Competent Democracy

Changing the political system for remarkably better execution and accountability

"Many forms of Government have been tried, and will be tried in this world of sin and woe. No one pretends that democracy is perfect or all-wise. Indeed it has been said that democracy is the worst form of Government except for all those other forms that have been tried from time to time.…" — Winston Churchill

Draft 1.5 © 2023–2026 All rights reserved

What is Competent Democracy

Competent Democracy is a vision and solution to the limitations inherent in traditional democratic frameworks, characterized by short-term focus, consensus-building challenges, and a deficit in accountability. It advocates for the establishment of measurable, expert-approved goals prior to their public dissemination. The model is designed to counteract the sway of populism and the pitfalls of unattainable promises, ensuring a governance system grounded in realism and practicality. Emphasizing the necessity of continuity and resilience, Competent Democracy safeguards the persistence of long-term strategies, irrespective of political transitions, thus preventing the derailment of essential plans due to fluctuating political landscapes. Moreover, it introduces a system of sanctions for failures to meet set objectives or fulfill promises, reinforcing the commitment to follow through on political pledges. At the heart of this model lies the strategic deployment of AI technology, which enables an unbiased, data-centric evaluation of progress and provides the public with transparent, immediate insights into governmental operations.

Competent Democracy aims to restore public trust in governance, ensuring that elected leaders are genuinely committed to delivering on their promises and effectively addressing the needs and challenges of their society. It is for you, for your family, friends and our entire society.

Introduction

The history of democracy spans thousands of years. The ancient Greeks are often credited with inventing democracy. In Athens, around 500 BCE, a system of direct democracy was established, where all male citizens over the age of 18 could participate in the Assembly, which was responsible for passing laws and making decisions on important issues. However, this system only applied to a small portion of the population, as women, slaves, and foreigners were excluded from political participation.

Today most democracies in the world are representative democracies, in which all citizens can vote and are represented by political parties and politicians that form a legislative body that elects a government. Many countries implement the same system at a municipal and sometimes regional level.

Democracy brought tremendous improvement to the quality of life of billions of people ensuring freedom, equality, and justice for all citizens. While we are living in an unprecedented era of prosperity relative to anytime in history, the feeling that modern democracies do not execute well is a sentiment that is shared by many individuals around the world.

Modern democracies are facing problems that are undermining their efficacy and credibility and make them unfit for purpose. Three of these problems are particularly acute: a focus on short-term popular agendas, a difficulty to reach consensus and to get from decision to actual execution, and a lack of accountability.

The first problem, a focus on short-term popular agendas, is a consequence of the competitive nature of modern democracies. Politicians are often under pressure to deliver quick results and win popularity with voters. This can lead to a narrow focus on short-term goals, such as improving the economy or creating jobs, rather than addressing longer-term and global challenges like climate change, income inequality, or social justice. This short-term focus can also lead to a lack of investment in public infrastructure, education, and health, which are crucial to long-term prosperity and well-being.

The second problem, difficulty in reaching consensus and getting from decision to actual execution, is another significant challenge facing modern democracies. In many cases, decision-making processes are slow, cumbersome, and prone to gridlock. This is because modern democracies are often fragmented, with multiple political parties and interest groups competing for influence and power. This can make it difficult to reach consensus on important issues, resulting in a lack of action and progress. Even when decisions are made, they can be subject to delays and roadblocks, such as legal challenges, bureaucratic hurdles, or resistance from interest groups.

The third problem, a lack of accountability, is a significant issue facing modern democracies. Elected officials are often held accountable only once every few years, during elections. This means that they may be less responsive to the needs and concerns of their constituents in between elections. Moreover, political leaders and decision-makers may be shielded from scrutiny and accountability by complex bureaucracies, opaque decision-making processes, and a lack of transparency. This can result in a sense of disempowerment and disillusionment among citizens, who feel that their voices are not being heard.

The issues facing modern democracies are often multifaceted and interconnected, requiring a great deal of expertise, coordination, and resources to address. This complexity can lead to a lack of clarity about who is responsible for addressing problems, and can create confusion and delays in decision-making.

A British think tank Institute for Government found that the UK government had only fully delivered on 39% of its major manifesto pledges in the period between 2010 and 2017. In many countries, the public is highly divided and often it is difficult to form stable coalitions. In presidential democracies, there is often slim or no majority to the ruling party in parliament. This phenomena can be seen all over the world:

• Belgium: divided along linguistic lines, with Dutch-speaking Flanders and French-speaking Wallonia, and this has led to complex coalition-building. It took more than a year and a half to form a government after the 2019 elections due to disagreements over issues such as migration and climate policy.
• Italy: has a long history of political fragmentation and instability, with frequent changes of government and parties. The need to form coalitions in order to govern has been a key factor in this instability, as parties struggle to find common ground.
• India: With its diverse population and regional politics, forming a coalition government in India can be challenging. It is not uncommon for multiple parties to join together to form a government, with negotiations often taking several weeks or even months.
• United States: While the US operates under a two-party system, gridlock and political polarization can make it difficult for parties to work together and form a coalition. This can result in government shutdowns and other political stalemates.

In many democracies, politicians are known for making promises to their constituents during election campaigns but then failing to follow through on those promises once they are elected. Examples include:

• Justin Trudeau: During his campaign for the Canadian Prime Minister in 2015, Trudeau promised to balance the federal budget by 2019, but the deficit continued to grow under his leadership. He was elected again in 2019.
• Boris Johnson: During his campaign for the leadership of the UK's Conservative Party and later during the general election of 2019, Johnson promised to "get Brexit done" and negotiate a new trade deal with the EU by the end of 2020. However, the negotiations took longer than expected and the UK ultimately left the EU with a last-minute trade deal in December 2020, which critics say does not deliver the benefits that were promised.
• Narendra Modi: The Prime Minister of India made several promises during his 2014 election campaign, including creating 10 million new jobs a year, cleaning up the Ganges river, and providing electricity to all households in India. However, many of these promises were not met during his first term as Prime Minister. Nevertheless, he was re-elected in 2019 with a larger majority.
• Emmanuel Macron: The President of France made several promises during his 2017 election campaign, including reforming the labor market, reducing unemployment, and boosting economic growth. While his government has made progress on some of these promises, such as labor market reforms, it has faced criticism for not doing enough to address poverty and inequality. Despite these challenges, Macron's party won a majority in the 2017 parliamentary election.
• Matteo Renzi: The former Prime Minister of Italy made several promises during his 2014 election campaign, including labor market reforms, reducing bureaucracy, and increasing investments. However, his government was criticized for not doing enough to address unemployment and economic stagnation. Nevertheless, his party won a majority in the 2018 general election.
• Viktor Orban: The Prime Minister of Hungary made several promises during his 2014 election campaign, including reducing taxes, creating jobs, and improving healthcare. However, his government has been criticized for restricting media freedom, undermining the judiciary, and cracking down on civil society. Despite these challenges, Orban's party won a majority in the 2018 parliamentary election.
• Mariano Rajoy: The former Prime Minister of Spain made several promises during his 2011 election campaign, including creating jobs, reducing the budget deficit, and improving healthcare. However, his government faced criticism for its handling of the economic crisis and for not doing enough to address corruption. Nevertheless, Rajoy's party won a majority in the 2011 and 2016 general elections.
• Alexis Tsipras: The former Prime Minister of Greece made several promises during his 2015 election campaign, including ending austerity measures, reducing unemployment, and renegotiating Greece's debt with international creditors. However, his government faced criticism for not doing enough to address poverty and inequality. Nevertheless, Tsipras' party won a majority in the 2015 parliamentary election and he was re-elected as Prime Minister in 2019.

This lead to frustration and disillusionment among some voters, who feel that their voices are not being heard and that their needs are not being addressed and in turn leads to low voting rates and to rise of populism.

One reason politicians are able to get away with breaking their promises is that many voters do not hold them accountable. Some voters may not pay close attention to what politicians say during campaigns, or they may not have the time or resources to research the candidates' backgrounds and track records. Others may be more swayed by a candidate's personality, charisma or even their personal success than by their actual policies. Further, while politicians are subject to criminal law supervision in terms of corruption, due to their immunity as elected officials there are no consequences for presentation of inaccurate or false information.

Furthermore, politicians are often under pressure from interest groups, lobbyists, and other stakeholders to deliver results quickly. These groups may have their own agendas and priorities, which may not align with the long-term interests of the country as a whole.

Another reason for the feeling that modern democracies do not execute well is the role of special interests and corruption. In many democracies, the influence of money in politics and the power of special interest groups can lead to decisions that benefit a few at the expense of the many. This can create a perception of unfairness and erode trust in the democratic process.

