Joseph Mariathasan: Factor investing – is it overcrowded?
The basis of active management is that markets are not efficient. Finding those inefficiencies has been the bread and butter of active managers for decades. What the last decade has shown, though, is that there are structural effects that give rise to inefficiencies, and these can be accessed systematically.
Much of this started with quant firms, whether hedge funds or long-only strategies. The rise of smart beta exchange-traded funds dramatically reduced the price of investing in such ‘factors’, meaning factor-based strategies are estimated to cater for as much as $1trn (€0.8trn).
Is there a danger that the marketplace is saturated and overcrowding will give rise to reduced returns? Amundi and Willis Towers Watson in a recent report argue that “a key skill in factor portfolio construction is to recognise at what point a factor strategy starts to run into capacity or liquidity constraints, or when it might risk a degradation in performance as a result of crowding”.
La Française Investment Solutions (LFIS) in a recent seminar argued that the answer depended on what type of risk premia you looked at. At one end lie risk premia, and at the other lie style premia.
Risk premia include the well-known value and size effects. LFIS argues that they exist because they compensate the rational investor for taking on extra risk. With value stocks, it is the risk that share prices fall further because the reason for the low valuations intensifies – the ‘value trap’. Small cap stocks tend to have businesses that are more concentrated by sector and geography, and are therefore more vulnerable to disruptions than large cap companies.
Both value and small cap stocks have a natural cap to the amount of funds that can be placed in them. As money is poured into these strategies value stocks no longer look valuable and small caps no longer look small.
Style premia do not have a natural cap. They reward certain types of investors who are able to use their lack of investment constraints to invest in strategies that are impacted by structural biases seen in their major client bases – some clients can’t use derivatives or leverage, for example. Such premia can be accessed by arbitraging relative valuations of credit default swaps and the underling bonds. LFIS likens these to a cake that has to be shared amongst many guests: the more guests there are, the less is available for each one.
There is an ongoing debate as to whether style premia are overvalued or not. Rob Arnott, founder and chairman of Research Affiliates, argues they are, while Cliff Asness, co-founder and managing principal of AQR, sees valuations as reasonable by historic standards.
The results a manager gets from analysis of valuations appear to be highly dependent on the exact method used. Two different managers both claiming to run momentum strategies will get very different results depending on several inputs, such as the exact time period used to generate trading signals.
LFIS admits that certain factor strategies are prone to overcrowding, but it argues that the issue may be overstated. Many of the industry’s most famous investors, such as Warren Buffett have performance that can be explained by a combination of style and leverage – in Buffett’s case an exposure to value, low-risk and quality factors with a leverage of about 1.6.
LFIS argues that it is not factor investing per se that creates dislocations in markets. That argument is an excuse for the real issues of data mining, overfitting, cost-ineffective implementation, and changing correlations between securities. In August 2007, for example, quantitatively managed equity funds as whole lost over 5% in just 3 days as a result of them all having similar portfolios with a need to deleverage at the same time due to volatility and correlations increasing in the marketplace.
For investors in factor strategies, LFIS’s view may be comforting. But perhaps investors should also heed the advice of Amundi/Willis Towers Watson and monitor potential crowding by assessing the current valuation of a factor relative to its long-term history. A sudden and strong multiple expansion may be a good sign to avoid the crowds.
Click here for all the content from IPE’s Factor Investing Special Report.