A theory of the state earns its authority in the same way a bridge earns its load rating, by being tested against the heaviest traffic available. For Marxist state theory the heaviest traffic of the past four decades is not hard to identify. Governments committed to reducing the role of the state presided over an expansion of its rule making. Financial systems built on the doctrine that markets discipline themselves were rescued, at the moment of collapse, by public money and public guarantees on a scale no postwar intervention had matched. States described as powerless before global capital redesigned their central banks, rewrote their competition law, built new regulators, and bound their own future budgets with rules of their own choosing. Each episode is an embarrassment for somebody’s theory. The argument of this article is that the embarrassments fall unevenly, and that the Marxist tradition, at its most careful, predicted the shape of several of them in advance, because it never accepted the description under which they appear paradoxical.

The description in question is the retreat story. In its popular form, shared oddly by celebrants and critics of market liberalisation alike, the decades after the late 1970s saw the state step back: industries were returned to private hands, markets were freed from political interference, and public authority shrank before the advance of private choice. The story has a surface plausibility, because some things genuinely were withdrawn. State owned enterprises were sold. Price controls and exchange controls were dismantled. Whole categories of economic decision were removed from ministerial desks. Yet the story fails at the level of mechanism, and it fails in a way that can be checked, because everything that was withdrawn had to be withdrawn by an act of public power, and most of what replaced it had to be built, staffed, funded and enforced by the same power. Selling a nationalised industry required legislation, a regulatory office to police the resulting private monopoly, a body of competition law, and courts willing to enforce contracts the state had written. The state did not leave the field. It changed its position on the field, from player to rule maker, referee and guarantor of last resort, and the change of position was itself a vast exercise of state capacity.

That observation, developed across the political economy literature of the past three decades and sharpened by Marxist writers into a general claim, is the organising claim of this article, and it deserves the name the brief for this article gives it. The reconstruction thesis holds that market liberalisation was accomplished by states through legislation and institutional design, so that describing it as the state stepping back inverts the actual mechanism, and any analysis built on that inversion will misidentify where decisions are made. The thesis is not a slogan. It generates expectations that can be compared with the record: rule making should grow rather than shrink in liberalised sectors, enforcement budgets should follow deregulation rather than fall behind it, central banks should gain formal powers even as governments disclaim responsibility for outcomes, and crises should reveal reserves of public capacity that the retreat story says no longer exist. Where those expectations are met, the thesis is doing work. Where they fail, the article will say so, because the same tradition that produced the thesis also produced the tools for testing it, and the honest finding of the empirical literature, that constraint and capacity vary far more than the strong versions of the theory predicted, is treated here as part of the theory’s content rather than as an embarrassment to be footnoted.

A boundary should be fixed before the argument proceeds, because this article applies a framework it does not rebuild. The framework itself, the long argument between instrumentalist and structural accounts of the capitalist state that runs from the Miliband and Poulantzas exchange through the state derivation debates to the strategic relational synthesis, is set out in the marxist theory of the state guide, and readers who want the theory in its own terms should begin there. What follows assumes that framework and puts it to work on six hard cases: financial rescue operations, fiscal consolidation, central bank independence, regulatory capture, platform regulation and industrial policy. The financial mechanism that connects several of those cases, the growth of finance into the commanding position from which rescues became unavoidable, belongs to the companion treatment of financialisation marxist theory, and the wider account of the economic order within which these states operate belongs to marxism and modern capitalism. This article keeps to the state side of the street: what governments did, what constrained them, what they built, and which account of their behaviour survives contact with what happened.

Two accounts of the state, and what each one expects

Application requires prediction, and prediction requires that the rival accounts be stated in forms definite enough to be wrong. The Marxist tradition contains two broad families of explanation for why capitalist states behave as they do, and the families diverge precisely on the cases this article examines. The instrumentalist family, associated in its classic form with Ralph Miliband, explains state action by the connections between the people who run the state and the people who own and direct capital: shared social backgrounds, overlapping careers, directorships, lobbying, campaign finance, and the plain pressure of organised business on particular decisions. On this account the state is an instrument, and the interesting questions concern whose hands are on it. The structural family, associated with Nicos Poulantzas and developed by writers such as Fred Block, explains state action by the position of the state within a capitalist economy, whatever the biographies of its personnel: a state that depends on private investment for employment, output and its own tax revenue will protect the conditions of that investment whether or not any capitalist asks it to, because the costs of failing to do so arrive automatically, as unemployment, fiscal crisis and electoral punishment.

The two families are often taught as rivals for the same territory, and on some questions they are. Yet for application their value lies in the different expectations they generate about observable events, and the discipline of this article is to write those expectations down before consulting the record. An instrumentalist account of a bank rescue expects the rescue to be traceable to access: institutions with the closest ties to treasuries and central banks should receive the most generous treatment, the design of the rescue should bear the fingerprints of the rescued, and the aftermath should show personnel moving between the rescued firms and the rescuing agencies. A structural account expects something colder. It expects the rescue to happen regardless of the government’s composition or ideology, because the alternative, the uncontrolled collapse of the payments system, threatens the tax base, the currency and the employment on which the state’s own existence rests. It expects the rescue to be accompanied by rhetoric hostile to the rescued, since the state’s need is to preserve the system rather than to reward its managers, and it expects the terms to be set by the logic of stabilisation rather than by gratitude. Both patterns can appear in the same episode, which is why single cases rarely settle the argument, and why the article grades each phenomenon in the application grid by fit across the whole pattern rather than by a single dramatic detail.

Block’s refinement of the structural position deserves separate statement, because it supplies the mechanism that cruder versions leave mysterious. Writing in the late 1970s, Block argued that state managers need no instructions from capitalists, because their own rational interest in a healthy revenue base and a quiet society leads them to protect business confidence, the shared expectation among investors that investment will be profitable and safe. Business confidence operates as a filter on policy before policy is ever debated: proposals expected to destroy it are abandoned in drafting, ministers learn without being told which measures are said to be impossible, and the rare government that proceeds anyway discovers the mechanism in the form of capital flight, a currency crisis or an investment strike. Przeworski and Wallerstein later gave the claim a formal model, showing that even a government wholly devoted to workers’ interests, facing rational owners of capital who control investment, would be driven toward policies that sustain profits, because the material welfare of its own supporters depends on investment it does not control. The model matters for this article because it converts a slogan about structural power into a claim with conditions: the constraint binds in proportion to capital’s mobility, the credibility of its exit threat, and the dependence of employment on its decisions, and each of those variables can be examined in a given case.

Why are bailouts better explained structurally than instrumentally?

Bailouts fit the structural account better because the pattern repeats across governments of opposed composition and ideology, because rescuing states acted against their own stated principles, and because the object preserved was the payments and credit system on which revenue and employment depend, not the fortunes of particular managers, many of whom were removed or ruined by the terms imposed.

The point can be sharpened by considering what would count against each account. Evidence of access, of bankers drafting term sheets or ministers taking calls, is real and documented in the inquiries that followed the rescues of 2008 and 2009, and an honest structuralism does not deny it. Access explains the distribution of advantage within a rescue: which institutions were saved whole, which were allowed to fail, which creditors were protected and which shareholders were wiped out. What access does not explain is the existence of the rescue itself, its scale, or its simultaneity across countries whose financial sectors had very different relationships with their governments. Systems with arm’s length traditions rescued as readily as systems famous for intimacy between banks and ministries. Governments that had campaigned against state intervention intervened on the largest scale in their histories. The variable that predicts the rescue is not intimacy. It is exposure: the degree to which the state’s own finances, currency and employment base were entangled with the failing institutions, so that letting them fall meant letting the entanglement pull the state down with them. That is the structural mechanism, stated in its least mysterious form, and it is the same mechanism that explains why the rescues were followed, in most jurisdictions, by an expansion of supervisory law rather than by the grateful quiet the instrumentalist story might lead one to expect.

None of this makes the instrumentalist family useless, and the article does not treat it as a defeated opponent. Lobbying is a real process with measurable effects on the fine print of financial regulation, on tax treatment, on the enforcement budgets of agencies, and on the hundred small decisions that distribute advantage among capitals. The structural claim is about the boundary conditions within which lobbying operates: the lobbyist can bargain over the terms of a rescue, the design of a rule, the timing of a requirement. What the lobbyist is rarely required to do is persuade the state to preserve the financial system, because the state’s dependence on that system has settled the question before the meeting begins. Application, in this tradition, therefore begins by sorting decisions into those where the structural constraint binds tightly, where prediction is strong and personnel are nearly irrelevant, and those where the constraint is loose, where access, ideas, coalitions and electoral calculation move outcomes and where the instrumentalist and mainstream accounts recover their force. The sorting is the skill. The rest of the article practises it case by case.

The retreat that never happened

Is neoliberalism a retreat of the state?

On the Marxist analysis, no. Market liberalisation required extensive legislation, new regulators, competition authorities, central bank redesign, and enforcement. The state was reorganised toward market-supporting functions rather than reduced, which is why measures of state capacity and rule-making did not fall as the retreat account predicts.

Begin with the legislative labour itself, because it is the least deniable part of the record. Privatisation is remembered as a subtraction, assets leaving the public balance sheet, but each sale was an addition to the statute book before it was anything else. A publicly owned utility could be directed by a ministry with a telephone call. A privatised utility operating as a private monopoly had to be governed by something else: a licensing regime, a price formula, a regulator with powers to demand information, an appeals process, and a body of law defining the obligations the private owner owed to consumers who had nowhere else to go. Jurisdictions that privatised fastest built regulators fastest, and the sequence was not accidental. The market in telecommunications, energy or water did not exist in nature, waiting for the state to leave it alone. It was drafted, clause by clause, by parliamentary counsel, and the drafting created a new species of public authority whose entire function was to make a private market behave as a competitive one would. Political scientists who studied the process across Europe gave the resulting formation a name of its own, the regulatory state, and documented its growth sector by sector: where the old state had owned, the new state ruled, and ruling turned out to require more law, more specialised staff, and more continuous intervention in prices and investment plans than ownership had demanded in its final, sleepy decades.

Steven Vogel’s comparative study of advanced economies, published in the mid 1990s under a title that states its finding, documented the pattern beyond the privatised utilities. Liberalisation in finance, telecommunications and broadcasting across Britain, Japan and the United States was accompanied not by the removal of rules but by their replacement: old controls that restricted entry and fixed prices gave way to new rules governing disclosure, competition, prudential soundness and consumer protection, and the total volume of binding rule making rose. Vogel called the combination freer markets and more rules, and the phrase has survived because it describes something anyone can verify in the statute book of any liberalising country. The Marxist literature arrived at the same finding from the other direction and gave it a different emphasis. Where the mainstream account saw a technical adjustment, the reconstruction thesis saw the state’s function being redefined: from directing production to constructing and policing the conditions under which private production takes place, a shift that required the state to know more about the industries it governed, not less, and to intervene continuously in the name of the market rather than against it.

Central banking supplies the clearest single illustration, because it is the reform most often described in the language of depoliticisation. Granting a central bank operational independence is presented, in the retreat story, as the state surrendering a power, removing monetary policy from politics. Examine the institutional facts and the description inverts. Independence is created by statute and can be altered by statute. The independent bank is given a mandate written by legislators, a target chosen by the government, a governor appointed through a political process, reporting obligations to a parliament, and legal protections for its officials that earlier arrangements never needed. What has been built is not an absence of the state but a particular architecture of it, in which one part of the state disciplines the others: the finance ministry is bound by the bank’s reaction to its budgets, future governments are bound by the cost of reversing the settlement, and wage bargainers are bound by the knowledge that excessive settlements will be answered with interest rates. Marxist writers on the monetary authorities, and the broader literature on the new constitutionalism associated with Stephen Gill, read the arrangement as a locking in of market discipline by legal means, an attempt to insulate the core commitments of economic policy from electoral pressure by placing them under institutional guard. Mainstream economists read the same architecture as a solution to a time inconsistency problem, and the readings are not as far apart as their vocabularies suggest: both agree that the arrangement exists to make a certain kind of policy reversal expensive. They differ on whose interests the expensiveness serves, which is a genuine and arguable difference, but neither reading permits the conclusion that the state retreated. A power reorganised is not a power abolished, and the budget of the independent bank, its staff of economists, its supervisory reach and its crisis lending capacity all grew across the decades in which its independence was being celebrated as a withdrawal.

Enforcement completes the picture, because rules without enforcement are literature. The decades of liberalisation saw competition authorities acquire powers to fine, to block mergers, to break up cartels and to enter premises, powers that finance ministries of the interventionist era had rarely possessed in usable form. Tax administration was rebuilt around the audit of complex private structures that privatisation and financial innovation had created. Labour markets described as deregulated were in fact re-regulated in detail: the legal position of unions, the enforceability of restrictive practices, the definition of unfair dismissal, the minimum terms below which contract cannot go, each specified in legislation that grew longer as the rhetoric of withdrawal grew louder. None of this is an argument that the changes were secretly interventionist in their effects on distribution; many were designed to weaken organised labour and strengthen employers, and they did. The argument is narrower and, for that reason, harder to answer. Whatever its purposes, liberalisation was a programme carried out by an active, legislating, enforcing state, and describing the period as one of state retreat mistakes the programme’s self description for its institutional reality. The reconstruction thesis simply insists on the institutional reality, and insists that analysts who begin from the self description will keep being surprised by capacities, central bank balance sheets used as instruments of policy, fiscal rules suspended overnight in a crisis, industrial subsidies announced by governments elected to abolish them, that the retreat story had assured them no longer existed.

The state’s share of national income tells the same story in aggregate for those who distrust institutional description. Across the advanced economies, public spending as a proportion of output ended the liberalising decades roughly where it began them, and in many countries higher, despite repeated programmes announced in the language of retrenchment. Composition changed profoundly: less on state industry and housing, more on transfers, health, debt interest and the apparatus of regulation and enforcement. A theory that predicts shrinkage confronts forty years of non shrinkage and must either complicate itself or retire. The structural tradition in Marxist state theory, whatever its other difficulties, never made that prediction, because it never located the state’s importance in the size of its budget. Its claim was always about function and dependence: a capitalist state must secure the conditions of accumulation and the cohesion of the social order, and it will expand, contract or reorganise its instruments as those tasks require. Liberalisation, on that reading, is not a puzzle. It is one of the forms the securing takes when the previous form, direct ownership within national boundaries, has ceased to serve. The puzzle belongs to the retreat story, and the application grid later in this article records how each hard case scores once the two descriptions are set against the record.

