Momentum investing is a quantitative technique that can be used to invest in the financial markets. In its essence, momentum investing is based on the notion that what has moved in the past will continue to do so in the future. This style of investing has proved to be quite beneficial to its practitioners. The reason why this phenomenon exists has not yet been properly explained and it remains one of the most remarkable anomalies to the efficient market hypothesis. Momentum strategies can be of two types, absolute or time series momentum and relative or cross sectional momentum.
In absolute or time series momentum, a stock or asset classes’ own past returns is considered independent of returns of other asset classes and is used to predict its future performance. In relative or cross sectional momentum, a stock or asset classes’ performance relative to other stocks or asset classes is used to determine its future performance.
When a stock starts moving with some force/momentum, it continues to do so before it loses steam and slows down or falls. Smartvalues algorithms attempt to identify this momentum, using quantitative parameters and ride the growth wave and exit when it determines the momentum may be going away and maybe a reversal is on the cards. This can be done much more efficiently and with speed using algorithms running on computers
Since the strategy requires measurement of momentum, it requires it is best done using computers. For long term momentum strategies very advanced computers are really not required. But shorter duration momentum strategies may require quicker response time and therefore more advanced computers.
Conventional momentum strategies are rule based and rely on certain fixed time interval to measure momentum. Most if not all of the parameters tend to be fixed. While creating a portfolio of stocks based on momentum, there are a few parameters that need to be input such as look back period which forms the basis of calculating momentum, the universe of stocks or the grouping of stocks from which the portfolio stocks are selected and of course the number of stocks in the portfolio.
Thus these methods have a fixed approach which can limit their effectiveness. The advent of AI can and will change the way technology plays a role in the implemetation of momentum strategies.
Let us take the example of our own smartvalues and the AI model that powers it. The configuration is more flexible and adaptable and can therefore operate with multiple momentum time frames as well as adapt to various human defined universes. Momentum is also not a homogeneous construct and there are different ways of calculating momentum. An advanced model like the one used by smartvalues can look at momentum in multiple ways and calculate relative strength to construct portfolios.
Momentum trading strategies will evolve with the evolution of technology. The traditional methods will continue to be used but in different ways. The practical implementation approach will continue to involve modeling, building and testing momentum strategies.
There will be a proliferation of cross momentum strategies as more and more datafeeds and databases enable systems to test momentum trading strategies across asset classes. With the proliferation and increasingly easy access to data, quantitative trading and investing approaches will continue to dominate, a trend that has been in play for sometime now.
Momentum strategies will be a centrepiece of this phenomenon.
Mr. Soumitra Sengupta