Independent initiative · The Baukultur standard
RomânăThe Baukultur standard for Romania — aligned with Davos 2018 & the New European Bauhaus.
Romania has no architecture policy. Concretely: someone who wants to build a house has no one to ask for good, free advice, and public money builds, too often, ugly and expensive. We propose a policy worthy of the stakes — on a par with Europe’s best, for the good of all.
9 chapters · 20 objections anticipated · every figure, sourced
national architecture policies in force in Romania (2026)
since the only policy we ever had (OAR, 2010–2015) lapsed
countries in the region that already HAVE an architecture policy or law
01The starting point
The only one we ever had was a profession-led document, for 2010–2015. Its term lapsed and it was never renewed. The profession is regulated — the quality of the place we live in is not.
We have a solid legal framework for who may practise architecture. We have nothing for what and how we build — no quality rubric, no state architect, no free advisory network, no competition service, no report to Parliament.
We have the “profession” half of Europe’s established model. We are missing the “quality of place” half — exactly the piece we set in place here.
02Romania on the world map
Estonia since 2002. Czechia since 2015. Lithuania, Slovenia, Croatia, Hungary — all have adopted a state instrument for the quality of the built environment. Romania stands beside Poland: a profession-led document that never became a policy of the state.
The built environment as national wealth
PASK — measures with an owner + a deadline (the closest model)
A Law on Architecture
“Architecture for People”
Aligned with Davos + NEB
A comprehensive law (a centralised model, criticised)
A profession-led draft, never adopted — the same trap
No policy in force — the last one lapsed in 2015
We are among the last in the region. But it is precisely those who arrive last who can start straight from the proven solutions, without repeating the others’ mistakes.
proven mechanisms, from 7 countries, that we are the first to wire into a single system
We are not the first with a policy — Finland (1998), Estonia (2002), Czechia (2015) and Croatia (2019) got there before us. The novelty is different: no one has gathered exactly these mechanisms TOGETHER into one self-funding, capture-resistant national system. Czechia has measures + an owner + a deadline; Croatia is Davos- and NEB-aligned; the EU OMC report gathers them on paper. We wire them into one apparatus that works — funded from three sources, so it doesn’t fall at the first budget.
how quickly Sweden’s much-praised policy unravelled once its funding was withdrawn
what CABE (UK) cost as a share of construction spending — abolished all the same, in a single budget revision (2011)
03The thesis
The global pattern is a trap: instruments of principle (declarations) endure, but they state without binding; binding instruments compel, but fall at the first budget. The answer is to weld them into a single system — self-funding and capture-resistant.
England abolished CABE — a flagship institution — in a single budget revision. Sweden’s policy unravelled within six years of its funding being cut. France is watching its advisory network slide into liquidation because it depends on a single funding source. These are all lessons already paid for.
Two things here are different from Flanders or the Netherlands, and the policy has to face them. First: most small communes have neither an architect nor an urban planner of their own — “design capacity” simply isn’t there; this is why the free advisory network (Pillar 5) is not a frill but the layer without which the rest stays a dead letter. Second: public procurement has an integrity problem that the “quality, not lowest price” rule must withstand — which is why the juries are public, the criteria written in advance, and jury reports published as a matter of course. We don’t import a system; we cut it to our own ground.
We arrive among the last — and that is exactly why we can take from each what has proven itself and wire it, for the first time, into a single system that funds itself and cannot be captured. We don’t invent new mechanisms; we finally make them work together.
04The blueprint
Each pillar takes the mechanism that has proven the most robust in its role — and each brings architects a clear advantage, without adding red tape.
If you remember one thing: anyone — a citizen, a mayor, a small builder — will be able to ask, free of charge, at the same counter, for an architect’s advice before spending a single euro. The other pillars support this one gain.
The quality of the built environment = the public interest
A short clause, in law, declaring the quality of architecture and of the built environment to be in the public interest.
What gives it force: It is the cheapest measure with the greatest leverage: it legitimises every other mechanism without adding cost or red tape.
The eight Davos criteria — with thresholds, a score, and binding effect
We adopt the eight Davos criteria (Governance, Functionality, Environment, Economy, Diversity, Context, Sense of Place, Beauty) as a national instrument of assessment.
What gives it force: Unlike Davos, it gains thresholds, a score, and a procurement grid: from a statement it becomes a working instrument. A common language for advice, competitions and codes, kept free of arbitrariness by its very design — and never imposing a style: “Beauty” is judged as fit with place and how people live the place, not as personal taste. The binding effect applies only in public procurement and funding, through a government decision or a law — never over private projects.
A steward of quality + an independent college
A national chief architect (the Rijksbouwmeester model) + a multidisciplinary college of advisors.