A dated prediction: authoritarian statism assessed

Most theoretical claims in the social sciences are protected from assessment by vagueness about when they might be tested. The claim at the centre of this section is an honourable exception, because it was published with a date on it and describes a direction of travel that the intervening decades could have falsified by going the other way. In State, Power, Socialism, published in French in 1978 and in English the following year, Nicos Poulantzas argued that the capitalist states of the advanced economies were moving toward a formation he named authoritarian statism. The term misleads readers who reach for it as a label for dictatorship, and correcting that misreading is the first task of any application. Authoritarian statism, in Poulantzas’s usage, described a transformation inside formally democratic states: the intensified concentration of power in the executive at the expense of legislatures, the decline of political parties as places where decisions are actually made, the growth of parallel administrative networks answering to the executive, the restriction of formal liberties through ordinary law, and the increasing role of the administration in organising the consent that parties once organised. The ballot box remains. Parliament sits. What changes is where the decisions that matter are prepared, and how much of them arrives in the legislature as a settled matter requiring ratification rather than a question requiring judgement.

What did Poulantzas predict about the executive?

Poulantzas predicted that executive power would concentrate at the expense of parliaments and parties: decision making would migrate to summits, cabinets and administrative networks, legislatures would ratify what they had not shaped, parties would hollow into electoral machines, and the administration would absorb functions of representation and consent building that democratic institutions had performed, all within continuing formal democracy.

The value of a dated prediction is that it can be marked. Take the components in turn. The migration of decision making toward executive summits and small inner groups is among the best documented features of advanced government across the period: major economic settlements negotiated by heads of government and finance ministers in restricted formats, framework legislation drafted in ministries and presented to legislatures under time pressure, and, within the European Union, a mode of governance in which councils of executives and an unelected commission originate most binding law while a parliament amends at the margin. The hollowing of parties is measurable in membership figures that collapsed across Western Europe over the same decades, in the professionalisation of party staffs, and in the transformation of parties from mass organisations with internal policy lives into cartels competing for office, a development the mainstream literature on party decline describes in vocabulary nearly interchangeable with Poulantzas’s own. The growth of administrative discretion, of quangos, agencies, independent authorities and public private bodies that exercise public power at one remove from electoral answerability, is the regulatory state already described, seen now from the side of democratic accountability rather than of market construction. The restriction of liberties through ordinary statute, in public order law, in the legal management of industrial action, in the expanding law of surveillance and preventive policing, proceeded in many jurisdictions by increments too small to announce a rupture and too consistent to describe as drift in any other direction.

Marking honestly also requires recording where the prediction underperformed. Legislatures proved capable of revival as sites of scrutiny, and select committees in several parliaments grew more assertive as executives grew more dominant, a counter tendency Poulantzas’s schema does not anticipate. Judicial review expanded across the same period, creating a rival centre of constraint on executives that sits awkwardly in a model of unidirectional concentration, even if courts discipline governments in the name of property and contract as readily as in the name of liberty. Devolution and regional government in several states dispersed power downward while it concentrated upward, complicating any simple gradient. And the prediction’s mechanism, a functional requirement of monopoly capital in crisis, is the weakest part of the original statement, asserted more often than demonstrated; the trend can be confirmed while the explanation of the trend remains contestable, and a careful application keeps the two judgements separate. What survives that separation is nonetheless remarkable. A theorist writing before the liberalising decades began described their constitutional physiognomy, executive dominance, party decline, administrative proliferation, managed legislatures, more accurately than the celebrants who expected the era to shrink the state and disperse its power. The assessment matters for method as much as for content. It shows the tradition at its best doing something its critics say it cannot do: risking a claim in advance, in print, with enough specificity to be wrong.

Two cautions govern the use of authoritarian statism in application, and both follow from the brief’s discipline for this article. The concept names a tendency within institutions, not a verdict on any particular administration, and deploying it as a label for a government one dislikes converts an analytical term into an insult and destroys its discriminating power, since a tendency measured across decades cannot be inferred from a single controversial measure. The concept also has conditions: it describes states in which the executive’s administrative dominance grows while electoral forms persist, and it loses its grip where legislatures genuinely recover agenda setting power or where parties rebuild themselves as deliberative organisations. An analyst using the term should therefore be able to state the indicators being observed, the direction of their movement, and the evidence that would count as reversal. Used that way, authoritarian statism is the strongest single exhibit for the usefulness of the tradition this article is applying, precisely because it was a prediction before it was a description, and because its partial failures are as instructive as its confirmations. Used as a slogan, it becomes an exhibit for the other side, evidence that the vocabulary survives by being stretched over whatever happens, and the objection section of this article takes that danger at full strength.

Structural dependence: discipline without orders

The mechanism on which the whole application rests is structural dependence, and it deserves the most careful statement in the article, because it is at once the tradition’s most powerful idea and the one most often rendered ridiculous by careless summary. The claim is that a state presiding over a privately owned economy depends for its own purposes on decisions it does not make. Employment depends on firms choosing to hire. Output depends on firms choosing to produce. The tax revenue that funds every public purpose depends on incomes and transactions that private decisions generate. A government of any stripe therefore faces a standing constraint that no lobbyist needs to deliver in person: policies that lead owners of capital to expect poor returns will be answered by reduced investment, relocation, or the simple refusal to expand, and the consequences, falling employment, a shrinking revenue base, a weakening currency, arrive as economic facts rather than as political threats. Charles Lindblom gave the underlying asymmetry its classic mainstream statement in 1977, observing that business occupies a privileged position in market oriented systems because its cooperation is not one demand among others but a precondition of governmental success itself. A minister can ignore a petition. A minister cannot ignore an investment strike, because the strike is indistinguishable, in its early stages, from ordinary commercial prudence, and because its costs are paid by the government’s own supporters in lost jobs before they are paid by anyone else.

Capital mobility intensifies the mechanism by shortening the distance between displeasure and consequence. In an economy of capital controls and fixed exchange rates, a government could sustain policies investors disliked for years, financing deficits behind a wall of regulation while the slow pressure of stagnation built. The liberalisation of capital movements, itself a legislative act of states and one of the reconstruction thesis’s central exhibits, replaced that slow pressure with a fast one. Portfolio capital can leave a currency in an afternoon. A bond issue can fail, or succeed only at interest rates that rewrite the government’s budget. Credit rating agencies, private firms whose judgements were written into public regulation, can raise the cost of borrowing for a state that has broken no law and missed no payment. None of this requires conspiracy, instruction or even intention. It requires only that many holders of mobile assets, each protecting a portfolio, respond to the same signals in the same direction at the same time, and the discipline that results is all the more effective for being nobody’s decision. Analysts in the tradition call this the structural power of capital, and contrast it with the instrumental power of lobbying: the first operates through the anticipated reactions of markets to policy, the second through direct pressure on policy makers, and confusing them produces most of the misunderstandings that surround the theory.

How does capital mobility discipline a government in practice?

In practice the discipline arrives as prices rather than commands: a currency sold down when a budget is announced, bond yields rising until the cost of borrowing rewrites the budget itself, investment postponed by firms citing uncertainty, rating outlooks revised downward, and ministers revising proposals in advance because advisers can forecast each reaction before any market participant has acted.

The anticipatory character of the mechanism is what makes it both powerful and difficult to study. The most consequential exercises of structural power leave no event to observe, because the proposal was never made. A finance ministry official who strikes a wealth tax from a draft budget after modelling the likely market response has not been lobbied and has not surrendered to anyone; a rational calculation has been performed inside the state, and the veto has been exercised by an expectation. Researchers who look only for visible conflicts between governments and markets therefore systematically underestimate the constraint, while researchers who treat every abandoned proposal as evidence of it make the theory untestable, and the methodological section of this article returns to that trap. The documented episodes in which discipline became visible are instructive precisely because they are the exceptions where a government proceeded anyway and the mechanism stepped into the open. The classic case in the literature is the experience of a socialist government elected in a large Western European economy at the start of the 1980s on a programme of nationalisation, reflation and redistribution, pursued in a world of falling demand. Within two years, capital flight, repeated pressure on the currency and the rising cost of defending it forced a reversal toward austerity and monetary discipline so complete that the episode entered the textbooks as the demonstration case of structural dependence under open capital markets. No instruction was issued by any capitalist. The constraint arrived through the balance of payments, and the government’s own economists administered it.

A second class of episode involves governments that were never anywhere near socialism and discovered the mechanism anyway, which is why the theory is not a claim about left governments alone. A British government in the mid 1970s, facing a sterling crisis and a sceptical market for its debt, accepted a loan from the International Monetary Fund with conditions that restructured its fiscal policy, an episode still argued over by historians who note that the Treasury’s own forecasts overstated the emergency. Four decades on, a government in a small euro area economy, elected to reject the terms attached to its creditors’ support and confirmed in that rejection by a referendum, accepted within weeks a programme harsher than the one refused, because its banking system depended on emergency liquidity that its creditors controlled and its exit from the currency would have destroyed the savings of its own citizens. The details differ; the structure is constant. In each case the decisive resource, investment, currency stability, banking liquidity, was controlled outside the government, and the government’s formal sovereignty over policy proved compatible with a narrow corridor of feasible action. Sovereignty and room to move are different variables, and confusing them is the beginner’s error this article exists to correct.

Defiance is nonetheless possible, and the theory is only respectable if it can say when. Governments have imposed capital controls in crises and survived, most instructively where controls were temporary, targeted and defended by a state with strong administrative capacity and reserves. Governments have defaulted on debts and returned to markets sooner than the discipline story implies, because lenders have short memories and few alternatives. Governments with their own currency, deep domestic bond markets and debt denominated in that currency have sustained deficits through episodes that would have broken a state borrowing in foreign money. Each qualification refines the same variable: the corridor of feasible action widens as dependence on mobile external capital narrows, and narrows to a passage as that dependence grows. Structural dependence is therefore a claim about degrees, not a law of impossibility, and the governments that fare worst under it are those that mistake the corridor for a wall in good times and for open ground in bad ones. The practical upshot for anyone advising or analysing a government is a set of questions rather than a verdict: in whose currency is the debt issued, who holds it, how fast can it leave, what does the central bank control, and how many weeks of imports do the reserves cover. The Marxist framing contributes the insistence that these are not merely technical questions, because the distribution of the corridor’s costs, when discipline bites, falls along class lines that the technical vocabulary is designed not to mention, a point the austerity section examines in full.

The competition state: reorganised for mobile capital

If structural dependence names the constraint, the competition state names the adaptation, and the literature that developed the concept in the 1990s, above all the work of Philip Cerny, began from a paradox that the retreat story cannot process. States exposed most directly to global markets did not shrink. They reorganised, acquiring new capacities in some directions while abandoning old ones in others, and the pattern of acquisition and abandonment was intelligible once the state’s task was redescribed. A state competing for mobile investment must offer more than low taxes. It must offer enforceable contracts and efficient courts, infrastructure that logistics firms can schedule around, a workforce trained at public expense to specifications employers set, stable money guaranteed by an independent central bank, and a regulatory environment predictable enough to price. Each offering is a state activity, several of them expensive, and the competition state is therefore not a smaller state but a differently purposed one, its growth sectors clustered around the supply side of the economy and its retreat concentrated in the guarantee of outcomes: full employment abandoned as a commitment, income security thinned, public provision reframed as a platform from which individuals compete.

Cerny’s own framing was not Marxist, and the concept’s usefulness to this article lies partly in that independence: a mainstream scholar of international political economy, examining the same record, concluded that globalisation transforms states rather than eroding them, and that the transformation increases state intervention in the areas where markets are made and maintained. Marxist writers took up the concept and gave it a class inflection the original lacked. On their reading, the competition state’s supply side investments are socialised costs of private production, training, infrastructure, research and legal order provided publicly so that returns can be appropriated privately, while its abandonment of outcome guarantees shifts the risks of economic life from collective provision onto households. The reading is arguable, and the strongest mainstream rejoinder deserves stating: supply side capacity benefits citizens as well as firms, an educated workforce and functioning courts are public goods by any standard, and a state that refused to compete would not thereby escape dependence, merely meet it poorer. The rejoinder lands against crude versions. It does not restore the retreat story, since both sides of the argument describe a state doing a great deal, and the disagreement concerns the distribution of the doing’s benefits, which is exactly where a class analysis claims its competence.

The competition state also explains a feature of the period that purely national theories handle badly, the convergence of policy repertoires across countries whose governments profess opposed philosophies. Corporate tax rates fell in steps across the advanced economies in a sequence no single government chose, each reduction altering the calculation of the next finance ministry. Investment incentives, special economic zones, accelerated depreciation and patent regimes multiplied as instruments of a rivalry conducted in tax codes. Convergence was never total, and the variation that persisted, in welfare generosity, in labour protection, in the public share of health and education, is the empirical finding the evidence section weighs, since it marks the boundary of what structural pressure explains. But the direction of the competitive instruments is uniform enough to require explanation, and the explanation that requires fewest coincidences is the structural one: states facing the same mobile factor reorganise toward the same attractors, and parties alternate in office within a policy corridor whose walls the competition builds. Voters choosing between parties in such a system are not choosing nothing, since the corridor has width and lives are lived differently at its edges. They are, however, choosing inside a structure that none of the parties on the ballot built and none proposes to leave, and the democratic question that raises is taken up in the companion treatment of marxism and democracy critique, which owns the argument about what the corridor does to the meaning of elections.

The consolidation state: creditors and voters

Wolfgang Streeck’s analysis, developed in essays from the start of the 2010s and in the book Buying Time, adds a temporal dimension that the competition state literature lacks, and it supplies the framework within which the austerity case becomes intelligible. Streeck begins from a tension he treats as constitutional for democratic capitalism. The postwar state answered to two constituencies whose claims conflict: citizens, who vote and who expect public services, social security and responsiveness to electoral majorities, and creditors, who hold the state’s debt and expect repayment, stable money and fiscal discipline. While growth was rapid, the conflict could be deferred, and Streeck’s historical sequence describes the deferrals: inflation in the 1970s quietly taxing creditors, public debt in the 1980s and 1990s shifting the bill to the future, private household debt in the 2000s sustaining consumption as wages stagnated, and finally, after the financial crisis, the consolidation state, in which the claims of creditors are institutionalised as the prior charge on public policy and the claims of citizens are managed within what remains. Fiscal rules written into constitutions and treaties, independent fiscal councils, debt brakes and the conditionality attached to rescue lending are the machinery of the consolidation state, and each element is, once again, an artefact of legislation, a state binding itself and its successors by law in order to be believed by the holders of its bonds.

What does the consolidation state add to the analysis?

The consolidation state adds the identity of the audience. It specifies that a heavily indebted democratic state performs fiscal policy before two constituencies with opposed demands, its voters and its creditors, and that consolidation is the institutional settlement in which creditor confidence is secured first, by rules and independent watchdogs, leaving electoral politics to distribute what the settlement permits.

