Full Text
MINISTRY OF ADVANCED DESIGN, RESEARCH AND DEVELOPMENT
Companies maintain research and development departments to improve products, yet almost no government maintains a permanent institution whose full-time mission is to improve government. Our paper proposes a Ministry of Advanced Design: a permanent institution mandated to measure, redesign, pilot, and improve every government institution and every government procedure, from local government offices to national digital government services, and obligated to improve the ministry itself, so the institution created to remove bureaucracy can never become bureaucracy. Citizens cannot exit government services, so competitive pressure never forces procedures to improve; tax compliance consumes nearly two percent of gross domestic product, and administrative burdens fall hardest on the poorest citizens. Existing remedies, including innovation laboratories and digital service units, remain narrow and advisory, and several units were dissolved when political priorities changed. Through literature synthesis and comparative analysis, our paper specifies five mechanisms: a public register measuring every procedure's time, cost, and error rate; service redesign around life events under a once-only data rule; proactive delivery of entitlements; mandatory pilots before national rollout; and shared digital infrastructure reused by every agency. Four empowerment instruments give the ministry enforceable standing: measurement rights, binding service standards, design review before funding, and published metrics. Elected governments keep full authority over policy goals. Anticipated findings include savings of several percentage points of public spending, higher benefit take-up, and a trust dividend treated as a testable hypothesis. Our paper recommends building the ministry first, because every reform presupposes an institution able to design institutions.
1. INTRODUCTION: THE MISSING RESEARCH AND DEVELOPMENT DEPARTMENT OF GOVERNMENT
Every large manufacturer maintains a research and development department, because products left undesigned fall behind. Governments operate the largest and most consequential service portfolios in every country, spending amounts approaching half of national output in many developed economies, yet almost no government maintains a permanent, empowered institution whose only product is a better government. Improvement happens, when improvement happens at all, through episodic reform commissions, temporary innovation laboratories, or heroic individual projects, and each of the three forms dies young. Our paper proposes to close the gap with a Ministry of Advanced Design: a permanent institution of ministry rank whose mandate is to measure, research, redesign, pilot, and continuously improve every other government institution and every government procedure, from the smallest local government office to the largest national digital government service, and whose mandate explicitly includes the ministry itself.
The intellectual warrant for treating government improvement as a design discipline comes from Herbert Simon, who defined the science of design in universal terms: "Everyone designs who devises courses of action aimed at changing existing situations into preferred ones." Simon insisted the intellectual activity producing material artifacts does not differ fundamentally from the activity devising a social welfare policy for a state, which places government procedures squarely inside design's domain. The word "advanced" in the ministry's name signals state-of-the-art method: service design, operations research, behavioral science, software engineering, statistics, and artificial intelligence, applied with the same seriousness manufacturers apply to products. The research question of our paper follows directly: can a permanent, self-improving, whole-of-government design institution reduce the cost and raise the quality of public administration more reliably than the episodic and fragmented remedies used today?
Our paper proceeds as follows. Section 2 establishes the structural problem. Section 3 reviews existing remedies and documents the fragility gap. Section 4 defines the proposed institution. Section 5 specifies five delivery mechanisms. Section 6 specifies the empowerment instruments and democratic boundaries. Section 7 answers objections. Section 8 defines anticipated findings and the evaluation procedure. Section 9 concludes with the build-order argument.
2. THE STRUCTURAL PROBLEM: WHY GOVERNMENT DOES NOT IMPROVE ITSELF
2.1 No exit, no pressure. Private companies improve because dissatisfied customers leave. Albert Hirschman's classic framework holds that exit dominates as the corrective force in competitive markets, and in the extreme case of a total monopoly the value of exit is severely diminished; Hirschman named the resulting pathology the "lazy public monopoly". Government is the limiting case: citizens cannot exit the passport office, and government agencies will never go bankrupt or get defunded as a result of project failure. Charles Wolf generalized the observation into a theory of nonmarket failure, arguing a theory of public policy requires an account of government failure symmetrical to market failure, and identifying four sources of nonmarket failure: internalities, redundant and rising costs, derived externalities, and distributional inequity. "Redundant and rising costs" is the academically named disease our proposal treats. Where the disciplining mechanism is structurally absent, improvement pressure must be created deliberately, by an institution built for the purpose.
2.2 The cost of standing still. The costs of unimproved administration are measured, large, and rising. William Baumol's cost disease predicts rising relative costs wherever human-hour services fail to gain productivity, and the cost and price behavior of the affected services has persisted for as long as statistical data exist, in some cases well over a century, with none of the counteracting programs adopted by various countries having succeeded. Compliance overhead alone is enormous: the Tax Foundation estimates 7.9 billion hours of annual tax compliance in the United States, worth roughly 413 billion dollars in lost productivity plus 133 billion dollars of out-of-pocket costs, totaling 546 billion dollars, nearly two percent of gross domestic product, and a European Parliament study reports business compliance cost estimates between 1.2 and 1.8 percent of gross domestic product. Old systems devour the budgets meant to replace old systems: the United States Government Accountability Office reports more than 100 billion dollars of annual federal information technology spending, of which about 80 percent goes to operations and maintenance of existing systems, and only three of the ten most critical legacy systems flagged in 2019 had been modernized by February 2025. The prize for fixing the pattern is proportionate: a United Kingdom government review found 45 to 87 billion pounds per year of unrealised savings and productivity benefits, equivalent to 4 to 7 percent of public sector spend.
