Evidently, six parts wasn't enough in my series on REITs! A question remains - individual REITs appear to be better than a utilities index fund, but that begs the question: are individual utilities better than REITs? What about small-cap value? Here I put three scenarios to the test.
Facts are indeed stubborn things, as John Adams says. When we err with how we present those facts, though, we shouldn’t be stubborn in correcting our mistakes. Alas, that’s me with a bit of egg on my face after I mislabeled two portfolios in my comparison of individual REITs and VPU, the utilities index fund. I wish it had just been a cosmetic mistake that I could have fixed in a few minutes, but it went way deeper than that. You see, I had based the whole conclusion of that post on my error. Yikes. Long story short: that piece required a through revision. The new version is at the link above, and there is a link to the original text at the end.
Furthermore, the mistake also has meant that my summary opinion on REITs in the post “Verdict on REITs” needs to be re-thought. I undersold the upside of individual REITs along the way, so now that summary is off too. I have taken down the post (which is why it isn’t linked here), and will now start re-writing it with the hope of getting it out soon.
This whole process inspired some other questions, though: if individual REITs are better than REIT index funds like VNQ, and also better than utilities index funds like VPU, what about an apples-to-apples comparison of selected utilities stocks to selected REITs? What about a comparison with ten small-cap value stocks, since we've already seen that you can get better results from them compared to REITs? If the move from index funds to the ten individual REITs with lower correlation numbers to stocks made a difference, might a similar process work with utilities and small-cap value?
Consider this a postscript to the six previous posts on REITs. In this one, I’ll choose ten individual utility stocks, and then ten small-cap value stocks from any sector, both following a similar process that I used for selecting individual REITs. Using a generic RP portfolio as a kind of chassis, I’ll then compare the basket of individual REITs, the basket of utilities and the basket of small-cap value to see how they fare in terms of compound annual growth rate, standard deviation, and Sharpe ratio.
Selecting Individual Utilities and Small-cap Value Stocks
I’ll look for ten of each type, and invest equal amounts in each. To find these ten, I’ll follow the same process to what I used to choose individual REITs:
- To find utilities and small-cap value stocks that are less correlated with the market, I’ll focus on those with a beta (or correlation to the stock market) the same as I set for REITs, which was .58. This number was the maximum I set when choosing the REITs, based on the beta for VPU.
- For utilities, I’ll follow the same process as REITs and take the ten largest by market capitalization with a market beta of .58 or lower.
- For small-cap value, I’ll set a maximum market cap of $6.54 Billion and a maximum P/E ratio of 12.65. Those figures are the averages for VBR, the most popular ETF of this type.
- I did run into a small problem once I started working with the small-cap value stocks. As I discovered later when I ran the correlation matrix for my original list, many had just started in the past year or two. To make the correlation matrix and the later backtests more reliable, I excluded stocks younger than the youngest REIT, which was February, 2018.
These are pretty simple screens, for sure. If anyone just points out that I didn’t choose the ten “best” REITs, utilities, or small-cap value, then point taken. Ideally, I’d like to just pick the ones that were going to go up in the future (the Will Rogers approach to investing), but that is a game with no end, so my simple criteria will have to do.For utilities, here are the ten to make my list (from Barchart):
It turns out that the restriction on beta made little difference, as this is basically the list of the top ten utilities by market cap. WEC is actually eleventh, as only one of the original top 10 had a beta higher than .58: Sempra (SRE) at .67. This list is all electric utility companies, though my search was not limited to just those. I would have included a water or gas utility company had they met the criteria.