The framework earns its place in this article because it converts the austerity argument from a quarrel about motives into an enquiry about machinery. Whether the finance ministers who imposed fiscal consolidation after the financial crisis believed in it, feared the bond markets, welcomed it as leverage for reforms they already wanted, or some mixture that varied by country and by week, is a question biographies can argue over indefinitely. The consolidation framework asks instead what the institutional settlement made easy and what it made nearly impossible: which expenditures sat inside the protected perimeter and which outside it, whose income streams were treated as contractual and whose as discretionary, whether the adjustment fell on public investment, on transfers, on public sector wages or on the taxation of mobile wealth, and how the choices were distributed across classes with different capacities to absorb them. Those questions have answers that can be documented from budgets, and the answers, across the consolidation episodes of the euro area and beyond, display a regularity that the class reading predicts and the purely technocratic reading struggles to state: the costs of restoring creditor confidence were concentrated on the users and employees of public services, while the claims of bondholders, depositors and the shareholders rescued through bank recapitalisations were honoured in full or restructured only where the alternative was systemic collapse. A framework that directs attention to that distribution, and predicts it in advance of examining any particular budget, is doing analytical work whatever one thinks of its author’s politics, and the grid below scores the fiscal consolidation row accordingly.

Streeck’s own verdict on the settlement’s stability is pessimistic, since a state that answers its creditors reliably and its citizens residually will eventually face the citizens’ withdrawal of consent, and his framework therefore contains a prediction about legitimation crises that parallels the older Marxist literature on the fiscal crisis of the state, in which James O’Connor argued in the early 1970s that the capitalist state must fund both accumulation and legitimation from a revenue base that capital resists enlarging, so that fiscal stress is structural rather than accidental. The convergence of a sociologist of the mainstream and a Marxist economist on the same unstable triangle, growth, creditor confidence and democratic consent, is one of those moments where the applied analyst should pay attention, because frameworks built from opposed premises that identify the same pressure point are usually pointing at something real in the object. The applied method for turning such frameworks into policy analysis, testing them against budgets and legislative records rather than against rhetoric, is developed in the companion article on marxist analysis public policy, and readers who want the workshop rather than the survey should go there next.

The application grid: six hard cases scored

Frameworks earn confidence the way witnesses do, by surviving questions they did not choose. The application grid is the article’s discipline for ensuring that the cases are not chosen by the theory’s friends. Each row takes a phenomenon that any account of the contemporary state must explain, states what an instrumentalist reading expects, states what a structural reading expects, summarises what the record shows, and records which account fits better. The grid is printed in full here so that the scoring can be inspected rather than asserted, and the discussion that follows argues the two hardest rows at length. Readers assembling notes for teaching or examination purposes can keep the grid as a one page summary of the article’s claims; the study notes companion at VaultBook is organised to hold exactly this kind of table alongside the definitions it compresses.

Phenomenon	Instrumentalist prediction	Structural prediction	What happened	Better fit
Financial rescue operations	Rescue follows access and lobbying ties	Rescue follows systemic exposure regardless of government	Governments of opposed ideologies rescued on comparable scales; terms varied with access	Structural, with instrumental effects inside the rescue
Fiscal consolidation	Consolidation reflects the preferences of governing parties and their donors	Consolidation protects creditor claims first and distributes costs downward	Adjustment concentrated on services, wages and transfers while creditor claims were honoured	Structural, with national variation in pace and mix
Central bank independence	Independence is a technocratic reform adopted on expert advice	Independence locks in market discipline against electoral pressure	Mandates written by legislatures, targets set by governments, powers and staff expanded	Structural on purpose, mainstream on mechanism
Regulatory capture	Regulators serve the industries they police through revolving doors	Capture operates within boundaries set by systemic stability	Documented capture in detail and enforcement that repeatedly overrode industry preference	Shared; instrumental inside structural limits
Platform regulation	Regulation follows lobbying strength of platforms and rivals	States regulate where platforms threaten public order, tax base or domestic champions	Major jurisdictions built new regulatory powers over platforms after a long permissive period	Structural on timing, instrumental on design
Industrial policy	Subsidies follow donor and regional interests of incumbents	States subsidise where competition with rival states threatens strategic capacity	Governments elected against intervention adopted large subsidy programmes under competitive pressure	Structural on adoption, instrumental on allocation

Financial rescue operations were argued in the opening sections and the grid records the verdict reached there: the existence, scale and simultaneity of the rescues are structural facts, while the distribution of advantage within them, which institutions were saved whole and which creditors escaped loss, bears the marks of access that the instrumentalist tradition taught everyone to look for. The scoring is deliberately not a shutout. A grid in which one column won every row would be evidence of a rigged enquiry, and the platform regulation row in particular resists any simple verdict. For two decades the large digital platforms operated under a permissive settlement: light taxation of profits booked across borders, minimal liability for hosted content, competition law that struggled to define their markets. An instrumentalist reading of the permissive period is strong, since the platforms lobbied effectively, funded friendly research and faced fragmented opponents. The turn toward regulation, when it came, arrived through structural channels: finance ministries confronting a tax base leaking into jurisdictions of convenience, competition authorities confronting gatekeepers whose private rule making governed whole markets, legislatures confronting the discovery that a handful of firms administered a large share of public communication. States built new powers, new units and new offences, and are building them still, which is the reconstruction pattern in its purest form, a state expanding its regulatory capacity in response to a form of private power it had earlier chosen not to govern. The design of the resulting regimes, their thresholds, exemptions and enforcement budgets, remains a field in which lobbying operates with its usual effectiveness, and the row is scored accordingly.

Industrial policy supplies the grid’s most instructive row, because it records a reversal performed by the actors least expected to perform it. Governments elected on programmes of market liberalisation, some of them within living memory of denouncing subsidy as the road to ruin, adopted industrial strategies involving large public funds for semiconductor fabrication, battery production, green technology and defence capacity. The instrumentalist account can explain the geography of the resulting plants, which tracks electoral marginality and established lobbying channels with embarrassing fidelity. It cannot explain the adoption itself, timed as it was to the subsidy programmes of rival states and to security assessments about supply chains, since no domestic lobby for fabrication plants existed at the required scale before the strategic alarm sounded. The structural reading, in its interstate form, expects exactly this: states competing for the industrial capacity on which military security and technological position depend will subsidise when rivals subsidise, whatever their governing ideology, because the cost of abstention is dependence. Jessop’s work on the state’s strategic selectivity and the older literature on the national state in an international system both anticipate the pattern, and the crisis dimension of it, the way emergencies convert reluctant states into directing ones within weeks, is the subject of the marxist crisis theory guide, which supplies the background this article assumes. The row also carries a warning the tradition should keep: explaining adoption structurally does not launder allocation, and a Marxist analysis that celebrates the return of industrial policy without auditing who received the funds and on what conditions has forgotten the instrumentalist half of its own inheritance.

Regulatory capture, scored as shared, deserves a paragraph of its own because it is the row where the popular version of Marxist analysis is most confidently wrong. The popular version treats capture as proof that the state is capital’s committee and stops thinking. The documented record is more interesting. Capture is real, measurable in revolving door appointments, in the dependence of regulators on industry data, in enforcement budgets set by legislatures attentive to industry complaint. Yet captured regulators repeatedly bite: competition authorities have blocked mergers the parties had paid fortunes to promote, financial supervisors have imposed capital requirements the banking sector resisted as a bloc, environmental agencies have closed profitable operations. The structural account explains the biting without denying the capture. A regulator is charged with the stability and legitimacy of a sector, not with the profits of its loudest firms, and where a firm’s conduct threatens the sector, through fraud, systemic risk or public scandal, the state’s interest in the whole overrides the firm’s interest in itself, sometimes ruinously for the firm. Poulantzas had a phrase for the underlying logic, the relative autonomy of the state, and the capture row of the grid is where the phrase stops being a scholastic subtlety and becomes a working tool: autonomy is relative, exercised within a dependence on the economy the state regulates, and it is precisely the dependence that makes the autonomy necessary, since a state that never disciplined individual capitals would soon have no orderly accumulation left to depend on.

The evidence audit: how strong is the constraint, really?

An applied tradition that reports only its confirmations is a public relations exercise, and this article has promised the other kind. The structural dependence claim has been tested, formally and empirically, by scholars with no interest in rescuing it, and the results complicate the strong version in ways that improve the theory by forcing it to specify itself. The strong version, common in the polemical literature, holds that globalisation and capital mobility have stripped governments of meaningful choice, that welfare states must converge downward, and that elections change administrations without changing policy. If that version were correct, the comparative record would show it. The comparative record shows something else, and this section states the something else at full strength before considering what remains standing.

Geoffrey Garrett’s systematic comparison of the advanced economies through the 1990s found that partisan differences in macroeconomic policy and welfare effort persisted under conditions of rising capital mobility, and that countries combining strong labour movements with encompassing unions sustained higher public spending and lower inequality without paying the growth penalty the strong version predicts. Duane Swank’s work on welfare state restructuring found retrenchment to be real but partial, path dependent and politically mediated, with globalisation pressures filtered through domestic institutions that amplified them in some countries and absorbed them in others. Dani Rodrik documented the compensation pattern from the trade side: more open economies historically built larger public sectors, because exposure to external risk increased the demand for social insurance, a correlation that runs in precisely the opposite direction from the race to the bottom. Research on bond markets added a further refinement. Layna Mosley’s interviews and market data showed that investors discipline governments coarsely rather than finely: they punish macroeconomic disorder, inflation and unsustainable deficits, while granting wide latitude over the composition of budgets, the structure of welfare and the level of provision, a finding summarised in the phrase strong on some things, room to move on others. Mosley’s governments faced a corridor, as this article has argued, but the corridor was wider than the strong structural claim allowed, and its walls stood at the level of aggregate discipline rather than at the level of social policy detail.

How strong is the evidence that globalisation constrained welfare states?

The evidence supports a real but moderate and uneven constraint. Comparative studies found capital mobility pressing on tax structure and macroeconomic stance while welfare effort, partisan differences and domestic institutions continued to produce variation between countries, so dependence operates through specific channels, with measurable strength that varies, rather than as a uniform force flattening every welfare state toward one model.

The correct response to this literature, within the tradition, is not retreat into unfalsifiability but a revision that the better writers had already begun. Constraint varies, and the variation is itself explicable by the theory’s own variables once they are measured rather than assumed. Dependence binds hardest where a state borrows in foreign currency, where its banking system is large relative to its fiscal capacity, where its debt is held by mobile foreigners, and where its export base is narrow. It binds softest where debt is domestic and long term, where the currency is a reserve currency, where welfare institutions command organised constituencies capable of defending them, and where encompassing labour movements can deliver the wage moderation that makes a generous state compatible with competitiveness, the combination Garrett’s corporatist cases display. A theory that predicts variation from measurable conditions, and is embarrassed when the variation fails to follow the conditions, is a theory in healthy shape. A theory that predicts everything predicts nothing, and the distance between those two conditions is the distance between the strategic relational literature and the pamphlet version of structuralism, a distance this article has tried to keep visible throughout.

The austerity debate supplies the test case for the revised claim, because it is the point at which the class reading of state action meets its most serious mainstream rival on ground both sides accept as factual: the budgets themselves. Is fiscal consolidation a class project? The question, as the FAQ section answers it in compact form, divides into a strong form that the evidence cannot carry and a weaker form that it can. The strong form alleges intention and coordination: that consolidation was designed, by identifiable actors acting for a class, as an instrument for shrinking social protection and weakening labour. For some episodes and some actors the documentary record supports elements of that reading, since the use of crisis to advance pre existing reform programmes is a matter of published record rather than inference. As a general explanation, though, the strong form fails on the variation: consolidation differed in speed, composition and severity across countries with similar class structures, and the differences track creditor exposure, currency membership and institutional veto points more closely than they track the enthusiasm of local elites. The weaker form survives that failure and remains a substantial claim: whatever its motives, consolidation as practised distributed its costs by class, protecting the value of financial assets and the incomes attached to them while withdrawing income, services and security from those dependent on wages and public provision, and it did so through mechanisms, the prior claim of debt service, the legal protection of contracts, the mobility of capital against taxation, that are structural features of the state form rather than choices made afresh by each cabinet. Mark Blyth’s history of austerity as an idea, whatever its author’s distance from Marxism, documented the distributional pattern across a century of episodes and concluded that consolidation functions as a protective device for the top of the distribution, a judgement reached from the archives by a scholar whose framework is ideational rather than Marxist. Convergence of that kind, like the convergence over the consolidation state, is the sort of evidence an applied article should weight heavily, because it cannot be explained by shared allegiance.

The honest audit therefore ends in a differentiated verdict, and the differentiation is the finding. Structural dependence is among the best supported mechanisms in the study of contemporary states, provided it is stated as a variable constraint operating through identified channels, and it is among the worst supported when stated as an iron law of convergence. The tradition’s serious literature, Block on business confidence, Przeworski on the formal logic of dependence, Jessop on the selectivity of state forms, mostly states it the first way. Its popular literature mostly states it the second way. An article applying the theory has an obligation to apply the first and to say plainly that the second is a caricature which the evidence refutes, even when the caricature is wearing the tradition’s own colours. That obligation discharged, the remaining question is one of method: whether the tradition possesses a framework that builds variation in from the start rather than bolting it on after the evidence objects. It does, and the framework is the subject of the next section.

The strategic relational approach: the tradition’s working tool

Bob Jessop’s strategic relational approach, developed from the 1980s and stated most fully in The Future of the Capitalist State in 2002, is the framework toward which this article has been converging, because it was designed to solve the exact problem the evidence audit exposed: how a state can be structurally biased without being structurally determined. Jessop takes from Poulantzas the idea that the state is not a thing but a social relation, an institutional condensation of the balance of forces, and adds the concept that makes the idea operational, strategic selectivity. State institutions are not neutral terrain on which all strategies are equally possible. Their forms, their divisions of competence, their rules of access, their inherited routines and their characteristic time horizons privilege some actors, some strategies and some time scales over others, the way a landscape privileges some routes and forbids others without determining any traveller’s journey. A finance ministry organised for credibility with bond markets hears creditor arguments in its native language and union arguments through translation. A corporatist bargaining institution hears them the other way round. Neither ministry nor institution decides the outcome; both tilt the field on which outcomes are decided, and the tilt can be described, compared and, within limits, measured.

What makes the strategic relational approach different?

The approach differs from both families by treating the state as institutions with built in selectivities rather than an instrument to capture or a structure that dictates. Outcomes emerge from strategies pursued by actors with unequal resources across terrain favouring some strategies over others, so variation between economies becomes a result to analyse rather than an anomaly to explain away.

