skip to navigation skip to content

Brain food


The five key ‘factors’ in factor investing

With factor investing and smart beta strategies on the rise, Professor Elroy Dimson of Cambridge Judge Business looks at five central factors that all investors should monitor closely.

financial graph on technology abstract background

Factor investing and smart beta strategies increasingly are “in vogue”, according to a new article in the Journal of Portfolio Management co-authored by Professor Elroy Dimson, Chairman of the Newton Centre for Endowment Asset Management at Cambridge Judge Business School.

Professor Elroy Dimson

Professor Elroy Dimson

These strategies seek to benefit from long-run premiums highlighted by academic researchers that look beyond traditional asset classes or indexes, in order to select a portfolio based on more specific or granular “factors”.

Almost three-quarters of asset owners are using or actively evaluating smart beta strategies, and the use of smart beta indexes is rising quickly, says the article, entitled “Factor-based investing: the long-term evidence” – which explores various research into these issues, including studies by Professor Dimson and his article co-authors, Professor Paul Marsh and Dr Mike Staunton of London Business School.

While researchers have identified more than 300 factors that may fit into such strategies (though most would not survive independent testing of their long-term effectiveness), there are five key factors that all investors are exposed to – whether knowingly or unknowingly: size, value, yield, momentum, and risk.

The article co-authored by Professor Dimson was included in a special issue of the Journal of Portfolio Management dedicated to different areas of factor investing. The issue included an opening editorial comment by Professor Stephen A. Ross of the Massachusetts Institute of Technology, who sadly passed away in March 2017, on the “Theory, statistics and practice” of factor investing.

Drawing from the Journal of Portfolio Management paper, Professor Dimson discusses the research history surrounding five smart beta factors that investors should monitor closely:

1. Size

Research spanning over six decades in the United Kingdom (and even longer in the United States) has found evidence that smaller quoted companies have provided the best long-term returns. Though the advantage of small-capitalisation companies has been uneven and inconsistent, there has been a general outperformance by small caps over the long term.

With dividends reinvested, £1 invested in UK large caps at the start of 1955 would have grown to £1,087 by the end of 2016, or an annualised return of 12 per cent. The same investment over the same period would have yielded £3,220 for mid caps, £6,861 for small caps, and £27,256 for micro caps.

2. Value

Investing in value stocks has paid off handsomely in markets like the US and UK over the long-term. Value stocks sell for low multiples of earnings or book value, and such stocks may have suffered setbacks or be mature and unexciting businesses – but extensive research in many countries shows that the long-term performance of such stocks has been far superior to growth stocks. In the UK, a £1 investment in the growth index in 1955 would have seen an annualised return of 10.3 per cent to £419 by the end of 2016, while the same £1 invested in the value index would have generated £9,173 or more than 21 times as much.

That said, there have been some relatively disappointing periods for value stocks, including much of the 1990s and again after 2007 – and there is continuing debate over whether the historical overall superiority of value investing over recent decades can be expected to persist.

3. Yield

Above-average dividend yield has been linked to a historical return premium in a series of papers since the 1970s. The evidence provided by Dimson, Marsh and Staunton, spanning 117 years of UK equity market returns, is the longest such study of the yield factor. An investment of £1 in a low-yielding stocks in 1900 would have provided £6,810 by the end of 2016, for an annualised return of 7.8 per cent – while the same investment in high-yield stocks would have generated a 10.8 per cent annualised return, or £158,727.

There are four possible reasons why high yielders have outperformed: mere chance that is unlikely to persist (the authors consider this to be doubtful); national tax policies that have caused growth stocks to sell at a premium by favouring capital gains; investor enthusiasm for growth stocks that bid prices to unrealistic levels; and as a return for risk (the professors find this explanation hard to accept because high-yielding stocks had lower historical volatility).

4. Momentum

From 2000 through the end of last year, winners outperformed losers by 10.2 per cent each year among the 100 largest UK stocks – providing some long-term backing for momentum strategy. But it’s important to mention two key caveats about momentum strategies: they’re costly to implement due to frequent rebalancing, and there are risks of volatility and even whiplash when markets sharply reverse direction — as seen in 2009. Globally, the three researchers found that momentum investment has been profitable in 21 of 23 studied markets since 2000, with the important exceptions being Japan and the US.

5. Volatility

Several researchers have shown the historical superiority of low-risk investing compared with higher-risk strategies – although other studies report that excess returns accrue disproportionately in the first month after a portfolio is assembled, and that this largely evaporates when low-priced stocks (less than $5) are removed from the mix. Dimson, Marsh and Staunton examine risk estimates based on both short term (60 day) and long term (60 month) windows. When they switch to long-term risk estimates, they find similar performance from three different risk-based portfolios until the technology bubble burst in 2000 – with high-risk stocks performing disastrously from 2000 to 2003 but outperforming since then. Although low-risk stocks beat high-risk stocks over the full period 1960 to 2016, this is entirely due to the 2000 to 2003 tech bust. The authors express caution about extrapolating the past performance of low-volatility strategies into the future.