Every assumption behind the model

If you think building costs more or less than the country’s own figure, move this: 100% keeps it, 120% makes the works a fifth dearer in every city.

Our design choice Somewhat sure · we’re testing this · one of the twelve that matter most

A reader’s disagreement with the country’s own construction-cost data; 100% = that figure.

How far the works cost more (or, below zero, less) than their target. An overrun: the reserve for surprises pays the first part, and the developer shares the rest. A saving: the developer earns a share of it.

Our guess Not sure yet · we’re testing this · one of the twelve that matter most

Zero in the base case; the bad-decade scenario sets 15%, of which the reserve for surprises absorbs the first part. Below zero the works come in under their target and the saving is shared with the developer (‘Developer’s share of a saving’).

The whole floor area of the converted building, walls, stairs and shared rooms included.

Our design choice Somewhat sure

A city office block of the kind that stands half-empty in most European cities; this is the reference building the model runs, not a market average.

How long permits and building works take, during which money is spent and nothing comes in.

Market benchmark Somewhat sure

Typical time for an office-to-housing conversion (permits + works, ~3 years)

Architects, permits, surveys and project management, as a share of the building works.

Market benchmark Somewhat sure

Usual allowance for soft costs on a conversion

An allowance for what turns up once walls are opened (hidden structure, services, asbestos), as a share of the works.

Market benchmark Somewhat sure

Usual contingency on existing-building works (higher than for new build)

What it costs to fully furnish one flat for members on the move — beds, sofa, kitchen, storage, lighting.

Market benchmark Somewhat sure

Hospitality-grade fit-out for a ~60 m² serviced flat

What a m² of new housing costs to build in the country, land excluded — the yardstick the conversion cost is measured against.

Sourced data Somewhat sure

Curated national construction cost (excluding land), with its source on each country

What turning an office into homes costs compared with building new: 0.7× means converting costs 70% of building new, 30% cheaper. A poor candidate can cost more than 1×.

Sourced data Somewhat sure · we’re testing this

Curated per-country conversion-cost factor from conversion studies; 0.7× where a country has none

How much of the building is paid for with members’ and investors’ own money rather than a bank loan — more own money means a smaller loan, a lower rent and a loan a bank will actually give.

Our design choice Somewhat sure · we’re testing this · one of the twelve that matter most

The range runs to 100% because the bad-decade scenario needs about 92%. The value on this card is chosen so that a bank lends on its usual terms without counting the residents’ volunteer hours; when those hours left the bank’s test (2026-09-28) the reference building in Berlin needed 42%, and at around 35% about nine otherwise-workable cities fell just short.

The yearly interest a bank charges on the building’s main loan — one rate for the euro area, and each country’s own rate where it borrows in another currency (Swiss francs, zloty, pounds…).

Sourced data Somewhat sure · we’re testing this · one of the twelve that matter most

ECB cost-of-borrowing indicator / national central banks (2026), 2026

Of the building’s own money, the part put in by the residents (with any local co-op or ground-floor tenants); the Fund puts in the rest.

Our design choice Somewhat sure

The residents provide a large part of the own money so the building stays theirs; the Fund’s slice carries the early risk and helps the building pass the bank’s test.

Whether an existing local housing co-operative puts money of its own into the building alongside the residents.

Our design choice Somewhat sure

Off until a real local partner is found — a young co-operative usually has nothing to invest.

The share of the building’s own money a local co-operative puts in, which lowers what the residents must pay in.

Our design choice Not sure yet

Used only when a local co-operative is switched on above. We do not count on a partner we have not found, so this is a placeholder for the day one appears.

Whether the businesses renting the ground floor also buy co-operative shares, in proportion to the space they use.

Our design choice Not sure yet

The businesses on the ground floor belong to the building’s co-op alongside the households, and mixed-use co-operatives often ask their commercial tenants to buy in, in proportion to the space they use. Not yet tested with any real tenant.

How long the building takes to pay back what it borrowed; after that, the loan payments stop.

Market benchmark Somewhat sure

Typical repayment terms for European housing loans (25–40 years)

Banks fix the rate for a set number of years, not for the whole loan; after that the rest of the loan is re-priced at the rate of the day.

Market benchmark Somewhat sure

German and Austrian banks usually fix housing loans for 10–15 years

Points added to the rate when it resets after the fixed period (1 = one percentage point dearer, negative = cheaper); the rest of the loan is then repaid at the new rate.

Our design choice Somewhat sure

Zero by default: nobody knows the rate in ten years, so this is a stress lever to see whether the building survives a dearer loan, not a forecast.

The yield a valuer divides the building’s yearly income by to put a value on it — it sets the loan-to-value a bank sees, not the rent (a higher yield means a lower value).

Market benchmark Somewhat sure

Residential yields for stabilised German and Dutch assets, ~4–5%

Whether part of the borrowing comes from a co-operative, ethical or public bank at below-market interest.

Our design choice Somewhat sure

Off in the baseline: the model must work at market borrowing costs before it counts on cheaper money.

The yearly interest on the cheaper loan, when it is switched on.

Market benchmark Not sure yet

Typical terms of co-operative, ethical and public development banks

How much of the borrowing the cheaper lender provides, when it is switched on.

Our design choice Not sure yet

A placeholder until a lender tells us how much it would take. It is a weight, not a fixed share: the shares of the lenders that are switched on are scaled to add up to 100%, so this lender’s real part depends on the others that are on.

Whether part of the borrowing comes from one European facility the network negotiates for all its buildings (the kind the EIB runs).

Our design choice Not sure yet

A mature-network prospect, not something a first building can count on.

The yearly interest on the European framework loan, when it is switched on.

Our guess Not sure yet

Set below the euro-area bank rate; no such facility has been negotiated yet.

How much of the borrowing the European facility provides, when it is switched on.

Our design choice Not sure yet

A placeholder share until a facility exists. It is a weight, not a fixed share: the shares of the lenders that are switched on are scaled to add up to 100%, so the facility’s real part depends on the others that are on.

Whether part of the borrowing is a bond sold to members and the public, on which only interest is paid until it matures.

Our design choice Not sure yet

Off by default: it is an option we are exploring, not a plan.

The yearly interest the bond pays its holders, when a bond is used.

Our guess Not sure yet

Between subsidised money and a bank loan, as co-operative bonds tend to price.

How much of the borrowing the bond provides, when it is used.

Our design choice Not sure yet

Zero until a bond is issued — so switching the bond on does nothing until a share is set here; a co-operative bond would plausibly carry 10–20% of the borrowing. It is a weight, not a fixed share: the shares of the lenders that are switched on are scaled to add up to 100%, so the bond’s real part depends on the others that are on.

Whether the construction is paid for by a short-term loan that the long-term loan replaces once the building is full.

Our design choice Not sure yet

Off by default: the works are financed at the long-term rate until a construction lender is lined up.

The yearly interest on the separate construction loan, when it is used.

Our guess Not sure yet

Construction lending prices above long-term debt; the rate here is a cautious placeholder.

A one-off fee for setting up the construction loan, as a share of the loan.

Market benchmark Not sure yet

Typical arrangement fee on development loans (~1%)

The part of the construction loan funded by donations or grants that take the first losses, making the rest safer to lend. Not priced in yet: the model records the cushion’s size, but it changes no rate, fee or other number.

Our design choice Not sure yet

Zero until a foundation or public body offers such a cushion; until a lender tells us what it is worth, the model does not let it lower the loan’s price.

Whether the building uses the city’s or country’s social-housing grant and cheap loan, where a verified one exists.

Our design choice Fairly sure

Off by default: the model must stand on its own economics first, and only go further where a real scheme exists.

A grant that never has to be repaid, per m² of the long-term homes, from the local social-housing scheme.

Sourced data Somewhat sure

Verified national and city housing-subsidy schemes (2024–2026), 2026

How much cheap public loan the social-housing scheme offers per m² of the long-term homes.

Sourced data Somewhat sure

Verified national and city housing-subsidy schemes (2024–2026), 2026

The yearly interest on the public housing loan — often zero.

Sourced data Somewhat sure

Verified national and city housing-subsidy schemes (2024–2026), 2026

How many years the public housing loan runs.

Sourced data Somewhat sure

Verified national and city housing-subsidy schemes (2024–2026), 2026

The highest starting rent the subsidised homes may charge in return for the public money (zero means no cap).

