Opening
The International Monetary Fund concluded, on 24 June, another Article IV consultation, and the report it produced confirms, with the composure of documents that no longer expect to surprise anyone, what successive diagnoses had been signalling: Portugal has recovered stability, posted a third consecutive budget surplus, and brought public debt below ninety per cent of output — and yet remains bound to the same structural limits as ever, with living standards below the European average, stagnant productivity, an ageing population, and a potential growth rate that stubbornly refuses to keep pace with the country's ambitions.1
The diagnosis is, as usual, impeccable. It lacks only the one virtue diagnoses are not meant to possess: that of transforming what they describe. Portugal does not suffer from a shortage of diagnosis; it suffers, rather, from a certain ease in the art of diagnosing itself — it produces reports with the competence other countries reserve for reforms, and too often mistakes the lucidity of the portrait for the beginning of the cure.
Nor should one suppose the Fund stops at the portrait: it prescribes too, and generously. It calls for reforms to raise productivity, support for innovation and the growth of firms, less bureaucracy, more flexible product and labour markets, better education and — a sign of the times — that the country prepare for the diffusion of artificial intelligence and for a more efficient allocation of resources. These are sound recommendations, and none of them surprises. But to observe that productivity ought to rise is still diagnosis, and to name the desired reform is still second-generation. What the report does not say — because it is not its place to say it — is how one builds an instrument that does these things and goes on doing them, measuring itself and correcting itself as it runs. Between knowing what is needed and designing what secures it lies, precisely, the distance that separates diagnosis from architecture.
Public policy learned, over the last century, first to diagnose problems and then to design solutions for them. What the twenty-first century demands of it is a third gesture, harder than either: to design the economic architecture within which those solutions are to operate — a structure that is not merely evaluated before it is born, as good practice requires, but that measures, simulates and corrects itself while it runs.
This is not, let it be clear, to promise that a single mechanism will resolve the contradictions the Fund enumerates. It is to show, in one concrete and circumscribed case, what it means to design for efficiency rather than to compensate inefficiency — precisely the gesture the report calls for without knowing how to name it, and which so rarely allows itself to be translated into engineering.
The Hybrid Territorial and Talent Continuity Mechanism (MCTH) was born small, and from a modest problem: that of mobility between the mainland and the Autonomous Regions. It was during its design that it became clear the real problem was not to reform an instrument of compensation, but to redesign the entire economic architecture of territorial continuity — to exchange a system that pays declared costs for one that recognises only efficient costs, computed, audited and shared.
What is presented here is not, therefore, a proposal for the aviation sector. It is a concrete economic architecture — the first of a family I shall, elsewhere and at greater leisure, call Productive Architectonic Economics. This article does not formalise that discipline; it demonstrates one of its instances. And perhaps it is enough, for now, to show that a public policy can cease to be an administrative mechanism of compensation and become a measured, simulated and audited economic structure — not for air mobility alone, but for the way in which, in this century, we are to conceive public policy.
1The three generations of public policy
It should be said, first of all, that the three generations spoken of here are not a law of history, nor a chronology I propose to defend against anyone who might contest it with dates. They are a way of reading — a convenient ordering of capacities that, in practice, coexist and overlap more than they succeed one another. To take them for rigorous stages of an evolution would be precisely the kind of pretension this article does not have. But as a way of grasping where we stand, and what we still lack, the reading serves.
The first capacity governments learned was to diagnose. They learned it well, and Portugal better than almost anyone: the Fund's report with which I opened this text is the perfected form of that knowledge, a lucid, honest and almost moving portrait in the precision with which it enumerates ills it has been describing for twenty years without a single one having allowed itself to be cured by the description. Diagnosis is a virtue; among us it has become almost a vice — we produce reports with the competence other countries reserve for reforms.
The second capacity was to design solutions. Governments learned to respond to the problems they could describe, and did so with instruments: programmes, incentives, subsidies, mechanisms of compensation. To this generation belongs, strictly, almost everything we today call public policy — and to it belongs, as we shall see, the old system of compensation for mobility in the Autonomous Regions, with its simple logic of paying after the fact the cost that someone declares to have borne.
The temptation, on reaching the third step, is to define it as the capacity to measure and simulate before implementing — and that would be a mistake, because that capacity is nothing new. Regulatory impact assessment has been common practice for decades; cost-benefit analysis is older than those who invoke it; and the design of incentive structures ex ante, far from being an invention of our century, earned a Nobel in economics for the man who formalised it more than fifty years ago. Anyone presenting simulation as a novelty would merely be revealing an ignorance of the history of their own discipline. It is not, then, in measuring or in simulating that the third gesture resides.