As for dividends, since they aren’t a big concern of mine, I didn’t prioritize them in my search for utilities but did list them since they are part of the total return. The yield for an equally-weighted portfolio of these ten came to 3.51%, higher than what I found with my basket of ten REITs (2.61%). It is also more than VPU’s dividend yield of 2.69%. In the backtests, these dividends will be included through re-investment.For small-cap value, here are the ten I’ll use (from Barchart):
This is quite the motley crew: insurance, healthcare, fin tech, and a grocery store chain. With all the parameters I had to put on the search: beta, market cap, P/E, age of the equity, it was quite the process to come up with the ten. It reminded me again why I have given up trying to pick stocks and prefer index investing instead. The ten here may well be crappy stocks, and I accept the critique that I just didn’t pick the right ones for the exercise. As I wrote in an earlier post, I’m a big fan of the Will Rogers approach to investing, I just can never manage to pull it off. By the way, dividends for this group were paltry, just 1.13%.
As for the basket of individual REITs, you can read all about that portfolio in this post, but briefly, the ten REITs are: AMT, CCI, EQIX, PSA, SBAC, DLR, EXR, AMH, CUBE, and COLD.
The Test Portfolios and the Comparisons
To explore how they might do, I’ll again use a modified version of a Risk Parity portfolio with a large allocation do the particular type of individual stocks so that we can really see their impact:
- 40% Total Stock Market (VT)
- 20% Long-term Treasuries (TLT)
- 10% Commodities (PDBC)
- 30% in the basket of REITs, utilities and small-cap value, so 3% each to the ten companies in the respective baskets.
I’ll throw in the benchmark of a traditional 60/40 as well for good measure.
As for correlations:
- Here is the correlation matrix for these ten utilities compared with VT, TLT, PDBC, and VPU.
- For small-cap value, I substituted VPU with VBR to represent the index fund version of this asset class. Here is the correlation of the ten small-cap value stocks compared with those reference funds.
- For reference, here is the correlation matrix I did earlier with the REITS.
- All correlation matrices go back to February, 2018, since that is the date of the youngest REIT in the portfolio, COLD.
With a maximum beta for all the funds of .58, it's no surprise that correlations are rather low in all three matrices. With so many funds in play, it is hard to synthesize the data, other than to say we’ll see how it plays out in the backtests, when lower correlation should lead to lower portfolio volatility.
The first backtest we’ll do is the one I started with in the look at individual REITs vs. VNQ and VPU, which goes back to February 2018. Click here to see the link to the full backtest.
- The best portfolio of the bunch, for this time period, is the basket of small-cap value stocks. Its CAGR exceeded the REIT portfolio by .96% and the utilities portfolio by 1.55%. It had the lowest volatility of these three contestants, and consequently, the highest Sharpe ratio.
- The basket of REITs was better than the basket of utilities over the past four years or so in terms of CAGR, surpassing it by .59%. The volatility was slightly higher than that with utilities, though, resulting in nearly identical Sharpe ratios.
- All three variations of the RP portfolio trounced the traditional 60/40, so score that one for Risk Parity, writ large.
In the earlier post about REITs, I did a second backtest, dripping COLD and allocating its 3% allocation evenly between AMT, CCI, and EQIX so that I could pull the backtest back to January, 2015. I tried this again with the new baskets of individual equities, though here I did run into a small issue: other stocks then constrained the data, so if I wanted to get it back to the same timeframe I did earlier, I would have had to substantially change the small-cap value basket since two stocks in that would have dropped off. In the end, I got it back to September of 2015 without having to do any other discretionary alterations, so that’s what I present here (with the link to the backtest if you’re interested):
Key points for the backtest to 2015:
- Once again, the RP portfolio with the basket of small-cap value had the best performance. Its CAGR was .6% over the one with REITs and 1.58% over the one with utilities.
- The utilities portfolio did have the lowest volatility of the three, just a hair lower than small-cap value, but due to its lower growth, fell short of small-cap value in terms of Sharpe ratio.
- The RP portfolio with REITs beat the one with utilities again, and this time by an even larger margin: 98 basis points. It was slightly more volatile, though, resulting in very close Sharpe ratios between the two versions.