Sourced data Somewhat sure

Verified national and city housing-subsidy schemes (2024–2026), 2026

Whether the building uses a public programme that rewards turning offices into homes, where one exists.

Our design choice Fairly sure

Off by default, like all public money: the baseline assumes none.

A grant covering this share of the cost of turning the offices into homes.

Sourced data Somewhat sure

Verified office-to-housing conversion incentives, e.g. KfW 266 (DE), PHB 2.0 (FR), 2026

The ceiling on the conversion grant for each home created. Zero means no grant at all, not “no ceiling” (unlike the subsidy’s rent cap) — set a figure here or the grant pays nothing.

Sourced data Somewhat sure

Verified office-to-housing conversion incentives, e.g. KfW 266 (DE), PHB 2.0 (FR), 2026

How much cheap public loan a conversion programme offers per m² of homes.

Sourced data Somewhat sure

Verified office-to-housing conversion incentives, e.g. KfW 266 (DE), PHB 2.0 (FR), 2026

The yearly interest on the conversion loan.

Sourced data Somewhat sure

Verified office-to-housing conversion incentives, e.g. KfW 266 (DE), PHB 2.0 (FR), 2026

How many years the conversion loan runs.

Sourced data Somewhat sure

Verified office-to-housing conversion incentives, e.g. KfW 266 (DE), PHB 2.0 (FR), 2026

Whether the co-operative leases the land for a yearly ground rent, buys the site outright, or an owner puts the building in for a stake.

Our design choice Somewhat sure · one of the twelve that matter most

A long lease keeps the land cost out of the loan and the land off the market for good; buying is the fallback where no lease is offered.

What the co-operative pays each year to lease the land, as a share of what the land is worth.

Market benchmark Somewhat sure

A cautious middle for a co-operative deal on a city office site; a council concession would be nearer 2%. The value on this card is the site default.

Heritable building rights (Erbbaurecht): market ground rents ~3–5%, municipal and co-operative leases ~1.3–2%

How many years the land lease runs — shown to lenders beside the loan’s term, because a bank wants the lease to outlast its loan by a wide margin.

Our design choice Somewhat sure

99 years, the usual length of a heritable building right; it is shown to lenders and checked against the loan, and does not change the rent.

The yearly return an owner who puts the building in (instead of selling it) receives on the part of it that is still theirs — zero would be a gift. Once the investors are out, the building buys the owner out at what the building was worth when it was put in, so the return shrinks and the co-op ends up owning it.

Our guess Not sure yet

A mission-aligned owner accepting about a ground rent’s worth; untested with any real owner.

What an empty office site is worth compared with land for new homes in the same city — below 100% means the site comes cheaper. It sets the site’s value however the land is held: the price if it is bought, the ground rent if it is leased, the owner’s stake if it is contributed.

Our guess Not sure yet

Half-empty office buildings trade below prime housing land; the results across cities lean on how far below, so this is one to test.

How many m² of floor the building has for each m² of land it stands on — it turns a land price into the value of the plot, and so sets the price, the ground rent or the owner’s stake, whichever way the land is held.

Market benchmark Somewhat sure

Typical density of an inner-city office block

Transfer tax, notary, registration and checks, as a share of the price, when the site is bought.

Market benchmark Somewhat sure

Property purchase costs across European countries (~4–8%)

What a m² of land for housing costs in the city — it sets the ground rent, or the price if the site is bought.

Sourced data Somewhat sure

Curated city data: residential land prices, shown with its source on each city

If you think the city’s land prices are too high or too low, move this: 100% keeps the city’s own figure, 120% makes land a fifth dearer in every city.

Our design choice Somewhat sure

A reader’s disagreement with the city’s own data; 100% = the city’s figure.

The share of the homes let furnished, for months at a time, to members moving between cities — they earn more and help carry the long-term homes.

Our design choice Somewhat sure · we’re testing this · one of the twelve that matter most

A quarter keeps the building mostly long-term homes while the furnished flats earn enough to matter. It is one of the numbers we are testing.

The share of the year the furnished flats are lived in.

Our guess Not sure yet · we’re testing this · one of the twelve that matter most

A growing member network should keep them full, but no building has run yet — a fixed input we are testing with partners and pilots.

What members on the move pay for a furnished flat, as a share of the six-month furnished market (scaled for the flat’s size) — every year: the flats’ price follows that market (‘Yearly rise in market rents’).

Our design choice Somewhat sure · we’re testing this · one of the twelve that matter most

Set below the market so the flats stay attractive for their flexibility and community, yet high enough that they help carry the long-term homes; it is not a deep discount. One of the numbers we are testing.

The share of rent from occupied homes that is paid late or never, apart from empty homes.

Market benchmark Somewhat sure

Typical arrears for well-run rented housing (~1–2%)

What a whole furnished flat costs per m² when you rent it for about six months on the platforms people use (HousingAnywhere, Wunderflats, Spotahome…), bills included — the price the mid-term flats compete with. It is quoted for a flat of about 45 m²; a bigger flat rents for less per m², so the model scales it down for the size of our furnished flats (by roughly 8–12% for a flat of 60–70 m²) before pricing them.

Sourced data Not sure yet

Measured for the focus cities; elsewhere it is estimated from the city’s ordinary rent, using the ratio the focus cities show. The size step is gentle (the 45 m² basis already leaves out pure studios) and never raises the price; the city’s own figure is what the city pages show.

Curated city data: mid-stay furnished rents, shown with its source on each city

What a furnished flat on a year-long lease costs per m² — not what a member staying a few months could get, so it is not the comparison. Used, raised by 40%, to estimate the six-month rent only where a city has no ordinary rent either.

Sourced data Not sure yet

The 1.4× step is the median across the 36 focus cities whose six-month rent was measured; it is an estimate wherever it is used.

Curated city data: furnished rents, shown with its source on each city

What a neighbourhood shop or café space rents for per m² each month, worked out from the city’s housing rent.

Market benchmark Not sure yet

No Europe-wide source for neighbourhood shop rents exists; broker data covers prime high streets only.

1.2× the new-lease housing rent (higher in cheap cities), checked against secondary ground-floor rents in Amsterdam, Berlin, Vienna, Madrid and Paris/Lyon

Whether the building has a ground floor let to a café, workshop or shop.

Our design choice Fairly sure

Office buildings come with street-level space that earns more as a business let than as homes.

How much of the lettable space goes to the ground-floor business.

Our design choice Somewhat sure

Roughly the ground floor of a five-to-seven-storey office block.

What the ground-floor business pays, as a share of the market rent for such a space — every year: its rent follows that market (‘Yearly rise in market rents’).

Our design choice Somewhat sure · we’re testing this

Let a little under market to keep a mission-aligned business in place.

The share of the year the ground-floor space is let.

Our guess Not sure yet · we’re testing this

Stands in for how many member companies take space; untested until a floor is let.

Whether the ground floor is simply let, or also receives events and bookings from the member network for a commission to the organisation.

Our design choice Somewhat sure

The event channel is an upside we have not tested; the value on this card is the site default.

The share of the year the long-term homes are lived in.

Market benchmark Fairly sure

Vacancy in affordable co-operative and social housing (near-full)

What utilities cost per m² each month in the city — added to the rent to show a household’s full monthly cost, and paid by the building for the furnished flats, whose price includes bills, so dearer utilities leave those flats less to spare.

Sourced data Somewhat sure

Curated city data: utility costs, shown with its source on each city

If you think repairs, tax, insurance, cleaning and the shared rooms cost more or less than the cards below, move this: 100% keeps them, 120% makes them all a fifth dearer (the savings for big repairs and the city’s utilities stay as they are).

Our design choice Somewhat sure

A reader’s disagreement with the running-cost figures in one move; 100% = the figures as they stand.

Routine maintenance and small repairs, per m² each month.

Market benchmark Somewhat sure

Typical running repairs for a well-kept multi-family building

Money put aside each year for the roof, façade, heating and lifts, as a share of what the works cost.

Market benchmark Somewhat sure

Renewal-fund practice in established housing co-operatives (~0.8%/yr)

Local property tax per m² each year.

Market benchmark Not sure yet

A European middle value — it varies widely by country

Insuring the building, per m² each year.

Market benchmark Somewhat sure

Typical buildings insurance for multi-family housing

Replacing and repairing furniture and equipment in the mid-term flats, per m² each month.

