Sound Investing For Every Stage of Life
Best-in-Class ETFs for Ultimate Buy & Hold Portfolio (Updated March, 2019)
By Chris Pedersen
See the Best-in-Class ETF Portfolio Recommendations 2019
To implement these portfolios using M1 Finance commission-free services:
For more Q & A regarding ETFs:
What’s the best set of ETFs to use to implement Paul’s Ultimate Buy and Hold Portfolio?
What’s the best set of ETFs to use if we want to implement Paul Merriman’s Ultimate Buy and Hold Portfolio as a DIY investor? That’s the question this article aims to address. The resulting recommendations are available for DIY investors at www.paulmerriman.com/etfs. They will also show up in Paul's M1 Finance and Motif Investing portfolios soon.
To start, we need to define what we mean by “best.” Obviously, keeping expenses low is a priority, but the Ultimate Buy and Hold Portfolio is based on academic research that says there’s been a higher return in small (low average market capitalization) and value (low price-to-book) equities. That higher return has come with higher volatility, but by combining several different asset classes that are at least somewhat uncorrelated, or better yet negatively correlated, a higher return per unit of risk is possible. So, we’re looking for low-cost, broadly-diversified ETFs in each asset class which collectively produce a tilt towards small and value.
Other attributes matter as well. In this year's analysis, we considered the following ETF characteristics: expense ratio, average market cap, price-to-book, number of stocks, bid-ask spread, turnover, impact on overall portfolio expected returns, yield, momentum, quality, tax efficiency, pros and cons of the underlying indexes, fund style-consistency, weighting methodology and process of maintenance including reconstitution. Here is a summary of the resulting changes.
These changes lower the average company size and price-to-book for the portfolio thereby increasing long-term expected returns and volatility slightly, but the unit of return per unit of risk stays about the same. Expense ratios increase from 0.20% to 0.28% and the number of companies held decreases from ~12,500 to ~8,000. Here are the Morningstar Style Boxes.
Individual investors should consider whether the advantages of the new recommendations justify trading and tax costs for their particular circumstances before switching.
Should I switch to the new recommendations?
This is something only you can decide. If your funds are in taxable accounts, you should consider the tax implications of selling the old funds vs. the potential benefits of the new ones. If your funds are in tax-deferred accounts, the taxes don’t matter, but your confidence in the recommendations does. If you believe the new recommendations will serve you better based on the rationale given and any additional research you do, then go ahead and switch. In the end, it’s probably more important that you have an investment strategy you believe in and can stick with than that you have exactly the right funds for that strategy.
What things do I look for when evaluating and selecting the best-in-class ETFs?
I start by searching for candidate ETFs for the 10 asset class categories that make up the Ultimate Buy and Hold portfolio. Some of them have been suggested by Paul’s listeners. Thank you for that! This year there were over 60 candidate funds. I evaluate each of them individually and as part of a portfolio using Portfolio Visualizer and Morningstar X-Ray. I look at the number of holdings, size of companies, price-to-book ratios, expenses , turnover and taxes, liquidity, weighting methodology, construction and reconstitution rules, and the historical performance of the funds. It’s an iterative process that goes back and forth between individual fund and overall portfolio optimization until I’m confident the changes are likely to help investors do better than the previous recommendations.
What’s a “factor,” and does it matter to Paul’s portfolios?
Factors in investing are attributes that academics have found which were historically significant indicators of better returns. When we say investing in the market has higher return than fixed income, or that smaller or cheaper companies have done better than large and expensive companies we’re talking about the market, size and value factors. Other factors worth considering include momentum (investments which have been going up, tend to continue to go up), profitability/quality, and low-volatility. Though the Ultimate Buy and Hold Portfolio is built to primarily get exposure to the market, small and value factors, I also evaluate every fund for exposure to these other factors as well. As Larry Swedroe points out in his book “Your Complete Guide to Factor-Based Investing,” the more broadly a portfolio is diversified across many factors, the less likely it is to underperform.
Do we really need so many funds? Doesn’t a global fund get me everything?
We only have 10 years of history for the Vanguard Total World Stock Index ETF, but here’s a comparison of the past performance of the Ultimate Buy and Hold portfolio using DFA funds. The short answer is that long-term past returns for the Total World Stock Index ETF (portfolio 1 in graph) have been good at 4.95%, but the Ultimate Buy and Hold portfolio (portfolio 2 in graph) has been even better at 5.81%.
Some of you are probably wondering how you can own the whole worldwide market and not get all of the small and value benefits. The reason is that the added returns for these factors only come when a portfolio has a bias or tilt towards them that is greater than the overall market. When you own the whole market, you don’t have a tilt to value or small. Here’s the summary view of VT showing little or no loading on size (-0.02) or value (HML: 0.05) factors.
How have the 2019 BIC ETFs performed vs. all-DFA and all-Vanguard portfolios in the past? Here’s a backtest from the Portfolio Visualizer website. I’ve substituted similar funds (VTRIX for VYMI in Vanguard and SFILX for FNDC in 2019 BIC) to get a longer history. The compound annual growth rates were 7.18% for all Vanguard, 7.46% for all DFA, 7.71% for the 2018 BIC ETFs and 8.08% for the 2019 BIC ETF Recommendations.
Backtest of Ultimate Buy and Hold Portfolios implemented with DFA, all-Vanguard (Portfolio 1), 2018 BIC ETFs (Portfolio 2) and 2019 BIC ETFs (Portfolio 3)
When is it worth it to pay more for funds with smaller and cheaper companies? Also, what are factor-predicted return premiums?