Critiquing politicians is easy, but it is not easy to be a politician. The political system itself can make it difficult to enact real change. Even when a politician genuinely wants to keep their promises, they may face resistance from other members of their party or from outside interest groups. Bureaucratic hurdles and budget constraints can also make it difficult to implement certain policies. By making the political system work much better, politicians will have much better reputation and we will also be able to attract more of the best people to do this important job.

To address these acute problems, modern democracies need to adopt a more holistic and long-term approach to governance. It is critical for politicians to balance short-term needs with long-term goals, and to be transparent about the trade-offs involved in their policy decisions and to embrace greater transparency, accountability, and citizen participation, such as through open data, citizen juries, and participatory budgeting. Creating trust in the system can only be based if measurement of success is done objectively by comparing objectives to results.

As will be explained in the next sections, Competent Democracy is a new philosophy, process and technology that remains committed to the democratic process but aims to create an effective system of governance that would allow dramatically better goal setting, execution and accountability.

Competent Democracy — Main Principles

Politicians, in their ideal role, are not just policymakers but visionaries and consensus-builders. They are tasked with the complex job of bridging the gap between diverse opinions and formulating a path forward that honors the society's collective will while being feasible in execution. This requires a combination of creative problem-solving, empathetic leadership, and strategic management.

Competent Democracy upholds the fundamental democratic principle of Representation, ensuring that voters elect parties or individuals who reflect their values and ideals, based on:

• Philosophical Views: voters' choices are influenced by their upbringing and life experiences, leading to a preference for political ideologies such as conservatism, liberalism, socialism, libertarianism, environmentalism, and more.
• Political plans and Agenda: voters are inclined to support parties or individuals that put forward strategies and agendas aligned with their own visions for the future.
• Reputation and Performance: The historical performance and reputation of political parties and individual politicians play a crucial role in the decision-making process of voters.

Building on the above Competent Democracy suggests several major principles listed below in high level. For more details, see the tables in the next pages.

• Manifesto/Platform (§2): to stand for election, a party must publish a manifesto of measurable goals — each with timeline, metrics, and budget — including long-term goals (e.g. environmental, educational). Goals must be reviewed and approved by the relevant expert committees before the manifesto is locked 90 days before election day. Once locked, no goal may be added, removed, or modified except through Change Management (§6).

• Experts (§4): experts serve on standing national committees by field. All eligible experts appear on a single public Expert Roster maintained by the independent Expert Roster Commission. Each expert publishes their philosophical political views and declarations of interest. Parties pick committee members from the roster but no expert may be a party member.

• Parliament and Government Formation (§3): after elections, a legislative body is formed and a government is formed by a single party or coalition. Expert committees are constituted by parties selecting from the Expert Roster — they are not appointed by individual ministers.

• Expert Committees (§4): standing national committees by field. Each governing-coalition party selects members in proportion to its coalition share; each opposition party is entitled to at least one seat per committee. Composition must satisfy minimum diversity floors (academic, private-sector, international).

• Stability and Continuity (§5): to avoid "political blackmail", governments can be dismissed by simple majority only every 2 years, or with a 70% supermajority¹ at any time. A government missing too many milestones at the annual review is dismissed automatically. Long-term goals must be carried by successor governments unless voted down by a 70% supermajority.

• Change Management (§6): in response to significant external events (pandemic, war, disaster, abrupt international-agreement change), parliament may, by 60% supermajority with expert concurrence, suspend scoring and sanctions for specifically-enumerated affected goals. Time-limited and renewable in six-month blocks, with multiple independent termination paths.

• Objective Measurement (§7): progress is scored by purpose-built open-source AI software. The framework is developed by an international non-profit (D2DEV) and operated nationally by a separate non-profit (D2OPS). Code and per-goal parameters are public; scores can be appealed by government, opposition, expert committees, and the Academic Supervisory Committee (ASC).

• Transparency and Involvement (§8): a government App provides citizens with real-time status updates on every goal and supports electronic voting via national e-ID. Citizen sentiment polls on the App are advisory only and do not feed any score.

• Accountability and Sanctions (§9): accountability runs on two tracks. Track A is collective and political — governments missing milestone thresholds lose office and forward eligibility. Track B is individual and adjudicative — the independent Office of the Manifesto Prosecutor prosecutes false-promise cases before a specialised Manifesto Tribunal, with tiered standards: civil for fines and ineligibility, criminal for imprisonment.

Use of Artificial Intelligence — overview

Competent Democracy uses AI for two narrow purposes only: suggesting candidate goals to expert committees for review, and measuring progress against goals that humans set and humans approved. AI surfaces evidence, summaries, and anomalies for human investigation; every consequential decision in the system is made by named institutions against published rules.

The substantive mechanics of measurement — scoring architecture, the D2DEV/D2OPS split, change control, open-source scope, appeals, and safeguards against capture — are specified in §7. The broader background on what today's frontier AI can and cannot reliably do, and how the system is designed around those limits, is in Appendix B.

Glossary — key institutions and instruments

For quick orientation. Each entry points to the section where it is fully specified.

Table 1 — Competent Democracy Major Principles

The principles below are described in detail. The measurements should be read as examples, as they should be debated and may vary between implementations of the system in different jurisdictions (countries or municipalities).

1. Representation

Rationale. People nature is that based on upbringing and life experience they have certain political philosophical views such as conservatism, liberalism, socialism, libertarianism, environmentalism, etc. This is maintained in Competent Democracy. People vote to political parties that represents their beliefs and present goals and execution plan.

The role of the politicians is to present the best ideas, seek agreement based on consensus and manage execution.

2. Measurable Short and Long Term and Global Goals and Plans

Rationale.

i. In order for a party to be listed for election, it must present measurable goals and plans. Including time line, metrics and budget.

ii. Naturally many of the goals will be short term (for the term of the government).

iii. To ensure execution on topics that take time and may be important in the future, some of the goals must be long term. Long-term goals must be carried by subsequent governments unless they are voted down by a 70% parliamentary supermajority (see §2 details).

iv. As the world is facing global issues (first and foremost climate issues) that take both time and international commitment to impact, some of the goals must be global. Global goals follow the same rule: they must be carried by subsequent governments unless voted down by a 70% supermajority of parliament or unless modified through binding international agreements.

Details and Measurements (example).

• Manifesto-lock. Manifesto goals must be finalised, reviewed and approved by the relevant expert committees, and published in full 90 days before election day. From the moment of lock, no goal may be added, removed, or modified except through the §6 Change Management process. The locked manifesto is published on the citizen App with cryptographic timestamping; subsequent edits are visible as logged amendments and tied back to the §6 motion that authorised them.
• Execution plan. A quarterly-milestoned execution plan for each goal must be published within 90 days after formation of a government.
• Long-term and global goal share. At least 25% of goals set must be for the long term (10–25 years) or for global initiatives.
• Carry-over threshold. A 70% parliamentary supermajority is required to vote down a long-term or global goal inherited from a previous government.
• AI suggestions. AI may be used to suggest goals and metrics. Suggestions are reviewed by the relevant expert committee before any goal is included in a manifesto.

3. Government

Rationale.

i. Government formation follows ordinary parliamentary practice: the legislative body is constituted from election results, and a single party or coalition that commands its confidence forms a government. Competent Democracy modifies the lifecycle and accountability of governments, not the act of forming one.

ii. The head of government (President or Prime Minister, depending on the constitutional model) appoints ministers, who may be either politicians or experts. Ministers exercise executive authority subject to the expert-committee recommendation requirement under §4 — they do not appoint their own expert committees. Expert committees are standing national bodies whose members are selected by parliamentary parties from the Expert Roster maintained by the independent Expert Roster Commission (§4.1).

iii. Within 90 days of formation, the government must publish a quarterly-milestoned execution plan for every locked manifesto goal (§2). The execution plan is the operational expression of the manifesto — the artefact against which Track A milestone-miss accountability is later measured (§5, §9.1).

iv. Coalition stability and dismissal cadence are governed by §5; Change Management for crisis events by §6; sanctions for individual politicians by §9.

Details and Measurements (example).

• Legislative body: may be unicameral or bicameral.
• Head of government: may be a President or Prime Minister, per the constitutional model.
• Ministers may be politicians or experts; expert-ministers are subject to the same accountability rules as politician-ministers.
• A serving minister may not sit on the expert committee of their own portfolio. The separation between executive (ministers) and advisory (expert committees) is structural.
• Execution plan publication: within 90 days of government formation, in the same machine-readable format as the locked manifesto, on the citizen App.

4. Expertise based on philosophical political views of elected parties

Rationale.

i. Most rules, executive orders, and other government actions in a committee's field must be recommended by a majority of the relevant expert committee before they are voted by politicians. A narrow class of core executive prerogatives and time-critical emergency actions is exempt (see §4.8).

ii. Experts form standing committees by field. Fields can be added, merged, or retired only by a 70% parliamentary supermajority.

iii. Competent Democracy does not pretend that expertise is politically neutral. It is openly and honestly partisan-aware but not partisan-controlled: every expert publishes their philosophical political views, parties pick from that public roster, but no expert may be a party member and the integrity of the roster is held outside any coalition.

iv. In each committee, each governing-coalition party selects experts in proportion to its size in the legislature. In addition, each opposition party is entitled to seats in each committee.

v. Experts are paid competitively by the state and serve for a limited, non-renewable tenure. They are barred from outside paid work, gifts, or post-service lobbying for a defined cool-down period.

