From the desk:Public REITs (Capital Deployment)

Two index funds report data center weights of 11.8 and 4.2 percent

Building the base layer of a REIT sleeve and I've spent two days on fund documents instead of picking names.

Target weights I want to land on: industrial 25, data centers 15, healthcare with a senior housing bias 15, residential 15, net lease 10, retail 10, remainder wherever the index puts it.

Fund A's factsheet shows data centers at 11.8 percent. Fund B shows 4.2. Same broad US real estate mandate, similar size, expense ratios 12bp apart. Their industrial weights differ by 6 points as well. Specialized is 19 percent in one and 8 in the other and neither document says what's in it. Healthcare is a single line in both, which tells me nothing about how much is medical office versus senior housing operating structures.

So the choice in front of me is index plus two satellite positions to push data centers and senior housing toward my targets, or eight individual names and no fund at all. The satellite route needs positions large enough to move an 11.8 or a 4.2 up to 15, which starts to look like single-name risk dressed as a tilt. The eight-name route puts me on the hook for reading eight supplementals a quarter, which I'd do, though I'd like to know it's buying me something.

What I can't resolve is whether the two factsheets even disagree about weights or just about definitions.

Answer

1kevin_lombardPG1BPRO5

I read both funds' full holdings lists twice because your numbers looked impossible, and I don't think the funds disagree about weights at all. They disagree about who counts.

One of them treats tower companies as real estate and puts them in their own line, the other folds them into specialized, which is most of your 19 versus 8 gap. Data centers get split the same way. Two of the largest data center owners appear in one fund's data center line and in the other fund's specialized bucket. Same companies, different label, and your 11.8 versus 4.2 mostly evaporates once you re-tag them by hand.

The other thing I noticed is that one fund holds a small slice of mortgage REITs, around 3 percent, which own loans rather than buildings. That isn't in your target table anywhere.

So before choosing between satellites and eight names, I'd download both holdings files, tag every position with your own seven categories, and see what your actual exposure would be under each fund. It's tedious and it took me an evening, but the gap you're trying to close with satellite positions might be a third the size you think, which changes the sizing problem completely.

4 replies

2Rena Expert

Index sector weights are the output of classification rules, so a target tilt can only be compared to a fund once you have the fund's own definitions. Some broad real estate indexes treat tower and data center names as their own sectors, others route them into a specialized bucket that can also hold billboards, gaming assets, cold storage and self storage in a single line. Two funds reporting 11.8 and 4.2 for the same sector are often disagreeing about membership rather than about weight.

The second item to settle is whether the fund holds mortgage REITs. Those own loans rather than buildings, book value moves with spreads and hedge performance, and a few percent of that inside an equity REIT sleeve changes the rate sensitivity of the whole position.

On index plus satellites versus eight names, the tilt a satellite can express is bounded by its size. Moving a 4 percent data center exposure to 15 across an entire sleeve requires a satellite large enough that single-name risk becomes the dominant risk in the sleeve. An eight-name portfolio delivers the weights precisely and hands you the work of tracking each company's supplemental disclosures and debt maturities every quarter.

Which structure fits depends on whether the edge you're claiming sits in sector selection or in name selection. Sector selection is the claim the current setup supports most directly, since dispersion between sectors has been wider than dispersion within several of them. If sector selection is your edge, you want the cheapest instruments that give clean sector exposure, and paying for single-name research is a cost with no matching return. Healthcare is the exception worth hand-checking, because leased medical office and operating senior housing behave very differently and sit on the same index line.

2laurenkim53Passive

Two funds with the same mandate and a 7 point difference in one sector is the clearest argument for reading holdings files I've seen posted here.

4stephen_ok9CO2CPRO6

I went through this with a smaller sleeve and ended up on index plus two satellites, with a rule that no satellite exceeds 8 percent of the sleeve. That capped my achievable data center weight around 12 instead of 15, and I decided the last 3 points weren't worth carrying company-specific risk for. Juniper's re-tagging exercise is the step I skipped, and if the funds are closer than the factsheets suggest, the whole satellite question might be smaller than either of us assumed.