Draft REIT sector weights for 400k that are hard to defend

Your table has two hidden overlaps and one category error, and fixing those matters more than the entry schedule.

The overlaps: several large industrial owners have meaningful data center or cold storage development arms, and a chunk of what gets labeled healthcare in a broad fund is medical office, which trades on physician group credit rather than on senior housing demographics. So your true data center exposure is probably higher than 25 and your true senior housing exposure lower than 15.

The category error is senior housing itself. A lot of that exposure is structured so the REIT takes the operating result of the buildings instead of a lease payment. That means you own labor costs, occupancy ramp and insurance renewals. It can be the best part of the sector when occupancy is recovering and the worst when wage growth outruns rate increases. If you're weighting it at 15 because of demographics, check how much of each name's NOI is operating versus leased before you commit.

On the schedule, stop using the calendar. Tie the tranches to something observable, like the spread between the sector's implied cap rate and where comparable private assets are trading, or a simple discount-to-consensus-NAV level. Deploy a third now, a third if the spread widens by some amount you write down today, and a third on a fixed backstop date so you can't stall forever.

Then write your sell rule in the same document. Most people who buy a valuation gap have no definition of the gap closing, so they hold through the convergence and out the other side. Mine is a stated NAV premium level and a payout ratio ceiling. Yours can be different, it just has to exist on paper before you buy.

Nothing on that list resembles typical direct market exposure, and that is more or less the argument for owning real estate this way instead.

Underwrite it the way a project gets underwritten, that instinct is fine, the numerator just needs to be the right one. For each name take market cap plus net debt plus preferred to get enterprise value, then divide forward cash NOI by that. The result is an implied cap rate that can be held against private trades for the same asset type in the same markets. Running that on a handful of industrial names typically puts the implied caps in the high 5s against private deals clearing in the low 5s. Call it 70bp. On a 5 cap that is roughly a 13 to 14 percent difference in asset value for the same square footage, before any leverage effect on the equity. That number is the convergence trade stated in a form that can be defended. The same exercise on data centers should show the premium just as plainly, which answers the second question without any hand waving. And use AFFO multiples rather than P/E anywhere names get compared, because depreciation makes REIT earnings close to meaningless.

Entry schedule depends on which mechanism you think closes the gap. If it closes because REIT share prices re-rate upward, spreading entries over six months costs you part of the move. If it closes because private marks drift down toward public pricing, your tranches buy at flat or lower levels and you give up little. The gap between public REIT pricing and private real estate pricing has been unusually wide and unusually persistent, the longest such stretch since the early 2000s, and the direction of convergence is the open question rather than the existence of the gap.

On the 25 percent data center weight, sector-level cheapness does not transfer to every name inside a sector. Check implied cap rates on the specific names against private pricing for comparable assets, and check the development pipeline, because a REIT trading at a premium is often being paid for NOI that has not been built yet.

Two constraints apply across the whole sleeve. Interest-rate sensitivity shows up twice, in the discount rate the market applies to the cash flows and in refinancing cost on the debt. Sector dispersion means the sector average tells you very little, so your weighting decision does most of the work in your result.

The cash sleeve inside the REIT allocation is a separate decision from your household cash. Holding both is defensible as long as you can name what that cash is waiting for and what event releases it. If the answer is a better price, write the price down now.

REIT distributions are generally taxed differently from qualified dividends and the details depend on your circumstances, so confirm treatment with your tax professional before sizing around after-tax income.

Worth seeing this from the private side, because that is where the other half of the gap lives. Non-traded NAV vehicles carry appraisal-based marks that move slowly, and when redemption requests pile up the sponsor can limit withdrawals. Investors who wanted out in that world have spent a long time in line. Public REITs offer the same underlying exposure in a form that can be exited on any trading day at a printed price. That is the real trade in a decision like this, and it is worth more than 50bp of tilt accuracy. Just be clear that the printed price is honest about bad news the same day, while the appraisal-based one takes three quarters to admit it.

I've been stuck on the same fork with a smaller number. What I landed on for now is a broad REIT index fund as the base and two satellite names to express the tilt, sized so a 40 percent drawdown in either one costs me under 4 percent of the sleeve. It means my data center weight ends up around 12 rather than 25, and I've accepted that as the price of not needing to read supplementals every quarter. Girder, you might actually enjoy that reading, which changes the answer. Do you want a portfolio or a research habit.

For individual REIT names, the quarterly supplemental is where the real argument gets settled. Pull the AFFO reconciliation and look at what's being added back. Straight-line rent, capitalized interest on development, and non-cash compensation can flatter the number supporting the dividend. Recurring capex per square foot should be checked against whether it looks like a real reserve for a portfolio of that age. The lease expiration schedule needs to line up against a supply thesis, since pricing power only shows up once leases actually roll. And whether management is internal or external matters, since external advisory fees scale with assets rather than returns, and that shows up in every future year.

Two numbers in this post fight each other. Calling the sector a valuation trough where the gap is likely to close, while also holding 10 percent cash inside the sleeve and staging entries over six months, cannot both be the real view. If the trough conviction is real, the cash is drag. If the cash reflects doubt about that conviction, cut the sleeve to 300k and hold the rest as cash called cash. Separately, the implied cap rate work is the piece that would move a decision here. If data centers screen at a premium to private pricing while industrial screens at a discount, a 25 and a 30 allocation are pointing in opposite directions on the same thesis.

Before you buy any name for income, pull five years of dividend history and the AFFO payout ratio. A payout above roughly 90 percent of AFFO leaves nothing for capex or a soft quarter, and names that cut in 2020 tend to have structural reasons that never went away. I also check the dividend against taxable income, since the distribution requirement is set against taxable income rather than cash flow, and those diverge. On your table, healthcare and net lease are where I'd expect coverage to be thinnest, and net lease is also the sleeve most directly priced off rates.

The thing that would break this for me is checking prices every morning. Liquidity is a tool and also a temptation.

On the structure side, read the charter before taking a large single-name position. REIT charters carry ownership limits designed around the requirement that shares be held by a minimum number of holders and that a small group not own too much, so oversized positions can trip provisions that catch people off guard. At $400k spread across six sleeves that's nowhere near a concern, so it's mostly academic here. More useful is checking whether the REIT is an UPREIT with an operating partnership, because that governs how it pays for acquisitions with units instead of cash, which matters if the accretive-acquisition part of the thesis is the point. How any of it lands on an individual return depends on facts a tax professional should look at directly.

Half the tickers that come up when screening REITs turn out to own loans rather than buildings. Worth checking whether any of the 400k is pointed at mortgage REITs, or whether the whole table is equity REITs. The yields on the loan side run much higher, and the reason isn't always obvious, sometimes it's compensation for spread risk and hedging, sometimes it's just leverage stacked on paper. Understanding the difference between book value and NAV first makes the rest of the screening easier to trust.

Seconding beacon on senior housing. The wage side of those buildings tends to move faster than any rent increase schedule that's been written for them, and it's worth underwriting that pressure directly rather than assuming rents alone carry the model.