4.1 The Expert Roster and the Roster Commission

The Expert Roster is the single public list of all individuals eligible to serve on any expert committee. Parties select committee members only from this roster.

The roster is maintained by an Expert Roster Commission, a standing body of nine members whose only function is roster integrity:

• 3 members nominated by the supreme court from sitting or retired senior judges.
• 3 members nominated by the academic supervisory committee (see §7.6) from senior academics, with at least one international academic.
• 3 members nominated by parliament, each requiring a 70% supermajority for confirmation.

Commissioners serve a single non-renewable seven-year term. Commission decisions on admission and removal require a simple majority; published with reasons.

The Commission's only job is roster admission and removal against the published conditions. It has no role in committee operations, in selecting committee members from the roster, or in any expert's substantive work.

4.2 Admission to the roster

Any individual meeting the conditions below may apply. The Roster Commission decides on admission against the conditions and publishes its reasoning.

Admission conditions:

A. Experience. At least seven years of demonstrated experience in the relevant field, in any combination of academic, public sector, or private sector roles. B. No lobbyist history. Must not have been employed by, or contracted as a lobbyist for, any organisation in the preceding five years. C. No competing role. Cannot hold any other paid role during expert tenure, except for academic teaching and research up to one day per week with declared employer. D. No party membership. Must not be a member of any political party at the time of admission or during tenure. A two-year cool-down from any prior membership is required. E. Peer endorsement. Must be endorsed by at least five existing or former roster members in the relevant field. F. Declared philosophical views. Must publish a written declaration of political philosophical views (e.g. conservative, liberal, environmentalist, libertarian, social-democratic), updateable but publicly versioned. G. Declared interests. Must publish a declaration of all relevant economic, family, and institutional interests, updated annually and within 30 days of any material change. H. Clean record. No unspent criminal convictions for offences of dishonesty, corruption, or public-trust violations.

The roster is published annually and updated continuously. Applications and decisions are public.

4.3 Removal from the roster

The Roster Commission may, by simple majority and with published reasons, remove an expert from the roster on any of the following grounds:

• Breach of any admission condition discovered after admission, or arising during tenure.
• Materially false or misleading declaration under conditions F or G.
• Conviction of an offence under §9, or any criminal offence of dishonesty.
• A finding of professional misconduct by the academic supervisory committee or by the expert's professional regulator.
• Failure to attend a defined proportion of committee meetings without good cause.

Removal during active committee service triggers immediate replacement by the nominating party from the roster, following the standard selection rules.

A removed expert may appeal to the supreme court on questions of fact and law.

4.4 Party selection from the roster

When committee seats are to be filled (after an election, on a vacancy, or on a field change):

1. Each governing-coalition party selects experts from the roster, in proportion to its share of coalition seats, until that party's allocation is filled.
2. Each opposition party with seats in parliament selects experts from the roster — at least one seat per opposition party per committee.
3. Selections are checked by the Roster Commission only for compliance with composition requirements (see §4.5).

Parties may consult published philosophical-view declarations in making their selections. Selections are made public on the date the committee is constituted.

4.5 Committee composition

Each committee must satisfy the following composition rules. These are minima; categories may overlap (an international academic counts toward both "academic" and "international").

• At least 30% academic background — primary career in research and teaching at recognised institutions.
• At least 30% private-sector background — substantial career in industry, professional services, or relevant non-profit operations.
• At least 30% international experience — sustained professional experience outside the country, or holding the citizenship of another state.

The Roster Commission certifies that committee composition meets these minima before the committee may take any binding vote. If composition cannot be met from the current selections, the parties must re-select.

4.6 International experts

International experts (the ~30% share with international experience) are integrated into committees on the same footing as nationals, with three structural differences:

• Source register. International experts are selected from a parallel D2DEV International Roster maintained by the international D2DEV organisation under the same admission conditions as the national roster, augmented by international peer endorsement. National parties select from this register through the national Roster Commission.
• Funding. International experts are funded by the state at the same competitive rates as nationals, with allowances for relocation and dependants where applicable. They do not accept compensation from D2DEV, their home state, or any other source for their service.
• Dual removability. An international expert may be removed by the national Roster Commission (on any §4.3 ground) or by D2DEV (for violations of international roster standards). Either suffices. The rule prevents capture by either side from trapping a compromised expert in service.

4.7 Expert independence in service

While serving on a committee, experts are bound by additional rules:

• No outside paid work beyond the academic exception under condition C.
• No gifts or hospitality above a published de-minimis threshold.
• No communication with parties about committee business outside the committee's formal proceedings.
• Recusal from any vote where a declared interest is material.
• Post-service cool-down: no paid lobbying or paid party-affiliated work for three years after leaving the committee.

Breach is grounds for removal under §4.3 and may also be a separate criminal offence depending on severity.

Details and Measurements (example).

• Number of committees: 21 — Finance, Foreign Affairs, Defense, Interior, Justice, Education, Health, Agriculture, Transportation, Energy, Labor and Social Affairs, Industry and Trade, Tourism, Culture, Environment and Climate Change, Information and Communications, Science and Technology, Artificial Intelligence, Youth and Sports, People and Gender Equality, Next Generation.
• Adding, merging, or retiring a field: 70% parliamentary supermajority.
• Scope of expert involvement: at least 90% of executive actions in a committee's field must obtain a recommendation by simple majority of that committee before being voted by politicians. The remaining ~10% is a narrowly defined class of core executive prerogatives and time-critical emergency actions, listed in implementing legislation.
• Committee size: 15–25 members per committee depending on the field, set by parliament.
• Composition floors: ≥30% academic, ≥30% private-sector, ≥30% international (overlapping).
• Opposition entitlement: at least one seat per opposition party per committee.
• Roster Commission: 9 members; 3 judiciary + 3 academic + 3 parliamentary; seven-year non-renewable terms.
• Expert tenure: up to 8 years or two terms (whichever is shorter), non-renewable thereafter.
• Cool-downs: 2 years from prior party membership before admission; 3 years from leaving committee before any paid lobbying or party-affiliated work.
• Roster, selections, declarations of view and interests, and Roster Commission decisions are all public on the citizen App.

5. Stability and Continuity

Rationale.

i. In order to allow the government to function continuously and not be paralysed by routine coalition stability fights, governments can only be dismissed at defined intervals or with a supermajority — except where they fail to deliver on their milestones, in which case they are dismissed automatically.

ii. Dismissal therefore operates on three independent rails: (a) a regular biennial confidence opportunity by simple majority; (b) a supermajority anytime; (c) an annual automatic dismissal triggered by missed milestones under §9 Track A.

iii. Long-term goals set by previous governments must be carried by following governments unless voted down by a 70% parliamentary supermajority, consistent with §2.

Details and Measurements (example).

• Regular dismissal (biennial). A confidence vote requiring only a simple parliamentary majority is available at every two-year anniversary of the government's formation. Between these biennial windows, simple-majority dismissal is not available.
• Supermajority dismissal (anytime). A 70% parliamentary supermajority — or the government's own resignation — may dismiss the government at any time.
• Automatic dismissal (annual milestone-miss). At the annual review, if fewer than 75% of milestones due in the preceding year have scored as "met," the government is dismissed automatically without a vote. This is Track A of §9 and applies independently of the biennial and supermajority rails.
• Long-term goal carry-over. A 70% supermajority is required to drop a long-term or global goal set by a previous government (see §2).

6. Change Management

Rationale.

i. The system is designed to make promises stick. But events outside any government's control — pandemic, war, natural disaster, major economic shock, abrupt change in binding international agreements — can render previously-feasible goals materially unachievable.

ii. Change Management is the controlled, supermajority-gated mechanism that allows the system to adapt to such events without dismantling the accountability it normally enforces. It is deliberately narrow, time-limited, and subject to multiple independent termination paths so that "crisis" cannot become an indefinite excuse for suspending the rules.

iii. The trigger is an external event of sufficient magnitude to materially affect a defined set of national goals. Triggers include but are not limited to: a declared state of emergency under constitutional or statutory grounds; an armed conflict directly or indirectly involving the country; a public-health emergency declared by the national health authority or the World Health Organization; a major natural disaster; a sudden material change in binding international agreements affecting committed goals.

iv. A Change Management invocation must enumerate the specific goals it affects. Goals not so listed continue to score and to carry sanction risk as normal. The override is a scalpel, not a curtain.

