Identifying the Hierarchical Influence Structure Behind Smart Sanctions Using Network Analysis

Hiroshi Iyetomi; Ryohei Hisano; Takayuki Mizuno

REVIEW 4 major objections 6 minor 13 references

Identifying the Hierarchical Influence Structure Behind Smart Sanctions Using Network Analysis

T0 review · 4 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that applying Helmholtz–Hodge decomposition to a network built from the order in which institutions add the same targets to smart sanctions lists exposes the hierarchical influence structure behind those lists, placing…

desk verdict First HHD application to a new sanctions-list influence network; interesting but the temporal-precedence-to-influence assumption is unvalidated and the data are not yet available. read the letter →

arxiv 1909.00847 v4 pith:B2C75OL4 submitted 2019-09-02 cs.SI

Ryohei Hisano , Hiroshi Iyetomi , Takayuki Mizuno This is my paper · ORCID

classification cs.SI

keywords smartsanctionsHelmholtz-Hodgedecompositioninfluencenetworkanalysislistshierarchicalstructuretemporalprecedenceinternationalpolitics

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

Smart sanctions lists—who is barred from doing business with whom—are issued by many countries and organizations, and the paper's idea is that the order in which institutions add the same named entity to their lists carries information about who is influencing whom. From 1,700 lists issued by 85 institutions, the authors construct a directed network in which an edge points from the institution that listed an entity first to the one that listed it later, then apply Helmholtz–Hodge decomposition to extract a scalar potential for each institution. The paper argues that this potential is a meaningful ranking of upstream initiators versus downstream followers, and that looking within issue-specific categories produces interpretable hierarchies—for example, the UN Security Council sits near the bottom for Iran–North Korea sanctions while some national and intergovernmental bodies sit near the top. The payoff, if the claim holds, is a data-driven window into the hidden political power structure behind international sanctions, independent of confidential diplomacy.

What carries the argument

The load-bearing object is the Helmholtz–Hodge decomposition of a weighted directed network: it writes the observed flow $F_{ij}$ along each edge as a gradient flow $w_{ij}(\phi_i-\phi_j)$ plus a circular loop flow, choosing the node potentials $\phi_i$ to minimize the squared difference, so $\phi_i$ becomes the hierarchy score of institution $i$. The input network is the paper's own construction: for each pair of institutions, whenever one institution lists an entity that another institution listed earlier, one unit of weight is added to the directed edge from the earlier to the later lister, and same-day listings are discarded because the direction is unclear. Modularity-based community detection then defines the issue categories (financial crimes, Libya, Africa, Burma, terrorism in general, Al-Qaeda, Iran–North Korea) used to build category-specific networks, and the gradient/loop ratio summarizes how much of the flow is hierarchical rather than cyclic.

What would settle it

Shuffle the listing dates within each sanctions category while preserving the content of each institution's lists, recompute the Helmholtz–Hodge potentials on many shuffles, and compare the resulting rankings with the reported ones; if the upstream/downstream ordering survives random date reassignment, then list timing is not what the hierarchy measures and the paper's interpretation would fail that test.

Watch

Extended reading notes

Core claim

The central claim is that a Helmholtz–Hodge potential computed from the who-lists-first network recovers a real hierarchy of influence among sanctions issuers, not just list similarity or conventional centrality. In the network built from all 1,700 lists, the OECD and the International Criminal Tribunal for Rwanda receive the highest potentials, which the authors read as meaning these institutions are least influenced by others, while the G7 countries sit in the middle of the ranking. The category-specific networks shift the ranking in ways the paper argues are interpretable: Switzerland has positive potential in nearly every category, Australia is mostly a follower, the United States ranks high against Al-Qaeda, Libya, Africa, and terrorism in general but low for financial crimes, and the UN Security Council sits at the bottom for Iran–North Korea sanctions, a position the authors tie to that body's complex approval mechanism. The paper presents the gradient-to-loop ratio as a companion measure of how hierarchical versus reciprocally circular each subnetwork is.

Load-bearing premise

The entire ranking rests on treating 'listed the same entity earlier' as 'influenced the later lister'; if institutions act independently, lag for bureaucratic reasons, or copy a shared third source, the edges—and the potentials built from them—misrepresent the influence structure.