It resides elsewhere, more discreet and more difficult. In the classical arrangement, measurement and simulation are external to the policy: they are done beforehand, in a study a separate office drafts, files and archives, and the policy, once approved, departs to live its administrative life forgetful of the model that preceded it. Evaluation and mechanism divorce — one happens in a report, the other at a counter, and they rarely speak again. The third gesture consists in dissolving that divorce: in making measurement, simulation and optimisation not a prior study but permanent organs of the mechanism itself, recalculating as it runs. The policy ceases to be evaluated once and then executed; it comes to execute itself by measuring itself. The architecture does not precede the policy — the architecture is the policy.
This is the difference that separates a second-generation instrument from an economic architecture. The instrument compensates and administers; the architecture compensates, measures what it compensates, simulates what it would compensate under other conditions, adjusts what it is to compensate, and does so in a closed circuit, without evaluation ever detaching from execution. This is exactly what the MCTH does with its two synchronised modules — the operational one, which settles in real time, and the strategic one, which projects the envelope and sustainability — feeding one another so that neither is ever an archived study.
I do not claim the MCTH inaugurates an era; I claim, more modestly, that it is a concrete instance of this third way of conceiving public action. For now it is enough to show the instance; the theory will come when it deserves its own document. What matters here is the mechanism, and what the mechanism demonstrates by working.
2From the SSM to the MCT and the MCTH
To measure what the MCTH does that is new, it is worth looking first at what it came to replace — and best to do so without condescension, for the instrument that preceded it was not ill-conceived for its time; it was merely conceived according to a logic that time has since exhausted.
The Social Mobility Subsidy, created in 2015, solved a real problem with the second-generation tool par excellence: compensation after the fact. The resident bought the ticket at the market price, bore it in full, and then received from the State the reimbursement of the difference between the eligible cost of the passage and a maximum fare fixed by ordinance. It was simple, it was just in intention, and it had the virtue and the vice of all after-the-fact compensation: it paid whatever had been charged, without ever asking whether what was charged was what ought to have been charged.
The 2026 reform — Law 23, which rechristened the subsidy as the Mechanism of Territorial Continuity — touched, with a generous hand, everything but the essential.2 It changed the name, digitised the process on gov.pt, did away with proof of contributory standing and the presentation of receipts, and promised, for the end of the year, that the passenger would no longer advance the full price but pay only their share at the point of purchase. In matters of administration, it was a serious reform. In matters of economics, it was no reform at all: the MCT goes on reimbursing the declared cost, exactly as the SSM did, and the one question that mattered — whether that cost is efficient — remains unasked.
There was, moreover, a gesture that amounts to a confession. In reforming, the State eliminated the ceiling on eligible cost — the four hundred euros of Madeira, the six hundred of the Azores — which was, however crude, the only cost discipline the system possessed. With the ceiling removed, the treasury bound itself to reimburse the difference without limit, whatever the price charged. The President of the Republic himself, on promulgating the law, warned that removing the ceiling might have effects calling for careful regulation and exacting oversight3; and the reason for the warning lies in the perversity of the design: with no anchor in an efficient cost, every euro added to the commercial price converts mechanically into more public compensation — and what this creates is not the certainty of an abuse, but an incentive structure that rewards the high price. It is the second generation carried to its logical conclusion: to compensate more generously without compensating more intelligently, and to discover, too late, that generosity without discipline is merely an invitation to another's inefficiency.
It is precisely here that the MCTH breaks — and the break is not where it first appears to be. It is not in the resident's paying only the fare, nor in the State's settling directly with the operator: in that the MCT itself is heading, and rightly, for it is a matter of payment logistics and not of the nature of the expense. The break is in what the State pays. The MCT pays the cost declared to it; the MCTH pays the cost shown to be efficient. Where the Mechanism of Territorial Continuity removed an arbitrary administrative ceiling and replaced it with nothing, the MCTH replaces it with a technical ceiling — not a figure set by rule of thumb, too high for some and too low for others, but a recognised price that tracks what the route ought, efficiently, to cost. The subsidy compensated inefficiency; the architecture refuses to finance it.
