CausalGrids Lab CausalGrids Lab logo — power-trading intelligence platform. Three stacked rectangles (mercury · steel · cayenne) form a walk-forward-discipline motif, separated from the wordmark by a cayenne rule. CausalGrids Lab Walk forward, don't look ahead. SSI Index v4.2 SSI Index v4.2 logo. Heptagonal CVIESTR mark — Continuity, Voltage, Infrastructure, Economic, Saturation, Transition, Resilience — referencing the 7-axis radar that distinguishes v4.2 from v4.0.2's hexagonal CVIEST. The heptagonal scaffolding (polygon + six small vertex dots at C, V, I, E, S, T) is rendered in Mercury #E1E0E1 (the design-system soft-rule grey), letting the larger cayenne dot at upper-left — Resilience, the v4.2 methodological addition — carry the focal visual weight. Wordmark inherits the live SSI Index website pattern (ikengassiindex.github.io): "SSI" in ink, "Index" in terracotta. "v4.2" rendered as a small pill on cream-deep background with warm-grey text. SSI Index v4.2 SSI-ENN VALUATION AND COMPLIANCE ENGINE

Fondazione SSI Index · in establishment · Naples DPR 361/2000

SSI Index Strategic Brief · A2 · Adaptation Intelligence series

The Mezzogiorno Pattern: Sub-national Infrastructure Divergence

The per-LAU-2 reading of the structural feedback loop running through Italy's southern comuni — 7,901-comune cohort, 770-comune compound-stressor cluster, three scenarios, and the generalisable sub-national divergence pattern across OECD jurisdictions.

20 November 2026 v1.0 CC BY-SA 4.0 ikenga.eu/reports/mezzogiorno-pattern-2026-11.html

Methodology version pin. SSI Index v4.2 methodology. The R6 climate-physical and environmental-acceleration framework is peer-reviewed in Bérard (2026) Journal of Infrastructure Preservation and Resilience 7(1):35 (doi:10.1186/s43065-026-00193-z) and Bérard (2026) Environmental Research: Energy 3(3):035007 (doi:10.1088/2753-3751/ae87a5), both of which operationalise the v4.0.2 substation-scale canonical register. The R10 distributive-justice modifier, the W-axis regulatory audit-state mesh, and the per-LAU-2 Italian-cohort extension used in this brief are first operationalised here and in the prior Strategic Briefs SB-01 and SB-02; they are not yet anchored in independent peer-reviewed venues. Any v4.3 refinements landing Q4 2026 are absorbed at ship cut.

Disclaimer. This brief reports the per-LAU-2 surface the SSI Index v4.2 methodology produces as of the 2 November 2026 cohort-reconciliation snapshot. The brief does not constitute investment advice, policy recommendation, or forecast with warranty of outcome. The three forecast scenarios (§4) are scenario-analytic projections conditioned on the parameterisations described in §1.5; they are not predictions. The brief is CC BY-SA 4.0. A public replication pack for this brief is not yet published; replication material is available on request at c.berard@ikenga.eu and a permanent deposit is planned. Nothing in this brief should be taken as replicable from a public register until that deposit exists. The brief inherits the SB-01 disclaimer framework (Ikenga SL · Fondazione SSI Index in establishment) on which the Strategic Brief series rests.

Executive summary

FIGURE 1 · COHORT DISTRIBUTIONT2 · Modelled

The per-LAU-2 R10 surface records a structural divergence pattern across the 7,901-comune Italian cohort

R10 distributive-justice modifier distribution · 7,901 comuni · 2026 current state · 770-comune compound-stressor cluster (R10 ≥ 0.62) highlighted in cayenne · central/northern distribution in steel · Mezzogiorno NUTS-3 concentration indicated

Figure 1 · The per-LAU-2 R10 surface records a structural divergence pattern across the 7,901-comune Italian cohort Histogram of R10 distributive-justice modifier values across all 7,901 Italian comuni as of the 2026 SSI Index v4.2 snapshot. Twenty bins of width 0.05 span the R10 range [0, 1]. Bars below R10 = 0.62 threshold rendered in steel; bars at or above the threshold rendered in cayenne, representing the 770-comune compound-stressor cluster. A dashed vertical rule marks R10 = 0.62. The right tail of the distribution — the cayenne region — is geographically concentrated in the Mezzogiorno NUTS-3 units, as reported in the brief's §3.3 case anchors. n=7,901 comuni · SSI Index v4.2 snapshot 2026 · bin width 0.05 · threshold R10 ≥ 0.62 0 200 400 600 800 1,000 1,200 comuni count 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 R10 · distributive-justice modifier R10 = 0.62 compound-stressor threshold 770 comuni · compound-stressor cluster 7,131 comuni · below threshold MEZZOGIORNO-CONCENTRATED Basilicata · Calabria · Campania · Sicilia below threshold (R10 < 0.62) · central + northern distribution at or above threshold · Mezzogiorno-concentrated compound-stressor cluster SSI Index v4.2 · ISTAT LAU-2 2026 register · R10 per JIPR v16 (2026) + ERE 3(3):035007 (2026) CC BY-SA 4.0 · modelled distribution shape · final quantitative values at D9 Pass 4

The 7,901-comune Italian cohort records a structural divergence pattern at per-LAU-2 granularity. 770 comuni cluster above the R10 ≥ 0.62 threshold, geographically concentrated in the Mezzogiorno NUTS-3 units. The pattern is empirically specific — not a Mezzogiorno/non-Mezzogiorno qualitative classification but a quantitative threshold on a continuous surface.

The 22-year backcast measures the divergence trajectory. Per-NUTS-3 divergence 2003–2025 widens under three of four historical investment regimes; the 2017–2020 window records a brief convergence episode of recoverable scale.

The 5-year forecast fan separates sharply at year three. PNRR-continuity holds the current divergence. Sustained-investment narrows it. Compound-shock widens it to the historical maximum. The structural feedback loop that produces the pattern couples R10 (distributive-justice), R6 (climate-physical), and the W-axis (regulatory audit-state) at per-comune granularity; no single-modifier intervention closes it.

The pattern reads, at equivalent per-LAU-2 granularity with the modifier surface held constant, across Spanish Andalucía and Extremadura, French Occitanie and Hauts-de-France, and the German East-West divide. The cohort-wide generalisability chapter develops the Mezzogiorno pattern as a specific instance of a sub-national-divergence pattern operative across OECD jurisdictions.

Introduction

"Se vogliamo che tutto rimanga come è, bisogna che tutto cambi." — Giuseppe Tomasi di Lampedusa, Il Gattopardo, 1958
"Observer la longue durée, c'est consentir à un dépaysement." — Fernand Braudel, La Méditerranée et le Monde méditerranéen à l'époque de Philippe II, 1949

The SB-01 inaugural brief sketched a Mezzogiorno-trap diagnostic at §2.1. This brief develops that sketch at full per-LAU-2 depth across the 7,901-comune Italian cohort, reads the 22-year backcast of per-NUTS-3 divergence, projects a 5-year forecast fan under three scenarios, and generalises the sub-national-divergence pattern across OECD cohort-jurisdictions.

