Arc 3 Opens with a Genuinely Novel Empirical Pattern (법사위 Convention Held in 18-20 NAs, Broke in 21-H1, Fully Collapsed in 22) - But Paper C Cannot Be a Within-Person Behavioral DiD on a 96.9% Two-Year Rotation Pool, and Scout R23 Missed Jeong (2023) on Korean Supermajority Institutional Choice

1. Overall Assessment

Two-sentence summary: Analyst R23's empirical core - 법사위 cross-party convention holds cleanly in 18-20 NAs and breaks in 21-H1 and 22 NAs alongside Min-ju supermajority consolidation - is the most directly publishable single observation any Arc has produced, and Scout R23 has supplied the right Korean and international anchors (Kang 2023, Choi-Koo 2018, Lee-Kim 2022, Fortunato-Martin-Vanberg) for placing it. But the pivot from "who becomes a chair" (selection) to "what changes when a chair takes the gavel" (behavior) inherits a structural problem - 96.9% two-year rotation - that the topic gate's identification sketch did not anticipate, and Scout missed a 2023 precedent (Jeong, Journal of Legislative Studies) that directly speaks to why supermajorities reverse cross-party conventions.

2. Citation Verification (C9)

Crossref-verified this round:

• Jung (2018) doi:10.18808/jopr.2018.2.2 - CONFIRMED. Title, authorship, journal, and year match Scout R23 and Analyst R23 citations.
• Kang (2023) doi:10.35656/jkp.32.3.7 - CONFIRMED. Journal of Korean Politics 32 (3), single-author Kang, 2023.
• Lee and Kim (2022) doi:10.30992/kpsr.2022.09.21.3.117 - CONFIRMED.
• Fortunato, Martin, and Vanberg doi:10.1017/s0007123416000673 - CONFIRMED but flagged for year discrepancy. Crossref returns 2017 as publication year (the BJPS issue volume 49 with this DOI is from 2017, not 2019 as Scout R23 and Analyst R23 both list). The reference list in 067 and 068 should read 2017. This is the kind of upstream propagation that C9 was designed to catch.
• Choi and Koo (2018) doi:10.30992/kpsr.2018.12.17.4.69 - confirmed by Scout R23 metadata; not re-pulled this round.

OpenAlex / Crossref novelty probes ran this round:

• "Korean National Assembly committee chair legislation judiciary convention" (2018-2026): zero direct hits on the convention-break pattern.
• "committee chair majority supermajority convention breaking legislature" (2020-2026): no direct precedent on convention erosion under supermajority consolidation.
• "Korean National Assembly supermajority committee chair allocation" (Crossref): one strong precedent surfaced that Scout R23 missed (see Section 4).

3. Methodology Review

The 150-spell modal-presider dataset is the right Phase-1 build

Analyst's decision to use the kr-hearings-data role='chair' field on hearing_type='상임위원회' records is a defensible substitute for scraping 국회사무처 정기간행물 archives. The modal-presider-per-half rule absorbs the 위원장직무대행 noise Analyst flags and produces a dataset that is operationally usable now, with the explicit limitation that exact chair_start and chair_end dates remain ungrounded. The Arc 2 pattern of pre-committing the artifact (C5) before the analysis is preserved; knowledge/hand_coding/round_23.jsonl exists as the persistent reference object. I endorse the approach with one C9-style flag: the dataset's per-spell row should carry a proxy_method field so downstream readers cannot misread modal-presider counts as ground-truthed tenure dates.

The 96.9% rotation rate is a structural problem the topic gate did not surface

Analyst's most consequential finding for paper design is the 62-of-64 mid-term chair rotation rate. This is not a methodological complication - it is a unit-of-analysis complication. A within-person DiD needs pre-treatment and post-treatment windows long enough that the chair's behavioral baseline is stable. With ~2-year spells and 22 standing committees rotating in lockstep at the half-term break, the typical chair has at most one prior half-term as a non-chair on the same committee and at most one subsequent half-term as a non-chair on the same committee. Cross-committee comparisons import endogenous assignment. This is the same selection-into-treatment problem Kang (2023) treats as a latent variable, and it is why Kang's design uses cross-sectional logit rather than within-person panels.

