Pith. sign in

REVIEW 2 major objections 4 minor 43 references

Elitism in the Aisle: A Long-Run Surname Measure of Legislative Elite Composition in Chile, 1834-2020

T0 review · 2 major / 4 minor · reviewed 2026-07-10 · grok-4.5

Pith's one-line read Chile's Congress went from about half enduring-elite surnames in the 1860s to about 12% by the 2010s, with a sharp 11–13 point drop around 1925, and in 1910–1950 that composition tracks the agenda net of party.

desk verdict A reusable long-run elite-composition instrument for Chile, with a well-supported 1920s break and a net-of-party agenda co-movement result that holds under the paper's own checks. read the letter →

arxiv 2607.08520 v1 pith:ZKZZKSVT submitted 2026-07-09 cs.SI stat.AP

classification cs.SIstat.AP
keywords descriptiverepresentationsubstantivesurnameselitepersistencelegislativespeechChilestructuraltopicmodel1925reform
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Class background of legislators is hard to measure over long historical periods because occupational and biographical records are thin. This paper offers a single, replicable instrument: the enduring-elite surname measure, which flags surnames that sit at the top of today's socioeconomic hierarchy (or among business leaders) and also appear in nineteenth-century elite registers. Applied uniformly to every member of the Chilean Congress from 1834 to 2020, and set against a dynamic population reference built from 22.65 million birth registrations, the elite share falls from a peak near half in the 1860s to about 12% in the 2010s, with an abrupt level drop of 11 to 13 points around the 1925 constitutional reform. In the steep-change window 1910–1950, structural topic models of floor speeches show that, net of party and chamber, common-surname legislators emphasize labor and related social themes while elite-surname legislators emphasize a statecraft cluster of defense, foreign affairs, and administration. The paper's claim is that this long-range surname ruler makes the descriptive-to-substantive link tractable historically: who sits in Congress moves together with what Congress attends to.

What carries the argument

The enduring-elite surname measure: the intersection of a contemporary criterion (surname socioeconomic index ≥65 or strong over-representation among company directors) with appearance in at least one nineteenth-century elite register, yielding a fixed list of 298 surnames applied uniformly across nearly two centuries of the legislative roster and compared to a dynamic population surname stock.

What would settle it

Rebuild the elite list from the same historical registers without the contemporary socioeconomic or business-leader gates, or under substantially different cutoffs and coding rules, and check whether the long decline, the 1925 level shift, and the 1910–1950 labor-versus-statecraft agenda contrast survive at comparable magnitude against the same population reference and speech corpus.

Watch

Extended reading notes

Core claim

A surname measure that requires both contemporary elite standing and historical elite attestation shows that the enduring-elite share of the Chilean Congress declines from about half in the 1860s to about 12% in the 2010s, with a sharp 11–13 point level shift around the 1925 reform; and that in 1910–1950 this composition contrast co-moves with legislative topic emphasis net of party, common surnames toward labor and elite surnames toward statecraft.

Load-bearing premise

That surnames which are both elite today (by socioeconomic index or business over-representation) and attested in nineteenth-century elite registers form a stable long-run proxy for elite descriptive representation rather than mainly reflecting families that survived as elite into the present or the particular cutoffs used to build the list.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper introduces a replicable enduring-elite surname measure (intersection of a contemporary SES/business-leader criterion with nineteenth-century elite registers) and applies it to the Chilean Congress, 1834–2020. Against a dynamic population reference from 22.65 million birth registrations, the elite share falls from ~52% in the 1860s to ~12% in the 2010s, with an interrupted-time-series level shift of 11–13 points around the 1925 constitutional reform. In 1910–1950, a structural topic model of floor speeches shows that, net of party, chamber, and year, common-surname legislators emphasize labor while elite-surname legislators emphasize a statecraft agenda (defense, foreign affairs, administration); within-group decade slopes confirm the gap widens rather than narrows. Secondary text-based ideal points and contemporary NOMINATE checks are reported as corroboration. The contribution is methodological (a long-range, externally anchored elite proxy) plus the Chile application linking descriptive composition to substantive agenda.