What gives it force: Open recruitment, a fixed five-year term, statutory protection on both appointment AND dismissal, transparency through public reports. It advises government, with no veto: it cannot block a permit.
Competition on quality, free for the public client
A national competition service for public clients, under EU procurement law (the design contest — art. 105–110 of Law 98/2016), followed by a negotiated award of the contract to the winner: ≥4 teams selected on design capacity (not turnover), paid for a vision, awarded on quality (not price).
What gives it force: The highest-yield mechanism — this is where a policy of principle becomes built quality. Free for the client, with a dedicated track for small practices and young architects.
Free advice, with no veto
Free, impartial architectural advice for citizens, mayors and small builders, folded into the permit through a single window, limited to one session, on clear and objective criteria.
What gives it force: The rule “advise, don’t refuse”, written into law. It is the capacity layer that staffs small town halls — extended through AI so it can reach everywhere.
Quality as an accelerator, not a brake
We tie public money and subsidies to the quality rubric; a pattern book of award-winning, fast-tracked designs; design embedded in social-housing finance.
What gives it force: Quality speeds delivery up (the NSW model: approval in 10 days) and gives small practices a cheap, predictable route — rather than slowing them down.
The loop that makes underfunding expensive
A national biennial report — from surveys of town halls + public opinion + expert workshops — presented formally to Parliament with a government response and a debate.
What gives it force: It turns a declarative policy into a loop of political accountability — and makes the quiet withdrawal of funding visible and costly. It is also what publishes the indicators by which results are measured.
05Objections, anticipated
This is where policies like this usually fail: red tape, delays, costs, subjectivity, new committees, politicisation. We took them all and, for each, set the mechanism that — by its very design — takes out of the equation the way such policies usually fail. We don’t promise nothing can go wrong — we show why precisely the known traps become hard to trigger.
The advice sits inside today’s permit, at the same desk, in a single meeting, on clearly written criteria. No new file, no extra application.
Source: France — CAUE ↗A fixed statutory deadline, with penalties for overrun. And the advice comes before the file, not after — so the permit gets faster, not harder.
Source: France — CAUE + statutory deadline ↗For the citizen and the town hall it is and stays free at the counter — that is the deliverable. In the government-decision phase it is covered from the existing budget and an independent fund (a government decision cannot create a tax). Only the law, later, can tie funding to a small special-destination share of the authorisation fee already paid today (Tax Code, art. 474) — proportional to the investment, not a new cost on every file.
Source: France — CAUE (development levy) ↗Advice replaces decisions made on a whim; it does not add a body on top of today’s. No one gains the power to block.
Source: France — “advise, don’t refuse” ↗The same public rubric for everyone — the 8 Davos criteria, scored by an independent panel. Never an imposed style.
Source: Davos Quality System ↗It asks for a result — performance, fit with place — not a particular look. You can step outside the norm if you prove your solution delivers the same result.
Source: Davos system (performance, not style) ↗The pattern book is optional and varied: a menu of award-winning designs, used as a shortcut — not a mandatory typology.
Source: Australia — NSW Pattern Book ↗Teams are chosen on the quality of the design, not on turnover. There is a dedicated track for graduates and free advice.
Source: Flanders — Open Call + Meesterproef ↗It hooks onto the existing procedure, through the same desk, and aligns with the building Code. It puts order into what exists — it does not add a new layer.
Source: Aligned with the Code (CATUC) ↗One national rubric, plus a network that brings skilled people into small town halls. The same standard everywhere.
Source: France — CAUE + the Davos rubric ↗Free advice, given early, lowers the cost of building well. And AI tests the options faster and cheaper.
Source: France — CAUE ↗A role set in law, open recruitment, a fixed term, protection on appointment AND dismissal. The juries are public.
Source: Netherlands — board of advisors ↗Public criteria, a multidisciplinary panel, and jury reports published as a matter of course. Anyone can check why a project won.
Source: Brussels — BMA (public juries) ↗The Davos rubric weighs function, cost, environment and context equally — not just “beauty”. And “beauty” means fit with place, not personal taste.
Source: Davos Quality System ↗Quality speeds things up, it doesn’t brake them: award-winning pattern-book designs are approved in 10 days, plus a fixed deadline for the rest.
Source: Australia — NSW (approval in 10 days) ↗It regulates objectives and performance, not fixed recipes. You can innovate as much as you like, if you prove an equivalent result.
Source: Davos system (objectives, not recipes) ↗The funded advisory network is exactly the help small town halls lack today. AI reaches the places that have no architect.
Source: France — CAUE ↗We fund the capacity first, and only then ask for obligations (the UK “design deficit” lesson). The money comes from a fund kept at arm’s length from politics.
