Government Failure

The patterned ways government intervention produces outcomes worse than the market failure it meant to correct, derived by modeling the state as self-interested actors under institutional constraints — capture, rent-seeking, electoral myopia, bureaucratic bloat — and set symmetrically against market failure on one comparative surface.

Core Idea

Government failure is the public-choice and public-economics term for the patterned ways in which government intervention in markets produces outcomes that are inefficient, perverse, or worse than the market failure the intervention was designed to correct — introduced as the symmetric counterpart to market failure to force comparative institutional analysis rather than uncritical prescriptive intervention. The recognized failure modes form a catalogue of mechanisms: regulatory capture, in which regulated industries gain sufficient influence over their regulators that the regulatory apparatus serves industry rather than the public (documented by Stigler 1971); rent-seeking, in which actors divert productive resources toward lobbying and influence to capture policy benefits rather than creating economic value (Tullock 1967); short electoral horizons, in which politicians underinvest in long-run public goods and overweight visible near-term transfers; bureaucratic budget-maximization, in which agencies expand beyond their welfare-optimal size because managers' prestige and compensation rise with budget (Niskanen 1971); information loss across the administrative hierarchy, in which the centre cannot observe what the periphery knows, so centrally designed interventions misfire in local conditions; and the cobra-effect class of incentive backfires, in which the measured and targeted behavior responds to the intervention while the underlying problem migrates or worsens.

The unifying analytical move — developed through the public-choice program of Buchanan, Tullock, and their collaborators — is to treat the government as a set of self-interested actors operating under specific institutional constraints rather than as a benevolent social planner. Once the intervener's own incentive structure is modeled, the symmetry between market failure (the market mechanism producing suboptimal outcomes) and government failure (the political-administrative mechanism producing suboptimal outcomes) becomes the organizing frame: any argument of the form "the market fails, therefore the government should act" must be completed by an assessment of whether the government's likely action fails worse, less, or differently than the market. The framework does not prescribe minimal government — it requires placing both candidate mechanisms on a common evaluation surface.

Structural Signature

Sig role-phrases:

• the social problem — the market failure or welfare shortfall the intervention was meant to correct
• the modeled intervener — the government recast not as a benevolent planner but as a set of self-interested actors under specific institutional constraints (incentives, horizons, information limits)
• the principal-agent gap — the wedge between the institution's mandate and its actors' private agendas, the misalignment from which each pathology follows
• the failure-mode catalogue — the closed list of recurring institutional pathologies: regulatory capture, rent-seeking, electoral myopia, budget-maximizing bureaucracy, centre-periphery information loss, cobra-effect incentive backfires
• the binding misalignment — for any proposed intervention, the specific incentive the design leaves unaligned with the mandate, which the catalogue is run as a checklist to locate
• the comparative-institutional surface — the symmetry with market failure that places both mechanisms side by side, so the prescription reads off "which fails less, here, on this dimension?" rather than a blanket disposition
• the matched-counter set — the standard fixes (independent agencies, sunset clauses, transparency rules, term limits, decentralized provision), each realigning a specific channel and inert against the others
• the two-sided boundary — the discipline the label polices: a real market failure does not by itself license intervention, and a possible government failure does not by itself forbid it; both mechanisms' costs must be priced

What It Is Not

• Not a case for minimal government. The concept does not assert that the state should do less; it demands that the costs of the chosen mechanism be priced. It installs a comparative surface on which a real market failure can still warrant intervention that fails less, so reading it as a brief for laissez-faire mistakes a requirement for symmetry as a disposition against the state.
• Not the claim that intervention is futile. Government failure is possible, not inevitable. The label is built to block the over-reading "the state will botch it anyway" just as firmly as it blocks "intervention is costless because the market failure is real"; a possible government failure does not by itself forbid action any more than a real market failure licenses it.
• Not "markets beat governments." It is not a verdict that one mechanism wins; it is the symmetric counterpart to market failure that places both on a common evaluation surface. The prescription reads off "which mechanism fails less, here, on this dimension?" — neither side gets a free pass, and the framework is silent on which generically prevails.
• Not incompetence or bad luck. The outcomes it names — capture, rent-seeking, electoral myopia, budget-maximizing bureaucracy, information loss, cobra-effect backfires — are recurring institutional pathologies with specific diagnoses (whose incentive is misaligned with the mandate, through what channel), not one-off scandals attributable to the wrong people. The point is that the same self-interested-actor structure generates them predictably.
• Not a standalone cross-domain prime. Strip the state-specific vocabulary — regulators, voters, legislators, public budgets, electoral horizons — and the recognized modes are compositions of principal_agent, collective-action, information_asymmetry, and goodhart_s_law applied to one substrate. Exported to a firm, NGO, or platform, "government failure" is mostly vocabulary borrowing; the portable content flows through those parent primes, not through "government."