Significance. If the measure is accepted as a valid long-run proxy, the paper supplies the longest uniform estimate of legislative elite composition of its kind and a concrete historical instance of co-movement between who sits and what is attended to, net of party. Strengths that raise the contribution include: a public replication repository with the deposited 298-surname list and construction scripts; extensive robustness (window sweeps, break-year placebo, coding-rule and business-route leave-outs, SES cutoffs, persistence-depth gates); external validation (Club de la Unión, elite schools, nineteenth-century kinship centrality, Senate vs Chamber); and careful inference on the agenda claim (theme aggregation, legislator-clustered SEs, method-of-composition for STM θ uncertainty, within-group slope tests). The coarseness (elite vs residual, not middle vs working class) is stated honestly. The design is portable under three stated conditions (lineage-transmitting surnames, long roster, historical + contemporary status signals).

major comments (2)
  1. [Appendix A; H.2–H.6] The central composition and 1925-break claims rest on the measure construction (Appendix A: contemporary SES≥65 or business over-rep ≥8/≥5 appearances ∩ at least one historical register). The paper already bounds the main threats: H.2 shows the unconditional 19c series lies strictly above the enduring-elite series (early share is a lower bound, never inflated by survivorship); H.3–H.6 show the level shift remains large and significant under SES cutoffs 60/65/70, bearer floors, depth≥2–4, OR/AND/paternal coding, and dropping the 13 business-route surnames (−13.3 → −11.1). No further load-bearing inconsistency is identified; residual risk is the acknowledged coarseness already stated in the conclusion. I do not require additional major redesign.
  2. [§6; Table 1; Appendix H.7–H.8] The agenda claim is carried by the within-group decade slopes (labor +0.90, statecraft +1.34 pts/decade, p≤0.002; Tables H15–H16), which survive legislator clustering and method-of-composition for STM θ uncertainty, not by single-theme stars (only labor, foreign relations, and elections clear 5% on the contrast column, and fewer once STM uncertainty enters). That is the right inferential weight. One remaining presentational risk: the main-text Table 1 stars sit on emphasis-vs-other, not on the elite−common contrast the paper interprets. Make that explicit in the table notes and lean harder on the slope tests in the main text so readers do not over-read individual theme magnitudes.
minor comments (4)
  1. [Figure 3; §4] Figure 3 and the over-representation ratios start only in 1884 (civil registration). A short note in the caption or text on how pre-1884 population shares are (or are not) benchmarked would avoid confusion.
  2. [§6; Appendix D] The ideal-point analysis (Appendix D) is correctly demoted to secondary; the categorical common effect is directional but not significant once chamber and tenure enter. A single sentence in the main text stating that the continuous SES index is the more precise of the two stance checks would help readers who skip the appendix.
  3. [§3; Appendix B] Biography coverage is balanced (Table B3), but the corpus itself is not redistributable. The replication note already documents the retrieval procedure; a one-line pointer in the main validation paragraph would be useful.
  4. [Appendix D, E] Minor: arXiv date and some figure filenames in the appendix (e.g., puntos_apellido_revised.png) are internal; clean for journal production.

Circularity Check

0 steps flagged · score 1.0 of 10

No load-bearing circularity: the enduring-elite list is a fixed external construction applied to independent Congress and speech data; self-citations support measure components and one validation, not the time-series or agenda claims.

full rationale

The paper's central results are empirical applications of a fixed surname list (298 names from contemporary SES/business over-representation intersected with independent nineteenth-century registers of landowners, nobility, merchants, wealth holders, etc.) to the full legislative roster and to OCR'd floor speeches. The list is not built from Congress membership, so the observed elite share path (peak ~52% in 1860s to ~12% in 2010s), the 1925 level shift of 11–13 points, and the net-of-party theme contrasts (labor vs. statecraft) are not forced by construction. Validation markers (Club de la Unión, elite schools from Library of Congress biographies; W-NOMINATE) and the population reference (22.65 M birth registrations) are external. Self-citations (Bro & Mendoza 2021 SES index; Bro 2023 kinship network; Bro & Morav 2024) supply measure inputs or a secondary centrality check and are not required for the composition or agenda findings; the paper already bounds survivorship and cutoffs. No fitted parameter is re-labeled a prediction, no uniqueness theorem is imported, and no known result is merely renamed. Score 1 only for the minor, non-load-bearing self-citations that appear in measure construction and one validation figure.