The approach changes what an analyst writes down. An instrumentalist analysis lists connections. A crude structural analysis lists constraints. A strategic relational analysis lists the selectivities of the specific state form in question and the strategies that actors, knowing or sensing the tilt, adopt in response: which ministry holds the pen on a given policy, which interests have routine access and which must storm the building, which arguments are treated as responsible before they are heard and which must prove their sanity, which time horizon the budget process imposes and whose projects fit inside it. Applied to the cases in the grid, the framework explains why the same structural pressure produced fast consolidation in one country and negotiated gradualism in another, why one state’s industrial policy runs through a development bank and another’s through tax credits, why platform regulation took a competition form in one jurisdiction and a content form in another. The differences are not noise around a structural signal. They are the product of distinct state forms selecting among strategies under a common pressure, which is what a mature application should expect the world to look like. Jessop’s own later work extended the framework to the state’s scalar reorganisation, the shifting of competences upward to supranational bodies, downward to regions and outward to networks, and to the rise of meta governance, the state’s growing occupation with governing the governors it has created, which is the reconstruction thesis stated in a systems vocabulary: every act of delegation generates a requirement for oversight, and the overseer of last resort remains, in crisis after crisis, the national state whose obituary the globalisation literature kept drafting.

Honesty about the framework’s costs belongs in an applied article. The strategic relational vocabulary is demanding, its key terms resist one line definition, and its analyses can degenerate into redescription, the event retold in the framework’s language with the tilt asserted rather than demonstrated. Jessop himself warned against using selectivity as an incantation. The discipline that keeps the approach empirical is the discipline this article has applied throughout: state the selectivity in advance, name the actors it should favour, identify the evidence that would show the field level or tilted the other way, and compare across cases where the forms differ. Performed that way, the approach is the strongest current answer the tradition offers to the charge examined next, the charge that structural explanation in application is a machine for being right whatever happens.

The strongest objection, at full strength

The objection must be stated as its best proponents state it, because answering a weak version would waste the reader’s time and the tradition’s credibility. It runs as follows. Structural Marxist explanation, in application, is unfalsifiable. When a government serves the interests of capital, the theory claims confirmation. When a government acts against the expressed interests of capital, taxing it, regulating it, imprisoning its executives, the theory claims confirmation again, redescribing the act as serving capital’s long term interest in stability, a manoeuvre available for every conceivable outcome and therefore informative about none. When outcomes differ between countries with similar economic structures, the theory reaches for auxiliary factors, institutions, histories, balances of force, which mainstream comparative politics explains more parsimoniously with electoral systems, veto points and coalition dynamics, without the metaphysical overhead of structural power. And when the theory’s strong predictions fail, as the convergence prediction failed, its defenders quietly substitute weaker versions while retaining the rhetoric of the strong ones, so that the research programme degenerates in exactly the pattern Lakatos described: a hard core protected by a belt of excuses, absorbing anomalies instead of anticipating facts. Pressed fully, the objection concludes that applied Marxist state analysis is not a theory being tested but a vocabulary being applied, and that its apparent successes, the bailout explained, the austerity distributed, are retellings whose explanatory work is done by ordinary mechanisms that other traditions named first and stated more cleanly.

When does the structural account become unfalsifiable?

The account becomes unfalsifiable at the moment its user permits redescription to substitute for prediction: when serving accumulation is defined so widely that austerity and stimulus, rescue and refusal, nationalisation and privatisation all count as confirmations after the fact, and no observation is specified in advance that would count against the claim being made.

The answer begins with a concession that should not be grudging. A great deal of applied writing in this tradition does exactly what the objection describes, and the objection’s proponents can fill a shelf with offenders. The concession matters because it locates the fault correctly, in a practice rather than in the core claim, and practices can be reformed by standards. The standards exist, and this article has tried to observe them in public: predictions written before outcomes are consulted, as in the grid; mechanisms specified as variables with measurable strength, as in the treatment of dependence through currency, debt composition and banking exposure; defeaters named for the article’s own central claims, as in the marking of the authoritarian statism prediction, where legislative revival and judicial expansion were recorded against it. Where those standards are observed, the structural claim is as falsifiable as any in social science, and the formal literature proves the point by having been, in part, falsified and revised: Przeworski’s model generated implications about the conditions under which social democratic governments would moderate, the comparative literature tested convergence and reported against the strong version, and the tradition’s serious journals printed the results. A programme that publishes its own refutations and rebuilds around them is not the programme the objection describes, even if its pamphleteers behave as though it were.

The variation clause of the objection deserves a direct answer rather than a deflection, because it is the clause with the most truth in it. Institutions, electoral systems and coalitions do explain a large share of the variation between similar economies, and the strategic relational framework does not compete with that finding so much as absorb and reframe it: institutions are the very medium in which selectivity operates, and the question the Marxist analyst adds is why the institutions take the forms they do, whose strategies those forms have historically privileged, and why the corridor within which electoral systems and coalitions operate sits where it sits, bounded by creditor exposure and capital mobility on one side and by organised social forces on the other. That added question is not metaphysical overhead when it generates the article’s testable content, the prediction, borne out in the consolidation episodes, that adjustment costs would be distributed by class across very different institutional settings, a regularity that coalition analysis alone does not lead one to expect. Where the added question generates nothing, where it merely renames what institutionalism already explained, the objection wins that round and the analyst should concede the vocabulary its keep only where it earns it. That is the verdict this article defends: the structural account, disciplined by stated predictions and measured variables, survives its strongest objection as a going concern with a mixed and improvable record, and the reconstruction thesis at its centre survives more strongly than most, because it was a claim about legislation and institutions that anyone can count, and the counting has been done, by friends and by strangers, with the same result.

Inside the corridor: the framework read from a minister’s desk

Theory that cannot survive translation into the working week of the people it describes is literature, and the brief for this article directs that the practical layer be expanded substantially, so this section performs the translation. Imagine the desk of a finance minister in a medium sized open economy, indebted in a currency partly held abroad, governing with a majority that expects both sound money and protected public services. The structural literature this article has surveyed is, from that desk, not an interpretation of events. It is a description of the furniture. The morning begins with prices: the yield on the government’s bonds at yesterday’s close, the currency’s level against its trading partners, the spread over the bonds of the safest borrower in the system. Those numbers are not information among other information. They are the terms on which every other item in the day’s business will be transacted, because a movement in them reprices the debt the state must roll over, the imports the economy must buy, and the credibility with which any announcement will be received. A minister who treats them as weather, regrettable but beneath political attention, discovers within a quarter that they are climate.

The corridor the minister inhabits has identifiable architecture, and mapping it is the first service the framework performs for a practitioner. Its near wall is the refinancing schedule: the volume of debt maturing in the next twelve months, the currency it is denominated in, and the identity of its holders, since debt held by domestic pension funds behaves in a crisis like a family obligation and debt held by foreign portfolio managers behaves like a hotel bill. Its far wall is the banking system: the size of bank balance sheets relative to national income, the exposure of domestic banks to the government’s own bonds, and the doom loop, as the euro area literature named it, in which a weakening state weakens its banks and weakening banks weaken their rescuer. Between the walls lie the variables the minister partly controls: the maturity structure chosen at issuance, the reserve position, the credibility accumulated by previous budgets, the administrative capacity to collect tax from those best able to hide income. The Marxist contribution to this map is not the map itself, which any competent debt management office can draw. It is the insistence that the map has a politics: that the corridor’s width was set by earlier settlements between classes and states, that its walls are patrolled by actors whose interests are served by narrowness, and that when the corridor forces a choice, the choice’s costs are distributed through society along lines the bland language of adjustment is designed to leave unexamined. A minister who knows the corridor is a corridor, built and maintained rather than natural, negotiates differently inside it, widening where widening is possible in good times, refusing to mistake a cyclical widening for freedom, and preparing the politics of the bad times while the good times still fund the preparation.

The staffer’s and analyst’s version of the same lesson concerns anticipation. Governments in the corridor rarely meet the market’s verdict for the first time in the market. They meet it in the drafting room, in the form of the official who models the gilt reaction, the adviser who reports the rating agency’s mood, the permanent secretary who recalls what happened to the last minister who announced an unfunded measure on a Friday. Structural power, experienced from inside, is a discipline of the imagination before it is a discipline of the budget: the range of proposals that reach paper is pre filtered by expectations about reactions, and the filtering is performed by people who regard themselves as technicians and would be offended by the suggestion that they exercise a veto. The analyst trained in this tradition learns to read that filtering as data. The proposals that never appear, the options annexes that list a measure and dismiss it in a sentence citing market reaction, the difference between what a party’s economists model in opposition and what its ministers propose in office: each is evidence about where the corridor’s walls stood at a given moment, and a history of the period written from those traces will often explain more than a history written from the public quarrels, which concern only the proposals that survived the filter. Journalists, whose layer of this article appears below, face the same evidence in degraded form: a market reaction reported as a verdict on competence, when the framework suggests it should be reported as a measurement of dependence, two descriptions of one event with opposite implications for where responsibility lies.

None of this instruction requires the practitioner to accept the tradition’s larger claims, and honesty requires saying so. A debt manager can use the corridor map while believing the corridor just, and a minister can prepare for discipline while intending to satisfy creditors gladly rather than resentfully. What the framework removes is innocence about the structure of the situation: the belief that a government’s room to move is set by its mandate alone, that market confidence is a judgement on virtue rather than a price on dependence, or that the distribution of adjustment costs is a technical residue rather than the main political event. Governments that proceed in that innocence tend to learn the framework’s lessons from experience, at tuition rates charged in currency crises, and the literature of every such episode, read carefully, is a set of footnotes to arguments that Poulantzas, Block and Lindblom had published before the ministers involved entered parliament.

The guardian institution: central bank independence revisited

The retreat section treated central bank independence as an exhibit for the reconstruction thesis, a power reorganised rather than surrendered. The institution deserves a second visit, because its crisis behaviour supplied the single most dramatic confirmation of the article’s central claim, and because the mainstream and Marxist readings of it diverge in a way that clarifies both. The orthodox account of independence is a story about credibility: politicians, tempted to inflate before elections, bind themselves through an independent bank the way a sailor in the old story bound himself against a song he knew he could not resist. The account is coherent and captures something real about the design’s intention. What it cannot describe without strain is the institution’s behaviour when the financial system it guarded began to fail. In the crisis years the independent banks of the advanced world did not behave as bound sailors. They behaved as sovereigns: inventing lending facilities over weekends, purchasing government and private securities in quantities that rewrote the boundary between monetary and fiscal policy, guaranteeing markets that had ceased to function, and coordinating swap lines among themselves that amounted to a collective management of the world monetary system. The capacity to do all this had not been diminished by two decades of supposed state retreat. It had been accumulated, legally and technically, inside an institution whose independence turned out to mean freedom to act without waiting for a legislature, a freedom that in an emergency is worth more than any instruction a legislature could have issued.

Political economists across traditions noticed the paradox at once, and the literature that followed, on central banks as powerful and unaccountable actors, on the new politics of central banking, on quantitative easing as distributive policy conducted by other means, reads in long stretches like an empirical appendix to the older Marxist claim that monetary authority is a form of class relevant state power whatever its formal charter says. The Marxist reading adds two specifications the mainstream literature tends to state and then drop. The first concerns the direction of the emergency action: bank rescues and asset purchases stabilised the value of financial wealth and the institutions holding it, while the wage earners and benefit recipients whose incomes the subsequent consolidation reduced had no equivalent guardian institution authorised to purchase their security over a weekend, an asymmetry of institutional design rather than of anyone’s intentions. The second concerns accountability’s geometry: independence insulates the bank from electoral pressure while leaving it permeable, through its need for market confidence and its recruitment from financial careers, to the expectations of the asset holding public, so that the selectivity Jessop describes is built into the institution’s information flows. Neither specification requires bad faith anywhere in the system. Both follow from asking the tradition’s standing question of an institution everyone else was describing as technical: whose reactions must it anticipate to succeed at its own mandate, and whose reactions can it safely ignore. A guardian that must anticipate the bond market’s mood each morning and need never anticipate the mood of the unemployed is a political institution in the only sense that matters for application, the distribution of whose anticipations count.

The counter case should be given its weight. Independent banks have at times defied the preferences of financial markets, tightening into booms their host institutions were enjoying, and have at other times defied governments of the right as readily as of the left, which is evidence that the institution’s selectivity serves price stability as a goal in itself and not any simple roster of interests. The framework absorbs that evidence without damage, since relative autonomy was always part of the claim: the bank serves the long run conditions of the monetary order, which include disciplining particular capitals and particular governments when they threaten it. What the framework denies, and what the crisis record supports denying, is the description of the whole arrangement as the state’s withdrawal from economic government. The guardian institution is the state in one of its most concentrated forms, and the decades that built it are among the strongest evidence that the era’s direction was reconstruction under a rhetoric of retreat.

An obituary reconsidered: globalisation and the national state

Few predictions in the social sciences have been repeated as confidently, or survived as poorly, as the announcement that globalisation was hollowing out the nation state. The announcement had a strong form, in which mobile capital and integrated production would reduce governments to local administrators of decisions taken in markets and boardrooms, and a scholarly afterlife in which the strong form was quietly replaced by studies of transformation that the original obituarists rarely acknowledged as a retreat. The Marxist tradition’s relationship to the obituary is double, and an applied article should state both halves. One current within the tradition amplified the obituary, theorising a transnational capitalist class and an emerging global state apparatus in which national governments would figure as transmission belts. Another current, associated with Leo Panitch and with the historical sociology of state capacities, objected from the start that globalisation was not a force acting on states from outside but a project authored by states, above all by the most powerful of them, through treaties, financial liberalisation and the legal construction of world markets, and that the authors retained the capacities they had exercised in the authoring. The second current’s account has worn better. The treaty architecture of world trade was negotiated by states and is enforced by their courts. The liberalisation of capital accounts was legislated capital by capital. When the architecture cracked in crisis, the repair crews were national treasuries and central banks, and the international institutions that coordinated them operated, as they always had, as instruments of their member states rather than as a government above them.

Linda Weiss gave the scholarly verdict its canonical form in the late 1990s, arguing that state capacity is not a fixed quantity that integration depletes but a variable that strong states deploy, and that the most internationalised economies were often those whose governments intervened most skilfully in industrial and technological development. Her distinction between the power to insulate domestic policy from external pressure and the power to transform an economy’s position within the international order remains the cleanest way to state what the evidence shows: globalisation reduced the first kind of power for most states, unevenly and in proportion to their dependence, while it increased the returns to the second kind, which the capable states collected. The nation state that emerges from the serious literature is neither the sovereign of the textbooks nor the husk of the obituaries. It is the indispensable author and guarantor of the global order, internally reorganised by the competition and consolidation pressures this article has described, and stratified: a small number of states write rules that others must price into their policies, a middle group negotiates its corridor, and the weakest inhabit corridors so narrow that the distinction between advice and instruction from creditors loses practical meaning. Stratification among states is itself a structural fact the tradition has the vocabulary to describe, through its older theories of imperialism and dependency, and applying those theories to a world of bond spreads and reserve currencies, rather than to gunboats, is among the more promising directions the applied literature has taken.

