are more likely to annihilate with someone else. That's the argument for why selling of order flow benefits you. However, I tend to believe there are currently too many loopholes and complex regulations. A single lit exchange with sub penny pricing would be the best of all worlds IMO.
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

matto wrote:Just so nobody accuses me of picking on equities, the following are much worse:

1. Municipal bonds
2. Forex
3. Corporate bonds
Very, very much true. Equities in the US, in general, have a very efficient market compared to other asset classes.
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

matto wrote:
livesoft wrote:@ftobin, thanks. So can I then say that the idea to break up an oddlot >100 shares order (103, 246, 318, e.g.) into at least two orders of a round-lot and a separate sub-100 share oddlot makes some kind of sense?
I believe mixed lots (> 100 shares) are broken down into round and odd lots. But there's probably another loophole somewhere.
No loophole for mixed lots -- anything >= 100 shares is treated equally.

livesoft: Well, if you break up your >100-share order, you leak information. If it's just two orders, it probably doesn't matter. So if you have 150 shares to trade, maybe two 75s. But I wouldn't worry about it, it's not worth your time. Even if you lost two pennies trading 100 shares of a $100 stock ($10,000 of stock), that's only $2 lost. That's like a one-time expense ratio hit of .02. I used to care about getting the perfect execution, but it's not worth my time for such small trades. The market rules and quality requirements are working on my behalf.
A single lit exchange with sub penny pricing would be the best of all worlds IMO.
Whoa, whoa, whoa. QQQ used to trade out to four decimals for a while about 10 years ago. It was a mess, with so much feedback looping from people subpenny-ing each other. Absolute mess. Manning especially.
matto
Posts: 163
Joined: Fri Jan 03, 2014 2:51 am

Re: Case study Broker trade executions

Post by matto »

ftobin wrote:
matto wrote:But because the rules (Reg NMS) are terrible, there are a million loopholes.
I agree with almost everything you wrote, but I'd disagree here. The rules aren't perfect, but I worked pre- and post-RegNMS, and I think it added a lot of sanity to the market. Personally, I think not allowing people to trade outside the market is a good thing. People who believe otherwise are free to send not-held orders.
I don't mean Reg NMS is so terrible that the absence of Reg NMS would be better.

Like I said, it certainly is better than other markets (bonds/currencies), but seems worse than futures.

Again, all I know about equities is from studying for my Series exam. I kept on thinking that the current regulations were needlessly complex and it would just be simpler to have a single exchange, which would eliminate the need for a lot of regulations. That and complete separation of prop trading and client trading.

There are obviously much worse problems in finance, which is why I think HFT equities gets unfairly singled out. Fidelity was quoting me a 0.4% spread on treasuries! It's shameless.

I am also a huge fan of ETFs because leaving trading up to the market seems much more fair than offloading the costs to the other shareholders.

The market is also more efficient than the mutual fund manager when it comes to pricing of foreign securities, for example. (I've heard there is a drag on international mutual funds because the fund manager must price at close of US hours but holds securities that do not trade during US hours).
matto
Posts: 163
Joined: Fri Jan 03, 2014 2:51 am

Re: Case study Broker trade executions

Post by matto »

ftobin wrote:
A single lit exchange with sub penny pricing would be the best of all worlds IMO.
Whoa, whoa, whoa. QQQ used to trade out to four decimals for a while about 10 years ago. It was a mess, with so much feedback looping from people subpenny-ing each other. Absolute mess. Manning especially.
Okay maybe four decimals is excessive :)

And on thinking about it, I'll agree it isn't a good idea. After all, companies can control their splits to target whatever bid/ask spread they want at penny pricing.
Tanelorn
Posts: 2370
Joined: Thu May 01, 2014 9:35 pm

Re: Case study Broker trade executions

Post by Tanelorn »

The new nickel pilot is taking US equities back in the wrong direction, but I don't hear the internalizers complaining about that.
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

Tanelorn wrote:The new nickel pilot is taking US equities back in the wrong direction, but I don't hear the internalizers complaining about that.
There are 3 test groups and there is a lot more than just nickel increment quoting going on with the ticket size pilot -- I don't recommend painting with a broad brush. The side effects of some of the changes might take some time to play out. This is a fairly reasonable pilot, having control groups and data collection.