Editorial extensions

If this is right

If the potentials are meaningful, the ordering of sanctions issuers from upstream initiators to downstream followers can be estimated from public list data alone, without access to internal decision records.
Category-specific decompositions separate issue areas: the Iran–North Korea network puts the UN Security Council at the bottom, which the paper attributes to that council's slow approval mechanism, while the Al-Qaeda network shows the UN and the EU acting as information-gathering hubs.
The gradient/loop ratio quantifies whether a sanctions domain is driven by hierarchical diffusion or by reciprocal influence; terrorism-in-general lists are almost purely hierarchical (loop ratio about 0.02), while Al-Qaeda lists carry the most loop structure (about 0.21).
Because the estimated potentials are not redundant with PageRank, the decomposition adds a dimension to standard network-centrality tools for political analysis.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The temporal-precedence assumption is testable: if a few documented cases show a 'later' lister acting on independent intelligence before the earlier list appeared, or both copying a third source, those edges are miscoded and the magnitude of the resulting ranking error could be estimated.
Since the lists carry dates, the same decomposition could be run in a sliding window to trace how influence hierarchies shift after major events such as new UN resolutions or terrorist attacks; the paper only reports static rankings.
The who-acted-first construction should transfer to other domains where organizations publish denial or warning lists over time—travel advisories, product-safety recalls, cybersecurity threat lists—where the resulting network could expose de facto standard-setters.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Colleague,

Quick take: this paper applies Helmholtz-Hodge decomposition to a newly constructed influence network of smart-sanctions issuers. The math is standard and correctly applied; the result is a new empirical ranking of institutions (OECD and ICTR at the top, UN Security Council near the bottom for Iran-North Korea). If you work on computational political science or sanctions, it's worth a look.

What's genuinely new is the dataset construction: 1,700 lists from Dow Jones, aggregated into directed edges by temporal precedence of listing the same entity. That is a legitimate new object of study. The HHD itself is not new (Jiang et al. 2011), but applying it here is reasonable, and the authors are careful to separate gradient from loop flow.

The soft spot is load-bearing: the assertion that if B lists an entity after A, then A influences B. The paper's own introduction lists alternative explanations—independent intelligence, shared third-party sources, or copying to project cooperation. None of those are ruled out. There are no null models, no significance tests, no external validation. The loop ratios in Table 2 are non-negligible (0.11 to 0.21), which is consistent with non-hierarchical mechanisms. And the data are proprietary; footnote 4 says the authors are 'recollecting the data for open use,' so as published the network cannot be independently reproduced.

The post-hoc exclusion of categories A and E in Section 4.2 is minor but should be better justified. The same-date exclusion rule deserves a sensitivity analysis.

None of this kills the paper. The derivation is self-contained, not circular, and the results are interesting enough to prompt follow-up. But the interpretation of the potentials as 'influence' needs external grounding or a serious robustness section.

Who is this for? Political scientists studying sanctions, and network analysts looking for an applied example. I would give it a serious referee, with the referee pushing for data release and validation.

Bottom line: worth engaging, but as an invitation to further work, not a settled result.

Referee Report

4 major / 6 minor

Summary. The paper analyzes 1,700 smart-sanctions lists from 85 institutions (73 countries and 12 international organizations) and constructs two directed networks: one at the level of individual sanctions lists and one at the level of institutions. An edge is placed from A to B when institution/list B includes an entity that A listed earlier; cases with identical dates are ignored. The paper then applies Helmholtz–Hodge decomposition (HHD) to these networks, extracting a gradient component with node potentials and a loop component. The authors report potentials for institutions overall and for seven sanction categories, and interpret the potential ordering as a hierarchy of influence: for example, OECD and ICTR are at the top in the overall analysis, Switzerland appears as a consistently influential player, and the UN Security Council is unusually low in the Iran–North Korea category. The main claim is that HHD of this network yields meaningful insights into the hierarchical influence structure behind smart sanctions.

