Pith. sign in

REVIEW 3 major objections 5 minor 20 references

Trade power is who can substitute when a link is cut, not who runs the surplus—and the world network is steeply lopsided.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.5

2026-07-14 01:12 UTC pith:5ERNAVSM

load-bearing objection A serious, usable short-run measurement of Hirschman power on the full 2022 world IO network; the position-vs-deficit claim holds under the stated closures. the 3 major comments →

arxiv 2607.09990 v1 pith:5ERNAVSM submitted 2026-07-10 econ.GN cs.SIq-fin.EC

Political Power in International Trade

Ashwin Bhattathiripad , Vipin P Veetil This is my paper
classification econ.GN cs.SIq-fin.EC
keywords National PowerWorld TradeSanctionsProduction NetworkHirschman matrixinput-outputtrade dependencecore-periphery
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper measures economic power in trade as the capacity of one country to impose loss on another by withdrawing from a relationship. It builds a short-run model on the world input–output network: each restriction is a pattern of barred links; both buyers and sellers reallocate over partners already in the network under a hard ceiling that no producer expands beyond its pre-shock scale; a CES nest turns the resulting bundle mismatch into output loss. Across 9,480 counterfactual severances on the 2022 table, mutual dependence is anything but mutual. The average bilateral asymmetry is about 0.6 on a zero-to-one scale; the United States holds the favourable side of every relationship and China of all but one; a cut with Russia would cost Belarus more than a tenth of activity and Russia only half a percent. That asymmetry tracks core–periphery position in the network far more than bilateral trade imbalance. A sympathetic reader cares because the numbers reframe leverage: deficits and bilateral balances are weak guides, while network position decides who can hurt whom by walking away.

Core claim

Across 9,480 counterfactual severances on the 2022 world input–output network, mutual trade dependence is highly asymmetric (average bilateral asymmetry about 0.59 on a scale from balance at zero to complete lopsidedness at one). The United States holds the favourable side in all its relationships and China in all but one. The same hierarchy runs far down the system. The asymmetry correlates only weakly with bilateral trade imbalance but closely with core–periphery position computed from the benchmark network alone. Power is therefore a property of network position, not of deficits.

What carries the argument

The Hirschman matrix of bilateral power: for each country pair, both sides’ own-size-normalized activity losses under the same bilateral severance, summarized as a signed asymmetry on [- 1, 1]. Losses come from a short-run model that attenuates barred entries in the world input-share matrix, reallocates both buyer and seller margins together by RAS (biproportional) balancing under pre-shock capacity ceilings, then maps realized bundles into output via complementary CES nests and a fulfillment fixed point.

Load-bearing premise

The short-run rule that no country-sector can expand beyond its pre-shock scale, with prices fixed and no new trading links created, so every loss is nonnegative and substitution is limited to partners already in the network.

What would settle it

Re-solve the same 9,480 severances after relaxing the capacity ceiling or allowing new links and price adjustment within the horizon; if the average asymmetry falls sharply, signs flip for major hubs, or alignment with core–periphery position collapses while alignment with bilateral deficits rises, the central claim fails.

Watch this falsifier — get emailed when new claim-graph text bears on it.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper defines economic power in trade as the capacity of one country to impose loss on another by withdrawing from a relationship, and measures it on the 2022 OECD world input-output network. A short-run model represents each restriction as attenuated barred entries in the input-share matrix; both sides reallocate over retained benchmark links via biproportional (RAS) balancing under capacity ceilings that no country-sector exceeds pre-shock scale; dual CES aggregators map composition and scale gaps into output; and a benchmark-selected fulfillment fixed point yields country-sector losses. Bilateral asymmetries form a Hirschman matrix H and index Ψ(H). Across 9,480 counterfactual severances, average absolute asymmetry is about 0.59; the United States is favourable in all 79 relationships and China in all but one; and H correlates only weakly with bilateral trade imbalance (~0.19) but strongly with core-periphery position (Spearman ~0.86). Two checks are offered: a calibrated reconstruction of 2022 Russian energy redirection, and a descriptive comparison of H with realized coercion dyads.