Oh, and by the way, I assure you at least one big internalizer was absolutely vociferous against some of the data collection going on.
User avatar
grabiner
Advisory Board
Posts: 35307
Joined: Tue Feb 20, 2007 10:58 pm
Location: Columbia, MD

Re: Case study Broker trade executions

Post by grabiner »

livesoft wrote:Just a note to myself:

Some lower volume small-cap emerging markets ETFs have only had one transaction this morning -- namely at the opening cross -- and no subsequent trades 9 minutes hence. I don't believe "price discover" has really happened.
Price discovery doesn't require trades, just offers. If someone is offering to buy at 49.95 and someone (possibly the same trader) is offering to sell at 50.05, you have a pretty good idea of the price even if nobody takes either order. If the bid is 49.50 and the ask is 50.50, or if the bid is just 100 shares and level 2 quotes show that nobody else is close to the bid, then price discovery may not have happened.

And I wouldn't recommend buying a low-volume ETF with a market order at the open; your order might move the opening cross against you, or there might not even be an opening cross unless your market order creates one. I don't hold these ETFs any more, but I think that there were days I bought EEMS in which there were no trades at the open and I had to wait quite a while to get a reasonable spread.
Wiki David Grabiner
Caduceus
Posts: 3527
Joined: Mon Sep 17, 2012 1:47 am

Re: Case study Broker trade executions

Post by Caduceus »

I was thinking that one form of comparison would be to look at dividend reinvestments for the same ETF. I've noticed that Vanguard consistently fills automatic dividend reinvestment of its own ETFs (i.e. Vanguard ETFs) at better prices than other brokers. It doesn't always reinvest at a better price, but it does so significantly more so for it to be attributable to random chance. When Vanguard did offer superior reinvestment, the range was usually in the 0.2% - 0.3% range. When it offered poorer reinvestment prices, the range was usually below 0.2%. 8 out of 10 times, Vanguard offered a better (lower) reinvestment price.

It doesn't seem likely to me that Vanguard has a superior trading algorithm than other brokers (or perhaps it does?) so I am wondering if this has something to do with the volume of Vanguard ETFs being traded at Vanguard in comparison to other brokers. Is it also possible Vanguard has greater insight into the day's order flow given that the bulk of Vanguard ETFs pass through Vanguard's order system?
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

Caduceus wrote:I was thinking that one form of comparison would be to look at dividend reinvestments for the same ETF. I've noticed that Vanguard consistently fills automatic dividend reinvestment of its own ETFs (i.e. Vanguard ETFs) at better prices than other brokers.
Vanguard themselves are not filling the orders -- they are passing it to a market maker with some instructions on how to fill it. I would hope in these situations they would be requesting either full-day VWAP or just sending it to the close, but I don't know. Other brokerages would be doing the same, sending it to different market makers with different instructions.
User avatar
alec
Posts: 3181
Joined: Fri Mar 02, 2007 1:15 pm

Re: Case study Broker trade executions

Post by alec »

livesoft wrote:A lot of what you wrote is also described in "Flash Boys" by Michael Lewis and the earlier book by Scott Patteron "Dark Pools" which went into more depth.

Any estimate of how much Schwab makes from all this routing first to places like Citadel?
It might be disclosed in Schwab's 606 report, or in Schwab's 10-Qs or 10-Ks.

Also, odd-lot limit orders are not required to be displayed by market makers in their quotes on exchanges - see SEC Rule 604. So, your odd-lot limit order (below 100 shares) might never be routed outside the market maker.
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

alec wrote:Also, odd-lot limit orders are not required to be displayed by market makers in their quotes on exchanges - see SEC Rule 604. So, your odd-lot limit order (below 100 shares) might never be routed outside the market maker.
In practicality, though, they're going to post an oddlot and try to get a rebate (passive limit orders are generally profitable with little risk). Also, if the oddlot is not exposed and the NBBO moves through it without an execution, the customer is going to call up a couple of weeks later and ask for it to be fixed (especially if it has now moved 5% in their favor). The guaranteed rebate is worth it to just post.
User avatar
alec
Posts: 3181
Joined: Fri Mar 02, 2007 1:15 pm

Re: Case study Broker trade executions

Post by alec »

ftobin wrote:
alec wrote:Also, odd-lot limit orders are not required to be displayed by market makers in their quotes on exchanges - see SEC Rule 604. So, your odd-lot limit order (below 100 shares) might never be routed outside the market maker.
In practicality, though, they're going to post an oddlot and try to get a rebate (passive limit orders are generally profitable with little risk). Also, if the oddlot is not exposed and the NBBO moves through it without an execution, the customer is going to call up a couple of weeks later and ask for it to be fixed (especially if it has now moved 5% in their favor). The guaranteed rebate is worth it to just post.
good point. thanks
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

ftobin wrote: Also, if the oddlot is not exposed and the NBBO moves through it without an execution, the customer is going to call up a couple of weeks later and ask for it to be fixed (especially if it has now moved 5% in their favor).
I am trying to wrap my head around "… the customer is going to call up …."