Assumptions & free parameters 5 free parameters · 4 assumptions · 1 invented entities

The central claims rest on a constructed binary elite/common/other partition whose cutoffs and source intersection are chosen by the authors, plus standard assumptions that surnames transmit lineage and that historical registers mark elite standing. No new physical entities; the invented object is the measure itself. Free parameters are the SES threshold, business over-representation thresholds, topic count K, and speech-count floor for ideal points—all subjected to some sensitivity but still load-bearing for the exact magnitudes.

free parameters (5)
  • Surname SES index cutoff = ≥65 (with ≥30 bearers)
    A surname clears Route 1a if mean SES index ≥65 and ≥30 bearers; sensitivity table varies 60/65/70 and changes list size and break magnitude.
  • Business-leader over-representation thresholds = over-rep ≥8 and ≥5 appearances
    Route 1b requires ≥5 appearances and over-rep ratio ≥8 among 5,411 directors/managers; joint sweep of thresholds moves the 1925 break within a band but the cutoffs are hand-chosen.
  • STM topic count K = 80
    Structural topic model fit with K=80; authors argue theme aggregation makes K non-consequential but K is still a modeling choice.
  • Ideal-point speech floor = ≥20 speeches
    Legislators restricted to ≥20 speeches for stable text ideal points, reducing sample from 597 to 281.
  • Segmented-regression break year = 1925
    Break placed at 1925 as best-fitting year; placebo shows global max near early 1920s plateau, so year is data-informed rather than purely exogenous.
assumptions (4)
  • domain assumption Surnames carry ancestry and therefore social-status information across generations in Chile.
    Stated in Introduction and Section 2; underpins the entire measure and is supported by cited surname-mobility literature but not proved here.
  • ad hoc to paper Appearance in at least one nineteenth-century elite register (plus contemporary status) identifies the enduring elite rather than merely contemporary fortunes.
    Criterion 2 in Appendix A; required for every surname including business-route additions.
  • domain assumption Common (most frequent) surnames are the appropriate population denominator for elite over-representation in Congress.
    Section 2: roughly half the population carries at least one common surname.
  • domain assumption Legislative speech topic prevalence, after party/chamber/year controls, is a valid proxy for agenda emphasis relevant to substantive representation.
    Section 6 and Appendix C; standard in text-as-data legislative studies but still an interpretive bridge from words to 'what Congress attends to.'
invented entities (1)
  • Enduring-elite surname measure (298-surname list) independent evidence
    purpose: Provide a single, long-range binary instrument for elite descriptive composition of a legislature when occupational class records are thin.
    Defined as intersection of contemporary SES/business criterion and historical elite registers; validated against external markers but is a new constructed object of this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Elitism in the Aisle: A Long-Run Surname Measure of Legislative Elite Composition in Chile, 1834-2020." pith.science (2026). https://pith.science/paper/ZKZZKSVT

@misc{pith2026260708520,
  author       = {Pith},
  title        = {Pith review of: Elitism in the Aisle: A Long-Run Surname Measure of Legislative Elite Composition in Chile, 1834-2020},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZKZZKSVT}},
  note         = {Machine review of arXiv:2607.08520}
}
read the original abstract

The link between descriptive and substantive representation is well established in the literature but is hard to trace historically, where class records are thin. We introduce a replicable enduring-elite surname measure, pairing a contemporary socioeconomic criterion with historical elite registers, and apply it across the Chilean Congress, 1834-2020. Against a dynamic population reference built from 22.65 million birth registrations, the enduring-elite share of Congress falls from about half in the 1860s to about 12% in the 2010s, with a sharp drop of 11 to 13 points around the 1925 constitutional reform. In 1910-1950, composition co-moves with the legislative agenda, net of party: common-surname legislators emphasize labor foremost, elite legislators a statecraft agenda of defense, foreign affairs, and administration. Across this window, who sits in Congress moves together with what Congress attends to.

Figures

Figures reproduced from arXiv: 2607.08520 by the authors.

Figure 1
Figure 1. Biographical markers of elite status by surname group. Bars give the share of legislators in each surname group (elite, common, and other) who display each of two markers drawn from legislator biographies: membership in the elite Club de la Unión and attendance at an elite twentieth￾century private school. On both markers the common pole sits below the other group and far below the elite pole. Source: legislator bio… view at source ↗
Figure 2
Figure 2. Network centrality of legislators by surname group in the nineteenth-century genealogical network of the political class. Bars give mean centrality (with standard-error whiskers) for elite, common, and other surname legislators on four measures: degree, betweenness, closeness, and eigenvector centrality. Source: genealogical kinship network of the Chilean political class (Bro, 2023). 4 Nearly two centuries of legisl… view at source ↗
Figure 3
Figure 3. Elite, common, and other surname shares of the Chilean Congress, 1834–2020, against a dynamic population reference. new commercial and industrial fortunes), part of the elite decline is the replacement of an old elite by families the measure does not see, turnover as well as democratization. Against a population in which elite surnames are a thin, stable slice (around 2% through￾out), this is dramatic over-represent… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Interrupted time series of the elite surname share of Congress around 1925. Points are the elite share of sitting legislators by year, 1880–1970; the fitted lines are a segmented (interrupted time-series) regression with the break set at 1925. Source: elite-surname cod…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages

  1. [1]

    , title =

    Bailer, Stefanie and Breunig, Christian and Giger, Nathalie and W\"ust, Andreas M. , title =. British Journal of Political Science , year =

  2. [2]

    and Saxton, Gregory W

    Barnes, Tiffany D. and Saxton, Gregory W. , title =. Political Research Quarterly , year =

  3. [3]

    , title =

    Carnes, Nicholas and Hansen, Eric R. , title =. American Political Science Review , year =

  4. [4]

    American Journal of Political Science , year =

    Carnes, Nicholas and Lupu, Noam , title =. American Journal of Political Science , year =

  5. [5]

    American Political Science Review , year =

    Carnes, Nicholas and Lupu, Noam , title =. American Political Science Review , year =

  6. [6]

    Annual Review of Political Science , year =

    Carnes, Nicholas and Lupu, Noam , title =. Annual Review of Political Science , year =

  7. [7]

    Annual Review of Political Science , year =

    Gulzar, Saad , title =. Annual Review of Political Science , year =

  8. [8]

    and Pierson, Paul , title =

    Hacker, Jacob S. and Pierson, Paul , title =. Presidential Studies Quarterly , year =

Show all 43 references
  1. [9]

    International Political Science Review , year =

    Murray, Rainbow , title =. International Political Science Review , year =

  2. [10]

    Comparative Political Studies , year =

    O'Grady, Tom , title =. Comparative Political Studies , year =

  3. [11]

    and Nedreg

    Fiva, Jon H. and Nedreg. Group Identities and Parliamentary Debates , journal =

  4. [12]

    and Hansen, Eric R

    Treul, Sarah A. and Hansen, Eric R. , title =. Political Research Quarterly , year =

  5. [13]

    and Feigenbaum, James J

    Thompson, Daniel M. and Feigenbaum, James J. and Hall, Andrew B. and Yoder, Jesse , title =. 2019 , doi =

  6. [14]

    PLOS ONE , year =

    Bro, Naim and Mendoza, Marcelo , title =. PLOS ONE , year =

  7. [15]

