Following on from the last Newsletter outlining some of the results from the paper titled “Value and Momentum Everywhere” today I’ll provide its main results on momentum for the US, UK, continental Europe and Japan.
First, a general note on momentum research
Academics have been studying momentum for decades now. There are various possibilities, but in price momentum the fundamental question is whether shares showing high returns over a few months then go on to show high returns; and whether shares showing poor returns over a few months continue that way.
A commonly chosen period is three months prior period returns to classify shares into high returns or low return groups, i.e. rank companies with the highest return at the top, falling to the lowest return shares. Then split into groups, such as one-tenth of all the shares (this one is called decile analysis).
Then the subsequent, “test period”, returns are observed for those portfolios. The test period might be 3, 6, 9 or 12 months after portfolio formation.
Alternative prior period returns might be 6, 9 or 12 months, followed by various test period returns.
If momentum studies go beyond 12 months they run into the confounding problem of return reversal. This is the observation that shares with extreme return over YEARS (not months) reverse the pattern, i.e. prior period losers on the whole outperform, whereas prior period winners under-perform – see earlier Newsletters (search under return reversal).
Momentum studies have been conducted in dozens of countries and over many different time periods.
It’s a phenomenon strongly believed to hold true by many in the financial community. But, I have my doubts, which I discuss after the results from the Asness et al paper.
How momentum is defined in this paper
A very simple approach is taken to classify shares as having Low, Middle or High momentum. Share returns over one year are taken. Then they are put in order and allocated to equally-sized portfolios, Low, Middle, High.
A reminder of the sample used:
Only the most liquid shares in the markets studied are included for annual portfolio formations. The portfolios are held (statistically speaking, i.e. in a computer programme that can go back in time and “buy” shares) for a year and then sold.
In the case of the US markets, 17% of its quoted shares are included from 1972 to 2011. There are 724 for an average year in the annual portfolio formation.
For the UK 13% of shares included from 1972 to 2011. The average number is 147 companies.
For continental Europe 20% were included, 1974 to 2011, with an average of 290.
For Japan 26% were included, 1974 to 2011, with an annual average of 471.
All the results shown below are statistically significant in that the returns on the portfolios are statistically above zero with at least a 99% probability. The differences between the High and Low portfolio returns are statistically significant different for US, UK and Europe, but not for Japan (t-statistic of only 0.57 – we are looking for at least 1.96).
The United States of America – One year Momentum
In America we see that High momentum shares out-performed growth, giving an average annual return above the Treasury bill rate of 14.2% compared with 8.8% for the Low portfolio.
United Kingdom – One year Momentum…..
To read the rest of this article, and more like it, subscribe to my premium newsletter Deep Value Shares – click here http://newsletters.advfn.com/deepvalueshares/subscribe-1.