6.1 Invocation

Invocation requires three concurrent conditions:

• Proposal. A formal motion to invoke may be brought by the government, by a petition of at least 25% of parliament, or jointly by two or more affected expert committees.
• Expert concurrence. Each relevant expert committee must, by simple majority, certify that the cited event materially affects the identified goals and that adjustment is reasonable.
• Parliamentary supermajority. Parliament must approve the motion by at least 60%.

The motion must specify (a) the triggering event, (b) the goals affected, (c) the nature of the adjustment for each (suspension of scoring, sanctions exemption, or both), and (d) the initial duration, not exceeding six months.

Approved invocations are published in full on the citizen App within 24 hours and entered as a logged parameter change under §7.3.

6.2 Effects

While an invocation is in force, for each affected goal:

• Scoring is suspended. The goal's score is frozen at "under §6 review" on the App. No automated-dismissal or forward-ineligibility consequence under Track A of §9 attaches to that goal during the period of suspension.
• Sanctions are exempted. No false-promise sanction under Track B of §9 may be initiated or prosecuted in respect of the affected goal for promises rendered unachievable by reason of the cited event. This is a defence to specific charges, not a general immunity for unrelated misconduct.
• Expedited goal-addition. New crisis-response goals may be proposed by the government and added to the manifesto without waiting for the next election cycle, provided they are reviewed and approved by the relevant expert committee (simple majority) and ratified by parliament (simple majority). Crisis-response goals are themselves scored and sanctionable from the date of addition; they do not inherit the override.

Goals not specifically listed in the invocation continue to score and to carry sanction risk as normal.

6.3 Duration, renewal, and termination

• Initial duration. Up to six months, as specified in the invoking motion.
• Renewal. Further six-month extensions require a fresh motion meeting the same conditions as the original invocation (proposal, expert concurrence, 60% parliamentary supermajority). Each renewal must reaffirm the affected goals and may add or drop goals from the list.
• Auto-expiry. If no renewal motion passes by the expiry date, the invocation lapses automatically. Affected goals revert to live scoring and full sanction risk from the lapse forward. Promises made during the override remain protected by it; promises made after the lapse do not.

Three independent paths to early termination operate alongside auto-expiry:

• Government motion. The government may end its own override at any time by simple parliamentary majority.
• Parliamentary force-end. A petition by 40% of parliament tables a force-end motion, which passes by simple majority. Force-end is deliberately easier than invocation: ending an override should not be harder than starting one.
• Academic supervisory committee recommendation. The academic supervisory committee may publicly recommend termination. The recommendation is advisory but must be debated in parliament within 30 days.

6.4 Post-override review

Within 90 days of the end of any invocation (whether by expiry, force-end, or government motion), the academic supervisory committee publishes a review covering:

• Whether the cited event in fact materially affected the listed goals.
• Whether the duration and any renewals were proportionate to the event.
• Whether any crisis-response goals added under §6.2 should be retained, modified, or removed.
• Whether any politician's conduct during the override warrants referral to the OMP under §9 for conduct not protected by the override.

The review is debated in parliament within 60 days of publication. Adverse findings do not retroactively unwind the override but may trigger §9 referrals and inform parliamentary judgement on any subsequent invocation.

Details and Measurements (example).

• Invocation threshold: 60% of parliament + relevant expert committee majority concurrence.
• Initial duration: up to 6 months. Renewals: 6 months each, same invocation threshold.
• Affected goals must be enumerated in the motion; non-listed goals continue to score and sanction normally.
• Expedited crisis-response goals: relevant expert committee (simple majority) + simple parliamentary majority.
• Early termination paths: government motion (simple majority); 40% parliamentary force-end (simple majority to pass); academic supervisory committee recommendation (advisory, debated within 30 days).
• Post-override academic review: published within 90 days of end; debated within a further 60 days.
• All invocations, renewals, terminations, and post-override reviews are public on the citizen App.

7. Objective Measurement of Progress

Rationale.

i. To ensure objective assessment of progress and meeting goals, the determination is done by an AI software that produces a score for every goal and milestone set.

ii. The system is built on a deliberate separation of duties: an international non-profit develops the open-source framework (D2DEV), and a national non-profit operates it locally on each country's data (D2OPS). Neither can act unilaterally on the other's domain.

iii. Both the code and every per-goal configuration are public and inspectable. Citizens, opposition, and experts can read exactly how each score is computed and from which sources.

iv. Scoring is the result of cross-checking many independent data sources rather than any single feed:

• Public information repositories
• Press
• Social media
• Public opinion polls
• Expert opinion
• Politician statements
• Global data sources

v. Because objective measurement is the system's most consequential design choice, it is the most heavily checked: every framework change, every per-goal parameter change, and every score is logged publicly, and four classes of stakeholder (government, opposition, expert committees, academic supervisory committee) have formal standing to challenge it.

7.1 Scoring architecture (hybrid)

Scoring uses a central framework with per-goal customization:

• Central framework (owned by D2DEV). A versioned, open-source library of:

  • Metric types — rate, count, ratio, composite index, qualitative survey result, time-to-event, etc. — each with a defined mathematical mapping from observed value(s) to a 0–100 score given a promise, deadline, and baseline.
  • Data-source whitelist — every source that may feed a score, with documented provenance, update cadence, known limitations, and approved fallbacks if the primary source becomes unavailable.
  • Scoring math — the deterministic functions that convert observed values into a score, including milestone interpolation, baseline normalisation, and adjustment for declared external shocks (see §7.3).

• Per-goal customization (owned by parties, approved by experts). When a manifesto goal is locked, the proposing party submits a measurement spec consisting of: the chosen metric type, the specific data sources from the whitelist (with declared fallbacks), the threshold that counts as "met," the quarterly milestones, and the baseline. The spec is reviewed and must be approved by the relevant expert committee including its opposition members before the manifesto is locked.

This split keeps comparability high (the same metric type means the same thing across goals and across governments) while letting parties tailor measurement to the specifics of what they promised.

7.2 D2DEV and D2OPS — separation of duties

The split exists so that a single political coalition cannot capture both the rules of measurement and the operation of measurement. A government that wants the framework changed must persuade the international D2DEV process; a government that wants its own running configuration changed must clear the national D2OPS process. Each is a check on the other.

7.3 Change control

Changes to anything that affects a score are versioned and logged publicly. Approval thresholds depend on what is being changed:

• Framework version (D2DEV). New releases require independent academic review and D2DEV board approval. The diff between releases must be published in a human-readable change-log alongside the code diff.
• Deployment of a new framework version (D2OPS). Deploying a new D2DEV release into the running instance requires a 70% supermajority of the D2OPS board and concurrence of the relevant expert committee (or, for cross-cutting upgrades, the academic supervisory committee).
• Per-goal spec change mid-term. A locked spec may only be changed by 70% of the D2OPS board with concurrence of the relevant expert committee. A written justification is published on the App alongside the change.
• Emergency data-source replacement. If a whitelisted data source becomes unavailable, the D2OPS director may substitute a declared fallback within 14 days; the substitution must be ratified by the board within 14 days of being made or it reverts.
• External-shock adjustment. If a goal is materially affected by an external event (pandemic, war, disaster, international-agreement change), parliament may, under the Change Management procedure (§6), declare a scoring adjustment for that goal. The adjustment is itself a logged parameter change.

7.4 Open-source — what is actually public

"Open-source" in Competent Democracy means three things, in increasing order of operational detail:

1. Code. The full D2DEV codebase is published under a permissive licence and developed in public.
2. Parameters. Every per-goal spec — metric type, chosen data sources, thresholds, milestones, baseline — is published on the citizen App at the moment of manifesto-lock, and every subsequent change is published with its justification (§7.3).
3. Inputs. Input data is published in full where it is already lawfully public (national statistics, government registers, anonymised administrative data). Where input data is individual-level and not lawfully publishable (health records, education records, tax data), the methodology, aggregations, and intermediate computations are published, while the raw data remains access-controlled to the academic supervisory committee under research agreements.

The premise is not that every citizen will read the code or the parameters, but that any citizen, journalist, opposition researcher, or academic can — and that the system's legitimacy depends on this being practically true, not just nominally true.

7.5 Appeals

Standing. Four classes of party may file an appeal:

• The scored government, on any score affecting its own goals.
• Any party with seats in parliament, on any score.
• Any expert committee, on any score within its field of competence.
• The academic supervisory committee, on any score.

Citizens may petition any of the above to file on their behalf; an expert committee that declines to file must publish a brief written reason.

Grounds. An appeal must specify one or more of:

1. Factual error — the input data is wrong or misapplied.
2. Methodological flaw — the per-goal spec does not measure what it claims to measure.
3. Framework defect — the underlying D2DEV release contains a bug or, more seriously, embeds a normative assumption that has been presented as a technical choice.
4. Extraordinary circumstance — an external event has made the score materially unfair and no §6 adjustment has yet been declared.

