Office Occupancy and the Jobs That Fill Offices
Nareit office occupancy against the level implied by the 2000-2019 relationship between occupancy and office employment — Information plus Professional and Business Services. Each quarter is read off that quarter's employment alone, so nothing accumulates.
The implied line is the 2000-2019 relationship between office occupancy and office employment, read off each later quarter's employment on its own. Nothing is chained or accumulated, so an unusual quarter affects that quarter alone, and the line does not depend on where a starting point is placed. It is a benchmark, not a forecast, and it does not establish what caused the relationship to change — remote work, sublet space, lease expiries and the mix of listed REIT portfolios are all consistent with it.
What does it show?
Office employment has recovered; office occupancy has not. By 2026 Q2 occupancy sat 5.2 points below the level that much employment implied before 2020 — a gap that opened to 9.0 points and has been closing since.
Methodology
Fit occupancy = a + b1 × Information employment + b2 × Professional and Business Services employment over 2000 Q1 to 2019 Q4 (R² 0.89, both weights positive), then evaluate the same fixed coefficients on each later quarter's employment. Held out cold — fitted to 2000-2011 and asked for 8 unseen years — it lands within 0.95 points RMSE, worst quarter 3.00. Professional and Business Services employment now runs above its fitted range; holding it at that range's top moves the current gap by 0.86 points, so the finding does not rest on the extrapolation. The implied line peaks at 96.0%, below the 96.9% high the series has actually recorded.