If a client submits a limit order for an odd lot that does not execute, how do they even know why it didn't execute? They would have to be watching trades or maybe go back and look at daily history or log of trades. What kind of client would even notice? And notice enough "to call up a couple weeks later"? For instance, is triceratop supposed to call up and ask for their 10 cents back?

Can you give any examples with details of things that customers actually called up and complained about? I'll give one complaint I reported: "Where are my dividends that were paid overnight? All my OTHER brokers show them? Why don't YOU show them? I need that money to buy shares today? Where is my money?" At least with my complaint it is easy to document a) dividends showed up at my other brokers and b) dividends didn't show up at the broker I am complaining to. :)
Wiki This signature message sponsored by sscritic: Learn to fish.
ftobin
Posts: 1071
Joined: Fri Mar 20, 2009 3:28 pm

Re: Case study Broker trade executions

Post by ftobin »

livesoft wrote:If a client submits a limit order for an odd lot that does not execute, how do they even know why it didn't execute?
Many clients have tools that help surveil this activity. It's not that difficult to do historically. If we had technical issues and had problems routing to the market, we would run this report on ourselves to fill the clients who would have otherwise gotten executions.
They would have to be watching trades or maybe go back and look at daily history or log of trades. What kind of client would even notice? And notice enough "to call up a couple weeks later"?
It not that uncommon to do error corrections that are worth thousands of dollars weeks after the fact.
For instance, is triceratop supposed to call up and ask for their 10 cents back?
triceratop is not the market maker's customer -- if triceratop is routing through Ameritrade, Ameritrade is the customer.
Can you give any examples with details of things that customers actually called up and complained about? I'll give one complaint I reported: "Where are my dividends that were paid overnight? All my OTHER brokers show them? Why don't YOU show them? I need that money to buy shares today? Where is my money?" At least with my complaint it is easy to document a) dividends showed up at my other brokers and b) dividends didn't show up at the broker I am complaining to. :)
All the complaints you mention are from a retail customer to their broker. A market maker does not deal with the end customer, but rather with brokers. They call up all the time asking for corrections or courtesy busts.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

Muchas gracias!
Wiki This signature message sponsored by sscritic: Learn to fish.
saver007
Posts: 236
Joined: Fri Nov 07, 2014 8:18 pm

Re: Case study Broker trade executions

Post by saver007 »

TD ameritrade's 10K provides pretty good insight into how much money brokers makes by routing orders to a few closed group of market makers (internalizers like citadel) as opposed to sending to exchanges or building their own order routing system with price improvement as goal.

TD Ameritrade's 2016 10K state it makes 299 million in "order routing revenue". Thats about 20% of their total transactional revenue (revenue from trades).

Their average commission per trade is $11.76 while average order routing revenue per trade $2.56. That means they collect $2.56 per trade from internalizers (citadel,etc) for executing trades for clients they already collected commission for.

When a broker is collecting 2.56 per trade from closed group of market makers, its incentive to build and invest in an order routing system focused on price improvement for client's order is limited, shall we say?

Most retail brokers follow TD Ameritrade's business model.

Interactive brokers is the rare exception. If you care about execution quality, better go with IB. They not only, route almost all orders to exchanges, they will pass on rebates provided by exchange to clients and have built an order routing designed to maximize price improvement. IB is also the only broker that i know of that publish monthly execution quality/expense report that calculate total execution expense for client orders.

https://finance.yahoo.com/news/interact ... 00372.html
https://www.interactivebrokers.com/en/index.php?f=1685
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

saver007 wrote:TD ameritrade's 10K provides pretty good insight ….

Their average commission per trade is $11.76 while average order routing revenue per trade $2.56. That means they collect $2.56 per trade from internalizers (citadel,etc) for executing trades for clients they already collected commission for.
Thanks for that. but the numbers don't smell right. I guess that a significant fraction of TDAmeritrade customers don't pay the published $9.99 per trade commission if the average commission is $1.77 more. Perhaps it is the mutual fund trade commissions and the broker-assisted trades that bump it up?

Also for a retail investor who uses no-commission ETFs with spreads of about 1 cent, it seems that unless 256 or more shares are traded in an order, then the $2.56 is not going to happen either.
Wiki This signature message sponsored by sscritic: Learn to fish.
saver007
Posts: 236
Joined: Fri Nov 07, 2014 8:18 pm

Re: Case study Broker trade executions

Post by saver007 »

livesoft wrote:
saver007 wrote:TD ameritrade's 10K provides pretty good insight ….