2.3 The human cost. Administrative burden research conceptualizes citizens' encounters with the state as a function of learning, psychological, and compliance costs, and documents two facts central to our argument. First, burdens are often chosen: administrative burdens are not simply unintended byproducts of governance but the result of deliberate policy choices. Second, burdens are regressive: people with the fewest resources and the greatest needs struggle most to overcome burdens, so the frictions reinforce existing inequality, and cognitive research sharpens the point into a trap, since the poor, the sick, and the elderly constitute the largest groups targeted by government programs, leaving many citizens in a human capital catch-22 when interacting with the state. Time itself is the tax: the United Kingdom's Digital Efficiency Report priced an hour spent interacting with government at 14.70 pounds for the average citizen, while a digital transaction costs roughly twenty times less than a telephone transaction, thirty times less than a postal transaction, and fifty times less than a face-to-face transaction. Good design eliminates whole categories of burden: the United States Social Security program, engineered from the start to minimize claimant effort, achieves a take-up rate of about 100 percent because the administration tracks earnings and calculates benefits automatically. Sunstein's sludge research generalizes the remedy, defining sludge as excessive or unjustified frictions such as paperwork burdens and recommending routine sludge audits by government agencies to catalogue costs and decide when and how to reduce sludge.
3. EXISTING REMEDIES AND THE FRAGILITY GAP
3.1 The fragments exist. The recognition of the problem is global, and remedies exist in fragments. Public sector innovation laboratories multiplied worldwide; the first comprehensive academic mapping identified thirty-five such organizations across the globe. Digital service units produced audited returns: the United Kingdom Government Digital Service reported 3.56 billion pounds saved over three years, and Estonia's prime minister summarized the national experience in one line: "Estonia saves two percent of GDP by signing things digitally." Behavioural teams scaled into a movement: the Organisation for Economic Co-operation and Development has mapped over 300 institutions applying behavioural science to public policy in 63 countries, including more than 200 inside the governments of 56 countries, and the flagship experiments were published in leading journals, with social norm messages in reminder letters raising payment rates for overdue tax across field experiments covering more than 200,000 individuals. The United Arab Emirates even reached for the ministry frame, launching the world's first virtual ministry applying design thinking and experimentation, the Ministry of Possibilities, in April 2019, though the ministry is virtual, has no minister, and is managed by cabinet members collectively.
3.2 The fragments die. The same record documents systematic mortality. The Observatory of Public Sector Innovation summarizes two decades as the emergence, spread and occasionally death of public innovation labs across the globe. Denmark's MindLab, the world's first and most famous public sector innovation lab, shut down after sixteen years following a swift shift in political priorities, and a former director drew the structural lesson: in the public sector a handful of people, sometimes one pen stroke, decides whether an institution exists. The United States 18F team was eliminated on the cusp of an eleventh anniversary, deemed non-essential, despite having completed 455 projects for 34 agencies plus the Executive Office of the President, the Library of Congress and the House of Representatives. Academic lifecycle research confirms the pattern beyond anecdote: comparative study of cross-departmental laboratories from birth to death found closures at Australia's DesignGov and the Helsinki Design Lab, and the global mapping study found media presence did not prevent close-downs; the most significant survival factor was chief executive support, and loss of support through the political process triggered the end.
3.3 The gap, stated precisely. Fragments exist at unit level: audits, laboratories, digital teams, behavioural squads. Sunstein comes closest to our proposal by suggesting dedicated sludge reduction teams, on the model of nudge units, yet even the closest proposal stops at team scale. No state combines the three properties our paper argues are jointly necessary: permanent ministry rank, whole-of-government scope covering every institution and procedure, and a recursive mandate requiring the improver to measure and improve the improver. One apparent counterexample deserves honest treatment. Estonia's celebrated digital state was built through informal cooperative networks, and the leading academic case study argues the success relied on networks and design principles not being institutionalised, while the most significant challenge Estonia now faces is whether such foundations can deliver the next evolution. The counterexample therefore concedes the core point: informality can build a first generation, and informality struggles to sustain improvement across political generations. Sustained improvement is exactly the property a permanent ministry supplies.
4. THE PROPOSAL: MINISTRY OF ADVANCED DESIGN
4.1 Definition. The Ministry of Advanced Design is a permanent institution of ministry rank whose sole product is improvement of government. The ministry runs one cycle, continuously and forever: measure every government procedure; research causes of cost, delay, and error; redesign; pilot at small scale; roll out what survives evidence; monitor; repeat. The mandate covers every institution and every procedure of the state, explicitly including local government offices, national ministries, and digital government services, and explicitly including the Ministry of Advanced Design itself.
4.2 The recursive mandate. Recursion is not ornament; recursion is the anti-bureaucracy guarantee. An institution demanding efficiency from every other institution must publish the same metrics about the institution's own operations, or legitimacy collapses. Recursion also breaks a documented reflex: when large government failures occur, the standard response is to spend more money and add more bureaucracy rather than reevaluate why the existing bureaucracy did not prevent the failure. A ministry whose own budget and staffing face the same published scrutiny the ministry applies to others cannot quietly grow into the disease the ministry was created to cure.