Source: Netherlands — the Stimuleringsfonds ↗Three sources, in steps, not one: first a multi-year budget line and an independent fund (doable without a law), then — by law — a small special-destination share of the existing authorisation fee (art. 139 of the Constitution requires a law for that). So funding doesn’t fall at the first cut, and no new tax appears while we are in the government-decision phase.
Source: The CAUE / Sweden collapse lesson ↗The Czech model: every measure has an owner and a deadline, plus a biennial report to Parliament and public indicators. It self-corrects.
Source: Czechia — PASK ↗A policy is not defended with good intentions. It is defended with mechanisms that allow no abuse, whoever is in power.
06AI, cooperative
AI will soon take on much of architects’ work — soon, whole projects, generated and checked end-to-end. The policy does not pretend this isn’t happening: it sets the rules that make the architect the accountable operator of the tool, not its spectator. The value stays with the profession.
We don’t say “never”. We say: accountability for a project stays with a person — a registered, insured, trained architect — however much AI generates and checks. We don’t ban the tool; we require demonstrable human oversight, in the spirit of Europe’s AI rules (art. 14 of Regulation (EU) 2024/1689 on AI; meaningful human control in the Council of Europe Framework Convention): the architect must be able to reconstruct and justify the solution, not merely confirm it. For architectural design, AI is not today a “high-risk system” under Annex III — which is exactly why we require this oversight by policy.
We use AI for faster, cheaper research-through-design — testing more options against the Davos-8 criteria before committing the money — and for the single window with a deadline.
The free advisory network uses AI to reach the small town halls that have no architect — closing the design deficit faster. The advisor stays human, and their advice cannot block a permit — only improve it.
We propose updating national training and continuing-education standards (Law 184/2001 + the OAR statute), inspired by the logic of Directive 2005/36/EC, with the coordination of AI tools and whole-life carbon assessment — so the architect is the certified operator of AI in the built environment, not a third party that bypasses the profession. AI in public procurement stays transparent and auditable.
The edge case the policy anticipates rather than dodges: a project generated and checked end-to-end by AI. The answer is not a ban but a ladder: how much the tool may work on its own depends on the stakes of the decision.
Massing options, energy and daylight simulations, renders, working through regulations. AI works autonomously; the architect reviews.
Code compliance, planning fit, technical solutions. AI proposes, the architect confirms with reasons — documented, not tacit.
Fire and structural safety, the fit with place and community, the final signature. No automated output decides until a person who is accountable owns it.
07The evidence
Not “pay us more”, but “here is what you gain” — and what we lose by doing nothing. Four figures, each with its source, each able to survive checking.
of how much a child learns in a year comes down to how their classroom is designed — light, air, shape.
Univ. of Salford, the HEAD project — 3,766 pupils, 153 classrooms, 27 schools (a large sample; generalising beyond UK primary schools warrants caution). Strong evidence.
salaries vs rent vs energy, in an office’s costs. People are almost everything — which is why their comfort beats any saving on energy.
Because people dominate the cost, a ~1% productivity gain from light/air/view outweighs the entire energy bill. That is why good design pays for itself.
is what Britain is now paying to fix unsafe cladding — the “saving” made at design stage cost, in the end, incomparably more.
NAO central estimate (range £12.6–22.4bn), in the wake of the Grenfell tragedy. The saving made at design stage (“cost optimisation”) cost, over the life cycle, orders of magnitude more than it saved up front. Strong evidence.
of the carbon “hidden” in materials is saved by renovating instead of demolishing — as much as 20+ years of the building’s emissions.
Embodied carbon ≈ 20+ years of operational emissions (RIBA 2030 / LETI). To be judged case by case — retrofit is not always greener.
Patients with a window onto nature — rather than a brick wall — recovered faster after surgery and asked for fewer painkillers. It is the study that started, four decades ago, all of evidence-based design; it has been confirmed ever since.
A child who learns better, a patient who heals faster, a building that needn’t be repaired at great cost — all begin with the same design decision. Good design is not a luxury: it is the cheapest investment a state can make.
The policy + the evidence, as an open, citable document — not “the first policy”, but the first to wire into a single self-funding, capture-resistant system the mechanisms others use separately.
The public–private coalition (WEF secretariat) that turns “high-quality Baukultur” into practice. Romania thereby takes its place on the international map of built-environment quality.
Romania already takes part at project level (regional development agencies, NEB funds) — but with no national framework. The policy provides that framework and opens access to European co-funding.
08Abroad
A policy that wires into a single system the quality mechanisms used separately across Europe does not stay a domestic matter. We make it known abroad — and it becomes the international credential of Romanian architecture.
Romania already takes part in the New European Bauhaus — but at the level of scattered projects, with no national framework. The policy provides that framework: immediate European legitimacy and a co-funding lever. And we enter the Davos Baukultur Alliance not as an observer, but with what others have done in pieces and we have wired into one system.