Scope of Application

Government failure lives across public economics and the disciplines that study the state — wherever the substrate the catalogue is built from is present (voters, legislators, regulators, agencies, public budgets, electoral horizons); its reach is that one family of institutions. Exported to a firm, NGO, or platform it is mostly vocabulary borrowing — the portable content flows through principal_agent + information_asymmetry + goodhart_s_law + the comparative-institutional symmetry, so non-state institutions fall outside this map.

• Public choice — the home program: Buchanan, Tullock, Niskanen, and Stigler's catalogue of regulatory capture, rent-seeking, budget-maximizing bureaucracy, and political business cycles.
• Political science — principal-agent gaps between voters and elected officials, agency drift in regulatory bodies, and coalition-bargaining inefficiencies.
• Development economics and aid — subsidy distortions, aid-conditionality pathologies, and price-control shortages (fuel, food, rent) as documented intervention backfires.
• Regulatory and administrative law — the capture, revolving-door, and rule-making-bias literature on how agencies drift from their mandate.

Clarity

Naming government failure makes the intervener a modeled object rather than a benevolent stand-in invoked whenever the market falls short. Welfare-economics arguments characteristically run "the market fails here, therefore the government should act"; the concept forces the missing second clause — "and there is reason to believe the government's action will not itself fail in a worse way" — by insisting that the state be analyzed as a set of self-interested actors under specific institutional constraints, with their own incentives, horizons, and information limits. That move dissolves a pervasive asymmetry in policy reasoning, in which market imperfections are scrutinized in detail while the corrective machinery is treated as frictionless and well-intentioned. With the term in hand, a flat preference for intervention (or, symmetrically, for laissez-faire) is exposed as an incomplete argument: the sharper question becomes not "does the market fail?" but "which mechanism fails less, here, on this dimension?" — placing market failure and government failure on a common comparative-institutional surface where neither side gets a free pass.

It also gives a name and a structure to a class of outcomes that otherwise look like bad luck or incompetence. Rent dissipation, regulatory capture, electoral myopia, budget-maximizing bureaucracy, information loss between centre and periphery, and incentive backfires of the cobra-effect kind become recognizable, recurring institutional pathologies rather than one-off scandals — each with a diagnosis (whose incentive is misaligned with the mandate, and how) and a corresponding family of counters such as independent agencies, sunset clauses, transparency rules, and decentralized provision. Crucially, the concept clarifies what it does not assert: it is not a case for minimal government, but a demand that the costs of the chosen mechanism be priced. That keeps the practitioner from sliding between two confusions the label is designed to prevent — assuming intervention is costless because a market failure is real, and assuming intervention is futile because government failure is possible.

