Summary
- ARIN now distinguishes General Members, Service Members and Trustee Members. Service Members can receive the full suite of registry services but cannot vote in ARIN elections; General Members can vote if they satisfy status, payment, contact and deadline requirements.
- The change improved the apparent health of elections. Participation rose from 13.93% of eligible General Members in 2022 and 15.48% in 2023 to about 48.85% in 2024 and 47.2% in 2025. The absolute number of participating organizations did not rise in proportion: it was 886, 959, 959 and 772 respectively.
- The denominator changed dramatically. ARIN reported 7,141 General Members before the first large inactivity conversion, then moved 5,179 organizations to Service Member status in January 2024. After the November 2025 conversion it reported 1,467 General Members and 25,029 Service Members.
- The reform can be defended as an honest separation between customers that want to govern and customers that want services. It also means that the word membership no longer identifies the constituency that elects most of the Board. Public claims about participation should therefore show the electorate, all membership classes and the wider service population together.
- Percentage-based quorum, recall and petition rules become easier to satisfy in absolute numbers when the eligible electorate contracts. ARIN should publish cohort flows, application outcomes, affiliations, blank-ballot use and both narrow and broad denominators so that a higher participation rate is not mistaken for wider representation.
The numerator did not explain the headline
Institutional participation is often reduced to one percentage: ballots divided by eligible voters. That ratio is useful only if readers know what the denominator represents and whether it has changed. ARIN's recent election history is a unusually clear demonstration. In 2022, 886 General Member organizations participated in the Board and Advisory Council elections. ARIN reported that as 13.93% of 6,361 eligible organizations. In 2023, 959 participated, 15.48% of 6,197 eligible organizations. In 2024, 959 participated again, but the eligible denominator was 1,963.
The same absolute participation as the prior year therefore represented approximately 48.85%. In 2025, 772 participated out of 1,637, producing the reported 47.2%.
The later percentages look much stronger. They are stronger in one meaningful sense: nearly half of the organizations that had completed the conditions for electoral participation submitted a ballot. The register was no longer dominated by thousands of organizations that had not voted for years. Candidate communications, election administration and quorum calculation were aimed at a group much more likely to respond.
But the change was not mainly an expansion of the numerator. Between 2023 and 2024 the number of participating organizations was exactly 959 in both published result sets. What changed was the institutional definition of the eligible population. ARIN had created a Service Member class, required affirmative qualification for General Member status and applied an inactivity rule that moved thousands of organizations out of the voting class.
That distinction does not make the percentage false. It changes what the percentage answers. The old-looking denominator asked how many eligible organizations on a broad General Member roll voted. The new denominator asks how many organizations that remained or became active General Members voted. The second is a measure of engagement among a selected electorate. It is not a measure of the share of all organizations receiving ARIN services that helped choose the institution's leadership.
What changed on 1 January 2022
The New ARIN Membership Structure for 2022 followed a consultation that ran from 28 October to 29 November 2021. The board exhibit says the distinguishing characteristic between Service and General membership was commitment to vote in ARIN elections. Existing end-user customers were to become Service Members on 1 January 2022, while a Service Member could request General status if it met the criteria. Organizations already listed as General Members at the end of 2021 retained that status initially.
The conversion design was not merely a new label. An applicant for General status had to be a Service Member in good standing with IPv4 or IPv6 resources receiving services under a valid registration services agreement. It had to agree to participate in annual elections, accept inclusion on public membership and voter lists, and agree that its conduct would follow ARIN's stated behavioral and mailing-list rules. The reform therefore linked franchise to a continuing service relationship, an affirmative governance choice, public identification and expected conduct.
The exhibit also announced the inactivity rule. After the 2023 annual election and each election thereafter, a General Member that had not cast a ballot in any of the previous three ARIN elections would revert to Service Member status. ARIN added an abstain or no-vote choice so an organization could record participation without supporting a candidate. It also removed a proposed waiting period: an organization reclassified for inactivity could request General status again without waiting one year.
This was a considered model rather than an accidental database cleanup. Its stated objective was to distinguish customers that intended to participate in governance from those that did not. The institutional question is not whether ARIN had a reason. It is what follows when the active voting class becomes only a small fraction of the organizations called members and an even smaller fraction of the networks and users affected by registry decisions.
