'Het kan helemaal misgaan met smart beta'

Het kan helemaal misgaan met smart beta, waarschuwt Robert Arnott, topman van Research Affiliates, het Amerikaanse bedrijf dat een van de eerste smart beta-indices ontwikkelde.
Robert Arnott,  Research Affiliates

In een vorige week verschenen rapport waarschuwt Arnott (foto) dat de toenemende populariteit van deze strategieën de komende drie tot vijf jaar wel eens tot fikse dalingen van de performance kan leiden.

In relation to the Research Affiliates paper that is cited in the article, which calls into question our Risk Efficient strategy (also termed Efficient Maximum Sharpe Ratio). Here are some remarks on that paper and on the comments that concern us in particular: 1) Concerning the EDHEC-Risk Efficient Indices. The analysis of smart beta strategies corresponds to pseudo strategies that do not respect the methodologies of the actual indices that are available. Clearly, relative to the pseudo strategy tested in the document, the EDHEC-Risk Efficient Indices are based on a much more liquid universe, and include turnover control to ease investability. Using pseudo strategies creates the risk that results may be dependent on the specification choices retained by Research Affiliates more than the actual strategy as used in the EDHEC-Risk Efficient indices. In addition, it should be noted that this Risk Efficient strategy is a diversification strategy and not a strategy that aims to capture a premium associated with a particular factor. 2) On the subject of factor indices. We have other offerings drawn from our research which are not studied by RAFI. These smart factor indices, published by ERI Scientific Beta, an EDHEC-Risk Institute initiative, rely not only on alternative weighting schemes but also tilt to well-known factors documented in the academic literature, including value, momentum, low risk, size, profitability and investment. It has been well-documented, both empirically and theoretically, that such factors deliver risk premia in the long term because they expose investors to additional types of risk. There is no consensual finding in empirical asset pricing that such factors generate premia because they provide exposure to undervalued stocks. Thus, the claims in the Research Affiliates document are inconsistent with the entire field of asset pricing research which has documented the capacity of such factors to deliver long term premia without relying on periods in which they are supposedly undervalued. More generally, academic research is careful not to rely on a specific valuation model allowing what is overpriced or underpriced to be identified. The Research Affiliates article takes a different standpoint and pursues a more non-academic approach by trying to identify factors and strategies as being currently overpriced. Of course, trying to identify things that are over and underpriced and using this information for trading decisions is a common approach among active traders but it has nothing to do with academically-founded smart beta approaches. 3) It should be noted that the insights of Research Affiliates that many factors are now overpriced comes as the fundamental equity indexation strategy promoted by Research Affiliates has led to live underperformance over recent years, relative to a cap-weighted benchmark, while many of the strategies qualified as being now overpriced in their document, have actually outperformed. FTSE reports in the latest edition (as of 31 January 2016) of its "FTSE Alternative Weighted Indices Monthly Overview" that the five year mean annualised live excess return relative to the cap-weighted index of the RAFI 1000 Developed index was -2.2% which compares to +2.5% for the EDHEC-Risk Efficient Index for Developed, also over the same live period. The corresponding numbers for emerging markets are -3.1% (RAFI) and +2.8% (EDHEC), again for live periods. An implication of the results presented in the Research Affiliates document seems to be that now is a good time to buy the underperforming strategy as it will be surely undervalued, and to avoid any strategy that has actually delivered its promise of outperformance over the past. While one may understand that this implication may be desirable from the perspective of providers of underperforming smart beta strategies, it should be recalled nevertheless that the fundamental equity indexation strategy was initially promoted as delivering long term outperformance, not as delivering underperformance which would eventually make it a good buy at some point. Ultimately, when investors now have serious live track records, it is curious that those who criticise factor mining or data mining are reduced to creating "false concept mining" in order to justify their poor track record.