Their average commission per trade is $11.76 while average order routing revenue per trade $2.56. That means they collect $2.56 per trade from internalizers (citadel,etc) for executing trades for clients they already collected commission for.
Thanks for that. but the numbers don't smell right. I guess that a significant fraction of TDAmeritrade customers don't pay the published $9.99 per trade commission if the average commission is $1.77 more. Perhaps it is the mutual fund trade commissions and the broker-assisted trades that bump it up?

Also for a retail investor who uses no-commission ETFs with spreads of about 1 cent, it seems that unless 256 or more shares are traded in an order, then the $2.56 is not going to happen either.
Numbers are from their 2016 annual report directly. Page 37.
boglerocks
Posts: 675
Joined: Sun Apr 03, 2011 9:08 pm

Re: Case study Broker trade executions

Post by boglerocks »

FWIW I love Interactive Brokers. Just started rolling my IRAs over there too. First one went off without a hitch. Been with them about a year and execution quality still sometimes surprises me.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

Just a note about Vanguard Brokerage [dis]Services. It was painful trying to get some money invested this morning compared to other brokerages. I expect that if I have made a few transactions and want to see what's happened, that I can click on a "Transaction history" or Transactions" or "Activity" tab and see what's happened. Not so at VBdS:

Image

There were transactions yesterday and today, but neither one of them show up. VBS: This tab needs to be updated in REAL-TIME by your software and not in Vanguard time. Seeing this "… no transaction …" makes me think I either made a mistake, lost my money, or went out of my mind! That's not a good way to treat a customer. Thanks!

Furthermore, the new opening page has a "Recent activity" tab which is missing all these transactions as well.

Yes, there is another tab called "Pending activity" which does not have these transactions either.

Where did my money go?
Wiki This signature message sponsored by sscritic: Learn to fish.
*3!4!/5!
Posts: 1256
Joined: Fri Sep 02, 2016 1:47 pm

Re: Case study Broker trade executions

Post by *3!4!/5! »

livesoft wrote:Just a note about Vanguard Brokerage [dis]Services. It was painful trying to get some money invested this morning compared to other brokerages. I expect that if I have made a few transactions and want to see what's happened, that I can click on a "Transaction history" or Transactions" or "Activity" tab and see what's happened. Not so at VBdS:

Image

There were transactions yesterday and today, but neither one of them show up. VBS: This tab needs to be updated in REAL-TIME by your software and not in Vanguard time. Seeing this "… no transaction …" makes me think I either made a mistake, lost my money, or went out of my mind! That's not a good way to treat a customer. Thanks!

Furthermore, the new opening page has a "Recent activity" tab which is missing all these transactions as well.

Yes, there is another tab called "Pending activity" which does not have these transactions either.

Where did my money go?
The "Order status" tab.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

*3!4!/5! wrote:The "Order status" tab.
That is missing everything, but "orders", so it doesn't have all my transactions.
Fortunately, this account should have no more activity for 2017.
Wiki This signature message sponsored by sscritic: Learn to fish.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

I made a trade last week that I found remarkable, so I am going to remark on it.

I submitted a limit order to sell 50 shares of an etf with the limit about 4 cents higher than the current ask and about 12 cents higher than the current bid. Of course, this was not a marketable limit order. I was watching the real-time Level II quotes and trades (similar to previous pics posted in this thread) casually. Eventually my order was executed, but because it was a sub-100 share odd-lot it did not show up in the list of trades. What is remarkable is that the highest price transactions that showed up were 4 cents lower for a few minutes before my transaction until about 10 minutes after my transaction. That is, the posted trades were all at least 4 cents per share lower than the price I sold for. Furthermore, the high price of the day (up to that time) was still posted as 4 cents lower than my trade. So my trade executed at 4 cents higher than the high price of the day (until later in the day when trading was higher).

I am not saying that my order executed outside of a normal bid/ask spread. I cannot say that because I was doing other things on my computer and not watching the real-time Level II quotes closely. Here's how that might work:

order submitted: bid/ask 100.14/100.22; my limit 100.26
when order executed: bid/ask could have been 100.18/100.26, with transactions happening at 100.22 and below. My order executed at 100.26.
the bid/ask could even have gone momentarily to 100.20/100.28 or 100.26/100.30, but no executions except mine before falling back down.