Three membership classes, only one organizational franchise
ARIN's current Membership page names three types: Service Members, General Members and Trustee Members. It also states two boundaries that are easy to miss. Membership is not required to obtain direct Internet number resources, and membership is not required to join policy-development discussions, submit consultation suggestions or participate generally in public consultations. Service, policy participation and corporate voting are related, but they are not the same entitlement.
The ARIN Bylaws provide the sharper legal distinction. General Members are entities with a valid ARIN registration services agreement that wish to participate in number-resource policy development and satisfy eligibility and fee requirements established by the Trustees. Service Members also hold a valid agreement and pay applicable fees, but either do not seek General status or do not meet the Board's eligibility requirements. The bylaws expressly say that Service Members do not have the right to vote in ARIN elections.
A General Member in Good Standing is current on fees and has the right to vote and join members-only discussions. Each must designate a Voting Contact. A Service Member in Good Standing may participate in specified lists and discussions and exercise additional rights that the Board grants, but the bylaw text does not give that class the organizational franchise. A Trustee may possess voting rights only if also acting as Voting Contact for a General Member in Good Standing.
The result is not a simple opposition between members and non-members. Service Members really are members under ARIN's own terminology. They can have active contracts, resource records, RPKI access, IRR access and a continuing economic relationship with the institution. What they lack is electoral status. Saying that members elect the Board is therefore incomplete unless membership is qualified to mean eligible General Members acting through valid Voting Contacts.
The first large conversion redrew the map
ARIN's 30 January 2024 reclassification announcement makes the scale visible. It said that, effective from 1 January 2022, General Members had to vote in at least one election within the last three years to maintain status, beginning with the 2021 election. ARIN then reported transitioning 5,179 General Member organizations that had not cast a ballot since 2021 to Service Member status. After that action it counted 1,903 General Members and 23,368 Service Members, with ASN-only organizations included in the latter figure.
This was not a marginal removal of stale records. More than five thousand organizations changed governance class in one action. The count called General Membership moved from the 7,141 reported at the September 2023 election snapshot to fewer than two thousand after the conversion, although those snapshots are not perfectly aligned and should not be treated as a direct reconciliation. The post-conversion Service Member count was more than twelve times the General Member count.
The voting constituency then grew modestly before the 2024 election. ARIN's 2024 election results reported 2,023 General Members as of 9 September and 1,963 eligible organizations as of the later Voting Contact deadline. That suggests additions or returns to General status, but the public result does not provide a complete flow table. It does not say how many applications arrived, how many were approved, how many were declined, how many reclassified organizations returned or how long approval took.
After the election, ARIN moved another 426 organizations that had not cast a ballot since 2022 to Service status. The published post-action totals were 1,642 General Members and 24,464 Service Members. The recurring rule was operating as designed: the voting class would continue to lose organizations that did not participate within the rolling three-election window, while willing organizations could request entry again.
A narrower electorate can be more active
There is a serious case for ARIN's model. A nominal electorate filled with organizations that never intended to vote can distort both institutional planning and public interpretation. Election notices go to unmonitored contacts. Candidate outreach is diluted. A low percentage can suggest apathy even where a stable core is highly engaged. Quorum can be threatened by entities that receive services but have made no governance commitment. Separating an active General Member register from a service register can make expectations explicit.
The 2022 and 2023 figures support the diagnosis of a broad inactive roll. The 2022 results recorded 7,386 General Members and 6,361 eligible organizations, but only 886 participating organizations. The 2023 results recorded 7,141 General Members, 6,197 eligible organizations and 959 entities. More than five thousand eligible organizations did not participate in each year. A membership title alone was not producing an active franchise.
After conversion, participation among those eligible was much higher. In 2024, 959 participating organizations amounted to nearly half the eligible electorate. That is not merely cosmetic. A candidate addressing 1,963 eligible organizations, of which 959 respond, faces a more attentive constituency than one addressing more than six thousand with the same response. Member discussion may become more reciprocal, and an organization that affirmatively requests General status has signaled some willingness to accept governance duties.
The blank-ballot option is also defensible. Status does not depend on endorsing a candidate from an unsatisfactory slate. An organization can register presence while withholding candidate support. Reapplication at any time reduces the risk that inactivity becomes permanent exclusion. These protections show an effort to cultivate participation rather than punish opinion.
