Trend Prophets Academy: May 14, 2024
Hi reader
Welcome to another edition of the Trend Prophets Academy Newsletter!
In this edition, we look at:
- The truth behind different exchange traded funds that promise better results than traditional index funds.
Since we know your time is valuable, we summarize each section with its key points on top so that you can get all the important information in less than 2 minutes. Those that want to learn more and see the graphs and visuals, can continue reading further down.
Are Equally Weighted ETFs Truly Superior to Market Cap Indexes?
The summary:
- The goal of this article is to share the first step of an investment analysis process used for portfolio construction.
- The article explores the common claims that equally weighted (EW) S&P 500 ETFs, like RSP, outperform market cap weighted indexes like SPY. Through a detailed quantitative analysis, I find that EW’s advantage is not as clear-cut as commonly portrayed:
- Rolling 3-Year Analysis: In 217 three-year periods, RSP outperformed SPY in just under 52% of cases, illustrating near-parity between the two.
- Rolling 5-Year Analysis: Similar results, with performance being essentially a coin toss over any 5-year period.
- Rolling 10-Year Analysis: RSP outperformed SPY 57% of the time, which is not statistically significant from a random outcome (50%).
- We can conclude that without strong statistical evidence of outperformance, investors should be cautious of claims favouring one weighting strategy over another.
Why is this important?
- The most important aspect of asset allocation and selecting investments is understanding behaviour. One cannot make conclusions by looking solely at one time period. This article provides the initial framework of how you can analyze different ETFs over rolling periods that will help you make evidenced-based decisions.
See below to read the full article.
The Big Picture
Are Equally Weighted ETFs Truly Superior to Market Cap Indexes?
It is very easy to trash market cap weighted indexes and say that you are better off with an equally weighted (“EW”) ETF. Why?
The arguments are:
- EW provides greater diversification since you aren’t overly exposed to the top 10 holdings.
- EW should outperform during market rallies when gains are broad based across sectors.
- Quarterly rebalancing of EW ETFs results in profit taking from winners and reinvestment into the losers.
- Less technology sector exposure (at least in the past few years).
- Market cap weighted indexes tends to have greater weights in companies that are likely overvalued.
This list could go on. And these are valid arguments. But when presented with the statement “equally weighted S&P 500 ETFs have performed better than a market cap weighted index”, I cringe. And then I get ready for a battle, armed with all the knowledge at my disposal because I am a quant that can do my own analyses in Python in a manner of minutes.
I worked as a registered investment advisor for several years, and we managed over $1 Billion in CAD assets. I used to have an endless flow of wholesalers from mutual fund and ETF companies marching through our door, coming to sell me the latest and greatest investment.
But when someone comes to you and says something is better, that is quite a ridiculous statement. Everyone will have a different definition of what “better” means. It’s one of the relative terms in which the frame of reference is of paramount importance. When I ask people to back up such claims, the first piece of evidence comes in the form of a graph: “Look how our Equal-Weighted ETF has outperformed over the past 10-years”.
And this leads me to the key message of this article: When evaluating investments, you should never base a conclusion on one time-period. That doesn’t tell you anything. It is a small sample from a larger population, and this is just cherry picking for information that makes you look good. And sadly, this is the norm in financial services. We blindly look at one or two periods only and base our conclusions solely on that.
In this article, I will present you with a starting methodology for how any investment should be analyzed.
If we want to get a true assessment for an investment’s behaviour, especially for comparison purposes, you need to evaluate performance over rolling periods. Why?
- Not everyone will start investing in that investment on the same day. So, you need to see what experience an investor would have no matter when they started investing.
- By looking at rolling periods, and using different rolling windows (for example, 3, 5, 10-year windows) you will get a much better understanding of when one investment did well, and how it is expected to perform during different market environments.