Process. The appeal is filed with the D2OPS appeals committee — a standing sub-committee of the D2OPS board chaired by an independent legal officer (not a board member). During an appeal the score remains in effect but is publicly flagged "under appeal" on the App. The appeals committee has 60 days to decide; it may commission an expert re-review or an independent technical analysis.

Decision. A 70% supermajority of the appeals committee is required to override or modify a score. Possible remedies are:

• Re-score with corrected inputs or methodology.
• Suspend the score pending a D2DEV framework fix.
• Refer the goal to parliament as unmeasurable, requiring re-specification under §6 or removal from the manifesto with sanctions consequences for the proposing party.

Every decision and its reasoning is published.

Retrospective audit. An expert committee, the academic supervisory committee, or 25% of parliament may trigger a retrospective audit of any past score within four years of its publication. Retrospective audits follow the same process and standards as live appeals; they cannot reverse a past election outcome but may trigger sanctions under §9 and revisions to the framework or to standing specs.

7.6 Safeguards against capture

Because the scoring system is, by design, the system's most powerful instrument, four cross-cutting safeguards apply:

• Opposition review at manifesto-lock. Per-goal specs cannot be approved by a majority that excludes the committee's opposition members; a spec must clear at least one opposition expert's vote.
• Independent academic review of framework changes. No D2DEV release reaches deployment without a published academic review of its diff, including an explicit search for normative choices embedded as technical defaults.
• Annual normative audit. The academic supervisory committee publishes an annual report identifying any framework or parameter choices that, in its judgement, encode contested values rather than neutral measurement. Findings do not automatically force change but are debated in parliament.
• Whistleblower protection. Any D2DEV or D2OPS employee, contractor, or expert may file a confidential concern with the academic supervisory committee; retaliation is a criminal offence.

7.7 The Academic Supervisory Committee

The Academic Supervisory Committee (ASC) is the national standing academic body that oversees D2OPS and the country's participation in D2DEV. It is the system's intellectual conscience — charged with publicly evaluating whether the technical machinery of measurement is doing what the system claims, and whether it is drifting in ways the political process is unlikely to detect on its own.

D2DEV (the international organisation) has its own International Academic Supervisory Committee under analogous rules. References to the "academic supervisory committee" elsewhere in this document mean the national ASC unless otherwise specified.

Composition. Seven sitting senior academics drawn from at least four different disciplines relevant to measurement, policy, statistics, ethics, and law. At least two members must be international academics. Members serve a single non-renewable six-year term, staggered so that the committee is never fully replaced at once.

Appointment. Members are nominated by the national academies of sciences and humanities and confirmed by a 60% parliamentary supermajority. The chair is elected by the committee from among its members.

Responsibilities.

• Publish the annual normative audit of the scoring framework (§7.6).
• Co-nominate three members of the Expert Roster Commission (§4.1).
• Receive whistleblower disclosures and refer matters to the OMP under §9 where warranted.
• Publish the post-override review under §6.4.
• Recommend termination of an override under §6.3 where, in its judgement, the override is no longer warranted.
• Hold standing in appeals under §7.5.
• Administer an annual research grant programme funding master's and doctoral work analysing D2OPS operations, the scoring framework, and Competent Democracy in comparative perspective.

Funding and independence. Direct line-item from the national budget; no other source of funding accepted. Members are paid at the senior judicial scale and cannot hold other paid roles for the duration of their term. The committee may not be dissolved or reduced in membership by any single coalition. Removal of an individual member follows the procedure in §4.3 with appeal to the supreme court.

Details and Measurements (example).

• D2DEV is funded by participating countries and municipalities through subscription fees; D2OPS by national taxation. Neither accepts party, corporate, or anonymous donations.
• D2OPS board: one representative per parliamentary party plus an equal number of experts (drawn proportionally from the expert committees), chaired by a supreme court judge with a casting vote only on procedural matters.
• Appeals committee: rotating sub-committee of five — two D2OPS board members nominated by government, two nominated by opposition, plus the independent legal chair.
• Threshold to override a score on appeal: 70% of appeals committee (i.e., 4 of 5).
• Threshold to deploy a new D2DEV release into D2OPS: 70% of D2OPS board.
• Maximum decision window on a live appeal: 60 days. Maximum window on a retrospective audit: 120 days.
• Lookback window for retrospective audit: 4 years from the published score.
• Annual normative audit by academic supervisory committee published within Q1 of each year, debated in parliament within 60 days.

8. Transparency and Involvement

Rationale.

i. Transparency is not a single feature but the operating mode of the entire system. Every locked manifesto goal, every per-goal measurement spec, every score, every parameter change, every appeal, every override, every sanctions referral, and every academic supervisory committee report is published on the citizen App.

ii. The App is also the system's principal channel for citizen participation: electronic voting in elections and referendums, sentiment polling on specific goals, and a structured channel to petition expert committees, the OMP, and the Roster Commission.

iii. The same App, however, must not become a vector for gaming the score. Citizen participation and scoring are therefore deliberately separated: opt-in App polls are advisory and visible to parliament and the public, but they do not directly feed any score. Score inputs come only from data sources on the published whitelist (§7.1), which includes commissioned random-sample surveys conducted to statistical-validity standards — never the App's opt-in polls.

8.1 What the App publishes

For every active goal, the App publishes:

• The original manifesto text, the locked measurement spec (metric type, data sources, thresholds, milestones), and the cryptographic timestamp of the lock.
• The current score and a full history of score changes.
• Every per-goal parameter change, with the §7.3 justification.
• Any active appeal under §7.5 (status: under appeal).
• Any active §6 override affecting the goal.

For the system as a whole, the App publishes the Expert Roster, all expert declarations of view and interest, Roster Commission decisions, OMP referrals and outcomes, ASC reports, and the running change-log of framework versions.

8.2 Voting and identity

• Electronic voting in elections and referendums is available through the App alongside traditional balloting. Identity is verified by the national electronic ID. A paper-ballot alternative is always available.
• Sentiment polls on specific goals are advisory only. Opt-in polls (where any App user may respond) are clearly labelled as opt-in and do not feed any score.
• Random-sample surveys commissioned by D2OPS use the App as one possible response channel but are run on statistically-drawn samples with identity verification, weighting, and methodology published. These can feed scores where the per-goal spec includes them.

8.3 Anti-manipulation

• D2OPS publishes a quarterly manipulation report identifying anomalous response patterns in opt-in polls (coordinated bursts, bot signatures, geographic anomalies).
• App-based opt-in polls cannot be referenced in any score appeal as evidence of public sentiment.
• Coordinated manipulation of App inputs is a criminal offence under implementing legislation.

Details and Measurements (example).

• Score inputs: only data sources from the §7.1 whitelist. App opt-in polls are not on the whitelist.
• Random-sample surveys: minimum sample size, weighting methodology, and field period published before fielding; results published with full microdata where lawful.
• Quarterly manipulation report by D2OPS: published in the same release cycle as the score updates.
• Electronic voting: requires national e-ID; paper-ballot alternative always available.

9. Accountability and Sanctions

Rationale.

i. Accountability operates on two distinct tracks. The first is collective and political: if a government fails to deliver, it loses office and its executive politicians lose eligibility for the next term. The second is individual and adjudicative: if a politician makes a promise that an expert committee determines was not achievable, they face fines and, in serious or repeated cases, imprisonment.

ii. Both tracks share a common principle — that the cost of failing to deliver, or of making promises that should never have been made, must fall on the individuals who chose to make those promises, not on the public that believed them.

iii. The adjudicative track is run as a regulatory-criminal proceeding, not as a political one. Charges are brought by an independent prosecutor; cases are tried by a specialised tribunal; standards of proof and protections for the accused are no lower than those in other criminal or regulatory matters.

9.1 Track A — Collective political accountability (milestone-miss)

A government that fails to meet a defined proportion of its quarterly milestones is held to account through its political fate, not through individual prosecution:

• Auto-dismissal threshold. If, at the annual review, fewer than the threshold proportion of due milestones have scored as "met," the government is dismissed automatically (see §5).
• Forward-ineligibility threshold. If, at the end of a term, the government's executive politicians (head of government and first-level ministers) have collectively delivered fewer than the threshold proportion of all milestones for the term, those politicians are ineligible to stand for executive office for the following term.
• No individual fault required. This track does not require any finding of wrongdoing. It is a clean, ex-ante rule: if you do not deliver what you promised, you do not govern.

9.2 Track B — Individual sanctions for false promises

The second track addresses promises that should not have been made — promises that an expert committee, applying its professional judgement, finds were not achievable within the stated time and budget.