The difference between the MCT and the MCTH is not, therefore, one of administration — it is one of generation. One is the second generation in its final refinement, more convenient, more digital, more generous, and blind exactly where it has always been blind. The other asks what the first never asked. How that technical ceiling is built, and how it is ensured it recognises efficient cost without decreeing it, is what the economic architecture of the model must explain next.
3From compensation to economic architecture
The architecture begins with a gesture of separation. In the old system, price and compensation were a single quantity seen from two angles: the State reimbursed a fraction of what the passenger had paid, and the market price governed everything — what the resident advanced, what the treasury returned, and the margin the operator took in between. The MCTH severs what was fused. On one side, the commercial price, which the operator sets as it sees fit, in the market, at its own risk; on the other, the public compensation, which the State ceases to calculate from the price charged and comes to calculate from the cost it recognises. With the two quantities separated, the market price no longer commands public expenditure: the State no longer pays a fraction of what was charged — it pays a figure it determines itself. It remains to be seen against what that figure is determined.
The point on which the whole architecture turns — and at which a sceptical reader would dismantle it with a single question — is this: how does one compute the efficient cost of a route served by one, at most two, operators? With no competitors to compare against, a "sectoral benchmark" risks being an elegant word for an arbitrary number.
The MCTH's answer is not to compare the operator with rivals that do not exist, but with its own idealised version. The eligible reference cost is neither declared by the operator nor decreed by the administration: it is built from the bottom up, item by item, from observable quantities largely external to the accounts of the one providing the service. Fuel enters as the normed consumption of the reference aircraft multiplied by the market price; navigation and airport charges, by published tariffs; crew, by collective-agreement tables and normed ratios; lease, by market rates for the aircraft type; maintenance, by the cost per block hour (the hour during which the aircraft is in operation, from departure stand to arrival stand) of the same type; the administrative structure, by a capped allocation. To this route-common cost is added, when — and only when — proven, an individual adjustment; absent proof, the adjustment is nil.
Upon this efficient cost sits a margin — the return the operator is entitled to — but a normed margin, fixed in advance, neither negotiated case by case nor indexed to what was spent. The difference is deeper than it appears. When the margin falls on the declared cost, as in the old cost-plus regulation, the operator that spends most earns most, and the perverse incentive installs itself of its own accord: to inflate expenditure in order to inflate profit. By fixing the margin on the efficient cost rather than the declared one, the MCTH reverses the sign of the incentive — the only way to earn more becomes to operate better, because the base on which the margin is computed has ceased to reward waste and come to reward efficiency.
The effect is as simple as it is decisive. The items that weigh most in the cost — fuel, charges, lease — do not live in the operator's books; they live in the market, and the market does not let itself be inflated for convenience. Efficient cost thus ceases to be something the operator demonstrates and becomes something the regulator computes. And the incentive to inflate declared costs, which corrodes any compensation resting on self-declaration, loses its target: there is no longer a declared price or cost capable, by itself, of determining the compensation. The operator still reports costs, volumes and adjustments — but reports them within an audited and normatively bounded architecture, where no declaration, on its own, sets what the State pays.
It is in this, and not in any sharing formula, that the true break resides. The State ceases to compensate the cost declared to it and comes to recognise the cost shown to be efficient. Compensation is not abolished — to promise that would be naïve; it is rebuilt on a foundation the operator does not control. It remains to be seen who governs that foundation, and it is there that discretion, expelled from the costs, will reappear — displaced, but not suppressed. With that the institutional architecture will deal, further on.
4The revolution in financial settlement
With the efficient cost determined, it remains to carry it from paper to counter — and it is in the way the money circulates that the MCTH makes its second break, less visible than the first but, for the traveller, more immediate.
Under the regime the MCTH replaces, the resident was, without having asked to be, the system's banker. They bought the ticket at the market price, advanced it in full — five hundred euros, six hundred, whatever the season and demand dictated — and then waited for the reimbursement of the difference, weeks or months, against receipt and proof of attendance. The State compensated, but late; and the resident financed, in the meantime, with their own money, an expense that was not theirs. It was a system that asked the weaker to advance money to the stronger.
The MCTH undoes this absurdity at the root. The resident pays only the territorial-continuity fare — a regulated amount, uniform for the route and known in advance, unrelated to the swings of the market price — and nothing more. They do not advance, do not accumulate receipts, do not wait. The compensation does not pass through their hands: the State settles it directly with the operator, and the passenger leaves the financial equation the moment they board.