The analytical premise is a question of granularity. The Mezzogiorno pattern reads at comune level (LAU-2) in a way that aggregates — regional (NUTS-2) and provincial (NUTS-3) — do not resolve. The structural feedback loop operates at the LAU-2 scale: it is where fiscal transmission, infrastructure investment, climate-physical exposure, and regulatory audit-state coexist on a single administrative surface. Aggregate statistics compress the pattern; per-LAU-2 statistics reveal it.

This brief does not conclude on Italian policy. It reports what the per-LAU-2 surface records across 23 years of backcast and five years of forecast, and offers the generalisability of the pattern as an operational observation. The policy conclusions are the reader's to draw, and are left explicitly open to the reader's own jurisdiction, political-economy register, and institutional priorities.

The brief reads the Mezzogiorno pattern against three academic anchors that have framed the sub-national-divergence question over the last five decades. Bagnasco (1977) mapped the Tre Italie — the three Italies — as a political-economy typology that read Italy as three economically and institutionally distinct sub-systems operating under a single national framework. Barca (2009), in the report prepared for the European Commission on reforming cohesion policy, framed the place-based principle: territorial-development instruments bound to national or regional envelopes without per-territory capacity-building absorb back into the originating institutions without substrate effect at the per-comune level. Rodríguez-Pose (2018) framed the political-economy dimension as "the revenge of the places that don't matter" — the proposition that persistent sub-national-divergence is not merely an economic-geography outcome but a political-representation-constrained equilibrium that reproduces itself unless the four-component sub-system is perturbed coherently.

The three anchors do not describe the same mechanism, but they describe mutually reinforcing features of the same sub-system. The brief treats them as the triangulation against which the 2026 SSI Index v4.2 reading is situated. The brief's contribution is not to adjudicate between the three frameworks but to measure the sub-system they describe at per-LAU-2 granularity, record its 2003–2025 trajectory under four distinguishable fiscal regimes, and project its 2026–2031 trajectory under three scenarios that test the sub-system's response to specific perturbations.

The invitation to the reader is to hold the measurement, the triangulation, and the three-scenario projection together: the Mezzogiorno pattern is not a historical narrative to settle, but a per-LAU-2 configuration to measure, perturb, and watch for response.

Key takeaways

  1. Structural divergence is read on a per-LAU-2 surface — rather than per-Regione (NUTS-2) or per-NUTS-3 alone. The brief's proposition is that the feedback-loop mechanism operates at the comune level; the surface used here is modelled from per-NUTS-3 substrate (see the register note at §1.2) and illustrates that proposition rather than evidencing it. Establishing within-province variation empirically requires a per-comune measured substrate, which is identified as the next measurement rather than claimed here.
  2. The compound-stressor cluster is identifiable by a single quantitative threshold — 770 comuni above R10 ≥ 0.62 — not by qualitative Mezzogiorno/non-Mezzogiorno classification.
  3. The 22-year backcast measures widening divergence under three of four historical investment regimes. The 2017–2020 window is the one documented exception.
  4. The 2026–2031 forecast fan separates sharply at year three: PNRR-continuity holds, sustained-investment narrows, compound-shock widens the divergence to the historical maximum.
  5. The structural feedback loop couples three modifiers (R10, R6, W-axis) at per-comune granularity. No single-modifier intervention closes the pattern.
  6. The pattern is generalisable across OECD jurisdictions at equivalent per-LAU-2 granularity: Spanish Andalucía/Extremadura, French Occitanie/Hauts-de-France, German East-West divide.

§1 · Methodology

§1.1 · Cohort construction

The brief's empirical surface is the 7,901-comune Italian cohort drawn from the ISTAT LAU-2 2026 register. The register is the canonical Italian LAU-2 instrument and is updated annually to reflect comuni mergers, de-mergers, and administrative-boundary changes recorded by the Italian national statistical institute. Per-comune matching to the SSI Index v4.2 modifier surface runs against this register as the identity anchor.

The cohort is read at three nested granularities throughout the brief. LAU-2 (comune) is the 7,901-unit surface where the feedback-loop mechanism operates and where the compound-stressor cluster manifests. NUTS-3 (provincia) is the 107-unit aggregate at which the historical backcast and forward forecast are resolved — the granularity at which per-regione policy decisions typically land on quantitative measurement. NUTS-2 (Regione) is the 21-unit administrative surface at which political-economy decisions are made; the brief refers to this granularity selectively and shows explicitly where the per-NUTS-2 view compresses the pattern.

Matching integrity between the ISTAT register and the SSI Index v4.2 modifier surface is measured at 100% at the Italian LAU-2 level per the cohort-reconciliation pass of 2 November 2026. No null-matched comuni; all 7,901 units carry a modifier surface reading.

§1.2 · R10 + R6 + W-axis definition pin

The SSI Index v4.2 modifier surface comprises ten R-modifiers (climate-physical through adaptation-governance) and a W-axis regulatory audit-state mesh (W1–W10 refined to a 5/1/1/3 class split per the 11 June 2026 codification). This brief reads three of those surfaces in combination.

R10 · distributive-justice modifier. R10 measures the distributive substrate of adaptation-choice at per-comune granularity — the share of the population that bears the cost of adaptation decisions disproportionately, measured against per-NUTS-3 socioeconomic substrate indicators drawn from ISTAT BES (Benessere Equo e Sostenibile), OIPE (Osservatorio Italiano della Povertà Energetica), and Eurostat regional-accounts registers. R10 is bounded continuous [0, 1]. The threshold R10 ≥ 0.62 defines the compound-stressor cluster this brief centres on.

Register note on R10 · T2 · Modelled
This brief describes the substrate above as per-NUTS-3 (province) resolution, while its own source register describes the OIPE energy-poverty indicators as per-comune. The two descriptions disagree, and that is recorded here rather than resolved in favour of either. Until it is resolved, the per-comune R10 surface is read as modelled at comune level, not measured there: to whatever extent the substrate is provincial, comuni in the same province inherit one value and the surface carries no within-province empirical variation from that part of it. It is read in this brief as an illustrative distributive surface sufficient to locate a pattern, and it is not claimed as a per-comune measurement. The SSI Index's own published per-substation distributive-justice field is a separate quantity and is not the surface used here. Every figure in this brief that rests on the modifier surface carries the T2 · Modelled badge for the same reason.

R6 · climate-physical modifier. R6 measures the climate-physical exposure the substrate faces, drawn from the per-substation hazard surface: heatwave frequency, flood-recurrence interval, drought-severity index, and compound-event probability. R6 is bounded continuous [0, 1]. The modifier was anchored empirically in the 2026 European heatwave reads reported in SSI Index flash briefs F-01 and F-02.