Significance. If the temporal-precedence edges were a valid proxy for influence, this would be a novel and interesting application of HHD to international political data, with the potential to inform political science and policy analysis. The mathematical framework is standard, the decomposition is correctly described, and the paper is honest about data limitations (footnote 4). However, the central interpretation rests on an unvalidated assumption that listing an entity earlier causes later listing of the same entity by another institution. The paper does not provide null models, significance tests, external validation, or uncertainty quantification for the potentials, and the data are not currently available for independent replication. These gaps currently limit the strength of the causal claims and the generalizability of the results.

major comments (4)

[Section 2 (network construction)] The load-bearing assumption is that a directed influence edge from A to B exists whenever B first lists an entity after A first listed it. This equates temporal precedence with causal influence, but the Introduction itself acknowledges alternative mechanisms: institutions might act independently, follow a shared third-party source, or copy lists to project international cooperation. Under any of these alternatives, the edge set is not a faithful influence relation and the HHD potentials inherit the misspecification. To support the central claim, the authors should validate this edge definition against external evidence (e.g., case studies or official documents), test robustness against null models such as random date permutations or configuration models that preserve list overlap, and analyze sensitivity to the same-date exclusion rule. Without such checks, the interpretation of the potentials as influence is unsupported.
[Section 4.1, Table 3, and Table 2] The potentials and gradient/loop ratios are reported as point estimates with no uncertainty quantification or statistical significance tests. For example, the claim that the UN Security Council is at the bottom of the Iran–North Korea hierarchy (Fig. 8b and Section 5) could be within noise, especially given the small number of entities in that category. A bootstrap over listed entities or a permutation test that randomly shuffles list dates would provide confidence intervals for the potentials and p-values for the observed ordering. Similarly, the gradient/loop ratios in Table 2 have no error bars; the non-negligible loop ratios (0.11–0.21) are consistent with non-hierarchical mechanisms such as shared external drivers, which undermines the interpretation of the gradient component alone as representing true influence.
[Section 4.2] The analysis excludes two of the nine detected categories (A: Japanese bureaucracy, and E: embargoes) because 'only a few countries were involved in the country-level network.' This post hoc selection is not accompanied by a principled criterion or threshold, and it weakens the generality of the claim that the analysis covers smart sanctions broadly. The authors should either include all categories with appropriate caveats, or provide a prespecified exclusion rule and show that the main findings are robust to alternative inclusion criteria. The current presentation leaves the impression that categories were selected to produce more interpretable results.
[Footnote 4] The data are proprietary Dow Jones datasets, and footnote 4 states that the authors are 'recollecting the data for open use.' As published, the core network construction cannot be independently reproduced or tested, which is especially problematic because the edge definition and the resulting HHD potentials are the entire basis of the paper's claims. The authors should make the data available in some accessible form (e.g., a curated list of institution-entity-date triples for the analyzed lists) or provide a detailed replication script and summary statistics so that reviewers and readers can check the construction and run the suggested null-model analyses.

minor comments (6)

[Section 2, Fig. 1] In the text, 'Table 1b summarizes the characteristics of each community' appears to refer to a panel in Fig. 1; the reference should be clarified.
[Fig. 2] The label 'Word Bank' is a typo and should read 'World Bank.'
[Section 4.1] The sentence 'The arrow shows the location of where the y-axis being 0' is ungrammatical and should be revised, e.g., to 'The arrow indicates where the y-axis value is zero.'
[Section 5] The phrase 'we could shed some light' should be 'we can shed some light' for consistency with the rest of the paper.
[Fig. 4a] The scatterplot comparing PageRank and HHD potential would benefit from reporting the Pearson or Spearman correlation coefficient and a regression line to make the claim of independence quantitative.
[Section 3] The HHD presentation would be clearer if the authors stated the exact least-squares objective and the normalization convention for the potentials (e.g., mean zero), and explained how disconnected nodes are treated in the decomposition.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the Helmholtz-Hodge potentials are computed from the constructed network via a parameter-free least-squares decomposition, and the one self-cited method reference is not load-bearing.

full rationale

The paper's derivation chain is self-contained: the influence network is built from temporal precedence in the sanctions datasets, and the Helmholtz-Hodge potentials are then uniquely determined by minimizing the squared difference between the observed flow and the gradient flow, with no free parameters fitted to the conclusions. The narrative interpretations of which institutions are 'key players' are post-hoc readings of the computed potentials, not inputs to the computation. The temporal-precedence-to-influence edge construction is an externally testable assumption about what the edges mean, but assuming a causal interpretation does not make the mathematical output circular; the potentials would be the same even if the interpretation were wrong. The only self-citation, reference [7] by one of the authors, is used to point to the Helmholtz-Hodge decomposition method, which is also supported by the standard reference [6] and is described in the paper; it does not import the paper's empirical conclusions or forbid alternative methods. No equation or fitted parameter is, by construction, equal to the output claim. The absence of external validation or robustness checks is a correctness-risk concern rather than a circularity concern. Therefore no circular step is identified.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No free parameters are fit in the analysis; the HHD potentials are the solution of a least-squares problem. The main burden is the modeling assumption that temporal co-listing of an entity implies influence, plus post-hoc category choices. No invented entities are introduced.

