Essay

Before the Border

Why a State Must Begin by Protecting the Political Community It Represents

Victor Ávila — August 2026Economist | Strategic Fiscal & Institutional Architect | AI-Era Public Finance, Productivity & Governance Transformation

Portugal has grown used to discussing immigration the way one discusses a door: whether it should be open, shut, or merely ajar. The metaphor arrives too late. By the time the conversation reaches the border, almost every decision that matters should already have been taken — which jobs the country intends to create, which wages it considers compatible with a decent life, how many homes it can supply, how many doctors and teachers it needs, what duties it asks of those who arrive. An open door in a house with no rooms is not generosity; it is improvidence. A shut door in an ageing house, with hospitals short of staff, is imprudent in equal measure.

The reasonable question is not whether we are "for" or "against" immigration. It is whether the country has a policy capable of selecting the flows it needs, limiting those it cannot absorb, and integrating with dignity those it admits. It does not. What follows is an attempt to design that architecture, with the caveat that should open any document of this kind: the sequence I propose describes the order of political deliberation, not a hierarchy of values. The Constitution is not a stage in the process. It is the boundary within which every stage takes place.

The figure that reframes the debate

Let us begin with the strongest objection that can be made to this essay: Portugal's migration growth has already slowed sharply.

Portugal had 11,424,031 residents at the end of 2025, of whom 1,597,539 were foreign nationals — 14.0 per cent of the population. But the population increase that year was only 36,809 people, against 330,587 in 2022. Among foreign nationals, the annual increase fell from 326,090 in 2022 to 59,113 in 2025, and net migration settled at 70,862 against a natural balance of minus 34,053.

It is worth being precise about what these figures from the national statistics institute do and do not say. A change in a stock incorporates entries, exits, deaths, naturalisations and registration adjustments; net migration is by definition a net concept. Neither allows us to claim that gross entries fell in the same proportion, because gross entries are not observable in these data. What is certain is that the exceptional expansion of 2022 to 2024 lost intensity.

A critic will say that proposing a containment architecture now is legislating for a crisis that has passed. But the slowdown coincided with the abrupt removal of the expression-of-interest route and with a period of procedural paralysis: it was braking, not steering. And braking is reversed by political decision or by circumstance, precisely because nothing in the legal order determines how many people the country admits, or against what capacity. It is the absence of a mechanism, not this year's number, that constitutes the problem of state.

The stock, moreover, has not gone away: between 2021 and 2025 the foreign population more than doubled, an increase of 849,384 people, and the pressure on housing, schools and services comes from the accumulated total, not from last year's marginal entry. And one fact deserves to be stated slowly: the 2025 estimates were the first produced entirely from administrative sources, and they forced a revision of the whole 2021–2024 series. Through the period of the greatest demographic transformation in Portugal's democratic history, the country did not know with any precision how many people lived on its territory.

What the numbers show and what they do not

The transformation has documented benefits. The Bank of Portugal recorded that average employment of foreign workers registered with social security grew 22.6 per cent in 2024, reaching 653,000 people out of 3,985,000 posts — 16.4 per cent of employees. Immigration widened the working-age population and the contributory base, and sustained sectors under strain.

Here the difficult part begins, and I prefer to state it as an open question. In 2023, median monthly earnings for foreign workers were €769 among the youngest and €781 among those over 35, against €902 and €945 in the corresponding groups of nationals. It is tempting to read wage compression into those figures. But they are raw medians, without controls for sector, occupation, tenure, qualification, region or firm — and the international literature shows that most of the observed gap in comparable situations is explained by composition, not by displacement.

There is, moreover, evidence in direct tension with the hypothesis. The Bank of Portugal reported consistently stronger wage growth in the lowest earnings brackets, and the statistics institute put the unemployment rate for the second quarter of 2026 at 5.3 per cent — the lowest since the series began in 2011 — with employment at a fifteen-year high. If the wage floor were being compressed by an excess supply of labour, one would expect the opposite.