- I’m happy to see that all three once again beat the traditional 60/40. I know the “test” here is of the three variations of the RP portfolio, but it's good to take a step back and see that they consistently outshine the old way of doing things, whichever one you choose. So a well deserved cheer for risk parity - no matter how much I mess with the portfolios, they seemingly always beat the 60/40.
Stock Picking Questions
As I tested these individual stocks, a question kept popping up in my head: is this all just a function of the stocks I chose for each basket? I could have made the maximum correlation lower - would that have meant taking out two strong performers and adding in two weaker performers? I could have changed the criteria for what counted as small-cap, and what counted as value - would I have found ten that would have underperformed the REITs and the utilities? I did my best to use reasonable search criteria, but at the end of the day, every stock picking method involves choices, some which will work out and some which won’t.
To set a baseline, I wanted to see if the performance I found testing individual stocks would have been replicated by just choosing an index fund to represent those asset classes. I then ran two backtests, with the same time frames as those above.
The first is from February, 2018 (see the backtest here):
This was flipped from the test of individual equities, which went SCV, REITs, utilities. This suggests I did a great job picking individual equities for Small-cap Value but a terrible job picking them for utilities, as I beat the trend in both directions.
The second is from September, 2015 (see the complete backtest here):
In the test of individual equities, it went SCV, REITs, then utilities. Once again, I seem to have done a terrible job picking utilities stocks on their own, as I parlayed a first overall finish in the indexed version into third place when I made active choices.
First lesson: these backtests really depend on the individual equities you choose. Second lesson: time frames matter too. Sometimes a particular asset is going great, sometimes now. Third lesson (for me, anyway): don’t pick individual equities!
We see from this quick look that there are mixed results for individual REITs. The inquiry began with a curiosity: indexed utilities beat indexed REITs, but individual REITs beat indexed utilities, so would individual utilities then beat individual REITs? Although the margin wasn’t huge, the answer is no. The RP portfolio with ten REITs beat its counterpart with utilities in the two backtests. The particular backtests are somewhat of a limiting factor for our conclusion, though, as one test was just four years and the other just seven. I’m sure we could have found other backtest periods where individual REITs would have beaten the two others, but just as likely could have found times when they would have trailed both.
The larger point may be that once we opened the inquiry to individual equities, we had to let small-cap value into the mix, as even a partial allocation to indexed small-cap value proved as good or even better than an allocation to indexed REITs. If we’re going to depart from index investing, then the world of possibilities opens up quite a bit and we might as well base our selections on risk factors that we understand through extensive academic research, especially this one that led to a Nobel Prize. As I summarized in my post about the research on REITs, the case behind REITs is on shakier foundations, if you’ll excuse the pun:
In all, definitely some good news for individual REITs in beating utilities, but they take the L when faced with small-cap value. I suspect what might be the better strategy all around would be to look for low-beta stocks regardless of what sector they may be in. The individual REITs that outperformed the REIT index fund were notable in having a much lower beta than the average holding within VNQ, and going forward, that lower beta would be the thing to focus on, not that they are REITs.
One last thought: working through all of these individual stock choices really reinforced just how difficult it is to deal with equities one by one. The backtests and correlation matrices tend to be shorter. At least with ETFs, if you find yourself constrained by one, you can often use a similar fund (even in mutual fund form) to give you a rough proxy. I’ll be returning to a reliance on ETFs, post haste.
Just for giggles, I ran another backtest, this time using that same basic RP skeleton, with the 30% devoted towards VNQ for REITs, VPU for utilities, or VBR for small-cap value, just to see how a portfolio on my preferred ground of passive ETFs would fare with a much longer time frame. With a few changes (DBC for PDBC, VTI for VT, and VBR in place of VIOV as above), I managed to lengthen the backtest back to 2006. All were pretty close, with the utilities one the best. Once again, all three beat the 60/40 in terms of return, but the 60/40 did finish second-best in Sharpe ratio, so it has got that going for it, which is nice.