At stock market flash crash, prices fall precipitously within a short time. For example, on May 6, 2010, the Dow Jones Industrial Average fell by more than 1,000 points, or nearly 9 per cent, in just 10 minutes, deleting about $1 trillion in market value .
Retail investors sustained huge losses, as some shares briefly exchanged at extreme lows before rebounding.
Sudden Market Volume Spikes
At 2:45 p.m. on May 6, 2010, an extraordinary surge in trading volume was registered in the stock market. The New York Stock Exchange data showed that the average number of trades per hour went up from roughly 5 million to over 20 million, contracted within the 15 minutes. The change could only be driven by trading algorithms working with a time horizon of less than 15 minutes and generating and executing hundreds of trade decisions.
High-frequency trading, in this way, can be seen as an activity that produces substantial market imbalances. The flash crash registered an approximately 9% drop of the Dow Jones Industrial Average, or over 950 points, recovering in minutes and finishing in the positive territory on the same day.
Advanced analytics could well track the irregular increase in volume. Alerts set to volumes higher than 10 million trades per hour would give a sufficient indication that the situation is likely volatile. It coincides well with the famous expression by John Maynard Keynes that “The market can stay irrational longer than you can stay solvent.” The concept reflects the idea that individual traders cannot compete in holding losing positions long term, as the market will eventually bring their balance to zero. In this way, the volume is an impulse indicating some irrationality about the change in trends. It means that the best solution to flash crashes and spikes in trading volumes is to get ready for changes, as they will occur affecting the investor.
Unusual Trading Activity from Bots
Over the recent decades, with the rising prominence of high-frequency trading , the roles of trading bots in the stock market have expanded. Consequently, there is a plethora of incidents when unusual activities of such bots could portend the impending market crisis. Such is the case with U.S. Treasury securities on October 15, 2014, when they experienced an unprecedented 37 basis point drop in yield over an extremely short period of time, and the phenomenon was reportedly triggered by algorithmic trades.
The most conspicuous signal that precedes crashes of such nature is the simultaneity of trades, and it was the most notable pattern in the course of this event, as trades that accounted for a considerable proportion of the overall trading volume occurred within a millisecond after one another, which is evidently beyond the reach of human capabilities. Moreover, they were not manually triggered but rather the product of market algorithms that prompt bots to react in a certain way under certain conditions, hence, the term algorithmic trades. It is particularly dangerous because it does not only imply more pronounced market movements but also exacerbates them, leading to flash crashes from time to time .
There are anomaly detection systems that are designed to predict and prevent similar instances of automated trading, which could pose dangers to the stability of the market. They are fed with the data regarding usual patterns of trading, and when they detect something that goes out beyond established norms, they signify the possibility of crisis by means of flashing. J.P. Morgan reportedly said that “the wise man, at the beginning of the path of wisdom, was only a doodoo head” . Using his logic, it seems that those who understand and anticipate the signals that bot activities send will be able to effectively avoid or survive potential market crises.
Rapid Shifts in Market Sentiment
The financial markets are extremely sensitive to shifts in sentiment. This can usually be quantified in rapid changes in certain metrics, with the VIX being the most famous in financial circles. The VIX or the ‘fear index’ is investors’ outlook on the market where low values indicate confidence and high values fear or panic.
On August 24, 2015, the VIX spiked by an astounding 90% in a single day due to the massive shift in market sentiment flow confidence to fear . This was in response to growing fears of economic instability and recession in China which led to a 5% drawdown in the S&P 500 .
Following the VIX allows one to watch crucial shifts in market sentiment. A quick spike usually signals increased volatility and often precedes drawn-downs.
Traders use such knowledge to adjust their positions accordingly, for example, by switching to safer asset classes or taking short positions . A particularly useful way to track sentiment is to employ real-time analytics tools that analyze a vast amount of data.
These tools go through everything from economic news to financial indicators and even social media sentiment. To be able to predict potential early signs of the shift, they employ a mix of high-frequency algorithms to quickly identify changing patterns in markets. Winston Churchill said, “Those that fail to learn from history are doomed to repeat it” . This has become somewhat of a mantra on behalf of the data-first style investing, especially in the high volatility of financial markets.
