taylor@hplabs.HP.COM (Dave Taylor) (01/26/88)
The following is based on an article off the Associated Press newswire that appeared in todays Santa Cruz Sentinel. - - - - - Shearson Lehman Brothers Inc. Hailed in Limiting Computer Program Trading on Wall Street New York - A Wall Street assault on one form of computer-driven stock trading intensified Friday when one of the biggest stock brokerage firms, Shearson Lehman Brothers Inc., announced a limit on their use of computer program trading. The company blamed it for fanning what they call `post crash volatility' and weakening investor confidence in the stock market. Many brokers hailed the announcement as a step towards limiting what is widely seen as the harmful impact of `index-arbitage program trading' [this is defined in the article as the use of computers to simultaneously buy and seel large quantities of stocks and stock-index futures]. An institutional stock manager at another firm put it this way; ``I think Shearson wants to show that we as an industry group are not going to let one tool destroy the confidence the investing public has enjoyed for so many years. I think it's an extremely good move.'' The blame for the October 19th crash of the stock market has been placed on this index-arbitage program trading system by many analysts, with even more agreeing that the resultant wild swings and panicked buying and selling of stocks were caused by this system. There was even a presidential commission created to study the collapse, their report stating that this program trading `certainly played a role' in what happened. A member of a firm that is boycotting brokerages that use index-arbitage trading states; ``The institutions that aren't involved in index-arbitage program trading are bringing their muscle to the plate to try and stop it. That's what causes the violent fluctuations in the marketplace, and that's running people out of the market, both the instutions and individual. They just don't want to be involved in a market that moves that much.'' The New York Stock Exchange, in an attempt to allay investor concern, has also taken steps to limit index-arbitage program trading via its main computer system when the Dow Jones industrial average moves more than 75 points in a single day. They are also considering temporary trading halts in volatile stocks which would make it more difficult for program traders to calculate prices accurately [which would result in the system being considerably less useful]. Other exchanges, including the Chicago Stock Exchange, are considering similar limitations. Shearson, in their statement to the press, said that ``index arbitage program trading had been suspended in response to concerns expressed by clients that program trading may exacerbate market volatility.'' - - - - - So it appears that the speed and power of the computer indeed has finite limitations in certain areas...anyone care to surmise what the ultimate long-term effects of this sort of limitation will have? --- Dave Taylor
reggie@pdn.UUCP (George W. Leach) (01/29/88)
First, a general disclaimer on what follows. I know absolutely nothing about the stock market other than I lost a few bucks on a company a few years ago. So all that I have to say is speculation on the situation and not an informed opinion. Corrections to any assumptions made are appreciated. I think that the limitations are due to the inability for human beings to become involved in the decision making process. The trading decisions are made and executed so fast that manual overrides may not be possible. The human being is taken out of the loop. If there is anything that one learns from a human factors point of view it is the need for providing manual overrides for any automated system. However, if the process does not allow enough time for human reaction to the situation, it may be too late. The other angle on all this is just what factors go into making a buy or sell decision? Are all them quantifiable? Decision support systems can't emulate some human qualities that go into a decision making process like gut feelings, emotions, etc....... George W. Leach
Gloger.es@Xerox.COM (Paul Gloger) (02/02/88)
In reply to the posting by Dave Taylor in Computers and Society Digest, regarding "Computers and the Stock Exchange"... Arbitrage trading really consists simply of buying and selling essentially identical securities or packages thereof, at closely matched times, but in different markets or different forms, where the spread between the buying and selling prices is sufficient to make the effort profitable. For example, I would buy an index future based on the Dow Jones Industrial Average, and separately sell all the individual stocks comprising that average, if I had an opportunity to do so at a profit. Notice that this opportunity could only exist if the index were undervalued and the individual stocks overvalued relative to each other; that the arbitrage trade tends to drive the prices into relative alignment; and that the arbitrage trading is therefore inherently both self-limiting and stabilizing of the market. Arbitrage trading is essentially the same activity as retailing / wholesaling / dealing / brokering / distributing / trading, except that these latter terms are usually applied where the underlying commodity is a tangible item rather than a financial security. However, the activity in all cases lubricates the market, makes the market more efficient; and it is in all cases inherently stabilizing and productive. The widespread "agreeing that the resultant wild swings and panicked buying and selling of stocks on Oct. 19 were caused by this system (arbitrage trading)" is totally unjustified and unfounded. The most elementary understanding of the economics of the situation shows that it has beneficial effects exactly opposite those for which it is blamed. Therefore we have a most interesting issue regarding "Computers and Society," namely that this entire issue is a massive exercise by society in pure scapegoating of computers, falsely blaming computer trading for the unsettling behavior of the financial markets. However, I was provoked to this response, not by the newspaper article quoted by Dave Taylor, which sort of media trumpeting of massively fraudulent political allegations is too common to admit of a reply; but by Dave's own closing paragraph: So it appears that the speed and power of the computer indeed has finite limitations in certain areas... anyone care to surmise what the ultimate long-term effects of this sort of limitation will have? I would ask Dave to either please explain just what these "finite limitations" are and offer evidence of their operation; or else to please admit that what we have here is our own Computers and Society moderator casually buying into this massive scapegoating of computers by society. Paul Gloger <Gloger.es@Xerox.com>
franka@mmintl.UUCP (Frank Adams) (02/02/88)