Our guess Not sure yet

Longer stays wear furniture less than a hotel; untested until a building runs.

A professional clean about once a month in the flats for members on the move, per m² each month.

Our guess Not sure yet

Members stay months, not nights, and keep the flats tidy themselves, so there is no hotel-style turnover.

Heating, lighting, cleaning and upkeep of the shared kitchen, workspace and lounges, per m² of shared room each month — shared between the long-term homes and the furnished flats by their floor area and paid as part of their running costs (the ground floor pays none).

Our guess Not sure yet

Shared rooms are used hard, so they cost more per m² than a home; residents run them themselves, so there is no manager’s wage.

How much the residents’ monthly fee is allowed to rise each year — slower than market rents, and never below what the building costs to run. It sets the homes’ fee only: the furnished flats and the ground floor follow their own market (‘Yearly rise in market rents’). It matters mainly in the first years: once the solidarity cap (‘Solidarity part: at most this share of what the homes cost’) binds — in years 7 to 9 in every city — the fee follows the building’s costs plus at most the cap, and this escalator no longer moves it.

Our design choice Somewhat sure · we’re testing this · one of the twelve that matter most

A steady rise set below market growth, so the gap opens over time. Where it runs above the building’s cost the difference is declared as a solidarity contribution to the network, and the fee is never set below cost. From the year the solidarity cap binds (years 7–9 in every city) the fee tracks cost plus at most the cap, so this number only matters before then — or if no cap applied.

How fast rents in the city rise each year. Faster growth widens the gap between the residents’ fee (whose yearly rise is set on ‘Yearly rise in residents’ fee’, and which after the first years follows the building’s costs) and the market, and it raises what the furnished flats and the ground floor earn, since they are priced as a share of their market every year — and with it EHC’s platform fee.

Our design choice Not sure yet · we’re testing this · one of the twelve that matter most

A planning assumption, not a sourced per-city figure; the stress scenario cuts it to 1%.

How fast the building’s running costs rise each year.

Our guess Not sure yet

Set a little under the ECB’s 2% inflation target because part of the cost base is fixed; untested over a long horizon.

What a new tenant pays per m² each month in the city today — the rent our homes are compared with.

Sourced data Somewhat sure

Curated city data: rents on new contracts, shown with its source on each city

The average rent across every current lease, older ones included — used where no new-lease figure exists, and to show what a long-standing tenant pays.

Sourced data Somewhat sure

Curated city data: average rents across the stock, shown with its source on each city

If you think the city’s market rents are too high or too low, move this: 100% keeps the city’s own figures, 90% makes ordinary, serviced and ground-floor rents a tenth lower in every city.

Our design choice Somewhat sure

A reader’s disagreement with the city’s own data; 100% = the city’s figure.

Once the homes’ loan is repaid, the share of the money freed that comes off the residents’ fee; the rest goes to the network to fund the next building. It works whether or not the solidarity cap binds.

Our design choice Somewhat sure

Our choice (2026-09-29, 80% by default): most of it goes to the households who paid the loan off — the moment an owner feels when the mortgage ends — and the network keeps the rest (the Mietshäuser Syndikat’s solidarity transfer is the precedent). The residents’ share comes off the fee and the rest goes to the network whether or not the cap on the solidarity part (‘Solidarity part: at most this share of what the homes cost’) binds; that cap limits a different thing and does not change this split.

The fee rises each year, but the part above what the homes cost — the solidarity part that buys the investors out and then helps build the next building — is never more than this share of that cost.

Our design choice Somewhat sure

Our choice (2026-09-29): residents carry a small, visible part of the network’s growth; most of it comes from the people who use the network — the furnished flats, the ground floors, the members’ dues. Without a cap the solidarity part grew to over half the fee after 40 years in Berlin.

How fast prices in general rise — used to show what a member’s money is worth in today’s money when they leave, to size any inflation compensation on refunds, and to keep a household’s loyalty account at its real value.

Market benchmark Fairly sure

A leaving member’s stake keeps its value in today’s money (the building’s co-op adds inflation), and so does a household’s loyalty account; an EHC member share is refunded at what was paid. None of them ever carries a market gain.

ECB medium-term inflation target (2%), 2021

How fast the organisation’s wages and costs — and, at the same pace, its fees and dues — rise each year.

Market benchmark Somewhat sure

ECB medium-term inflation target (2%), 2021

The average private floor area across the building’s long-term homes, singles’ flats and family homes together. The one-person flat the page uses as its example is a separate, representative size.

Our design choice Somewhat sure

Compact on purpose: generous shared rooms make up the difference.

The private floor area of a furnished flat for members on the move.

Our design choice Somewhat sure

The same size as a long-term home, so families and older members can use them too — not a studio.

How many adults live in a typical home.

Market benchmark Somewhat sure

EU household size ≈ 2.3 people, big cities ≈ 1.8–2.0 (Eurostat)

How many children live in a typical long-term home.

Market benchmark Somewhat sure

EU household size ≈ 2.3 people, big cities ≈ 1.8–2.0 (Eurostat)

How many children live in a typical furnished flat.

Our guess Not sure yet

Set equal to the long-term homes because the flats deliberately welcome families.

The share of the usable floor given to rooms everyone shares — a big kitchen, a place to work, a workout room, a kids’ corner.

Our design choice Somewhat sure

Small private homes whose living space is extended by excellent shared space is the heart of the idea.

Zurich co-operatives Kalkbreite and Mehr als Wohnen run far above the usual ~5%

How much space the average person in the city has at home — the yardstick our compact homes are compared with.

Sourced data Somewhat sure

Curated city data: average living space per person, shown with its source on each city

How many hours each adult resident gives each month to light upkeep, admin and community work, which lowers what has to be paid for. The saving is capped at 85% of the day-to-day repairs line (a paid professional still oversees the work), so at the standard value of an hour it stops lowering the fee at about one hour a month; more hours change no number.

Our design choice Not sure yet

Everyone contributes and everyone benefits; two hours a month is modest, and it never counts towards the bank’s test. Hours beyond the cap are community life, not a further saving in the model.

The paid work one hour of residents’ help replaces, in euros. The total saving is capped at 85% of the day-to-day repairs line, so once hours × value reach that cap a higher value changes nothing.

Our guess Not sure yet

Valued well below a tradesperson’s rate, since it is untrained community work.

Whether better-off households pay a little more for their stake so that households on lower incomes can pay less.

Our design choice Somewhat sure

Off by default (everyone pays the same) until a building’s members choose it.

The share of households who pay extra for their stake, when the solidarity pool is on.

Our guess Not sure yet

Roughly the better-off third of a mixed building; for each building’s members to decide.

The extra each contributing household pays, as a share of the normal stake.

Our design choice Not sure yet

A quarter more is a noticeable but bearable contribution; for each building’s members to decide.

The share of households whose stake the pool reduces.

Our guess Not sure yet

Roughly the lower-income third of a mixed building; for each building’s members to decide.

The middle income in the country after tax, per person (adjusted for household size, so it is what one person living alone has) — used to show what share of that income the rent on a one-person flat takes.

Sourced data Fairly sure

Eurostat EU-SILC median equivalised net income

What buying a flat costs in the city, per m²: the midpoint of the asking prices in the centre and outside it. Used only to compare with buying.

Sourced data Somewhat sure

The model uses the midpoint of the asking prices in the centre and outside it; where a city has only one of the two it uses that one, and where it has neither, a national estimate.

Numbeo, apartment prices per m², city centre and outside (2026), shown on each city

The share of the price a buyer pays from their own savings; the rest is the mortgage.

Our guess Somewhat sure

Most European banks lend up to about 80% of the price to a first-time buyer.

Transfer tax, notary, land registry and agent, as a share of the price — paid on top of the deposit.

Our guess Not sure yet

Between about 3% and 15% across Europe; 8% is typical of Germany, France and Spain.

The fixed interest rate on a buyer’s mortgage.

Our guess Somewhat sure

Set at the building’s own bank rate; a household usually pays a little more, so lift it to test a dearer mortgage.

How many years the buyer takes to repay the mortgage.

Our guess Somewhat sure

Thirty years is the common term for a first home.

What owning costs each year besides the mortgage — repairs, the building’s shared charges, insurance and property tax — as a share of the flat’s value. Heating and utilities are left out, as they are in the fee and the rent.

Our guess Not sure yet

Around 1–2% of the value a year is the usual rule of thumb for a flat.