This raises the standing of the Romanian architect: not a slogan, but a document that others cite.
09The path to adoption
A policy cannot be decreed. We designed the path so it does not hang on a single vote, a single minister, a single moment. It begins with what we can do ourselves, today, and advances toward law only as each step proves its worth. If it stops halfway, what has been done still stands.
The realistic target: a policy taken on by the state through a government decision within 18–24 months — without waiting for a law. The groundwork (steps 0–2) we lay this year, whatever the political calendar.
The document in public-policy form (measures, an owner, a deadline), the Davos-8 rubric calibrated, a mandate from the National Conference. It depends on no one outside.
Within our control — starts immediatelyThe standard enters the Order’s competitions and reviews; the “quality–price” guide under Law 98/2016 (the design competition, Art. 105–110). Without asking permission.
Within our control — in parallel with Step 0A handful of real competitions and free advice for small town halls, with the numbers to show for it. We don’t wait for the pilot to finish before asking for the political step — it runs in parallel.
The variable: a willing public clientThe government mandates the relevant ministry — MDLPA (the Ministry of Development, Public Works and Administration) — to draft the policy. It commits politically, not legally; zero cost. It is requested in parallel with the pilot — it need not wait for it.
The first political variable: a willing ministryThe policy adopted by government decision (the HG procedure, GD 775/2005), with measures, owners and deadlines. This is where it is actually adopted — with no parliamentary cycle. A government decision takes on the vision and the standard for state investment; it cannot, without a law, oblige town halls or levy taxes.
The key threshold · the fastest real “yes” — depends on political willProtection of the State Architect against dismissal, multi-source funding (an earmarked levy, through a law — Art. 139 of the Constitution), the biennial report — anchored in law and in the Code. It does not block adoption; it strengthens it afterwards.
Slow (a parliamentary cycle) — which is why it is NOT a precondition for adoptionA window forces the pace: the Territorial Planning Code comes into force in 2026, and its implementing rules are being written right now — that is where we graft on, while they are still open. Which is why steps 0–3 start in parallel, this year.
Honestly, without idealism: steps 0–2 depend on us alone and start immediately; steps 3–4 depend on political will; step 5 (the law) is slow. That is why the adoption target is a government decision — not a law — and we start, in parallel, everything that needs no one’s consent. In the government-decision phase no new tax is created: the cost is covered from the existing budget and an independent fund; a proportional share of today’s authorisation fee can be tied to quality only later, by law.
How it actually passes
A policy is not adopted because it is good, but because someone can sign it without losing. So we designed it to start small and to disarm, by design, the predictable opposition.
The smallest signable “yes” asks only this: the statutory declaration (the quality of the built environment = the public interest) plus the Davos-8 rubric applied only in state-funded procurement — public money only, no new tax, no obligation on municipalities or on private projects. The rest (protecting the State Architect, multi-source funding, the biennial report) requires a law — art. 139 of the Constitution. The natural owner ministry: MDLPA (the Ministry of Development, Public Works and Administration).
Who opposes it — and what they get in return
“It slows me down.”
The pattern book of award-winning designs, with fast-track approval (the NSW model). Quality becomes a shortcut, not a brake — whoever meets the rubric clears faster.
“It tramples my autonomy.”
The government decision steers only central-state investment; local autonomy stays intact (art. 120–121 of the Constitution). Nothing imposed on town halls without a law.
“It costs.”
In the government-decision phase: no new office, no new tax — only the existing budget and an independent fund. The real cost comes only with the law, and then from a fee already paid today.
If it doesn’t pass in time: even if the government decision is not signed within 18–24 months, the OAR standard, the Transylvania pilot and a ready-to-adopt document still stand. The path does not hang on a single political moment.
For the young architect: “design capacity” means portfolio and vision, not turnover — exactly the criterion that opens the door turnover closes for you today.
This document is synthesised from research into the architecture policies of dozens of countries, with primary sources and verification. Every mechanism links to its origin; figures are calibrated to the evidence (large, peer-reviewed studies are marked as such). Where a fact is an estimate or still being confirmed, it is flagged explicitly.
This is an independent prototype built to demonstrate the digital modernisation of OAR (the Romanian Order of Architects). It is not an official platform of the Order of Architects of Romania.
What we ask
The document is not an inert PDF. It is alive: you can read it article by article, amend and comment on it, and the public consultation becomes its very legitimacy. This is what a policy made by those who build looks like — not one made about them.
This page — and the document — were built as proof: we do what ought to be done, rather than merely ask that it be done.
A mayor of a small commune? Be a pilot — at no cost →
An architect starting out? A path that is yours too: teams are chosen on the project, not on the balance sheet.