By Dr Peter Hopkins, principal, Style Research. Style Research is a provider of factor-based analytics for investment professionals.

Factor ETFs – Eat before you pay.

There are lots of ways to entice people into parting with their money – and attractive packaging and descriptions are near the top of the list. ETFs are no exception to the aggressive marketing taking place; neither are the factors that are often used to calibrate the performance of these investments. But can you really trust the products? And what should you look out for?

While there are regulations that require certain amounts of information to be laid out, ultimately this may not give enough insight into the actual product. The choice of ingredients and how they’re blended really matters, and the end-quality will depend on the recipe, the baking process and the skill of the chef.

As more and more people get into the factor world, exploiting bigger data, an increasing number of strategies are being put forward. A recent paper presented at Inquire Europe looked at two million combinations of financial accounting data and tested strategies based on random combinations of variables. Needless to say, many of these produced good back-tests, but the ultimate strategies surviving a series of increasingly stringent statistical tests were composed of nonsense criteria with no economic rationale for their success.

Care is required to sift through the noise to ensure that factor choices make sense. If you take a look at league tables of factor performance, you’ll see that there is a wide variation in return success across different categories of factors. Take value-oriented factors. They were a poor contributor to performance in 2015, came top in 2016, but have dropped to bottom again this year. Compare this with momentum, which has done just the opposite.

Given the difficulties in forecasting factor returns, multifactor funds have emerged, seeking to diversify the factor mix whilst attempting to retain exposures to the desired factors. Imagine now that you are looking to build a multifactor product. How are you going to combine the factors? How will you weight the stocks and rebalance? How do you control risk along with maintaining exposures to the targeted factors while avoiding others? There are many choices, each one adding to the challenges of really understanding the make-up of the fund.

And that’s where a consistent factor framework for ETFs can help.

First, we need to be parsimonious. You should use well-established, transparent fundamental factors that would be recognized by all investment practitioners, not just quants or factor specialists. The finer granularity found at the factor level is important because there can be great variety in outcomes even within popular style categories.

For product creators, even the most systematic of approaches can give rise to many variations, contingent on which factors and styles are selected, and how they are put together. Alongside those choices, portfolio construction and implementation make a huge difference. Taken together, these variations lead to many investment structures and outcomes. Viewing these in the context of recent factor behaviour and competitor offerings in a standard factor framework can help fine-tune the approach and identify future market opportunities.

For fund investors, relying on the label just won’t cut it – there are far too many possibilities and not all stated exposures make it to the final mix. To truly understand what’s in a fund it has to be sampled in some way. And in a factor ‘bake-off’, some of the labels match whilst others do not, with perhaps different exposures altogether coming to light.

Multifactor ETFs in the spotlight

We took a look at three US ETFs as an example, and used the Russell 3000 as the common benchmark, to put everything on a neutral playing field. The ETFs – a fund from Goldman Sachs’ ActiveBeta range; a Global X Scientific Beta ETF, and an iShares offering – all have relatively low risk and active share.

In the case of the Goldman ETF, the ingredients are described as value, momentum, quality and low vol. That seems broadly right, but the fund shows no bias to short or medium-term momentum currently. While the description seems fair, it likely wouldn’t be your choice if a strong momentum bias was important to you.

The Global X ETF has significant smaller cap and low volume biases, in line with the descriptions in its summary prospectus. But the claimed exposures to value and momentum have been much more variable over the period since its launch.

Finally, the iShares ETF is most in line with its stated ingredients. It describes itself as being quality, value, size and momentum. On value there is no doubt: it has by far the most significant exposure of our selected ETFs. It has a small-cap bias, and a positive, if not pronounced exposure to momentum. And though negatively biased to profit margin, it has biases to stability measures of quality. Of all funds reviewed today it is closest to what’s on the label.

Factor investing is not immune from subjectivity, and ultimately the choice of factors requires a belief that factors will continue to work in the future. But whichever factors you believe in, you want to make sure that the ETF you create or buy has those exposures, and a factor framework gives you the tools to do just that. If you don’t do that, do you really know what you are investing in?

(The views expressed here are those of the author and do not necessarily reflect those of ETF Strategy.)