How fast the flat’s value rises each year. It decides most of what an owner has after thirty years.

Our guess Not sure yet

It moves with market rents in every weather: 3.5% a year in a good decade, 1% in a bad one, and the same rate as market rents in our version. Prices have risen faster than rents in many cities since 2010, and fallen in some after 2022 — this is the number to test.

Agent and notary when the flat is sold, as a share of its value.

Our guess Not sure yet

Around 3% where the seller pays an agent.

Whether the people who move in put some of their own money into the building, as co-owners rather than tenants.

Our design choice Fairly sure · a rule we don’t bend

Residents owning their building together is the co-operative principle the whole model rests on.

Would a bank lend? Only if the building’s spare income covers each year’s loan payments this many times over.

Market benchmark Fairly sure · a rule we don’t bend

Volunteer hours are left out of the income the bank tests: a bank does not lend against unpaid work.

Standard senior lending floor for rented housing and commercial property (1.25–1.35×)

The most a bank will lend as a share of the building’s whole cost, interest during the works included.

Market benchmark Somewhat sure · a rule we don’t bend

Typical development-lender loan-to-cost ceiling

How the borrowing splits between the bank and any cheaper lender that is switched on; with only the bank loan on, the bank lends it all.

Our design choice Somewhat sure

Not a free choice: it is a weight against the cheaper lenders that are switched on (the shares of the lenders that are on are scaled to add up to 100%), so with only the bank on the bank lends everything whatever this says. The public levers are the cheaper-money switches and their shares.

How much of the construction loan is, on average, borrowed while the works run — it sets the interest paid before anyone moves in.

Market benchmark Somewhat sure

Usual approximation for a loan drawn down evenly over a build

Whether banks in the city will give a long-term loan at all; where they will not, the building borrows from a central European facility instead.

Our guess Somewhat sure

Assumed true across our focus cities; some markets in South-East Europe may need the central facility.

The yearly interest on the central European loan used where local banks will not lend long-term.

Our guess Not sure yet

A pooled European rate a little below the euro-area bank rate; no such facility exists yet.

How many years the central European loan runs, where it is used.

Our guess Not sure yet

Shorter than a local mortgage, as central facilities tend to be; a placeholder.

Whether the model asks for more own money where a land lease is weaker security for a bank.

Our design choice Somewhat sure

Where land leases are rarely used for housing, lenders want a bigger cushion; ignoring that would flatter those countries.

Leasehold viability research across 30 countries, 2026

How many extra percentage points of own money a building needs in a country where land leases work but are thin for housing.

Our guess Not sure yet

Which countries fall in each tier is researched; the size of the extra cushion is our judgement.

Leasehold viability research across 30 countries, 2026

How many extra percentage points of own money a building needs where land leases are short, hard to borrow against or barely used for housing.

Our guess Not sure yet

Which countries fall in each tier is researched; the size of the extra cushion is our judgement.

Leasehold viability research across 30 countries, 2026

How many m² are built for each m² people can use: 1.25× means 1.25 m² built for each m² you can use, so about 80% is usable and the rest is stairs, corridors and walls.

Market benchmark Somewhat sure

BDA Bayern, Office to Housing (2025): living area about 80% of gross floor area in converted offices (75% when subsidised); TU Delft, Adaptive Reuse for Housing (2025): office lettable area loses 10–20% as homes

A target the model checks, not a price: the members’ row for the furnished flats passes only if their price (set on ‘Furnished flats priced at this share of the market’) is at least this far below the market.

Our design choice Somewhat sure

A check on the price card, not a second price: just below market is enough, because the flats’ value is flexibility and community, not a deep discount.

How many members live in or want to come to the city — shown beside the furnished flats as a check on demand.

Our guess Not sure yet

A placeholder until the appetite test gives real numbers per city.

The share of members who, at some point, move to another city for a few months.

Our guess Not sure yet

Half is our guess for a network that attracts people who move for work; untested.

A member on the move takes a stay once every this many years.

Our guess Not sure yet

Once every three years is our guess for members who move for work; untested.

How long a typical stay in a furnished flat lasts.

Our guess Not sure yet

About six months, the middle of the three-to-twelve-month range the flats are designed for.

Below this much private space per person the model warns that homes are becoming overcrowded.

Our design choice Somewhat sure

A guard-rail, not a cost: compact is fine, cramped is not.

A test, not a price: the residents’ row passes if the homes start at least this far below a new lease in the city, OR if the fee grows more slowly than market rents so the gap opens over time. The fee itself is always the building’s cost, so in year one it can sit close to the market.

Our design choice Somewhat sure

The homes charge what they cost, not a discount, so being level with the market in year one is allowed; what the test insists on is that residents get ahead of the market over time.

The yearly return paid on the money the Fund puts into a building — it is part of what the rent has to cover, so a lower return keeps the homes cheaper.

Our design choice Somewhat sure · we’re testing this · one of the twelve that matter most

3.5% is enough to interest a broad range of impact investors, individuals included, while keeping the rent low.

The ceiling on the yearly return on all the investors’ money in a building — the part in the homes and the part in the furnished flats and ground floor alike. Anything a building earns above it stays with the building and the network, never with investors.

Our design choice Fairly sure

A capped return is what keeps the model non-speculative.

How long investors’ money stays in the building before the building’s co-op starts buying them out.

Our design choice Somewhat sure

Five years protects the building, and its bank, through the risky first years after opening.

After the lock-up, how many years EHC takes to buy — with its own money, in equal yearly slices — the part of the investors’ stake that becomes its lasting share (‘EHC’s lasting share of each building’). It does not pace the rest: the building’s co-op buys the other investors out as fast as the building’s surplus allows.

Our design choice Somewhat sure

Ten years spreads EHC’s own purchase so it never strains EHC’s cash. When the investors are fully out depends on the building’s surplus, not on this card.

The small share of each building EHC itself buys up to and keeps for good, as the anchor that stops it ever being sold off — EHC’s own stake, not the investors’ money (that is a separate internal setting).

Our design choice Somewhat sure

About 10%: enough to hold the mission lock, small enough that the building’s co-op owns the rest. EHC buys up to it from its own money, spread over the years on ‘Years over which EHC buys its lasting share’.

What the local partner that leads the design, permits and works earns, as a share of the works plus the planning fees and the reserve for surprises — a one-off, paid from the building’s budget.

Our design choice Somewhat sure

A modest margin on a target-cost contract; the developer is paid, it takes no stake in the building.

The part of any cost overrun beyond the reserve for surprises that the developer bears, never more than its margin.

Our design choice Somewhat sure

Half each way under a target-cost contract, capped at the margin, so the developer shares the risk without being ruined by it.

When the works come in under their target cost, the part of the saving the developer earns on top of its margin — at most its margin again. The building keeps the rest.

Our design choice Somewhat sure

Half, the mirror of the overrun rule: a target-cost contract shares the gain as it shares the pain, so the developer has a reason to keep the works below plan.

Whether a professional housing manager runs the building for a fee, or the building’s co-op runs it itself.

Our design choice Somewhat sure

A new building’s co-op is not assumed to run a whole building alone from day one.

What the building manager is paid, as a share of the rent collected, for running the building.

Market benchmark Somewhat sure

Kept low because the platform and residents’ help carry part of the work.

European management-fee band (~6–12% of collected rent)

How much of the manager’s fee goes on the manager’s own staff and costs; the rest is its margin.

Our guess Not sure yet

A guess until an operator shares its numbers.

Whether the building is run by a different organisation from the one that developed it.

Our design choice Somewhat sure

By default one local partner builds and then runs it. It changes who the money-flow map names, not the numbers: the pay is the margin and fee cards.

The share of a normal salary the founders take while EHC is being built. The rest is owed to them as deferred pay.

Our design choice Somewhat sure

The founders carry the risk that EHC never works. Taking less pay early leaves the money in EHC when it is scarcest; what they give up is owed to them and repaid only out of surplus.

How much the pay the founders are owed grows each year while they wait for it: their reward for carrying the risk.

Our design choice Not sure yet

Deferred wages, not a share of ownership or a dividend: nothing is owed if EHC never makes a surplus. The premium is set like a subordinated loan’s interest, because the founders are paid after every other cost and only once EHC has earned back its losses.

The most the founders can ever receive for their deferred pay, as a multiple of what they gave up.

