Options: Open Outcry vs. Electronic

[FDAXHunter]

Hehe... nicely messed with my Eurex representative today.



Basically, it was about the success of screen trading and I wanted to point out to her that in options, screen based trading actually is less efficient. I threw in a free suggestion on the proper way to set it up.



Let's review some facts in a market where we can make a comparison:



Treasury complex on the CBOT: essentially 97% of all options get traded in the pit, even though over 90% of all futures are traded electronically.



European Options scene: Most volume goes through brokers, rather than directly on the screen.



Now, here is my reasoning on why and how and all that:



- Front Month markets are easy to trade electronically. You put up a couple of prices and you're off to some active trading.

- Spread markets are more difficult to trade electronically, [i]unless[/i] the combinatorials of spreads are easy and everybody trades the same spreads. (Euribor Futures, Eurodollar Futures)

- If the spreads are all essentially non-standardized, you need an auction system.



The problem with option spreads versus futures spreads is the vast number of combinations that people want to trade: the 50-60-70 put fly versus the 55-60-65 iron fly, etc.

The exchanges (Eurex and LIFFE.Euronext) try to tackle this problem with running essentially seperate order books for each strategy. This is, quite frankly a ridiculous idea and always upsets me when I think about it.



Let's recognize an important element:

All options strategies are like Legos. You have a bunch of pieces lying around, then put them together. Everybody that knows Legos knows that. You don't "pre-build" several pieces, then stash them away in your big box for use later.



So what happens in the open outcry pit? Someone asks for a particular structure, then everybody goes to their lego box, builds the thing and presents one with a price what they think it's worth.



What happens on the screen? You call a broker, who in turn calls a bunch of market makers, who pull out their lego box and return their prices to the broker, who returns them to you. Or you call the market makers yourself.

So where is the efficiency in this process? First of all, calling a bunch of market makers takes a fucking long time, the competition is less intense, because they can't see the prices everyone else is putting up (nobody trusts a broker, as they are all scumbags), and then the broker might do something in there to switch you with his cousin or sumptin'.



All in all, a ridiculously inefficient scheme. Small wonder the pit still does 95% of the option volume.



So, how does Eurex & LIFFE.Euronext and all the rest fix that?



Here's my idea:



- Populate prices for the individual pieces (the Lego pieces). i.e. all strikes and all maturities (as far as commerically reasonable of course).



- Recognize the fact that a straddle is no different from a iron strangle swap and introduce and auction system in the Quote Request sense. Essentially, each underlying instrument has a single ongoing Quote Request and answering session where each strategy can be requested, no matter if straddle, butterfly or a 12-legged jelly roll strip. Each quote request would instantiate a new auction, where market makers would join if the want to. Then the auction would disappear after a time out or a quote request originator taking a price.



Problem solved.

[Man]

Good to see you FDAX!



"Each quote request would instantiate a new auction, where market makers would join if the want to. Then the auction would disappear after a time out or a quote request originator taking a price."



Question about the instantiation of new auctions, would there not be so many that it would be overwhelming to particpate in many auctions? Would each shop have a dedicated screen trader for a few markets of instantiated option auctions?  Is this an online pit?

[FDAXHunter]

There would be a new auction for each instance of a quote request, yes. All market makers respond to quote requests pretty much in an automated fashion, you cannot answer quote request manually.



You are right, that this is essentially an online pit where the exchange asks: What are your markets for the Jun 30-35-40 call fly. Market makers then autoquote this.



This is no different than what the brokers do, except that they split liquidity and drive up costs.



You have to realize that Europe is more advanced as far as screen trading goes. We only have one exchange, so we can concentrate on lots of bells and whistles. In the US, where you have 6 options exchanges, the major technological challenge is integrating them on the firm level, no time for trading strategies, quoting vols, etc.