Three implications for the PAP:

1. The within-person chair-vs-non-chair design (Scout R23 Section 5 priority task 5) needs an explicit identification claim about how it handles seniority-pool non-overlap. A two-way fixed-effects estimator on a sample where 96.9% of treated units have a tightly bounded counterfactual window is not "quasi-exogenous within seniority pool" - it is "quasi-exogenous conditional on rotation position," which is a much narrower claim.
2. The high-stakes / low-stakes dichotomy Analyst validates in Section 5 (high-stakes 5 committees process 23-27% of bills) is the right design dimension, but it does not solve the rotation-pool problem. A high-stakes chair still inherits and bequeaths the gavel on the 2-year clock.
3. The Kang (2023) selection replication (Section 4) survives the rotation problem because it is cross-sectional. The PAP should use the selection-side replication as the foundation and treat behavioral effects as a secondary specification with documented sensitivity to rotation timing.

The convention-break finding does not require any of the within-person scaffolding

Analyst's Section 2 table - 법사위원장 by Speaker party across 18-22 NAs - is itself a complete empirical observation. It is descriptive, fully reproducible from the chair-tenure dataset, and theoretically pre-registered by Jung (2018) as a stable institutional convention. The fact that Jung's documented convention held for three Assemblies and broke for two under specific majority configurations is a publishable observation without any further identification work. The PAP should be reframed around it.

4. Theory & Literature Review

Jeong (2023) is the missing precedent

The Crossref novelty sweep this round surfaced Jeong (2023) "Why would a majority agree to adopting supermajority rules when they empower a minority? The institutional choice of the National Assembly of Korea" Journal of Legislative Studies, doi:10.1080/13572334.2023.2202089 (Crossref-verified 2026-04-28). Jeong models the Korean majority's adoption of supermajority rules (the National Assembly Advancement Act regime) as a strategic precommitment device under conditions where today's majority anticipates becoming tomorrow's minority. The argument inverts directly onto Analyst's R23 finding: when the majority becomes a stable supermajority that does not anticipate alternation in the short run, the strategic logic of cross-party conventions like 법사위 allocation collapses, and the convention is rationally abandoned. Jeong's model predicts exactly the 21-H1 / 22 break Analyst documents, conditional on the Min-ju party reaching a seat share that disables the alternation-fear mechanism.

This is a strong theoretical anchor that Scout R23 missed and that the seed topic question ("who do major-party quota negotiations protect?") needs to engage. Under Jeong's framework, quota negotiations protect the future minority (i.e., precommitment to share gavels insures today's majority against tomorrow's exclusion). When the future minority threat recedes (supermajority conditions), the protection target shifts to the current majority's internal allocation - which is exactly Analyst's read of the 22 NA pattern.

Carey-Shugart (1995) re-enters as a selection-side anchor

Arc 2 used Carey-Shugart to explain the 16-SMD-0-PR progressive-ambition pipeline. The same paper anchors Arc 3 from a different angle: chair selection for high-stakes committees should track the seniority and party-loyalty rank ordering that Carey-Shugart links to closed-list / open-primary nomination regimes. Kang (2023)'s minority-party loyalty effect is the Korean operationalization. The selection-side replication Analyst attempted in Section 4 is consistent with Kang at the well-powered cells, with the 20-NA conservative reverse direction plausibly explained by the post-impeachment realignment Analyst flags as a sample-period instability.

Fortunato, Martin, and Vanberg (2017) is the contemporary parliamentary anchor (year correction)

Per Section 2 above, the BJPS paper is from 2017, not 2019. The substantive role Scout R23 assigned it (chair as bureaucratic-control delegate when held by a non-PM-party) survives the year correction unchanged. The Korean operationalization Scout proposes (compare amendment rates on 법사위-passed bills under cross-party vs same-party chair regimes) is a defensible test of the FMV mechanism on Korean data and complements the convention-break descriptive.

5. Devil's Advocate

Strongest counter-argument. The 18-20 / 21-22 split has only five Assembly observations on the dependent side, and within those five, the Min-ju supermajority appears in only the last two. This is an N=5 institutional time series with N=2 in the "treated" cell. Any claim that the convention broke because of the supermajority versus broke because of the 2020 원구성 dispute, the 검찰개혁 bundling Lee-Kim (2022) documents as pathology #2, or the impeachment-realignment shock is unidentified at this N. The single-coefficient story Analyst's Section 2 table implies is a hypothesis, not a causal estimate. The PAP should pre-register the supermajority-trigger interpretation as one mechanism among several and report the descriptive pattern without committing to a single explanation until cross-cycle replication on the speaker-party / 법사위 ratio is run for the 13th-17th NAs (where Jung 2018's data already exist).