Here a measure of honesty is owed, lest the article claim as its own what already belongs to many. The convenience of paying only the fare at the point of purchase is not the MCTH's invention: the Mechanism of Territorial Continuity is itself heading there, promising the point-of-sale discount by the end of 2026. If the revolution were only that — sparing the resident the advance — the MCTH would merely be arriving first at a destination the State is also seeking. The difference, as always in this model, lies not in what the passenger sees, but in what happens beneath.
Beneath, the MCT will go on settling against the declared cost: it will discount at the counter, yes, but it will discount a fraction of whatever price the operator charged. The MCTH settles against the efficient cost determined in the previous section. The gesture, at the counter, is the same; the quantity settled is another. One discounts what was charged; the other pays what ought to have been charged. And it is for this reason that, in the MCTH, operational settlement can be done at the point of sale and by objective rules: because the amount settled there does not depend on an invoice to be checked, but on parameters the system already knows at the instant of sale. That reckoning is not, however, final at that instant — it is followed by an audited reconciliation, in which costs, volumes and adjustments are confirmed and squared; what disappears is not the audit, but the resident's wait for it.
This last property is more than operational elegance. When compensation is computed and settled in the moment, the customary distance between the expense authorised and the expense discovered disappears. The State knows, with every ticket, how much it has spent and how much remains — and public expenditure, which in reimbursement systems is reckoned only at the year's close, becomes legible as it happens. It is this real-time legibility that opens the door to what the next section addresses: a system that does not merely pay, but watches itself pay, and makes of that watching an instrument of government.
5A public policy founded on simulation
It was said, in closing the previous section, that the MCTH watches itself pay. It is time to show how — for it is in that watching, and not in the ingenuity of the calculation nor the convenience of the counter, that what distinguishes an architecture from a mere instrument resides.
Recall the thesis with which this article opened. The novelty of the third generation lies not in measuring or in simulating — economics has been capable of that for decades — but in making measurement and simulation not a prior study to be archived, but a permanent organ that lives within the mechanism and recalculates as it runs. The MCTH gives that idea a concrete body: an integrated architecture of economic simulation, made of two synchronised modules that do not succeed one another in time — they coexist, and speak to each other without cease.
The first is the operational module, the one we have already seen at work. It is what fixes the resident's fare, determines the efficient cost, settles the compensation with the operator, and shares the efficiency gains by the agreed rule. Every ticket it processes is at once a transaction and a datum: in settling, it measures; in paying, it records. The operational module is not merely the arm that executes — it is also the sense that gathers.
The second is the strategic module, the one that looks at what the first has gathered. It takes the data of the operation and projects with them the financial envelope, the budgetary forecasts, the alternative scenarios — what would happen if demand rose, if fuel surged, if a route gained or lost traffic — and watches, over all of it, the system's sustainability across the multi-year horizon. It is here that the prudential execution trigger resides: not a fixed threshold that fires at ninety per cent of the envelope, but a watch that compares cumulative expenditure against the expected budget trajectory and, when a material deviation opens between the two, decomposes it — into volume, cost, traffic mix and efficiency — before any correction. To spend ninety per cent of the budget in November may be normal; to spend it in July is not, and it is that difference the trigger reads. Of its nature — a lookout that warns, not a helm that steers — the next section will treat; for now it is enough to record that it exists and that it watches.
The essential, however, is not what each module does in isolation, but the fact that they work in a closed circuit. The operational feeds the strategic with the real — what was sold, what was settled, what was saved; and the strategic returns to the operational the probable — the projections and limits that are to calibrate the decisions to come. One measures the present, the other rehearses the future, and the boundary between them is in perpetual crossing. Nowhere in this design is there the classical moment at which evaluation ends and execution begins: one evaluates by executing, executes by evaluating. And lest this remain a matter of intention — the architecture is built, in two synchronised simulators that reproduce, piece by piece, the circuit described here.
It is in this synchrony that the most ambitious sentence of this article ceases to be rhetoric and becomes description. The architecture, we said, does not precede the policy — the architecture is the policy. In the MCTH one sees why: remove the simulation and what remains is not a simpler mechanism but a blind one, able to pay but unable to know whether it can go on paying. Measurement is not an accessory bolted onto the system; it is the system coming to be aware of itself. And a public policy that knows itself while it acts is, by definition, a different thing from what we were accustomed to call public policy.
