Price momentum is widely cited as a factor in stock selection, though I am not a natural born believer. Purchasing something because it has become more expensive does not put me at ease. A blog post at alphaarchitect.com [1] was recently published that cited how price momentum with small changes offers favourable returns. This blog post shares the results of using Equities Lab to test the idea that smoothly increasing prices can form a good momentum strategy.
The screener
The screener requires that a stock be sufficiently liquid with the following definition
- Market cap > 500 M$
- minimum average volume within 20 days > 250k
- minimum price / share > $2.5
The top 20 stocks are chosen from a score defined by the sum of:
- 1 * rank of price ratio (the price ratio is the current to 1 year ago price ratio)
- -1 * rank of price variance (the statistical variance of daily variations)
This score favours that have risen in price over the past year with small fluctuations.
Reality factors
- slippage: I have assumed a conservative value of 0.5% rebalancing loss due to bid/ask spread and commissions.
- rebalancing: Monthly at the beginning of the month.
- delay: the performance assumes a delay of 1 day from Equities Lab screen results to updating the portfolio.
Over the time period of 1996-2015, the screener returns an insane CAGR of nearly 32% with an gut wrenching standard deviation of 43%. Looking at the log-scale graph of returns is troubling: it appears that the returns are less favourable with time. In fact, the last 2 years have negative returns.
Is the momentum effect getting perhaps arbitraged away?
Notes
The standard deviation function within Equities Lab seems to be defunct at the moment, so here is how the variance of the daily price fluctuations was found:
