Exploring the Shape of Economics: A Multilayer Network Analysis of Social Communities and Intellectual Similarity Among Journals Before and After the 2008 Financial Crisis

Alberto Baccini; Carlo Debernardi; Lucio Barabesi

arxiv: 2508.09079 · v2 · submitted 2025-08-12 · 💰 econ.GN · cs.DL· q-fin.EC· stat.OT

Exploring the Shape of Economics: A Multilayer Network Analysis of Social Communities and Intellectual Similarity Among Journals Before and After the 2008 Financial Crisis

Alberto Baccini , Lucio Barabesi , Carlo Debernardi This is my paper

Pith reviewed 2026-05-18 23:25 UTC · model grok-4.3

classification 💰 econ.GN cs.DLq-fin.ECstat.OT

keywords economics journalsmultilayer networksinterlocking editorshipbibliographic couplingtextual similarity2008 financial crisisjournal communitiesknowledge production

0 comments

The pith

Editorial networks dominate hierarchies and knowledge legitimacy in economics journals with little change after the 2008 crisis.

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

The paper maps the social and intellectual structure of economics journals by building a multilayer network from all EconLit-indexed outlets in 2006, 2012, and 2019. Four connection types—shared editors, shared authors, shared references, and article text similarity—are fused into single similarity networks to reveal journal communities. The analysis finds high continuity in these communities across the periods, even as research topics shifted after the crisis, and identifies editorial networks as the strongest factor in ordering journals and determining what counts as legitimate work. A sympathetic reader would care because the result points to social mechanisms, rather than pure intellectual content, as the primary stabilizers of an academic field during external shocks.

Core claim

The study models journals as nodes in a four-layer multiplex network whose layers capture interlocking editorship, interlocking authorship, bibliographic coupling, and textual similarity. Similarity Network Fusion integrates the layers to produce unified similarity networks from which communities are extracted for each of the three years. Comparison across periods shows remarkable structural continuity in both social and intellectual relationships, with the editorial layer emerging as the dominant influence on journal hierarchies and the legitimation of knowledge production inside the discipline.

What carries the argument

Four-layer multiplex network of journals that fuses interlocking editorship, interlocking authorship, bibliographic coupling, and textual similarity through Similarity Network Fusion to detect stable communities.

If this is right

The fundamental relationships among economics journals persist across the 2008 crisis even while article topics evolve.
Editorial overlaps exert stronger influence on journal status than authorship overlaps or citation patterns.
Community detection on fused networks can track long-term stability in the organization of a scientific field.
Knowledge production in economics is shaped more by social selection through editors than by direct intellectual overlap.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same multilayer method applied to physics or biology journals could test whether editorial dominance is specific to economics or common across fields.
Altering the composition of editorial boards might be a more direct route to changing field directions than altering research incentives alone.
The observed stability implies that academic disciplines can absorb large external shocks without rearranging their internal hierarchy of outlets.

Load-bearing premise

The four selected relationship types plus Similarity Network Fusion give an unbiased and complete picture of the social and intellectual structure of the economics journal system.

What would settle it

Repeating the identical four-layer construction and fusion on journals from another discipline or a later economics period and finding a different dominant layer or a sharp structural break tied to an external event.

read the original abstract

This paper develops a multilayer network approach for exploring the evolution of scientific disciplines, using the case of economics before and after the 2008 global financial crisis as a large-scale empirical testing ground. The units of analysis are journals, linked by social and intellectual relationships. The analysis covers all journals indexed in EconLit across three years (2006, 2012 and 2019). In the most recent year (2019), the dataset includes 909 journals, over 30,000 editorial board members, more than 260,000 authors, 134,000 articles, and nearly 2 million cited references. For each period, we model journals as connected in a four-layer multiplex network: the social relationships are based on shared editors (interlocking editorship) and shared authors (interlocking authorship), while the intellectual ones are based on shared references (bibliographic coupling) and textual similarity between articles. These four layers are integrated using Similarity Network Fusion to produce unified similarity networks from which journal communities are identified. Comparing the field across the three periods reveals a high degree of structural continuity. Although research topics changed after the crisis, the fundamental social and intellectual relationships among journals remained remarkably stable. A major result of the analysis is that editorial networks play the dominant role in shaping hierarchies and legitimize knowledge production within the discipline. Whether this finding holds in other scientific disciplines remains an open question for future research.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper finds stable journal communities in economics across the 2008 crisis and flags editorial overlaps as the strongest layer, but the fusion step leaves the dominance claim under-justified.

read the letter

The main thing here is that economics journal networks held their community structure steady through the 2008 crisis even while topics shifted, and the authors point to interlocking editorship as the layer that most shapes hierarchies in their fused model. They built this from the full EconLit corpus for 2006, 2012, and 2019, with four layers covering shared editors, shared authors, bibliographic coupling, and textual similarity, then combined them via Similarity Network Fusion before detecting communities. That scale and the before-after comparison around a real external shock are the concrete pieces worth noting. The continuity observation is straightforward to follow from the time slices and gives a usable empirical snapshot for anyone tracking how a discipline absorbs a major event. The dataset size and the choice to mix social and intellectual ties in one multiplex setup are the parts that feel like actual work rather than routine extension. The soft spot is the dominance claim for the editorial layer. The abstract presents it as a major result, yet the description gives no layer weights, ablation runs, or importance scores from the fusion process. Without those, it is hard to tell whether editors really drive the outcome or whether the method simply amplifies the densest input graph. Parameter choices for the fusion and any robustness checks on community detection are also not visible in the summary, which leaves the attribution open to the concern that it could be an artifact. This is mainly for readers in scientometrics or the sociology of economics who already work with journal networks. Someone looking for a descriptive case study of field stability would get value from the continuity part. It is worth sending for peer review so referees can examine the fusion details and the exact method used to assign dominance to one layer.

Referee Report

2 major / 2 minor

Summary. The paper develops a multilayer network approach to analyze the structure and evolution of economics journals before and after the 2008 financial crisis. It constructs a four-layer multiplex network for journals indexed in EconLit (2006, 2012, 2019) using interlocking editorship, interlocking authorship, bibliographic coupling, and textual similarity; fuses the layers via Similarity Network Fusion to obtain unified similarity networks; detects communities; and reports high structural continuity across periods together with the dominance of editorial networks in shaping hierarchies and legitimizing knowledge production.

Significance. If the results hold under transparent robustness checks, the work provides a large-scale empirical contribution to scientometrics and the sociology of economics by documenting post-crisis stability in journal relationships and by quantifying the outsized role of editorial boards. The dataset scale (909 journals and >30k editors in 2019) and the explicit comparison across three time slices offer a useful benchmark for future studies of disciplinary change.

major comments (2)

[Methods (SNF integration)] Methods section on Similarity Network Fusion: the paper does not report the layer weights, similarity thresholds, or any ablation/importance scores produced by the fusion algorithm. Because the headline claim of editorial dominance and the continuity results are derived from the fused network, the absence of these quantities leaves open whether the reported hierarchy is driven by the editorship layer or by default algorithmic behavior and relative graph densities.
[Results (community structure and hierarchy metrics)] Results section on hierarchy and legitimation: the assertion that 'editorial networks play the dominant role' is not supported by an explicit decomposition (e.g., layer-specific centrality, modularity contribution, or leave-one-layer-out comparisons). Without such evidence the dominance conclusion cannot be distinguished from an artifact of the fusion procedure or data density.

minor comments (2)

[Data description] Clarify whether the >30,000 editorial board members reported for 2019 are unique individuals or total board positions; this distinction affects the interpretation of interlocking editorship density.
[Community detection] Provide the precise community-detection algorithm (e.g., Louvain, Leiden) and resolution parameter used on the fused networks, together with any modularity or stability diagnostics.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major comment below and outline the revisions we will make to improve methodological transparency and empirical support for our claims.

read point-by-point responses

Referee: Methods section on Similarity Network Fusion: the paper does not report the layer weights, similarity thresholds, or any ablation/importance scores produced by the fusion algorithm. Because the headline claim of editorial dominance and the continuity results are derived from the fused network, the absence of these quantities leaves open whether the reported hierarchy is driven by the editorship layer or by default algorithmic behavior and relative graph densities.

Authors: We acknowledge that the manuscript does not explicitly document the SNF parameters or conduct ablation studies. The implementation follows the standard SNF procedure from the cited reference with equal layer weights and default kernel settings, but these details and robustness checks are not reported. In the revised manuscript we will add a dedicated subsection in Methods that specifies all SNF hyperparameters (including layer weights, number of iterations, and similarity thresholds), reports the resulting fused-network statistics, and includes leave-one-layer-out comparisons that quantify the contribution of the editorship layer relative to the others. revision: yes
Referee: Results section on hierarchy and legitimation: the assertion that 'editorial networks play the dominant role' is not supported by an explicit decomposition (e.g., layer-specific centrality, modularity contribution, or leave-one-layer-out comparisons). Without such evidence the dominance conclusion cannot be distinguished from an artifact of the fusion procedure or data density.

Authors: We agree that an explicit layer-wise decomposition would strengthen the claim. While the current results show that the fused communities and hierarchy metrics align most closely with the editorship layer, we did not present separate centrality scores or modularity contributions per layer. In the revision we will add these analyses to the Results section, including layer-specific centrality rankings, the change in modularity when the editorship layer is omitted from fusion, and a direct comparison of community stability across single-layer and fused networks. These additions will provide quantitative evidence that the editorial layer exerts the strongest influence. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical multilayer construction and SNF integration remain independent of headline claims

full rationale

The paper builds four empirical layers (interlocking editorship, interlocking authorship, bibliographic coupling, textual similarity) directly from journal metadata and article content for fixed years, fuses them via the standard Similarity Network Fusion procedure, extracts communities, and compares structural metrics across 2006/2012/2019 slices. The dominance attribution for the editorial layer is presented as an observational outcome of the resulting fused network's community structure and hierarchy measures rather than a quantity fitted to or defined by that outcome. No equations reduce the reported result to its own inputs by construction, no self-citation supplies a uniqueness theorem that forces the conclusion, and no ansatz is smuggled in. The derivation chain is therefore self-contained as a descriptive network analysis.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The analysis rests on standard network-science modeling choices and the assumption that the selected layers capture the relevant relationships; no new physical entities are postulated and only a modest number of tunable parameters are expected in the fusion step.

free parameters (1)

Layer weights and similarity thresholds in Similarity Network Fusion
These control how the four layers are combined and are typically chosen or optimized on the data.

axioms (1)

domain assumption Journals can be represented as nodes whose relationships are adequately captured by shared editors, shared authors, shared references, and textual similarity.
This modeling premise underpins the entire multiplex construction and is stated in the abstract's description of the four-layer network.

pith-pipeline@v0.9.0 · 5811 in / 1347 out tokens · 49399 ms · 2026-05-18T23:25:43.690078+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

four-layer multiplex network: interlocking editorship, interlocking authorship, bibliographic coupling, textual similarity ... integrated using Similarity Network Fusion
IndisputableMonolith/Foundation/AbsoluteFloorClosure.lean absolute_floor_iff_bare_distinguishability unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Partial Distance Correlation (PDC) ... contribution of each layer to the fused network

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.