6The architecture of incentives
All the engineering described so far — the computed cost, the normed margin, the point-of-sale settlement — serves, in the end, a single purpose: to reduce public expenditure structurally, not by decree nor by cuts, but through efficiency. And efficiency, unlike cuts, cannot be imposed; it must be cultivated. Someone must have an interest in being efficient. This whole section is about how the MCTH manufactures that interest.
Begin by separating two things subsidies tend to confuse. The first is the operator's remuneration — the efficient cost recognised to it, plus the normed margin that pays for capital and risk. That remuneration covers the efficient cost and a reasonable return on capital and risk — but it is the efficient service it pays for, not the badly run one: controllable inefficiency above the benchmark stays with the operator, not the treasury, save a recognised exogenous shock. What is guaranteed is not the recovery of any cost, but the remuneration of a service conducted as it ought to be. The second is the sharing of efficiency gains — and this exists only when the operator manages to provide the service for less than the recognised reference cost. They are two distinct mechanisms, and keeping them distinct is what prevents the operator from capturing, by way of margin, what ought to be a productivity gain.
When that gain appears, it is shared by a deliberate key: of every euro saved, sixty cents lower the resident's price, thirty relieve the treasury, and ten reward the operator that saved. The proportion is not arbitrary; it is a declaration of principles. In a system sustained by public money, the greater part of the fruit of efficiency must return to those who fund it — to the citizen who travels and the State that compensates — and only a fraction, sufficient but modest, stays with the one who generated it. Ten per cent seems little; it is, however, pure reward, added to the remuneration already recognised to it, so that the operator never loses by being efficient and always gains something by being so. It is enough to keep the incentive alight without handing to a private party what belongs to the community.
And it is precisely this measure that avoids the classic criticism of transport support: that all the productivity wrung from a public system ends up in the operator's pocket. Here it does not. More: the key aligns, at one and the same time, three agents that traditional instruments set at odds — the resident wants the low price, the State the low expense, the operator the profit, and all three come to want the same thing, which is efficiency. It is rare to find a design in which each one's interest serves the interest of all; that is the small miracle of the 60/30/10 key.
Even so, the key must not be taken for an automatism. The 60/30/10 is the general configuration, not a vested entitlement of the operator: the generator's share exists only where the saving springs from demonstrable effort, innovation or initiative — it is that causality the rate β, between zero and ten per cent, measures. A saving that falls from the sky by external or regulatory means carries no private prize at all; in that case the freed share is divided between the resident and the State, two parts to one. And one must distinguish the gain that has come to stay from the one that passes: structural efficiency lowers the fare permanently, transitory efficiency yields a temporary credit, and, once the fare reaches its floor, the resident's share that no longer fits in the price reverts to relieve public expenditure. The key, in short, is not a switch with three fixed positions; it is a rule that adjusts to the origin, the duration and the destination of each gain.
Consider now the circle that follows from it. Greater efficiency on the operator's part means a lower actual cost; lower cost drags down, over time, the reference cost itself; a lower reference means a lower ceiling and smaller compensation; smaller compensation means less pressure on the budget; and less budgetary pressure means a more sustainable system, one that can last without alarms. The circle closes upon itself and turns the right way — each revolution leaves the system a little cheaper and a little more solid than the one before. Note that none of its steps requires an order from above: the operator pursues efficiency out of self-interest, and self-interest, here, has been designed to coincide with the public interest.
There remains the debt the previous section left open: the prudential execution trigger. Where does it sit in this circle? Outside it — and deliberately so. The trigger does not govern the system; it watches it. It exists for the case, always possible, in which the virtuous circle jams — a demand crisis, a cost shock, a scenario the projections did not foresee — and the envelope begins to drain faster than it should. In that case, and only then, the threshold lights up and forces correction. But to confuse the trigger with the helm would be to invert the logic of the model. What steers the system, day to day, is not the threat of a brake; it is the pull of a gain. The helm is efficiency, and the one who holds it, out of self-interest, is the operator itself. The trigger is merely what ensures that, should the hand slip from the helm, the system does not sink before anyone notices.
There is, besides, a virtue this architecture offers for free, and which a European regulator will not fail to appreciate. Because the liquidation recognised to the operator can never exceed the technical ceiling — the efficient cost plus the margin — the system prevents, at the outset, payment above what is efficient, and structurally reduces the risk of over-compensation. This is not, however, an absolute arithmetical impossibility: because total commercial revenue, ancillary revenue and common costs must still be weighed, a global ex post test verifies, at operator–route–year level, whether total revenues and public compensation together exceeded the cost of the service plus reasonable remuneration. The guard against excess is thus twofold: a ceiling that bounds each unit, and a test that closes the accounts at the end. And in a domain where public support to private operators lives under the close scrutiny of competition rules, holding this twofold safeguard is no technical detail: it is the difference between a mechanism that is approved and one that is contested.