This is the most difficult question I grapple with each time I update the Best-in-Class ETF recommendations. One of the tools I use to try and answer the question is the Portfolio Visualizer fund factor regression analysis , and historical factor statistics. Using them, it’s possible to calculate expected return premiums (above risk-free investments) based on past fund and market behavior. Here’s an example showing expected return premiums for several US large-cap-value funds going back to 2010. The total expected premium including expenses is the number on the right of each stack of bars.
The highest expected return premium (7.69%) in this chart is for the Invesco S&P 500 Pure Value ETF (RPV) which has an expense ratio of 0.35%. This is one of the main reasons RPV made it into our recommendations this year. The fund with the lowest expense ratio is SCHV at 0.04%, but it has has a much lower expected return premium (6.10%). The second choice based on this analysis would have been the DFA fund, and the second choice ETF would be VTV at an expected premium of 6.98% and expense ratio of 0.05%. It’s impossible to predict with certainty which fund will do best in the future, but the factor analysis seems like a useful indicator. It’s also reassuring to see that RPV has outperformed both DFLVX and VTV on average by about 0.75% per year over the last 11 years.
Why aren’t there factor-predicted return premiums for REITs and Emerging Markets?
As far as I know, there is no historical factor performance data for REITs and Emerging Markets. Without those data, it’s impossible to calculate a factor-predicted return with any confidence.
What’s a fundamentally-weighted fund and aren’t they actively-managed?
Fundamentally-weighted funds use non-price measures of company size (e.g. sales, cash flow and dividends plus buybacks) to select and weight securities in a broad-market index. Though this implies some trading activity, the fact that the trading is done based on transparent rules and a public index means this is a passive approach. The Schwab fundamentally-weighted fund FNDC included in this year’s recommendations has historically delivered well on the value and size premiums after expenses. Though the average company size and discount can vary over time, it has historically had more exposure to small and value than VSS which is the fund it replaces. VSS has also lagged in overall return — likely due to the emerging markets part of the VSS holdings.
Are all of the BIC ETFs available for commission-free trading at Vanguard and M1 Finance?
If the BIC ETFs are available commission-free at Vanguard, why do you still have the all-Vanguard portfolio?
I’ve left it in this year for two reasons. First, there may be people who are brand loyal or want to stick with Vanguard for their low costs. Second, people may want to see the comparison. The all-Vanguard portfolio has a much smaller percentage of its holdings in the small-value corner of the Morningstar style boxes, and a much larger average company size and price-to-book which should lead to lower returns over the long-term.
Why aren’t the expected return premiums higher?
The asset classes with the highest expected premiums are some of the hardest to get cost-effectively. Here are some US Small-Cap-Value charts that help illustrate the point.
Factor-predicted return premiums vs. expense ratios for US Small-Cap Value ETFs
VBR is the cheapest fund by far with an ER of 0.07%, but it also has the largest average company size and price-to-book ratio. For a small step up in cost (to 0.15%), SLYV provides smaller companies and lower price-to-book ratios, but it’s still on the high-end of the Morningstar definition of value. If we reach further to get a true small and value, fund like PXSV, the expected returns from small and value go up, but they are offset by higher (0.39%) expenses, a substantial negative alpha for this historical timeframe and a 0.39% average bid-ask-spread due to the funds small daily trading volume. Yes, DFA’s DFSVX mutual fund has the highest predicted premium, but it’s only 0.06% per year higher than SLYV, and this analysis doesn’t include management fees. Add it all up, and I think you can see the law of diminishing returns at work. This doesn’t mean we shouldn’t try to get the small and value premiums the academics say are there, but it does suggest we should temper our expectations regarding how much of them we’ll be able to get.
The expense ratio for the portfolios is going up by a lot (0.08%) — is it worth it?
Time will tell. Historically, the small and value factors we’re seeking to get are worth more than 1% per year, so increasing exposure to them should be worth and increase of 0.08% per year to a total of 0.28% to 0.38% for the various portfolios.
When will Paul’s Motifs, M1 Pies and calculators be updated to the new ETFs?
They will all be updated by the end of March.
Porfolio Vis. Links
Paul's All Vanguard ETF's - Click to See
Comparable DFA Portfolio links- Link1 | Link2
Old Best-in-Class ETF Portfolio (Nov, 2018)- Link1 | NA
New Best-in-Class ETF Portfolio (Mar, 2019) - Link1 | Link2
All-Value BIC ETFs - Link1 | Link2
All-SCV BIC ETFs - Link1 | Link2
From Morningstar X-Ray Analysis
- Expense Ratio: The average annual expense ratio of the entire portfolio. Smaller is better.
- Price-to-Book: The average price/share divided by book-value/share of the companies. Smaller is better.
- Avg. Mkt Cap: Average market capitalization (price times shares) of companies in the portfolio. Smaller is better.
- Proj. EPS Growth 5yrs: The projected earnings growth over the next 5 years for the overall portfolio. Higher is better, but uncertainty is high.
- Yield %: Annual percentage paid out as dividends or bond interest. We’d rather have it in earnings, but higher is still better.
From Portfolio Visualizer Fama-French Factor Regression (using AQR 4-Factor including Quality & BAB factors)
- Market (Rm-Rf): Correlation to market-returns minus risk-free returns. Higher is generally better. Bonds lower the number.
- Size (SMB): Correlation to small-minus-big or small premium. Higher is generally better.
- Value: Correlation to high-minus-low or the value premium. Higher is generally better.
- Momentum: Correlation to up-minus-down or the momentum premium. Higher is generally better.
- Quality (QMJ): Correlation to quality-minus-junk premium. Higher is generally better.
- Low Volatility (BAB): Betting Against Beta factor. Higher is generally better.
- R^2: Coefficient of determination. Higher is generally better.
- CAGR based on backtesting at www.portfoliovisualizer.com with annual rebalancing and no management fees.