4.3 Organizational template. For internal design, our paper borrows the best-evidenced template for high-performing public research organizations, the ARPA model, whose key elements are organizational flexibility on an administrative level and significant authority given to program directors to design programs, select projects and actively manage projects. The ministry's program units are led by empowered directors on fixed terms, staffed by multidisciplinary design teams, and evaluated on published outcome metrics. Dynamic-capabilities scholarship supplies the theoretical frame: leading work identifies dynamic capabilities in the public sector as perhaps the key missing element in the new generation of innovation policies, and pandemic-era analysis lists the needed capacities directly, including capacity to adapt and learn, capacity to align public services and citizen needs, and capacity to govern data and digital platforms. The Ministry of Advanced Design is the organizational home such capabilities have lacked. Within the accepted taxonomy of public sector innovation, which distinguishes new service, administrative process, technological process, conceptual and governance innovation, the ministry institutionalizes administrative and technological process innovation as a permanent state function, completing the trajectory digital-era governance scholarship began when Dunleavy and colleagues argued the new public management wave had stalled or reversed, chiefly through cumulative increases in institutional and policy complexity, and proposed reintegration, needs-based holism, and digitization in place of the older paradigm.
5. FIVE DELIVERY MECHANISMS
5.1 The public procedure register. No one can state, for any country today, how many government procedures exist, how long each takes, what each costs, and how often each fails. The register repairs the blindness: every procedure of every institution appears with measured completion time, citizen hours consumed, unit cost, step count, error rate, and satisfaction score, published openly and refreshed continuously. The register operationalizes Sunstein's program, whose sludge audits quantify burdens by measuring time, money, and steps, identify unnecessary requirements, assess costs against benefits, examine who bears the burden, and publish findings for accountability, and whose lookback principle demands periodic review of the existing stock of burdens to eliminate the outmoded and pointless. Methodology for measuring output and productivity in government exists and is documented at book length; the leading treatment shows productivity in government, long treated as static, should be rising rapidly under digital change, with the decisive factors including management practices, information technology use, organizational culture, strategic mis-decisions and political churn. League tables built on the register create the internal competitive pressure the state otherwise lacks; sunset review deletes procedures unable to justify measured cost; the worst-ranked procedures receive ministry redesign help first.
5.2 Life-event redesign under a once-only rule. The once-only principle, in the European Commission's canonical definition, requires citizens and businesses to supply the same information only once to a public administration, with an estimated annual net savings of as much as 5 billion euros per year from extension across the European Union. Working national implementations already return measured value: the United Kingdom's bereavement service saves over 20 million pounds a year by sharing data across organisations including local government, with decade-scale benefits estimated at 255.8 million pounds between 2011 and 2021, split between government and public. The design destination has a name and a validated stage model: the no-stop shop, in which the citizen performs no action and fills no form to receive a government service, progressing along three dimensions of data collection, data storage, and data use. The ministry's redesign teams organize services around life events, birth, moving, marriage, bereavement, starting a business, and drive each event toward the no-stop stage.
5.3 Proactive delivery. Proactive government inverts the direction of service: from the pull approach, whereby the citizen must seek out services, toward a push model delivering just-in-time services shaped around needs, preferences, and circumstance. Working cases exist on three continents. Austria has issued family allowance automatically upon the birth of a child since 2015, with no citizen action required. Estonia switched family benefits to proactive delivery in October 2019, after which parents receive a proposal from the government instead of applying, the birth entry in the population register activating all follow-on services; before the redesign only 15 percent of benefits were offer-based and manual processing took up to 65 minutes per application, while today 99.99 percent of registered births are checked automatically for eligibility. The equity payoff is explicit in the Estonian framing: proactive services ensure vulnerable groups, such as the elderly and low-income families, do not miss benefits through ignorance, confusion, or stigma, a direct answer to the regressivity documented in Section 2.3.
5.4 Pilot-first policy engineering. Large public projects fail in a specific, statistically documented way: fat tails. The landmark study of 1,471 projects found an average cost overrun of 27 percent masking the real danger, with one project in six a Black Swan carrying a 200 percent cost overrun and almost 70 percent schedule overrun, and the public sector companion analysis found an 18 percent probability of a project spinning out of control, with typical Black Swan overruns of 130 percent. Catastrophes at national scale illustrate the tail: the United Kingdom National Health Service record program was canceled after at least 9.8 billion pounds with only two percent of promised benefits realized, and the FiReControl consolidation was canceled after 469 million pounds against an initial 100 million pound plan. The countermeasure is procedural: nothing rolls out nationally without a small pilot, randomized where feasible, and every deployment carries a mandatory post-implementation review with rollback authority. Government experimentation at scale is proven practice, codified in the Test, Learn, Adapt methodology and demonstrated by tax-letter field experiments in which more than 3 million pounds arrived within a 23-day window attributable to redesigned messages at trivial cost. Pilots convert catastrophic tail risk into cheap, fast learning.
5.5 Shared infrastructure and upstream law design. Every agency separately procuring identity, payments, notifications, and data exchange multiplies cost and failure. The alternative carries an established name, government as a platform, under which the public sector provides shared components and a common infrastructure supporting efficiency and reuse. The returns are documented at national scale: Estonia's X-Road exchange layer handled more than 1.3 billion queries in a year, of which citizen-submitted queries were only 3 percent, with automated queries saving over 1,100 working years annually; the United Kingdom Government Digital Service's spend-control and platform work contributed to the multi-billion-pound savings cited in Section 3.1. Shared infrastructure also rescues local governments, which individually can never afford national-grade systems. The final mechanism operates upstream, before laws are born: Denmark has required, since 2018, all new legislation before parliament to be digital-ready, with a mandatory impact assessment of digital implementation consequences accompanying every draft, run through a secretariat reviewing roughly 200 legislative proposal summaries months before public consultation; the screening has covered over 800 laws and 340 proposals, while pre-2018 legislation remains largely unreviewed, and the model is spreading, with the United Kingdom's 2025 blueprint committing to work towards all legislation being digital-ready, drawing on Denmark's approach. Inside the Ministry of Advanced Design, the legislative review function attaches a complexity cost estimate to every bill, preventing bureaucracy at the source instead of demolishing bureaucracy after construction.