This track applies in two contexts:

• Manifesto goals. All manifesto goals must pass expert-committee review at manifesto-lock (see §2). If the committee determines, at the time of review, that a goal is not achievable, the party may withdraw or modify the goal, or contest the determination through the appeals process. A goal that fails review cannot appear in the manifesto. Sanctions on a manifesto goal therefore only arise where new evidence later shows the original expert review was misled — for example, where a party submitted false data to obtain approval.
• Extra-manifesto promises. Public promises made outside the locked manifesto — in campaign speeches, debates, paid advertising, official social-media channels, or formal interviews — are not pre-vetted and remain the principal source of sanctions cases. Any politician who makes such a promise may, on retrospective expert review, be found to have promised something not achievable.

9.3 The Office of the Manifesto Prosecutor

Sanctions cases are brought by the Office of the Manifesto Prosecutor (OMP) — a specialised division of the state prosecutor's office, statutorily independent of the government of the day and reporting only to parliament for budget and to the constitutional court for legality of process.

The OMP's responsibilities are:

• Receiving and acting on referrals from the D2OPS appeals committee, expert committees, the academic supervisory committee, parliament (by simple majority), and whistleblowers.
• Initiating investigations on its own motion.
• Deciding which referrals warrant prosecution and which do not. Declined referrals must be answered with a published written reason.
• Conducting investigations with the assistance of expert committees as needed.
• Prosecuting cases before the Manifesto Tribunal.

The Chief Prosecutor is appointed by parliament for a single, non-renewable term of seven years, removable only for cause and only by a 70% parliamentary supermajority.

9.4 The Manifesto Tribunal

Cases are heard by the Manifesto Tribunal — a specialised judicial panel of three senior judges (with two reserves), nominated by the supreme court from among judges with demonstrated expertise in regulatory, economic, or constitutional matters. Tribunal judges sit only on Manifesto cases for the duration of their term; they cannot also hear unrelated matters in other courts.

Tribunal decisions may be appealed to the supreme court on questions of law only.

9.5 Standards of proof and state of mind

Sanctions are tiered, and the prosecutor's burden rises with the severity of the sanction sought:

"Negligence" is anchored to a reasonable-politician standard, informed by what the relevant expert committee, the AI suggestions, and the party's own internal advice would have indicated to a politician acting in good faith.

9.6 Available sanctions

• Fine. Up to two years' net ministerial salary for a first offence, paid into the public treasury. Fines are personal liabilities and cannot be reimbursed by the party, by the state, or by any third party.
• Ineligibility. Bar from holding executive office (head of government, minister, deputy minister) for one electoral term. A repeat finding within eight years extends the bar to two terms or to permanent ineligibility, at the tribunal's discretion.
• Imprisonment. For convictions on the criminal standard with proven recklessness or intent: up to three years for a first offence, up to seven for a repeat offence within eight years. Imprisonment also carries automatic permanent ineligibility for executive office.

The tribunal may impose any combination of the above. Sanctions are independent of, and additional to, the political consequences under Track A.

9.7 Defences

Available defences include:

• Expert-approved spec. A manifesto goal that passed expert-committee review under §2 cannot itself ground a sanction against the proposing politician, unless the prosecutor proves the politician obtained that approval through false or materially misleading submissions.
• External shock under §6. Where parliament has declared a Change Management adjustment under §6 for a relevant event, promises rendered unachievable by that event are exempt from sanction for the duration of the adjustment.
• Reliance on the AI suggestion. Where the original AI-suggested goals corroborated the politician's promise, this is a strong (but not conclusive) factor against a finding of negligence or recklessness.
• Good-faith disagreement. A promise that was genuinely contested among experts at the time of making — for example, where the committee was split — does not support a finding of recklessness, though it may still support negligence.

9.8 Procedural protections

Politicians facing sanctions have:

• The right to know the case against them and the underlying evidence.
• The right to counsel (state-funded where required).
• The right to call expert witnesses, including dissenting experts from the relevant committee.
• The right to cross-examine the prosecutor's witnesses.
• The right to a public hearing, save where the tribunal directs otherwise on narrow grounds (e.g. classified material).
• The right to written reasons for any adverse decision and to appeal on questions of law.

Statutes of limitations: four years from publication of a manifesto goal or from the public making of an extra-manifesto promise, whichever is later — matching the retrospective audit window in §7.5.

Details and Measurements (example).

• Track A milestone threshold (auto-dismissal): 75% of due milestones met at annual review (see §5).
• Track A milestone threshold (forward-ineligibility for executive politicians at end of term): 60% of total milestones for the term.
• Maximum fine (first offence): 2 × annual net ministerial salary.
• Imprisonment range: up to 3 years (first offence), up to 7 years (repeat within 8 years).
• Ineligibility: 1 electoral term first offence; 2 terms or permanent on repeat within 8 years.
• Statute of limitations: 4 years.
• Manifesto Tribunal: 3 sitting judges + 2 reserves, nominated by supreme court.
• Chief Prosecutor term: 7 years, non-renewable, removable only by 70% parliamentary supermajority.
• OMP must publish quarterly statistics: referrals received, investigations opened, prosecutions filed, outcomes by category. Declined referrals: published with written reasons.

On real-world rollout, adoption strategy, and funding of the Competent Democracy initiative itself (separate from the in-world system specified above), see the companion document concept-implementation-1.5.md.

Appendix A: Borrowing concepts from the business world — management and governance of public companies

When people invest money in a business, it is because they reasonably trust the management and the governance system to protect their money. While not perfect, there is a lot to be learned from the way that public companies operate and the methods that were developed to ensure that companies create value for investors, act in transparency and are accountable for results.

Executive Management

In Public Companies, the Executive Management team and CEO (Chief Executive Officer) of a company has important responsibilities towards the board of directors and shareholders.

Towards the board, the CEO has the responsibility of communicating the company's performance, strategy, and plans, and seeking approval and guidance from the board on major decisions. The CEO must also ensure that the company operates in compliance with laws, regulations, and ethical standards, and that there is adequate risk management in place to protect the interests of the company and its stakeholders. Additionally, the CEO is responsible for managing the company's resources and ensuring that the company's financial statements are accurate and transparent.

Towards the shareholders, the CEO has the responsibility of creating long-term value. This means that the CEO must develop and execute a strategy that maximizes the company's profitability and growth potential while minimizing risk. The CEO must also ensure that the company's financial performance is transparent and that the shareholders are informed of any significant developments or changes in the company's operations. Additionally, the CEO must balance the interests of shareholders with those of other stakeholders, such as employees, customers, suppliers, and the wider community.

CEOs must act in the best interests of the company and its stakeholders, while maintaining high ethical standards and promoting transparency and accountability.

Information disclosure

Public companies are required by law to provide accurate and timely financial and other information to their shareholders and to the public. This information is typically provided in the form of annual and quarterly reports, which contain financial statements and other disclosures that allow investors to evaluate the company's performance and make informed investment decisions.

There are several reasons why accurate reports are important for public companies. Firstly, accurate reports help to build investor confidence and trust in the company. When investors believe that a company is providing reliable information, they are more likely to invest in the company and hold onto their investments over the long term. This can help to stabilize the company's share price and reduce volatility in the market.

Secondly, accurate reports are important for regulatory compliance. Public companies are subject to strict reporting requirements under securities laws and stock exchange regulations. Failure to comply with these requirements can result in penalties, fines, and legal liability for the company and its officers and directors.

Thirdly, accurate reports are essential for effective corporate governance. Boards of directors and management teams rely on accurate and timely financial and operational information to make informed decisions about the company's strategy, operations, and risk management. Without accurate reports, these decisions may be based on incomplete or misleading information, which can have serious consequences for the company and its stakeholders.

Overall, accurate reports are critical for public companies to maintain the trust of investors, comply with regulatory requirements, and make informed decisions about the company's operations and strategy.

Supervision

Public companies are supervised by internal and external entities and processes.

Board of Directors. CEOs report to a board of directors. Selecting board members in a public company requires careful consideration of several criteria:

• Relevant experience: Board members should have relevant industry experience and expertise that can provide valuable insights and guidance to the company's management team. This may include experience in areas such as finance, marketing, operations, technology, and corporate governance.
• Independence: Board members should be independent and able to provide objective advice and oversight to the company's management team. They should not have any conflicts of interest that could compromise their ability to act in the best interests of the company and its shareholders.
• Diversity: Board members should bring diverse perspectives and backgrounds to the table. This may include diversity in terms of gender, ethnicity, age, and professional experience.
• Leadership qualities: Board members should have strong leadership qualities, such as the ability to communicate effectively, make tough decisions, and work collaboratively with other board members and the management team.
• Reputation: Board members should have a good reputation and be held in high regard by their peers and the broader business community. This can help to enhance the company's credibility and reputation.
• Availability: Board members should be able to devote sufficient time and attention to their responsibilities as directors. They should be willing and able to attend board meetings, participate in committees, and provide guidance and oversight to the company's management team.

Auditing. Auditing of public companies refers to the independent examination of a company's financial records and operations by a qualified third-party auditor. The purpose of auditing is to provide assurance to stakeholders, including investors and regulators, that a company's financial statements and other disclosures are accurate and reliable.