W-axis · regulatory audit-state mesh (W1–W10). The W-axis measures the instrument-state of the comune's administrative apparatus against ten regulatory audit conventions, refined to a 5/1/1/3 class split (five canonical · one candidate · one in review · three candidates). The axis captures the regulatory plurality that the Italian multi-level governance system produces at per-comune level by constitutional design — different ARERA engagements, different PNRR absorption pathways, different MASE adaptation-planning anchors across ordinary- and special-statute regioni.

The R6 climate-physical modifier and the environmental-acceleration framework are peer-reviewed in Bérard (2026) Journal of Infrastructure Preservation and Resilience 7(1):35, doi:10.1186/s43065-026-00193-z, and Bérard (2026) Environmental Research: Energy 3(3):035007, doi:10.1088/2753-3751/ae87a5. These anchors operationalise the v4.0.2 substation-scale canonical and do not themselves define the R10 distributive-justice modifier or the W-axis regulatory audit-state mesh. The R10 and W-axis conventions used in this brief are first operationalised in the SSI Index v4.2 extension described here and in the prior Strategic Briefs SB-01 and SB-02; peer-review anchors for R10 and W-axis are in preparation.

§1.3 · The compound-stressor cluster · R10 ≥ 0.62

The threshold R10 ≥ 0.62 defines the compound-stressor cluster at 770 comuni. The 0.62 figure is the current canonical threshold per v4.2 methodology; the derivation follows the statistical procedure in the ERE companion paper (§3.4) and is pinned to the natural inflection in the cohort-wide distribution.

FIGURE 2 · THREE-MODIFIER SURFACET2 · Modelled

The 770-comune compound-stressor cluster on the Italian map

Three-panel strip restricted to the 770-comune compound-stressor cluster · Panel A R10 kernel density · Panel B R6 kernel density · Panel C W-axis 5/1/1/3 instrument-class mix by regione cluster

Figure 2 · R10 · R6 · W-axis three-modifier surface for the 770-comune compound-stressor cluster Three-panel strip of per-comune modifier readings restricted to the 770-comune compound-stressor cluster. Panel 1: R10 distributive-justice density. Panel 2: R6 climate-physical density. Panel 3: W-axis instrument-class mix. Illustrative distribution shapes. n=770 comuni · SSI Index v4.2 snapshot 2026 · modelled density and class mix Panel A · R10 distributive-justice kernel density · cluster subset R10 = 0.62 0.62 1.00 Panel B · R6 climate-physical kernel density · cluster subset 0.0 1.0 Panel C · W-axis class mix 5/1/1/3 instrument distribution CAM PUG CAL SIC 5 active 1 candidate 1 review 3 candidates Three-modifier surface restricted to compound-stressor cluster · densities on the modelled surface · class-mix per ARERA 2025 baseline · CC BY-SA 4.0

§1.4 · Backcast methodology

The backcast runs 22 annual snapshots 2003–2025, resolved at per-NUTS-3 (107 units). The historical investment-regime classification distinguishes four regimes: pre-crisis 2003–2008 · austerity 2009–2016 · PNRR-precursor 2017–2020 · PNRR-active 2021–2025.

§1.5 · Forecast fan methodology

The forecast fan runs five annual snapshots 2026–2031, resolved at per-NUTS-3. Three scenarios parameterise the investment-rate multiplier and shock exposure: PNRR-continuity (×1.0), sustained-investment (×1.4), compound-shock (×0.7 + climate-physical shock + fiscal-transmission shock). Monte Carlo uncertainty bounds at N = 5,000 draws per-NUTS-3 per-scenario per-year.

§1.6 · What is peer-reviewed, what is model-derived, and what is pending

The brief rests on three distinct categories of claim that the reader is invited to hold apart.

Peer-reviewed. The R6 climate-physical modifier and the Markov degradation + multi-hazard environmental-acceleration framework that generates its per-asset distribution are anchored in Bérard (2026) Journal of Infrastructure Preservation and Resilience 7(1):35 and Bérard (2026) Environmental Research: Energy 3(3):035007. Both papers operationalise the v4.0.2 canonical register at per-substation resolution. The Markov five-state transition specification, the per-asset Expected Time to Critical distribution, the compound environmental-hazard non-linearity, and the validation against 1,220 independently documented substation failures are peer-reviewed. The thresholding of R6 against percentile anchors is peer-reviewed.

Operationalised in-brief, not yet peer-reviewed. The R10 distributive-justice modifier and the W-axis regulatory audit-state mesh are first operationalised in the SSI Index v4.2 extension — in this brief, in SB-01 (July 2026), and in SB-02 (October 2026). The R10 construction procedure (aggregating ISTAT BES indicators, OIPE energy-poverty composites, and Eurostat regional-accounts substrate to a [0,1] scalar at per-LAU-2 resolution) is documented in the SB-02 appendix. The W-axis "5/1/1/3 instrument distribution" (five active instrument classes, one candidate, one in review, three candidate-under-development) is the current per-regione codification; the per-comune aggregation to a scalar follows the 11 June 2026 codification note. Peer-review anchors for both R10 and the W-axis are in preparation. The three-modifier coupling specification (§3.1) and the four-component sub-system framework (§3.2) are conceptual extensions proposed by this brief and are not themselves measurement procedures.

Model-derived scenario outputs. All forecast values in §4 (corridor widths, cluster counts, cascade coefficients) are outputs of the three-scenario Monte Carlo simulation under the parameterisation stated in §1.5. They are not predictions. The quantitative values reported in Figures 1, 2, 3, 4, 7, 9, 10 and in the corresponding prose are current modelled-surface readings pending binding to the canonical v4.2 register; the final bind is scheduled for v1.1 chapter-open amendment. The figures carry an explicit modelled-surface annotation. The reader is invited to attend more to the pattern than to the digit.

What this means for the brief's claims. The structural finding — that the per-LAU-2 surface records a compound-stressor cluster concentrated in the Mezzogiorno NUTS-3 units — rests on the R6-anchored peer-reviewed framework plus the R10 and W-axis extensions operationalised in-brief. The cross-jurisdictional generalisability finding (§5) is pattern-consistent on the modelled surface; quantitative cross-cohort binding is pending. The forecast separation at year three (§4.4) is a modelled scenario separation, not a prediction.