6. EMPOWERMENT: FOUR INSTRUMENTS AND THE DEMOCRATIC BOUNDARY
Advisory bodies die ignored; Section 3.2 is the obituary collection. The Advanced Design Ministry therefore holds four administrative instruments. First, measurement rights: a legal entitlement to operational data from any government procedure, the fuel of the register. Second, binding service standards: floor requirements for speed, error, accessibility, and data reuse, which agencies must meet. Third, a design review gate: new administrative systems and procedures cannot receive funding before passing ministry review, generalizing the United Kingdom spend-control precedent, where controls on digital and technology spending saved 391 million pounds in a single year, and echoing the stage-gated funding discipline of the United States Technology Modernization Fund, where agencies compete for funding, sharpening problem definition, and funding is released in stages against milestones. Fourth, a publication duty: every metric the ministry computes about others, and about the ministry, is public by default.
The democratic boundary is equally explicit. Elected governments keep full authority over policy goals: what to provide, to whom, at what generosity. The ministry holds authority only over implementation quality: how fast, how cheap, how accurately, how humanely the chosen policy is delivered. The division answers the concentration-of-power objection structurally, and the lean-government literature explains why the institutional home matters at all: improvement methods transplanted as toolkits fail, since implementation of Lean in public services focused on technical tools without an over-arching logic to validate the tools. A permanent institution carries the logic; scattered projects cannot.
7. OBJECTIONS AND ANSWERS
7.1 "Reorganizations waste money." True as documented: the United Kingdom National Audit Office counted over 90 central government reorganisations in four years, unable to demonstrate value for money because objectives were vague and costs and benefits went untracked, at almost 200 million pounds per year on average, with 25 departments created since 1980 of which 13 no longer exist; companion research estimates at least 15 million pounds of cost for each wholly new department and warns of a productivity dip and a minimum two-year wait before concrete benefits. Three answers. The documented waste describes rushed, unplanned, unmeasured political churn, the precise opposite of a metrics-first permanent institution; the cited studies themselves blame absent planning and absent measurement, the two things the ministry exists to supply. A fifteen-million-pound setup cost stands against a documented prize of 45 to 87 billion pounds per year in one country alone. And the ministry must pass a version of the ministry's own design review at birth: published baseline metrics, a piloted organizational design, and a pre-registered evaluation plan, applying recursion from day one.
7.2 "Estonia succeeded without such a ministry." Answered in Section 3.3: the strongest academic study of the Estonian case concedes the informal-network model now faces a sustainability problem the model cannot solve alone. First generations can ride enthusiasm; permanence requires institutions.
7.3 "Improvement programs failed before." Lean, total quality management, and reinvention waves failed in the documented pattern of tools without institutional logic. The ministry is not another toolkit; the ministry is the standing owner of the logic, the memory, and the evaluation discipline the failed waves lacked.
7.4 "The trust payoff is speculative." Partly conceded, and deliberately so. The causal arrow between service performance and political trust is contested; the classic analysis warns the mechanistic reading only partly corresponds to reality, because existing trust also shapes perceived performance. Our paper therefore states the trust dividend as a testable hypothesis, not a promise, and Section 8 names the measurement instrument.
7.5 "Baumol implies services cannot improve." The critical literature answers directly, arguing the cost-disease frame misconceptualizes public services and fails to recognize evidence of substantial scope for productivity improvement. Our proposal needs only the weaker claim: administrative processing, unlike bedside nursing, automates without destroying any relational value, so the ministry automates administration aggressively while treating human-contact services with piloted care.
8. ANTICIPATED FINDINGS AND EVALUATION PROCEDURE
As a position paper, our paper states expectations, open parameters, and the procedure by which the expectations can be falsified. Outcome variables: the procedure register itself supplies the primary panel, citizen hours per life event, unit transaction costs, error rates, and take-up rates, with channel-cost ratios and compliance-hour totals benchmarked against the figures of Section 2. Anticipated magnitudes: recoverable administrative waste of several percentage points of public spending, anchored by the 4 to 7 percent of public sector spend estimate and the near-two-percent compliance figure; take-up movement toward the Social Security benchmark in proactively delivered entitlements; per-transaction cost collapse toward the digital channel ratios. The trust hypothesis is measurable on an existing cross-national instrument: the Organisation for Economic Co-operation and Development Trust Survey, which fielded nearly 60,000 responses across 30 countries in the second wave and 76,927 valid responses across 38 countries in the third, measuring reliability, responsiveness, integrity, openness and fairness, the exact dimensions the ministry's mechanisms target. Open parameters for national adaptation: constitutional rank versus statutory agency status; phasing of the four instruments; the legal basis of measurement rights under national data protection law; and small-state versus federal-state variants of the register.
9. CONCLUSION: BUILD THE DESIGNER FIRST
Government improvement today is an accident of political weather: a lab blooms, a team ships, an election arrives, a pen stroke ends the experiment, and the procedures grind on unmeasured. The evidence assembled above supports one structural conclusion. The returns to improvement are large and documented; the existing carriers of improvement are small and mortal; and the missing element is not method, money, or talent, but an institution: permanent, empowered, whole-of-government, and pointed at every procedure including the institution's own. A country that builds the Ministry of Advanced Design first acquires the capability every subsequent reform presupposes, because every ambitious public project, fiscal, digital, or social, is at bottom an act of institutional design, and acts of design go better with a designer. Time returned to citizens is life returned to citizens; the ministry is the machine that returns the time, and the machine measures the machine.
