|
|
发表于 2010-5-13 08:52 PM
|
显示全部楼层
Since I cannot 发新话题, I am pasting something I came across.
Program Trading Perils: Systemic Risk and Unjust Enrichment
Note: This is a long post, but hopefully worth the read.
When I analyze difficult problems, I often look through a lens of extremes. I’m going to compare the structure of the stock market in the United States in 1970 to its structure in 2010. I argue that a program trading oligopoly has changed the structure of the market over the last forty years and has introduced systemic risk across the market. I also argue that this program trading is unjustly enriching the firms (and thus the individuals comprising these firms) in this elite oligopoly.
1970: CENTRALIZED MARKETS CONTROLLED BY HUMANS.
In 1970, the New York Stock Exchange was the primary stock exchange in the United States. Computers were new meaning that if programmed trading existed it only played a very nominal role. Additionally, the New York Stock Exchange was a specialist system where humans handled the order flow.
We can conclude that in 1970 the stock market in the United States was concentrated (mainly just the NYSE) and that human beings handled the order flow.
2010: FRAGMENTED MARKETS CONTROLLED BY COMPUTERS.
Fast forward to May 6, 2010. The day of the “flash crash.” The NYSE is a hybrid system. We still have the humans handling some trades, but the NYSE also utilizes computers to handle order flow. Additionally, there are numerous electronic exchanges handling transactions outside the walls of the NYSE. Of course, we also have the Nasdaq which is an electronic stock exchange.
Further, because of the computerized exchanges, we also have a market dominated by programmed trading. A Wikipedia article on programmed trading notes:
According to the New York Stock Exchange, program trading accounts for about 30% and as high as 46.4% of the trading volume on that exchange every day.[3] These historical percentages show the dominance of Program Trading listed on the NYSE.
AND:
Program Trading is a strategy normally used by large institutional traders such as Goldman Sachs (the largest program trading firm), Credit Suisse First Boston, UBS Securities, Barclays Capital, SG America’s Securities, and Morgan Stanley. During the second quarter of 2009, Goldman Sachs recorded record trading profits, with much of those gains ascribed to program trading, according to heavy press coverage.
We can conclude that in 2010 the stock market is highly fragmented among numerous electronic exchanges and that computers handle most of the order flow.
THE FLASH CRASH OF 2010
There have been ridiculous claims that the flash crash was caused by a fat finger trader at Citigroup. I previously posted why this made no sense. The real explanation is that there was a liquidity crisis. The most sensible explanation for the liquidity crisis has to do with having a fragmented marketplace and a market dominated by program trading.
Having a fragmented market was a problem because the NYSE went into “slow mode” to calm the markets. But this arguably exacerbated the problem because the other electronic exchanges did not follow suit. There was no system of coordination to make all of the markets slow or stop trading. The NYSE is still the largest marketplace. Therefore, when the NYSE slowed down a ton of liquidity was removed from the market. The other exchanges continued trading with low levels of liquidity and as bids dried up prices crashed.
A market dominated by programmed trading also likely played a role in the crash. Proponents (generally a concentrated group of banks) have argued that program trading is a public good. Proponents argue that program trading increases market liquidity allowing tighter spreads for both retail and institutional investors. I think this argument has merit although the “public good” provided does not even come close to outweighing the harm caused by program trading.
Opponents of program trading have argued (in part) that program trading is “weak liquidity.” When the going is good, these providers hang around and let their computers trade the market. However, opponents argue that there is nothing to keep this weak liquidity in the market during times of crisis. Therefore, if the market starts to free fall, it is quite possible that the computers will stop trading. I would argue that Lloyd Blankfein has the power to crash the market. I think a call to his trading group instructing them to “turn off the computers” would cause an instant liquidity crisis.
It is quite plausible that the computers stopped trading last Thursday. Part of it may have been because the NYSE slow mode limited their trading. Part of it might have been because the banks controlling them shut them down (either by pressing a button or because they are programmed to do so). In a sense, it really doesn’t matter what happened last Thursday. The point is that the few banks controlling the computers could have shut them down.
Unfortunately, like a bad drug, non-computer based market participants have grown accustomed to having this liquidity. So when the weak liquidity suddenly leaves the market (why they leave really doesn’t really matter), the humans are left to fend for themselves. A huge void is left where the computers once traded. All those bids go away and we enter free fall mode. Especially when we are in a situation like last Thursday when momentum already strongly favors selling.
Moreover, other innovations developed since 1970 such as index futures likely exacerbate the problem. If you have a long portfolio and can’t sell stock because there are no bids what do you do? You goto the futures market and hedge your portfolio using index futures. This kind of cascading effect makes the crash even worse (short index futures is often cited as a reason for the 1987 crash).
We can conclude that the flash crash was caused by an instant loss of liquidity. It is quite likely that fragmented markets and program trading were causes of loss liquidity which in turn caused an instant drop in the market.
SOLVING THE FRAGMENTED MARKET PROBLEM
The fragmented market problem can be solved. Regulators are already working on a system to coordinate all of the exchanges. Because this fragmented market problem was so foreseeable, it is somewhat surprising that there was not already a system in place. But that is a another issue in itself. Regulation should allow coordination which would permit circuit breakers to allow a cool down period during a free fall.
PROGRAM TRADING IS A SYSTEMIC PROBLEM
Unfortunately, I’m not very confident that our regulators will solve the program trading problem. The stock market has a major structural problem that did not exist in 1970. A huge portion of the trading volume is concentrated among just a few firms.