I would like to take this opportunity to combat the myth that program trading was responsible for the stock market crash. It may have exacerbated the problem, but if so, it was not by very much. The biggest cause of the crash, as far as I can tell, is the practice by various big institutions of a technique misnamed "portfolio insurance". The idea of portfolio insurance is that if the stock market goes down a bit, you sell some of your stock, and invest the proceeds elsewhere. The idea is that if you sell quickly enough, your losses are limited. It should be obvious that if enough of the investment capital in the market is playing this strategy, it will stop working. All will try to sell at once, and the market crashes. This is what happened in October. Program trading comes in as follows. The people who engage in portfolio insurance have adopted the practice of selling futures on stock market indexes as the fastest way of getting rid of their stocks. There is another group of traders, who use computers to watch the relationship between the futures market and the underlying stocks. When these get out of line, they sell one and buy the other. (These people are known as arbitrageurs using the original meaning of the word, which has nothing to do with takeovers.) Now, when the institutions responded to falling prices by selling futures, this made the futures price drop. The program traders' response was to buy futures, and sell the stocks. Note what happens if the program traders don't exist. The futures market crashes even more heavily than it did. Eventually (read, later that day), the institutions stop selling futures at such a big loss, and sell the stocks instead. Thus the same crash happens, delayed only a few hours. Frank Adams
demers%beowulf@sdcsvax.ucsd.edu (David E Demers) (02/19/88)
George Leach writes: > The other angle on all this is just what factors go into making a buy or > sell decision? Are all them quantifiable? Decision support systems can't > emulate some human qualities that go into a decision making process like > gut feelings, emotions, etc....... There are a number of different types of programs. One of the major ones is designed to hedge index futures. Although it may not be possible to correctly predict the motion of a securities price (other than the famous maxim "It will fluctuate" - J.P. Morgan?), the relationship between a security's price and the price of an option on that security CAN be predicted, in so far as they bear an optimum relation to each other (some of the ingredients in the formula are empirically derived numbers, thus only known to some error tolerance.). Thus if the option price is too high, RELATIVE TO THE SECURITY PRICE, the option should be sold and the security purchased. Both will go up or down, closely paralleling each other, but eventually the option price will decrease with respect to the security's price. Since the S & P index is a good approximation to the market as a whole, such analyses are often done on the S & P and its index future contract, which is traded as a security. When the program kicks in a "buy stock - sell future" message, each of the securities underlying the S & P is purchased in its pro-rata quantity. That can add up to a lot of "buy" orders, especially when programs from a number of houses all act nearly simultaneously. Speed is of the essence, since you have to get the proper prices. Once all the orders come in, supply & demand operates and the prices rise. The sudden rise in prices may cause other programs to issue orders... hence the volatility. This is not the only thing going on, & sorry for the length - just wanted to point out that some things CAN be fully quantified (given the right assumptions!) in the market. Dave DeMers, J.D., M.B.A. {My methods may not be the best, but they had me sell in August...}
kurt@tc.FLUKE.COM (Kurt Guntheroth) (02/19/88)
Frank Adams writes:
I would like to take this opportunity to combat the myth that program
trading was responsible for the stock market crash. The biggest cause
of the crash...is the practice of..."portfolio insurance".
Both these names are symptoms of the same problem. Big institutions can
transact a great deal of business suddenly by trading stocks, options,
futures, and other financial instruments with computers.
The price of a financial instrument (say a share of stock) is based on the
market's perception of the financial health of the underlying company plus
the market's perception of the financial health of the general marketplace.
As this perception changes gradually with time, the price of the share
changes. At any given moment, shares of a stock go for a particular price,
and the market makers keep an inventory of that stock to sell at that price.
When orders come in at a reasonable rate, for individual stocks, this all
works pretty well.
When a computer trade dumps a whole bunch of stocks on the market, individual
market makers don't see the entire trade right away. Even though the
emergence of a bunch of shares for sale indicate a change in the perception
of the marketplace, this information is hidden from the individual market
makers, who buy up shares of stock AT THE PRE-PROGRAM PRICE. Within a few
minutes, the traders notice the sudden increase in selling volume and
decline in broad indices. By then it is too late. Shares have already
changed hands at "too high" a price. Of course the human-assisted part of
the program is to set the trade in motion when things are busy, or when a
period of upward motion will mask the program trade's effect for awhile.
What has happened is that the trading system is too efficient. The trading
system runs faster than the trader's ability to keep up with the trades and
set prices accordingly. This imbalance in the system causes a sort of
ringing; faster-and-higher-than-expected rises and falls in price. If the
system becomes unstable, you get 500 point down days.
Eliminating program trading practices, whether it is program stock
transactions or futures transactions would partially rebalance the system.
Slowing order entry would slow the system down to the point where the human
market makers could set the price effectively. People who say program
trading is the cause of the crash probably are misinformed as to the
nomenclature, but I think they are right that it is the computerized trading
system that permits the disaster to happen.
Kurtfranka@mmintl.UUCP (Frank Adams) (02/26/88)
>> I would like to take this opportunity to combat the myth that program >> trading was responsible for the stock market crash. The biggest cause >> of the crash...is the practice of..."portfolio insurance". > > Both these names are symptoms of the same problem. Big institutions can > transact a great deal of business suddenly by trading stocks, options, > futures, and other financial instruments with computers. > > When a computer trade dumps a whole bunch of stocks on the market, individual > market makers don't see the entire trade right away. Whatever speed transactions take place at, the market makers don't see the entire trade "right away". Markets get on fine in spite of this. Faster reactions by traders just make the market better. It doesn't matter if it's a human trader or a computer trader. The problem is that certain major trading institutions have been engaging in destabilizing trading practices. (The people who have been doing so are the ones who took the biggest losses in the crash, by the way.) It is probably the case that access to computers tempted those people to think they could get away with those practices -- indeed, as long as they could react faster than anybody else, and there weren't too many of them, they could. But this is the only way in which computer trading is the problem. Frank Adams