Companies maintain research and development departments to improve products, yet almost no government maintains a permanent institution whose full-time mission is to improve government. Our paper proposes a Ministry of Advanced Design: a permanent institution mandated to measure, redesign, pilot, and improve every government institution and every government procedure, from local government offices to national digital government services, and obligated to improve the ministry itself, so the institution created to remove bureaucracy can never become bureaucracy. Citizens cannot exit government services, so competitive pressure never forces procedures to improve; tax compliance consumes nearly two percent of gross domestic product, and administrative burdens fall hardest on the poorest citizens. Existing remedies, including innovation laboratories and digital service units, remain narrow and advisory, and several units were dissolved when political priorities changed. Through literature synthesis and comparative analysis, our paper specifies five mechanisms: a public register measuring every procedure's time, cost, and error rate; service redesign around life events under a once-only data rule; proactive delivery of entitlements; mandatory pilots before national rollout; and shared digital infrastructure reused by every agency. Four empowerment instruments give the ministry enforceable standing: measurement rights, binding service standards, design review before funding, and published metrics. Elected governments keep full authority over policy goals. Anticipated findings include savings of several percentage points of public spending, higher benefit take-up, and a trust dividend treated as a testable hypothesis. Our paper recommends building the ministry first, because every reform presupposes an institution able to design institutions.
1. INTRODUCTION: THE MISSING RESEARCH AND DEVELOPMENT DEPARTMENT OF GOVERNMENT
Every large manufacturer maintains a research and development department, because products left undesigned fall behind. Governments operate the largest and most consequential service portfolios in every country, spending amounts approaching half of national output in many developed economies, yet almost no government maintains a permanent, empowered institution whose only product is a better government. Improvement happens, when improvement happens at all, through episodic reform commissions, temporary innovation laboratories, or heroic individual projects, and each of the three forms dies young. Our paper proposes to close the gap with a Ministry of Advanced Design: a permanent institution of ministry rank whose mandate is to measure, research, redesign, pilot, and continuously improve every other government institution and every government procedure, from the smallest local government office to the largest national digital government service, and whose mandate explicitly includes the ministry itself.
The intellectual warrant for treating government improvement as a design discipline comes from Herbert Simon, who defined the science of design in universal terms: "Everyone designs who devises courses of action aimed at changing existing situations into preferred ones." Simon insisted the intellectual activity producing material artifacts does not differ fundamentally from the activity devising a social welfare policy for a state, which places government procedures squarely inside design's domain. The word "advanced" in the ministry's name signals state-of-the-art method: service design, operations research, behavioral science, software engineering, statistics, and artificial intelligence, applied with the same seriousness manufacturers apply to products. The research question of our paper follows directly: can a permanent, self-improving, whole-of-government design institution reduce the cost and raise the quality of public administration more reliably than the episodic and fragmented remedies used today?
Our paper proceeds as follows. Section 2 establishes the structural problem. Section 3 reviews existing remedies and documents the fragility gap. Section 4 defines the proposed institution. Section 5 specifies five delivery mechanisms. Section 6 specifies the empowerment instruments and democratic boundaries. Section 7 answers objections. Section 8 defines anticipated findings and the evaluation procedure. Section 9 concludes with the build-order argument.
2. THE STRUCTURAL PROBLEM: WHY GOVERNMENT DOES NOT IMPROVE ITSELF
2.1 No exit, no pressure. Private companies improve because dissatisfied customers leave. Albert Hirschman's classic framework holds that exit dominates as the corrective force in competitive markets, and in the extreme case of a total monopoly the value of exit is severely diminished; Hirschman named the resulting pathology the "lazy public monopoly". Government is the limiting case: citizens cannot exit the passport office, and government agencies will never go bankrupt or get defunded as a result of project failure. Charles Wolf generalized the observation into a theory of nonmarket failure, arguing a theory of public policy requires an account of government failure symmetrical to market failure, and identifying four sources of nonmarket failure: internalities, redundant and rising costs, derived externalities, and distributional inequity. "Redundant and rising costs" is the academically named disease our proposal treats. Where the disciplining mechanism is structurally absent, improvement pressure must be created deliberately, by an institution built for the purpose.
2.2 The cost of standing still. The costs of unimproved administration are measured, large, and rising. William Baumol's cost disease predicts rising relative costs wherever human-hour services fail to gain productivity, and the cost and price behavior of the affected services has persisted for as long as statistical data exist, in some cases well over a century, with none of the counteracting programs adopted by various countries having succeeded. Compliance overhead alone is enormous: the Tax Foundation estimates 7.9 billion hours of annual tax compliance in the United States, worth roughly 413 billion dollars in lost productivity plus 133 billion dollars of out-of-pocket costs, totaling 546 billion dollars, nearly two percent of gross domestic product, and a European Parliament study reports business compliance cost estimates between 1.2 and 1.8 percent of gross domestic product. Old systems devour the budgets meant to replace old systems: the United States Government Accountability Office reports more than 100 billion dollars of annual federal information technology spending, of which about 80 percent goes to operations and maintenance of existing systems, and only three of the ten most critical legacy systems flagged in 2019 had been modernized by February 2025. The prize for fixing the pattern is proportionate: a United Kingdom government review found 45 to 87 billion pounds per year of unrealised savings and productivity benefits, equivalent to 4 to 7 percent of public sector spend.