I therefore take an explicit position: wage compression in specific segments of the Portuguese economy is a plausible and important hypothesis, not a demonstrated fact. What the data show with confidence is more modest and still significant: a substantial share of recent immigration was absorbed by the economy's wage floor rather than by its productivity frontier. Whether that reflects displacement of resident workers, deferred automation, or simply the sectoral structure of a low value-added economy is precisely what a serious measurement system would exist to determine. It is the difference between a political argument, which selects the reading that confirms the prior position, and a public policy, which builds the instrument capable of refuting it.

The same release contains, however, the figure that underpins everything that follows: youth unemployment, among those aged 16 to 24, remained at 18.3 per cent. A labour market at historic highs with nearly a fifth of young people out of work is not a market without internal slack. It is a market that is not mobilising the slack it has.

The order of responsibilities

A government does not administer humanity in the abstract. It governs a country, and it must protect, first, the political community that gave it its mandate — including the foreign nationals lawfully resident within it who, through work, taxes and family, are already part of its social reality. This priority is not a licence to discriminate: the Constitution enshrines human dignity, equality, and the extension of rights and duties to resident foreign nationals. It is not about creating an ethnic queue for dignity; it is about ordering the state's decisions by the interest of the population for which it answers.

Legal equality does not oblige the state to ignore that order. A hospital treats an emergency regardless of passport, but a government, in planning the hospital network and migration flows, has a duty not to deliberately create demand the network cannot meet. The first is a requirement of humanity. The second is an obligation of governance.

For "national interest" not to remain rhetorical, it must be converted into verifiable criteria — income per head, real wages, productivity, housing capacity, access to health and education, territorial cohesion, linguistic and civic integration. It then ceases to be a flag waved in the wind and becomes an account subject to audit.

The principle also demands intergenerational honesty. Immigration slows the decline of the working-age population, but it does not repeal the arithmetic of ageing: it buys time, it does not substitute for policies on birth rates, housing, skills and productivity. Giving priority to the Portuguese means, above all, refusing to treat the emigration of qualified young people as a fatality while their absence is offset by the arrival of poorer workers. A country that exports engineers and nurses, imports low-wage labour, and presents population growth as sufficient proof of success is confusing movement with progress. It is possible to fill the station and still let the most important trains depart.

Governed flows

The claim that all immigrants able to work are welcome sounds benevolent, but it is not a policy. The capacity to work is a characteristic of billions of adults, and no state has housing, services or productive jobs for all of them. Turning a near-universal human disposition into an admission criterion is installing a lock that opens with any key.

The relevant criterion is the country's concrete, demonstrated need. A "labour shortage" can, in fact, mean four very different things: genuine scarcity of specialised skills; a geographic mismatch between vacancies and workers; conditions so poor that residents sensibly refuse the post; or a business model that is only viable at wages that are too low. Authorising immigration identically in all four cases would be prescribing the same medicine for fever, fracture, fatigue and hunger.

From this follows a demanding labour-market test: an employer recruiting from outside the European mobility area should demonstrate that the vacancy was advertised with transparent pay, that conditions comply with collective bargaining, that resident workers were sought, that training or reskilling was considered, and that decent accommodation is available in areas under pressure. And from it follow differentiated limits: fast tracks for the highly qualified, circular arrangements in seasonal work, ceilings and specific contributions in low-productivity, high-turnover sectors.

The student route deserves its own note, because it is the typical blind spot of these systems. An international student is not a labour quota, but occupies housing, uses services and, in most cases, subsequently moves into the labour market. It was precisely that flow that broke Canada's capacity while the country believed it was controlling permanent migration.

One final point, and it is not a matter of opinion. Worker mobility between employers has ceased to be debatable: the European single-permit directive, applicable since May 2026, recognises the right to change employer, caps the initial tie-in at six months, and protects the permit during at least three months of unemployment. A visa dependent on a single employer turns an administrative border into economic servitude — and protecting the immigrant against exploitation is the most direct way of protecting the Portuguese worker against competition built on the artificial reduction of rights.