§2 · The 22-year backcast

FIGURE 3 · DIVERGENCE TRAJECTORYT2 · Modelled

Per-NUTS-3 divergence 2003–2025 · 10 Mezzogiorno NUTS-3 + 10 Central/Northern NUTS-3

Twenty-line trajectory · annual R10 median at per-NUTS-3 level · Mezzogiorno in cayenne · Central/Northern in steel · divergence corridor shaded · investment-regime boundaries vertical rules

Figure 3 · Per-NUTS-3 R10 trajectory 2003–2025 · twenty NUTS-3 units Time-series of R10 modifier values for twenty Italian NUTS-3 units from 2003 to 2025. Ten Mezzogiorno NUTS-3 units in cayenne. Ten Central/Northern NUTS-3 units in steel. Divergence corridor shaded. Illustrative shape. n=20 NUTS-3 units · SSI Index v4.2 2003–2025 backcast · corridor delta 0.19 (2003) → 0.31 (2025) · modelled working surface 0.2 0.4 0.6 0.8 1.0 R10 modifier 2003 2008 2013 2018 2023 2025 2017–2020 convergence window Mezzogiorno median Central/Northern median Mezzogiorno: Napoli · Caserta · Salerno · Potenza · Matera · Cosenza · Reggio Calabria · Palermo · Catania Trapani · Centre/North: Milano · Torino · Bologna · Padova · Firenze · Genova · Verona · Varese · Bergamo · Brescia

§2.1 · Per-NUTS-3 divergence trajectory 2003–2025

Figure 3 reads the per-NUTS-3 R10 trajectory for twenty provinces across the 22-year backcast. Ten Mezzogiorno NUTS-3 units — Napoli, Caserta, Salerno, Potenza, Matera, Cosenza, Reggio Calabria, Palermo, Catania, Trapani — are rendered in cayenne. Ten Central / Northern NUTS-3 units — Milano, Torino, Bologna, Padova, Firenze, Genova, Verona, Varese, Bergamo, Brescia — are rendered in steel. The divergence corridor is the vertical band between the Mezzogiorno median trajectory and the Central / Northern median trajectory.

At the 2003 baseline the divergence corridor is moderate, the two median trajectories separated by an R10 delta that reads as a measurement but not yet as a structural signature. By 2025 the corridor has widened materially; the Mezzogiorno median has moved upward on the R10 axis while the Central / Northern median has held near its 2003 position. The pattern is not uniform within each group — Napoli widens faster than Caserta; Milano holds more tightly than Torino — but the median-level separation reads consistently across every annual snapshot. The corridor width — measured as the R10 delta between the Mezzogiorno median trajectory and the Central / Northern median trajectory on the W-axis-weighted surface — reads at approximately 0.19 at the 2003 baseline and approximately 0.31 at the 2025 reading in the current modelled surface, a widening on the order of 50–70% over the 22-year record. Final values pending binding to the v4.2 canonical register; values reported here are model-derived from the working surface and are flagged illustrative in Figure 3.

The trajectory shape is not monotonic. Three sub-periods of convergence appear in the record — brief windows during which the Mezzogiorno median moved downward or held while the Central / Northern median held stable — before the widening resumes. The 2017–2020 window is the most pronounced of the three and is developed separately in §2.3. The pre-crisis 2003–2008 micro-convergence and the austerity-era 2013–2014 micro-convergence are documented in the record but do not alter the long-run trajectory direction.

The reader is left to draw the trajectory's implication for the five-year forecast that follows. The brief does not conclude from the 22-year widening that the widening is inevitable; it reports that the widening is measurable, patterned, and operates at the per-comune granularity Figure 1 surfaces.

§2.2 · Four historical investment regimes

The backcast distinguishes four investment regimes across 2003–2025. Each regime is defined by its governing fiscal instrument and its per-NUTS-3 infrastructure-investment rate.

Pre-crisis · 2003–2008. The Legge Obiettivo (Law 443/2001) and the second EU Cohesion Policy programming cycle (2000–2006) frame the period. Per-NUTS-3 investment rates run high in absolute terms but unevenly in geographic distribution; the Mezzogiorno absorption rate against allocated Cohesion funds reads below the national average, a pattern later analysed in the Barca 2009 Report on the reformed Cohesion Policy.

Austerity · 2009–2016. The 2008 financial crisis, the 2011 sovereign-debt stress, and the Monti / Letta / Renzi fiscal-consolidation sequence frame the period. Per-NUTS-3 investment rates fall sharply and asymmetrically; the Mezzogiorno absolute investment volume compresses faster than the Central / Northern volume. The 2013–2014 micro-convergence window appears here, driven by the decreti sblocca-Italia short-term stimulus package before the broader austerity trajectory resumes.

PNRR-precursor · 2017–2020. The Industria 4.0 Plan (2016), the Stabilità 2017 fiscal package, and the Piano Nazionale per la Ripresa 2018–2020 collectively frame a transitional regime in which infrastructure-investment rates partially recover and specific per-comune programmes — fibre broadband, ZES (Zone Economiche Speciali), energy-infrastructure upgrade — concentrate investment in the Mezzogiorno. This is the regime producing the 2017–2020 convergence window developed in §2.3.

PNRR-active · 2021–2025. NextGenerationEU adoption in 2021 and the Italian PNRR (Decreto 77/2021 · Missione 2 Rivoluzione verde e transizione ecologica) frame the current regime. Per-NUTS-3 investment rates run at the highest absolute level of the 22-year record. The Mezzogiorno share is formally pinned to minimum thresholds by the PNRR structural rules. Early-period absorption 2021–2023 reads below the formal targets; late-period absorption 2024–2025 improves but remains asymmetric across NUTS-3 units within the Mezzogiorno. Aggregated per-NUTS-3 investment-rate indices, normalised on the 2003–2008 pre-crisis mean (=1.00), read on the modelled surface as approximately 1.00 pre-crisis, 0.55–0.60 austerity, 0.80–0.90 PNRR-precursor, and 1.40–1.50 PNRR-active across the Mezzogiorno NUTS-3 cohort; the Central / Northern equivalent reads approximately 1.00 / 0.70 / 0.90 / 1.25–1.30. Final values pending binding to the v4.2 register and DEF historical fiscal-stance cross-walk. The widening PNRR-active gap on the investment side is the pattern to hold against the §4 scenario separation.

FIGURE 4 · REGIME OVERLAYT2 · Modelled

Investment regime overlay 2003–2025 · four regimes against per-NUTS-3 divergence trajectory

Time-series overlay · investment-rate per regime shown as band · divergence corridor shown as line · four regime labels annotated

Figure 4 · Investment regime overlay 2003–2025 · four regimes against the per-NUTS-3 divergence trajectory Time-series overlay of four investment regimes (pre-crisis 2003-2008, austerity 2009-2016, PNRR-precursor 2017-2020, PNRR-active 2021-2025) against the Mezzogiorno vs Central/Northern R10 divergence corridor. Illustrative shape. Mezzogiorno normalised investment-rate index (2003–2008 pre-crisis = 1.00): 1.00 / 0.58 / 0.84 / 1.43 · modelled working surface pre-crisis 2003–2008 · ×1.00 austerity 2009–2016 · ×0.58 PNRR-precursor 2017–2020 · ×0.84 PNRR-active 2021–2025 · ×1.43 2003 2008 2013 2018 2023 2025 Δ = 0.19 Δ = 0.31 partial convergence window Line · per-NUTS-3 R10 corridor-width (Mezzogiorno vs Central/Northern medians, W-axis-weighted) · Bands · Mezzogiorno investment-rate regime Legge Obiettivo (2001) · Monti/Letta/Renzi austerity (2009–2016) · Industria 4.0 + ZES (2017–2020) · PNRR Missione 2 + 5 (2021–2025) · CC BY-SA 4.0

§2.3 · The 2017–2020 convergence window

Across 2017–2020 the per-NUTS-3 divergence corridor narrows. The brief records the measurement; it does not conclude that the convergence was engineered, that it was replicable at scale, or that its underlying conditions can be reconstructed. Four observations frame the reading.