Significance. If the short-run closures are accepted as the relevant measurement frame, the paper delivers a concrete, network-based quantification of Hirschman's influence effect that the literature has long lacked. The two-sided RAS reallocation, monotone least-loss selection (Propositions 2–5), and full set of 9,480 equilibria with tight residuals and diagnostics are genuine technical contributions; the parameter sweep showing that signs, rankings, and position alignment are stable while magnitudes scale is a useful robustness check. The central empirical claim—that power tracks network position rather than bilateral deficits—is falsifiable against the same table and would discipline both geoeconomics theory and sanctions design. The work sits cleanly at the intersection of production networks, quantitative sanctions, and political economy of interdependence.

major comments (3)
  1. [Sections 2.2–2.4, Lemma 4, Section 6] The short-run capacity ceiling (no country-sector exceeds pre-shock scale), fixed prices, and no-new-links RAS reallocation (targets (3)–(4), Lemma 4, activity box [0,s], Sections 2.2–2.4) are load-bearing for nonnegative losses and for the measured substitution gaps that define H. The paper correctly flags this in Section 6, but the central claim that "power is a property of network position, not deficits" is stated as a general finding. The manuscript should either (i) reframe the claim explicitly as short-run under these closures, or (ii) provide at least one alternative closure (partial capacity expansion, limited new links, or a simple price adjustment) showing that the weak imbalance / strong coreness ranking survives. Without that, the ranking is a well-specified computational result under maintained assumptions rather than a robust structural fact.
  2. [Section 5.1, Table 8, Appendix D] Section 5.1 uses the 2022 Russia sanctions both to help discipline τ (and inform δ) and as the main redirection check. The paper is transparent that this is a calibrated contemporaneous reconstruction, not independent out-of-sample validation, and that the closer fit at τ≈0.10 is calibration-consistent. Even so, the abstract and introduction still lean on the episode as empirical discipline for the mechanism. Either move the episode fully into calibration (and drop the "validation" framing) or add a leave-one-episode-out / pre-war benchmark exercise so that the two-sided rule is tested on held-out reallocation patterns.
  3. [Definitions 6–7, Proposition 3(iii), Appendix F] The outer CES nest (Definition 6, ϱ) and the fulfillment rule φ=λ (Definition 7) are reduced-form finite-shock closures, not derived from a CES technology with fixed primary factors. Remark 1 and Appendix E correctly show degeneracy at unit pass-through, and the cushion condition (24) is carefully stated. For the power ranking, however, the paper should report how often the cushion fails (the text says 18 of 9,480) and whether any of those failures involve the headline pairs that drive Tables 1, 3, and 5. If the ranking is driven only by interior, cushion-satisfying equilibria, that should be stated; if not, the selection and positivity claims need a short robustness note for those 18 cases.
minor comments (5)
  1. [Abstract, Section 1, Table 4] Abstract and introduction round the average asymmetry to 0.6; Table 4 reports Ψ(H)=0.589. Use a consistent figure (e.g., 0.59) throughout.
  2. [Figure 1] Figure 1 is log-scale over 6,320 bilateral vulnerabilities; a short note on how many pairs sit near machine zero would help the reader interpret the left tail.
  3. [Section 5.2, Table 9] The coercion comparison (Section 5.2, Table 9) is descriptive and correctly caveated; still, reporting the share of favourable-side initiations at the episode-cluster level (already mentioned in text) in the table itself would make the "six of eight" claim easier to check.
  4. [Section 3.3] Notation for the Hirschman objects switches between H_cc', h_cc', and Ψ(H); a single display of the three definitions early in Section 3.3 would reduce friction.
  5. [Sections 3.4, 4.7, Appendix F] Appendix F is repeatedly deferred to for diagnostics and the full parameter sweep; a one-paragraph summary of the sweep's main stability result in the main text (near the end of Section 3.4 or 4.7) would help readers who do not open the appendix.