Reception capacity is not a sentiment

"Reception capacity" has become a comfortable phrase because everyone can hear it without knowing what it means. A serious state policy must take it out of rhetoric and put it into a balance sheet.

I propose an Annual Reception Capacity Budget, adopted with the seriousness of a financial budget, estimating region by region the new affordable homes, school places, health teams, transport capacity, validated labour demand and integration services. The limit would follow from the binding constraint — the tightest — and not from an optimistic sum of intentions: a bridge carries the load set by its weakest element, not by the average of the rest.

The accounting is where errors take hold. The capacity available for controllable admissions is the year's new capacity, less the accumulated deficit, less household formation within the resident population, less net internal mobility, and less non-contingentable entries — European free movement, international protection, family reunification — plus capacity released. It is not the net growth of the foreign population, which already incorporates exits and naturalisations and therefore systematically understates pressure. And the result should not be presented as an exact point but as a range with a central scenario and published sensitivity analysis: a ceiling that feigns precision it does not have loses credibility at the first challenge.

This rule has, however, a defect that would make it perverse if left uncorrected. If housing is the binding constraint along almost the entire coast, the rule produces admissions close to zero — including for those who would build housing and for the nurses who would relieve the health service. Capacity limits immigration; but capacity depends, in part, on immigrant labour.

The correction is a distinct category of capacity-creating flows, admitted outside the general ceiling: health professionals assigned to new teams in territories with demonstrated shortfall, teachers in posts unfilled by national competition, workers contractually assigned to affordable housing works or to infrastructure within an approved multi-annual programme. They do not count against the constraint they relieve, but they do count against the others — the nurse admitted occupies housing and her child occupies a school place. And they count with a lag: the construction worker occupies a home today and delivers homes in two or three years, so the exception must depend on a specific project, demonstrated additionality and a delivery schedule.

The resulting rule is simple to state and demanding to apply: reception capacity is not a fact of nature; it is a decision variable. A state that treats it as fixed turns planning into resignation.

Financing follows the decision

Those who obtain the economic benefit of migration cannot transfer to the taxpayer the entire cost of degraded housing, language learning, labour inspection and emergency assistance. If an activity asks the country to admit workers, it must share in the costs that make that admission sustainable.

The formulation sounds technical; it is a matter of honest accounting. A tomato, a parcel or a service can look cheap because part of the real cost was sent to an overcrowded room, to the municipality, or to the health system. The firm records efficiency; society receives the invoice off balance sheet.

Five instruments suffice: mandatory licensing of intermediaries, with verified fitness; joint and several — not subsidiary — liability of the ultimate beneficiary along the chain; liability of the Portuguese employer for the acts of the agents it contracts abroad, with recruitment fees charged to the worker prohibited; separation of the employment contract from the accommodation contract; and a refundable deposit per worker recruited, returned on proof of certified training, wage progression, compliant accommodation and absence of infringements.

I do not overstate the last. The deposit improves incentives; it does not eliminate economic incidence, since the employer can pass it through into starting wages, prices or informal intermediation. It must be calibrated to demonstrated incremental costs, with a ceiling proportional to pay, an express prohibition on deduction, and auditing — and the enabling legislation will have to settle its legal nature, incidence, limits and refund conditions, because an instrument halfway between a price, a fee and a guarantee does not survive litigation. Its defensible basis is the internalisation of identifiable incremental costs generated by recruitment; not the privatisation of financing for the health service, schools or public housing, which are general public goods funded by taxation.

That said, the predictable objection — that it raises costs in thin-margin sectors — describes precisely the intended effect. A sector whose viability depends on not paying the social cost it generates is not profitable: it is subsidised by third parties who never voted for that subsidy.

A hundred capacities, not one

Portugal does not have a reception capacity. It has hundreds, and the proof lies in the regional figures that the national debate ignores.