Still, active-rate improvement and representational breadth answer different questions. A smaller, self-selected electorate can deliberate well while remaining socially or economically unrepresentative of the service constituency. Institutional honesty requires reporting both qualities rather than allowing one percentage to stand in for both.
The denominator called eligible has several gates
Even General Member status does not by itself place an organization in an election denominator. ARIN's current Elections page explains that an organization must be a General Member in Good Standing 45 days before a ballot or election. It must also designate a Voting Contact with a valid ARIN Online account. The organization votes through that individual.
Each gate has a legitimate administrative purpose. A valid registration agreement establishes an ongoing relationship. Fee currency prevents an expired or delinquent entity from exercising current rights. A named Voting Contact supplies authority and auditability. The 45-day record date lets ARIN certify a list and prevents opportunistic last-minute changes. An authenticated online account supports security.
Together, however, the gates produce at least four relevant counts: all organizations receiving services, all organizations in any membership class, all General Members, and eligible General Members with compliant status and contacts at the record date. A fifth count is participating General Members. The public language often moves between these populations too quickly.
The difference between total General Members and eligible organizations was 1,025 in 2022 and 944 in 2023. After the large conversion it was only 60 at the published 2024 snapshots and 42 in 2025. This narrowing suggests that the active General Member class was also better prepared administratively. Yet the public figures do not break down why an organization failed eligibility. Some may have been late on fees; some may have lacked a valid Voting Contact; some may have had account-association problems. Without categories, the gap cannot be treated as political abstention.
The contact layer adds another complication. The 2023 result attributes 959 participating General Members to 789 Voting Contacts. One person can represent more than one organization, and ARIN explains that a shared contact can receive a single ballot on behalf of represented organizations unless separate accounts are used for different ballots. Organizational ballots are therefore the right election unit, but contact concentration is relevant to independence and should be disclosed in aggregate.
The absolute electorate matters as much as turnout
ARIN's 2025 figures illustrate why percentages should be paired with counts. The 2025 election results report 1,679 General Members at the September status date, 1,637 eligible organizations at the October contact deadline and 772 entities. The 47.2% rate is robust by comparison with 2022. Yet 772 is fewer participating organizations than in any of the preceding three published elections reviewed here.
That does not show a failed election. It shows that election health has several dimensions. A high share of a small electorate can produce clear procedural validity and still rest on fewer organizational decisions. A lower absolute count may be acceptable if the electorate is genuinely open, diverse and accountable to the wider constituency. It is more concerning if entry is poorly understood, concentrated by sector or difficult for small organizations.
After the 2025 election, ARIN's 19 November 2025 announcement reported another 259 organizations moved to Service Member status because they had not participated since 2023. It then counted 1,467 General Members and 25,029 Service Members. These post-election totals cannot be used to recalculate 2025 turnout because they were measured after the ballot and conversion. They do, however, show the scale of the two constituencies at the end of the cycle: roughly seventeen Service Members for every General Member.
The larger group is not disenfranchised in the same sense as citizens denied a statutory vote. Service Members occupy a voluntarily structured corporate relationship and may request General status if eligible. But their operational dependence and fees remain relevant to legitimacy. A Board elected by the smaller group can make decisions affecting services, budgets, contracts and priorities across the larger group. The mandate should be described with that institutional geometry visible.
Quorum becomes easier when the electorate contracts
The bylaws set election quorum at ballots from at least 5% of total eligible voters. Percentage quorum is common because it adjusts to organizational size. It also means that conversion changes not only the reported turnout rate but the absolute minimum required for a valid election.
Using published eligible counts, 5% of the 6,361 eligible organizations in 2022 was 318.05, so at least 319 organizational ballots would clear a whole-ballot threshold. Five percent of 1,637 in 2025 was 81.85, so 82 would do so. Actual participation exceeded both levels comfortably. The point is not that any reviewed election approached failure. It is that the constitutional safety margin now protects against absence among a much smaller class.
Other percentage rules are affected too. The bylaws allow a recall process to begin with signatures from at least 10% of General Members in Good Standing. Petition nominations require at least 2% of eligible General Members, subject to a minimum of 100. The fixed minimum moderates denominator shrinkage for nomination petitions, but recall remains directly sensitive to the size of the class. A governance reform that changes who counts can therefore alter the practical cost of using member remedies even without editing the percentages.