2.3 The human cost. Administrative burden research conceptualizes citizens' encounters with the state as a function of learning, psychological, and compliance costs, and documents two facts central to our argument. First, burdens are often chosen: administrative burdens are not simply unintended byproducts of governance but the result of deliberate policy choices. Second, burdens are regressive: people with the fewest resources and the greatest needs struggle most to overcome burdens, so the frictions reinforce existing inequality, and cognitive research sharpens the point into a trap, since the poor, the sick, and the elderly constitute the largest groups targeted by government programs, leaving many citizens in a human capital catch-22 when interacting with the state. Time itself is the tax: the United Kingdom's Digital Efficiency Report priced an hour spent interacting with government at 14.70 pounds for the average citizen, while a digital transaction costs roughly twenty times less than a telephone transaction, thirty times less than a postal transaction, and fifty times less than a face-to-face transaction. Good design eliminates whole categories of burden: the United States Social Security program, engineered from the start to minimize claimant effort, achieves a take-up rate of about 100 percent because the administration tracks earnings and calculates benefits automatically. Sunstein's sludge research generalizes the remedy, defining sludge as excessive or unjustified frictions such as paperwork burdens and recommending routine sludge audits by government agencies to catalogue costs and decide when and how to reduce sludge.
3. EXISTING REMEDIES AND THE FRAGILITY GAP
3.1 The fragments exist. The recognition of the problem is global, and remedies exist in fragments. Public sector innovation laboratories multiplied worldwide; the first comprehensive academic mapping identified thirty-five such organizations across the globe. Digital service units produced audited returns: the United Kingdom Government Digital Service reported 3.56 billion pounds saved over three years, and Estonia's prime minister summarized the national experience in one line: "Estonia saves two percent of GDP by signing things digitally." Behavioural teams scaled into a movement: the Organisation for Economic Co-operation and Development has mapped over 300 institutions applying behavioural science to public policy in 63 countries, including more than 200 inside the governments of 56 countries, and the flagship experiments were published in leading journals, with social norm messages in reminder letters raising payment rates for overdue tax across field experiments covering more than 200,000 individuals. The United Arab Emirates even reached for the ministry frame, launching the world's first virtual ministry applying design thinking and experimentation, the Ministry of Possibilities, in April 2019, though the ministry is virtual, has no minister, and is managed by cabinet members collectively.
3.2 The fragments die. The same record documents systematic mortality. The Observatory of Public Sector Innovation summarizes two decades as the emergence, spread and occasionally death of public innovation labs across the globe. Denmark's MindLab, the world's first and most famous public sector innovation lab, shut down after sixteen years following a swift shift in political priorities, and a former director drew the structural lesson: in the public sector a handful of people, sometimes one pen stroke, decides whether an institution exists. The United States 18F team was eliminated on the cusp of an eleventh anniversary, deemed non-essential, despite having completed 455 projects for 34 agencies plus the Executive Office of the President, the Library of Congress and the House of Representatives. Academic lifecycle research confirms the pattern beyond anecdote: comparative study of cross-departmental laboratories from birth to death found closures at Australia's DesignGov and the Helsinki Design Lab, and the global mapping study found media presence did not prevent close-downs; the most significant survival factor was chief executive support, and loss of support through the political process triggered the end.
3.3 The gap, stated precisely. Fragments exist at unit level: audits, laboratories, digital teams, behavioural squads. Sunstein comes closest to our proposal by suggesting dedicated sludge reduction teams, on the model of nudge units, yet even the closest proposal stops at team scale. No state combines the three properties our paper argues are jointly necessary: permanent ministry rank, whole-of-government scope covering every institution and procedure, and a recursive mandate requiring the improver to measure and improve the improver. One apparent counterexample deserves honest treatment. Estonia's celebrated digital state was built through informal cooperative networks, and the leading academic case study argues the success relied on networks and design principles not being institutionalised, while the most significant challenge Estonia now faces is whether such foundations can deliver the next evolution. The counterexample therefore concedes the core point: informality can build a first generation, and informality struggles to sustain improvement across political generations. Sustained improvement is exactly the property a permanent ministry supplies.
4. THE PROPOSAL: MINISTRY OF ADVANCED DESIGN
4.1 Definition. The Ministry of Advanced Design is a permanent institution of ministry rank whose sole product is improvement of government. The ministry runs one cycle, continuously and forever: measure every government procedure; research causes of cost, delay, and error; redesign; pilot at small scale; roll out what survives evidence; monitor; repeat. The mandate covers every institution and every procedure of the state, explicitly including local government offices, national ministries, and digital government services, and explicitly including the Ministry of Advanced Design itself.
4.2 The recursive mandate. Recursion is not ornament; recursion is the anti-bureaucracy guarantee. An institution demanding efficiency from every other institution must publish the same metrics about the institution's own operations, or legitimacy collapses. Recursion also breaks a documented reflex: when large government failures occur, the standard response is to spend more money and add more bureaucracy rather than reevaluate why the existing bureaucracy did not prevent the failure. A ministry whose own budget and staffing face the same published scrutiny the ministry applies to others cannot quietly grow into the disease the ministry was created to cure.
4.3 Organizational template. For internal design, our paper borrows the best-evidenced template for high-performing public research organizations, the ARPA model, whose key elements are organizational flexibility on an administrative level and significant authority given to program directors to design programs, select projects and actively manage projects. The ministry's program units are led by empowered directors on fixed terms, staffed by multidisciplinary design teams, and evaluated on published outcome metrics. Dynamic-capabilities scholarship supplies the theoretical frame: leading work identifies dynamic capabilities in the public sector as perhaps the key missing element in the new generation of innovation policies, and pandemic-era analysis lists the needed capacities directly, including capacity to adapt and learn, capacity to align public services and citizen needs, and capacity to govern data and digital platforms. The Ministry of Advanced Design is the organizational home such capabilities have lacked. Within the accepted taxonomy of public sector innovation, which distinguishes new service, administrative process, technological process, conceptual and governance innovation, the ministry institutionalizes administrative and technological process innovation as a permanent state function, completing the trajectory digital-era governance scholarship began when Dunleavy and colleagues argued the new public management wave had stalled or reversed, chiefly through cumulative increases in institutional and policy complexity, and proposed reintegration, needs-based holism, and digitization in place of the older paradigm.