First. Not every Mezzogiorno NUTS-3 unit participated. Napoli and Bari narrowed materially. Reggio Calabria and Caltanissetta did not. The convergence is not a Mezzogiorno-wide phenomenon but a per-NUTS-3 one · its geography matches the geography of the specific investment programmes active in the period.

Second. The investment-rate shift was asymmetric. Per-NUTS-3 absolute investment volumes rose faster in the participating Mezzogiorno NUTS-3 units than in the Central / Northern ones · the delta compressed because the lagging units were catching up, not because the leading units were slowing.

Third. The regulatory substrate showed measurable W-axis activity. ZES designations in Campania, Calabria, Puglia, and Sicilia activated during the window; ARERA regulatory-audit activity on grid-infrastructure projects in Mezzogiorno NUTS-3 units ran at a higher rate than in Central / Northern NUTS-3 units of comparable scale.

Fourth. The convergence did not survive the regime boundary. The transition into the PNRR-active regime 2021–2025 brought larger absolute investment volumes but a less-focused per-NUTS-3 distribution; the corridor widened again in the record even as aggregate investment rose. Why this happened is a question the brief does not answer; that it happened is what the measurement records.

§2.4 · What the backcast records (reader-inference anchor)

The backcast records the trajectory. The reader draws the trajectory's implication for the five-year forecast that follows. The brief does not conclude whether 2017–2020's convergence conditions were replicable at scale.

§3 · The compound-stressor cluster · structural feedback loop

§3.1 · The three-modifier coupling (R10 · R6 · W-axis)

The structural feedback loop couples three SSI Index modifiers at per-comune granularity. R10 (distributive-justice) measures the substrate of adaptation-choice distribution. R6 (climate-physical) measures the exposure the substrate faces. The W-axis (regulatory audit-state mesh) measures the regulatory instrument-state of the comune's administrative apparatus. The three co-amplify under the mechanism described below.

The coupling runs in one direction most of the time and in both directions during shock windows. Under steady-state conditions — the condition governing most of the backcast record — R6 climate-physical exposure modulates R10 indirectly through the W-axis: climate exposure concentrates regulatory audit-state activity on particular classes of infrastructure, that activity consumes administrative capacity, and the distributive-justice modifier reads the residual capacity (or its absence) as a per-comune substrate signal. The feedback cycle is annual at the measurement level — SSI Index snapshots are produced annually — but the underlying policy-administrative cycle on which the W-axis operates runs faster (ARERA quarterly cycles, MASE semi-annual) and the civic-governance substrate R10 anchors on operates slower (electoral cycles, demographic trajectories, inter-generational substrate).

Under shock conditions — the regime the §4.3 compound-shock scenario measures — the coupling runs in both directions simultaneously. An acute R6 event (heatwave, flood, drought) perturbs per-comune infrastructure-failure probability directly, which triggers a fiscal-transmission response captured on the W-axis, which concentrates distributive consequences on the R10 surface, which in turn conditions the regulatory capacity available for the next R6 event. The Buldyrev et al. (2010) interdependent-network cascade mathematics operationalises this coupling in the Systemic Layer substrate (convention SY.11); Figure 8 schematises the cascade at the per-NUTS-3 resolution the brief reads.

The product of amplification coefficients around the loop determines whether the cycle self-amplifies or self-dampens. Under the brief's three scenarios, the loop reads as self-dampening under sustained-investment (coefficient product below 1.0), near-stationary under PNRR-continuity (product near 1.0), and self-amplifying under compound-shock (product above 1.0). These are regime-dependent properties of the coupling, not fixed properties of the modifier definitions — a point the brief returns to in §3.4.

FIGURE 5 · FEEDBACK-LOOP SCHEMATICSchematic

Three-modifier coupling at per-comune granularity

Schematic diagram · R10 · R6 · W-axis positioned as vertices · coupling arrows with directional amplification · per-comune feedback-cycle period annotated

Figure 5 · Three-modifier coupling at per-comune granularity Conceptual schematic diagram showing the structural feedback loop that couples three SSI Index modifiers at per-comune level: R10 distributive-justice modifier, R6 climate-physical modifier, and the W-axis regulatory audit-state mesh. Three modifier vertices are positioned as a triangle around a central feedback-cycle period annotation. Directional coupling arrows between each pair of modifiers indicate the amplification channels. The schematic is not quantitative; it illustrates the mechanism the §3.1 text describes. schematic · not to scale · mechanism illustration for §3.1 R10 distributive-justice modifier R6 climate-physical modifier W-axis regulatory audit-state mesh (W1-W10) amplifies compounds conditions FEEDBACK CYCLE operates at per-comune (LAU-2) on annual snapshot cadence LEGEND primary coupling (amplification) secondary coupling (modulation) The three couplings co-amplify under the mechanism §3.1 describes. No single-modifier intervention closes the loop. SSI Index v4.2 · R10 · R6 · W-axis per JIPR v16 (2026) · ERE 3(3):035007 (2026) · CC BY-SA 4.0

§3.2 · The political-economy-fiscal-talent-infrastructure sub-system

The three-modifier coupling described in §3.1 does not by itself distinguish the Mezzogiorno pattern from a generic under-investment case. The specificity sits one level deeper, in the sub-system that produces the per-comune substrate the R10 modifier reads. The sub-system has four coupled components — political-economy, fiscal-transmission, talent-flow, and infrastructure-stock — whose joint operation generates the compound-stressor configuration the brief measures.

The political-economy component inherits a long history of regional-asymmetric institutional capacity that Bagnasco (1977) first mapped as the Tre Italie — the three Italies — distinguishing the industrial North-West, the small-firm North-East-Centre, and the dependent-development South. The 2026 SSI Index per-comune readings do not confirm the Bagnasco tripartition directly (the LAU-2 resolution is finer than the NUTS-2 clusters he worked with), but the W-axis distribution across regioni reads as consistent with it: the regulatory-mesh density per comune is higher in the small-firm NUTS-2 clusters than in the Mezzogiorno NUTS-2 clusters even where the comuni are of comparable size. The W-axis difference is not an artefact of population — it is an artefact of the regulatory-administrative substrate the comuni operate within.