Manages Complexity

The historical record of interventions gone wrong is an endless parade of seemingly disconnected episodes — a rent-control ordinance that empties the housing stock, a fuel subsidy that breeds shortages and smuggling, a regulator that ends up working for the industry it polices, an agency that swells past any welfare-justified size, a target-driven program that hits its metric while the underlying problem migrates, a long-run public investment that no politician will fund. Read one at a time, each looks like its own scandal, its own incompetence, its own bad luck, and the policy analyst confronting a fresh proposal has no leverage on the question "will this intervention misfire?" beyond case-by-case suspicion. Government failure compresses that parade by supplying a closed catalogue of recurring institutional pathologies — regulatory capture, rent-seeking, electoral myopia, budget-maximizing bureaucracy, centre-periphery information loss, cobra-effect incentive backfires — each with a sharp diagnosis (whose incentive is misaligned with the mandate, and through what channel) and each with a standard family of counters (independent agencies, sunset clauses, transparency rules, term limits, decentralized provision). A centuries-deep sprawl of bad policy thus reduces to a short, finite list of named mechanisms the analyst can run as a checklist against any proposed intervention, reading the likely failure mode off which incentive the design leaves unaligned rather than re-deriving the pathology from scratch for each case. Beneath the catalogue sits the deeper compression: the single analytical move of modeling the government not as a benevolent planner but as a set of self-interested actors under specific institutional constraints — incentives, horizons, information limits — from which every catalogued mode follows as a consequence of some particular misalignment. That move collapses the whole asymmetric tangle of policy argument onto one comparative surface. Instead of scrutinizing market imperfections in detail while waving the corrective machinery through as frictionless, the analyst places market failure and government failure side by side and reads the prescription off a single comparison: not "does the market fail?" but "which mechanism fails less, here, on this dimension?" The branch structure is correspondingly clean and symmetric — a real market failure does not by itself license intervention (the intervention may fail worse), and a possible government failure does not by itself forbid it (the market may fail worse still) — so the practitioner reasons from the relative magnitude of two priced failure modes rather than from a blanket disposition toward or against the state, and the high-dimensional question of when government should act reduces to locating the binding misalignment on each side and comparing their costs.

Abstract Reasoning

The concept's foundational move is comparative-institutional completion: it intercepts the argument "the market fails here, therefore the government should act" and forces the missing second clause — "and the government's action will not itself fail in a worse way." The reasoning is symmetric and runs in both directions. FROM a documented market failure, the analyst does not infer that intervention is warranted, because the corrective machinery may fail worse; FROM the bare possibility of government failure, the analyst does not infer that intervention should be withheld, because the market may fail worse still. The inference that licenses action is always relative: place both candidate mechanisms on one evaluation surface and reason FROM the relative magnitude of two priced failure modes TO the prescription, asking not "does the market fail?" but "which mechanism fails less, here, on this dimension?" This is the move that the concept exists to install, and it dissolves the asymmetry in which market imperfections get scrutinized while the state is waved through as frictionless.

A diagnostic move runs the catalogue as a screen against any proposed intervention. The analyst reasons FROM the design's incentive structure TO its likely failure mode by locating the binding misalignment: if the regulator is funded by and revolving-doors with the regulated industry, predict capture; if the benefit is concentrated and the cost dispersed, predict rent-seeking effort diverted to capture it; if the payoff is long-run and the electoral horizon short, predict underinvestment in the durable public good; if the agency's prestige and budget rise together, predict expansion past the welfare-optimal size; if the relevant facts live at the periphery and the rule is written at the centre, predict misfire in local conditions; if a single metric is targeted, predict the cobra-effect migration of the problem while the metric is hit. Each prediction is read off which incentive the design leaves unaligned with the mandate, so the analyst names the expected pathology before the intervention is enacted rather than diagnosing it after the scandal.

The interventionist move pairs each diagnosed misalignment with a counter and a predicted effect. The reasoning runs FROM the identified pathology TO the institutional fix that realigns the offending incentive: independent agencies and revolving-door restrictions to blunt capture; sunset clauses and term limits to lengthen the effective horizon against electoral myopia; transparency and contestable rule-making to raise the cost of rent-seeking; decentralized provision to put decisions where the local information is. The predicted payoff is specific to the channel — a sunset clause does nothing about capture, an independent agency does nothing about electoral myopia — so the move is not "add safeguards" generically but "match the counter to the binding misalignment," and its effectiveness is judged by whether it actually moves the incentive that the diagnosis flagged.

The concept's boundary-drawing is unusually explicit and itself a reasoning discipline, because the label is built to prevent two symmetric over-readings. It does not license the inference "intervention is futile" — government failure is possible, not inevitable, and a real market failure can still warrant action that fails less. And it does not license "intervention is costless because the market failure is real" — the chosen mechanism's costs must be priced. The single move that grounds all of this is treating the government as a set of self-interested actors under specific institutional constraints rather than a benevolent planner; the boundary the concept polices is exactly the line between modeling the intervener (in scope, where the catalogue applies) and assuming it away (out of scope, where the whole analysis collapses back into the asymmetry the concept was built to correct).