In 2025, foreign nationals were 27.9 per cent of the population of the Algarve and 22.6 per cent of Greater Lisbon, but 7.3 per cent in Madeira and 3.8 per cent in the Azores — the two lowest figures among the country's regions. Between the extremes there is a difference of more than sevenfold. To discuss "immigration in Portugal" as a single phenomenon is an error of scale comparable to discussing a continent's climate through the average of its temperatures.

The Azores had 245,328 residents at the end of 2025, of whom 9,333 were foreign nationals. The number looks small until one observes the trajectory: there were 3,772 in 2021, a rise of 147.4 per cent in four years. The composition differs from the national one — the Brazilian community accounts for 23.6 per cent of the total, well below the national 35.9 per cent, followed by the Cape Verdean, the American and the German.

In Madeira the pattern is sharper: 266,130 residents, of whom 19,371 foreign nationals, after a 161 per cent increase in four years, with Funchal alone holding more than half. And here is a figure that captures Portuguese demographic arithmetic better than any argument: in 2025, foreign residents increased by 1,779 people while residents of Portuguese nationality fell by 349. All of the region's population growth came from immigration.

From these figures follow conclusions the national architecture must absorb, and the first two exist to prevent the wrong reading of the rest.

Statistical headroom is not capacity. A share of 3.8 or 7.3 per cent indicates that the archipelagos are not saturated by comparison with the Algarve; it does not indicate that they have capacity available. Island capacity is constrained by other routes: a small housing stock contested by tourist accommodation in Funchal and Ponta Delgada, a construction sector of limited scale with costs inflated by freight, narrow and seasonal labour markets, and health and education networks running on minimum teams.

And priority is not suspended in the regions. The difference in stock between territories redistributes flows within a ceiling; it does not create a ceiling, nor does it dispense with the first principle. The order of decision is the same in São Miguel, in Funchal and in Setúbal: first one checks whether there are Portuguese nationals and legal residents — unemployed, underemployed, workers in reskilling, emigrants willing to return — who could fill the post with adequate training and compliant pay; then one establishes the binding constraint in housing and construction capacity, employment, health and education; only what remains defines external admission. A region with little immigration and high youth unemployment does not have headroom: it has a domestic employment problem that immigration does not solve and may aggravate. In the archipelagos, where the departure of qualified young people is structural, admitting from abroad while exporting one's own working population would be repeating at regional scale the error this essay attributes to national policy.

Insularity also alters the parameters and not merely the values: housing supply does not respond with mainland elasticity, and demand for specialised services is indivisible — an island hospital needs a full team even with a small population, and a single anaesthetist can be the difference between an operating theatre open and one closed. The regions should certify their own capacity and distribute the sub-quota internally; sovereign decision over admission and borders remains with the organs of the Republic.

One thing remains to be said that no Portuguese document on migration usually says: the island diaspora is part of the solution. No other part of the territory has an emigration so dense, so organised and so close in language and family ties. To ignore the qualified return of the descendants of emigrants is to search far away for what one has near at hand, and at greater cost of integration.

Constitutional integration

A democracy does not require of those who arrive that they admire every local custom. It requires something more limited and more serious: respect for constitutional law, for individual liberty and for equality between women and men. Multicultural integration is sustainable only where there is a normative centre that does not change with each community — and that centre is the Constitution.

Pluralism is compatible with different beliefs, dress, cuisines and freely chosen family structures. It is not compatible with forced marriage, genital mutilation, domestic violence, religious coercion, withdrawing girls from school, or the persecution of homosexuals. The boundary of tolerance is the right of the concrete person: a liberal community welcomes many cultures, but not that any of them build private territories where the Constitution ceases to apply. The rule holds for everyone, including the Portuguese.

