Which Real Estate Sector Actually Pays the Most on Your Asset?
Put a dollar into a piece of real estate. Hold it steady-state. Maintain it so it keeps producing. Set aside appreciation and leverage. How many cents does it pay you back each year?
That number is called Free Cash Flow Yield: NOI minus recurring capex, divided by property value. Unlevered. It's different from REIT total return, which bundles share price moves, cap rate compression, and dividends into a single number. What the building actually produces gets lost in the mix.
NCREIF has tracked FCFY across more than 420,000 quarterly property observations since 1978. Here's the 37-year ranking for the classic five sectors (See Chart Above)
Retail first. Then industrial, apartment, office, and hotel. If you ask commercial operators to rank these from memory, you tend to get a different order. The data says otherwise.
The hidden cost inside the chart
Look at the capex column. Retail runs 1.29% of property value in recurring capex each year. Apartments 1.15%. Industrial 1.43%. Those are the sectors on top.
Office runs 2.36%. Hotels run 3.12%.
Put those capex numbers against each sector's NOI and you see the real drag. In the NCREIF data, recurring capex consumes 39% of a hotel's NOI before it reaches the investor. Office surrenders 32%. Retail and apartments give up 16%. Industrial, 19%.
The sectors with the highest unlevered yield on the asset are the ones that need the least capital to keep producing.
Everybody knows their cap rate. Fewer know their capex ratio. Almost nobody underwrites the 10-year reserve accurately until the roof goes.
How capex actually shows up
The NCREIF ratios capture averages. What the averages hide is that capex looks very different sector by sector in how it hits an owner's cash flow.
Storage runs low because the physical asset is simple. Concrete, steel, roll-up doors, and a gate. The building can go 30 or 40 years between major capital events. Ongoing spending is modest.
Manufactured housing is lower still. The operator owns the land, the utility infrastructure, and the roads. The resident owns the home on top. Everything that breaks inside the home is the resident's problem. The operator maintains infrastructure, not buildings.
Apartments require continuous reinvestment. Appliances rotate through replacement cycles. HVAC compressors fail. Roofs need sections replaced. Unit turns eat money every time a tenant leaves. Lenders typically underwrite $300 per unit per year in reserves. Actual stabilized requirements often run $500 to $800. Older assets with deferred maintenance can demand $1,000 to $2,000 per unit per year. The gap between underwritten reserves and real requirements is one of the most common ways apartment deals underperform their pro formas. You don't get to defer. You only get to underfund until something fails and a tenant leaves.
Industrial spending is episodic rather than continuous. Building envelopes last a long time, but tenant-specific improvements on rollover can be large, especially for specialized users. A logistics tenant needs very different buildouts than a light manufacturer.
Retail turns on tenant mix. A well-leased center with long remaining term runs modest capex. A center facing lease rollover, cotenancy issues, or anchor turnover can need heavy repositioning capital.
Office and hotels carry continuous spending plus calendar-driven large events. Office needs fresh tenant improvements and leasing commissions every time a lease rolls. Hotels require brand-mandated full renovations every five to seven years or the brand pulls the flag. These events come on top of ongoing maintenance, not in place of it.
The point isn't that any sector can run on fumes. The point is that the annual bill for keeping the asset productive varies widely, and it adds up over a hold period.
What about storage and manufactured housing?
Neither is in NCREIF's classic five-sector index. Both were added to the expanded index more recently, but the 37-year apples-to-apples history only covers the original five.
The operating evidence is consistent. PGIM puts manufactured housing recurring capex at 7 to 10% of NOI. PREA and Hines put storage at 8 to 12%. Both below every sector in the classic five.
The economics explain why. A single-story self-storage facility costs $50 to $65 per square foot to build and rents for $18 to $25 per square foot. A manufactured housing community works differently: the operator owns the land and the utility infrastructure, and the resident owns the home. Moving a manufactured home costs $5,000 to $15,000, if you can find someone willing to do it. So residents stay. Or they leave the home behind, and the operator inherits the improvement on top of land the operator already owned.
Public Storage's 2024 10-K: self-storage NOI of $3.26 billion, maintenance capex of $234 million, gross real estate at cost of $28.5 billion. Operating yield on cost: 10.6%. Higher than any classic-five sector. On market value, where storage has seen heavy cap rate compression, FCFY likely sits in the 7 to 8% range. Still higher than retail at 6.67%.
What this ranking doesn't answer
The FCFY ranking doesn't project rent growth. Apartments may outrun retail on growth over the next decade even while producing less cash today. It doesn't predict cap rates either — storage compressed from 8% to 5% between 2012 and 2022, inflating total returns for anyone holding, and an expansion gives that back. Operator skill sits outside this number entirely and can swamp sector economics on any given deal.
What it answers is the base question: all else equal, which sector pays the most cash per dollar of physical asset?
One variable at a time
This post isolates one variable. It's not portfolio advice.
The question I'm asking is what's the structural ROA of each real estate business model. Not what happens if you pay a 4 cap for a storage facility today, where compressed cap rates have already priced in the sector's structural advantages. Paying a 4 cap for any asset means the yield you earn is 4, regardless of whether the sector averaged 7% over 37 years. The structural ranking tells you which business models are inherently more productive per dollar of replacement cost. It doesn't tell you what your entry price is buying.
A real analysis of any deal requires working through several more variables. Rent growth. Cap rate direction. Leverage, rate, and terms. Operator capability. Market supply dynamics. Tax treatment. The cost basis you're entering at relative to replacement cost. Each deserves its own analysis. Combining them into a single "which sector is best" answer is how people confuse themselves.
What this piece does is lock every variable except one and ask the cleanest possible question. One dollar invested in the physical asset. One year of steady-state operations. How much cash came back?
On that question, the 37 years of NCREIF data say: retail, industrial, and apartments lead the classic five. Office and hotels trail. Storage and manufactured housing probably sit above the whole list based on available operating data.
A piece of the picture worth having. The rest is a different analysis.