The question the FAQ section answers in brief, whether globalisation weakens the nation state, thus dissolves into better questions. Weakens it for what, for whom, and compared with which counterfactual. For insulating a full employment commitment against capital flight, most states are weaker than their postwar predecessors, by their own legislative choice compounded over decades. For constructing markets, disciplining populations of firms and citizens, projecting legal order across borders and rescuing financial systems, the leading states are stronger than any predecessors in history, a combination that only looks paradoxical if the state’s strength is imagined as a single quantity rising or falling. The reconstruction thesis was built for exactly this combination. It predicts a state simultaneously less able to defy markets and more able to build and police them, and the decades since the obituary was first pronounced have supplied that combination in abundance.

A worked application: reading a rescue package clause by clause

Frameworks prove themselves on documents, and this section applies the article’s apparatus to the anatomy of a financial rescue package of the kind assembled in the crisis of 2008 and 2009, treating the package as a text whose clauses answer the tradition’s questions in sequence. A rescue package, stripped of its press releases, contains a small number of recurrent elements: a recapitalisation fund through which the state purchases shares in failing banks, a guarantee scheme underwriting bank borrowing, an asset purchase or insurance facility removing impaired assets from balance sheets, emergency liquidity from the central bank against widening classes of collateral, and, attached to each element, conditions whose strictness varies in revealing ways. The analyst’s first question is the structural one: which elements were non negotiable across every jurisdiction that faced systemic failure, whatever its government. Recapitalisation, guarantees and liquidity appear everywhere, scaled to the size of the banking system in peril, and their universality across governments of left, right and centre is the pattern the structural account predicts from the state’s dependence on a functioning payments and credit system. No treasury on earth opened the question whether the banking system should survive. The question that politics was permitted to answer concerned the terms.

The terms are where the second and third questions operate. The instrumentalist question asks who was in the room and whose fingerprints the conditions bear, and the inquiries, memoirs and litigation that followed the rescues documented that access mattered at exactly the level the theory assigns it: the pricing of the state’s shares, the treatment of existing shareholders, the decision of which institution would be made an example of and which absorbed quietly, the survival or removal of particular boards. One large investment bank was allowed to fail in the American sequence, and its failure, which very nearly took the system with it, is best read as the exception that maps the boundary: the authorities either misjudged the interconnections or intended a demonstration, and the consequences taught every treasury watching that the boundary of the permissible lay well inside the territory of rescue. The strategic relational question asks the subtler thing: how the institutional form of each state selected among the available rescue designs. States with strong finance ministries and weak legislatures produced packages by decree over a weekend. States with powerful central banks routed the rescue through the bank’s balance sheet, keeping it off the fiscal accounts and out of parliamentary votes. States inside a currency union discovered that their corridor had been drawn elsewhere, since the decision to save their banks created debts their own central bank could not underwrite, and the rescue of the banks became, within two years, the crisis of the sovereign, the doom loop closing in public. Same pressure, three architectures, outcomes distributed by institutional form: the pattern is the framework’s signature, and a single case read this way teaches more of the method than a chapter of definitions.

The final question the analyst asks of the package is the distributive audit, conducted after the emergency has passed and the accounts can be read whole. Who, in the end, bore the cost of the rescue. The answers vary by jurisdiction, and the variation is data, but the recurring shape is consistent with the consolidation framework: the direct fiscal costs were socialised through budgets that later justified consolidation, the losses of shareholders were real but partial and often recovered as asset prices reflated under public support, the creditors of failed institutions were in most jurisdictions made whole, and the populations whose services and transfers were then reduced to repair the public finances had not been parties to the risks the rescue retired. Stating that shape is not yet an explanation, and the careful analyst distinguishes the three claims the package supports. It supports the structural claim that the rescue itself was overdetermined by dependence. It supports the instrumental claim, in qualified form, that access shaped its internal distribution of mercy. And it supports a claim about legitimation that belongs to the oldest stratum of the tradition: that a state seen to move with unlimited energy for finance and with disciplined reluctance for everyone else will spend authority it cannot easily replenish, a prediction the subsequent decade of political turbulence in the rescuing countries did nothing to falsify. A rescue package, read clause by clause, is thus the applied method in miniature: dependence first, access second, institutional form throughout, distribution last and loudest.

Austerity as a governing programme

Fiscal consolidation appeared in the grid and again in the evidence audit, where its class character was divided into a strong form the evidence refuses and a weaker form it sustains. The subject merits one more pass at the level of the programme rather than the verdict, because austerity in the consolidation states was not merely a budget stance but a technique of government, and techniques of government are this article’s home ground. Consider what a consolidation programme actually contains when it is implemented rather than announced. It contains a fiscal rule or a memorandum that fixes the path of the deficit and removes whole categories of decision from annual politics. It contains a spending review in which each ministry defends its budget against a centre armed with the rule. It contains the reclassification of obligations: commitments once treated as entitlements of citizenship reappear as costs to be managed, indexed less generously, means tested more finely, delivered later in life. It contains, characteristically, a public sector pay policy that uses the state’s position as employer to set an example for private wage bargaining, and a programme of asset sales and outsourcing that transfers functions to providers whose workforces stand outside the public settlement. Each element is administrative. Taken together they compose a reorganisation of the state’s relationship to its citizens as thorough as the reorganisations performed on industry in the privatising decades, and performed by the same means: legislation, institutional design, and the patient use of fiscal emergency to settle questions that ordinary politics had left open.

The Marxist reading of the technique focuses on the asymmetry of the perimeter, and the asymmetry is checkable in any consolidation budget. Obligations owed to creditors sit outside the review: debt service is paid first, in full, and its sanctity is the premise of the entire exercise, since the point of consolidation is to be believed by the holders of the debt. Obligations embodied in contracts with private providers are protected by law and by the compensation clauses the providers’ lawyers wrote into them. Obligations owed to citizens, by contrast, are statutory rather than contractual, and statute can be amended by the same parliament that is conducting the consolidation, so pensions are reindexed, benefits are frozen, eligibility is narrowed, and services are withdrawn by the stroke that contract law forbids against a bondholder. The asymmetry is not a conspiracy and rarely a preference expressed aloud. It is a property of the legal forms in which the state’s promises are held, and it guarantees that a fiscal adjustment conducted under pressure will travel down the path of least legal resistance, which is the path through the incomes of those whose claims on the state are merely political. A class analysis earns its keep by noticing that the path of least resistance is not a fact of nature: the forms were chosen, their choosing had a history, and the history runs through the reconstruction decades, in which the claims of capital were progressively contractualised and entrenched while the claims of citizens were left in the amendable currency of statute. Streeck’s consolidation state is the name for the settlement once it hardens. Poulantzas, reading the earlier hardening from 1978, would have recognised the technique and asked after its authoritarian margin: how much of the programme required parliament’s genuine deliberation, and how much arrived as the arithmetic consequence of rules adopted earlier, under other emergencies, precisely so that the deliberation would never have to occur.

Against this reading the strongest rival is the claim of necessity, and it should be met without flinching. States that cannot borrow must adjust, arithmetic is not an ideology, and several consolidation episodes began in genuine fiscal emergencies where the alternative to an ordered programme was a disorderly one imposed by empty coffers. The tradition’s answer is not to deny the arithmetic but to dispute the distribution of its burden and the uses to which the emergency was put: whether the adjustment’s timetable was chosen for economic or for presentational reasons, whether revenue from mobile wealth was pursued with the same energy as savings from services, whether the programme’s architects used the window to enact permanent changes, in pension ages, in labour law, in the structure of provision, that the emergency did not require and that outlasted it by decades. Those are empirical questions with different answers in different countries, and the applied literature that asks them country by country, budget by budget, is the tradition working as it should. Where the answers show a programme confined to the emergency, the class reading of that episode weakens and should be recorded as weakened. Where they show the emergency used as a solvent for a generation of social settlements, the reading stands, and the grid’s scoring of the fiscal consolidation row stands with it.

What rival explanations handle better

Neutrality in an applied article is not a garnish of even handed sentences. It is the discipline of recording where the rival frameworks win, and this section records three such territories plainly. The first is the explanation of policy innovation and timing at the level of the individual measure. Why a particular country adopted a particular pension formula in a particular year, why one regulator moved against a practice while its neighbour studied it, why an industrial subsidy took the form of a tax credit rather than a development bank: on questions of this grain the Marxist apparatus is blunt, and the fine work is done by the literatures of policy diffusion, bureaucratic politics, party competition and ideas. The career of the independent central bank as a policy fashion, spreading through emulation and expert networks as much as through structural need, is better told by the scholars of diffusion than by any account in which each adoption expresses a local balance of class forces, and an honest application says so, reserving the structural claim for the pattern the fashion served once adopted.

The second territory is electoral politics in its own right. Parties win and lose elections for reasons that run through leadership, scandal, media dynamics, economic weather and the accidents of campaign timing, and the outcomes matter: the corridor has width, and governments of different parties distribute its width differently, as Garrett’s persistence findings showed. A Marxist analysis that treats elections as theatre will fail to predict the differences that citizens experience between administrations, and will deserve the failure. The tradition’s defensible claim is about the agenda rather than the result: the range of programmes that can be carried to office intact, the filtration that occurs between manifesto and budget, and the systematic direction of the filtration under market pressure. Within the filtered range, politics is politics, and its students in the mainstream departments possess tools the Marxist tradition has mostly borrowed rather than built. The third territory is the internal life of the state itself: the turf wars of ministries, the conservatism of legal advisers, the empire building of agencies, the sheer administrative incapacity that defeats programmes no class interest opposed. Bureaucratic politics explains failures of implementation that a class reading would strain to reach, and the strain shows. A state that cannot collect its taxes from the informal economy or staff its inspectorates is not exercising relative autonomy. It is exhibiting weakness, and the vocabulary of capacity, developed by historical sociologists such as Michael Mann and Theda Skocpol from premises outside Marxism, describes that weakness better than the vocabulary of class. The applied analyst keeps all three rival toolkits in the bag, reaching for the structural frame when the question concerns the boundary of the possible and for the finer tools when the question concerns movement within the boundary, and the maturity of the tradition in any given work can be judged by how cleanly it knows which question it is answering.

Testing a structural claim in a single country

The objection section promised standards, and this section converts them into a procedure a researcher can run on one country and one episode, since method stated abstractly persuades nobody. Begin by writing the structural claim in its variable form: not that the state serves capital, but that a specified dependence, on bond finance, on a banking system of a specified size, on export earnings in a specified sector, constrained a specified government during a specified interval, through a specified channel. Each specification is a place the claim can break, and the researcher should welcome the breakage points, since a claim with none is a mood. Next, derive the observations the claim leads one to expect before opening the archive: the abandoned proposals and the language of their abandonment, the sequencing of announcements against market movements, the differential treatment of obligations by legal form, the behaviour of the same government on issues where the dependence does not bite, which serves as the within case control. A government that defies creditors on Monday and employers on Tuesday is telling the analyst that the corridor model is wrong for that episode, and the diary of such defiances is worth more than a shelf of confirmations.

Then consult the sources that record anticipation, because anticipation is where structural power hides. Budget submissions and their annexes, the minutes in which officials forecast market reaction, the correspondence of debt management offices, the memoirs of participants checked against the market data of the days they describe: these are the strata in which a veto exercised by expectation leaves fossils. The researcher should also interview against the claim, seeking out the officials who will say, and document, that the market reaction was invoked as a convenience to defeat proposals the ministry disliked for its own reasons, a misuse of structural language that certainly occurs and that the honest literature must be able to detect. Detecting it is possible precisely because the claim is specified: if the invoked reaction was implausible given the debt structure of the moment, if the same ministry embraced larger risks elsewhere without anxiety, if the forecast reaction failed to materialise when a comparable measure slipped through, then the structural explanation of that decision fails and an institutional or ideological one takes its place. The procedure generalises beyond any single tradition, which is the point. A Marxist claim that survives it has earned the same epistemic status as any other claim in comparative politics, no better and no worse, and a tradition confident in its mechanisms should route its students through the procedure rather than around it. The examination and teaching layers below take up how this discipline is taught; the researcher’s working library for it, the models, the case literatures and the market studies cited across this article, is best kept as a single evolving file of notes and clippings, the purpose for which the VaultBook study notes companion was built.

Why the welfare states did not converge

The persistence of large, divergent welfare states under four decades of globalisation is the fact that did most to discipline the tradition, and an applied article should be able to explain the persistence with the tradition’s own tools rather than treating it as an anomaly donated by critics. Three mechanisms, each developed inside or alongside the Marxist literature, do the work. The first is the functional indispensability of social provision to the economic order itself. Welfare states train the workforce, maintain its health, socialise the costs of reproduction and old age that wages alone do not cover, and stabilise demand across the cycle, and a state that dismantled them would not liberate accumulation but vandalise its conditions, a point Claus Offe built into his account of the welfare state’s contradictions: capitalism can neither do without it nor do comfortably with it. Structural dependence, properly understood, therefore protects as well as constrains, since the dependence runs through the reproduction of labour and social peace as well as through the confidence of bondholders, and the analyst who models only the creditor channel has read half the balance sheet.

The second mechanism is institutional sedimentation, which the strategic relational framework absorbs as selectivity operating over time. Programmes create constituencies, administrative capacities and legal expectations that raise the cost of retrenchment above the cost of drift, so that even governments ideologically committed to dismantling discover that the feasible operation is trimming at the margins, delaying indexation, tightening eligibility, the politics of blame avoidance that Paul Pierson documented from outside the tradition and that any honest Marxist application now cites as a mechanism of the state’s form. The third mechanism is compensation, the Rodrik pattern: exposure to open markets increases the political demand for insurance against risks that markets generate and cannot price for individuals, so integration itself manufactures the constituency for the provision it supposedly forbids. Put together, the three mechanisms predict the record with embarrassing neatness: aggregate social effort persists, its structure shifts from passive transfers toward activation and supply side investment in the competition state’s image, its generosity stratifies by labour market position, and its politics becomes a politics of defended perimeters rather than of expansion or abolition. Variation between countries persists because the mechanisms’ strengths differ: where unions remain encompassing, compensation is negotiated rather than residual; where welfare institutions are contribution based and legally entrenched, sedimentation is deeper; where a state borrows abroad in another’s currency, the creditor channel can, in a crisis, override all three at once, which is why the euro area consolidations cut deeper than the pressures felt by states with monetary sovereignty. The welfare case, handled this way, stops being the theory’s embarrassment and becomes its demonstration that structural claims, stated as variable mechanisms in specified institutional settings, can explain persistence and change with the same apparatus.