Knowledge Transfer

Within public economics and the disciplines that study the state, government failure transfers as mechanism. The comparative-institutional completion move, the named-pathology checklist (regulatory capture, rent-seeking, electoral myopia, budget-maximizing bureaucracy, centre-periphery information loss, cobra-effect backfires), the diagnose-the-binding-misalignment step, and the matched-counter intervention logic all carry intact across the home fields because they share the substrate the catalogue is built from — voters, legislators, regulators, agencies, public budgets, electoral horizons. So the framework runs without retuning across public choice (Buchanan, Tullock, Niskanen, Stigler), political science (principal-agent gaps between voters and officials, agency drift), development economics and aid (subsidy distortions, conditionality pathologies, price-control shortages), and regulatory and administrative law (capture, revolving doors, rule-making bias). Across these the analysis is not re-applied by analogy; it is the same model of the intervener as self-interested-actors-under-institutional-constraints, the same symmetry with market failure, and the same pathology catalogue, applied to different organs of the same kind of institution. The transfer is gated on the state substrate being present, and within it the diagnostic and its counters travel exactly.

Beyond the state the honest report is case (B): the structural skeleton genuinely recurs across coordinating institutions, but it travels through the parent primes, not through "government failure." Two things must be separated. First, the catalogue's distinctive force is locked to its substrate: regulatory capture, electoral myopia, budget-maximizing bureaucracy, and pork-barrel allocation presuppose regulators, elections, public budgets, and legislators, so importing the named modes wholesale into a firm, an NGO, a platform, or a household is mostly vocabulary borrowing — the words do not survive the strip-the-jargon test. Second, and more usefully, there is a real substrate-independent structure underneath, and it is the genuinely portable lesson: every coordinating mechanism — market, state, firm, community, platform — has characteristic failure modes, and comparative institutional analysis requires putting all candidates' failure modes on the table together. That principle travels, and government failure is one named slice of it (the institutional counterpart of market_failure). But where the skeleton recurs across substrates, the load-bearing structure is the family of primes the catalogue is composed of: principal_agent gaps (mandate versus the actor's private agenda), collective-action problems on the agent side, information_asymmetry and information loss across hierarchical levels, incentive-compatibility failures, and — for the cobra-effect / target-driven class — goodhart_s_law and the cobra-effect pattern. Strip the state-specific terms and what remains is exactly those primes applied to one substrate; the recognized government-failure modes are compositions of them, which is why the cross-domain content "flows through" the underlying primes rather than through "government" as a substrate. The home-bound cargo government failure leaves behind is its institutional specificity — the electoral horizon, the regulatory agency, the public budget, the revolving door, the voter-as-principal — and the corresponding standard counters (independent agencies, sunset clauses, transparency rules, term limits) that presuppose exactly those organs. So the correct cross-domain lesson carries principal_agent + collective-action + information_asymmetry + the symmetry principle that every mechanism has priced failure modes (the candidate generalization institutional failure modes / intervention pathologies), not the named concept; "government failure," exported to a non-state institution, is at best the parent principal-agent-plus-information-loss analysis wearing political vocabulary. That is precisely why it is a domain-specific abstraction — canonical inside political economy, decomposing into existing primes outside it (see Structural Core vs. Domain Accent).

Examples

Canonical

George Stigler's 1971 "The Theory of Economic Regulation" is the seminal demonstration of a government-failure mode. Against the view that regulators impartially serve the public, Stigler argued — and marshaled evidence — that regulation is often acquired by the industry it governs and operated for that industry's benefit, because concentrated producers have far stronger incentives to organize and lobby than dispersed consumers. The paradigm case is the Civil Aeronautics Board, which for decades set U.S. airline fares and routes and blocked new entrants, protecting incumbent carriers' profits at travelers' expense until deregulation in 1978. The regulator meant to discipline the industry had come to serve it. Modeling regulators and firms as self-interested actors, rather than assuming a benevolent agency, is what made the outcome predictable rather than surprising.