    The British Journal of Sociology , year =

    Bro, Naim and Morav, Liran , title =. The British Journal of Sociology , year =

  8. [16]

    Theory and Society , year =

    Bro, Naim , title =. Theory and Society , year =

  9. [17]

    Party Politics , year =

    Byrne, Kevin and O'Malley, Eoin , title =. Party Politics , year =

  10. [18]

    Explorations in Economic History , year =

    Clark, Gregory and Cummins, Neil and Hao, Yu and Diaz Vidal, Dan , title =. Explorations in Economic History , year =

  11. [19]

    Dolores and Ortu\ no-Ort\'in, Ignacio and Romeu, Andr\'es , title =

    Collado, M. Dolores and Ortu\ no-Ort\'in, Ignacio and Romeu, Andr\'es , title =. Investigaciones Econ\'omicas , year =

  12. [20]

    Jaramillo-Echeverri, Juliana and \'Alvarez, Andr\'es and Bro, Naim , title =

  13. [21]

    and Yoo, Choong-Yuel , title =

    Jung, Jay Heon and Kumar, Alok and Lim, Sonya S. and Yoo, Choong-Yuel , title =. Journal of Accounting and Economics , year =

  14. [22]

    Cardoso, Fernando Henrique and Faletto, Enzo , title =

  15. [23]

    Correa, Sof\'ia , title =

  16. [24]

    , title =

    Madrid, Ra\'ul L. , title =. Comparative Politics , year =

  17. [25]

    , title =

    Rueschemeyer, Dietrich and Stephens, Evelyne Huber and Stephens, John D. , title =

  18. [26]

    Salazar, Gabriel , title =

  19. [27]

    , title =

    Scully, Timothy R. , title =

  20. [28]

    Samuel , title =

    Valenzuela, J. Samuel , title =

  21. [29]

    , title =

    Dahl, Robert A. , title =

  22. [30]

    2023 , doi =

    Coppedge, Michael and others , title =. 2023 , doi =

  23. [31]

    2023 , note =

    Registry of directors and managers of Chilean publicly traded companies , howpublished =. 2023 , note =

  24. [32]

    and Stewart, Brandon M

    Roberts, Margaret E. and Stewart, Brandon M. and Tingley, Dustin and Lucas, Christopher and Leder-Luis, Jetson and Gadarian, Shana Kushner and Albertson, Bethany and Rand, David G. , title =. American Journal of Political Science , year =

  25. [33]

    , title =

    Vafa, Keyon and Naidu, Suresh and Blei, David M. , title =. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL) , pages =. 2020 , doi =

  26. [34]

    and Rosenthal, Howard , title =

    Poole, Keith T. and Rosenthal, Howard , title =

  27. [35]

    Clark, Gregory , title =

  28. [36]

    Pitkin, Hanna Fenichel , title =

  29. [37]

    Journal of Politics , year =

    Mansbridge, Jane , title =. Journal of Politics , year =

  30. [38]

    and Levitt, Steven D

    Fryer, Roland G. and Levitt, Steven D. , title =. The Quarterly Journal of Economics , year =

  31. [39]

    Journal of Political Economy , year =

    Abramitzky, Ran and Boustan, Leah Platt and Eriksson, Katherine , title =. Journal of Political Economy , year =

  32. [40]

    , title =

    Yeager, Gertrude M. , title =. The Americas , volume =. 1979 , pages =

  33. [41]

    The Review of Economic Studies , year =

    Dal B\'o, Ernesto and Dal B\'o, Pedro and Snyder, Jason , title =. The Review of Economic Studies , year =

  34. [42]

    American Economic Review , year =

    Cruz, Cesi and Labonne, Julien and Querub\'in, Pablo , title =. American Economic Review , year =

  35. [43]

    , title =

    Smith, Daniel M. , title =. 2018 , publisher =

Pith tools

Reviewed July 10, 2026 · model on record in the stance chip above.