The platform economy and the tax base

One further case extends the grid’s logic into territory the classic literature never saw, and it is worth a section because it tests whether the framework travels or merely reminisces. The large digital platforms presented states with a new species of problem: firms whose most valuable assets are intangible, whose profits can be booked where the tax rate is kindest, whose market power rests on network effects that competition law was not drafted to reach, and whose services had become part of the infrastructure of daily life and public communication. The permissive first response of states fitted the competition state model to the point of parody, jurisdictions bidding against one another with tax rulings and light touch regimes to host the platforms’ regional headquarters, socialising the costs of the educated workforces and legal orders the platforms required. The turn, when it came, followed the fiscal and sovereign logic the framework predicts. Finance ministries watching the corporate tax base erode into booking jurisdictions cooperated, against their competitive instincts, on a global minimum tax for large multinationals, negotiated through the OECD over years and adopted in principle by more than a hundred jurisdictions: the competition state discovering the limits of competition when the prize is the revenue on which the state itself depends. Competition authorities, confronting gatekeepers that set the terms on which other firms reach customers, built new ex ante regimes rather than waiting for harm to be litigated after the fact. Legislatures wrote duties of care and transparency obligations for hosted content, converting private moderation systems into objects of public law.

Read the episode with the article’s questions and the pattern is by now familiar enough to state briefly. The adoption of regulation was structural in timing, driven by threats to the tax base, to the governability of public communication and to domestic firms competing against untaxed gatekeepers, and it occurred under governments across the spectrum in the major jurisdictions at roughly the same moment, which is the signature of a pressure rather than a preference. The design of the regimes bears instrumental marks throughout, in thresholds set just above or below particular firms, in enforcement resources that lag the statutes, in the revolving door between platform policy teams and the ministries that regulate them. The institutional forms differed by state tradition, competition law in one place, communications law in another, tax treaty machinery in a third, which is selectivity doing its usual work. And the reconstruction pattern completes the case: an activity once governed by contract and courtesy is now governed by agencies, reporting duties, fines and appeals, which is to say by more state, built in the name of disciplining a market. A framework formulated to explain steel nationalisations and bond strikes that can also organise the facts of platform regulation, without forcing them, has passed the only test that distinguishes a living theory from a period piece.

Crisis as revelation: what emergencies disclose

Emergencies are the applied analyst’s natural experiments, because they strip away the routines that let every theory claim the ordinary times. In a crisis the state must act at speed, its choices reveal its priorities in an order that peacetime budgets conceal, and capacities that official doctrine declared obsolete are either produced on demand or confessed as lost. The financial crisis disclosed the reserves of the fiscal and monetary state: guarantees issued in days that dwarfed annual budgets, central bank balance sheets expanded by decisions taken in hours, and parliaments discovering that the sovereignty they had lent to rules could be recalled by the same vote that lent it. The pandemic emergency that followed a decade later disclosed a different register of the same truth: states closed economies by decree, paid the wages of millions through furlough machinery built in weeks, directed production, procured at any price, and ran deficits that the consolidation rules of the previous decade had defined as impossible, before the rules were reimposed, in several jurisdictions, with a speed that itself requires explanation. War in Europe then disclosed the energy and industrial register: states that had treated energy supply as a market outcome rebuilt it as a strategic system within a year, subsidising, requisitioning and rationing by price cap, while defence industrial capacity, after decades of procurement governed by market logic, became an object of direct state planning. No party or leader is characterised by observing the sequence, which is institutional rather than biographical. The observation is that each emergency was met by the rapid expansion of exactly the capacities the retreat story had retired, and that the expansion was fastest where the pre crisis state had maintained administrative depth beneath the rhetoric of withdrawal.

The tradition reads such episodes through a concept the crisis literature has made its own: the state of exception reveals the normal, because the hierarchy of protected claims, visible for a moment with the mask off, does not reassemble in a different order when calm returns. In the financial emergency the first claims protected were those of the payments system and the creditors attached to it; support for households arrived later, was smaller relative to exposure, and was withdrawn sooner. In the pandemic emergency the order briefly inverted, household income protected first and unconditionally, because the threat ran through public health and social order rather than through bank balance sheets, and the inversion is as instructive as the rule: the hierarchy of protection tracks the channel through which the crisis threatens the reproduction of the whole, which is the structural claim stated in its most general form. Analysts can use the sequence prospectively. When the next emergency gathers, in climate damage, in a payments failure, in a supply shock, the framework predicts the questions that will decide the response before the response is announced: which claims are treated as systemic, whose balance sheet the state will stand behind without a vote, which obligations will be declared contractual and untouchable and which statutory and revisable, and how quickly the emergency expansion will be reclassified as excess once the systemic channel is safe. A theory that tells its user where to look during the fog of an emergency has passed a test that purely retrospective frameworks cannot sit.

The fiscal constitution: rules that bind their authors

Between the corridor’s walls and the consolidation state’s machinery lies a legal stratum that repays close attention, because it is where the reconstruction thesis becomes literally constitutional. The decades of liberalisation were also the decades in which economic policy was progressively juridified: deficit and debt limits written into treaties and in some countries into constitutional text, fiscal councils established to police budgets with the authority of expertise, central bank mandates entrenched in statute and treaty articles amendable only by unanimity, trade and investment agreements granting investors enforceable rights against future regulation. Each instrument shares a design: a political majority, at a moment of strength or of crisis, binds its successors by raising the cost of reversal above the cost of compliance. The mainstream literature describes the design as commitment technology, a solution to the instability of democratic preference over time. The Marxist literature, in the work associated with Gill’s account of disciplinary neoliberalism and with the German debates on the market conforming constitution, describes the same design as the insulation of accumulation’s conditions from democratic revision, and the descriptions differ mainly in their evaluation of whose instability is being corrected. Voters, after all, are the variable the instruments treat as the risk: no fiscal rule binds a bond market to lend, no treaty obliges capital to invest, while every instrument obliges or constrains the organs through which citizens act. The asymmetry is a matter of drafting, open to inspection, and stating it requires no theory at all, which is why the fiscal constitution has become the applied literature’s most persuasive teaching case: students who resist every abstraction in the tradition can be brought to the primary texts of a debt brake or a fiscal compact and asked the tradition’s question of them, whose future choices this document exists to prevent, and why those choices, rather than others, were defined as the danger.

Application follows the document trail in both directions. Forward, the analyst reads a proposed rule for its incidence: which expenditures it will squeeze when revenue falls, since rules bind totals and the composition of the squeeze is decided by the legal hierarchy of obligations examined in the austerity section; which generation of politicians it empowers, typically the finance ministry and the fiscal council against the spending ministries and the legislature; and which constituencies acquire, through it, a veto they could not win in an election, typically the holders of the public debt, whose confidence the rule exists to purchase. Backward, the analyst reads the rule’s origin: the crisis or negotiation that supplied the majority, the alternatives that were available, and the interests that drafted. None of this reading practice presumes the rules are wicked. Rules that bind governments can protect citizens from inflation, from the looting of pension funds by a temporary majority, from the chaos of defaulted promises, and several of the instruments were supported across party lines by politicians with honourable motives. The analytical claim is narrower, that the ensemble of rules is not neutral among futures: it selects, in Jessop’s sense, for strategies of credibility with creditors and against strategies of redistribution through the budget, and it raises the price of the second until only a government willing to fight on constitutional terrain can pay it. Whether that selection is a safeguard or a cage is the political question the framework refuses to answer for its user. That it is a selection, designed and legislated, is the fact the retreat story cannot see, and seeing it is the beginning of competence in reading the contemporary state.

The fiscal council, the least dramatic of these instruments, repays attention precisely because of its modesty. A council commands no budget, vetoes no bill and commands no battalion of inspectors. It publishes forecasts, costs the proposals of governments and parties, and pronounces on the plausibility of the numbers on which budgets rest, armed with nothing but a statutory remit and a reputation for independence. Its power is the power of the authoritative embarrassment: a government may override its council, but it must do so in public, against an official finding that its arithmetic does not hold, and the markets whose confidence the entire apparatus exists to secure will read the override as information. Here the selectivity of the fiscal constitution operates in its purest form, through anticipation alone. Party economists drafting a manifesto cost their promises against the council’s known methods before a word is published. Officials strike the pledge whose costing would be savaged. The council never meets the proposal it has already defeated, and its annual report records a world of fiscal rectitude that its own existence silently manufactured. A researcher hunting structural power in the drafting record will find few cleaner specimens, and a citizen wondering why manifestos of opposed parties converge on the same fiscal arithmetic will find, in the council’s remit, a large part of the answer: the arithmetic is guarded by an institution built to guard it, staffed by a profession trained to one standard of soundness, and answerable, in the last instance, to the confidence of the creditors whose claims the whole constitution was written to protect. None of that makes the council a fraud or its forecasts worthless; honest arithmetic is a public good, and governments that cook their books injure their citizens first. It makes the council a political institution of a recognisable kind, one whose neutrality is real at the level of calculation and structured at the level of what counts as a question, which is exactly the combination the strategic relational framework predicts for the mature institutions of a reconstructed state.

The corridor met from both directions

A theory of constraint that only ever constrains one side of politics would be a partisan instrument wearing analytical dress, and the record, read whole, is reassuringly even handed in a bleak way: the corridor has disciplined governments of the right with the same indifference it shows to governments of the left, whenever their programmes collided with the expectations of those who fund states. The episodes differ in costume and agree in structure. A conservative government in Britain in the early 1990s, committed by ideology and treaty to a fixed exchange rate against the currencies of its partners, spent its reserves and raised its interest rates in a single day of futility before abandoning the parity, because the holders of sterling had concluded that the domestic cost of defending it exceeded the government’s capacity to pay, and had sold accordingly. No socialist measure was anywhere in sight; the programme being disciplined was a monetarist government’s own centrepiece, and the discipline was administered by the same mechanism, at the same speed, as in the textbook cases of left governments forced to reverse. Earlier in the same country’s postwar history, a Labour government had met the corridor over devaluation and an IMF loan; later, governments across the euro area of every colour met it over bond spreads. The catalogue matters because it refutes the lazy use of structural dependence as a synonym for the frustration of the left. Dependence is a property of the state’s position, not of its occupants’ intentions, and its political meaning is almost the opposite of the lazy use: by showing that the constraint binds governments that capital might be expected to favour, the record supports the structural reading against the instrumental one, since a committee of the bourgeoisie would presumably exempt its friends.

What varies with the governing party is not the existence of the corridor but the location of its walls and the distribution of the pain when a government tests them. Markets extend differing credit to differing programmes, and the political economy literature has measured a partisan premium: governments of the left historically paid higher interest rates on taking office, as holders of capital priced an expected redistribution, and earned the premium back only by demonstrating restraint, a pattern documented across the postwar decades and weakening as left parties pre committed themselves to the fiscal constitution described above. Governments of the right receive a presumption of soundness that survives until a specific measure destroys it, whereupon the discipline arrives with a violence proportional to the surprise. The asymmetry is real, it is itself a structural fact about whose intentions are trusted by the owners of mobile wealth, and the tradition is entitled to cite it as evidence about the class character of confidence. But the same literature records the discipline’s arrival in both directions, and the applied analyst should hold both findings at once: confidence is partisan in its presumptions and structural in its enforcement, generous with its benefit of the doubt along class lines and ruthless, when the doubt is exhausted, without regard to them. Governments that understand the combination govern accordingly, purchasing credibility where it is cheap for them and spending it where their purposes require, and the rare governments that mistook the presumption of soundness for a permanent endowment have supplied the case literature with its most dramatic weekends.

State theory at street level: the everyday state examined

Application in this article has operated at the altitude of treasuries and central banks, and a reader could finish it believing the theory speaks only of summits. The tradition’s mechanisms, however, run at street level with the same logic in smaller denominations, and a section at that altitude tests whether the apparatus is a method or a mood. Consider the government of a city whose largest employer announces a relocation. The instrumentalist questions arrive first and are legitimate: the meetings between executives and mayors, the incentives offered, the personal networks through which the threat was communicated. The structural questions follow and dominate: the city’s budget depends on the employer’s payroll and the spending it supports, its bonds are rated against that dependence, its service commitments assume the revenue, and so the negotiation occurs inside a dependence that no meeting created and no mayor can resign from. The city will offer what it can, tax abatements, infrastructure, land, and the offer will be described as partnership. The framework describes it as the corridor in municipal form, widened or narrowed by the same variables as the national case: the mobility of the employer, the diversity of the tax base, the legal powers the higher levels of the state permit the city to use. Consider, next, the administration of welfare in the same city, where the consolidation state’s logic arrives as software: eligibility systems that sort claimants by risk, payment schedules designed around fraud prevention, sanctions regimes whose severity tracks fiscal pressure at the national level rather than local need. The street level official applying the sanction is not executing a class conspiracy, and the framework does not ask anyone to believe otherwise. The official is executing a selectivity built into forms, targets and budgets set at altitudes where the creditor’s claim is a premise and the claimant’s income is a variable, and the distance between those altitudes is precisely what the theory exists to make visible.

The street level also supplies some of the framework’s best correctives, and an honest application collects them. Local states innovate in the corridor’s slack: municipalities have built public banks, remunicipalised water and energy systems that privatisation had sold, and constructed housing programmes against national policy currents, using legal and fiscal room that the strong version of structural dependence says should not exist. Each case refines the variables rather than refuting the frame: the room existed where the debt was small, the asset was immobile, the law permitted, or the organised constituency was strong, and the innovations cluster exactly where the theory, stated in its variable form, predicts slack. A framework that can explain both the discipline of the indebted city and the defiance of the solvent one, by reference to measurable features of their positions, has earned the right to be called a method. One that explains only the discipline is a lament, and the tradition’s pamphlets are full of laments; its literature, at the standard this article has tried to keep, is something better.

The plumbing of the world economy: reserves, payments and statecraft

Below the level of treaties and summits, the contemporary state’s power runs through plumbing: the payment systems through which money moves, the reserve assets in which states store value, the correspondent banking networks that connect national systems, and the technical standards bodies whose decisions determine whose institutions may transact with whom. The plumbing was built by states, a fact the retreat story’s focus on markets tends to hide. The dollar’s position at the centre of the system rests on the depth of the market for United States government debt, which is to say on the fiscal capacity of a state; the messaging networks that instruct transfers operate under the law of the jurisdictions that host them; the reserve status that lets one state borrow in its own currency at privileged rates is the widest corridor in the world economy, the exorbitant privilege of the older literature, and its possessor experiences structural dependence in a form so softened that its governments can sustain deficits through crises that would discipline any other borrower within weeks. Stratification, in other words, is not a metaphor laid over the system. It is machined into the pipes, and the Marxist tradition’s older vocabulary of hierarchy among states, built for an age of colonies and concessions, translates into the new material with unusual directness once the objects are renamed: dependence measured in reserve composition and swap line access rather than in garrisons, discipline administered by exclusion from payment networks rather than by blockade.