Circularity Check

1 steps flagged

Core Hirschman power matrix is an independent counterfactual computation on the IO network; the only material circularity is the Russia 2022 redirection check, which re-uses the same episode that disciplines τ (and informs δ) and which the paper itself labels non-independent.

specific steps
  1. fitted input called prediction [Section 5.1 (and Intro ¶ on empirical discipline; Appendix D)]
    "Evidence from the 2022 sanctions episode helps discipline the friction parameters τ and δ, and the 2022 annual table contains post-invasion flows from most of the calendar year. Conditional on those inputs, the exercise asks whether the two-sided balancing reproduces the realized reallocation... This is not an independent out-of-sample validation... Because this lower friction is read from the same episode, the closer level agreement is a calibration fit, not separate validation evidence."

    τ (and the scale of δ) is disciplined by the aggregate recovery/redirection of the same 2022 Russia sanctions episode that is then reconstructed. The overall scale of reconstituted trade is therefore not an independent prediction; only the partner-specific shares (conditional on the fitted frictions and the reference affinities) are generated by the RAS operator. The paper itself labels the exercise a calibrated contemporaneous reconstruction rather than external validation, so the ‘match’ on totals is partly by construction of the calibration.

full rationale

The central objects—bilateral losses under 9,480 severances, the Hirschman matrix H, Ψ(H)≈0.59, US/China hierarchy, and the weak correlation of H with bilateral imbalance (~0.19) versus strong alignment with core–periphery position (Spearman ~0.86)—are generated by applying a fixed short-run mechanism (RAS both-sides reallocation under capacity ceilings, CES bundle/output nests, benchmark-selected fixed point) to the 2022 OECD ICIO table. H is not defined as the trade deficit, not fitted to coercion outcomes, and not algebraically forced to equal any benchmark network score; the core-score gap is computed from the benchmark intensity matrix alone and then correlated with the counterfactual H. No uniqueness theorem or ansatz is imported from the authors’ prior work; references are external (Hirschman, Sinkhorn/RAS, Acemoglu–Baqaee–Farhi, etc.). The sole circular step is the Section 5.1 Russia reconstruction: evidence from the same 2022 sanctions episode helps set τ (and the scale of δ), the 2022 annual table already embeds post-invasion flows, and the model is then shown to reproduce the redirection signature (coalition collapse, non-coalition rise, China–India–Türkiye order). The paper repeatedly flags this as a calibrated contemporaneous reconstruction rather than out-of-sample validation. That is a genuine but limited fitted-input-called-prediction issue confined to the mechanism check; it does not render the main power claims circular. Parameter sweeps further show that signs, rankings, and position-versus-imbalance alignment are stable while magnitudes scale. Overall circularity is therefore modest and localized.

Axiom & Free-Parameter Ledger

4 free parameters · 6 axioms · 2 invented entities

The central claim rests on a short-run structural model with a handful of calibrated frictions and maintained closures (capacity ceiling, fixed prices, no new links, two-sided RAS, dual CES nests). The free parameters are few and partly disciplined by literature and the 2022 episode; the main load-bearing axioms are domain modeling choices rather than pure math. No new physical entities are invented; the Hirschman matrix is a constructed statistic.