assumptions (5)

domain assumption An entity appearing in list B after appearing in list A implies a directed influence A -> B.
This is the foundational modeling assumption of the influence network (Section 2). It equates temporal precedence with causal influence, which may overcount independent targeting or shared third-party sources.
domain assumption The number of overlapping entities between two institutions' lists is a meaningful edge weight.
The network uses an unweighted edge (1) for any shared entity, ignoring magnitude or entity-specific significance (Section 2).
domain assumption Helmholtz-Hodge decomposition potentials capture the true hierarchical ordering of influence.
Interpretation of the gradient component as 'influence' is assumed throughout Sections 4-5; the paper provides no external validation.
domain assumption Community detection (Louvain) yields meaningful categories of sanctions lists.
The nine communities found by modularity are treated as substantive categories 'financial crimes', 'Libya', etc., without validation (Section 2, Fig. 1).
standard math HHD least-squares minimization has a unique solution up to additive constant.
Standard linear algebra; the paper cites Jiang et al. (2011).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Identifying the Hierarchical Influence Structure Behind Smart Sanctions Using Network Analysis." pith.science (2026). https://pith.science/paper/B2C75OL4

@misc{pith2026190900847,
  author       = {Pith},
  title        = {Pith review of: Identifying the Hierarchical Influence Structure Behind Smart Sanctions Using Network Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B2C75OL4}},
  note         = {Machine review of arXiv:1909.00847}
}

read the original abstract

Smart sanctions are an increasingly popular tool in foreign policy. Countries and international institutions worldwide issue such lists to sanction targeted entities through financial asset freezing, embargoes, and travel restrictions. The relationships between the issuer and the targeted entities in such lists reflect what kind of entities the issuer intends to be against. Thus, analyzing the similarities of sets of targeted entities created by several issuers might pave the way toward understanding the foreign political power structure that influences institutions to take similar actions. In the current paper, by analyzing the smart sanctions lists issued by major countries and international institutions worldwide (a total of 73 countries, 12 international organizations, and 1,700 lists), we identify the hierarchical structure of influence among these institutions that encourages them to take such actions. The Helmholtz--Hodge decomposition is a method that decomposes network flow into a hierarchical gradient component and a loop component and is especially suited for this task. Hence, by performing a Helmholtz--Hodge decomposition of the influence network of these institutions, as constructed from the smart sanctions lists they have issued, we show that meaningful insights about the hierarchical influence structure behind smart sanctions can be obtained.

Figures

Figures reproduced from arXiv: 1909.00847 by the authors.

**Figure 2.** Influence network at institution level using all smart sanctions lists de [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Illustration of gradient and loop ratios. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Figures from the paper (4 more)

**Figure 5.** Figure 5: Influence network at institution level using the smart sanctions lists [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

**Figure 6.** Figure 6: Analysis of the community structure of the influence network at the level [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

**Figure 7.** Figure 7: Analysis of community structure of the influence network at the level of [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

**Figure 8.** Figure 8: Analysis of community structure of the influence network at the level of [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

13 extracted references · 12 canonical work pages

[1]

Ahn and Rodney D

Daniel P. Ahn and Rodney D. Ludema. The sword and the shield: the economics of targeted sanctions. Technical report, 2019

work page 2019
[2]

Fast unfolding of communities in large networks

Vincent D Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Etienne Lefeb- vre. Fast unfolding of communities in large networks. Journal of Statistical Me- chanics: Theory and Experiment , 2008(10):P10008, 2008

work page 2008
[3]

Brin and L

S. Brin and L. Page. The anatomy of a large-scale hypertextual web search engine. In Seventh International World-Wide Web Conference (WWW 1998) , 1998

work page 1998
[4]

The ruble between the hammer and the anvil: Oil prices and economic sanctions

Christian Dreger, Jarko Fidrmuc, Konstantin Kholodilin, and Dirk Ulbricht. The ruble between the hammer and the anvil: Oil prices and economic sanctions. Dis- cussion Papers of DIW Berlin 1488, DIW Berlin, German Institute for Economic Research, 2015

work page 2015
[5]