We can conclude that if anything goes wrong with one of the players in the program trading oligopoly that we will have a liquidity crisis. If for any reason the program trading stops (either willfully or accidentally), we will have a liquidity crisis.
And a liquidity crisis in the stock market is much worse than a fundamental event causing a crash. If there is a fundamental event, humans can revalue price taking into account the fundamental event. But if there is a liquidity crisis, bids can go away instantly. Any stock or index can lose virtually all of its value since there are simply not enough bidders to fill the void left by program trading firms providing the “weak liqudiity.”
RES IPSA LOQUITUR & UNJUST ENRICHMENT
In law, part of the negligence doctrine includes a concept called “res ipsa loquitur.” In short, negligence can sometimes be inferred by the very nature of an accident. For example, assume a person lives in an apartment building and a person is injured (or dies) when the elevator free falls 5 stories. We can infer negligence because we know that an elevator that is maintained properly should never free fall.
Likewise, we can infer that, Goldman Sachs, the largest program trading firm, has an unfair advantage solely based on the fact that they made money every single day of the first quarter of 2010.
We can safely conclude it is an unfair advantage and that the firm is being unjustly enriched. Just like an elevator should not free fall 5 stories, no firm or individual can statistically make money trading every single day of an entire quarter. Moreover, this is not just one quarter. Goldman in past quarters only lost money on at most a few trading days. And this isn’t just Goldman. I’m just picking on them because they are #1 in program trading. All of the program trading firms need to be looked at.
I’m not concluding that Goldman is doing anything outside the law. I am concluding that they have an unfair advantage and are being unjustly enriched. If the current law already prohibits their activities, then the law needs to be enforced. If the current law does not prohibit their activities, it must be retooled in a way to remove this unfair advantage.
PRAGMATIC SOLUTIONS FOR PROGRAM TRADING
I believe I have identified two significant problems with program trading. First, we have the systemic problem with program trading. Because it is concentrated among only a few firms and those few firms control such an enormous amount of the total volume on the exchanges, there is a major systemic risk if any of the program trading firms simply stop their computers (it could be willful or accidental). Second, we have a problem whereby firms, most notably Goldman Sachs, have on its face an unfair advantage permitting them to be unjustly enriched. The firms that control the program trading essentially have a license to steal money out of the marketplace. It may be currently legal. But if it is, the practice must end. Not only is it unjust, but the activities of the program trading oligopoly is a huge systemic risk.
The regulators are in a really tough spot fixing it. It is not as easy putting in regulations to terminate the program trading oligopoly. The market is addicted to program trading like a drug. If the regulators suddenly cutoff this liquidity through regulation, the market will crash. Even doing something like eliminating just a certain type of program trading, such as the “flash trade”, could cause a crash.
After all, the program trading oligopoly only likes doing their high frequency trading because they know that they will win. If you make it so they don’t always win, they might just voluntarily turn their computers off.
SO WHAT DO I DO AS A RETAIL TRADER?
Even if the regulators have the desire to stop the unfair advantage (that is putting it nicely) that these program trading firms have, it is seems like they will have their hands full taking this drug away from the market without suffering some major withdrawal symptoms. So my guess is that they won’t do anything to unwind this risk and that the oligopoly will continue to provide lavish bonuses to their employees.
Honestly, if I was an investment banker at Goldman, I would have mixed feelings. An investment banker actually provides a useful service and makes an honest living. So when people naysay about Goldman, the guys and gals doing real work on the IB side get the same label as the prop traders. On the other hand, the IB guys and gals get larger bonuses thanks to the unjust profits earned by the program trading group.
I must admit that before last Thursday I knew of the problems outlined in this post, but I mostly ignored it because it was really just a theoretical problem. But now that theory has gone to reality, I take the issue of program trading much more seriously.
Mutual funds often talk about investing in the market for the long term and provide tables showing how equities out perform over the long term. But much of that data comes from a time when systemic risk from program trading did not exist.
Because program trading controlling the markets is still relatively new, we really don’t know how often a liquidity event (and thus crash) will occur in this relatively new trading environment. Is May 6, 2010 a once in every year event? Will it happen every 5 years? Will it happen every 10 years? And to what degree? Will it just be a quick 10% drop? Or might it be 50% next time?
Because I’m out of the market right now, it is a good time for thinking about problems like this. I can’t put my money under my mattress.
I can say that as long as the casino is rigged, alternatives such as real estate and venture capital seem much more appealing. Alternatives are also appealing as a protest vote too. Another alternative is to only play the market during bear markets and to play from the short side. That might reduce opportunities, but you can then perhaps benefit from a liquidity event.
Other ideas include using a simulator to trade a liquidity crash so that you are prepared. But this only works if you trade virtually full time. You have to assume that hedging stock won’t work so it requires getting well versed with trading the ES mini. But you still must accept the assumption that shorting the ES mini will be available. Not a great solution.
The good news is that coordinating the markets so that the circuit breakers actually work (instead of making things worse) should also help quite a bit in mitigating the risk of a liquidity event. If one or more oligopoly firms causes a liquidity event, they can stop the trading until they are able to go back online. Assuming that the computers continue trading when trading resumes, this should be a big help.
But what if there is a major exogenous event? Perhaps the oligopoly firms find it too risky to turn their computers on when the market resumes trading. What then? Well, if you are long, you better hope that the PPT exists. |
|
狗仔卡