free parameters (4)
  • rerouting friction τ = 0.30 (baseline); lower ≈0.10 read from Russia episode
    Fraction of displaced mass removed from row/column targets; baseline 0.30 chosen from episode evidence and mixed fungible/specific goods; losses scale with τ.
  • barred-link attenuation δ = 0.10
    Reference weight on barred links in the disrupted matrix before balancing; baseline 0.10; not a hard cap on realized retention.
  • inner CES curvature ρ = −1 (baseline)
    Complementarity of intermediate sector bundles; baseline −1 (elasticity 1/2); sensitivity reported.
  • outer CES curvature ϱ = −1 (baseline)
    Complementarity between intermediate composite and non-intermediate block; baseline −1; sweep to −9; unit pass-through pole excluded to avoid degeneracy.
axioms (6)
  • domain assumption Short-run capacity ceiling: no country-sector activity or seller sales exceed pre-shock benchmark scale.
    Imposed via capped RAS targets and activity box [0,s]; forces nonnegative losses and shapes substitution (Section 2.2–2.4).
  • domain assumption Prices fixed at pre-shock levels; no price block or terms-of-trade channel.
    Stated in Section 2.3; losses are pure quantity losses.
  • domain assumption No new trade links: substitution only over the benchmark support (positive attenuation keeps support).
    Assumption 1 and construction of A_ω; buyers/sellers lean only on partners already used.
  • ad hoc to paper Two-sided biproportional (RAS) balancing is the correct short-run mutual reallocation under capacity.
    Central modeling device; compared to one-sided Leontief/Ghosh-style rescalings but not derived from a full bargaining or search model.
  • domain assumption Intermediate inputs are strictly complementary (ρ, ϱ < 0); outer nest is a maintained reduced-form finite-shock map.
    Section 2.3; outer nest not structurally derived from a full CES technology with fixed primary factors.
  • ad hoc to paper Benchmark-selected (monotone from full fulfillment) equilibrium is the relevant selection for power measurement.
    Definition 8 / Proposition 3; reports least-loss equilibrium among possibly multiple fixed points.
invented entities (2)
  • Hirschman matrix H and asymmetry index Ψ(H) no independent evidence
    purpose: Scalar bilateral power measure from paired own-size-normalized losses under the same bilateral severance.
    Constructed statistic (eqs. 34–36); not an independent physical object; independent_evidence false because it is defined by the model.
  • Benchmark-selected equilibrium under two-sided RAS + dual CES fulfillment no independent evidence
    purpose: Maps each restriction ω into a unique reported loss vector for vulnerability and power.
    Model object combining classical RAS with a specific fulfillment fixed point; falsifiable only via external episodes, which the paper treats carefully as non-independent.

pith-pipeline@v1.1.0-grok45 · 59966 in / 3900 out tokens · 38900 ms · 2026-07-14T01:12:01.236366+00:00 · methodology

0 comments
read the original abstract

Economic power in international trade is the capacity of one country to impose loss on another by withdrawing from a trading relationship. This paper measures it. A model of the short run represents each trade restriction as a pattern of barred entries in the world matrix of input shares and maps it into a vector of losses by country and sector. The asymmetry between the two countries' losses under the same severance is the measure of power: a gap in substitution, since a buyer's dependence turns on how easily it finds another source and a seller's on how easily it finds another market. When a relationship is barred, buyers lean on alternative suppliers and barred suppliers on alternative buyers already present in the benchmark network, under the ceiling that no producer exceeds its pre-shock scale. The reallocation is a RAS balancing of the disrupted matrix that lets \emph{both sides} adjust together. Across 9,480 counterfactual severances on the 2022 world input--output network, mutual trade dependence is anything but mutual. The average bilateral asymmetry is 0.6 on a scale that runs from balance at zero to complete lopsidedness at one. The United States holds the favorable side in all its relationships, China in all but one. The same tilt runs far down the hierarchy: a severance with Russia would cost Belarus more than a tenth of its economic activity, but Russia only half of one percent. The asymmetry bears only a weak relation to bilateral trade imbalance but closely tracks whether a country sits at the core or the periphery. Power is a property of network position, not deficits.

Figures

Figures reproduced from arXiv: 2607.09990 by Ashwin Bhattathiripad, Vipin P Veetil.