[TonyC]

since i only trade NYMEX screens arent very important to me . . . . . . but why couldn't one build one's own "engine" to assemple the lego pieces oneself i.e. gather up the bids and asks on the strikes, and figure out where the xmas tree is trading for oneself. yes, the numbers you figured out would be crossing the bid/ask on a lot of strikes, and would thus be a greater price than the pit might quote you, but couldnt you just submit a bid on one leg, and if you git hit on your bid lift an offer on the others [all the while doing nanosecond electronic cancel/replace adjustments of your bids and offers as the underlying future and options markets moved]? or is there a technical glitch of which i am unaware [one that comes to mind is that maybe cancel/replace orders are not so easy]

[TonyC]

and oh yeah, why arent my carriage returns showing up as paragraphs

[FDAXHunter]

TonyC,



Welcome and stuff. Good to see you Smiley



Yes, you'd absolutely build your machine for the individual lego pieces to add them up. However, as you said, you'd need to cross the bid-ask on every leg to get filled.

The alternative is to call up a market maker and say "what's your market it the Jun 03 30-35-40 Put butterfly?" Now, he's going to have an idea what it's worth, by punching it into his own pricer, but is he going to be competitive? No, because he can't see other participants prices.

In the pit, it's an auction, and that's a very efficient price discovery process.

You can see where other people are bid and offered, so a market maker might say: "Hm, these guys are 14 bid on this fly, so maybe, I'll join them, or heck, I could probably beat them an make 15".



As far as legging the strategy goes, you are right that is possible, especially if you only have two legs. Totally doable and this is what people do. However, there are technical limitations to this. You can't submit quotes every microsecond, the exchanges would be overwhelmed, so you have to put them in every couple of seconds, which in certain markets means you have to include enough safety margins, which leads to a possibilty of not getting filled.

Also, you're not going to get the best price, because people



Also, for more than two legs, that's not feasible, because some of the options you got in there might not trade for a long time. So you might get filled on your 30 put, but there isn't a buyer for the 35 put and no seller for the 40 put all day (unless you offer enough incentive of course, and then you are going to give up alot of edge in each). All the while, any market maker would have made you a price in the 30-35-40 fly happily.



Apart from that, legging always suffers from the problem that when you definitely don't want it, you're going to get it and when you need it most, you won't.



Don't you think my proposal makes alot of sense (normally my ideas don't, but this one, I think has merit, no?)



PS: I think you carriage returns don't show as paragraphs because you use Netscape or something? You can HTML paragraphs as "
", for example. In IE, you actually get quite a powerful text editor, where you can insert pictures and links and tables, etc. quite comfortably. (That's probably not what you want to hear though, I gather)

[dgn2]

Definitely your proposal makes sense. I wish that were the way things were because it would make screen-based trading that much more appealing and open a lot more opportunities for program trading in options. Any idea what prevents on exchange like EUREX from doing exactly this? I can't think of anything off the top of my head.

[dgn2]

Another thought: wouldn't dealers tend to set up their automated response (at least initially) to return the spread price implied by all the individual option prices.

[FDAXHunter]

Thanks dgn2.



What's keeping Eurex from doing this? I don't know, Eurex isn't the place it used to be. When it first started as Deutsche Terminbörse (DTB), it was a lean player, couple of really specialized techies tha knew their stuff. Now, it seems, everyone you talk to is some MBA student who was taught the basics of derivatives trading in a weekend course. You can see that with retarded ideas poping up like two seperate order books for Bund futures (one for Eurex and one for Eurex US). Clearly they don't know what they are doing. (they abandonded that notion now, thank god).



As far as quoting goes... well, a market maker quoting a straddle as the price of the spread of the two legs is not going to attract alot of business Smiley

The guys who know how to do this stuff: Timber Hill, Archelon, CMT, Wolverine, Susquehanna etc. will be all over this, rest assured.

[dgn2]

Thanks FDAX.



Yeah, I agree that such prices would not be too appealing...are Timber Hill, Wolverine, Susquehanna, etc actually making better prices using an algorithmic approach (presumably they are)? The only data I have on spreading comes from GLOBEX and it only just began to appear to be worth looking at. I am interested in learning more about automated spreading in options, but I have not yet had time to really figure out if I can effectively use any of the co-movement discovery and prediction techniques I designed and implemented for backtesting of other simpler strategies.