Kitacyosen Kaku no Shikingen Kokuren Sousa no Hiroku [Funding Source of North Korea: A Note on United Nation ’s Investigation] Funding source

Katsuhisa Furukawa. Kitacyosen Kaku no Shikingen Kokuren Sousa no Hiroku [Funding Source of North Korea: A Note on United Nation ’s Investigation] Funding source. Tokyo Shincyosya, Tokyo, Japan, 2017

work page 2017
[6]

Statistical ranking and combinatorial hodge theory

Xiaoye Jiang, Lek-Heng Lim, Yuan Yao, and Yinyu Ye. Statistical ranking and combinatorial hodge theory. Math. Program., 127(1):203–244, March 2011

work page 2011
[7]

Iyetomi H

Iino T. Iyetomi H. Inoue H. Kichikawa, Y. Hierarchical and circular ﬂow structure of the transaction network in japan. RIETI Discussion Paper Series , Aug 2019

work page 2019
[8]

Lambiotte, J

R. Lambiotte, J. C. Delvenne, and M. Barahona. Laplacian Dynamics and Multi- scale Modular Structure in Networks. arXiv e-prints , page arXiv:0812.1770, Dec 2008

arXiv 2008

Show all 13 references

[9]

R. Nephew. The Art of Sanctions: A View from the Field . Center on Global Energy Policy Series. Columbia University Press, 2017

2017
[10]

Uniﬁed quality measures for clusterings, layouts, and orderings of graphs, and their application as software design criteria, 2007

Andreas Noack. Uniﬁed quality measures for clusterings, layouts, and orderings of graphs, and their application as software design criteria, 2007

2007
[11]

M.B. Steger. Globalization: A Very Short Introduction . Very Short Introductions. OUP Oxford, 2017

2017
[12]

M. Stone. The Response of Russian Security Prices to Economic Sanctions: Policy Eﬀectiveness and Transmission. Technical report, 2019

2019
[13]

J. Zarate. Treasury’s War: The Unleashing of a New Era of Financial Warfare . PublicAﬀairs, 2013

2013

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Ahn and Rodney D

Daniel P. Ahn and Rodney D. Ludema. The sword and the shield: the economics of targeted sanctions. Technical report, 2019

work page 2019

[2] [2]

Fast unfolding of communities in large networks

Vincent D Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Etienne Lefeb- vre. Fast unfolding of communities in large networks. Journal of Statistical Me- chanics: Theory and Experiment , 2008(10):P10008, 2008

work page 2008

[3] [3]

Brin and L

S. Brin and L. Page. The anatomy of a large-scale hypertextual web search engine. In Seventh International World-Wide Web Conference (WWW 1998) , 1998

work page 1998

[4] [4]

The ruble between the hammer and the anvil: Oil prices and economic sanctions

Christian Dreger, Jarko Fidrmuc, Konstantin Kholodilin, and Dirk Ulbricht. The ruble between the hammer and the anvil: Oil prices and economic sanctions. Dis- cussion Papers of DIW Berlin 1488, DIW Berlin, German Institute for Economic Research, 2015

work page 2015

[5] [5]

Kitacyosen Kaku no Shikingen Kokuren Sousa no Hiroku [Funding Source of North Korea: A Note on United Nation ’s Investigation] Funding source

Katsuhisa Furukawa. Kitacyosen Kaku no Shikingen Kokuren Sousa no Hiroku [Funding Source of North Korea: A Note on United Nation ’s Investigation] Funding source. Tokyo Shincyosya, Tokyo, Japan, 2017

work page 2017

[6] [6]

Statistical ranking and combinatorial hodge theory

Xiaoye Jiang, Lek-Heng Lim, Yuan Yao, and Yinyu Ye. Statistical ranking and combinatorial hodge theory. Math. Program., 127(1):203–244, March 2011

work page 2011

[7] [7]

Iyetomi H

Iino T. Iyetomi H. Inoue H. Kichikawa, Y. Hierarchical and circular ﬂow structure of the transaction network in japan. RIETI Discussion Paper Series , Aug 2019

work page 2019

[8] [8]

Lambiotte, J

R. Lambiotte, J. C. Delvenne, and M. Barahona. Laplacian Dynamics and Multi- scale Modular Structure in Networks. arXiv e-prints , page arXiv:0812.1770, Dec 2008

arXiv 2008