Figure 1
Figure 1. Figure 1: The distribution of bilateral vulnerabilities [PITH_FULL_IMAGE:figures/full_fig_p037_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Bilateral vulnerability Gcc′ among the thirty largest economies: the loss of the row country, in percent of its activity, when trade with the column country is severed bilaterally. Baseline calibration: τ = 0.30, ρ = −1, ϱ = −1, δ = 0.10. Where do the losses live within an economy? [PITH_FULL_IMAGE:figures/full_fig_p038_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Sector composition of the loss under the USA–China bilateral severance, both sides, [PITH_FULL_IMAGE:figures/full_fig_p039_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Trade is recomposed under a severance, pooled across a random sample of forty bilateral [PITH_FULL_IMAGE:figures/full_fig_p040_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The Hirschman matrix Hcc′ among the thirty largest economies (eq. 35). Red cells: the row country is the more vulnerable side of the bilateral relationship; blue cells: the row country holds the advantage. Baseline calibration: τ = 0.30, ρ = −1, ϱ = −1, δ = 0.10. Ψ (H) Ψwithin Ψbetween pairs in E 0.589 0.588 0.590 3160 [PITH_FULL_IMAGE:figures/full_fig_p049_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: The autonomous-demand share γ ζ /γ of each economy’s mean bilateral loss (Lemma 5) against benchmark size. Baseline calibration: τ = 0.30, ρ = −1, ϱ = −1, δ = 0.10. We come to the comparison the section was built for. Is the power in H simply benchmark structure in counterfactual clothing, and if so, which structure? [PITH_FULL_IMAGE:figures/full_fig_p051_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Hirschman asymmetry against direct trade-flow asymmetry, all pairs in [PITH_FULL_IMAGE:figures/full_fig_p053_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The calibrated 2022 redirection check at a glance. Left: coalition share of Russian energy [PITH_FULL_IMAGE:figures/full_fig_p057_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Sensitivity of headline bilateral vulnerabilities to each calibrated parameter, the others [PITH_FULL_IMAGE:figures/full_fig_p102_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Rank-size (top row) and survival (bottom row) plots for the four power measures of [PITH_FULL_IMAGE:figures/full_fig_p110_10.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

20 extracted references

  1. [1]

    The Network Origins of Aggregate Fluctuations

    Acemoglu, Daron, Vasco M. Carvalho, Asuman Ozdaglar, and Alireza Tahbaz-Salehi.2012. “The Network Origins of Aggregate Fluctuations.”Econometrica, 80(5): 1977–2016. Atkeson, Andrew, and Patrick J. Kehoe.1999. “Models of Energy Use: Putty-Putty versus Putty- Clay.”American Economic Review, 89(4): 1028–1043. Bacharach, Michael.1970.Biproportional Matrices a...

    2012

  2. [2]

    The Macroeconomic Impact of Microeconomic Shocks: Beyond Hulten’s Theorem

    Baldwin, David A.1985.Economic Statecraft.Princeton University Press. Baqaee, David Rezza, and Emmanuel Farhi.2019. “The Macroeconomic Impact of Microeconomic Shocks: Beyond Hulten’s Theorem.”Econometrica, 87(4): 1155–1203. Barrot, Jean-No¨el, and Julien Sauvagnat.2016. “Input Specificity and the Propagation of Idiosyn- cratic Shocks in Production Network...

    1985

  3. [3]

    Power-Law Distributions in Empirical Data

    Clauset, Aaron, Cosma Rohilla Shalizi, and M. E. J. Newman.2009. “Power-Law Distributions in Empirical Data.”SIAM Review, 51(4): 661–703. Clayton, Christopher, Matteo Maggiori, and Jesse Schreger.2026. “A Framework for Geoeco- nomics.”Econometrica, 94(1): 105–136. Congressional Research Service.2025. “Iran’s Petroleum Exports to China and U.S. Sanctions.”...

    2009

  4. [4]

    A Lattice-Theoretical Fixpoint Theorem and Its Applications

    Tarski, Alfred.1955. “A Lattice-Theoretical Fixpoint Theorem and Its Applications.”Pacific Journal of Mathematics, 5(2): 285–309. Thompson, Anthony C.1963. “On Certain Contraction Mappings in a Partially Ordered Vector Space.”Proceedings of the American Mathematical Society, 14(3): 438–443. U.S. Energy Information Administration.2025. “Russia’s Oil Export...

