Resist it or embrace it, the age of technological innovation is upon us. Disruptive technological change pervades all aspects of the global economy. Growth in the conversion of huge quantities of information into digital format has been exponential during the last decade. Signals from a variety of equipment, including desktops, smartphones, cameras, cars, satellites, networks, and industrial and environmental sensors, have spawned a torrent of data, ushering in the era of Big Data. The plethora of intelligence unleashed by these advances has, amongst other things, improved efficiency, and converting intuition in to data-driven decision making.
Today many of these early entrants have become among the world’s most successful companies.
BRAVE NEW WORLD
The investment industry has not been immune to technological change. That is shown by the way it has embraced Smart Beta: rules-based, scalable portfolios created to avail of new techniques in data analysis, and purporting to offer investors intelligent exposure to selected factors, while limiting potential deviation from the market and aiming at outperforming a marketcap-weighted passive index. Just as many of the new technology business entrants have attracted mass consumer support, price sensitive investors have been entranced by the efficiencies and technological innovations demonstrated by Smart Beta strategies.
The ability to automate tasks and expand machine learning into the areas of pattern recognition, classification, and even prediction has engendered a litany of innovative start-ups. Digital platforms, through their control of customer experience and revenue matching techniques, have allowed new entrants to scale up rapidly and compete aggressively in industries such as car transportation (Uber), tourism (AirBnB), and messaging (WhatsApp). They have become self fulfilling aggregators of supply and demand in their respective universes.
While Smart Beta has attracted accelerating inflows over recent months, Naïve Beta ETFs, which track familiar indices, still dominate the ETF sector. Mesmerised by the expectation of improved efficiencies, whether from cheaper pricing or data driven investment decision making, few seem to have challenged the distorting impact that Smart and Naïve Beta have had on the fabric of the equity market. The advent of more frequent and severe rotations, at a style, sector and factor level, have broadly coincided with the increased flows into these strategies. Challenges in attaining efficient price discovery, and a proliferation of crowding across various parts of the market, have coincided with the ramp up in ETF and index strategies.
Investors have been blessed over recent years by a generally rising market environment. The return structure of most Smart Beta strategies has yet to be untested. Investors’ attention may have been distracted from robust scrutiny of the risk exposures generated by the proliferation of these strategies. Risk analysis and performance attribution (see graph) suggest that many Smart Beta strategies may actually contain different risk exposures, and offer a very different return profile, to that suggested by their particular label. For example, low volatility can lead to a short value exposure, generating unintended consequences in a broader portfolio context.
We argue that the careful selection of the right data inputs and performance metrics, especially the overall evaluation criterion, is an integral characteristic of successful data-driven decision making. Any system is likely to make mistakes and exhibit biases. The goal should be to select an approach that minimizes biases and errors, and allows them to be corrected swiftly and easily.
In a bid to meet these challenges, our funds are run with a dynamic alpha weighting scheme, utilising information on market state, rather than average factor returns over a predefined window length. Our weighting scheme is dynamic because we recognise factor returns’ short-term cyclicality, driven by the market environment. Cyclicality of factor returns can lead to alpha instability and downside risk (when the factor is out of favour); and as importantly, it can pollute the alpha with a lot of undesirable correlation. We seek to create factors that are robust and less buffeted by market state than the factors typically used in Smart Beta ETFs. We have undertaken a complete rethink of how we capture risk premia such as value, momentum, and growth. We believe that our current investment process is far more responsive to risk, and allows us better to manage changes in volatility or correlation.
So while the machines may be on the march, it may be best not to get too carried away by the latest fashions. There is a reason why the first letter in the Greek alphabet is not beta.