The use of the plumbing as an instrument, in sanctions regimes that sever institutions and whole economies from the networks, completes the demonstration that the state did not retire from the world economy it built. A sanctions programme is an act of pure statecraft conducted through market infrastructure: private banks enforce it because their access to the central networks is worth more than any prohibited client, compliance departments multiply inside firms that nominally celebrate the free market, and the target state experiences the corridor as a wall, its reserves frozen by decisions taken in foreign treasuries. Whatever one’s judgement of any particular programme, and this article offers none, the technique is unintelligible within a story of state retreat. It is equally unintelligible without the tradition’s central insight in reverse: the state that controls the infrastructure on which private accumulation depends possesses a power over accumulation itself, and the states that built the infrastructure with the deepest liquidity hold that power in its most concentrated form. Applied work in this area, on the weaponisation of interdependence as one prominent mainstream account names it, has converged from international relations theory toward conclusions the Marxist literature on imperialism reached by its own road, and the convergence, yet another instance of frameworks agreeing on the pressure point from opposed premises, marks the plumbing as territory where the applied analyst should expect the tradition’s questions to keep paying rent: who built the network, whose law governs it, whose access can be revoked, and whose accumulation is hostage to the revocation.

A map of the applied literature

A reader persuaded that the application is worth pursuing deserves a chart of the field as it stands, organised by the question each school answers best rather than by allegiance. Begin where this article began, with the classic exchange and its sequels: Miliband’s empirical sociology of state personnel, Poulantzas’s structural alternative and his final, predictive work on authoritarian statism, and the state derivation debates in which German scholars attempted to deduce the state’s form from the logic of capital itself, an attempt whose rigour exposed, better than any critic, the point where deduction must give way to history. Fred Block’s essays remain the shortest route into the mechanism of business confidence, and the formalisation by Przeworski and his collaborators the shortest route into the mechanism stated as a model with conditions. The fiscal sociology line, running from O’Connor’s fiscal crisis through Offe’s contradictions of the welfare state to Streeck’s consolidation state, answers questions about the budget as the state’s exposed nerve, and it is the line that has travelled furthest into mainstream citation, which the analyst may take as evidence about where the tradition’s claims were strongest all along.

A second shelf holds the comparativists and the testers, several of them outside the tradition and indispensable to it: Garrett on partisanship under globalisation, Swank on welfare restructuring, Mosley on the granularity of market discipline, Pierson on retrenchment as blame management, Vogel and Cerny and Weiss on the transformed rather than diminished state. A third shelf holds the institution builders of the framework itself: Jessop’s strategic relational corpus, the regulation school’s accounts of accumulation regimes, Gill on the new constitutionalism, Panitch and his collaborators on the state as author of globalisation, and the crisis literature that followed 2008, in which scholars of several traditions documented the return of the state they had never quite announced as departed. The applied student reads across the shelves in that order, theory first for the questions, testers second for the discipline of evidence, framework third for the synthesis, and keeps the grid habit throughout: for every claim encountered, from any shelf, the claims of the other shelves are written beside it before the evidence is consulted. That habit, more than any doctrine, is what the tradition at its best has to teach the study of the contemporary state, and it is teachable, examinable and usable by readers who will never call themselves Marxists, which the persona sections that follow take up in turn.

The next application field: energy, climate and industrial strategy

Every generation of the applied literature has been defined by the material its decade supplied, and the material now accumulating concerns the energy transition and the industrial strategies gathering around it, a field on which the framework’s questions can be posed in advance and its performance, for once, registered prospectively. The structural questions write themselves. Decarbonisation requires investment on a scale and at a pace that private returns, unassisted, have not historically delivered in infrastructure, so the corridor problem appears in a new form: states must induce investment they cannot command, from owners of capital whose alternative is to wait, while the fiscal constitution built in the consolidation decades limits the borrowing with which states might have invested directly. The competition state logic has already reasserted itself in the resulting subsidy rivalries, as the grid’s industrial policy row recorded, with governments bidding for battery plants, fabrication facilities and hydrogen infrastructure in a sequence each finance ministry experiences as forced by the previous one’s announcement. The selectivity question follows: within each state, which institutions hold the pen on transition policy, energy ministries captured by incumbent utilities in the instrumentalist pattern, treasuries applying the discount rates that make long dated public investment look wasteful, or new delivery agencies built, in the reconstruction pattern, to govern a market into existence. The distributive audit waits at the end of the process, in the structure of the subsidies and the ownership of the resulting assets: whether the transition’s costs are socialised while its returns are privatised, the pattern the rescue packages displayed, or whether states use their indispensability as underwriters to claim public stakes, cheaper energy as a social wage, and direction over the pace of closure in the regions the old economy sustained.

Registering the questions in advance is more than a rhetorical flourish. It converts this section into a deposit against the article’s own standards, a set of expectations a later reader can mark as the evidence accumulates: that the transition will be built by legislatures and regulators rather than left to price signals, that its pace will track interstate competition and fiscal capacity more closely than electoral preference, that incumbent energy capital will shape the design of the regimes that replace it without succeeding in preventing them, and that the states with the widest corridors, reserve currency issuers and fiscal heavyweights, will move first and set the terms on which narrower corridor states must follow. If the record instead shows price signals accomplishing the transition without state construction, or electoral preference predicting national pace better than fiscal position, the framework will have failed a test it set itself in print, and the failure should be recorded in the same tone this article used for the convergence prediction. That is what application means when the word is taken seriously, and it is the standard by which the article as a whole asks to be judged.

What would change the verdict

An applied survey owes its reader a closing account of its own exposure: the observations that, if they accumulated, should persuade the reader to put the framework down. Four would do serious damage. The first is a sustained episode in which a heavily indebted state, borrowing in foreign held currency with an exposed banking system, defied creditor preferences across a full cycle without paying the corridor’s toll in rates, currency or forced reversal; one such case is an anomaly, a pattern of them is a refutation of dependence as specified. The second is a reversal of the reconstruction pattern itself: a major economy in which liberalisation proceeded over a generation with shrinking rule books, falling enforcement, and central banks losing rather than gaining legal capacity, which would break the thesis at the centre of this article rather than at its edges. The third is the failure of the distributive regularity: consolidation programmes across differing institutions that distributed their costs without class pattern, falling on asset incomes as readily as on services and wages, would remove the finding that the consolidation framework and the ideational historians reached independently, and with it much of the case for treating the state’s form as class selective rather than merely constrained. The fourth is subtler and would be visible only across a literature: the strategic relational framework ceasing to generate analyses that surprise their authors, its applications becoming commemorative, rereading settled cases in settled vocabulary, in the pattern that marks a research programme’s degeneration more reliably than any single failed prediction.

None of those four observations describes the record as it stands, and the article has documented the record’s contrary indications at length: dependence binding measurably and variably, rule making growing through every liberalisation, consolidation distributing by class across institutions, and a framework still capable of being embarrassed by welfare persistence and corrected by it, which is the opposite of commemoration. The verdict the evidence supports is therefore neither the triumphalism of the pamphlets nor the dismissal the objection urged at full strength. It is the working verdict of the grid, row by row: a tradition whose strong forms failed and whose specified forms keep identifying, in advance and in the open, where decisions about the contemporary state are actually made, who can veto them without speaking, which claims on the public purse the law treats as sacred and which it treats as negotiable, and why the state that was supposed to retreat has been found, in every emergency and every statute book, rebuilt. For the analyst, the minister, the journalist and the student, those are not ornaments on the study of politics. They are most of its substance, and a theory that keeps supplying them has earned its place on the desk where the substance is worked on.

Pooled sovereignty: the supranational layer and its selectivity

The hardest test the framework faced in the application decades was the construction of a currency union among sovereign states, because the union appeared to realise the obituary’s prophecy in institutional form: national governments surrendering the monetary instrument entirely, binding their budgets by common rules, and submitting to a central bank no single electorate could instruct. The literature’s verdict, as the arrangements matured and then entered crisis, ran the other way, and the episode deserves treatment as the supranational counterpart of the reconstruction thesis. The union was not an escape from states but a construction by them, negotiated treaty by treaty, and its design encoded the preferences of its strongest members in its rules: price stability as the bank’s overriding mandate, fiscal limits as the condition of membership, and no mechanism of fiscal transfer by which a stricken member’s citizens could claim on the whole. Selectivity, in Jessop’s sense, was written at the constitutional scale. Strategies of national reflation were not merely discouraged by the design; they were deprived of their instruments, while strategies of internal devaluation, wage restraint, fiscal consolidation and export discipline were left as the only moves the board allowed, a tilt so pronounced that the crisis years merely made visible what the founding documents had arranged.

The crisis then administered the framework’s own examination. Member states in distress discovered that the corridor, in a currency they did not issue, narrowed to the width of their creditors’ patience, and that the union’s institutions, council, commission and central bank, operated as the corridor’s wardens with a hierarchy the treaties had left ambiguous and the emergency clarified: the central bank’s liquidity decisions became the effective constitution, since a banking system cut off from emergency assistance collapses within days whatever a parliament votes. States outside the union, retaining their own currencies and lenders of last resort, faced the same markets with wider corridors and used them, devaluing, easing, and spreading their adjustments across years rather than months, and the contrast between the two groups during the same global crisis is among the cleanest natural comparisons the applied literature possesses: dependence constant in kind, corridor width varying with monetary sovereignty, outcomes varying with the width. The supranational layer thus confirms rather than escapes the article’s argument. Sovereignty pooled is sovereignty reorganised, its pooling designed by states, its design selective among strategies, and its crises resolved by the reappearance, in Frankfurt meeting rooms and emergency summits rather than in national parliaments, of exactly the executive concentration that Poulantzas had described at the national scale a generation earlier. The analyst who has understood authoritarian statism as a claim about where decisions migrate will know where to look when the next treaty is drafted, and will read its articles for the selectivities they encode before the first crisis arrives to reveal them.

When states say no: default, controls and the price of refusal

The corridor metaphor risks implying that refusal is impossible, and the record of refusal deserves its own audit, because the cases in which governments defied creditor preferences and survived are the theory’s most instructive data: they price the toll rather than denying the road. A Southeast Asian government at the end of the 1990s, facing speculative collapse and an international programme it judged ruinous, imposed capital controls against the near unanimous advice of the financial establishment, stabilised its currency, and recovered with its policy autonomy partly intact; the episode is now taught, including by institutions that condemned the controls at the time, as evidence that the corridor has a gate that administrative capacity and reserves can hold. A South American government in the following decade defaulted on a mountain of debt, imposed a restructuring with deep losses on its creditors, was denounced as a pariah, and returned to growth and eventually to markets, having discovered that the punishment for default, exclusion from borrowing, matters less to a state that has decided to live without borrowing for a season. A small North Atlantic island state, its banks collapsed at a scale dwarfing its economy, let the banks fail, imposed controls, put the question of compensating foreign depositors to referendums that refused, and prosecuted bankers, emerging with its welfare state and its currency intact. Each refusal was real. Each also paid the toll in the corridor’s currency: years of exclusion or expensive credit, litigation pursued across foreign courts, investment forgone, emigration of the young. And each refusal occurred under conditions the framework can enumerate: a government willing to accept the toll, an administrative apparatus capable of enforcing controls, a public prepared by crisis to share the costs, and, in the island case, an economy small enough that its creditors’ losses were containable.

The audit’s lesson is the opposite of fatalism, and it is the lesson the FAQ on defying markets compresses. Defiance is a purchase, not an impossibility, and the applied analyst’s task is to price it honestly in the specific case: the reserves available to defend a currency during the transition, the maturity profile that determines how quickly creditors can retaliate, the import dependence that sets the public’s pain threshold, the legal exposure of state assets abroad, and the political cohesion required to hold the policy when the toll arrives in the form of shortages and headlines. Governments that priced refusal honestly have sometimes bought it and been vindicated. Governments that announced refusal without the reserves, the controls machinery or the cohesion discovered that the corridor’s walls are load bearing, and their episodes stand in the literature beside the others as the cost schedule’s other column. Structural dependence, to its credit, predicts both columns from the same variables, which is the final reason this article has treated it as the tradition’s most valuable possession: it does not tell governments what will happen to them. It tells them what to count before they decide, and it tells analysts which numbers will decide what happens.

Land, housing and the planning state

A final field, chosen because it is the one in which ordinary citizens meet the state as market maker most directly, confirms the pattern at the scale of the street and the title deed. Housing in the advanced economies is routinely described as a market failure, a realm where private supply and private demand mysteriously fail to meet, as though the market in question were a natural formation whose rules nobody wrote. The institutional record reads differently. What may be built, where, at what density and to what standard is determined by planning law administered by public authorities. The value of land, the asset on which the whole sector’s finance rests, is created in large part by public decisions: the zoning that permits development, the infrastructure that makes a district reachable, the green boundaries that make the permitted plots scarce. The finance that turns those permissions into buildings flows through mortgage markets whose terms are set by central bank policy and whose risks the state has repeatedly underwritten in crisis, while the tenure structure, the balance between owning, private renting and social provision, is the sediment of a century of legislation, subsidy and sale. A state so deeply the author of a market’s preconditions cannot plausibly be described as absent from it, and the housing question, which presents itself to citizens as the purest case of market power, the landlord, the developer, the price, dissolves under institutional inspection into a chain of public decisions, each of which selected among interests: the landowner against the renter, the incumbent homeowner against the newcomer, the construction industry against the treasury’s exposure to a price collapse on which the banking system’s collateral depends.

The framework’s questions organise the field without forcing it. The structural question asks why governments that profess horror at house price inflation preside over its continuation, and answers through dependence: household wealth, bank collateral, construction employment and the consumption that housing equity funds are load bearing elements of the economic order, so the state’s own fiscal and financial stability is mortgaged to the prices its citizens protest, a dependence as binding as any bond market’s and considerably more intimate. The instrumental questions locate the access: the developers’ presence in the drafting of planning reform, the revolving door between ministries and housebuilding firms, the local vetoes through which organised homeowners defend the scarcity that enriches them. The selectivity question explains the reforms that result, perpetually announced as supply revolutions and perpetually drafted so that their costs fall on the unorganised: obligations on developers negotiated away viability by viability, social housing quotas diluted in committee, the green protections of the articulate defended while the rents of the inarticulate rise. And the reconstruction pattern closes the case, since every attempt to solve the housing question by deregulating it has generated new apparatuses of permission, assessment and appeal, more state in the service of a market that was supposed to need less. The citizen who cannot afford a house and the minister who cannot afford a price fall are inhabitants of the same structure, and a theory that lets an analyst describe that structure, its dependencies named and its selectivities mapped, without reaching for conspiracy or for despair, has demonstrated its usefulness in the place where usefulness is hardest to fake.