Our design choice Not sure yet

A cap keeps the reward proportionate however long the wait. At 3× and our other defaults, about €0.8 million given up in the founding years is repaid as about €2.4 million, most of it in the fifteen years after that.

The most of last year’s surplus that goes to the founders in a year. Repayment starts only after the founding years, once EHC has earned back its losses and holds its operating reserve.

Our design choice Somewhat sure

Half, so the founders are repaid without starving the community fund or the price relief that also come out of surplus.

What each building pays EHC, as a share of what its furnished flats and ground floor earn, for the platform and the members it brings.

Our design choice Somewhat sure

6% funds the central organisation while leaving the building enough for investors and a reserve.

A one-off fee EHC earns for finding the building, putting the money together and bringing the project to life, as a share of its cost.

Our design choice Not sure yet

How EHC could earn without holding stakes; a judgement, not a market rate. Set it to zero to see the network without it.

A small fee, on top of the rent, that a member pays EHC when booking a furnished flat — it funds the platform, not the building.

Our design choice Not sure yet

Small enough that a member’s all-in price (the rent plus this fee) stays clearly below a comparable furnished flat on the open market.

What each individual member pays EHC every year.

Our design choice Not sure yet · we’re testing this

Low enough to join for the values and the network, not only for a home.

The one-off co-operative share an individual buys to become a member, refunded when they leave.

Our design choice Not sure yet

An amount most people can afford, and refunded when they leave; a plan value, not tested.

The one-off co-operative share a company buys to become a member.

Our guess Not sure yet

One share size for every company for now; a placeholder until the tiers are priced.

What a member company with fewer than 10 staff pays EHC each year.

Our guess Not sure yet · we’re testing this

Scaled to company size so a small firm can afford to join; a placeholder until companies tell us what they would pay.

What a member company with 10 to 49 staff pays EHC each year.

Our guess Not sure yet · we’re testing this

Scaled to company size so a small firm can afford to join; a placeholder until companies tell us what they would pay.

What a member company with 50 to 249 staff pays EHC each year.

Our guess Not sure yet · we’re testing this

Scaled to company size so a small firm can afford to join; a placeholder until companies tell us what they would pay.

What a member company with 250 to 9,999 staff pays EHC each year.

Our guess Not sure yet · we’re testing this

Scaled to company size so a small firm can afford to join; a placeholder until companies tell us what they would pay.

What a member company with 10,000 or more staff pays EHC each year.

Our guess Not sure yet · we’re testing this

Scaled to company size so a small firm can afford to join; a placeholder until companies tell us what they would pay.

Of all member companies, the share with fewer than 10 staff.

Our guess Not sure yet · we’re testing this

Weighted toward the employers who actually move staff between cities, not Europe’s company count (most firms are tiny); a placeholder until the appetite test. The five shares must add up to 100% (the controls keep them there).

Of all member companies, the share with 10 to 49 staff.

Our guess Not sure yet · we’re testing this

Weighted toward the employers who actually move staff between cities, not Europe’s company count (most firms are tiny); a placeholder until the appetite test. The five shares must add up to 100% (the controls keep them there).

Of all member companies, the share with 50 to 249 staff.

Our guess Not sure yet · we’re testing this

Weighted toward the employers who actually move staff between cities, not Europe’s company count (most firms are tiny); a placeholder until the appetite test. The five shares must add up to 100% (the controls keep them there).

Of all member companies, the share with 250 to 9,999 staff.

Our guess Not sure yet · we’re testing this

Weighted toward the employers who actually move staff between cities, not Europe’s company count (most firms are tiny); a placeholder until the appetite test. The five shares must add up to 100% (the controls keep them there).

Of all member companies, the share with 10,000 or more staff.

Our guess Not sure yet · we’re testing this

Weighted toward the employers who actually move staff between cities, not Europe’s company count (most firms are tiny); a placeholder until the appetite test. The five shares must add up to 100% (the controls keep them there).

Money EHC raises before its first building — grants, a founding round — to carry its costs through the years before the buildings’ fees and members’ dues cover them.

Our design choice Not sure yet

Zero in our version. EHC runs at a loss for its first decade or so; the members’ shares carry it because they stay in the bank until EHC covers its costs (the first-stake guarantee is off). Founding money would let EHC take its stakes in the first buildings sooner — or carry it if the guarantee is switched on (the “What EHC has in the bank” panel shows how much).

Hosting, data and payment costs of the platform, as a share of what it earns.

Our guess Not sure yet

Hosting, data and payment fees of a booking platform; a guess.

What one full-time employee costs in a year — salary, employer’s taxes and workplace — averaged across Europe.

Market benchmark Somewhat sure

Blended fully-loaded cost of professional staff across European countries

Offices, IT, travel and insurance, as a share of what the staff would cost at full pay. Audits and filings per country, the Fund’s running costs and automation tools are their own cards.

Market benchmark Somewhat sure

Standard overhead build-up for a professional-services organisation

Money spent on advertising and campaigns, as a share of EHC’s income.

Our guess Not sure yet

A small advertising budget, since most members arrive through the network; a guess.

What EHC spends on events, knowledge exchange and a solidarity fund for each member each year, once it can afford it (see ‘Pay the community fund only from surplus’).

Our guess Not sure yet

What EHC spends per member each year on events, exchange and a solidarity fund; a guess.

How much faster than prices the pay of EHC’s staff rises each year. The work left to people — negotiating, building trust, members’ democracy — gets dearer as the economy grows.

Our guess Somewhat sure

Real pay in Europe has grown by well under 1% a year on average this century; 0.75% is our central guess over sixty years. It is paired with ‘Pace of automation’: one makes each hour dearer, the other needs fewer of them.

Each year, software and AI take over this share of the routine work still left in each team. Each team has its own ceiling: most member support is routine, most deal-making and leadership is not.

Our guess Not sure yet

Nobody can say what AI will do over sixty years, so the model shows the team at half and at 1.75× this pace as a range. At 4% a year about a third of the routine work is automated after ten years and half after seventeen. It arrives with time, whatever EHC’s size; ‘How much more a mature team gets done’ is the separate gain from doing the same work many times.

Automation is not free: part of the pay it saves comes back as software licences and computing.

Our guess Not sure yet

A quarter is our guess; it keeps the saving honest rather than counting it in full.

Whether the events and community fund is paid only out of last year’s surplus (after the founders), rather than every year whatever the result.

Our design choice Fairly sure

It is money handed to members, so it waits until EHC can afford it. Paid every year, it would cost about as much as EHC’s whole cash cushion in the loss-making years.

One-off lawyers’, tax and bank costs of setting up EHC’s entity in each new country, with the regulatory advice. The first countries are set up one a year.

Our guess Not sure yet

A co-operative entity, tax registration, a bank and advice on housing and financial rules in a new jurisdiction; our guess.

Each country’s entity must be audited and file its accounts and tax returns every year.

Our guess Not sure yet

A small entity’s statutory audit, filings and registrations; our guess.

For every building that goes ahead, how many were checked with paid surveys and legal work and then dropped.

Our guess Not sure yet

Many more buildings are looked at than bought; this counts only the ones that got as far as paid checks.

Surveys, valuation, legal checks and a feasibility study spent on a building that is then dropped.

Our guess Not sure yet

Our guess for a mid-sized office conversion.

The Fund’s administration, audit, custodian and regulatory reporting, paid by EHC as the Fund’s sponsor. While the Fund is small only a third of it applies; all of it from €100 million, where the EU’s light regime for small funds ends.

Our guess Not sure yet

A regulated fund needs an administrator, an auditor and a depositary; small funds can register under the lighter regime of the EU’s fund-manager directive (AIFMD, article 3(2)).

The part of the Fund’s running costs that grows with the money it holds, in hundredths of a percent a year (10 = 0.1%).

Our guess Not sure yet

Administration and custody are usually charged as a small share of the assets; our guess.

Part of the solidarity part each household pays is kept in its name and paid back when it leaves.

Our design choice Somewhat sure

Our choice (2026-09-29): years of helping the network grow are not simply gone if a member moves away. It is a member credit on what they paid, never a return on money and never a vote.

The share of each year’s solidarity part credited to the household’s account. It keeps its value against inflation and is paid out when the household leaves; the rest stays with the network.

Our design choice Somewhat sure

Half: the other half still funds the next building. A credit the co-op owes its member, paid from the building’s surplus before the investors are bought out.