Auditing of public companies typically involves several key activities, including:

• Planning: The auditor will work with the company to develop an audit plan, which outlines the scope of the audit and the procedures that will be used to evaluate the company's financial records and operations.
• Fieldwork: The auditor will then conduct fieldwork, which involves gathering and analyzing data from the company's financial records and other sources. This may include testing internal controls, reviewing financial statements and disclosures, and interviewing key personnel.
• Reporting: Once the fieldwork is complete, the auditor will prepare a report that summarizes their findings and conclusions. This report is typically issued to the company's management, as well as to external stakeholders such as investors and regulators.
• Follow-up: In some cases, the auditor may also perform follow-up work to address any issues or concerns identified during the audit. This may include additional testing or analysis, or recommendations for improvements in the company's internal controls or financial reporting processes.

The auditing of public companies is important because it provides assurance to investors and regulators that the company's financial statements and disclosures are accurate and reliable. This, in turn, helps to maintain the integrity and stability of financial markets and promotes transparency and accountability in corporate governance.

Several countries have implemented legislation that sets standards for auditing practices, conducts inspections of audit firms, and imposes disciplinary actions for noncompliance. These laws are typically designed to improve corporate governance, increase transparency and accountability, and protect investors. Here are some examples of such laws:

• United States — auditing of public companies is regulated by the Public Company Accounting Oversight Board (PCAOB), which was established by the Sarbanes-Oxley Act of 2002.
• Canada — Canadian Public Accountability Board (CPAB): The CPAB is a non-profit organization that oversees the auditing of public companies in Canada. It was established in 2003 in response to accounting scandals at several Canadian companies, and it is responsible for setting auditing standards, conducting inspections of audit firms, and enforcing compliance with regulations.
• European Union — European Market Infrastructure Regulation (EMIR): EMIR is a set of regulations that were introduced in 2012 to increase transparency and reduce risk in the derivatives market in the European Union. It requires reporting of all derivatives trades, the clearing of certain types of derivatives, and the establishment of risk management procedures.
• United Kingdom — Companies Act 2006: The Companies Act 2006 is the primary legislation governing company law in the UK. It includes provisions related to director's duties, shareholder rights, and financial reporting requirements. In addition, the UK has established the Financial Reporting Council (FRC) to oversee the auditing of public companies and enforce compliance with accounting and auditing standards.
• Japan — Financial Instruments and Exchange Act (FIEA): The FIEA is a law that regulates securities and financial instruments in Japan. It requires public companies to disclose certain information related to their financial performance and operations, and it establishes rules for financial reporting, auditing, and disclosure.

While the specifics of these laws may vary, they share a common goal of improving corporate governance and increasing transparency and accountability in financial markets.

Regulatory Supervision. The supervision of public companies by regulators is a critical aspect of maintaining the integrity and stability of financial markets. Regulators are responsible for ensuring that public companies comply with applicable laws and regulations, and that they operate in a fair, transparent, and ethical manner.

The supervision of public companies by regulators typically involves several key activities, including:

• Registration and disclosure: Regulators require public companies to register with them and provide detailed information about their operations, finances, and management. This information is made available to investors, allowing them to make informed investment decisions.
• Compliance monitoring: Regulators monitor public companies to ensure that they comply with all applicable laws and regulations. This includes conducting regular audits and inspections, reviewing financial reports and disclosures, and investigating potential violations.
• Enforcement actions: Regulators have the power to take enforcement actions against public companies that violate laws or regulations. This can include imposing fines, revoking licenses, or even pursuing criminal charges in extreme cases.
• Investor protection: Regulators are also responsible for protecting investors from fraudulent or unethical practices. They may investigate complaints or tips from whistleblowers, take action against individuals or companies engaged in misconduct, and work to recover losses for affected investors.

The supervision of public companies by regulators is an essential component of maintaining the integrity and stability of financial markets. Regulators play a critical role in ensuring that public companies operate in a fair and ethical manner, and that investors are protected from potential harm.

Executive Accountability

Public companies have a strict set of governance rules to ensure proper conduct, information disclosure and supervision. But most importantly when a CEO of a public company fails to meet the goals set by the company's board of directors, the consequences can vary depending on the severity of the situation and the board's response. Here are a few possible outcomes:

• Reprimand: The board of directors may issue a warning to the CEO and ask them to take corrective actions to get the company back on track.
• Termination: If the board believes that the CEO's performance is not improving, they may terminate the CEO's employment. In some cases, this may involve a negotiated departure, including a severance package.
• Loss of compensation: A CEO who fails to meet their goals may lose some of their compensation, such as their annual bonus, stock options, or other incentives.
• Shareholder pressure: Shareholders may put pressure on the board to take action against the CEO. This can take the form of public statements, shareholder resolutions, or even attempts to vote out board members.

Ultimately, the board of directors is responsible for ensuring that the company is being managed effectively and that the CEO is meeting their goals. If the board determines that the CEO is not meeting their expectations, they may take any of the above actions or other measures they deem necessary to protect the company and its shareholders.

False promises made by CEOs can be considered as securities fraud, which is a violation of the federal securities laws. This can lead to civil or criminal charges, fines, and even imprisonment.

Criminal Liability

In public companies, the CEO and the board of directors have a legal responsibility to ensure that the company complies with all applicable laws and regulations. If the company engages in criminal activity, such as fraud or insider trading, the CEO and board members may be held criminally liable.

Criminal liability refers to the legal responsibility of an individual for a crime committed by themselves or by others under their direction or control. In the case of a public company, criminal liability can arise if the CEO or board members are found to have knowingly or recklessly engaged in criminal conduct or failed to prevent it.

The specific criminal liabilities of a CEO and board members will depend on the jurisdiction in which the company operates, as well as the nature and severity of the crime committed. Some common criminal charges that may be brought against CEOs and board members in public companies include:

• Securities fraud: This can include making false or misleading statements about the company's financial performance, insider trading, or manipulating stock prices.
• Environmental crimes: If a company violates environmental laws, the CEO and board members may be held criminally liable.
• Bribery and corruption: If a company engages in bribery or other corrupt practices, the CEO and board members may be held criminally liable.
• Antitrust violations: If a company engages in anticompetitive practices, the CEO and board members may be held criminally liable.

If a CEO or board member is found guilty of a criminal offense, they may face penalties such as fines, imprisonment, or both. In addition to criminal liability, CEOs and board members may also face civil liability, which refers to the legal responsibility for financial damages or other losses caused by their actions or inactions.

Overall, the CEO and board members of a public company have a significant responsibility to ensure that the company operates in a legal and ethical manner, and can face serious consequences if they fail to meet this obligation.

Downsides and criticism

Although there is much to learn from publicly traded business, there is justified criticism and several potential problems that must be mentioned.

• Short-term focus: Publicly traded companies are often under pressure to deliver strong financial results quarter after quarter, which can lead to a short-term focus on meeting or exceeding earnings targets. This can lead to decisions that prioritize short-term gains over long-term sustainability or investment in research and development.
• Conflicts of interest: Publicly traded companies are managed by a board of directors and executive team, who may have personal or financial interests that conflict with the interests of shareholders. For example, executives may prioritize their own compensation over shareholder value or may be reluctant to pursue actions that could harm their personal reputations.
• Lack of transparency: Publicly traded companies are required to disclose certain financial information to shareholders, but they may be less transparent about other aspects of their operations, such as their environmental or social impact. This can make it difficult for shareholders to fully evaluate the risks and opportunities associated with investing in the company.
• Pressure to meet market expectations: Publicly traded companies are subject to market expectations and may face consequences if they fail to meet or exceed those expectations. This can lead to a focus on short-term gains rather than long-term growth, and can discourage innovation or investment in new technologies or products.
• Pressure to maximize shareholder value: Publicly traded companies are often under pressure to maximize shareholder value, which can lead to decisions that prioritize profits over other goals, such as sustainability or social responsibility.

Examples of companies that have faced criticism for their management practices include Enron, which engaged in accounting fraud and ultimately went bankrupt in 2001; Wells Fargo, which faced a scandal in 2016 over its sales practices and lack of oversight; and Uber, which has been criticized for its poor treatment of drivers and lack of transparency around its business practices.

Borrowing concepts from the business world

Competent Democracy borrows some ideas from the way that public traded companies operate and the methods that were developed to ensure that companies create value for investors, act in transparency and are accountable for results, while trying to mitigate some of the issues of short-term planning and bias. Here are some key ideas:

Executive Accountability. Just as CEOs are accountable to the board and shareholders, national leaders and top government officials should be held accountable to the public. They should regularly communicate performance, strategies, and plans and be transparent about their decision-making processes.

