Playbook › Playbook
A REIT WACC has no interest tax shield, and CAPM inverts inside net-lease
Claim
A WACC built with the textbook defaults will understate a REIT's true cost of capital by 200–300bp. The agency made this error itself and carried it for four months.
Evidence — VICI, March vs August 2026
The 2026-03-20 moat report concluded VICI earned "200–300bp of genuine value creation" from an 8.2% ROIC against a ~5.5% WACC. Reconstructing that number found four independent errors, every one biased downward:
Ke = 3.97% (Feb trough Rf) + 0.687 × 4.4% = 7.0% ← 1: trough risk-free, not current
← 2: CAPM cost of equity
Kd = 4.45% × (1 − 0.21) = 3.52% ← 3: phantom REIT tax shield
← 4: embedded coupon, not marginal
WACC = 0.60 × 7.0 + 0.40 × 3.52 = 5.61%
Rebuilt correctly: WACC = 8.16%, against a marginal ROIC of 8.03%. The spread is −13bp. It was never +270bp. The entire "genuine value creation" conclusion was an artifact of the discount rate.
The two REIT-specific errors
1. There is no interest tax shield
A REIT distributes ≥90% of taxable income and pays essentially no entity-level tax — VICI's effective rate was 0.086% (FY25) and 0.355% (FY24). It still deducts interest, but the deduction shelters income that was never going to be taxed. The shield accrues to nobody.
Use the pre-tax marginal cost of debt.
Kd(after-tax) = Kd(pre-tax).
Applying a 21% haircut understated VICI's WACC by ~118bp × the debt weight ≈ 42bp on its own. Check the effective tax rate before applying any tax adjustment — this generalises to every pass-through: REITs, MLPs, BDCs.
2. CAPM inverts inside net-lease — an empirical refutation, not a preference
Three net-lease REITs, one day, 2026-08-03:
| Beta | Market-implied Ke | CAPM rank | |
|---|---|---|---|
| VICI | 0.687 | 9.85% | lowest |
| GLPI | 0.686 | ~9.6% | lowest |
| Realty Income | 0.720 | ~7.4% | highest |
Realty Income has the highest beta and the lowest cost of equity, by 245bp. CAPM predicts the exact opposite ordering. Within one industry, on one day, beta is not merely imprecise — it is inverted.
The mechanism: beta measures the volatility of the existing holder's mark in the secondary market. It says nothing about the price at which the company can issue primary equity — and a REIT that must issue equity to grow pays the primary price, not the secondary one. REIT betas are low because REIT returns are rate-driven and news-poor, not because REIT businesses are low-risk. VICI's 0.687 beta cannot see 69.8% two-tenant concentration, because those tenants are themselves low-beta.
Guard
For any REIT, MLP, BDC or other pass-through:
| Input | Do not use | Use instead |
|---|---|---|
| Cost of debt | Embedded coupon × (1 − 21%) | Marginal / new-issue coupon, pre-tax |
| Cost of equity | CAPM off beta | AFFO (or FFO/NII) yield, and the market-implied Gordon Ke = D/P + g — report both |
| Risk-free rate | A trough or an average | Today's 10yr |
| Sanity check | — | Solve P = D/(Ke − g) backwards. If it reproduces the market price, the inputs are coherent |
On VICI the backward solve reproduced $26.28 against an actual $26.27 at Ke 9.85%, g 3.0% — a one-cent tie that validated the whole input set.
Why it matters beyond one number
WACC is the hurdle. Get it wrong low and every downstream conclusion inverts: deals that destroy value look accretive, a business with no spread looks like a compounder, and a moat gets scored on a spread that is arithmetic rather than competitive. The March report did not misjudge VICI's assets, management or leases — it misjudged the discount rate, and that alone moved the verdict from WIDE 7.2/10 to NARROW 5.9/10.
Corollary worth keeping: when a moat score rests on a ROIC-minus-WACC spread, the WACC is load-bearing and must be shown, not assumed. A spread quoted without its WACC derivation is not evidence.
What would falsify this
Nothing about the tax arithmetic — it follows from REIT status. The CAPM claim is empirical and would be weakened by a period where net-lease betas do rank with observed cost of equity; that ordering should be re-checked rather than assumed permanent. The underlying reason it fails (secondary-mark volatility ≠ primary issuance price) is structural and will not go away.
Related
[[principle-currency-consistency-roe-vs-coe]] — the other way a cost-of-capital comparison silently breaks. [[principle-primary-source-beats-vendor]] — vendor ROIC for a REIT mis-handles the financing-receivable book; recompute from primaries.
History
- 2026-08-03 — found while auditing our own March VICI moat report during the
/analyzere-run. Four downward-biased errors compounding to a ~270bp understatement; the reported +200-300bp value-creation spread was actually −13bp.