The account becomes the household’s in equal steps over this many years of membership; a household that leaves sooner takes the part it has earned.

Our design choice Somewhat sure

Ten years: it rewards staying, not a quick move in and out.

A leaver’s account is paid out in equal parts over this many years, so no building is drained when several households leave at once.

Our design choice Somewhat sure

Three years, like a co-op’s usual notice period for paying back member shares.

When a household or a business downstairs leaves the building’s co-op, how much of general inflation is added to the stake it paid in. 0% = refunded at face value; 100% = it keeps its value in today’s money.

Our design choice Somewhat sure

Our choice (2026-09-28): full inflation, so a stake held for 20 years does not lose a third of its value — never more, so nobody profits from the building. The next member pays it (the next card).

Only matters when inflation is added to leavers’ stakes (the card above). On: the building pays it from what it has left over each year, before buying the investors out — a newcomer still pays the same stake, but less is left for the buy-out and for the next building. Off: the next member pays it, through an entry stake that rises with the same index — the building pays nothing, but joining gets dearer.

Our design choice Not sure yet

Our choice (2026-09-28): the next member pays, as in housing co-ops that index, so the building’s surplus still goes to the buy-out and the next building. The price is a way in that gets dearer over time, at the pace of inflation.

The share of the building’s households and businesses that move on in a year. It sets how much inflation compensation on stakes falls due each year and, when the loyalty account is on, how many accounts are paid out.

Our guess Not sure yet

The same 8% the network assumes for long-term homes coming free.

When a person or a company leaves EHC, how much of general inflation is added to the member share it bought. EHC sets that aside from its own result every year, so every leaver can be paid in full.

Our design choice Somewhat sure

Our choice (2026-09-28): face value. Set apart from the building’s stakes: the shares are EHC’s to pay, so compensating is a yearly cost on EHC’s own result that grows with its membership (full inflation costs it over a million euros a year by year 60), and it touches no building’s fee or buy-out.

Whether the building’s co-op buys the Fund’s investors out over time, so the building ends up owned by the people who live and trade there.

Our design choice Fairly sure · a rule we don’t bend

Ownership passing to the building’s co-op — its households and the businesses downstairs — is the point; investors are paid back, not bought in for ever.

Whether a veto held by EHC and the residents stops any sale of the building for profit.

Our design choice Fairly sure · a rule we don’t bend

It is what makes permanent affordability a guarantee rather than a promise; it has no cost in the numbers.

The least impact investors will take each year before the money stops coming.

Our guess Not sure yet

A check on the investors’ return (‘Return to the Fund’s investors’, the lever that sets it), not a second return: what foundations and mission investors have typically accepted; to be confirmed with real investors.

Which stage of EHC’s life a building belongs to — early on, impact investors provide almost all of the Fund’s money; later, EHC can put in more of its own.

Our design choice Somewhat sure

Honest about the early years: EHC’s own money keeps the organisation alive before it can fund buildings. The live page opens on the pilot stage.

How much of a pilot building’s Fund money EHC puts in itself; investors provide the rest.

Our design choice Somewhat sure

A token 1% first-loss stake: EHC is a real owner of the first building without betting its runway. The building uses the share for its own stage; the network, when each building keeps the stage it started in, uses all three.

How much of the Fund money EHC puts in itself for buildings started while it is scaling up.

Our guess Not sure yet

EHC’s own money grows as its operating reserve fills; the step is our judgement. The building uses the share for its own stage; the network, when each building keeps the stage it started in, uses all three.

How much of the Fund money EHC puts in itself once the network is mature.

Our guess Not sure yet

EHC’s largest share once it can borrow against its stakes; the level is our judgement. The building uses the share for its own stage; the network, when each building keeps the stage it started in, uses all three.

EHC’s share of the Fund money when it is set by hand rather than by stage.

Our design choice Not sure yet

Only used when the funding stage is set by hand.

Whether part of the Fund’s money comes from members and the public buying small, tradeable investment shares at a lower return.

Our design choice Not sure yet

Off while we test whether it is legally and practically workable.

How much of the Fund’s money members’ tradeable shares provide, when switched on.

Our guess Not sure yet

A tenth of the Fund money is a cautious guess for what members might invest.

The yearly return members’ tradeable shares pay — lower than impact investors’, because they can be sold and the mission is the point.

Our guess Not sure yet

Separate from the investors’ return (‘Return to the Fund’s investors’). Members accept about 1% because the shares can be sold and the mission is the point; untested.

How much of the money members pay for their co-operative shares goes into buildings; the rest is kept aside as a cash buffer.

Our design choice Somewhat sure

Four-fifths invested, a fifth kept liquid so leaving members can be paid back.

Whether members’ share capital is reserved for EHC’s stakes from the first building on, even while the organisation is not yet paying for itself.

Our design choice Somewhat sure

Off in our version (since 2026-09-30): members’ shares go into EHC’s stakes only once EHC covers its running costs and keeps its cash reserve, so its account never runs below zero; until then outside investors hold EHC’s stakes in the first buildings, and EHC buys them as its surplus allows. On, every member owns a piece of the first building — and EHC’s account runs short while it still loses money, unless it raises founding money.

Whether a building’s mix of EHC and investor money is set by the year it started, so early buildings stay investor-funded as EHC grows.

Our design choice Somewhat sure

Off in the code default, on in the live model: it is how EHC’s own share of the whole network rises over time.

Buildings started from this year on count as the scaling stage.

Our guess Not sure yet

Set around the time the first buildings have opened and EHC’s fees start to come in; a guess.

Buildings started from this year on count as the mature stage.

Our guess Not sure yet

Set for when the network has a track record and a standing book; a guess.

The share of a building’s money left after its investors are bought out that goes to the Fund, as own money for the next building; the rest stays with the building’s co-op.

Our design choice Somewhat sure

At 100% all of it revolves: each building helps start the next, so less outside money is needed as the network grows.

Money raised each year from members and the public through tradeable investment shares, separate from membership shares.

Our design choice Not sure yet

Zero by default: impact investors fill the Fund until such a raise is tested.

Whether investors hold a share of all buildings together rather than one building each, which spreads the risk.

Our design choice Not sure yet

Off by default: one building per investment is simpler to start with.

The part of the pooled fund, usually from foundations or public bodies, that takes any losses first and so protects everyone else.

Our design choice Not sure yet

A tenth from foundations or public bodies is typical of blended impact funds; untested with ours.

The part of the pooled fund that is repaid first and carries the least risk.

Our design choice Not sure yet

The protected slice most blended impact funds offer institutional money; untested with ours.

How likely it is, each year, that a building loses some of its investors’ money. Only matters when the pooled fund (‘Pool the buildings in one fund’) or the network-wide loan (‘A network-wide loan against the Fund’s stakes’) is switched on — both read it.

Our guess Not sure yet

A cautious placeholder; co-operative housing has a strong record, but a new network has none.

How much one building’s trouble goes with another’s (0 = independent, 1 = all at once) — lower means spreading across cities protects more. Only matters when the pooled fund (‘Pool the buildings in one fund’) or the network-wide loan (‘A network-wide loan against the Fund’s stakes’) is switched on — both read it.

Our guess Not sure yet

Buildings in different cities and countries should fail mostly independently; 0.2 allows for a shared downturn.

Whether the Fund borrows once, against all its building stakes together, to lend more cheaply than any single building could.

Our design choice Not sure yet

Off by default: EHC takes on no debt until it has a standing book to borrow against.

How many times the Fund’s lasting stakes it can borrow against.

Our guess Not sure yet

Three times the stakes, with the stakes as the cushion beneath; a placeholder until a lender prices it.

The rate a building borrowing alone would pay — the network-wide loan is priced below it because its risk is spread.

Market benchmark Somewhat sure

Only the network-wide loan reads it, as the rate it is priced below; the building’s own loan uses the bank loan rate (‘Bank loan interest rate’), which is the public lever. Keep the two equal unless you mean them to differ.

ECB cost-of-borrowing indicator / national central banks (2026), 2026

A yearly fee EHC charges the Fund for managing its buildings, as a share of what the Fund holds.

Our design choice Not sure yet

Zero by default; it only moves money between EHC and the Fund.

A yearly payout from the Fund to its outside investors, as a share of their money, to show the flow between the two.