For example, we are focusing here on the equally weighted S&P 500 ETF, RSP, and the market cap weighted SPY. Our first step will be to look at rolling 3-year periods and calculate out of the total number of rolling 3-year windows did RSP outperform SPY. For this analysis, I retrieved monthly returns using a common inception date of April 2023, or 253 monthly returns. A rolling 3-year return analysis will look at every 3-year period, increasing the rolling window 1 month at a time. In this case, there are 217 three-year periods.
Let’s start with a visual inspection. Figure 1 shows the rolling 3-year returns of SPY and RSP.
Table 1: The percent of total periods where RSP outperformed SPY, with the best and worst 3-year return for each investment. Source: EODHD, Trend Prophets.
At first glance, we can see that their respective rolling 3-year returns don’t show that much difference. It would appear that RSP outperformed in almost every three-year period starting in 2006, up until 2016 when SPY started beating RSP in almost every 3-year period (note that earliest 3-year period would be in 2006).
Table 1 shows the summary statistics for rolling 3-year periods. We can see that RSP outperformed in just under 52% of total rolling 3-year periods. That’s hardly a convincing statistic. Irrespective of which date you started investing, it was essentially a coin toss as to which one would do better over the next three years.
Let’s look at rolling 5-year periods. Figure 2 shows the performance and Table 2 shows the summary statistics.
Again, we see that over any 5-year period, it is a coin toss.
Now let’s look at rolling 10-year periods. Of course, we will have a much more limited sample size when looking at rolling 10-year periods, even with over 20 years’ worth of monthly returns for both investments. Figure 3 and Table 3 present the line graph and summary statistics.
Table 3: The percent of total periods where RSP outperformed SPY, with the best and worst 10-year return for each investment. Source: EODHD, Trend Prophets.
Now we are seeing a 57% rate of outperformance. That seems a little more enticing. We can also see that the worst return earned by RSP was still slightly better than the worst return of SPY, and that its maximum return is higher than SPY. But I am still not convinced. Just because we see a 57% rate of outperformance, it doesn’t mean it is statistically significant.
We need to test for statistical significance. In this case, is 57% really different than 50%, which is our random coin toss? And this is why the limited sample size of rolling 10-year period matters.
Table 4 shows the output of a Python function I created that conducts a binomial test on the rolling performance data to see if it is significantly different from 50% at a 5% significance level.
Table 4: This not so glamorous table is just the output of a binomial significance test. In this case, we are testing if the 10-year percent of periods with outperformance over SPY is statistically different than 50%.
Lo and behold, it is not!
So, I am not convinced that RSP is better than SPY.
Here is another key message I want everyone to remember. And this is something I learned from the retired Director of Research from Morningstar. He is a giant of asset allocation and the brains behind the math in Morningstar Direct:
When building factor ETFs or devising a new weighting scheme for stocks based on factors or whatever, what you are saying is that you found a better way to build a mouse trap. And indexes are no different from portfolios. They should be constructed (meaning names should be weighted) with expected returns in mind. When you build an equally weighted portfolio, or minimum variance for that matter, you are removing expected returns from the construction process, and that could be a bad idea.
Market cap weighted indexes by definition have an expected return component: momentum. Meaning, the high weights of the top names are a result of their strong performance, and the momentum theory is that the top performing names should continue to outperform. So, this inherently includes an expected return for each asset.
There is nothing wrong with factor ETFs and other index construction protocols. But if you decide to include them in a portfolio, you must understand their behaviour and justify why you are including that over a market cap weighted index. Don’t be fooled by a single point-in-time statistic.
Just remember the immortal words of Homer Simpson: “Oh, people can come up with statistics to prove anything, Kent. Forty per cent of all people know that.”
Don’t wait. Subscribe today and see what we can do for you.
That’s it for this edition of the Trend Prophets newsletter! Please contact us at info@trendpophets.com for any questions.
Cordell L. Tanny, CFA, FRM, FDP
President & Founder
Join the Trend
Disclaimers: Past performance is no guarantee of future results. This newsletter should not be considered as investment advice and is intended for information purposes only. Please see our Terms and Conditions for all disclaimers.