The term “Smart Beta” was developed by marketing in the last decade to describe what is essentially a quantitative strategy to capture a well-documented factor premium. Quants have been doing this for decades and while I grant you that the moniker “Smart Beta” may have contributed to their recent surge in popularity, comparing some of them with their quantitative elders is a bit like comparing today’s Boy Bands with the Beatles. Most fail and break-up never to be heard again (literally). Likewise, most of today’s Smart Beta products fail to gather enough assets to remain viable for their makers and are eventually retired from the shelf.
Investing is forecasting. There is no getting around that. And there is no doubt that factor investing works; it has for decades. But one factor does not outperform all others all the time. They take turns. Value, for example, had a great run of outperformance after the GFC but had a dreadful year last year, when Growth was the big winner. Successful factor investors use combinations of factors and/or factor rotation strategies. The key is researching which factors perform well in what type of economic and market environment then forecasting the current and near future to select the right factor premium to capture for that period of time, or to hedge by choosing a diversified set of factors.
Asset Management companies trying to raise their AUM, need to have a family of factor premium products to offer their investor clients so as to enable these factor rotation strategies. The democratization of factor investing was helped by Marketing’s spin name “Smart Beta” – it does roll off the tongue better than “quantitative strategy to gain exposure to a systematic source of return”, I’ll give them that – but not all Smart Beta products deliver what they promise. Investors need to first decide which factor premium they want to gain exposure to, then select the Smart Beta product that best delivers that exposure via a transparent and disciplined methodology.
What should Smart Beta designed focus on in order to ensure success for their product? If you ask the Marketing departments then transparency and simplicity are the most crucial ingredients, as it will help raise AUM faster if more investors (think they) understand it. But because I take a ruler to draw a line in the sand doesn’t mean it’s pointing in the right direction. With a lot of smart beta products, what’s in the can isn’t necessarily what it says on the label. That’s because there is a tradeoff between simplicity and the purity of the factor exposure you can deliver.
Factor investors whose research support their decision to invest in a specific factor (over others), want a smart beta product that delivers that factor premium alone, and not that factor plus, plus, plus. The goal should be to design a portfolio that delivers the highest exposure to the target factor given the investment constraints (e.g. Long-only, liquidity, etc.) but at the same time neutralizes all non-target factor exposures so as to ensure that it is the target factor return that drives the portfolio’s return, and not noise from non-target factors. This requires a disciplined quantitative portfolio construction process involving a well-defined multi-factor risk model and the use of an optimizer.
Factor purity and exposure cannot be delivered via a simple screening process involving a ruler and a number of stock target. For example, take a Value smart beta product with a significant small-cap bias and a large concentration in one or two sectors. In this case, what will drive this smart beta product’s performance? Is it the Value premium? Or is it the Small-Cap premium? Or is it simply one of the two sectors the product has most of its AUM invested in? If as an investor my research does not lead me to believe the small-cap premium will outperform in this environment, but that the Value premium will, investing in this product might not give me the performance I’m expecting because it will be dominated by the Small-cap premium and the two sector premia. So, while Value may very well outperform, as per my expectations, my smart beta product might not because of these other three significant sources of return. In this example, Value worked, but my Smart Beta product may not.
When it comes to delivering a high-quality smart beta product, you have to put your math where your mouth is. Too many Smart Beta products are built using dumb math. Dumb math is transparent and easy to explain, but it doesn’t deliver the kind of factor purity and exposure that serious investors require. Complex math can be just as transparent as the dumb kind, you just need to have to folks who truly understand it, explain it. Focus on delivering to investors exactly (not approximately) what they want and give up the idea that simplicity is a must-have. Simple is how you make a pound cake, not how you design a financial product. Just focus on delivering the desired outcome, and turn your black box into a glass one. Investors will come.
Smart Beta products have become hugely popular in the last decade. It started with the whole Low Volatility craze, then Value, and more recently investors shunned both of those for Growth, Momentum, and the IT sector theme – the so-called FANGS stocks. But be mindful that popularity does not equal quality. After your research has identified the factor premiums you want exposure to, you need to identify the best Smart Beta product to gain exposures to them (based on what we discussed above).
Below are two tables summarizing factor performance in 2017. The top table reports the individual factors’ (Y-axis) full year performance for each individual market (X-axis). We highlight the best (green) and worst (red) performance for each factor. Momentum had the best performance in Asia Pac ex-Japan and the worst in Japan (still a positive one though!). The bottom table puts those 2017 factor performances in perspective by reporting their percentile rank over our entire history. For example, the Size factor return (Large Cap premium) for the US market may not have been the highest Size factor return across all markets (the highest Size return was in Asia Pac ex-Japan since it is highlighted in green), but for the US market, this performance represents a 99 percentile rank (i.e. almost the highest in its history).