Mapped back: Recasting the regulator as a self-interested actor is the modeled intervener; capture is one entry in the failure-mode catalogue. The drift of the CAB from public mandate to industry service is the principal-agent gap, and the asymmetry between organized producers and dispersed consumers is the binding misalignment the diagnosis locates.

Applied / In Practice

Rent control is a widely studied instance of government failure. Introduced to fix a real affordability problem — a housing market pricing lower-income tenants out — rent ceilings have repeatedly produced a worse allocation than the shortfall they targeted: landlords cut maintenance on capped units, convert rentals to condominiums, and curtail new construction, so the rental stock shrinks and deteriorates while queues and side-payments replace the price mechanism. A study of San Francisco's controls by Diamond, McQuade, and Qian found that protected tenants benefited but landlords' responses reduced the city's rental supply and pushed rents up elsewhere. The intervention hit its nominal target — capped rents on existing units — while the underlying scarcity migrated and intensified.

Mapped back: The affordability shortfall is the social problem the intervention meant to correct. The supply contraction and quality decline are a cobra-effect entry in the failure-mode catalogue, and asking whether that outcome is worse than the affordability gap it addressed is exactly the comparative-institutional surface — pricing the chosen mechanism's failure rather than assuming the corrective is costless.

Structural Tensions

T1: Symmetry versus the two dispositions (anti-state and futility). The concept installs a comparative surface on which neither "the market fails, therefore intervene" nor "the government fails, therefore do not" is a complete argument — its entire discipline is symmetry. But that symmetry is unstable in use: it slides into an anti-state disposition (read as a standing case for minimal government) or, mirror-image, into fatalism (intervention is futile because failure is possible). The tension is that the construct polices a razor's edge between two dispositions it was built to prevent, and its very name — failure — tilts the unwary reader toward the anti-state side. Diagnostic: Is the analysis actually pricing both mechanisms' failure modes and comparing them here, or has it collapsed into a blanket disposition for or against the state?

T2: Modeling the intervener versus caricaturing it. The foundational move — treat government as self-interested actors under institutional constraints rather than a benevolent planner — is what makes the pathologies predictable and corrects welfare economics' frictionless-corrector asymmetry. But the model is itself a choice that can overshoot: assume every official is purely self-interested and you predict failure everywhere, erasing the cases where mandate and incentive genuinely align. The tension is that the analytical move which fixes the benevolent-planner asymmetry can, pushed hard, install the opposite one — the knave-only intervener whose failure is foreordained by assumption rather than diagnosed from structure. Diagnostic: Is the intervener modeled with its actual incentive structure, or caricatured as uniformly self-interested so that failure is guaranteed before any analysis?

T3: Closed catalogue versus its substrate-boundedness. The named-pathology checklist — capture, rent-seeking, electoral myopia, budget-maximizing bureaucracy, centre-periphery information loss, cobra-effect backfires — compresses centuries of bad policy into a finite screen the analyst runs against any proposal. But each mode presupposes the state substrate: regulators, elections, public budgets, revolving doors. The tension is that the catalogue's diagnostic force and its portability trade off directly — the more state-specific each named mode, the sharper it is in situ and the less it survives export, so importing "regulatory capture" or "electoral myopia" into a firm or platform is mostly vocabulary borrowing that fails the strip-the-jargon test. Diagnostic: Does the case actually have the organs each named mode presupposes (elections, regulators, public budgets), or is the mode being applied as a metaphor to a substrate that lacks them?

T4: Matched counter versus channel-specific inertness. Each standard fix realigns exactly one channel — an independent agency blunts capture, a sunset clause lengthens the horizon against myopia, transparency raises the cost of rent-seeking, decentralization puts decisions where the information is — and does nothing for the others. This precision is the point: the move is "match the counter to the binding misalignment," not "add safeguards" generically. But the same specificity means a well-intentioned safeguard aimed at the wrong channel is inert, and stacking safeguards adds cost and complexity without touching the binding misalignment. The tension is that what makes counters effective (channel-specificity) is what makes mis-targeted ones useless. Diagnostic: Does the proposed safeguard realign the specific incentive the diagnosis flagged as binding, or a different channel that is not the one actually failing here?

