The case behind having REITs as a distinct asset class in a Risk Parity portfolio is not looking too good so far. But what should one do instead? Here is a look at increasing allocations in other types of equities instead of allocating to REITs. Complete with backtests!
If a RP investor were to decide to move away from REITs, what can they invest in instead? Later, we’ll look at ways to invest in real estate directly, as well as a subclass of equities that looks like it will actually accomplish what REITs are supposed to, but for now, we’ll take the simpler route: using the allocation to REITs to invest in other types of equities instead.
Using Portfolio Charts, I offer two backtested portfolios below to explore alternatives to REITs:
- The first option is just take out REITs, and redistribute their allocation to other asset classes. This is pretty straight-forward, we’ll see what a difference the REITs make anyway. My hunch is that there won’t be much change, but if there is, the resulting portfolio will be a little less lucrative but a little less risky, since you’re exchanging dedicated equities for a more risk-balanced portfolio.
- The second option will be to replace the REIT allocation with Small-Cap Value and Intermediate-term bonds. Recall Kizer and Gordon’s finding that you could replicate the performance of REITs by dedicating 66% of it to small-cap value and 34% to corporate bonds instead. Actually, this wound up improving the risk/reward metrics and became a real source of doubt for the question of whether REITs are actually doing much in a portfolio. I have slightly changed Kizer and Gordon’s prescription from a corporate bond allocation to intermediate-term Treasuries for a few reasons: 1) my dislike of corporate bonds (in short, they are neither fish nor fowl); 2) the fact that I already have an allocation to IT Treasuries so it’s easy just to increase those; and 3) Portfolio Charts doesn’t have data on corporates.
The RPC Income is the only test portfolio of mine with an allocation to unlevered REITs, so I’ll be using that as the baseline. Note: the RPC Income uses a few 3X Leveraged ETFs such as UPRO, so its capital allocation is offered followed by its equivalent allocation in parentheses. The use of leverage is accounted for by a negative allocation to cash (“BIL”). Here is the asset allocation table for the portfolios:
To run the backtests, I followed the same guidelines I used when doing the original portfolio backtests. As usual, thanks to Portfolio Charts for making this possible.
The biggest takeaway would be that the portfolios are all pretty similar, and for all the hullabaloo about REITs, it doesn’t seem like much is gained or lost by including them. The RPC Income originally has an allocation to REITs of 15%, which is not substantial but it’s also not nothing, and 15% seems to be in the range of what most portfolios have when they have REITs.
By eliminating them completely and reallocating to existing asset classes, you see a slight dip in expected return and slightly worse risk metrics. But I suspect this has a lot to do with the increased expense ratio from having more leveraged funds in it. Before I adjusted the expense ratio up to .74% (it was .14%, but then I add basis points commensurate by the leveraged position, so plus 60 bps), it out performed the original RPC Income portfolio. Most notably, you see a dip in the projected perpetual withdrawal rate, so in this case we can say that there is an oh so slight benefit to having REITs in a portfolio.
But by the same token, comparing the original RPC Income portfolio with one that substitutes in Small-cap Value stocks and intermediate-term bonds points towards a similar benefit from not having REITs, at least if you can replace them in this way. The portfolio that used Kizer and Gordon’s advice resulted in a slightly lower expected return, but better risk metrics and, importantly, an increase in the projected perpetual withdrawal rate from 5.8 to 5.9%.
I have left out one key element of the REIT story, though, and that has been the dividends they pay out. This is the RPC Income portfolio, after all, so the 15% allocation to VNQ was done partially because of VNQ’s higher dividend rate. This means more of the monthly withdrawals are covered by those dividends, thereby altering the withdrawal rates.
These backtests were done without consideration to dividend rates, but since they are so close even without considering them, this would be a plus on the REIT side of the ledger, at least for investors who seek dividends. I do not, though, so for me it might not be what I would do, but it is fair to say that this is a benefit to including REITs in a portfolio. This is also the kind of question that could be answered more accurately with more data that was net of dividends. I guess I could do that by backtracking with real data through my monthly portfolios, but that’s a project for another day.
Like most investing decisions, it really depends on your particular goals and situation. Unlike the earlier two posts, where I assessed REITs and then where I assembled expert opinions on REITs, these backtests do paint a slightly more positive (if still not definitive) picture for REITs. Still want to explore a few more avenues before making up my mind.