Every escalation clause is a bet about inflation, placed years in advance, by two parties who usually spend more time negotiating the first year's rent than the mechanism that governs the next nine. The landlord who signs a fixed three percent bump has sold inflation protection at a fixed price. The one who signs CPI has bought it, and paid for it in predictability. Neither structure is right in general. Each is right against an inflation path, and nobody gets the path in advance.
The last few years made this concrete for everyone at once. Fixed three percent bumps written in a two percent inflation world looked generous until inflation ran hot, at which point every fixed-bump lease in the portfolio was quietly repricing itself downward in real terms, year after year, with no negotiation and no memo. CPI leases held their real value through the same stretch, and their tenants felt every point of it. The clause nobody read became the line item everybody felt.
Structure details do most of the work. CPI clauses rarely arrive naked: caps limit the landlord's upside in hot years, floors protect the downside in flat ones, and a capped CPI clause in a high-inflation run behaves almost exactly like a fixed bump set at the cap, which is to say the protection evaporates precisely when it was supposed to pay. Compounding does the rest. Small annual differences feel ignorable and are not: over a ten-year hold, one point of annual escalation spread compounds into a year-ten rent difference in the double digits and a cumulative rent difference larger than most people's intuition, which is why the widget below exists.
Run one lease through it both ways and pay attention to two outputs. The year-ten rent gap matters at exit, because the buyer of your building is capitalizing that final-year income. The cumulative gap matters during the hold, because it is cash flow you did or did not collect. A structure can lose on one and win on the other depending on where inflation lands, which is exactly why this is a modeling question rather than a preference.
And a portfolio of leases is a portfolio of these bets with different structures, caps, floors, and reset dates, layered across a rent roll. Blended-average models compress all of it into one growth assumption, which amounts to letting the spreadsheet renegotiate every lease into the same clause. PropCalc™ carries each lease's actual escalation structure through the full hold, so the building's income path is the sum of the bets that were actually signed. The widget below prices one bet. The rent roll is many.