5. FIVE DELIVERY MECHANISMS
5.1 The public procedure register. No one can state, for any country today, how many government procedures exist, how long each takes, what each costs, and how often each fails. The register repairs the blindness: every procedure of every institution appears with measured completion time, citizen hours consumed, unit cost, step count, error rate, and satisfaction score, published openly and refreshed continuously. The register operationalizes Sunstein's program, whose sludge audits quantify burdens by measuring time, money, and steps, identify unnecessary requirements, assess costs against benefits, examine who bears the burden, and publish findings for accountability, and whose lookback principle demands periodic review of the existing stock of burdens to eliminate the outmoded and pointless. Methodology for measuring output and productivity in government exists and is documented at book length; the leading treatment shows productivity in government, long treated as static, should be rising rapidly under digital change, with the decisive factors including management practices, information technology use, organizational culture, strategic mis-decisions and political churn. League tables built on the register create the internal competitive pressure the state otherwise lacks; sunset review deletes procedures unable to justify measured cost; the worst-ranked procedures receive ministry redesign help first.
5.2 Life-event redesign under a once-only rule. The once-only principle, in the European Commission's canonical definition, requires citizens and businesses to supply the same information only once to a public administration, with an estimated annual net savings of as much as 5 billion euros per year from extension across the European Union. Working national implementations already return measured value: the United Kingdom's bereavement service saves over 20 million pounds a year by sharing data across organisations including local government, with decade-scale benefits estimated at 255.8 million pounds between 2011 and 2021, split between government and public. The design destination has a name and a validated stage model: the no-stop shop, in which the citizen performs no action and fills no form to receive a government service, progressing along three dimensions of data collection, data storage, and data use. The ministry's redesign teams organize services around life events, birth, moving, marriage, bereavement, starting a business, and drive each event toward the no-stop stage.
5.3 Proactive delivery. Proactive government inverts the direction of service: from the pull approach, whereby the citizen must seek out services, toward a push model delivering just-in-time services shaped around needs, preferences, and circumstance. Working cases exist on three continents. Austria has issued family allowance automatically upon the birth of a child since 2015, with no citizen action required. Estonia switched family benefits to proactive delivery in October 2019, after which parents receive a proposal from the government instead of applying, the birth entry in the population register activating all follow-on services; before the redesign only 15 percent of benefits were offer-based and manual processing took up to 65 minutes per application, while today 99.99 percent of registered births are checked automatically for eligibility. The equity payoff is explicit in the Estonian framing: proactive services ensure vulnerable groups, such as the elderly and low-income families, do not miss benefits through ignorance, confusion, or stigma, a direct answer to the regressivity documented in Section 2.3.
5.4 Pilot-first policy engineering. Large public projects fail in a specific, statistically documented way: fat tails. The landmark study of 1,471 projects found an average cost overrun of 27 percent masking the real danger, with one project in six a Black Swan carrying a 200 percent cost overrun and almost 70 percent schedule overrun, and the public sector companion analysis found an 18 percent probability of a project spinning out of control, with typical Black Swan overruns of 130 percent. Catastrophes at national scale illustrate the tail: the United Kingdom National Health Service record program was canceled after at least 9.8 billion pounds with only two percent of promised benefits realized, and the FiReControl consolidation was canceled after 469 million pounds against an initial 100 million pound plan. The countermeasure is procedural: nothing rolls out nationally without a small pilot, randomized where feasible, and every deployment carries a mandatory post-implementation review with rollback authority. Government experimentation at scale is proven practice, codified in the Test, Learn, Adapt methodology and demonstrated by tax-letter field experiments in which more than 3 million pounds arrived within a 23-day window attributable to redesigned messages at trivial cost. Pilots convert catastrophic tail risk into cheap, fast learning.
5.5 Shared infrastructure and upstream law design. Every agency separately procuring identity, payments, notifications, and data exchange multiplies cost and failure. The alternative carries an established name, government as a platform, under which the public sector provides shared components and a common infrastructure supporting efficiency and reuse. The returns are documented at national scale: Estonia's X-Road exchange layer handled more than 1.3 billion queries in a year, of which citizen-submitted queries were only 3 percent, with automated queries saving over 1,100 working years annually; the United Kingdom Government Digital Service's spend-control and platform work contributed to the multi-billion-pound savings cited in Section 3.1. Shared infrastructure also rescues local governments, which individually can never afford national-grade systems. The final mechanism operates upstream, before laws are born: Denmark has required, since 2018, all new legislation before parliament to be digital-ready, with a mandatory impact assessment of digital implementation consequences accompanying every draft, run through a secretariat reviewing roughly 200 legislative proposal summaries months before public consultation; the screening has covered over 800 laws and 340 proposals, while pre-2018 legislation remains largely unreviewed, and the model is spreading, with the United Kingdom's 2025 blueprint committing to work towards all legislation being digital-ready, drawing on Denmark's approach. Inside the Ministry of Advanced Design, the legislative review function attaches a complexity cost estimate to every bill, preventing bureaucracy at the source instead of demolishing bureaucracy after construction.