    1955

  5. [5]

    The corresponding theoretical statement applies whenever a maintained parameter rectangle lies inside T ω∈Ω Pω

    Both are needed for the full Φω iteration to be well defined and geometrically convergent from every starting point in Dω. The corresponding theoretical statement applies whenever a maintained parameter rectangle lies inside T ω∈Ω Pω. On the 2022 table the computed intersection is empty under this conservative ℓ1 bound, so no reported scenario is certifie...

    2022

  6. [6]

    Condition (iii) probes a wider basin

    Conditions (i), (ii), and (iv) document terminal decay, residual accuracy, and positivity of the reported benchmark path. Condition (iii) probes a wider basin. The ratio in (i) is an observed terminal rate, not a Jacobian spectral radius or a contraction modulus. None of these finite numerical checks is a proof of contraction on a neighbourhood or of glob...

    2022

  7. [7]

    Contraction on a neighbourhood is not claimed, and what multiplicity survives sits on the starvation boundary. Proposition 4(Uniqueness and convergence on Pω).For every( τ, ρ, ϱ) ∈ Pω, the map Φω under the restriction is a contraction in a weighted norm, equivalent to the product ℓ1 norm, on the maintained domain Dω(hω), and sends Dω into itself. The fixe...

    2022

  8. [8]

    The fixed points of the monotone φ(ω) on [0,1]N form a complete lattice (Tarski, 1955), of which Proposition 3 exhibits the greatest element

    and the strict subhomogeneity φ(ω) i (f h) > f φ (ω) i (h), f∈ (0,1), inherited from the degree-one homogeneity of κ(ω) i through the outer aggregator (Lemma 3(iii)), already pin the equilibrium down everywhere except at exact starvation. The fixed points of the monotone φ(ω) on [0,1]N form a complete lattice (Tarski, 1955), of which Proposition 3 exhibit...

    1955

  9. [9]

    Certified rates on Pω, and the parameter continuity built on them (Lemma 6), therefore remain the province of the contraction analysis

    are a local property of the Jacobian along the computed paths, which the envelope (61) does not certify. Certified rates on Pω, and the parameter continuity built on them (Lemma 6), therefore remain the province of the contraction analysis. What no longer depends on Pω is which equilibrium the model means. Basin check (iii) of Definition 13, randomized st...

    2022

  10. [10]

    This contradicts the definition of t∗

    Monotonicity and the strict inequality of (65) give, at every coordinatei, hi =φ (ω) i (h)≥φ (ω) i (t∗g)> t ∗ φ(ω) i (g) =t ∗ gi, so, the index set being finite, mini hi/g i > t ∗. This contradicts the definition of t∗. Hence t∗ ≥ 1, that is,g ≤ h. Exchanging the roles ofgandhgivesh ≤ g, sog=h. This is the cone argument of Krasnosel’skii (1964) for monoto...

    1964

  11. [11]

    This is (61)

    The upper side is where the cone extension is used, since T hω may leave [0,1]N . This is (61). Because ¯e∗ < 1 and t0, T0 are positive and finite, t ¯e u ∗ 0 → 1 and T ¯e u ∗ 0 → 1, and the envelope squeezesh (u) → hω. For the metric form, nonexpansiveness on the positive orthant is the weak inequality in (65) applied two-sidedly, the standard property o...

    1963

  12. [12]

    It originates in statistics with Deming and Stephan (1940), who fitted a contingency table to known marginals by alternately rescaling its rows and columns (later called “raking”)

    A word on provenance, since the device has been discovered repeatedly across fields. It originates in statistics with Deming and Stephan (1940), who fitted a contingency table to known marginals by alternately rescaling its rows and columns (later called “raking”). In matrix analysis, Sinkhorn (1964) and Sinkhorn and Knopp (1967) proved that a positive ma...