T5: A comparative surface versus commensurability. The framework's payoff is placing market failure and government failure side by side and reading off "which fails less, here, on this dimension." But that presupposes the two failures can be priced on a common surface, and they frequently resist commensuration — a market's deadweight loss, a regulator's capture-driven distortion, and a rights or distributional cost are measured in different currencies, on different dimensions, over different horizons. The tension is that the comparison the concept demands is exactly the step it cannot guarantee is well-defined, so "which mechanism fails less?" can be genuinely indeterminate even after both failure modes are correctly named. Diagnostic: Are the two mechanisms' failures actually commensurable on the dimension being compared, or is a common evaluation surface being asserted where no shared metric exists?

T6: Autonomy versus reduction (a political-economy concept or its parent primes). Government failure is canonical inside public choice and public economics, carrying its state-specific cargo — the electoral horizon, the regulatory agency, the public budget, the revolving door, the voter-as-principal, and the counters (independent agencies, sunset clauses, term limits) that presuppose exactly those organs. Strip that vocabulary and the recognized modes are compositions of principal_agent gaps, collective-action problems, information_asymmetry, and goodhart_s_law applied to one substrate, over the symmetry principle that every coordinating mechanism has priced failure modes. Diagnostic: Resolve toward those parent primes (plus the comparative-institutional symmetry) when the case is a non-state institution — firm, NGO, platform, household; toward government failure when the state substrate and its named pathology catalogue are genuinely present.

Structural–Framed Character

Government failure sits on the framed side of the spectrum — best read as framed-leaning: a normatively-tinged analytical concept wholly constituted by the human institution of the state, whose portable core is a composition of agent-and-information primes with no observer-free structure beneath it. On evaluative_weight it reads framed, though with a disciplined qualification: the word "failure" (and "pathology") is evaluative, and the concept is a critique-enabling frame, yet the entry insists its whole discipline is symmetry — it prices the chosen mechanism's costs and compares, refusing to be a verdict against the state; so it carries real normative freight held under a comparative rule, more than a neutral mechanism but less than a flat condemnation. On human_practice_bound it reads strongly framed: the concept is constituted by the state and dissolves without it — strip away voters, regulators, legislators, agencies, public budgets, and electoral horizons and the named modes have nothing to attach to; its very subject is a human institution. Institutional_origin is likewise framed: government failure is a designed construct of the public-choice program (Buchanan, Tullock, Stigler, Niskanen), and what it analyzes is itself a set of human institutions, not a fact of unpeopled nature. On vocab_travels it reads framed: regulatory capture, electoral myopia, rent-seeking, budget-maximizing bureaucracy, the revolving door are pinned to the state, and the entry is explicit that exporting them to a firm or platform is "mostly vocabulary borrowing" that fails the strip-the-jargon test. And on import_vs_recognize the transfer is the entry's explicit "case-B" pattern — within the state-studying disciplines (public choice, political science, development economics, administrative law) the framework is recognized intact across different organs of the same institution, but beyond the state it travels only as the parent primes, "government failure" exported being at best a principal-agent-plus-information-loss analysis wearing political vocabulary.

There is no observer-free structural core here: the recognized failure modes are, as the entry states, compositions of principal_agent gaps, collective-action problems, information_asymmetry, and goodhart_s_law applied to the one state substrate, sitting over the comparative-institutional symmetry principle (every coordinating mechanism has priced failure modes; government failure is the institutional counterpart of market_failure). That composition and that symmetry principle are what genuinely travel, but they are exactly what government failure instantiates from its parents, not what makes "government failure" itself travel: the cross-domain reach belongs to the principal-agent / collective-action / information-asymmetry / Goodhart composition and the "every mechanism has priced failure modes" principle, while the electoral horizon, the regulatory agency, the public budget, the voter-as-principal, and the matched counters (independent agencies, sunset clauses, term limits) stay home. Its character: a symmetry-disciplined but normatively-tinged, state-constituted analytical frame — framed-leaning by its evaluative "failure" register, its institutional subject and origin, and its state-pinned vocabulary, structural only in the composition of agent-and-information primes it instantiates under the comparative-institutional symmetry it shares with market failure.