From this follows an integration compact, signed on the granting of the permit and renewed with it. It is not an oath of cultural obedience, but a written undertaking of determinate content — and its scope must be defined precisely, lest it appear to condition rights that are not at the state's disposal. It applies to discretionary residence permits for third-country nationals provided for in national law, with adaptations for age, disability and incapacity, and without prejudice to the distinct regimes of European free movement, international protection and family reunification. The state's side of the reciprocity, in turn, cannot be programmatic: decision deadlines, effective language provision, recognition of qualifications and protection against exploitation must be enforceable rights, with consequences for an administration that fails to deliver them.

Whoever signs undertakes to respect the Constitution and the law; to recognise fundamental rights and equality between men and women as norms admitting no exception of origin, religion or custom; to ensure their children's compulsory schooling, daughters included; to meet tax obligations; to learn Portuguese; and — in the clause that replaces any generic formula — not to use violence, threat, abuse of authority, or economic or family dependence to prevent another person from accessing the Portuguese courts, attending school, working, receiving health care, choosing or leaving a religion, marrying freely, or exercising the rights recognised by the Constitution and the law.

The drafting is deliberate. The obligation bears on conduct — using violence, threat, abuse of authority or dependence to prevent someone from exercising a right — and not on convictions, nor on the right to argue publicly for changing the law. This is not a concession: it is what makes the clause enforceable, and it is what reaches exactly what is at stake, which is not anyone's belief but the subjection of third parties to a rule that is not Portuguese law. What criminal law, family law and child protection already treat as unlawful when the author is Portuguese becomes explicit at the moment of admission. The compact creates no special regime for foreigners: it makes visible what applies to everyone.

Sanctions flow from the law, not from the compact. The instrument retains two effects of its own, both ordinary. The first is revocation for falsehood — but only where it bears on objective facts determinative of the grant, or where there is sufficient proof that the undertaking was given fraudulently at the moment of the declaration. Subsequent non-compliance does not, in itself, permit any presumption of initial falsehood, and the distinction is indispensable: without it, an administration would convert any breach into a retroactive fiction of fraud. The second is that failure to meet the civic and language obligations conditions progression to a more favourable status, which is a condition for granting a benefit and not a sanction.

Two admission conditions accompany the compact. The first is a criminal record check, with proportionate assessment — serious, violent and sexual crime, trafficking and documentary fraud are relevant; a traffic offence is not — and verification at defined moments. The European ECRIS-TCN system will allow automated identification of convictions recorded in other Member States; it does not, however, replace certificates issued by third countries of previous residence, and one should not promise automation the system does not provide.

The second is initial health cover, on those routes where the law permits or requires insurance, during the phase preceding effective registration with the National Health Service. This point requires care, because it is easily misread. It is not a public-charge test and nothing here turns on diagnosis. Entitlement to the National Health Service in Portugal derives from lawful residence and not from contributions; the requirement concerns only the financing of foreseeable costs in a transitional phase, does not condition urgent care, and does not exclude people with chronic conditions. Excluding by diagnosis would be contrary to the Constitution and to the Convention on the Rights of Persons with Disabilities, unworkable — "chronic condition" covers diabetes, hypertension and asthma — and self-defeating, since it would turn away the diabetic surgeon the region needs. Where recruitment is employer-driven, the cost falls to the employer.

The language obligation deserves a ladder rather than a single threshold: demonstrated participation and progress at temporary renewals, level A2 for permanent residence, level B1 for citizenship — which requires a separate amendment to the Nationality Act and cannot be introduced administratively. Time runs from the date on which the state actually made a course place available, not from the date of entry: without that safeguard, the immigrant is punished for the administration's waiting list. And the consequence of non-compliance is non-progression, not removal. Language thus ceases to be an examination that expels and becomes the condition of a citizenship that is earned.

None of this applies retroactively. The conditions govern applications made after entry into force, not those already lawfully resident nor pending cases. A state that demands predictability of those who arrive should begin by offering it to those who have already arrived.

As for the second generation, the question is simpler than the debate suggests: children born and schooled in Portugal are, in fact if not in law, daughters and sons of this community — and the discrepancy belongs to the state, not to them.