Focusing on Asia Pac ex-Japan, we see that the best performing factor returns have been Momentum, Size (Large Cap), Value, and Leverage (companies with higher leverage than the average in the market). These factor returns were also very significant if we look at their percentile ranking in the bottom table. We also see that Volatility (high-vol) didn’t do well at all, so investors were betting on Large Cap, Value stocks with a Growth characteristic, and that have outperformed in the past (i.e. high momentum), and have more debt (to fuel investment in growth) than the average stock in the market, but are less volatile than the average stock in the market (i.e. low-vol outperformed high-vol). In other words, GARP (Growth At a Reasonable Price) with downside risk protection, seems to be the flavor of the year in Asia Pac ex-Japan.
As to which smart beta product is the most popular, that can easily be monitored from inflows data. But, as we have discussed above, popularity based on AUM inflows doesn’t mean they were the best ones to own, maybe just the best ‘marketed/understood’. The best Smart Beta product to invest in is the one that gives you the most assurance that the returns of the portfolio will match what you are expecting in terms of target factor return. So first decide what factor premium you want to invest in, then find the Smart Beta product whose portfolio construction methodology and performance attribution analysis gives you the highest confidence that its returns are overwhelmingly driven by the desired factor exposure (and not something else).
You cannot look across the Smart Beta spectrum performance and decide which product performed best. If the Dividend Yield factor premium had a bad year, like in Australia in 2017, then Smart Beta products harvesting that factor premium will/should also underperform other smart beta product harvesting a winning factor premium, like Momentum, or Profitability in that market (Australia). That doesn’t mean that the Dividend Yield Smart Beta products are ‘bad’ and the Momentum ones are ‘good’.
When it comes to Smart Beta product performance, what matters is how the product performed vis-à-vis the pure factor return it is supposed to capture. If Dividend Yield under-performed the market in 2017, then those Smart Beta products marketing themselves as Dividend Yield products should also under-perform. If their under-performance comes close to matching that of the Dividend Yield factor return, then that is a well performing Smart Beta product. If the performance is uncorrelated with that of the factor it is tracking, in this case Dividend Yield, then it is a poorly performing product (i.e. it is not delivering the return it is meant to deliver).
So, when it comes to Smart Beta Products, performance is measured relative to the factor premium it is named after, not absolute performance. If it says “Chicken” on the label, but it smells like fish when you open the can, put it back on the shelf and move on.
Additionally, Smart Beta is not a one-size-fits-all investment proposition, and investors need to find the one-size-that-fits-just-me product. Factor investing is not done in a vacuum. Investors will have other portfolios and investments and may simply be looking to add a certain factor premium leverage to their existing plan. Maybe they are bullish on the APAC ex-Japan markets, but also believe Value will outperform Growth in the next two years. In this case, they may want to buy a core market ETF plus a Value Smart Beta ETF, but will want to make sure to add only the Value Premium to their core investment plan and not cause any unwanted style drift because their smart beta product is ‘infested’ with other non-disclosed style exposures.
They may also have a specific definition of value, which may or may not match the definition of the Value ETFs out in the market. Say you think of value as Earning Yield but the Value ETFs are all based on Book-To-Price. In this case, constructing your own Value portfolio is a better idea than buying one that is not based on what your research tells you will be a source of outperformance. The same goes for Momentum. If your definition of Momentum is a stock that has outperformed over the last 90 days, but the Momentum ETFs are all based on a definition of momentum of either 20-days or 240, then constructing your own factor portfolio makes more sense so you can avoid a big factor misalignment problem.
The issue of finding a smart beta product whose factor premium definition matches the investor’s exactly, is bigger with fundamental factors like Value and Growth, then it is with technical factors like Volatility, Momentum, or Size. That’s because Value or Growth, like beauty, is very much in the eyes of the beholder. Technical factors are much more objective. In my opinion, outsource the technical factor premia, but in-house the fundamental ones, to ensure they match your firm’s definition. There are, of course, a lot of other considerations when selecting an ETF (e.g. liquidity, volume, costs, etc.), but from a factor investing point of view, what sets them apart is a matching definition of the factor premium your seeking to harvest. And the lack of a disciplined portfolio construction methodology to capture it should be a deal-breaker.
This blog post was inspired by my conversation with Asia Asset Management journalist Paul Mackintosh around the performance of Smart Beta in practice. His article can be found here.