    1940

  13. [13]

    Absence of empty rows and columns alone is not sufficient

    At δ = 0 the severed entries leave the support, and existence and convergence require the resulting prescribed-margin transportation polytope to contain a point positive on every retained edge. Absence of empty rows and columns alone is not sufficient. 91 D Calibrating the rerouting frictionτfrom disruption episodes The model carries two free parameters b...

    2022

  14. [14]

    Malaysia

    and after the severance (t = 1). Recovery is measured against the no-disruption counterfactual intake eQrc, so as to net out demand destruction. The effective episode estimate of the rerouting friction is bτ= 1− bMrec (1− bδ)X 0 ,(79) with the seller-side estimate defined symmetrically. This is a calibration proxy for the target friction in (78), not an e...

    2023

  15. [15]

    A restricted equilibrium is exactly a fixed point of Γ , with activity given by the linear solution (72) (withK ω in place ofΛ ω) at that fulfillment

    Fix a restriction ω with severed mass τ(1 −δ ) > 0 and write Γ : [0 ,1]N → [0,1]N for the equilibrium fulfillment map, Γ (h) := φ(ω)(h) = κ(ω)(h). A restricted equilibrium is exactly a fixed point of Γ , with activity given by the linear solution (72) (withK ω in place ofΛ ω) at that fulfillment. Two facts, both already in the manuscript, drive everything...

    2022

  16. [16]

    A ρ-power mean of retentions at most one, one of them strictly below one with positive weight, is strictly below one, soκ(ω) j (1) < 1, a contradiction

    But with τ(1 −δ ) > 0 the intended retention of j falls strictly short of one in any used sector it sourced from the severed partner (Lemma 2). A ρ-power mean of retentions at most one, one of them strictly below one with positive weight, is strictly below one, soκ(ω) j (1) < 1, a contradiction. Hence m = 0,h ∗ =0,K ω(0) =0on the active set, and the activ...

    1978

  17. [17]

    The wipeout corner is therefore notempty in the data

    The wipeout corner in the data For the record behind the statement in Section 2.2, on the 2022 OECD ICIO table (80 economies, 50 industries), 1 ,821 of the roughly 1 .9 × 105 active buyer-sector cells are fully single-sourced (maxi∈Nr aij /α rj = 1), predominantly small domestic service cells. The wipeout corner is therefore notempty in the data. Under As...

    2022

  18. [18]

    Coercion episodes used in the empirical check The table below lists the 154 directed coercion dyads scored in Section 5.2

    0.000 0.997 1.000 1.000 measured contraction factor 0.032 0.594 0.659 0.733 fixed-point iterations 1 19 24 29 Sinkhorn sweeps 1 12 25 37 Table 11: Solver and stability diagnostics across all 9,480 baseline scenarios: the dissipation norm qω, the row-total bound ¯aω, the intended-retention floor, equilibrium fulfillment and bundle efficiency minima, the cu...

    2022

  19. [19]

    The distributions are heavy-tailed and Zipf-like

    Table 14 reports the fits and Figure 10 the rank-size and survival plots. The distributions are heavy-tailed and Zipf-like. Coreness has α≈ 2.3, within 1.3 standard errors of the Zipf value, and economic size α≈ 1.9, the textbook Zipf reading that serves as the sanity check. The dollar power to harm is somewhat steeper at α≈ 2.9, spread across a few more ...

    2022

  20. [20]

    ζω Autonomous-demand channel of the intermediate input decomposition (Lemma 5): ζω = (I −Λ ω(hω))z ≥ 0, the vector of autonomous-demand attenuations

    The second piece of the two-part loss decomposition (bundle penalty on intermediate sales). ζω Autonomous-demand channel of the intermediate input decomposition (Lemma 5): ζω = (I −Λ ω(hω))z ≥ 0, the vector of autonomous-demand attenuations. Nω Intermediate-input channel of the intermediate input decomposition:Nω = A(I−β)s−diag(hω)Aω(I−β)sω ≥ 0, the gap b...

    2022