Measuring in order to govern

Counting entries is necessary, but it amounts to judging a hospital by the number of people who walked through the door. Administrative statistics tell you how much movement there was; they do not tell you whether the system produced health.

The predictable error would be to answer a lack of measurement with a catalogue of indicators. A panel of fifty promised metrics is worth less than twelve published ones — and no indicator should enter the system without an identified source, a fixed frequency and a responsible body. Twelve suffice: hourly pay and wage progression by entry cohort, employer mobility, value added per hour and investment per worker in sectors under quota, rent burden and overcrowding, employer-provided accommodation inspected, patients without a family doctor, class sizes and the share of pupils with Portuguese as a second language, language certifications as a proportion of those required to obtain them, naturalisations and second-generation school outcomes, and — the two that are always missing — administrative backlogs and the execution of removal decisions.

Five methodological disciplines separate measurement from propaganda: a baseline established before the decision, so that the government cannot choose the starting point afterwards; analysis by entry cohort; comparison between like persons and like sectors, without which raw medians continue to govern the debate; distributional measurement, which asks who gains and who loses; and counterfactual assessment.

And the system needs an independent review rule: after three years, an external body would assess whether each route produced the announced effects. Sectors using immigration to expand output, pay better and invest would keep access; sectors with accumulated infringements, wages at the minimum and low productivity would see quotas cut. Immigration would cease to be an invisible subsidy to fragile business models.

The prior question: deciding and enforcing

All of this presupposes what Portugal has not had for a decade: a state capable of deciding within the deadline and enforcing what it decides.

Quotas are worthless in a state that cannot process. An annual quota presupposes decisions within the year to which it relates; if a decision takes two or three years, the limit becomes an accounting fiction laid over a queue. And the queue itself creates a de facto population — people present, working, renting, enrolling children, with their legal situation undecided. The state that does not decide ends up regularising by omission what it refused to authorise by decision: the most expensive form of migration policy, with all the costs of admission and none of the benefits of selection.

Nor are they worth anything in a state that cannot enforce. A selective system is credible only if refusal has consequences. Where the execution of removal decisions is residual, the signal sent to applicants, intermediaries and employers is that the operative criterion is not authorisation but presence.

I advance no figures here on the current backlog, and not through carelessness: as at the closing date of this essay, no public, monthly and sufficiently disaggregated series could be located that would allow the backlog to be quantified with rigour by case type and age. That is precisely the argument.

What Portugal can decide alone

The question of this architecture's legality has a more favourable answer than the debate assumes. Article 79(5) of the Treaty on the Functioning of the European Union expressly preserves the right of Member States to determine volumes of admission of third-country nationals coming to their territory in order to seek work. Setting an annual ceiling is not an eccentricity in tension with European law: it is a competence the Treaties deliberately reserved to states.

Three limits are insurmountable: free movement of Union citizens, to which no quota applies — less constraining than it sounds, given that 89.5 per cent of the foreign population is non-EU, but absolute; family reunification, which the European directive frames as a right and not a discretionary faculty; and long-term resident status, whose integration conditions are subject to proportionality review.

There remains what the Portuguese debate almost never mentions: the CPLP Mobility Agreement, and, in the Brazilian case, the Treaty of Friendship of 2000. The CPLP is the Community of Portuguese Language Countries — Portugal, Brazil, Angola, Mozambique, Cape Verde, Guinea-Bissau, São Tomé and Príncipe, East Timor and Equatorial Guinea — and its 2021 mobility agreement, together with the equal-status regime Portugal and Brazil established in 2000, is the instrument that most conditions the viability of any quota. It is almost absent from the discussion. It is of variable geometry, so national configuration is possible without denunciation — but any change carries an asymmetric diplomatic cost, because within that community Portugal is simultaneously a destination and a middle power.

