How the Flight Risk Index works
The Flight Risk Index combines searches, reading and reported tracker entries into a monthly indicator. It does not count departures or predict how many people will leave.
Method version 4. Data as at .
Sources and weights
Search makes up 50%. We use UK monthly search estimates from Ahrefs, retrieved on 29 October 2026. They are estimates, not counts of people. The ten phrases are moving to dubai, leave the uk, emigrate from uk, p85, statutory residence test, moving to portugal, moving to australia, non dom, moving abroad from uk and moving to spain.
Reading makes up 25%. We use monthly user page views from Wikimedia for the English Wikipedia articles the article on expatriate tax status, British diaspora and Golden visa. All 243 monthly values for January 2020 to September 2026 matched a provider check on 29 October 2026. We do not know when the original copy was retrieved.
The Flight Risk Capital tracker makes up 25%. It counts entries by the month they were first reported. The tracker snapshot was counted on 25 October 2026 and contains 82 verified records, including one with only a year that is excluded from monthly counts.
How is the score calculated?
Each search phrase and reading topic is compared with its own average for 2020 and 2021, set to 100. Each search phrase is rounded to one decimal before we take the middle value; the search middle value is also rounded to one decimal. Reading values keep their precision until the final score. The tracker uses the rolling total for the latest 12 months, compared with the mean of those rolling totals in 2022 and 2023. Each signal is capped at 300 before taking the weighted average. If a signal has not started, the available weights are adjusted to add up to 100%. Scores are rounded to one decimal place. Missing search or reading data stop publication.
These components have different reference periods. The combined score does not have a single baseline of 100 and is not a percentage of the population. The dial runs from 0 to 300 because the components are capped at 300. It is a scale for the score, not a probability or a forecast.
We also compare the current score with the average of the 24 monthly composite scores in 2020 and 2021. That period included pandemic travel restrictions and sparse tracker records. Tracker coverage can affect this comparison.
What are the limits?
Searches and reading may reflect curiosity or a single story. Wikipedia reading is global, not limited to UK residents. Tracker entries depend on research and press coverage and include plans, returns and corrected records. The tracker is a retrospective collection, not a complete count of reports or departures. Earlier records are sparse. Before the first known record, the early rolling windows assume zero entries, not that nothing happened.
The signals can overlap. A score above an earlier average does not establish that more people will leave. The 12-month average is the mean of the current and previous 11 scores, rounded to one decimal with half points rounded up. Neither series is adjusted for the time of year.
Revisions
On 29 October 2026 version 4 removed news because its historical source data could not be fully checked. The remaining signals retain their previous relative weights. All 81 months were recalculated. Version 3.1 and its correction history are archived as withdrawn data in the project records. Later tracker additions do not automatically change saved scores.
Index revision log. Latest Index and monthly values. Full series CSV. Official departure statistics and sources.