7Institutional architecture
It was said that the efficient cost ceases to be declared and comes to be computed. It would be dishonest, however, to present that calculation as a machine that dispenses with judgement. Discretion does not disappear — it is displaced. From the cost, where it was opaque and served the operator, to the reference parameters — which model aircraft, what normative load factor, how many block hours — where it becomes public, structured and reviewable. The integrity of the model resides, therefore, not in the audited accounts; it resides in the governance of the parameters that define the reference operator. Anyone who knows the history of "efficient-firm" regulation knows it was always on that ground, and not on that of costs, that the battles were fought. The MCTH does not escape the rule; it merely acknowledges it and frames it — confining that discretion to versioned, dated and auditable rules, and not to the will of whoever decides.
The most instructive example of this displaced discretion is the normative load factor. A flight may depart full or half empty, and the cost per passenger of a half-full aircraft is, arithmetically, twice that of a full one. If the system reimbursed the cost per passenger just as it emerges from the actual operation, it would reward the empty aircraft: the fewer people the operator carried, the dearer each seat would cost it, and the more the State would pay. The MCTH refuses this absurdity by fixing, instead of the actual occupancy, a normative one — the fill rate an efficient operator ought to achieve on that route. The reference cost is calculated as if the aircraft were filled to that level, and not to wherever the operator's management took it. Whoever flies full sits above the norm and gains; whoever flies empty falls below it and bears the difference. It is a discretionary parameter, no doubt — someone must decide what occupancy is deemed efficient — and that is why it belongs, like the others, to the domain governed by rule and not by will. But it is also one of the model's finest instruments, because it turns the load factor from a passive outcome into an active objective.
All of this — the audited cost, the reference parameters, the load factor — rests on a single condition without which nothing holds: the audit. And not an end-of-year audit, which arrives when the money has already gone, but a permanent audit, contemporary with the operation, that verifies the figures as they are produced. In a system that pays against costs, the credibility of what is paid is exactly the credibility of who audits; and the MCTH makes of the audit not a subsequent control but a continuous organ, without which the computed cost would be merely one more declaration, now under another name.
Around this core of parameters and audit, the architecture raises a set of defences that secure its integrity and its neutrality. The first are the anti-capture rules: the design provides that no one — neither the operator, in reporting costs, nor the regulator, in fixing parameters — can bend the system in their favour without leaving a trace, because every quantity is audited and every parameter versioned and dated. The system is built against manipulation, not entrusted to the good faith of those who operate it.
The second are the neutralities, and they are two. Competitive neutrality: the mechanism does not distinguish the incumbent from the newcomer, does not reward seniority nor penalise entry, and the same route is measured by the same reference, whoever flies it. Technological neutrality: the reference is not tied to any aircraft or generation of equipment, so that the operator who adopts more efficient technology reaps the benefit of that choice, rather than being imprisoned by a norm written for yesterday's aeroplane. Together, these two neutralities keep the door open: operators enter and leave without the system raising artificial barriers, and contestability — that discipline which a thin market, the market of a route whose traffic barely sustains one or two operators, seldom offers of its own accord — is preserved by design.
There is, amid these guarantees, one that deserves note for the elegance with which it solves an old problem: the single resident fare per route, regardless of which operator flies it. What the resident pays does not depend on who carries them — it is the same amount, on the same route, whichever operator flies it. The consequence is subtle but powerful: competition between operators ceases to be waged at the passenger's expense, who is protected by a uniform regulated fare for the route, and comes to be waged where it ought to be — on efficiency, the one ground on which the operator has anything to gain. The citizen is withdrawn from the price war, and the price war is converted into a race for efficiency.
Finally, what gives all this permanence: the multi-year contracts. The parameters, the neutralities, the sharing key, the audit — none of this can live at the mercy of each year's budget, on pain of the operator being unable to plan and the system unable to last. The multi-year contract fixes the rules over a horizon that gives the operator the security to invest in the efficiency expected of it, and the State the assurance that such efficiency will not evaporate at the first turn of the cycle. It is the casing that turns a set of good rules into a stable commitment.
