Seven Years of Assembly Seat Reduction Debates — The Time Structure of 'Candidate Shortage' and Seat Apportionment Ordinance Revisions in Japanese Local Assemblies

This note aggregates, by year and prefecture, speeches containing terms such as "assembly seat reduction" (giin teisū sakugen), "seat apportionment ordinance" (giin teisū jōrei), "candidate shortage" (narите busoku), "uncontested elections" (mutōhyō tōsen), and "assembly member remuneration" (giin hōshū) from the machikarte lab's assembly speech dataset — a corpus of over one million assembly speeches as of 2024. The aim is to read the time structure and geographic distribution of debates on assembly seat apportionment and candidate shortages. The purpose is not to evaluate individual assemblies or individual assembly members.

Over the seven years from 2018 to 2024, mentions of assembly-seat-related terms in Japanese local assemblies show neither a simple increase nor a simple decrease, but a wave-shaped variation. From 2018 to 2020, mentions held at a broadly stable level; after a relative decline phase spanning 2021-2022, they rose again in 2023-2024, with 2024 reaching one of the highest levels in the seven-year window.

Whereas debates on fiscal consolidation and young carer issues follow a monotonic increase accompanying "the emergence of new policy concerns," debates on assembly seat apportionment show a structural difference: as an already-established concern, they exhibit a periodic rise-and-fall cycle. Because seat apportionment ordinance revisions tend to occur roughly once per term, the relationship between those revisions and the electoral cycle likely influences the volume of assembly debate.

Comparing the 2024 single-year discussion density by prefecture — mentions per municipality, for prefectures with six or more recorded municipalities — the gap between the top and bottom groups spans approximately five-fold. In regions where population decline is advancing, a structure is observed in which pressure to reduce seat counts tends to become intertwined with fiscal consolidation debates.

Institutional Background of "Candidate Shortage"

The Ministry of Internal Affairs and Communications (MIC) Research Group on Local Assemblies and Assembly Members published its report in September 2020, identifying as structural challenges the increase in uncontested elections, the difficulty of securing assembly members in small-scale local governments, and the level of assembly member remuneration. The influence this report had on assembly mentions after 2020 can be observed as the same pattern seen in other policy cycles — such as young carers and LGBTQ issues — in which national policy debates percolate down to local assemblies.

The research group's issues summary proposed a shift from a simple debate of "should seats be reduced?" to a compound framework of "securing an appropriate number of members and maintaining assembly functions." The extent to which this shift in framing is reflected in local assembly mentions can only be verified by adding contextual classification.

The Debate Structure Around Seat Reduction and Fiscal Efficiency

Assembly seat and remuneration reductions frequently intersect with fiscal consolidation debates (see trends-fiscal-health-2018-2024). Cases in which seat reduction is raised in the context of fiscal compression, and cases in which seat retention or increase is raised from the perspective of strengthening assembly functions and ensuring diversity, are both included. Accordingly, "an increase in mentions of seat reduction" does not mean an increase in one-directional advocacy.

Correspondence With the Electoral Cycle

Because revising seat apportionment requires going through the procedure of amending the seat apportionment ordinance prior to elections, debate volume tends to concentrate in election years and the preceding year. Years that are the year before unified local elections (every four years) — such as 2018 and 2022 — show relatively higher mentions, while election-following years (2020 and 2024) are influenced by separate institutional milestones. The existence of this periodicity suggests that cycle effects need to be taken into account in time-series comparisons.

Regional Variation in Population Decline and Seat-Reduction Debates

The approximately five-fold prefectural difference in mention density may bear some relationship to the pace of population decline. In municipalities where population is declining rapidly, multiple problems tend to become simultaneous agenda items: a shrinking electorate, difficulty securing candidates, and the fiscal burden of maintaining current seat counts. However, the relationship between population decline and mention density is not one-to-one; differences in regional political culture and assembly practice operate in parallel.

The Impact of the 2023 Local Autonomy Act Revision on Assembly Mentions

The Local Autonomy Act (Chihō Jichi Hō) revision enacted in June 2024 included an expansion of the scope within which the national government can issue directives to local authorities (chihō jichitai), adding a new context to debates about the role of local assemblies. Whether this revision discussion contributed to the elevated mention level in 2024 can only be disentangled by adding keyword co-occurrence analysis.

Correspondence With the "Response Deferral" Structure

Because assembly seat and remuneration revisions are agenda items that directly touch the interests of assembly members themselves, the proportion of administrative responses in which a "under consideration" (kentō) reservation is entered may be higher compared with other policy subjects. A pathway exists to verify deferral patterns in seat and remuneration debates by combining this with the deferral rate analysis (see case-sakiokuri-rate).

• No separation by position: Speeches calling for seat reduction, seat maintenance, or seat increase are aggregated together; classification by stance has not been carried out.
• Electoral cycle effect not corrected for: The pre-election-year peaks in the four-year cycle complicate comparisons.
• No separation by speaker type: Assembly members' problem-raising statements and administrative responses are not distinguished.
• Coverage bias: Year-to-year variation in the number of municipalities included affects absolute counts.
• No naming of individual municipalities: This article adheres strictly to structural analysis; no ranking by municipality name or assembly member name is presented.

The query specification used for aggregation (spec_version v1-assembly-seats-2026-07) is stated explicitly at the end of this article, and the BigQuery (data warehouse) aggregation query is also available in the machikarte repository on GitHub.

Aggregation Query (spec_version v1-assembly-seats-2026-07)

Annual trend (2018-2024):

SELECT
  year,
  COUNT(*) AS assembly_seats_mentions,
  COUNT(DISTINCT municipality_code) AS municipalities
FROM `correlate-workspace.isvd_machikarte.speeches`
WHERE year BETWEEN 2018 AND 2024
  AND REGEXP_CONTAINS(body, r'(議員定数削減|議員定数.*条例|なり手不足|無投票当選|議員報酬.*削減|定数.*削減)')
GROUP BY year
ORDER BY year

Research Group on Local Assemblies and Assembly Members (地方議会・議員に関する研究会) — Ministry of Internal Affairs and Communications (総務省). Ministry of Internal Affairs and Communications

Issues Summary — Research Group on the Future of Local Assemblies and Assembly Members (地方議会・議員のあり方に関する研究会 論点整理) — Ministry of Internal Affairs and Communications (総務省). Ministry of Internal Affairs and Communications

Final Report — Research Group on the Future of Local Assemblies and Assembly Members (地方議会・議員のあり方に関する研究会 報告書) — Ministry of Internal Affairs and Communications (総務省). Ministry of Internal Affairs and Communications

machikarte — National Local Assembly Speech Search Platform (beta) — ISVD. ISVD

machikarte (GitHub) — schema, aggregation queries, license (MIT + CC BY 4.0) — ISVD. GitHub

Verification of a Role Classification Method for Diet Meeting Record Speeches Using a BERT-Based Classifier (BERTベース分類器を用いた国会会議録発言文の役割分類手法の検証) — Miyaki, Y. & Uchida, Y.. Journal of Japan Society for Fuzzy Theory and Intelligent Informatics (知能と情報), Vol. 37, No. 1, pp. 530-534