Apologies if this is a very simple question, but how exactly do firms make money from "flow trading"? Is it simply profiting from "selling" direct market access?
$\begingroup$
$\endgroup$
6
-
2$\begingroup$ Wall Street is like Beverly Hills: there is a lot of gossip from people who claim to be in the know. In Beverly Hills gossip it is about Hollywood actresses and movies, in Wall Street it is about what big investors are buying or selling. By being in the middle of the flow of trades some firms have short lived but valuable information that they can profit from themselves or pass on to clients or exchange with others for other pieces of information. $\endgroup$– Alex CCommented Oct 29, 2017 at 23:28
-
$\begingroup$ Of course they can't tell you "George Soros just bought MSFT" or they'll lose the account but they can say "we are seeing a lot of buying interest in MSFT this morning from major hedge funds". And some people will be interested in having a relationship with a firm that can tell them that. $\endgroup$– Alex CCommented Oct 29, 2017 at 23:28
-
$\begingroup$ @AlexC isn't this front-running? $\endgroup$– user997112Commented Oct 30, 2017 at 0:36
-
1$\begingroup$ It is front running only if you do your trade before doing the trade on behalf of George Soros. $\endgroup$– Alex CCommented Oct 30, 2017 at 1:03
-
2$\begingroup$ Alex C, you should put your comment as an answer so I can downvote it... ;) Seriously madilyn's answer is matter of fact and IMO accurate representation on how flow trading is profitable by taking advantage of the bid-ask spread. Your vision is more like a Hollywood depiction of trading in the 60's $\endgroup$– NegativeJoCommented Oct 30, 2017 at 20:34
|
Show 1 more comment
2 Answers
$\begingroup$
$\endgroup$
5
Flow trading is in spirit very similar to market making - such firms make a profit by earning a spread.
There are 3 common ways this is done. Suppose a client wants to buy 100k shares of XYZ, which is publicly quoted at [email protected] bid, [email protected] ask. For sake of simplification, assume sub-penny pricing is not accepted in the jurisdiction where XYZ is listed.
Match against inventory: The firm already has a position of 100k shares of XYZ at an average fill price of 10.015. In theory, the firm has an unrealized loss at liquidation price on the public market. Now, the firm can sell all 100k shares to the client at 10.02, which is an improvement over the public best offer of 10.03 for the client, while the firm also collects a spread of 0.005x10^5=$500 - a win-win situation.
Instantaneous: The firm does not have a position in XYZ. Nevertheless, the firm can sell all 100k shares to the client at 10.03 immediately. Since this is the best offer price on the public market, the firm benefits because it has likely skipped the queue at 10.03 to get a fill. Now the firm can work the offsetting order, wait for a client that wants to sell XYZ, hedge its position, or improve the public best bid by posting buy [email protected] where it would earn a spread of 1 cent.
Combination of the above: The firm already has an inventory of 80k shares at an average fill price of 10.01. It can sell 100k to the client at 10.02 and simultaneously buy 20k from the market at 10.03, thus realizing an average entry price of 10.014 and average exit price of 10.02, collecting a spread of 0.006x10^6=$600.
(1), (2) and (3) suggest that all previous replies to your answer are misinformed, it is entirely possible to do this in a way that neither frontruns the client nor disadvantages the client relative to NBBO. On the contrary, the client may be a noise trader and frontrunning its order may actually work against you.
(1) suggests that the firm needs to already be in the business of holding on to a large portfolio which it wants to be at net zero exposure, this is why many flow trading firms are also large broker-dealers.
(2) suggests that the flow trader collects profits at the expense of the otherwise market makers who were quoting 10.01-10.03 and would've loved to compete for the business of filling those 100k shares. The moral issue arises here as one may argue that the flow trader is collecting its profit at the expense of its client: perhaps a high-frequency market maker on the exchange would've been willing to cross the spread and sell at 10.00 if it saw a [email protected] bid post because the public market maker has prediction alpha. By internalizing the flow, we're removing competitive price discovery from the public market. This lends credence to why most high-frequency trading firms prefer lit and transparent markets, which are more competitive for the utility traders, investors and end users of these markets.
-
2$\begingroup$ Im sorry but how is this any different to market making? $\endgroup$– TrajanCommented Feb 2, 2018 at 21:25
-
$\begingroup$ Also could you explain where client flow trading fits between market making and prop trading please as this is regular question on forums $\endgroup$– TrajanCommented Jan 9, 2019 at 21:15
-
$\begingroup$ @Permian You should ask that as a separate question. $\endgroup$– madilynCommented Feb 2, 2019 at 23:52
-
$\begingroup$ @user1050421 Depends on what you're trying to do. I might know others who can help you as well. What's your email? $\endgroup$– madilynCommented Feb 3, 2019 at 0:14
-
$\begingroup$ @user1050421 Sent you an email. -AM $\endgroup$– madilynCommented Feb 3, 2019 at 1:23
$\begingroup$
$\endgroup$
2
As to the value that customer order flow provides, there is specific published research on order flow in the FX markets by Sarno et al. (Why) Does order flow forecast exchange rates?
-
6$\begingroup$ Could you give a summary of the paper concerning the question. Just providing links to papers is normally not considered a good answer. $\endgroup$– vonjdCommented Oct 30, 2017 at 13:23
-
1$\begingroup$ A one sentence summary: banks active in FX observe the order flow from their customers; this order flow can be shown to have predictive value for FX rates, so the bank could profit if it uses this proprietary information to speculate in FX for its own account. QED. $\endgroup$– Alex CCommented Oct 27, 2018 at 22:05