How Predictable Is a Company's Next Office Move?

The most cited empirical study of firm relocation, published by Brouwer, Mariotti, and van Ommeren in The Annals of Regional Science (2004), analyzed relocation behavior across firms in 21 European countries. Three findings matter for a working broker.

First, change predicts movement, in both directions. Firms whose workforce grew were more likely to relocate, and so were firms whose workforce shrank. Stability is what predicts staying put. Second, external growth is a trigger: relocations frequently follow mergers, acquisitions, and takeovers. Third, firms serving larger markets relocate more often than purely local operators.

The study also gives brokers something most vendor pitches never mention: a base rate. Among the large registered firms studied, roughly 2.7 percent relocated in a given year. Later empirical work adds texture. A UK government analysis found expansion was the single most common driver, cited in about 47 percent of moves, and a 2026 study of Beijing firms found that in dense megacity markets, larger and high-tech firms show different patterns than the European baseline. The direction of the signals travels well. The exact coefficients do not, which matters later.

Because the lease is a fixed claim on a changing organization. A company that added 40 people has outgrown its space; a company that cut 40 is paying for empty desks it can shed at expiry. Both have a live reason to be in the market, while the company that looks the same year over year has none.

This is the piece most prospecting instinct gets wrong. Brokers naturally chase visible growth, funding announcements, hiring sprees, new logos on the door. The research says contraction is just as strong a signal and far less competed for. A tenant quietly running at 60 percent utilization after a layoff is a mover the growth-chasers never call.

Start with 2 to 3 percent a year and the arithmetic of cold outreach becomes clear. On an untargeted list of 500 companies, something like 10 to 15 will move this year, and several of those are already represented. This is why an office lease being a once-every-three-to-five-years decision means most of any raw list is out of market at any moment, and why untargeted volume feels like shouting into a void. It mostly is.

The leverage is in reordering the list, not lengthening it. If the base rate is 2.7 percent and a company shows two or three evidenced signals, a meaningful headcount inflection, a recent acquisition, a lease approaching expiry, the odds for that specific company are multiples of the base rate. The broker who calls 50 signal-bearing companies is working a fundamentally different funnel than the broker who calls 500 random ones, at a tenth of the effort.

Every predictor in the academic work has a public, watchable proxy, which is convenient for brokers in Washington State:

Headcount change shows up in job postings, H-1B LCA filing trends, and state employment filings well before it shows up in a space requirement. WARN notices flag the contraction side. Funding rounds and M&A announcements are public the day they happen and are direct analogs of the study's internal and external growth triggers. Firm age comes from Secretary of State registration records. Tenant improvement permits at a company's current address can signal recommitment, while permits at other addresses can signal an impending move.

We wrote a companion piece on where each of those records lives in Seattle and King County specifically, and how early each one fires. The research here supplies the why behind that list: these are not random tea leaves, they are the variables that carried statistical weight when someone actually counted.

Carefully, and only after asking how it was built. Two honest cautions, and they apply to every vendor in this category, including us.

First, calibration is hard with thin data. Confirmed office moves are rare events, and models trained on small samples of them produce scores that rank companies usefully but should not be read as precise probabilities. A tool that says "72 percent likely to move" on thin evidence is projecting confidence the underlying data cannot support. A range or a ranked list is the honest output. The right question for any vendor is what the score was calibrated against and what base rate it assumes.

Second, the industry's own trust instincts here are sound. First American and DealGround's 2026 survey found 66 percent of CRE professionals use AI weekly or daily, but only 5 percent trust it to inform actual deal decisions. For deal-critical judgments, that skepticism is rational until a tool shows its work. Scores should start conversations and order call lists. They should not close judgments.

The academic foundation is real but dated and foreign. The core study draws on European firm data from 1997 to 1999, and the 2026 Beijing research shows the size effect can flip in dense markets: findings from Dutch and Italian firms 25 years ago are priors, not laws, for Bellevue in 2026. Post-pandemic hybrid work also changed the meaning of headcount itself, since a company can grow 30 percent and shrink its footprint. Anyone claiming the relocation literature transfers cleanly to present-day Puget Sound is overselling it.

The defensible position is more modest: the research identifies which variables have ever carried predictive weight, and those are the ones worth watching locally. The weights need local recalibration. The variable list does not.

What percentage of companies move offices each year?
The benchmark academic estimate is roughly 2.7 percent of established firms per year (Brouwer, Mariotti, and van Ommeren, 2004). Younger, growing, shrinking, or recently acquired firms move at meaningfully higher rates.

What are the strongest signals a company will relocate?
Workforce change in either direction, merger or acquisition activity, younger firm age, and larger market reach, per the empirical literature. Locally observable proxies include job-posting trends, H-1B LCA filings, WARN notices, funding announcements, and permit activity.

Do layoffs mean a company will move offices?
They raise the odds. Workforce decreases predict relocation in the research, because a shrunken company is paying for space it no longer needs and has a concrete incentive to act at lease expiry.

Can AI predict office moves?
AI can rank companies by move likelihood using the signals above, which is valuable for prioritizing outreach. Precise single-number probabilities from thin data are false precision; ranked lists and ranges are the honest form of the output.