REIT sector total return spreadsheet: the columns, the coverage, and where office actually lives

A REIT sector total return spreadsheet sounds like one artifact. It is four decisions wearing a single filename: the row grain (month-end or year-end), the return type (total, price or income), the sector taxonomy (investment-sector aggregates versus property sectors), and whether an index level rides beside the returns. Get one of the four wrong and the sheet still opens - it just answers a different question than the one you asked.

In the Datadory catalog the phrase resolves to a compact family of performance records rather than one file: a monthly panel running 665 month-ends deep, an annual record carrying the property-sector cuts, a sixteen-product report family that ties the frequencies together, and the appraisal-based private-market complement. Format-wise the spreadsheet is rarer than its reputation - of the 1,744 datasets Datadory catalogs, only 97 ship legacy XLS and 410 ship XLSX, against 734 offering CSV - which is exactly why the columns matter more than the container. What survives into production is a feed delivered daily, weekly, or hourly, not a workbook nobody can schedule.

The most common mistake is structural, not technical: assuming the flagship monthly history contains an office column. It does not. Its 42 columns organize into six investment-sector blocks - All REITs, Composite, Real Estate 50TM, All Equity REITs, Equity REITs and Mortgage REITs - and every block carries the same four sub-columns: Total Return, Price Return, Income Return and Index. Those are industry aggregates. Nothing in that panel is named Office.

Office enters through a different door: the Nareit Annual Index Values & Returns (1972-2025) record, whose Property Sector and Property Subsector sheets pair Total and Price return columns for fourteen sectors from 1994 onward - Office sitting beside Industrial, Retail, Residential, Diversified, Health Care, Lodging/Resorts, Self Storage, Timberland, Telecommunications, Data Centers, Gaming, Specialty and Mortgage, with subsectors down to Apartments, Regional Malls and Shopping Centers.

For monthly granularity, the property-sector series inside the Nareit / Nareit.com Corporate Site report family breaks returns out to subsector - which is where office components sit, at index-level grain rather than per-company resolution. One wrinkle worth planning around: there is no dedicated office product among the sixteen regularly issued reports. If a workflow assumes an office tab, the assumption fails before any analysis starts.

Depth is the quiet argument for the whole family. The Nareit Monthly Index Values & Returns - Complete History XLS (1972-2026) record runs from a December 1971 base - All REITs indexed to 100 - through 665 consecutive month-ends ending July 2026, with no gaps to fill. Measured returns begin January 1972: All REITs returned 1.2203525% that month, 0.3268945 from price and 0.893458 from income. At the modern end, the Composite index closed at 11,960.151 on July 31, 2026. Those two figures are the cheapest sanity check in the industry - any extract that misses either anchor is wrong somewhere upstream.

The annual record stretches the same series across calendar years: investment sectors 1972 through 2025, property sectors and subsectors 1994 through 2025. Office's first print is instructive - +2.86% total return in 1994, in a year Retail returned 18.66% - followed by +38.79% in 1995. The early-1990s office downturn reached the index tape late and left violently.

Private-side history runs deeper still. The NCREIF Property Index Office Returns record starts the NPI at Q4 1977 and the NFI-ODCE fund index at Q1 1978: composite total, income and appreciation returns by property type, subtype and geography down to Zip Code, one row per index per quarter.

Left alone, the family is three shapes: a wide monthly matrix, two narrow annual sheets, a live sector page. Delivered, they become one schema. Datadory packages the records as typed rectangular rows keyed on period and taxonomy - percent-typed return columns rather than display strings, index levels carried beside them, one row per period per block - so a sheet built on the feed appends next month instead of being rebuilt.

That combination is the point of a sector return series done properly: index history for the curve, the live tape to reconcile it against, and the operating layer explaining why the curve moved.

Every block in the monthly panel splits Total Return, Price Return and Income Return plus the Index level, and the split is not decorative. Price return omits dividends - fatal for REITs, where income is the thesis. January 1972 already made the case: 0.893458 of that month's 1.2203525 total came from income, roughly seventy-three percent. Compounded over five decades the gap is the whole story - an All REITs index rebased to 100 at December 1971 standing at 10,067.242 by 2024 is a total-return statement, and the same years on price-only columns tell a much thinner one.

Drawdowns read differently too. 1974's -42.23% for All REITs is a total-return figure, dividends reinvested straight through the loss - a price-only reconstruction of the same year would overstate the damage.

Yes - side by side, never spliced. FTSE Nareit series mark a traded market daily; the NPI is appraisal-smoothed and publishes about five weeks after quarter end. Set them in adjacent columns and the spread between them reads as sentiment versus fundamentals; merge them into one continuous series and you have invented a methodology neither source supports.

Beneath the composites, the NCREIF Property Index Office Returns record carries property-quarter rows - about 12,000 active properties with 67 fields each, drawn from a historical pool near 50,000, and NPI gross market value past $896 billion as of 1Q2024. Its income-return column is the private-market analogue of everything the listed income column measures.

For the physical stock underneath the rent roll, the EIA Commercial Buildings Energy Consumption Survey (CBECS 2018) contributes 6,436 sampled buildings standing in for roughly 5.9 million, across about 1,250 variables - vintage, tenancy and usage intensity, the inputs that decide what the returns can compound from a decade out.

Portfolio strategists and allocation teams use the 1994-forward property-sector cuts for rotation studies and the 1972-forward aggregates for cycle work - five decades is enough regime variety to test a thesis against something other than one bull market.

Researchers, journalists and analysts quote the anchors - the 1971 base, the 11,960.151 composite, the -42.23% year - because verifiable round trips beat unverifiable adjectives, and the T-Tracker supplies the operating footnote under every return column. Their page: office REITs for market researchers.

Across all of them the pattern repeats: the value is not any single annual print but the join - monthly grain, annual sector cuts and the live tape reconciling under stable keys.

Start with the question, then take the smallest record that answers it. Curves want the complete-history monthly record; sector-by-year cuts want the annual index values record; the live office tape wants the Sector Hub; private-market income wants NCREIF.

Then prove it on your own pipeline. Request a sample cut to the sectors, measures and months you name - real rows first, with the field dictionary and coverage statement attached. You keep all three regardless of what happens next.

For breadth, the best office-reits datasets ranking sorts the pool by quality score, the office reits data guide maps all 27 datasets in the slice, and sibling posts carry the adjacent questions: list of US office REITs explains who sits inside the sector columns, and what is the Nareit T-Tracker covers the operating layer that explains them.