The reasonable conclusion is not to restrict the Lusophone route. It is to demand of it what makes it valuable. A predominantly Lusophone flow substantially reduces, though it does not eliminate, the language barrier this essay identifies as the infrastructure of citizenship. The CPLP route should be the highest-quality route in the Portuguese system, not its tradesman's entrance — with swift recognition of qualifications, mobility between employers, verified progression and a predictable path to permanent residence.

As for territorial distribution, it meets a serious constitutional limit: Article 44 guarantees the right to move and settle anywhere in the national territory. A permit legally binding a person to a municipality would be of very doubtful constitutionality. Territorialisation must work through incentive and supply — sub-quotas where capacity allows, admission tied to contracts with verified accommodation, deposit revenue assigned to the destination territory, municipal funding based on effective residence — and it must publish its retention rates. An instrument of territorialisation that does not publish its retention rate is not a policy; it is an intention.

The international mirror

Three examples suffice, each for a distinct reason.

Denmark is in the process of adopting an admission route grounded in collective agreement, open to certified firms and to a fixed list of third countries, with a lower pay threshold — and usable only while unemployment stays below a defined level. It is the design closest to what is argued here: a trigger tied to the domestic labour market, which closes the route without political decision, without negotiation and without reputational cost. The bill lapsed with the dissolution of parliament and was reintroduced in June 2026, still unapproved.

Canada demonstrates that even an experienced state corrects targets when material capacity diverges from demographic ambition. The 2026–2028 plan cut new temporary arrivals to 385,000 in 2026 and 370,000 in each of the following two years, against earlier projections above half a million, after recognising that limiting only permanent migration was controlling the front door while leaving the windows open.

And Spain offers the closest warning and the cleanest statistical lesson. The extraordinary regularisation process that closed in June 2026 received 1,174,978 applications, with 262,475 regularised already registered with social security a month later. In that month of July, affiliation of foreign workers rose by 66,045 while affiliation of Spanish nationals fell by 24,318, and registered unemployment rose by 19,517 — a rise the Ministry of Labour itself attributes in part to the regularisation, because those regularised became able to register at employment offices. The same process simultaneously increased registered affiliation and registered unemployment, without those movements permitting any measurement of an equivalent change in the real volume of work. There is no cleaner demonstration that administrative statistics are not economic measurement. And a large-scale regularisation, repeated, sends a message about the expected value of irregularity: it treats the stock and, without control of flows, recreates the next one.

None of these countries has solved the equation. But no consolidated democracy regards the sentence "there is a firm willing to hire" as sufficient. Employer demand is indispensable information, not a sovereign mandate: a factory knows how many operators it needs next quarter, not how many homes, teachers and transport links the decision will require over a decade. To govern is to bring together the accounts that the market, quite legitimately, keeps apart.

Who gains, who loses, and what the balance is

A governance proposal is serious only when it says whom it serves, whom it costs, and what is left for the country.

Gaining are Portuguese nationals and legal residents on low incomes, because internal recruitment is tested before any external admission and competition built on the artificial reduction of rights becomes impossible. Gaining are admitted immigrants, who receive legality from day one, decisions within deadline, funded language learning, recognised qualifications, mobility between employers and a predictable path. Gaining are compliant firms, which acquire annual predictability and stop competing with those who do not pay the social cost they generate. Gaining are municipalities and the autonomous regions, funded by effective residence rather than by contract location, with a voice in certifying their own capacity. Gaining are the social partners, because collective bargaining becomes the wage filter for admission. Gaining is the exchequer, which trades annual contributory balances for a lifecycle account and the expensive cycle of regularisations for decisions on time. And gaining is the second generation, whose legal trajectory ceases to be indefinite.

Losing are the sectors whose viability depends on not paying the social cost they generate, because the deposit and the wage filter make that cost visible. Losing are opaque intermediaries, who will need licences, fitness and joint liability. And losing is short-term political convenience: a government that adopts an annual capacity budget and a panel of twelve indicators becomes obliged to demonstrate what it can today simply assert.