That may be desirable. A recall mechanism should not be impossible because thousands of service-focused organizations ignore governance mail. Conversely, a small active class should not be able to speak as though it were identical to every organization that funds and relies on ARIN. The answer is not necessarily a larger quorum. It is transparent constitutional accounting: every threshold report should state the exact denominator, snapshot date, class definition and corresponding share of the broader service constituency.
The Board defines eligibility for the electorate that elects the Board
The bylaws say General Members must meet eligibility requirements that the Trustees may establish from time to time. They also place ARIN's power, property and affairs under the authority of the Board, subject to corporate law and the Articles. Most Board seats are elected through the General Member franchise. This creates a familiar but important loop: the elected body has power to shape the eligibility conditions of the class that elects it.
Such a loop is not inherently improper. Corporate boards often administer membership criteria under a constitution. Someone must interpret good standing, approve forms, secure elections and adapt to changed risks. ARIN also consulted publicly before the 2022 structure and published the resulting design. The Board did not conceal the reform.
The legitimacy test is stronger because the power concerns the electorate itself. Material changes to entry, inactivity, publication, contact requirements or status loss should receive clear notice, a reasoned paper, consultation and a prospective effective date. The institution should distinguish a technical account rule from a franchise rule. It should publish how many organizations each proposed condition would include or exclude, with uncertainty stated.
Review is also necessary. A Service Member denied General status, or a General Member found ineligible near the record date, needs a fast route to correct factual error before the ballot closes. Ordinary customer support may solve many cases, but the published governance record should identify the decision-maker, timing standard and escalation path. If eligibility is a Board-established condition, aggregate challenge outcomes should reach the Board and membership.
No evidence reviewed here establishes systematic wrongful exclusion. The point is structural. An institution should not ask observers to infer fairness from the absence of visible disputes when the application and correction denominators are not published.
Service Members still form the operating constituency
The General Member franchise is not the only channel of influence. ARIN states that membership is unnecessary for taking part in public policy discussions and consultations. Network operators, end users, technical experts and Service Members can speak in those forums. The Advisory Council works in an open policy environment rather than as a closed parliament. This broader participation reduces the risk that electoral status monopolizes every substantive decision.
Yet open discussion is not equivalent to organizational voting. It does not choose Trustees. It does not automatically approve budgets, executive oversight or corporate strategy. The Board can listen widely while remaining formally accountable through elections to General Members. Policy openness therefore complements the franchise; it does not erase its boundaries.
Service Members also carry practical knowledge. They maintain resource records, sign agreements, pay fees, use routing-security services and serve customers. Some may deliberately avoid General status because governance participation is not central to their mission. Others may lack an attentive contact, find the benefit unclear or face internal authorization barriers. The inactivity classification does not tell us which explanation applies.
Calling these organizations passive would be unjustified. Non-voting in an ARIN election can coexist with active network operation, policy-list participation, consultation responses or regional community work. Conversely, submitting a blank ballot preserves General status without proving substantive engagement. The rule measures one observable act because it is administratively simple. ARIN should describe it as electoral activity, not as a complete measure of community contribution.
This distinction also protects the reform from overclaiming. If the purpose is a responsive voter register, the rule can be judged on that basis. If it is presented as identifying the organizations that care about ARIN, the evidence is too narrow.
One organization, one vote is equal only inside the gate
ARIN General Members cast one vote for each Board and Advisory Council vacancy and one vote on each ballot measure. Unlike APNIC's tiered voting, ARIN does not weight a General Member's ballot by its resource holdings. A small eligible organization and a large eligible organization have the same formal vote. That is an important equality within the electorate.
But equality after admission does not answer equality of access to admission. Organizations must understand the status distinction, hold the relevant agreement and resources, request General membership where necessary, accept public listing, maintain fees, designate a compliant contact and participate within the rolling window. A formally equal ballot can coexist with unequal administrative capacity to preserve it.
Large network operators may have legal, regulatory and public-policy teams that monitor elections. A small municipal network, nonprofit, university or local provider may have one technical contact who changes roles. This is a hypothesis about capacity, not a claim about the current electorate. ARIN's published figures do not provide enough sector and size data to test it.
The appropriate response is not to assign multiple votes to larger organizations. It is to measure access. ARIN could publish conversion and eligibility rates by anonymized resource-service tier, organization type, geography and years of relationship. It could report how often missing contacts, fee status or account links prevented eligibility and whether correction occurred before the ballot. If no meaningful disparity appears, the data would strengthen confidence. If one appears, targeted assistance could improve representation without weakening the activity rule.
