Daily Speculations
 |
James Sogi
Philosopher, Juris Doctor, surfer, trader, investor, musician, black belt, sailor, semi-centenarian. He lives on the mountain in Kona, Hawaii, with his family. |
8/10/2005
Gaps, by Jim Sogi:
S&P mini morning gaps this week:
8/ 4 -4.75 8/ 5 -3.00 8/ 8 +2.25 8/ 9 +3.00 8/10 +3.75
The discontinuity is one of the weak points of many mathematical market models assuming a continuous process. The gaps are a good percentage of the open to close move. Definitely worth study. Given the worldwide markets, the electronic markets, what is the purpose of market breaks?
Markets by their very definition are completely discontinuous. The basic auction process of bid and ask requires a jump from one side to the other to make a deal. There is no split the difference, meet in the middle. It's either bid or ask. There are no real economic reasons for the size of the spread, other than artificial internal market rules for stoppages that mandate spread sizes. The purpose of the exchange is to narrow spreads to attract customers, but not so narrow as to squash profits. The market makers need to gap price to maintain liquidity and not run out of either cash or inventory of stock to maintain a bid an offer at reasonable spread and maintain volume. Gaps are their way to let off steam and adjust the market price for maximum action and profit. This is another example of how price is not necessarily a metric of value, but rather 'a metric of liquidity or risk adjustment and allocation.
Hong Kong, Tokyo and Osaka markets close for lunch break. What a joke. Even more opportunity for mischief, with no real purpose, but to shake more shekels loose from the customers.
Vinh Tu responds:
In Trading and Exchanges, Larry Harris says the safe thing for market makers to do is raise prices when they need inventory, and lower prices when they need cash. Otherwise they risk being gamed by better informed traders. This is the opposite of the model I had just proposed. Both models therefore need be tested. Which brings us to an ecological hypothesis: I'll posit three market-making strategies, which I'll dub "rational", "perverse" and "counter-perverse".
- ("rational" model) lf the market makers are trading against a perfectly rational population, of the sort described in textbooks, then the market makers gap only to the extent needed to balance supply and demand, which in turn take into account all the available information. The market makers may set the price slightly higher if they need to replenish inventory, or slightly lower if they need to unload.
- ("perverse" model) If the market makers are trading against a naive population that trades at the price given them (lambs being believers in efficient markets), prone to using stops, and being spooked, etc., then the market makers can gap in the direction of their desired price. In other words, move price up when you want to sell, and move price down when you want to buy, in contradiction to the normal laws of supply and demand. In this situation, the market makers fleece the lambs... until some wolves eventually appear and start fading the gaps.
- (the "counter-perverse" model) If the market makers are trading against a large population that deliberately fades the gaps, the market makers can occasionally reverse the strategy in 2. By moving prices in the same direction as in 1, but perhaps gapping further than necessary, to draw out the wolves, and then skin them.
I.e., once again we find the phenomenon of ever-changing cycles. Being a wolf all the time you get skinned. But if you wait till the farmer is busy shearing his sheep, then you might get a meal. Unless the other wolves get there first. Then you had better run because the farmer's out to get you. And so forth.
Unfortunately, the more complex the model, the harder it is to concoct an adequate test.
Thus, we have two explanatory theories for why gaps happen. If we take Mr. Sogi's model (which, if I understand correctly, is that market makers cause a down gap if they run out of inventory and need to buy low or, conversely, gap up if they need cash and want to sell high), then we would expect gaps to close more often than by randomness. Hence the popular technician's counsel to fade the gaps. On the other hand, Mr. Kuik points out an alternative explanation, i.e. that gaps are due to information being generated exogenously while the market is closed. I'll call the theories the microstructure explanation versus the exogenous explanation.
For a given gap, if the exogenous explanation predominates, one would expect that correlated price movements would have occurred overnight in other time zones. Conversely, a market-maker induced gap would be uncorrelated to overnight moves. And so,
Hypothesis: gaps preceded by big moves in similar overseas instruments will close less often than gaps not preceded by such overseas instruments.
Test #1: From Nasdaq gaps following similar move in Japanese techs, count probability of gap closing. From Nasdaq gaps following an opposite move in Japanese techs, count probability of gap closing. Is the latter statistic significantly higher than the former, as predicted by the above line of reasoning?
Test #2: Do same for a market with known egregious market makers. Is the apparent opportunity, if any, larger where an oppressive vig prevents one seizing it?
P.S. Speaking of Lobagola, I was looking for a picture of him on the web, and could not find any. So if anyone has an image or illustration of Lobagola, I would be curious to see it. In my mind's eye, I imagine him astride a stampeding elephant, surrounded by chattering monkeys. Bulls and bears scatter before the herd's approach and lick their wounds as the dust settles. I am willing to be disillusioned by a much more prosaic picture.
Phillip J. McDonnell adds:
Much of M.F.M. Osborne's book, A Physicist looks at the Stock Market, is about market making, especially from the NYSE specialist's stand point. The key point that Osborne makes is that market makers do not care where the market goes. What market makers really seek to do is to remain flat.
If a market is trending up then the market maker is presumably going short at increasing prices as he is "forced" to sell at his ask price. If the market subsequently does a Lobagola the market maker will be forced to buy at ever lower prices as he covers his shorts on the bid. Assuming an equal number of buys on the way up and sells on the way down the specialist will automatically profit by the amount of his spread times the number of round trip orders.
For a market maker his ability to profit is largely determined by his capital and the spread. His capital translates into how much inventory he can carry before he gets in trouble. So his only real goal is to get his inventory flat at the end of the day. Profit will take care of itself. He just needs to keep from going broke.
When a new day starts there is rarely an equal balance between buy and sell orders. So the market maker is required by exchange rules to make up the difference. As long as he can justify a gap based on the imbalance to an exchange official he is relatively free to pick his own price - the price at which he is willing to equalize the imbalance. He is also constrained by any limit orders on his book. If he gaps the price too far he may possibly bring in more competing limit orders.
Quite often the gaps go too far simply because the market maker has extended his position too far and wants to help insure a reversal movement. Learning to read what gaps mean in terms of whether the market maker is long or short can be a very profitable exercise. Remember that the 1962 SEC study of NYSE specialists showed that they made over 100% per year on their capital. So the question becomes do you want to trade with the market maker or with the excitable public who caused the order imbalance?
c
Jim Sogi, May 2005
(address is not clickable)