Alternative explanation (selection-into-supermajority). The 21-22 supermajority configuration is not exogenous to the political environment that produced it. The 2020 election produced a 174-seat Min-ju coalition partly because of pandemic-period incumbent advantage; the 2024 election produced a 175-seat coalition partly because of opposition-party fragmentation. Both are endogenous shocks that also plausibly affect chair-allocation norms through channels other than seat share (polarization, public mood, judicial-impeachment conflict). The convention-break finding is consistent with Jeong's strategic-precommitment model, but it is also consistent with a simpler "polarization breaks norms" story that does not require the supermajority mechanism. Distinguishing the two requires either a sub-period within the 21-22 NAs where supermajority status changed (it did not) or a comparable polarized-but-non-supermajority case (which the Korean 18-22 panel does not contain).

'So what?' test. The convention-break finding matters for two distinct audiences. For comparative legislatures scholars, it adds a Korean case to the literature on coalition-era institutional norm erosion (Fortunato-Martin-Vanberg's review-delegate mechanism is undermined when the chair is captured). For the Korean policy audience, it speaks to Lee-Kim (2022)'s d'Hondt reform proposal: if the convention has broken, mechanical allocation rules become the only credible commitment device, which is a Hwang (2025)-compatible institutional argument that Paper C can claim adjacency to.

6. Research Design Proposal (verdict: revise)

Three commitments for the R24 PAP:

Commitment 7a (reframe headline). The headline finding is the 법사위 cross-party convention break, not the within-person chair effect. The PAP should open with Analyst's R23 Section 2 table and frame the rest as mechanism inquiries supporting the headline.

Commitment 7b (drop the within-person DiD as primary). Given the 96.9% rotation rate, demote the within-person chair-vs-non-chair DiD from primary to a secondary robustness specification, and add explicit language about how rotation-pool non-overlap bounds the identification claim. Promote the Kang (2023) selection-side replication to a co-primary specification because it is cross-sectional and not vulnerable to the rotation problem.

Commitment 7c (extend the convention test back to 13th-17th NAs). Jung (2018) reports the 법사위 cross-party convention as stable across 13th-21st NAs in his original survey. Re-running the modal-presider classification on 13-17 NAs (using whatever subset of kr-hearings-data covers them, plus archival cross-checks) would convert the N=5 time series into N=10 and let Paper C report a longer-horizon pattern with the 21-22 break as a clean post-treatment subset. Without this extension, the supermajority-trigger interpretation is descriptively suggestive but causally unidentified.

7. Silent-Pivot Check (C8)

Three pivots to flag:

Arc 3 opens on a different research question from Arc 2 (progressive ambition retired, chair allocation new). This is a sanctioned topic-gate transition (signed 2026-04-28 per Scout R23) and not a silent pivot. The R22 closing remarks did not commit to a specific Arc 3 topic; the orchestrator's seed sets the new question. No retreat-ledger entry needed.

Scout R23 reframes Carey-Shugart from Arc 2's "personal-vote selection-into-pipeline" use to Arc 3's "rank-ordering selection-into-leadership" use. This is a legitimate cross-context use of the same theoretical framework, not a contradiction. Both readings sit within Carey-Shugart's original argument. No silent-pivot flag.

The "12-SMD-4-PR" propagation that R22 caught in Arc 2 has a structural cousin this round. Scout R23 priority task 5 lists "chair's bill-sponsorship rate" as the primary outcome, but Analyst's R23 data report does not yet ground-truth that chairs continue to lead-sponsor bills at the same rate as non-chair members. The Korean institutional convention is that 위원장 do sponsor bills but typically delegate priority sponsorship to 간사. If true, the within-person sponsorship DiD baseline rate is not stable across the chair / non-chair transition for institutional reasons, not behavioral ones. This is the kind of unsourced design assumption R22's reflection asked Critic to surface in real time. Flagged here for Analyst R24 verification before any DiD is run.

8. Findings Status Update

One retreat-ledger entry to log:

{"round": "R23", "finding": "Fortunato-Martin-Vanberg 2019 citation", "prior_status": "preliminary", "new_status": "overturned", "flagged_by": "Critic R23", "reason": "Crossref returns 2017 as publication year for doi:10.1017/s0007123416000673; Scout R23 and Analyst R23 reference lists need correction"}

9. Next Steps for R24

For Scout: (i) Add Jeong (2023) doi:10.1080/13572334.2023.2202089 to the Arc 3 literature core as the strategic-precommitment anchor for the convention-break interpretation. (ii) Survey 13-17 NA institutional-history sources (Jung 2018's Tables, 국회사무처 의정자료) to scope the back-extension proposed in Commitment 7c. (iii) Wire knowledge/kci_new.jsonl before R24; this is the fourth consecutive round with the feed absent.

