Cap rate = NOI ÷ property value, the yield a private buyer demands on a building. Turn it around and Value = NOI ÷ cap rate. That's a perpetuity, which you've met before, with growth hidden inside the rate: cap rate ≈ required return − long-run NOI growth. Indian Grade-A offices have transacted at cap rates around ~7.5 to 8.5%, and prime US assets at 4 to 7% depending on class and era. Those are as of mid-2026, so verify them against recent deals, because cap rates move with bond yields.
Worked example: NAV and the implied cap rate
Now Horizon's NAV. Stabilized NOI of ₹3,040 cr ÷ 8% = ₹38,000 cr for the completed portfolio. Add under-construction and land at ₹2,500 cr and take off net debt of ₹11,500 cr, and you get NAV ₹29,000 cr = ₹305/unit. At a unit price of ₹215, P/NAV = 0.70, a 30% discount.
Implied cap rate: reverse the market's view (the reverse-DCF instinct). EV = market cap (215 × 95 = ₹20,425 cr) + net debt ₹11,500 cr − non-income assets ₹2,500 cr = ₹29,425 cr. Against NOI of ₹3,040 cr, that's an implied cap rate of 10.3% vs the appraiser's 8%. So the market is pricing the buildings ~23% below appraisal, and your job is to decide who's right. Are the appraisal rents stale? Or is the discount just illiquidity and rate fear? Indian REITs have in fact spent long stretches at 0.7 to 0.9× NAV. Deep discounts have historically been the entry points and premiums the caution zones, the same P/NAV logic you used on developers.
Advertisement
REIT yield vs the bond, and the LTV check
Distribution yield vs the bond. This is the retail investor's anchor, and a real one if you use it correctly. Horizon distributes ₹13.0 on ₹215, a 6.0% yield, against a 10-year G-sec at, say, 6.8%. That's a negative 80 bps spread. Is that madness? No. The bond's coupon is frozen, while Horizon's distribution should grow ~4 to 5%/yr on escalations and positive re-leasing spreads, so you're trading current yield for growth. The spread is your pricing dial. Wide positive spreads (REIT yield ≫ G-sec) signal either a bargain or a market smelling distribution risk, so check the AFFO payout and occupancy before you celebrate. Deeply negative spreads mean growth is being paid for upfront, so demand proof (contracted escalations, MTM upside) that it exists. Then there's duration. REITs are long-duration yield assets, so when the 10-year sells off 150 bps, REIT prices fall mechanically, whatever the buildings did that quarter. In the US the standard gauges are the dividend-yield spread to the 10-year Treasury and the implied-cap-rate spread (~150 to 350 bps historically).
LTV, the leverage check. LTV = gross debt ÷ gross asset value. For Horizon, 11,900 ÷ 40,500 = 29%, comfortable against SEBI's 49% cap. It's already past the 25% rating/approval tripwire (rated, approved). Headroom matters for growth too. 49% × 40,500 = ₹19,845 cr of permitted debt, which means ~₹7,900 cr of acquisition capacity without issuing units. But treat 49% as a ceiling to stay well clear of. At 45%+, a modest markdown in asset values breaches the cap and forces equity issuance at the worst possible price. A healthy operating zone is <30 to 35%.
Which multiples do REITs trade on?
The two multiples the sector actually trades on divide the unit price by the two cash measures you just built. On Horizon, FFO per unit = 1,760 ÷ 95 = ₹18.53 and AFFO per unit = 1,420 ÷ 95 = ₹14.95. So at ₹215 the units trade at 11.6× FFO and 14.4× AFFO. Flip them over and you get an FFO yield of 8.6% and an AFFO yield of 7.0%, against the 6.0% distribution yield. The gap between the AFFO yield and the distribution yield is the retained cushion.
Here it's thin, and that's the same 87% payout read a second way.
Prefer P/AFFO to P/FFO whenever the REIT discloses enough to build it. FFO adds back all depreciation and subtracts none of the money that keeps the rent flowing, so it's the flattering one. On Horizon, the 19% distance between the two multiples is the maintenance capex, tenant improvements, leasing commissions and straight-line rent that FFO ignores.
Can you compare a REIT multiple with a company's P/E?
Now for the comparison nobody should make. A REIT on 11.6× FFO isn't cheaper than an operating company on 20× earnings because of that lower number. Those two numerators are different kinds of quantity. For a manufacturer, depreciation is a real economic cost. Its machine tools wear out and must be replaced at rising prices, so adding depreciation back would overstate what the owner can take out. For a well-maintained property portfolio, the building's accounting depreciation is largely an artefact. Land doesn't depreciate at all, and Grade-A space in a supply-constrained micro-market often appreciates.
That's the reason the sector invented FFO in the first place.
Capital structures differ too. REITs run 25% to 50% leverage by design and distribute nearly everything they earn. So their equity is a thinner, more levered claim on a more predictable stream than an operating company's. Put a REIT multiple next to an operating company's and you're comparing a levered post-maintenance cash yield with a pre-maintenance accounting profit. Compare REITs with REITs on P/AFFO, compare either with a bond on yield, and cross-check both against NAV.
Where private-market valuation and taxes fit in
Two neighbouring topics pick this up. The private-market side belongs to alternatives and due diligence: the three approaches to valuing a property directly (income, sales comparison and cost) and the property-level discounted cash flow under a cap rate. Use those when the asset is unlisted and there's no unit price to invert. What the distribution does in the holder's hands belongs with products, costs and taxes. There you break an Indian REIT payout into its interest, dividend and repayment-of-capital components, each with its own tax treatment, and a 6.0% headline yield becomes a very different post-tax number depending on who owns it.