One boundary, however, must be drawn, lest the mechanism be asked for what it does not promise. The MCTH compensates the tariff differential of the eligible passenger — it is not a general instrument for covering a route's deficit. Where the State imposes minimum frequencies, reserve capacity or permanent availability on a connection the market would not, by itself, sustain, those obligations belong to an autonomous public-service obligation, with its own financing. The MCTH covers the price of the one who travels; the availability of the route, where imposed, is covered alongside it, and not within it.
To this foundation the model reserves, further, two directions of development that systematic formalisation will have to fix: a trajectory of reduction of the reference cost over time, so that today's efficiency becomes tomorrow's norm, and the periodic validation of that cost against comparable thin-route systems in Europe, which discipline it and keep it from drifting into complacency. Neither is yet closed mechanics; both are the direction in which this architecture is designed to grow.
8Far beyond air mobility
It would be a lesser ending to close this journey talking about aeroplanes. What was designed here was not an instrument for the aviation sector; it was a way of conceiving public action that the case of mobility merely had the advantage of making visible. Territorial continuity offered the right problem — small, bounded, with auditable costs and identifiable agents — to show, without the noise of greater ambitions, what it means to design for efficiency rather than to compensate inefficiency. But what was shown is not bound to the problem that served to show it.
Look at the method without its object. A cost computed rather than declared; a compensation rebuilt upon quantities the beneficiary does not control; a measurement that does not precede the policy but lives within it; a circuit that settles and simulates at once, adjusting as it runs. None of this is peculiar to aviation. They are operations recognisable, with due translations, in any domain where the State compensates, transfers, subsidises or regulates against costs — which is to say, almost everywhere. The question the MCTH leaves open is not whether there will be other routes; it is whether there will be other sectors where the same architecture can replace blind compensation with efficient recognition.
I do not answer it here, and deliberately so. To generalise a method from a single instance is the work of theory, and theory has demands — of formalisation, of proof, of delimitation — that an article cannot honour without ceasing to be an article. What I can affirm, and do, is more contained, and it is a claim of authorship, not of historical fact: the MCTH is the first fully developed application of the framework I call Productive Architectonic Economics. I do not pretend it is the first incentive regulation, nor the first system that adjusts itself — of these there are examples — but the first integral realisation of this particular method, which now asks to be systematised. That systematisation, when it comes — in another document, with the leisure and the apparatus the name demands — is where it will make itself a discipline. Here there remains only the instance, and the suspicion — well-founded, but yet to be demonstrated — that it will not be alone for long.
Conclusion
Let us return to the beginning, and to the report with which we began. The Fund will say of Portugal, next year and the year after, what it said in this one: that stability is secured and productivity is not; that the ills are known and the cures deferred; that what is wanting is to grow, to innovate, to reform. It will say it well, as it always has. And, as always, its diagnosis will not suffice — because between knowing what is needed and building what secures it remains that distance which no report, however lucid, crosses for us.
What this article has sought to show is that this distance has a name and a method. It is called architecture, and it consists in designing the economic structure within which a policy is to operate before setting it to operate — computing the cost rather than declaring it, settling in real time rather than reimbursing out of time, measuring itself while it acts rather than letting itself be evaluated when it is already too late. The twentieth century taught governments to administer programmes; the twenty-first asks them to learn to design architectures. It is a harder request, because administration tolerates routine and architecture demands the project — but it is also a more honest one, because it obliges them to answer, before spending, the question subsidies never ask: is this cost what it ought to be?
The MCTH was born small, and from a modest problem, which was to carry a resident from the Azores to the mainland without ruining them or squandering the treasury. It solved it; but, in solving it, it realised it was doing something else, larger than the commission. It was demonstrating that a public policy can cease to be an administrative mechanism of compensation and become an economic architecture that measures, simulates, audits and corrects itself — a policy that, for the first time, knows itself as it governs. Perhaps that is its true lesson, and it far exceeds the airport from which it set out.
I shall not maintain here that a science is born of this; I shall maintain it elsewhere, with the patience the sciences demand. For now, the instance suffices — this one, concrete, at work — and the conviction I draw from it: that the next generation of public policy will be distinguished from the last not by diagnosing better, for we already diagnose more than enough, nor by promising more, for of promises we are weary, but by building with the intelligence with which, until now, we have confined ourselves to describing. It was to show that this is possible, and not merely to make tickets cheaper, that this model was written.