Blank ballots are both a safeguard and a weak signal
The blank-ballot option resolves a genuine problem. A member should not be forced to endorse a candidate merely to preserve corporate status. Candidate slates can be uncontested, incomplete or unsatisfactory. Recording presence without a preference respects organizational autonomy and makes the inactivity rule less coercive.
At the same time, a blank ballot reveals how limited the activity metric is. It establishes that an authorized contact accessed and submitted the ballot. It does not establish that the organization reviewed candidate materials, discussed institutional performance or reached a collective view. A marked ballot does not prove those things either, but a blank ballot makes the distinction unmistakable.
ARIN should publish the aggregate number of blank ballots or explicit abstentions where ballot secrecy and system design permit. That figure would help separate participation needed to retain status from candidate choice. It should never identify organizations or contacts. The purpose is not to shame abstention, which is a legitimate choice, but to interpret the active-member denominator accurately.
The same principle applies to shared Voting Contacts. ARIN already reports participating organizations and the number of contacts through whom they acted. A fuller aggregate series could show contacts representing one, two, or more organizations without disclosing names. Concentration may be harmless where related entities use a common officer, but it affects the claim that each organizational ballot represents an independent deliberative center.
No public denominator reviewed here shows blank-ballot use across 2022-2025 or the distribution of organizations per contact. Those facts remain unknown and should remain unknown in the analysis rather than being estimated.
Reapplication makes the boundary permeable, not irrelevant
ARIN emphasizes that a reclassified organization may request General Member status at any time. This is an essential protection. Status loss after inactivity is not a permanent ban, and the board exhibit specifically rejected a one-year waiting period. The current membership page directs eligible organizations to a request action in ARIN Online and tells approved applicants to establish a Voting Contact.
Permeability reduces the severity of conversion. It does not make the boundary irrelevant. Timing matters. An organization that notices its status after the 45-day record date cannot become eligible for that election merely because reapplication is available in principle. Internal approval may take time. Public-listing consent may require legal review. Contact configuration can fail. A delayed return can mean one lost annual vote.
The quality of reapplication should therefore be measured. How many Service Members request General status each year? How many are approved, declined or left incomplete? What are the median and longest processing times? How many begin after election reminders? How many newly approved organizations complete a valid Voting Contact before the deadline? How many reclassified organizations return within one, two or three years?
These are administrative facts, not private political preferences. Aggregate publication would show whether the boundary is truly easy to cross. It would also clarify the source of changes between the post-conversion count of 1,903 General Members in January 2024 and the 2,023 reported before the September election. Without flows, observers can see net movement but not institutional accessibility.
The election series should be read as four different populations
The published figures support a concise longitudinal table, but only with careful labels:
| Election year | General Members at stated snapshot | Eligible organizations | Participating organizations | Participation among eligible organizations |
|---|---|---|---|---|
| 2022 | 7,386 | 6,361 | 886 | 13.93% |
| 2023 | 7,141 | 6,197 | 959 | 15.48% |
| 2024 | 2,023 | 1,963 | 959 | about 48.85% |
| 2025 | 1,679 | 1,637 | 772 | 47.2% |
The table should not be treated as a seamless census. Snapshot dates and eligibility definitions matter. The 2024 and 2025 results use a General Member status date and a later Voting Contact update date. Earlier result notes describe eligibility as good standing plus a properly registered Voting Contact at the September deadline. The election rules evolved administratively even where the basic franchise remained stable.
Nor should the table be joined directly to post-election conversion totals. January and November announcements describe the population after classification actions. They are evidence of the mechanism and scale, not substitute election denominators.
The strongest finding is therefore comparative, not causal in every detail. The eligible electorate fell from more than six thousand in 2022-2023 to fewer than two thousand in 2024-2025. The participation rate roughly tripled, while participating organizations were flat from 2023 to 2024 and then declined in 2025. The first large inactivity conversion accounts for much of the discontinuity. Public evidence does not isolate every addition, departure, fee issue, agreement change or reapplication.
What the Board's mandate does and does not represent
ARIN's Board has broad authority over corporate affairs. General Members elect most Trustees, and eligible General Members also elect the Advisory Council. The election mandate is therefore real. The reviewed elections met the published quorum, produced ranked candidate totals and were certified under the stated procedures.
