The use of ETF-based model portfolios may be rising on both sides of the Atlantic, however, a recent academic study has warned these strategies provided by asset managers create worse outcomes for end investors.

The report, titled Advising the Advisors: Evidence from ETFs, found asset managers regularly provided clients with portfolios based on their own ETFs boasting higher fees and worse returns than their equivalents.

Model portfolios run by asset managers have exploded in recent years, especially in the US. They provide financial advisers with one-stop, multi-asset solutions that can incorporate the firm’s own ETFs.

Authors from the University of Utah, Norwegian School of Economics and the Shanghai University of Finance and Economics stressed the odds ratio of ETFs affiliated with an asset manager being included in their models was 3.19 times that of unaffiliated ETFs.

Furthermore, affiliated ETFs with lower prior one-year performance and higher fees have a greater probability of being added to model portfolios. Evidencing this, the odds ratio of the worst-performing affiliated ETFs being included is 1.01 times that of the best-performing affiliated ETFs. Also, when the expense ratios of affiliated ETFs rise by one basis point (bps), their odds ratio of addition jumps 3%.

This creates a tangible impact on investor outcomes, the research said.

“On average, affiliated ETFs charge six bps higher expense ratios and generate 67 bps lower net year-to-date (YTD) returns than unaffiliated funds,” the research warned. “Their past performance, measured by the performance-rank percentiles over the previous one and three years, is also five and six ranks lower, respectively.”

These dynamics work in reverse for unaffiliated ETFs, the research said, where the odds ratio of inclusion in asset manager models fell by 1% for every expense ratio increase of one basis point.

Looking ahead, researchers examined the future performance of ETFs included in models and found no evidence to suggest affiliated products outperformed their unaffiliated counterparts in the longer term.

Baking in worse outcomes 

These dynamics would pose no threat if model portfolios existed in a vacuum. Unfortunately, they take place in a context where ETFs experience an average of 1.1% higher flows per month after being added to model portfolios.

Offering a real-world example of models driving asset flows, the research pointed to the closure of F-Squared – which ran 19 model portfolios – in 2015 and saw flows into the ETFs included in its shuttered models fall by an average of 6.3% versus others within their respective categories.

Outlining this problem of endogeneity, the paper referred to a quote by Todd Rosenbluth, head of ETF and mutual fund research at CFRA: “The concern is that somebody would look at an ETF today and think there is a broader following than it has.”

Another issue surrounds investor attitudes to poorer outcomes. The research pointed to two previous studies by Ben-David et al. (2021) and Dannhauser and Pontiff (2021), which found investors tend to rush towards ETFs with strong past performance and lower fees.

However, investors become less sensitive to these factors when looking at ETFs being recommended as part of a model. For instance, flows to model-recommended products are 2% less sensitive to fees and 1% less sensitive to past performance.

“It implies that some investors may completely follow the recommendations of model providers and pay less attention to the price and quality of recommended ETFs,” the report added.

This defies logic, given unlike traditional ‘turnkey’ asset management solutions – such as funds of funds and separate management accounts – model portfolios are merely a patchwork portfolio of different ETFs, and therefore financial advisers are able to swap out products as they please. 

Unsure why many advisers accept the oven-ready model portfolios they are offered, the research suggested asset managers’ clients might follow their recommendations to reduce reputation concerns or perhaps because they are better off following sub-optimal recommendations rather than none at all.

Concluding, the research said: “The findings on the choice of affiliated ETFs and their future performance resonate with the fee structure of model providers. The fees they get from financial advisers are independent of the performance of the models.

“Moreover, when model providers include their own funds into recommendations, they get indirectly compensated through asset management fees charged by these affiliated ETFs. Hence, the model providers might have incentives to recommend funds with high expense ratios.

“We find that these recommendations have a large and significant effect on ETF flows. However, conflicts of interest seem to affect the quality of these recommendations.

“Asset managers tend to include their own ETFs. These affiliated ETFs, on average, have lower past returns and higher fees than unaffiliated funds. We also do not find evidence that the affiliated ETFs provide superior performance after they are recommended.”

Interestingly, this harmful favouritism towards affiliate funds and lower sensitivity to fund performance in model recommendations mirror a study by Pool et al. (2016), which found similar dynamics existed in mutual fund families and the 401(K) plans they provided in the US.

Related articles