The fiscal-transmission component determines how national and EU-level fiscal instruments reach per-comune substrate. Barca (2009), in the report prepared for the European Commission on reforming cohesion policy, framed the question place-based: transfers bound to national or regional envelopes without per-territory capacity-building absorb themselves back into the originating institutions without substrate effect at the comune level. The PNRR Missione 5 per-comune outlays 2021–2025 read as a partial test of the Barca place-based framework at Italian national scale; §2.3 reports the four-regime backcast against which the test is measured.

The talent-flow component is the economic-geography sub-system Rodríguez-Pose (2018) framed as "the revenge of the places that don't matter." Per-comune talent out-migration, where it concentrates and sustains itself over multiple cycles, hollows the administrative-institutional capacity on which the W-axis depends and the political-representation capacity on which the fiscal-transmission depends. The brief treats talent-flow as a modifier on both W-axis evolution and R10 trajectory rather than as an independent SSI Index axis; the per-comune backcast records out-migration cohorts as one of the four principal R10 trajectory drivers 2003–2025.

The infrastructure-stock component is the cumulative legacy of per-comune infrastructure investment, deferred maintenance, and depreciation. The stock reads on the SSI Index as the baseline against which R6 climate-physical exposure translates into infrastructure-failure probability. Where the stock is thin, the same R6 exposure produces a larger W-axis and R10 consequence — the coupling §3.1 described. The four components couple pairwise: political-economy conditions fiscal-transmission, fiscal-transmission conditions infrastructure-stock, infrastructure-stock conditions talent-flow (through local employment conditions), and talent-flow conditions political-economy (through demographic-representation effects). The compound-stressor configuration the brief measures is the fixed-point of this four-component sub-system under the austerity (2009–2016) and PNRR-precursor (2017–2020) fiscal regimes. The §4 scenarios test whether the PNRR-active regime perturbs the fixed-point sufficiently to move the per-comune reading off the compound-stressor surface, or whether the perturbation absorbs back without substrate effect.

§3.3 · Three per-comune case anchors

Three comuni are selected to illustrate the feedback-loop mechanism against distinct substrate conditions. Selection balances geographic distribution across the Mezzogiorno, substrate-condition diversity, and narrative specificity.

Scanzano Jonico (Basilicata)

Scanzano Jonico is a comune of some seven thousand residents on the Ionian coast of Basilicata, in the province of Matera. In November 2003 a decree of the Berlusconi II government designated Scanzano Jonico as the national site for the long-term storage of high-activity radioactive waste. The decree was withdrawn within fifteen days following a local civic-society mobilisation that drew tens of thousands of participants to the comune and opened a national debate on siting decisions, democratic consent, and regional-infrastructure asymmetry. The episode left a durable civic-governance substrate in the comune that reads on the W-axis today.

On the R10 distributive-justice surface, Scanzano Jonico sits above the 0.62 threshold and inside the compound-stressor cluster. The R6 climate-physical exposure is moderate — Mediterranean temperature trajectory, coastal-erosion exposure, periodic drought. The W-axis state reads as engaged: current PNRR Missione 2 energy-infrastructure projects are active in the comune and are tracked publicly through the Agency for Territorial Cohesion's monitoring portal. The combination — high R10, moderate R6, active W-axis — makes Scanzano Jonico the brief's exemplar of a compound-stressor cluster comune with a visible civic-governance substrate.

Laviano (Campania)

Laviano is a comune of fewer than fifteen hundred residents in the Alto Sele valley of the province of Salerno, inland Campania. On 23 November 1980 the Irpinia earthquake (M 6.9) struck the area; Laviano recorded nearly three hundred deaths against its pre-earthquake population of roughly two thousand. The reconstruction that followed — administered through Law 219/1981 and a long sequence of successor instruments — is still legible in the current R10 surface forty-six years later. Several comuni in the Alto Sele valley record R10 values materially above the national median; Laviano is one of them.

On the R6 climate-physical surface, Laviano sits in a seismically active and hydrologically complex Apennine zone; the R6 compound-event probability reads near the upper quartile of the Italian cohort. The W-axis reads selectively engaged — long-cycle reconstruction programmes have maintained administrative activity but at a cadence that compounds the R10 substrate rather than relieves it. Laviano is the brief's exemplar of path-dependent substrate — the pattern by which a single historical event, imperfectly recovered, continues to shape the R10 reading four decades later.

Partanna (Trapani · Sicilia)

Partanna is a comune of some ten thousand residents in the province of Trapani, western Sicilia. The comune sits in the Belice valley; the 1968 Belice earthquake and its decades-long reconstruction frame its long-horizon substrate. The specificity that makes Partanna the brief's W-axis exemplar is more current: the comune operates at the intersection of three constitutionally distinct regulatory frameworks that it must navigate in combination — the Regione Siciliana's special-autonomous statute (Statuto Siciliano, 1946, with constitutional rank under the Italian Republic), which carries dedicated grid-infrastructure permitting competences; ARERA's national electricity-regulatory framework operating under the Italian Constitution's concurrent-competence provisions; and the MASE national adaptation-planning anchor operating under the Climate Law framework. The three frameworks are not unaligned by design — each is constitutionally established in its own domain — but the per-comune administrative interface across all three produces a W-axis class profile that reads distinctively in the cohort.

On the R10 surface, Partanna sits above the compound-stressor threshold. The R6 climate-physical exposure is high — the Sicilian drought-severity trajectory 2017–2025 is one of the most pronounced in the Italian record. The W-axis reads as a regulatory-mesh specificity: five W-conventions active, one candidate, one in review, three candidates — the full 5/1/1/3 class distribution expressed at per-comune level. Partanna is the brief's exemplar of a constitutionally plural regulatory environment producing a distinctive W-axis signature at per-comune resolution; the brief reports the measurement and expresses no view on the constitutional architecture that underlies it.

FIGURE 6 · CASE ANCHORST2 · Modelled

Three per-comune case anchors on a shared R10/R6/W-axis surface

Triptych · one panel per comune · R10 time-series + R6 time-series + W-axis state · 2003–2025 backcast with 2026–2031 forecast fan · same surface dimensions across panels for comparability

Figure 6 · Three per-comune case anchors on a shared R10/R6/W-axis surface Triptych. One panel per comune: Scanzano Jonico, Laviano, Partanna. Each panel shows R10 time-series and R6 time-series from 2003 to 2025 with 2026-2031 forecast fan, plus W-axis state indicator. Shared axes across panels. Three case anchors · 2003–2025 backcast + 2026–2031 forecast fan · shared R10 scale · modelled working surface Scanzano Jonico Basilicata · MT · R10↑ R6→ W engaged R10 = 0.62 2003 waste-site decree 2003 ———————— 2025 | 2026–2031 Laviano Campania · SA · R10↑↑ R6↑ W selective R10 = 0.62 1980 Irpinia M6.9 (legacy) 2003 ———————— 2025 | 2026–2031 Partanna Sicilia · TP · R10↑ R6↑↑ W 5/1/1/3 R10 = 0.62 2017–2025 drought trajectory 2003 ———————— 2025 | 2026–2031 R10 distributive-justice R6 climate-physical forecast fan · three-scenario envelope R10 R6 scaled on shared vertical axis per panel · W-axis class profile reported in sub-header · case selection per §3.3 · CC BY-SA 4.0 Scanzano Jonico: 2003 radioactive-waste designation + civic mobilisation Laviano: 1980 Irpinia earthquake path-dependence · Partanna: Belice valley + Regione Siciliana + ARERA + MASE regulatory mesh

§3.4 · What does not close the pattern

The backcast record 2003–2025 includes four categories of intervention that addressed a single modifier of the three-modifier coupling without perturbing the four-component sub-system §3.2 described. Each produced measurable local effects. None closed the convergence gap at per-NUTS-3 resolution. The brief catalogues them not to disparage the interventions, but to establish the structural claim: the pattern is a sub-system property, and single-modifier interventions, however well-executed, read as sub-threshold against the sub-system fixed-point.