The mandate is not a plebiscite of every resource holder, every Service Member, every network in the region or every Internet user affected by registry administration. It is a corporate mandate from eligible General Member organizations. That is neither trivial nor universal.
Precise description matters because legitimacy claims can expand in public retelling. Member elected can become community elected; community elected can become accountable to everyone affected. Each step adds a constituency that did not necessarily have a ballot. ARIN can avoid that inflation by pairing election statements with a short denominator note: elected by eligible General Members, while Service Members and the wider community participate through other channels.
The Board's responsibility can still extend beyond its electors. Trustees can owe duties to the corporation and pursue ARIN's mission across the service region. They can consult Service Members, users, governments and technical communities. A narrow electoral base does not require narrow judgment. It does require evidence that broader interests are heard and that the elected group does not mistake its own composition for the whole affected public.
Annual Board reporting could therefore distinguish electoral accountability from service accountability. The first covers candidate elections, member petitions and member communications. The second covers customer satisfaction, service performance, fee consultation, policy openness and impacts across all membership classes. Together they would describe the institution more faithfully than turnout alone.
Better disclosure would resolve much of the tension
ARIN already publishes unusually useful election counts and conversion announcements. The next step is a stable membership accountability table for every year. It should show, on consistent dates where possible:
- Organizations with active registration service relationships.
- Service Members, General Members and Trustee Members.
- General Member requests, approvals, declines, withdrawals and incomplete requests.
- Reclassifications for inactivity and returns after reclassification.
- General Members in Good Standing, those with valid Voting Contacts and final eligible organizations.
- Participating organizations, blank ballots and distinct Voting Contacts.
- Aggregate contact concentration and corporate-group concentration, using privacy-preserving bands.
- Quorum and petition thresholds in both percentage and absolute terms.
The report should retain snapshot dates rather than force false reconciliation. It should explain definitional changes and preserve historical series. It should also include anonymized sector, economy, organization-size and relationship-age distributions, with small cells suppressed for privacy.
These disclosures would let supporters show that the active electorate is open and diverse. They would let critics identify barriers with evidence rather than treating every Service Member as wrongly excluded. Most importantly, they would make it impossible to confuse a higher active rate with a larger democratic reach.
The right legitimacy claim is conditional
ARIN can reasonably say that its current General Member electorate is far more active than the pre-conversion roll. Nearly half of eligible organizations participated in 2024 and 2025, compared with roughly one in seven in 2022 and 2023. It can also say that Service Members retain services, that public policy participation is open more broadly and that reapplication is available.
It should say with equal clarity that the electorate is much smaller. The January 2024 action moved 5,179 organizations out of General status. By November 2025, ARIN reported 1,467 General Members beside 25,029 Service Members. A Board chosen by the former governs an institution serving the latter as well.
The reform's legitimacy therefore depends on four conditions. Entry into General status must be understandable and practically accessible. Reclassification and eligibility decisions must be correct and rapidly reviewable. The electorate must be sufficiently diverse to avoid capture by a narrow professional core. The Board must remain demonstrably responsive to the larger service and policy community that does not vote.
If those conditions are met, the shrinking denominator can be defended as a truthful active register rather than an exclusionary device. If they are not measured, a 47% participation rate offers less reassurance than it appears to. Denominator design is not clerical. It decides whose absence disappears from the statistic and whose consent is treated as institutional authority.
Evidence boundaries and watchpoints
The public record establishes the rule, the principal conversions and annual election counts. It does not establish why individual organizations abstained, whether General Member applicants were refused, how many blank ballots were submitted, how many voters were affiliated, or whether the post-conversion electorate better reflects the region by sector and geography. Those questions remain open.
The next election should be watched for both numerator and denominator. Relevant signals include the September General Member count, final eligible count, participating organizations, distinct Voting Contacts, any new reclassification total and any published reapplication flow. A rise in turnout with stable or growing electorate breadth would be stronger evidence than a percentage rise produced by another contraction.
Governance outcomes matter too. Candidate diversity, contested seats, member petitions, consultation responses and Board attention to Service Member concerns can indicate whether the active electorate improves accountability. None can be inferred from ballot volume alone.
ARIN's reform has made one fact easier to see: service dependence and electoral membership are different relationships. The remaining task is to make that difference visible every time participation is reported. A member-driven institution should never require the public to guess which members drove it.