Issue 074 - Refugee housing - Housing capacity
What would housing the displaced require?
UNHCR and European institutions continue to describe millions of refugees from Ukraine living outside the country. For this exercise, use 6 million refugees abroad as a rounded scenario benchmark.
Public systems and economicsAbout 1 minute
Sources checked October 12, 2026. Figures and circumstances may have changed.
The problem
If 6 million refugees from Ukraine lived abroad for several years, about how many homes or housing units would be needed to accommodate that population?
Then estimate how much existing housing might absorb, how much new or converted capacity might still be needed, and how that compares with annual homebuilding in a large European country.
Because Fermi problems target an order of magnitude, I normally use no more than two significant digits and write most calculations in scientific notation; the Fermi reference explains both conventions.
Grounding facts
UNHCR's 2026 planning material put refugees from Ukraine at roughly the 6 million scale globally, while Eurostat reported about 4.3 to 4.4 million people from Ukraine under temporary protection in the EU in spring 2026.
Eurostat's 2026 living-conditions report says the EU had about 203 million households in 2025, with an average of 2.3 people per household. That makes 2 to 3 people per housing unit a good first-pass peg.
Housing supply is not the same as housing stock. The European Commission says a substantial share of homes is unoccupied, but those units are unevenly distributed, may be holiday homes, may need repairs, or may not be where jobs and schools are. Germany completed roughly 250,000 dwellings in 2024 and closer to 200,000 in 2025, which is a useful large-country construction benchmark.
After checking sources
Checked answer and calculation
The total housing need starts with people per household:
housing units
~= 6 x 10^6 people / 2.3 people per unit
~= 2.6 x 10^6 unitsThat makes about 2.5 million homes the central estimate. If more refugees join existing households or share crowded accommodation, the separate-unit count might be closer to 1 million to 2 million. If families are smaller or many people need independent apartments, it stays near 2.5 million to 3 million.
The hard part is not that Europe lacks millions of dwellings in the abstract. The hard part is usable housing in the right places, at affordable rents, for years rather than weeks:
if half is absorbed by existing housing
remaining new/converted units ~= 1.3 millionif three-quarters is absorbed
remaining new/converted units ~= 650,000Compared with one large European country's homebuilding, the scale is substantial. Against Germany's recent pace of roughly 200,000 to 250,000 dwellings per year:
2.6 million total units
~= 10 to 13 years of German completions650,000 to 1.3 million net new/converted units
~= 3 to 6 years of German completionsDistributed across many host countries, this is not literally one country's whole construction program. But it is not a modest adjustment either. Even if existing stock absorbs much of the need, the pressure is large enough to tighten rental markets, strain social housing, and require serious conversion, subsidy, placement, or construction efforts.
Before checking sources
Matt's original estimate
This is the unverified estimate Matt wrote before checking sources, not the checked answer.
I don't have great mental models for how much housing inventory is available and how much is being built annually around the world.
I started with the assumption that there will need to be about one separate unit for every 2.5 people who have been displaced, which means about 2.4 x 10^6 units for 6 million people, assuming no cohabitating single units by multiple families.
I guessed that the average european country had a population of about 50 million people, and that the available housing represents about half a percent of the actual population, probably high, and I came up with about 500,000 available units per country - which means if NONE of that is taken by the domestic population, displaced refugees would consume the entirety of the available housing for about 5 average european countries. That's a pretty significant and unrealistic burden.
I then estimated that new housing development might occur at an annual rate of half the current available housing or about 250,000 new housing units per the average country, which means it would take all of the new housing development of 10 of the average european countries to house all 6 million displaced persons.
If we're able to distribute refugees across 20+ european countries, that might be about 120,000 housing units taken by refugees by each, which might be half of the new housing built or 1/4 of the available existing units, and it seems like split between those two without a tremendous change, it could be accommodated but it would have a big impact on each of those country's inventories in those years.
Reasoning score
Matt's reasoning score: 75 / 100
Higher is better: earn points for useful facts, a sound reasoning approach, correct math, and a final estimate close to the sourced answer. The owl meter shows percent full of it: 100 minus the reasoning score.
Useful facts: 20/30. The 2.5-person household peg was very good. The average-country population and housing-output assumptions were too high for Europe as a whole, though plausible for a large country.
Reasoning approach: 25/30. Separating total housing demand, existing absorption, and new construction was the right structure, but the average-country framing hid how uneven the capacity is.
Math: 10/10. The arithmetic followed cleanly from the assumptions.
Final estimate: 20/30. The total housing-unit answer was close. The difficulty of absorbing the demand through available or newly built homes was probably understated.
Post-check reflection
Matt's reflection
The estimate for total number of units needed to house that refugee population was good, everything else was maybe within an order of magnitude but still off. The population of the average European country was a high estimate, the stock of homes that are available was about right, the new housing production was good for a larger developed European nation but not good as an average across all.
Ultimately, my calculations led me to believe the housing issue was easier to solve than it probably is, in terms of availability of new and used homes.
Recommended memory peg
For refugee housing, start with housing units ~= people / 2 to 3 people per household. Then separately estimate how much existing housing can actually absorb before comparing the remainder with annual completions.
Reader results
Bars show how submitted estimates sort into the answer choices from the gut-check prompt.