The first category is infrastructure capital investment without regulatory-administrative capacity-building. Several Mezzogiorno NUTS-3 units recorded large per-capita infrastructure outlays over the 2003–2008 pre-crisis regime through national instruments operating ahead of the PNRR architecture. The R6 climate-physical and W-axis readings moved — new stock lowers failure probability on new assets — but the R10 distributive-justice trajectory did not. The sub-system reabsorbed the stock effect through maintenance-schedule reversion and administrative-capacity constraint on new-stock uptake. Infrastructure outlay without accompanying W-axis capacity-building is one of the four sub-threshold intervention classes.

The second category is regulatory-instrument introduction without infrastructure-stock support. The 2017–2020 PNRR-precursor regime introduced several per-regione regulatory instruments ahead of the fiscal envelope that would eventually support them. The W-axis class profile widened — the audit-state mesh grew across more comuni — but R10 did not respond, because the regulatory instruments lacked the physical-stock substrate on which they could act. Regulatory expansion without stock is a second sub-threshold class.

The third category is talent-attraction instruments without fiscal-transmission reform. Several regioni introduced per-comune talent-attraction instruments over the 2017–2025 interval — fiscal incentives on individual inward migration, academic-institutional co-location, start-up and SME supports. The talent-flow component responded measurably in the destination comuni. The R10 compound-stressor reading did not, because the fiscal-transmission component did not reform in parallel: the inward talent flows arrived into unchanged fiscal-administrative substrates and absorbed back without cumulative substrate effect.

The fourth category is political-economy reform without the other three components. The 2001 constitutional reform (Title V), re-opened in subsequent reform cycles, redistributed competences between national and regional levels. The W-axis records the reform as a regulatory-architectural shift measurable across all regioni. The R10 compound-stressor trajectory in the Mezzogiorno NUTS-3 units did not respond, because the fiscal-transmission, talent-flow, and infrastructure-stock components did not couple into the political-economy reform at per-comune scale. Political-economy reform without three-component coupling is the fourth sub-threshold class.

The four categories exhaust the single-modifier intervention record 2003–2025. The brief's §4 scenarios are built against this record: the PNRR-active (2021–2025) and projected 2026–2031 envelopes constitute the first sustained attempt to perturb all four sub-system components simultaneously. Whether the perturbation closes the convergence gap is the empirical question the forecast reads.

§4 · Three forecast scenarios · 2026–2031

§4.1 · Scenario A · PNRR-continuity

Scenario A projects the per-NUTS-3 trajectory 2026–2031 under an investment-rate multiplier of 1.0 (PNRR baseline sustained at current cadence), regulatory continuity (no major W-axis instrument additions or withdrawals), and no superimposed climate-physical or fiscal-transmission shock. The scenario is the brief's reference baseline — the trajectory that the four-component sub-system produces under its current operating regime.

Under the PNRR-continuity scenario, the 2017–2020 convergence window observed on Napoli and Bari (§2.3) extends partially into the forecast horizon: both NUTS-3 units continue to narrow against the central-north reference corridor through year three, then converge toward a reduced-rate asymptote as the PNRR Missione 5 and Missione 2 envelopes complete their current drawdown schedules. Per-comune R10 readings in the compound-stressor cluster hold their 2025 distribution within confidence band — no systematic cluster reduction, no systematic cluster growth. The Reggio Calabria and Caltanissetta NUTS-3 units continue to widen against the reference corridor through year five, with no reversion mechanism engaged. The three case-anchor comuni (§3.3) track their respective substrate-conditioned trajectories: Scanzano Jonico holds, Laviano continues its slow upward R10 drift, Partanna's W-axis regulatory mesh records no convergence toward the national pattern.

The corridor shape at 2031 under Scenario A is best read as pattern-stationary: the Mezzogiorno NUTS-3 divergence pattern neither closes nor widens materially, but the per-comune compound-stressor cluster retains its 770-unit scale with modest geographic reorganisation. The scenario is the brief's reference against which Scenarios B and C are measured.

§4.2 · Scenario B · Sustained-investment

Scenario B projects the per-NUTS-3 trajectory 2026–2031 under an investment-rate multiplier of 1.4 (PNRR baseline supplemented with the 2028–2034 EU Cohesion Policy cycle instruments plus Italian national co-funding at the levels proposed in the 2026 Documento di Economia e Finanza), regulatory expansion (W-axis instrument additions tracking the fiscal envelope), and the four-component sub-system coupled operationally in line with the Barca (2009) place-based framework.

Under the sustained-investment scenario, the 2017–2020 convergence window extends into the forecast horizon and widens to cover additional NUTS-3 units. Napoli and Bari continue narrowing through the full horizon. Catania, Palermo, and Salerno — NUTS-3 units that showed sub-threshold convergence signal in the 2017–2020 backcast — move to measurable convergence against the reference corridor by year three. The three Reggio Calabria–Vibo Valentia–Crotone units (the brief's furthest-divergence Calabrian cluster in the 2003–2025 record) record first stabilisation around year two, then slow convergence from year four. The 770-comune compound-stressor cluster contracts to approximately 620 comuni by year five, with contraction concentrated in the Campania and Puglia regioni and partial in Sicilia. Laviano (§3.3) records the slowest response of the three case anchors — the path-dependent substrate is resistant to the fiscal envelope alone.

Scenario B is the scenario in which the Barca place-based principle is operationalised at national scale: per-comune capacity-building paired with per-comune fiscal envelope, regulatory-instrument expansion paired with infrastructure-stock support, talent-attraction paired with fiscal-transmission reform. The scenario is not a projection of political-economy likelihood — it is a projection of the trajectory that would result from the sub-system perturbation the Barca framework prescribes.

§4.3 · Scenario C · Compound-shock

Scenario C projects the per-NUTS-3 trajectory 2026–2031 under an investment-rate multiplier of 0.7 (PNRR baseline sustained at reduced cadence following fiscal-transmission shock), combined with a climate-physical shock superimposed on the Southern NUTS-3 cohort in year two of the forecast, with cross-shock amplification operationalised per the Buldyrev et al. (2010) interdependent-network cascade mathematics (Systemic Layer convention SY.11).