Structural Core vs. Domain Accent

This section decides why government failure is a domain-specific abstraction and not a prime, and carries the case for its domain-specificity in one place.

What is skeletal (could lift toward a cross-domain prime). Strip the state and one thin relational structure survives — and here it is a composition rather than a single core: a coordinating mechanism whose actors' private agendas diverge from its mandate, who cannot see what the periphery knows, and who respond to targets rather than to the underlying problem, produces patterned suboptimal outcomes — and any argument for switching to it must price its own failure modes alongside the one it replaces. The portable pieces are abstract — a gap between mandate and agent (principal_agent), a collective-action asymmetry on the agent side, an information gap across levels (information_asymmetry), a target that displaces its objective (goodhart_s_law), and a comparative rule that every candidate mechanism has priced failure modes (the symmetry with market_failure). Nothing there requires voters or regulators. That composition and that symmetry principle are genuinely substrate-portable — every coordinating mechanism, market, firm, community, or platform, has characteristic failure modes — which is why the recognized government-failure modes are exactly compositions of those primes applied to one substrate. But that composition is what government failure shares, not what makes it government failure.

What is domain-bound. Almost all the distinctive content is state furniture, and none of it survives extraction. The intervener is not generic — it is the government modeled as self-interested regulators, legislators, and agencies. The failure-mode catalogue is worked political economy — regulatory capture (Stigler), rent-seeking (Tullock), electoral myopia, budget-maximizing bureaucracy (Niskanen), centre-periphery information loss, cobra-effect backfires — each presupposing regulators, elections, public budgets, or a legislature. Its counters are state-specific instruments — independent agencies, sunset clauses, transparency rules, term limits, decentralized provision — that presuppose exactly those organs, and its worked cases (the Civil Aeronautics Board, rent control) are interventions by a state. The decisive test, which the entry states outright: strip the state-specific vocabulary — regulators, voters, legislators, public budgets, electoral horizons — and the named modes fail the strip-the-jargon test; what remains is the bare principal-agent-plus-information-loss-plus-Goodhart composition, a looser thing that any coordinating institution can exhibit.

Why this does not clear the prime bar. A prime's vocabulary travels and its transfer is recognition of the same mechanism, not analogy. Government failure's transfer is bimodal. Within the state-studying disciplines it moves intact — the comparative-institutional completion move, the named-pathology checklist, the diagnose-the-binding-misalignment step, and the matched-counter logic all run without retuning across public choice, political science, development economics, and administrative law, because these study different organs of the same kind of institution and share its substrate. That is genuine within-domain mechanism transfer. Beyond the state it travels only as vocabulary borrowing: exporting "regulatory capture" or "electoral myopia" to a firm, NGO, or platform is at best a principal-agent-plus-information-loss analysis wearing political dress, and the words do not survive the substrate change. And when the bare cross-domain lesson is wanted — every coordinating mechanism has characteristic failure modes, and comparative institutional analysis must put all candidates' failure modes on the table together — it is already carried, in more general form, by the parent composition (principal_agent, collective-action, information_asymmetry, goodhart_s_law) under the "every mechanism has priced failure modes" symmetry it shares with market_failure. The cross-domain reach belongs to those parents and that principle; "government failure," as named, carries the electoral horizon, the regulatory agency, the public budget, the voter-as-principal, and the state-specific counters that should stay home.

Relationships to Other Abstractions

Current abstraction Government Failure Domain-specific

Parents (6) — more general patterns this builds on

• Government Failure presupposes Public Choice Domain-specific

  Government Failure depends on Public Choice's move of modeling the state as self-interested role occupants rather than a benevolent unitary planner.

• Government Failure is part of, conditional Goodhart's Law Prime

  Its cobra-effect branch contains Goodhart's proxy-under-pressure mechanism, while capture and electoral-myopia branches need not target a proxy.

• Government Failure is part of, conditional Information Asymmetry Prime

  Government Failure contains Information Asymmetry in its centre-periphery, regulator-industry, and voter-official information-loss branches.

