Cap Rate Cycles Don't Move in Straight Lines
The relationship between cap rates and interest rates is real, but it's not mechanical. Plenty of analysts learned this during 2021-2022, when the Fed began raising the federal funds rate in March 2022 but industrial cap rates continued compressing well into mid-year — deal velocity and institutional demand pressure kept pricing tight even as borrowing costs rose. Then, as rate hikes accelerated into late 2022, cap rates moved — but unevenly, and at different speeds across asset classes.
Understanding cap rate compression and expansion cycles means understanding what drives each half of the equation: income growth expectations and required return expectations. These don't always move together, and the lag between them is where acquisition pricing errors tend to cluster.
The Mechanics of Compression: What's Actually Happening
Cap rate compression — where cap rates fall and therefore asset prices rise — is most cleanly understood as an increase in the price investors are willing to pay per dollar of current NOI. That willingness has two drivers: expected income growth (if rents are projected to rise sharply, buyers will pay more today) and required return compression (if alternative investments offer lower yields, real estate looks more attractive on a relative basis).
The 2017-2022 industrial compression cycle illustrates both drivers operating simultaneously. E-commerce penetration created a structural demand story that justified aggressive rent growth assumptions — Inland Empire industrial rents increased roughly 80-100% from 2018 to 2023 on a nominal basis, though the pace was uneven across submarkets and lease roll timing. Simultaneously, the sustained low-rate environment compressed required returns across the capital stack, pushing institutional capital into industrial at cap rates that would have been unthinkable in prior cycles.
The practical result: going-in cap rates in top industrial markets fell to the 3.5-4.5% range by late 2021. For an acquisition underwritten at 4.0% going-in with an exit cap of 4.5% (a conservative 50 bps spread), the model still worked if the rent growth story held. For acquisitions that assumed exit caps stayed flat or compressed further, the 2022-2023 cap rate reset was painful.
Going-In, Reversion, and the Exit Cap Assumption Problem
Every acquisition model has two cap rate inputs that matter most: the going-in cap (what you're buying at today) and the exit or reversion cap (what you expect to sell at). The spread between them — the exit cap padding — is where disciplined underwriters earn their keep.
The standard practice in most institutional acquisition models is to underwrite the exit cap at 50-100 bps over the going-in cap. The logic is simple: you're buying at peak pricing, you should assume some mean-reversion at exit. But this convention can be applied mechanically without examining whether the spread makes sense in context. A 50 bps exit cap pad on an industrial acquisition at a 4.0% going-in cap — exiting at 4.5% — implies you still believe the exit buyer will price at a premium to most historical industrial cap rate ranges. In a rising-rate, oversupply environment, that's not a conservative assumption.
The reversion cap assumption also has to be tied to your hold period assumption. A 5-year hold model with a 50 bps cap rate expansion is different from a 10-year hold model with the same expansion — because the income growth from 10 years of lease rollovers at higher market rent changes the NOI base that gets capitalized at exit.
Terminal value sensitivity: where the value sits
In a typical 5-year hold, the terminal value (the resale proceeds) often represents 65-75% of the total projected return. That means a 50 bps error in the exit cap assumption has roughly 3-4x the impact on project returns as the same percentage error in Year 1 NOI. This is not a new observation — anyone who has built a CRE waterfall model understands terminal value sensitivity — but it is underweighted in how acquisition presentations are often stress-tested. Sensitivity tables that show ±100 bps on the exit cap are essential. Models that only show base case vs. downside on rent growth, without touching the cap rate, are not showing the right sensitivity.
How Compression and Expansion Differ Across Asset Classes
Cap rate dynamics in multifamily, industrial, and office don't follow the same cycle, even when they're moving in the same direction. The underlying demand drivers, tenant credit profiles, and lease structures create different absorption speeds and different floors.
Multifamily: Historically the tightest cap rate compression cycle because of a combination of institutional capital depth and long-term structural undersupply narratives in gateway markets. Multifamily cap rates in Sunbelt markets compressed from the mid-5s in 2018 to sub-4s in some corridors by 2022. The 2023-2024 reset brought cap rates back toward the high-4s to mid-5s range in many secondary markets, driven by new supply delivery overhang rather than pure rate movement. The key lesson: multifamily compression cycles are both rate-driven and supply-cycle-driven, and the two can offset or amplify each other.
Industrial: The fastest compression in modern CRE history, followed by the fastest rate of exit-cap-assumption obsolescence. The supply pipeline built on 2021-2022 economics — particularly big-box distribution centers above 500,000 SF in saturated logistics corridors — delivered into a period of absorption deceleration. Markets that saw net absorption near historical highs in 2021-2022 posted negative or near-zero net absorption in parts of 2023-2024. Cap rate expansion in some industrial markets has been sharper than initial models assumed.
Office: The structural story overrides the rate cycle for most investors. Even in a rate-cut environment that theoretically supports compression, trophy-class office cap rates in most gateway markets are pricing in long-term occupancy risk rather than moving in line with the broader cap rate environment. The effective floor for Class B suburban office cap rates has moved materially — where 7-8% would have been aggressive pre-pandemic, 8-10% is now a reasonable market range depending on leased percentage and WALT.
Market Cap Rate vs. In-Place Cap Rate: The Valuation Trap
We're not saying in-place cap rates are misleading — they reflect actual contracted income. What we're saying is that an acquisition team that only looks at the in-place cap rate without computing the market cap rate on the same asset is not doing complete analysis.
The market cap rate — NOI that would be generated if all space were leased at current market rent, minus market-level operating expenses — tells you what the building is worth if fully leased at today's rents. The spread between in-place and market cap rate signals whether you're buying below or above stabilized value. A building with in-place cap rate of 5.5% but market cap rate of 6.5% is telling you that in-place rents are significantly below market — either a value-add opportunity or a sign that below-market leases will roll in at current-market economics, suppressing near-term income.
This in-place vs. market spread analysis is particularly relevant for assets coming out of long-term NNN leases signed during compression cycles. A retail property locked into 10-year NNN leases at 2018 rents is generating above-market or below-market income depending on how rents have moved since. Either scenario has underwriting implications that go beyond the simple current-income cap rate.
Positioning Assumptions for the Phase of the Cycle
Acquisitions analysts working in a compression environment should underwrite exit caps conservatively — pad 75-100 bps over going-in and run a sensitivity that shows an additional 100 bps expansion. In an expansion environment, the discipline runs the opposite direction: be careful not to over-pad the exit cap in a way that understates project returns when the market stabilizes.
The most common error in either phase is using comparable sales from the prior phase to set assumptions for the current phase. Comps from peak compression don't set the exit cap for a deal underwritten today — they represent a data point from a different rate and credit environment. Adjusting comp data for current market conditions, rather than using it as a mechanical benchmark, is the actual analytical work.
Cap rate assumptions aren't the place to anchor on a single comparable sale and call it done. The acquisition model is a probability distribution with a point estimate — and the cap rate sensitivity table is the most important part of making that visible to the decision-maker reviewing the deal memo.