The expected balance is favourable, though unquantified — and the distinction matters in an essay that has just declared it did not calculate the lifecycle fiscal account. The claim is not that any immigration produces a positive balance. The claim is that a system which selects, holds to account, measures and corrects is more likely to produce an economic, social and institutional result superior to a regime governed by administrative accumulation.

Nor is it free. There is a real administrative cost; some sectors contract and some prices rise — intensive agriculture, hospitality, care work, part of construction; and there is an execution risk that should be stated before others state it: if enforcement does not keep pace, part of the activity migrates into informality, and the country is left with the costs of the system and none of its benefits.

The favourable judgement rests on three grounds. Productivity, because making the cost of cheap labour visible shifts business decisions towards investment and wages, which is precisely the transition the Portuguese economy has not made. The fiscal ground, deferred, because a country that decides and enforces stops paying for periodic regularisations and for a legally undefined population. And the institutional ground, the most valuable: a migration debate with a baseline, cohorts and independent review deprives radicalism of its best resource, which is the absence of proof.

And there is something the balance is not. This architecture does not predetermine whether Portugal will admit more or fewer people than it does today. If measurement shows no displacement effects and capacity is expanded, ceilings rise; if it shows the opposite, they fall. What it determines is that the number, whatever it is, becomes defensible before those who bear it.

The choice that begins before the border

The order of principles is decisive. First the political community and the national interest — a priority that precedes any other criterion, including territorial distribution. Then flows, selected by real needs. Then reception capacity, treated as a decision variable and not as a fatality. Then financing that follows the decision, constitutional integration, and measurement. And, once more: this sequence describes the order of deliberation, not the hierarchy of values. Fundamental rights and equality between men and women are not a stage in the process; they are the frame within which the remaining stages are legitimate.

This architecture does not diminish those who arrive: it prevents the immigrant from being reduced to a unit of labour, a fiscal contribution or a demographic solution. The disordered border is often presented as compassion, but it delivers the newcomer to the informal landlord, the opaque recruiter and the employer who knows that their legality depends on him.

Nor is there social justice in asking the Portuguese on the lowest incomes to bear alone the pressure of a policy whose benefits are concentrated elsewhere. Protected urban professionals celebrate abundant labour in restaurants, deliveries and care; property owners benefit from higher demand. Low-income tenants face the same scarce supply, and patients compete for the most congested services. If there is also an effect on the wages of comparable workers, that is a matter to be established with data and not decreed by conviction — but a policy that refuses even to examine distribution turns the cosmopolitan virtue of some into the material invoice of others.

The real test is not whether immigration is good or bad. Such a question has the useless elegance of asking whether rain is good or bad: it depends how much falls, where, when, on what soil and with what drainage. The difference is called architecture.

That is the architecture that must be built before the border, and now, precisely because growth has slowed: a containment system is built when pressure eases, not in the middle of the emergency. A migration policy worthy of the name does not measure its humanity by the number of entries, nor its firmness by the number of refusals. It measures them by its capacity to turn a sovereign decision into a life that is legal, productive and free — without sacrificing, along the way, the common house that made that life possible.

This essay summarises a longer working paper, available in Portuguese, containing the technical basis, the full indicator framework, the legal feasibility analysis and an accounting prototype of the capacity budget.

Principal sources: Statistics Portugal (INE), Resident Population Estimates 2025 and Labour Force Statistics, Q2 2026; Bank of Portugal, Economic Bulletins of June 2024, March 2025 and June 2026; DREM and SREA, regional estimates for 2025; Eurostat, migrant integration housing statistics; OECD, International Migration Outlook 2025; Directive (EU) 2024/1233; Directives 2003/86/EC and 2003/109/EC; TFEU, Article 79(5); Government of Canada, Immigration Levels Plan 2026–2028; Spanish Ministry of Inclusion and Ministry of Labour, July 2026 data; Congressional Budget Office (2024); Card (1990), Borjas (2003, 2017), Peri and Yasenov (2019), Dustmann, Schönberg and Stuhler (2016, 2017).