NASDAQ Data Loss And 1929

When there was a three-hour communications problem on Thursday, hitting the NASDAQ in New York, some may have thought that it was no big deal. After all, this was during the summer holidays, many business folks were out of town, and trading volumes were expected to be low. But others have seen the scale of how dependent traders are on electronic systems, and started to wonder.

There are two kinds of problems that “digital dependency” can cause. One, clearly, is the downtime and the consequential loss from all those trades that depend on being done immediately, that bet on, take advantage of, and generally depend on market movements that happen immediately after the time they are enacted. These need the computing power just to be possible.

Another problem with “digital dependency” is the development and use of trading algorithms, especially with the high-frequency trades that are a feature of hedge funds. These can cause price movements to act as a “hair trigger”, if you will, setting off large-scale selling that can wipe huge sums from stock values in frighteningly short spaces of time.

Added to these is the fear, maybe even panic, sparked by loss of information: the thought that events may start to move so rapidly that there is no time for rational consideration. The market sees a rush to sell, so investors join in, and if there is a data loss, or delay in reporting, the urge to sell does not abate, but gathers pace. In this way nothing has changed in 84 years, save the size of the sums involved.

In 1929, when data was carried in more primitive forms – literally, the ticker brought news of changes in the price of stocks – this was all superbly illustrated when, on what has become known as Dark Thursday, 24 October, there was an unexpected and sudden rush to sell stocks. Many investors had bought “on margin”, or loaned most of the money to pay for their investment. Brokers tried to contact them.

If investors could not be contacted, or were unable to offer up more margin, their holdings were sold. The ticker fell way behind. This loss of data merely resulted in a wild panic as yet more investors rushed to sell. True, there was some respite later in the day, but then, the following Tuesday, there was yet another torrent of selling. The market carried on down for almost four years.

Now we no longer depend on the ticker, but on the installed power of computers and communications networks. And this week’s NASDAQ crash has showed that this dependence may have its downsides. Data loss, hair trigger trading algorithms, and old fashioned panic could yet combine to set in train another Great Crash. The only improvement over the years is the likely speed at which it will happen.

Yet so much is invested in Governments snooping. Which is more important?