Bottom line: Some trades are not reported to the general public.
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
alec
Posts: 3181
Joined: Fri Mar 02, 2007 1:15 pm

Re: Case study Broker trade executions

Post by alec »

As of 12/9/13, odd lot trades started to be reported through the consolidated tapes (i.e. SIPs), but the do not update high, low and last sale price.

http://www.finra.org/industry/trade-reporting-faq#310
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

Thanks alec, good to know.
Wiki This signature message sponsored by sscritic: Learn to fish.
User avatar
nisiprius
Advisory Board
Posts: 52215
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Case study Broker trade executions

Post by nisiprius »

livesoft wrote:...I am pretty sure that small odd lot limit orders get worse execution at many brokers. They just don't go into the pipeline...
Do you think that's the reason why I was never able to get a limit order executed at Fidelity? I tried each time. Each time I bid the exact posted "ask" price--the Fidelity website would not let me bid above whatever it was posting as the current ask. But I was unable to get an order executed. I tried this on two or three separate occasions, and each time I failed several times in a row trying to place a limit order at the asking price. One of them would have been for $10,000 worth of iShares TIP which was probably trading at a bit above $100 and thus probably was not quite a round lot.

In the good old mutual-fund-versus-ETF discussions, people who have less than $3,000 to invest and thus can't meet the minimum purchase requirement for Vanguard index funds are often counseled to consider ETFs, but are never told about problems with odd lots.

I didn't use limit orders to eke out any kind of clever advantage, I just wanted protection against the extremely remote possibility of another "flash crash," because everyone said it was trivially easy, "just use a limit order."

By the way, I've never been clear as to whether any real-world small retail investors were actually hurt in the flash crash. It's clear that the flash crash did affect some of the ETFs a Boglehead might well invest in, but I don't know if anyone really got hurt... considering that the exchange unwound at least some of the affected transactions.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

nisiprius wrote:In the good old mutual-fund-versus-ETF discussions, people who have less than $3,000 to invest and thus can't meet the minimum purchase requirement for Vanguard index funds are often counseled to consider ETFs, but are never told about problems with odd lots.
I just described a plus for odd lots: I sold at 4 cents better than others trading at the same time, but surely someone bought my shares but maybe that someone had a large volume order and the rest of their order was a lower price. It is true that 4 cents ($2 total for 50 sh) was a very small percent of transaction.

I have never placed an order at Fidelity, so I cannot comment directly on your situation. I do have a Fidelity brokerage account, but I transferred shares "in-kind" into it several years ago and have never made a trade in the account.

My personal experience is that all the mumbo-jumbo about ETF orders is for the most part bunk. I place market orders at my brokers and usually get a price near the mid-point between bid and ask (even if spread is 1 cent) or price improvement in the right direction from those prices. I do not place limit orders to sell at the bid (the low price); nor do I place limit orders to buy at the ask (the high price).

If I did place a limit order to buy at the bid (the low price), I know that there may be no sellers at that price which is why it is the bid price. I can see before I place the order if trades are taking place every few seconds or tens of seconds or few minutes. If shares are actively traded every few seconds or less, then a market order is fine.

If shares are traded every tens of seconds, then a limit order to buy at the bid will just get in the queue and will not execute. Clearly, any limit order to buy at a lower price will also not get executed until something else happens. That something else could be folks(robots) cancelling their existing limit orders, so that one's order becomes the new best bid.

If shares are traded every tens of seconds, then a limit order to sell at the ask will just get in the queue and will not execute. Clearly, any limit order to sell at a higher price will also not get executed until something else happens. That something else could be folks cancelling their existing limit orders, so that one's order becomes the new best ask.

So my strategy depends on what I want to accomplish. If I submit a limit order for fear of getting ripped off and nothing happens, then I might change to a market order. In every such case that I have done that, I have never gotten a worse price than my limit and most of the time I received a better price somewhere between the bid and ask prices. This is not what bogleheads.org teaches folks.

This is why I usually tell new ETF users to experiment and try out various trades and see what happens.

PS: I just re-read what I wrote. It will probably give anyone who does not place a trade at least once a month a headache. Sorry about that. :)
Wiki This signature message sponsored by sscritic: Learn to fish.
*3!4!/5!
Posts: 1256
Joined: Fri Sep 02, 2016 1:47 pm

Re: Case study Broker trade executions

Post by *3!4!/5! »

I thought the problem with a market order is that you could end up selling for $0.01 per share or buying for $1,000,000 per share.
Topic Author
livesoft
Posts: 86076
Joined: Thu Mar 01, 2007 7:00 pm

Re: Case study Broker trade executions

Post by livesoft »

*3!4!/5! wrote:I thought the problem with a market order is that you could end up selling for $0.01 per share or buying for $1,000,000 per share.
I think the exchange and/or broker would "bust" such trades, so it is not something to worry about. More likely is that one can buy for up to 2% to 5% higher and sell for 2% to 5% lower than a non-naive investor. And that's only for large orders of low-volume issues.
Wiki This signature message sponsored by sscritic: Learn to fish.
Post Reply