For students and exam candidates

The examinable core of this article is a set of distinctions, and answers score by deploying them in order rather than by narrating the period. Open with the reconstruction thesis as a named claim and state its evidence in one sentence: liberalisation was legislated, regulators multiplied, central banks were redesigned, so the state was reorganised rather than reduced. Then separate the two classical accounts and attach a prediction to each, since questions on application reward candidates who can say what an instrumentalist would expect about a bailout and what a structuralist would expect, and why the record of rescues across opposed governments favours the structural expectation for the rescue itself while leaving access to explain its internal terms. Authoritarian statism is the highest value named concept in the set, because it is a dated prediction: give the year, the three components, executive concentration, party decline and administrative growth, and one mark against it, legislative or judicial revival, to show the evaluation habit markers reward. Close by handling the variation finding before the examiner raises it: dependence binds as a variable, Garrett and Mosley are the citations, and the corridor metaphor, stated once and used precisely, organises an answer better than any summary of governments behaving badly.

For teachers and seminar leaders

This article teaches best as an exercise in scoring rival accounts against a record the students assemble themselves. Set the application grid as the seminar’s working document with the verdict column blank, and assign each row to a pair of students required to argue the instrumentalist expectation, the structural expectation and the evidence in sequence, so that the class discovers the grid’s mixed scoring rather than receiving it. The authoritarian statism material supports a second exercise in intellectual honesty: distribute the components of the 1978 prediction before revealing the term or its author, and have the seminar grade each component against the intervening decades, including the counter tendencies, before the authorship is disclosed and the discussion turns to what it means for a theory to risk itself in print. For the evidence audit, Garrett against the strong convergence claim and Mosley’s granularity finding make a matched pair of readings short enough for a single week and sharp enough to generate genuine disagreement about where the corridor’s walls stand. The recurring classroom error to hunt is the slide between levels: students will treat structural dependence as a claim about ministers’ motives, and the correction, that the mechanism operates through anticipated reactions and would function with ministers of any motive whatsoever, is the single clarification that most improves the essays that follow.

For researchers, historians, and journalists

For researchers, the article’s exportable content is its procedure rather than its verdicts: state the dependence as variables, derive expectations before the archive, hunt the fossils of anticipated reaction in budget annexes and debt office correspondence, and interview against the claim by seeking officials who used market reaction as a convenient veto. The archival frontier of the field lies in the drafting record of the liberalising decades, where the reconstruction thesis can be tested clause by clause, and in the crisis archives now opening, where the sequencing of rescue decisions against market data can be reconstructed hour by hour; both frontiers reward the historian who arrives with the grid’s questions already written. For journalists, the discipline is one of inference. A market reaction to a budget is evidence about dependence, not about competence, and copy that reports a falling currency as a verdict on a government’s worth has adopted, unexamined, the standpoint of the asset holder, for whom the corridor’s narrowness is indeed good news about discipline. What can legitimately be inferred from a rescue is its structure: which claims were protected first, which were made whole, which were restructured, and on what timetable the costs were distributed to the wider public. What cannot be inferred is motive, coordination or class intention from outcomes alone, and the reporter’s version of the article’s honesty rule is to attribute the structural reading to the framework and its named scholars, to give the treasury’s account of constraint at equal length, and to reserve the word capture for cases where access and outcome are both documented rather than presumed. Analysts in markets and ministries, finally, can use the framework without its vocabulary: corridor mapping, selectivity auditing and the distributive reading of budgets are portable techniques, and the staffer who prices a government’s room to move from its debt structure rather than from its speeches will be right more often than the one who does the reverse.

Frequently Asked Questions

Q: Is Marxist state theory useful for analysing politics today?

Yes, where usefulness means generating expectations that the record can grade, and the article has graded them in the open. The reconstruction thesis predicted rule making growth through liberalisation and was confirmed by friends and strangers alike. Structural dependence predicted that rescue and discipline would follow exposure rather than ideology, and the crisis record fits. Authoritarian statism risked a description of executive concentration in 1978 that the decades largely bore out. The same audit records the failures: strong convergence claims about welfare states were tested and fell. A framework that supplies both columns, and a procedure for telling them apart in new cases, from platform regulation to industrial subsidy, is useful in the only sense scholarship recognises, which is that it improves on the descriptions it replaces and survives being wrong in instructive ways.

Q: Is neoliberalism a retreat of the state?

No, and the question is the article’s organising test, because the answer reorganises everything downstream of it. Liberalisation was accomplished by legislation: privatisation statutes, licensing regimes, competition law, central bank redesign, fiscal rules and enforcement agencies, each an exercise of public power that had to be drafted, staffed and funded. What changed was the state’s position, from owner and director toward rule maker, referee and guarantor, and its purposes, toward the construction and policing of markets. Measures of the supposed retreat refuse to cooperate with the story: public spending shares ended the period near where they began, rule making volume rose, and crisis revealed reserves of fiscal and monetary capacity that a retreating state would not have possessed. Analysts who begin from retreat keep being surprised by the state they actually encounter.

Q: What is authoritarian statism?

Authoritarian statism is the term Nicos Poulantzas gave, at the end of the 1970s, to a transformation he predicted inside formally democratic states: power concentrating in the executive, parliaments declining into ratifying bodies, political parties hollowing into electoral machines, administrative networks multiplying beyond electoral answerability, and formal liberties narrowing through ordinary legislation. The term does not describe dictatorship, and using it as a label for a disliked government destroys its value. It describes a tendency measurable across decades, through indicators such as where legislation is drafted, how much of it arrives settled, party membership and internal party life, and the growth of agencies exercising public power at a remove. Assessed against the decades since publication, the prediction scored strongly on executive concentration and party decline, and less well where legislatures and courts revived as counterweights.

Q: Does capital mobility constrain governments?

Yes, as a variable pressure rather than an iron law, and the variation is the finding that matters. Mobility disciplines through prices: currencies sold, bond yields rising until budgets rewrite themselves, investment postponed, ratings revised. Its force depends on measurable conditions, the currency and holders of the public debt, the size and exposure of the banking system, the breadth of the export base, the depth of reserves. Where dependence is high and mobility unrestricted, the corridor of feasible policy narrows sharply, as the documented reversals of governments that proceeded regardless demonstrate. Where debt is domestic, long term and in a reserve currency, or where capital controls are competently imposed, the corridor widens considerably. Comparative research found the constraint real on macroeconomic stance and tax structure, and much weaker on the composition of welfare and public provision, where domestic institutions and organised constituencies continued to decide.

Q: Why do states bail out banks?

Because the alternative, in a systemic failure, is the collapse of the payments and credit system on which the state’s own revenue, currency and employment base depend, and no government facing that alternative has yet chosen it. The structural explanation fits the pattern: rescues occurred across governments of opposed ideologies and traditions of bank state relations, on scales proportioned to exposure rather than to intimacy, and were frequently executed by governments elected against intervention and accompanied by hostile rhetoric toward the rescued. Access and lobbying, the instrumentalist variables, explain the internal distribution of a rescue: which institutions were saved whole, which creditors protected, which managers removed. They do not explain the rescue’s existence or simultaneity. The terms states imposed, dilution, removal, restructuring, also fit the structural logic, since the object preserved was the system and its functions, not the fortunes of the individuals who had endangered it.

Q: What is the competition state?

The competition state is the form described in the literature of the 1990s, above all by Philip Cerny, for a state reorganised to attract and retain mobile investment: offering enforceable contracts, efficient courts, infrastructure, publicly financed skills, stable money and predictable regulation, while withdrawing older guarantees of outcomes such as full employment and comprehensive income security. The concept matters because it refutes the retreat story from within mainstream scholarship: the competing state does more, in the areas where markets are made and maintained, even as it does less in the areas where citizens were insured. Its instruments converge across countries, corporate tax competition, investment incentives, supply side investment, because states facing the same mobile factor reorganise toward the same attractors. The Marxist reading adds that the supply side investments socialise costs of private production while the abandoned guarantees privatise risk onto households, a distributive claim the budgets can be used to test.

Q: Is austerity a class project?

In its strong form, the claim that consolidation was designed as a coordinated instrument for shrinking social protection, the evidence supports it only in episodes and elements, since programmes varied with creditor exposure and institutions more than with elite enthusiasm. In its weaker form the claim is substantially supported: consolidation as practised distributed its costs by class, protecting the value of financial assets and honouring creditor claims in full while withdrawing income, services and security from those dependent on wages and public provision. The mechanism is legal as much as political. Creditor claims are contractual and sit outside spending reviews, while citizens’ claims are statutory and amendable by the parliament conducting the consolidation, so adjustment travels down the path of least legal resistance. The honest verdict distinguishes motive, which budgets cannot prove, from incidence, which budgets document in detail.

Q: Can a government defy financial markets?

Yes, at a price that varies with its position, and the record of refusal is the best guide to the price list. Governments have imposed capital controls in crises, defaulted and restructured with heavy losses for creditors, refused to compensate foreign depositors, and let outsized banking systems fail, and several emerged with policy autonomy and social provision intact. The survivals share conditions: administrative capacity to enforce controls, reserves to survive the transition, a public prepared for the costs, and debts structured so that exclusion from borrowing is bearable. The failures share the opposite conditions, and the toll in every case was real, expensive credit, litigation, forgone investment, emigration. Defiance is therefore a purchase to be priced in the specific case, from the debt’s currency and maturity, the banking system’s exposure and the state’s import dependence, rather than an impossibility to be lamented or a slogan to be chanted.

Q: What is the strategic relational approach to the state?

The strategic relational approach, developed by Bob Jessop from the 1980s, treats the state neither as an instrument waiting to be captured nor as a structure dictating outcomes, but as an institutional terrain with built in selectivities. State forms, their divisions of competence, rules of access, inherited routines and time horizons, privilege some actors and strategies over others, the way a landscape privileges some routes without determining any journey. Outcomes emerge as actors with unequal resources pursue strategies across that tilted terrain, and actors learn the tilt and adapt, which is why the approach explains variation between similar economies as a product to be analysed rather than an anomaly. Its discipline is to state the selectivity in advance, name the actors it should favour, and identify the evidence that would show the field level or tilted otherwise, which keeps the framework empirical rather than incantatory.

Q: Does globalisation weaken the nation state?

It weakens some capacities, strengthens others, and transforms the rest, which is why the question survives better dissolved than answered. For insulating domestic commitments against capital flight, most states are weaker than their postwar predecessors, largely through their own accumulated legislation. For constructing markets, policing firms and citizens, projecting legal order across borders, operating the payment and reserve infrastructure of the world economy and rescuing financial systems, the leading states are stronger than any predecessors. Globalisation itself was authored by states, through treaties, liberalisation statutes and institutional design, and its crises are repaired by states, which is difficult to reconcile with the obituary the strong globalisation thesis pronounced. What integration genuinely produced is stratification: rule writing states, corridor negotiating states, and states whose corridors creditors effectively administer, a hierarchy machined into reserves, currencies and payment networks rather than imposed by any single decision.

Q: What should an analyst record before calling a bailout a class project?

The analyst should record the counterfactual and the distribution before reaching for the label. First, whether a feasible alternative existed: what refusal would have done to payments, deposits, employment and the public finances, given the banking system’s size relative to the state. Second, the internal terms: which creditors were made whole, which shareholders diluted or wiped out, which managers removed, and what conditions were attached and enforced. Third, the fiscal aftermath: over what period the costs were recovered, and from whom, through which taxes, service reductions or transfers forgone. Fourth, the comparison set: how states with different governments and different bank state relations handled comparable failures. A rescue that no plausible government would have refused, with losses imposed where the system’s logic allowed, supports a structural account with class incidence. The label class project claims more, an intention and a beneficiary by design, and it should be reserved for cases where the record of choices, against available alternatives, sustains it.

Q: How do bond markets discipline a government without issuing orders?

Through prices that rewrite the government’s budget from outside. When holders of a state’s bonds judge its plans unsustainable or uncongenial, they sell or demand higher yields at auction, and the state’s interest bill rises until the arithmetic of its own budget forces revision of the plans the market disliked. No instruction is sent and none is needed, because the finance ministry performs the discipline internally: officials model the likely reaction before announcements, ministers withdraw measures in drafting, and the veto is exercised by an expectation. The mechanism’s granularity is coarse, which limits it. Research on investor behaviour found markets punishing inflation, disorder and unsustainable aggregates while tolerating wide variation in the composition of spending and taxation, so the discipline fixes the budget’s outer frame more reliably than its contents. It also arrives unevenly, charging some governments a premium on arrival, on the strength of expected intentions, and extending others a presumption of soundness that survives until a specific measure destroys it.

Q: Why does authoritarian statism challenge the retreat of the state story?

Because the two accounts make opposite predictions about the same institutions, and the record discriminates between them. The retreat story expects the liberalising decades to disperse public power: thinner executives, revived legislatures as policy returned to politics, a shrinking administrative apparatus. Authoritarian statism, published before those decades began, expected the reverse physiognomy: decision making concentrating in executives and their administrative networks, parties hollowing, legislatures ratifying, agencies multiplying. The decades delivered multiplying agencies, professionalised and shrinking parties, framework legislation drafted in ministries, and summit governance at national and supranational levels, alongside counter tendencies in assertive committees and expanding judicial review that the honest assessment records. A prediction made in advance that fits the institutional record better than the era’s own self description is strong evidence for the framework that generated it, and the strongest single exhibit for the claim that this tradition saw the contemporary state clearly while its obituaries were being written.

Q: What separates a competition state from a welfare state in this analysis?

The direction of the guarantee. A welfare state, in the postwar form, guarantees outcomes to citizens: employment, income in sickness and old age, services as rights of citizenship, and it builds the fiscal and administrative apparatus those guarantees require. A competition state guarantees conditions to investors: contract enforcement, stable money, skilled labour, infrastructure and predictable regulation, and it builds the apparatus those guarantees require. The forms overlap in practice, since most states carry both inheritances, and the analytical use of the distinction is to ask, of any given reform, which guarantee it strengthens and which it thins. Training programmes tied to employer specifications, benefit systems redesigned as activation, tax codes written as bids for mobile capital, each extends the competition guarantee. Provision defended as a right, financed progressively and insulated from market performance, extends the welfare guarantee. States are scored by the balance and its direction of travel, not assigned to either box.

Q: How would you test whether austerity is a class project in one country?

By separating the claim into motive, mechanism and incidence, and testing each against a different record. For incidence, the budget documents suffice: reconstruct, across the consolidation years, the distribution of the adjustment among creditor payments, protected contracts, public investment, transfers, public wages and taxation by income source, and compare the burdens borne across the income distribution using the standard distributional analyses governments themselves publish. For mechanism, examine the legal hierarchy of obligations: which claims were contractual and untouched, which statutory and amended, and whether revenue from mobile wealth was pursued with the same administrative energy as savings from services. For motive, the hardest element, use the drafting record and the timing: measures enacted under cover of the emergency that the emergency did not require, and that outlasted it permanently, support the design reading, while a programme confined to the emergency and reversed as borrowing recovered counts against it. The comparison with a similar country that consolidated differently converts the exercise from narrative into test.