Our design choice Not sure yet

Display only — it draws the Fund-to-investor flow and does not change any building’s numbers. Zero means it follows the investors’ return (‘Return to the Fund’s investors’), which is the public lever.

The fee below which the model says a building does not cover its share of EHC’s costs.

Our design choice Somewhat sure

A check, not a price: the level at which a building’s fee is thin enough that EHC’s central costs would not be covered across the network. A judgement; the fee itself is the card above.

When a building earns less than planned, whose share gives way — the residents always pay the full cost of their homes either way.

Our design choice Somewhat sure

EHC is the mission-driven party, so its fee flexes first and only the investors’ return is a hard claim.

What EHC earns per m² each month by sending member and company events to the ground-floor business.

Our guess Not sure yet

Zero until the event channel is tested.

How many months of its own running costs EHC holds in cash before putting its members’ shares or its surplus into buildings. (With the first-stake guarantee on, members’ shares set aside for buildings go in whatever the reserve.)

Our design choice Somewhat sure

Fifteen months of runway first: an organisation that runs out of cash helps no building. With the first-stake guarantee off (our version) the reserve comes before any stake; with it on, it holds back EHC’s surplus only.

Yearly income from councils, other co-operatives or firms that use EHC’s digital platform.

Our guess Not sure yet

Zero until anyone outside the network pays for it.

The fee below which the model flags that running the building is no longer a viable business for the manager.

Our design choice Not sure yet

Below about 4% of rent a professional manager could not staff a building; our judgement.

Whether the network’s recycled surplus lowers the stake that residents of later buildings must pay in.

Our design choice Not sure yet

Off by default: it changes both the stake and the rent, so it is switched on deliberately.

How much smaller the residents’ stake is in a pilot building.

Our design choice Fairly sure

Zero: in the pilot years there is no surplus yet to pay for it.

How much smaller the residents’ stake is in a building started while the network scales up. Only read when this building’s own stage (‘Network stage when this building is funded’) is set to scaling — the page opens on the pilot stage, and the network view does not apply it to its later buildings.

Our guess Not sure yet

What a few years of recycled surplus could fund; our judgement.

How much smaller the residents’ stake is in a building started once the network is mature. Only read when this building’s own stage (‘Network stage when this building is funded’) is set to mature — the page opens on the pilot stage, and the network view does not apply it to its later buildings.

Our guess Not sure yet

What a mature network’s recycled surplus could fund; our judgement.

The average pace at which the network opens new cities after the first pilots — few early, more later. It sets how many cities are reached by the end (this pace × the years, up to the most-cities ceiling); the openings follow a curve that starts slow and speeds up, so after ten years far fewer cities are open than ten times this number.

Our guess Not sure yet · we’re testing this · one of the twelve that matter most

Three cities every two years on average: a pace a co-operative can staff, which keeps opening cities for most of the horizon.

When the Fund’s own money — what buildings send on and what EHC earns above its reserve — is not needed by any building yet, it pays for extra buildings, breadth first: the next cities in the plan’s order open early (at most two openings a year, one extra building each), and what is left adds buildings in cities that are already open and have room (at most one extra a year in any city). All starts in a year stay within this many times the planned pace. 1 turns it off.

Our design choice Not sure yet

Our choice (2026-09-29): money the network earns goes into more homes, not into a bank account. Twice the planned pace is the most by default, because teams and partners limit how fast buildings can really start.

Of the empty buildings in a city that could become homes, the share EHC can realistically win on workable terms — against private converters, hotels and other bidders.

Our guess Not sure yet

About one in twenty. Dutch housing corporations own 7% of office conversions (CBS 2024); Paris’s social landlords did 42% of large ones, but with the city behind them; Dublin’s council assessed 500+ buildings and ended with one scheme. The number that decides how many buildings each city can take.

CBS, Transformaties in de woningvoorraad 2023/2024; APUR 2023; Dublin City Council via Dublin Inquirer 2025, 2025

The share of a city’s empty office floor in whole buildings that can be converted AND pay their way at a workable price — available when EHC arrives.

Market benchmark Somewhat sure

About 30% of empty office space is technically convertible (ifo/Colliers 2024 for Germany’s seven largest cities; Cushman & Wakefield NL 2022: 32%), but only a fraction sits in whole empty buildings (≈20% in Brussels) and pays its way at undistressed prices — about 10%.

ifo/Colliers 2024; Cushman & Wakefield NL 2022; perspective.brussels 2024, 2024

The share of a city’s whole office stock that is converted into homes each year, by anyone — the new buildings that come up after EHC arrives.

Market benchmark Somewhat sure

Measured: about 0.15% a year in Paris (APUR 2023), about 0.6% in Brussels (perspective.brussels); the Netherlands converts some 2,800 buildings a year (CBS 2025).

APUR 2023; perspective.brussels 2024; CBS 2025, 2023

Once the network has the experience, other empty buildings add to what a city can supply each year — churches and their halls, schools, barracks and public offices, department stores, some hospitals and hotels — counted in buildings of our size and shown as a share of what its offices supply. More in the north-west of Europe, where churches close fastest; less in Central and Eastern Europe.

Our guess Not sure yet

About 40% (a range of 12% to 100%). Where homes are made from old buildings, other buildings give 0.5× (England, 2024-25) to 1.4–1.9× (the Netherlands, 2023–24) the homes offices do, but mostly in small buildings; at our size churches, schools, state property and department stores add about a third to a half. Surplus is well documented — a quarter of Dutch churches already have a new use, German churches plan to sell some 40,000 properties by 2060, Italy has closed 2,600 schools in a decade, and state owners (BImA, the UK and French state) sell at a discount for homes — but only 8 of 56 closed German department stores became homes.

CBS Transformaties 2023/2024; House of Commons Library 2025; RCE 2020; Il Sole 24 Ore; Empirica/ZIA 2024, 2025

The year of the network’s life from which it also takes on other empty buildings.

Our design choice Not sure yet

Eight years: public buildings and hotels can come earlier, churches and schools later — harder conversions, best taken on once the first buildings are running.

The year by which other empty buildings add their full share; they phase in evenly until then.

Our design choice Not sure yet

Twenty years: a gradual ramp as the network learns each kind of building, while the surplus of churches, schools and hospitals keeps growing through the 2030s–2050s.

However few empty buildings a city offers, the network can take at least this many in total there. With the supply limit on, it is the floor under every small city: raise it and all the smaller cities add buildings at once.

Our design choice Not sure yet

Two by default: opening a city for a single building rarely pays for the local work. It is more powerful than it looks, because it applies to every small city the network opens and so lifts the size of the network in the later years.

On: each city’s buildings follow its supply of convertible empty buildings. Off: a ceiling by population alone (the old rule).

Our design choice Somewhat sure

Our choice (2026-09-29): a plan should not count buildings nobody would offer.

How much of its own money the Fund keeps in hand, counted in years of what new buildings need. Above it, money goes first into more buildings and then back to members as lower prices.

Our design choice Not sure yet

Our choice (2026-09-29): three years, so growth never has to wait on outside investors again once the network pays for itself — and no more, so money is not simply kept.

The cities where a building adds up but the plan does not reach wait in reserve; when the Fund’s own money is idle, it opens them, in the plan’s order.

Our design choice Not sure yet

Our choice (2026-09-29): growth comes before handing money back — but only in cities where the numbers work. In our version it changes nothing: the plan already reaches every city where a building adds up, so no city waits in reserve. It would matter with a lower ceiling on the number of cities.

What is still idle above the buffer, after the extra buildings, is not collected the next year: lower dues for every member, cheaper furnished flats for members on the move, and a smaller solidarity part for the households — each in proportion to what they paid in.

Our design choice Not sure yet

Our choice (2026-09-29): a co-op should not keep what it does not need. It is a refund of use, never a dividend; EHC hands back only its own surplus, never below its costs.

The ceiling on how many cities the network opens in.

Our design choice Somewhat sure

A ceiling of roughly three or four cities in each of about thirty focus countries.

The most cities any one country opens, its best-suited first, so no country dominates.

Our design choice Somewhat sure

A few per country, close to the pilot shortlist, so the network stays pan-European rather than national.

How many buildings a city starts each year in the early years, before it tapers toward its ceiling. With the supply limit on, a city can only start what it can really source, so this sets how fast cities reach their supply, not how many buildings there are at the end.