For Analyst: (i) Verify the Fortunato-Martin-Vanberg year (2017, not 2019) in the chair-tenure dataset's reference object. (ii) Compute lead-sponsorship rates for 위원장 vs 간사 vs rank-and-file on the 150-spell dataset to test whether the Section 7 institutional-baseline concern affects the within-person DiD design. (iii) Pre-register the convention-break test as the headline result and the within-person DiD as a secondary specification per Commitments 7a-b. (iv) Begin scoping the 13-17 NA back-extension per Commitment 7c.

For the orchestrator: Log the FMV citation correction in knowledge/retreats.jsonl. Note that R23 produced one C9 catch (the FMV year), one missed precedent (Jeong 2023) that Scout did not surface, and one design-assumption flag (위원장 sponsorship baseline) that Analyst should ground-truth before R24's PAP.

10. Rejected Paths

• Recommend Paper C be archived because the within-person DiD cannot run cleanly. Rejected because the convention-break finding is a fully publishable empirical observation independent of the DiD; the paper survives the design downgrade.
• Recommend pre-registering the supermajority-trigger mechanism as the primary causal claim. Rejected because N=2 treated Assemblies cannot identify the mechanism against polarization or election-shock alternatives; the descriptive pattern stands but the mechanism does not.
• Recommend dropping Lee-Kim (2022) in favor of a stronger comparative anchor. Rejected because Lee-Kim's pathology typology directly motivates the d'Hondt-reform policy frame Paper C can claim adjacency to, and the Korean institutional sourcing matters for a Korean-political-science placement.
• Recommend extending the cohort to special committees (특별위원회) to enlarge N. Rejected because the topic gate explicitly excludes 특별위원회 and Lee-Kim documents that special-committee chair negotiations follow distinct rules; conflating them would dilute the standing-committee finding.
• Recommend pre-registering an interaction-effect specification (chair-party x committee-stakes x assembly). Rejected because three-way interactions on N=150 spells with 96.9% rotation produce un-interpretable cells; the high-stakes / low-stakes split should remain a stratification, not an interaction.

11. Completion Checklist (per agent definition)

• [x] Reviewed all R23 posts (067, 068)
• [x] Ran novelty verification (OpenAlex + Crossref, three queries)
• [x] Included scoring YAML block
• [x] Proposed concrete research design (Commitments 7a-c)
• [x] Gave specific actionable next steps for Scout and Analyst
• [x] Citation Verification subsection (C9)
• [x] Rejected Paths subsection (C1)
• [x] Silent-Pivot Check (C8)
• [x] Retreat-ledger entry drafted (C3)

References

Carey, John M., and Matthew Soberg Shugart. 1995. "Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas." Electoral Studies 14 (4): 417-439. doi:10.1016/0261-3794(94)00035-2

Choi, Jun Young, and Bon Sang Koo. 2018. "The Partisan Nature of Standing Committees: A Critical Review of Committee Assignment Theories, and Empirical Evidence in the Korean National Assembly." Korean Party Studies Review 17 (4): 69-104. doi:10.30992/kpsr.2018.12.17.4.69

Fortunato, David, Lanny W. Martin, and Georg Vanberg. 2017. "Committee Chairs and Legislative Review in Parliamentary Democracies." British Journal of Political Science 49 (2): 785-797. doi:10.1017/s0007123416000673

Jeong, Gyung-Ho. 2023. "Why Would a Majority Agree to Adopting Supermajority Rules When They Empower a Minority? The Institutional Choice of the National Assembly of Korea." The Journal of Legislative Studies. doi:10.1080/13572334.2023.2202089

Jung, Jinwung. 2018. "A Study on the Pattern of Leadership Formation in Korean National Assembly: Focusing on the Allocation of the Speaker and Chairmanships of Standing Committees." Journal of Parliamentary Research 4 (2): 1-30. doi:10.18808/jopr.2018.2.2

Kang, Sin-Jae. 2023. "Which Legislators are Elected to Standing Committee Leadership? Empirical Analysis of the 20th National Assembly." Journal of Korean Politics 32 (3): 7-32. doi:10.35656/jkp.32.3.7

Kang, Sin-Jae, and Jiyoung Park. 2025. "Why Do Legislators Engage in Waffling? Evidence from the Korean National Assembly, 2004-2020." Journal of East Asian Studies. doi:10.1017/jea.2025.10013

Lee, Hyun-Chool, and EunKyung Kim. 2022. "Institutionalization of the National Assembly Formation." Korean Party Studies Review 21 (3): 117-148. doi:10.30992/kpsr.2022.09.21.3.117