Board of Directors as an Oversight Mechanism. Drawing from the role of a board in public companies, Competent Democracy suggests an AI based mechanism that provides oversight and strategic guidance to ensure that government actions align with the long-term interests of the nation and its citizens. The observations of this AI mechanism are monitored by an independent council or committee established at the national level, comprising individuals with relevant expertise.

Transparent Reporting. Just as public companies are required to produce accurate financial reports, governments should produce and disseminate clear, comprehensive reports on their activities, budget usage, and policy outcomes. These should be made accessible to the public to enhance transparency and trust.

Diverse Representation. Board member selection in companies often considers diversity to ensure a range of perspectives. Similarly, public advisory boards or committees should reflect the diversity of the national population in terms of gender, ethnicity, age, and professional experience to enrich decision-making at the national level.

Appendix B: Artificial Intelligence

Artificial Intelligence (AI) has changed substantially since this concept was first drafted in 2023. Frontier large language and multimodal models — systems such as Anthropic's Claude, OpenAI's GPT, Google's Gemini, and equivalent open-weight models — can now read, summarise, analyse, and cross-reference very large bodies of heterogeneous information (statistical releases, regulatory filings, press, scientific literature, satellite imagery, audio, video) at a speed and breadth that was infeasible only a few years ago. Specialised research, statistics, and audit systems have begun to incorporate these models as components.

Competent Democracy uses AI for two narrow and well-bounded purposes: suggesting goals in support of expert review, and measuring progress against goals that humans set and humans approved. The system does not delegate goal-setting, policy choice, or execution to AI. That distinction is structural, not stylistic.

B.1 What modern AI is genuinely useful for here

Today's frontier models are well-suited to several tasks at the heart of the measurement problem:

• Aggregating heterogeneous evidence. Combining structured sources (national statistics, registers, sensor feeds) with unstructured sources (press, official documents, scientific literature, transcripts) into consistent, source-linked summaries.
• Cross-checking claims against sources. Verifying whether a stated outcome is supported, contradicted, or unaddressed by named sources, with explicit citations.
• Time-series analysis and trend detection. Identifying inflection points, seasonal effects, and anomalies in long-running indicators.
• Drafting candidate goals. Proposing measurable, time-bound, budgeted goals from a domain brief (see Appendix C), for review by an expert committee.
• Surfacing data anomalies for audit. Flagging patterns consistent with reclassification, definitional changes, or coordinated submission of low-quality data.

In each case the model produces drafts, summaries, and flags for human decision — not decisions.

B.2 What modern AI is not reliable for, and how the system handles that

Frontier models have well-documented limitations that bear directly on this use case:

• Hallucination. Models can produce confident, plausible statements that are not supported by their inputs. Competent Democracy's scoring framework therefore (a) cites every input that contributed to a score, (b) refuses outputs that cannot be traced to a whitelisted source, and (c) requires that score-changing claims be reproducible from public inputs.
• Sycophancy and prompt-induced bias. Models can be steered by the framing of their inputs. The framework therefore fixes the question schema per metric type and forbids ad-hoc prompting in production scoring.
• Embedded normative choices. Choices that look technical (which sources to weight, how to handle missing data, what to treat as "the same indicator") can encode value judgements. This is precisely the risk the annual normative audit by the Academic Supervisory Committee (§7.7) is designed to expose.
• Training-data contamination. A model that has been trained on the country's press and political discourse will inherit the biases of that corpus. The framework documents the provenance and date-range of training data for every model used in production scoring.

B.3 An illustrative use case

Consider monitoring acute-care overcrowding in a national hospital system. A modern AI scoring pipeline can:

• Continuously ingest emergency-department waiting-time data from the national health statistics agency.
• Cross-reference with structured outcome data from hospital information systems.
• Read unstructured clinical-coding bulletins and operational notices to detect reclassification events (e.g. patients being moved from "waiting" to "admitted-but-not-treated") that would otherwise inflate the headline indicator.
• Run trend analysis to distinguish genuine improvement from seasonal noise.
• Surface anomalies — including statistically improbable improvements — for review by the Health and Care expert committee.

The pipeline does not decide whether the goal has been met; it produces a score and a fully-cited evidence pack. The Health and Care committee, the appeals committee, and the academic supervisory committee can all interrogate that pack.

B.4 Bias and its mitigation

It remains essential to acknowledge that AI systems can be biased. Bias in AI systems arises whenever the data used to train or operate them is unrepresentative, incomplete, or systematically skewed. Common forms include:

• Sampling bias — training or input data drawn from a population that is not the population the system serves.
• Confirmation bias — system design that effectively privileges a prior assumption, including via the choice of indicators or sources.
• Algorithmic bias — model design or training procedures that produce systematically unfair outcomes across groups.
• Dataset bias — the input data itself encodes a historical inequity that the model then propagates.

Competent Democracy addresses these structurally rather than relying on the goodwill of any single operator:

• Whitelist-only inputs. Score inputs are restricted to the data-source whitelist published per §7.1; no model may ingest sources outside it.
• Reproducibility. Every score must be reproducible from public inputs and the published framework version.
• Open-source code and parameters. §7.4 makes the codebase, the per-goal specs, and the framework versions available for inspection.
• Independent academic review of every framework release, explicitly searching for normative choices embedded as technical defaults (§7.6).
• Annual normative audit by the ASC, published and parliament-debated (§7.6, §7.7).
• Diverse stakeholder involvement in design and testing, including international experts and opposition committee members.

B.5 Why open-source matters specifically here

For consumer AI products, open-source is a preference. For a measurement system that determines whether a government has kept its promises, open-source is a precondition. Citizens, opposition, journalists, and academics must be able to inspect how every score was computed; in a closed system, the word "objective" is rhetorical.

This is also why the system is deliberately verifiable rather than merely transparent. Publishing code that no one can actually run, or parameters in a format no one can read, would be transparency in name only. The system commits to publishing in human-readable form, with reproducible pipelines and worked examples, so that the practical cost of auditing a score is low enough that opposition researchers, civil society, and academia will routinely do so.

B.6 The non-decision principle

It bears repeating: Competent Democracy does not use AI to set goals, to govern, or to enforce. AI in this system has three roles only:

1. Suggest goals to expert committees for review.
2. Measure progress against goals that humans set and humans approved.
3. Surface anomalies for human investigation.

Every consequential decision in the system — admitting a goal to the manifesto, approving a score, dismissing a government, sanctioning a politician — is made by humans, by named institutions, against published rules. AI is an instrument of governance, never its agent.

Appendix C: Example goals generated by AI

Comprehensive program to alleviate overcrowding in acute care services in the Netherlands, with specific timelines, measurable metrics, and budget considerations.

1. Capacity and Workflow Optimization Program

• Timeline: 2-year rollout with quarterly milestones.
• Measurable Metrics: 20% reduction in patient wait times by Q4 Year 1; 10% increase in patient throughput by Q2 Year 2.
• Budget: €50 million, with funds allocated to infrastructure expansion and process improvement technology.

2. Health Workforce Support and Expansion

• Timeline: Recruitment over 18 months, ongoing support.
• Measurable Metrics: Hire 500 new healthcare professionals within the first year; achieve a 10% reduction in staff turnover by Year 2.
• Budget: €30 million for recruitment incentives, training, and support programs.

3. Technology and Digital Health Integration

• Timeline: Implementation within 24 months, initial rollout of telemedicine services in 6 months.
• Measurable Metrics: Achieve 15% decrease in non-urgent ED visits via telemedicine by end of Year 1; 25% of all triage to be assisted by AI by end of Year 2.
• Budget: €40 million for technology infrastructure and software development.

4. Public Health Education Campaign

• Timeline: Launch initial campaign within 3 months, ongoing for the duration of the program.
• Measurable Metrics: Reach 80% of the target population with educational materials by Q2 Year 1; demonstrate a 5% reduction in unnecessary ED visits by end of Year 1.
• Budget: €5 million for campaign development, dissemination, and evaluation.

5. Policy and Regulatory Improvements

• Timeline: Policy development in 6 months, implementation over the following 12 months.
• Measurable Metrics: Introduction of 3 new policies within the first year; 20% reduction in inappropriate ED utilization by Q4 Year 2.
• Budget: €10 million for policy development, stakeholder engagement, and implementation costs.

6. Intersectoral Collaboration

• Timeline: Establish partnerships within 12 months.
• Measurable Metrics: Form 5 new private sector partnerships by Q2 Year 1; measure a 10% increase in patient satisfaction due to diversified care options by Year 2.
• Budget: €15 million for partnership establishment, coordination, and integration efforts.

7. Evaluation and Continuous Improvement

• Timeline: Begin concurrent with program rollout, with the first assessment at 6 months.
• Measurable Metrics: Biannual improvement of 5% across all target metrics; annual reduction in wait times and staff burnout rates.
• Budget: €10 million for the monitoring system, data analysis, and continuous improvement initiatives.

Total Estimated Budget: €160 million for 2 years.