Our guess Not sure yet

Half a building a year: one co-op with one local partner manages about 0.3–0.5 a year in the cases we know (wagnis in Munich, Kalkbreite in Zürich, the City of Paris’s own landlord at ~0.7), and a first building takes 5–8 years after entering a city. Because each city is also limited by the buildings it can really source (‘EHC’s share of the convertible buildings in a city’), moving this barely changes the network in year 60; it changes how soon it gets there.

The most buildings the network puts in one city of two million people or more; smaller cities get a share of it by population. It matters only when the limit by what a city can really source is switched off.

Our guess Not sure yet

Used only when limiting each city to the buildings it can really source (‘Limit each city to the buildings it can really source’) is off, or for a city with no supply estimate; while the limit is on, each city takes what it can really source and this number does nothing.

How long, in the network view, it takes from finding a building to its first residents moving in.

Our guess Not sure yet

Longer than the works alone, because finding, negotiating and financing a building comes first.

Of the buildings converted across the network, the share that are easy (shallow floors, sound façade) and so cheaper.

Our guess Not sure yet

Half easy, a fifth hard lands the blend on each country’s own conversion factor.

The share that are hard (deep floor plans, new façade and services) and cost about as much as building new.

Our guess Not sure yet

A fifth hard; together with half easy it lands the blend on each country’s own conversion factor.

How much office floor stands empty in the city — it sets how many buildings the city could eventually host.

Sourced data Somewhat sure

Curated city data: vacant office stock, shown with its source on each city

How many people live in the city on stays of three to twelve months — a sign of demand for the furnished flats when choosing where to start.

Sourced data Not sure yet

Curated city data: mid-term residents, shown with its source on each city

How many people join before the first building to found the co-operative across Europe.

Our design choice Not sure yet · we’re testing this

The plan’s target for the appetite test: 1,000 founding individuals.

How many people who do not live in a building join through it for each home — the waiting list, future residents and exchange users.

Our guess Not sure yet

A guess: a well-liked building’s waiting list, future residents and exchange users can easily outnumber the people who live in it.

How many companies join before the first building — from anywhere in Europe, not only the first cities — and stay as a floor, leavers replaced.

Our design choice Not sure yet · we’re testing this

Our choice (2026-09-29): 50 companies paying before the first building — the appetite test aims for more, but the closest comparable network (Job & Wohnen Berlin) started with 8. Change it here.

How many member companies the network can expect in each city it is active in.

Our guess Not sure yet · we’re testing this

A placeholder until partners give real numbers: companies follow the network’s footprint rather than growing on their own.

How many more member companies each open building draws in — suppliers, managers, employers housing staff there.

Our guess Not sure yet

Suppliers, managers and employers housing staff in the building are the ones it draws in; a placeholder until our first buildings show who actually joins.

The share of the gap to the expected number of companies that closes each year.

Our guess Not sure yet

Companies decide over a budget cycle or two, not overnight, so only part of the gap closes each year; a placeholder until our own members tell us.

The share of member companies that leave in a year.

Our guess Not sure yet

Set a little below the 16–18% a year that membership bodies typically lose (they keep 82–84% of members: MGI 2025/26; i4a), on the bet that a company housing staff through the network has more reason to stay. A placeholder until our own members tell us.

The ceiling on member companies across the whole network.

Our guess Not sure yet

A ceiling so the company base cannot outrun the network; a placeholder (decided 2026-09-28).

The share of the furnished flats set aside for staff sent by member companies; any unused go back to all members.

Our design choice Not sure yet

Off in our version: every furnished flat stays open to all members. Raise it to set a share aside for staff sent by member companies, so the promise of housing for company staff means something while most flats stay open to all.

The share of what a building spends on works and running that member companies could bid to supply.

Our guess Not sure yet

Most of what a building spends on works and running — repairs, cleaning, energy, fit-out — is bought locally and could be supplied by member companies; the share that is realistically theirs is a placeholder until they bid.

How many times faster a city starts buildings once the organisation is experienced (1 = no faster).

Our guess Not sure yet

Our version lets an experienced organisation start somewhat more buildings per city, ramping up over the years on the next card; set it to 1 to not count on getting faster.

How many years it takes to go from the early pace to the mature one.

Our guess Not sure yet

Twenty years for an organisation to reach its full pace; only matters when the faster-pace card is above 1.

The share of long-term homes whose residents move on in a year, freeing them for other members.

Our guess Not sure yet

A member moves on about every twelve years.

How many nights a year a resident offers their home to other members while away.

Our guess Not sure yet

A few weekends and a holiday a year, times the share of residents who offer; a guess.

How long a short stay in another member’s home lasts.

Our guess Not sure yet

A long weekend.

The share of members actively looking for a place to live in a given year — many join for the network and the values.

Our guess Not sure yet

Many join for the network and the values, so only a quarter look for a home in any year; a guess.

The model warns when the homes that come free serve fewer than this share of the members looking for one (0.5 = half of them; 1 = all of them).

Our design choice Somewhat sure

Below half, the promise of access starts to ring hollow.

How many of a member company’s staff join as individual members themselves.

Our guess Not sure yet

Off in our version: we do not count the staff of member companies as individual members until companies tell us how many would join. Raise it to add members (with their dues and shares) through companies.

How many full-time people run the organisation in its first year.

Our design choice Somewhat sure

Three founders full-time in year one — the team we can realistically fund from founding members.

The most the team can grow in a year — hiring and settling people in takes time.

Our guess Not sure yet

Hiring and settling people takes time, so the team cannot double overnight; a placeholder for a real hiring plan.

How many years the founders’ way of working (more field work, less head office) takes to fade out.

Our guess Not sure yet

About eight years, until the first cities and buildings are established.

How much of the central head-office team does not exist yet at launch and is built up over the founding years.

Our guess Not sure yet

A founding team cannot staff a full head office on day one; 60% held back is our judgement.

How much of the founders’ time goes on opening cities and finding buildings in the first year.

Our guess Not sure yet

The founders do the first city-opening and building work themselves; our judgement.

The leadership and central team that exists whatever the network’s size — finance, legal, partnerships, communications, operations.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

Regional coordination, local law and partnerships, per country the network is active in.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

The ongoing presence in each city — supporting local members, the council relationship and the pipeline of buildings.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

Full-time years of work to open a new city — finding offices, the local co-op, the council and investors.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

Central work to find and structure each building while it is being developed.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

Central oversight of each building once it is open (0.05 = one person for 20 buildings); the building manager does the day-to-day work.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

How many members one member-support person looks after — higher means more self-service.

Our guess Not sure yet

A digital-first co-operative automates much of member support; a guess.

How many members there are for each person working on marketing and growth.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

Finance, legal and compliance work per country — each country needs its own regulated entity.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

The team that raises and looks after the Fund’s money from investors.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

Extra fundraising and investor-relations work for each open building (for example, 0.006 is one person for about 170 buildings).

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

Staff who organise members’ assemblies, councils and elections in each country; the taking part itself is voluntary.

Our guess Not sure yet

Lean, because members run their own assemblies; a guess.

The team that argues for better housing policy and shares what the network learns, once the network is fully grown.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

Work measuring and reporting each building’s social and environmental impact.

Our guess Not sure yet

A bottom-up staffing guess: with the other staffing cards, the team grows with the network to about eighty people by year 60. To be replaced by a real organisation plan.

How many times more work per person the central team does once the network is large, from doing the same work many times: templates, repeat deals, known partners. Automation is a separate card.

Our guess Not sure yet

Learning from repeat work should cut the effort per building by a third at 500 open buildings; untested. Until 2026-09-30 this card also carried automation (at 2×); that is now ‘Pace of automation’.

The engineers who build and run the digital platform, whatever the network’s size.

Our guess Not sure yet

A small core engineering team; a guess until the platform is specified.

Extra engineering for each open building the platform serves (for example, 0.02 is one engineer per 50 buildings).

Our guess Not sure yet

One engineer per few dozen open buildings; a guess.

HR, recruiting, internal IT and the layer of managers a team needs as it grows, per 100 other staff.

Market benchmark Somewhat sure

Typical HR-to-staff ratios of 1–2 per 100 in mid-sized organisations, plus internal IT and management

Engineers who build the digital platform before the first buildings open. The extra team shrinks to nothing over the founding years, as the buildings it serves open.

Our guess Not sure yet

The booking, membership and exchange platform has to exist before there is anything to book; our guess until the platform is specified.