6. EMPOWERMENT: FOUR INSTRUMENTS AND THE DEMOCRATIC BOUNDARY
Advisory bodies die ignored; Section 3.2 is the obituary collection. The Advanced Design Ministry therefore holds four administrative instruments. First, measurement rights: a legal entitlement to operational data from any government procedure, the fuel of the register. Second, binding service standards: floor requirements for speed, error, accessibility, and data reuse, which agencies must meet. Third, a design review gate: new administrative systems and procedures cannot receive funding before passing ministry review, generalizing the United Kingdom spend-control precedent, where controls on digital and technology spending saved 391 million pounds in a single year, and echoing the stage-gated funding discipline of the United States Technology Modernization Fund, where agencies compete for funding, sharpening problem definition, and funding is released in stages against milestones. Fourth, a publication duty: every metric the ministry computes about others, and about the ministry, is public by default.
The democratic boundary is equally explicit. Elected governments keep full authority over policy goals: what to provide, to whom, at what generosity. The ministry holds authority only over implementation quality: how fast, how cheap, how accurately, how humanely the chosen policy is delivered. The division answers the concentration-of-power objection structurally, and the lean-government literature explains why the institutional home matters at all: improvement methods transplanted as toolkits fail, since implementation of Lean in public services focused on technical tools without an over-arching logic to validate the tools. A permanent institution carries the logic; scattered projects cannot.
7. OBJECTIONS AND ANSWERS
7.1 "Reorganizations waste money." True as documented: the United Kingdom National Audit Office counted over 90 central government reorganisations in four years, unable to demonstrate value for money because objectives were vague and costs and benefits went untracked, at almost 200 million pounds per year on average, with 25 departments created since 1980 of which 13 no longer exist; companion research estimates at least 15 million pounds of cost for each wholly new department and warns of a productivity dip and a minimum two-year wait before concrete benefits. Three answers. The documented waste describes rushed, unplanned, unmeasured political churn, the precise opposite of a metrics-first permanent institution; the cited studies themselves blame absent planning and absent measurement, the two things the ministry exists to supply. A fifteen-million-pound setup cost stands against a documented prize of 45 to 87 billion pounds per year in one country alone. And the ministry must pass a version of the ministry's own design review at birth: published baseline metrics, a piloted organizational design, and a pre-registered evaluation plan, applying recursion from day one.
7.2 "Estonia succeeded without such a ministry." Answered in Section 3.3: the strongest academic study of the Estonian case concedes the informal-network model now faces a sustainability problem the model cannot solve alone. First generations can ride enthusiasm; permanence requires institutions.
7.3 "Improvement programs failed before." Lean, total quality management, and reinvention waves failed in the documented pattern of tools without institutional logic. The ministry is not another toolkit; the ministry is the standing owner of the logic, the memory, and the evaluation discipline the failed waves lacked.
7.4 "The trust payoff is speculative." Partly conceded, and deliberately so. The causal arrow between service performance and political trust is contested; the classic analysis warns the mechanistic reading only partly corresponds to reality, because existing trust also shapes perceived performance. Our paper therefore states the trust dividend as a testable hypothesis, not a promise, and Section 8 names the measurement instrument.
7.5 "Baumol implies services cannot improve." The critical literature answers directly, arguing the cost-disease frame misconceptualizes public services and fails to recognize evidence of substantial scope for productivity improvement. Our proposal needs only the weaker claim: administrative processing, unlike bedside nursing, automates without destroying any relational value, so the ministry automates administration aggressively while treating human-contact services with piloted care.
8. ANTICIPATED FINDINGS AND EVALUATION PROCEDURE
As a position paper, our paper states expectations, open parameters, and the procedure by which the expectations can be falsified. Outcome variables: the procedure register itself supplies the primary panel, citizen hours per life event, unit transaction costs, error rates, and take-up rates, with channel-cost ratios and compliance-hour totals benchmarked against the figures of Section 2. Anticipated magnitudes: recoverable administrative waste of several percentage points of public spending, anchored by the 4 to 7 percent of public sector spend estimate and the near-two-percent compliance figure; take-up movement toward the Social Security benchmark in proactively delivered entitlements; per-transaction cost collapse toward the digital channel ratios. The trust hypothesis is measurable on an existing cross-national instrument: the Organisation for Economic Co-operation and Development Trust Survey, which fielded nearly 60,000 responses across 30 countries in the second wave and 76,927 valid responses across 38 countries in the third, measuring reliability, responsiveness, integrity, openness and fairness, the exact dimensions the ministry's mechanisms target. Open parameters for national adaptation: constitutional rank versus statutory agency status; phasing of the four instruments; the legal basis of measurement rights under national data protection law; and small-state versus federal-state variants of the register.
9. CONCLUSION: BUILD THE DESIGNER FIRST
Government improvement today is an accident of political weather: a lab blooms, a team ships, an election arrives, a pen stroke ends the experiment, and the procedures grind on unmeasured. The evidence assembled above supports one structural conclusion. The returns to improvement are large and documented; the existing carriers of improvement are small and mortal; and the missing element is not method, money, or talent, but an institution: permanent, empowered, whole-of-government, and pointed at every procedure including the institution's own. A country that builds the Ministry of Advanced Design first acquires the capability every subsequent reform presupposes, because every ambitious public project, fiscal, digital, or social, is at bottom an act of institutional design, and acts of design go better with a designer. Time returned to citizens is life returned to citizens; the ministry is the machine that returns the time, and the machine measures the machine.