• Government Failure is part of, conditional Regulatory Capture Prime

  Regulatory Capture is an internal branch of the Government Failure catalogue when regulated interests redirect the agency meant to constrain them.

• Government Failure is a decomposition of Agency Problem Prime

  Removing state vocabulary leaves the mandate-versus-private-agenda wedge of the live Agency Problem, which already aliases principal-agent.

• Government Failure is a decomposition of Comparison Prime

  Its portable analytical operation places market and state mechanisms on one failure-cost surface and reads off which fails less in the case at hand.

Children (2) — more specific cases that build on this

• Fiscal Illusion Domain-specific is a kind of, conditional Government Failure

  Fiscal Illusion is a Government Failure when financing-induced cost opacity makes public spending worse than the informed comparative benchmark.

• Rent-Seeking Trap Domain-specific is a kind of, conditional Government Failure

  A Rent-Seeking Trap is a Government Failure when state rules create and defend the rent, but analogous institutional-allocation traps need not be governmental.

Hierarchy paths (17) — routes to 12 parentless roots

• Government Failure → Public Choice → Institution → Normativity → Constraint

Show alternative paths (16)

Not to Be Confused With

• Market failure. The symmetric counterpart the concept was coined against — the market mechanism itself producing inefficient outcomes (externalities, public goods, monopoly, information problems). Government failure is not a subtype of it but its paired opposite on the same comparative-institutional surface: the corrective machinery failing rather than the market failing. Tell: which mechanism is producing the suboptimal outcome — the price system left alone (market failure), or the political-administrative intervention meant to fix it (government failure)?
• State failure / the failed state. The political-science concept of a state that has lost the capacity to perform core functions — monopoly on violence, territorial control, basic service delivery — and slides toward collapse. This is a contrast case, not the same idea: government failure concerns a functioning state whose interventions misfire because its actors are self-interested under institutional constraints, not a state that has ceased to function. Tell: is the state intact but its policy producing perverse outcomes (government failure), or has the state's basic capacity itself collapsed (state failure)?
• Regulatory capture. One entry in the failure-mode catalogue (Stigler), where a regulated industry comes to control its regulator — not the whole concept. Treating capture as synonymous with government failure mistakes one catalogued pathology for the family that also includes rent-seeking, electoral myopia, budget-maximizing bureaucracy, information loss, and cobra-effect backfires. Tell: is the diagnosed misalignment specifically industry-over-regulator (capture), or one of the other channels the catalogue is run as a checklist to locate (some other government-failure mode)?
• The principal-agent problem (a parent prime it composes from). The substrate-neutral prime naming the wedge between a principal's mandate and an agent's private agenda. It is a parent, not a peer: government failure is largely principal_agent (plus collective action, information_asymmetry, and goodhart_s_law) applied to the state substrate. Tell: strip away regulators, voters, and public budgets and what remains is the bare mandate-versus-agenda gap — the parent prime, treated more fully elsewhere, not government failure.
• Goodhart's law / the cobra effect (parent prime for one class). The prime that a measure targeted as an objective ceases to be a good measure, so the metric is hit while the underlying problem migrates. It supplies the incentive-backfire class within the catalogue but is broader than the state: it fires in firms, schools, and platforms alike. Tell: is the point that a targeted metric displaces its objective generally (goodhart_s_law / cobra effect), or that a government intervention did so as one entry among the state's pathologies (that catalogue mode of government failure)?
• Public choice theory. The research program (Buchanan, Tullock, Niskanen, Stigler) that models political actors as self-interested and, in doing so, produced the concept of government failure. The program is the analytical apparatus; government failure is one of its central outputs — the catalogue of intervention pathologies it derives. Tell: are you naming the method of applying economic self-interest assumptions to politics (public choice), or the specific class of botched-intervention outcomes that method predicts (government failure)?

Neighborhood in Abstraction Space

Government Failure sits in a sparse region of the domain-specific corpus (70^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Public Choice & Policy Failure (5 abstractions)

Nearest neighbors

• Public Choice — 0.85
• Policy Design — 0.85
• Rent-Seeking Trap — 0.82
• Agenda Building — 0.82
• Tullock Paradox — 0.82

Computed from structural-signature embeddings · 2026-07-12