There are four different approaches to dynamic factor allocation (macro sensitivities, trend, fundamentals and risk) and one may not fit everyone's needs, said Hitendra Varsani (pictured), executive director, equity applied research, MSCI, speaking during the MSCI annual conference on Global Investing and Risk Management on May 16 in London.
"Perhaps it is a combination, perhaps it is two or three," said Varsani. Different factors have different sensitivities to macro scenarios, according to Varsani. "The first, or dominant drive of factor returns are growth prospects," said Varsani who noted that defensive factors such as minimum volatility, high dividend and quality produce excess returns during sharp decline whereas pro-cyclical factors such as enhanced value and equal weight deliver outperformance during sharp upticks. "Momentum is an odd one. It tends to benefit the most when there is stability in the market, there may be gradual changes but not extreme changes."
The second approach, trend, is an interesting one because it has been shown to work across stocks, sectors, countries and asset classes, according to Varsani. "Trend is very easy to calculate, very easy to back-test but it is depending on the parameter." There are different look-back periods with some investment managers looking at return between one and six months. "The market standard seems to be centred on 12-months in terms of momentum and in the short term it is all about reversals," said Varsani.
"We have seen that in the market recently. The first half of 2016 minimum volatility outperformed, leading up to Brexit, while post Brexit, expectations of growth increased, and value outperformed consistently for the next six months. This year quality has consistently outperformed," said Varsani, noting that this rotation is not random.
All of this looks great in terms of back-testing, however, in reality it is quite challenging to implement, according to Varsani. "Constructing an index, or a strategy, based on a one-month signal is going to have excessive turnover. If you translate this into a long-only implementation this will have approximately a 200% turnover and there are very few investors who are really willing to stomach that," said Varsani. "Returns are great but practicalities are questionable."
Fundamental is one of the most useful metrics to look at because you have an anchor of price and replacement value, said Varsani. "There is lots of talk about factor investing, smart-beta, becoming popular, it's crowded and becoming expensive. We actually find that valuations are in line with the long term averages." The only exception is dividend yield which has a very high price, according to Varsani. "That comes as no surprise since we have been in a low yielding environment for several years now, with bond yields, credit yields compressed," said Varsani.
"That search for yield has been very intensive and equity yield has been fairly compelling. That has raised the price of stocks and raised the valuations. It is the most expensive it has been for 16 years," said Varsani.
Whether or not it is fair to say that the valuation of momentum today gives you a good indication of the returns over the next two to three-years remains to be seen, according to Varsani. "Probably not," he said. "Momentum has a 100% turnover. The valuation I am looking at today is not reflective of the underlying securities of that index in the future. So that anchor is less useful. Whereas for low turnover strategies such as equal weight or value it might be more fruitful."
With the fourth approach, risk based allocations, you are putting a concentrated bet on a particular outcome, said Varsani. "One of the most attractive features of factor investing is diversification," he said. "Beyond beta there are additional return sources that I can tap into that are less correlated, help me diversify, help me give me a smoother ride, so this is great. However, something to be aware of is that the correlations between factors are also time variant. We have seen correlations among defensive factors - minimum volatility, yield and quality - reach extreme levels, most recently 0.6."
Momentum which is usually seen as the most uncorrelated factor actually became highly correlated, according to Varsani. "[Momentum became] almost correlation 1 immediately after Brexit," he said adding that because of the prolonged defensive cycle momentum attached itself to defensive stocks.
"Momentum became correlated with defensive stocks, and the diversification we thought we were getting didn't actually materialise. However 2017 is painting a different picture. We are now starting to see the markets normalise and correlations fall."
Click the link to view the full presentation and Dynamic factor allocation strategies.
Related stories:
ESG integration is a framework for the institutional investor, MSCI
The phenomenal rise and rise of multi-factor indices, LGIM
Patience and discipline is required when it comes to risk factors, SRP Europe