The climate-physical shock parameters are calibrated on the 2022 Po Valley drought and the 2023 Emilia-Romagna floods as reference events, re-projected southward across the Sicilian and Calabrian NUTS-3 units with the per-event return-period and intensity distributions the 2026 IPCC AR6 regional projection records for the Mediterranean basin under SSP2-4.5. The fiscal-transmission shock is defined as a sustained reduction in the post-PNRR fiscal envelope: specifically, a 2028–2034 EU Cohesion Policy adoption-and-implementation delay combined with Italian national co-funding reversion toward pre-PNRR (2017–2020) levels, translated into a per-NUTS-3 infrastructure-investment-rate multiplier of approximately 0.7 against the modelled sustained-investment trajectory in years four and five of the forecast horizon. The PNRR Missione 5 drawdown itself is modelled as closing on its statutory 2026 end-of-NGEU-cycle horizon in all three scenarios; the Scenario C fiscal shock operates on the successor envelope, not on an already-closed PNRR instrument.

Under the compound-shock scenario the modelled cascade runs as follows: a climate shock event in year two concentrates on the Sicilian and Calabrian NUTS-3 units; R6 readings spike at the per-event level; W-axis instrument capacity concentrates on the immediate response and the N = 5,000-draw ensemble records a loss of administrative bandwidth from parallel R10 substrate-support activity; the modelled R10 compound-stressor cluster widens from 770 toward the high-900s in central-tendency by end of year three; the fiscal-transmission shock lands in year four on the widened cluster via the post-PNRR successor envelope described above, with the reduced multiplier concentrating on the worst-condition comuni; the cascade feeds back into the climate-physical exposure horizon as insurance-market withdrawal from the affected cohort reduces the fiscal-transmission capacity for R6-preparation activities in year five. On the modelled surface the Reggio Calabria–Vibo Valentia–Crotone cluster records the most pronounced cascade; Partanna and Scanzano Jonico record substantial but not irreversible R10 shifts in central tendency; Laviano records a seismic-adjacency secondary event (not primary cascade) that compounds its already-elevated R10 reading. By 2031 the modelled surface in the central-tendency projection shows a widened divergence pattern against the 2025 baseline; the magnitude of the widening is sensitive to the climate-shock calibration, the per-step amplification coefficients, and the assumed post-PNRR successor-envelope trajectory, and is reported as a Monte Carlo band rather than a point estimate — see Figure 7 for the envelope shape, and the forthcoming Technical Companion for the per-step sensitivity analysis. The Buldyrev et al. (2010) cross-network cascade framework informs the coupling-structure specification only; the brief does not report a decommissioning projection and deliberately declines to publish a single compound-shock multiplier at the 2031 horizon, because any such headline figure would travel outside the scenario frame that generates it. Scenario C is a stress test of the sub-system under adverse conditions, not a prediction of the Mezzogiorno infrastructure stock.

FIGURE 7 · FORECAST FANT2 · Modelled

2026–2031 forecast fan · per-NUTS-3 × three scenarios

Fan chart · one trajectory per scenario · per-NUTS-3 central tendency + 90% Monte Carlo band · scenario separation emphasised at year three · Mezzogiorno NUTS-3 units highlighted in cayenne

Figure 7 · 2026–2031 forecast fan · per-NUTS-3 × three scenarios · corridor widths at year three 0.42 / 0.31 / 0.24 Fan chart showing per-NUTS-3 central tendency plus 90% Monte Carlo band for three scenarios: PNRR-continuity, sustained-investment, compound-shock. Scenarios separate sharply at year three (2029). Monte Carlo N=5,000 draws per NUTS-3 per scenario per year. Mezzogiorno NUTS-3 cohort · W-axis-weighted R10 corridor-width · 2026 baseline = 0.31 · N = 5,000 draws per scenario-year · modelled working surface 0.20 0.28 0.36 0.44 0.52 0.60 corridor width Δ 2026 2027 2028 2029 2030 2031 year 3 · separation 2026 baseline Δ = 0.31 Scenario C · compound-shock · Δ 0.42 → 0.51 Scenario A · PNRR-continuity · Δ = 0.31 Scenario B · sustained-investment · Δ 0.24 → 0.22 0.42 0.31 0.24 Three-scenario fan · 90% Monte Carlo band shown as shaded envelope · scenario central-tendency as solid line year-three separation window overlaps PNRR Missione 5 drawdown + 2028–2034 Cohesion cycle decision horizon Buldyrev et al. (2010) cascade mathematics operationalised in Scenario C (per Systemic Layer convention SY.11) · CC BY-SA 4.0

FIGURE 8 · CASCADE DIAGRAMT2 · Modelled

Compound-shock amplification cascade

Cascade diagram · climate-physical shock entry → R6 modifier perturbation → per-comune infrastructure-failure probability → fiscal-transmission response → R10 compound amplification · per-step amplification factors annotated

Figure 8 · Compound-shock amplification cascade (scenario C) Cascade diagram showing the compound-shock scenario amplification path. A climate-physical shock enters at the R6 modifier, perturbing the per-comune infrastructure-failure probability surface; the shock then propagates through a fiscal-transmission response captured by the W-axis regulatory audit-state mesh, re-entering the R10 distributive-justice modifier as a compound amplification. The schematic illustrates the Buldyrev interdependent-network cascade mathematics at per-NUTS-3 granularity. scenario C compound-shock · cascade path per Buldyrev et al. (2010) · schematic STEP 1 Climate- physical shock entry heatwave · flood · drought compound perturbs STEP 2 R6 modifier perturbation per-comune infrastructure- failure probability ↑ triggers STEP 3 Fiscal- transmission response W-axis state investment-rate ↓ regulatory lag compounds STEP 4 R10 compound amplification distributive- justice surface R10 shift ↑ feeds STEP 5 Next- period input t+1 cycle restart feedback loop · next period AMPLIFICATION COEFFICIENTS (indicative) α₁₂ · climate→R6 · ≈ 1.0 (direct proportional) | α₂₃ · R6→W-axis · 0.7–1.2 (compound rate) α₃₄ · W-axis→R10 · 0.5–1.5 (fiscal channel) | α₄₁ · R10→next-period climate-exposure · 0.8–1.3 Product Π αᵢⱼ > 1.0 under scenario C parameterisation → cascade self-amplifies Scenario C · compound-shock · Buldyrev et al. (2010) interdependent-network cascade · per-NUTS-3 resolution · CC BY-SA 4.0

§4.4 · Scenario comparison

The three scenarios separate sharply at year three (2029). Measured as the divergence-corridor width between the top-quartile Mezzogiorno NUTS-3 cohort and the bottom-quartile central-northern reference cohort on the W-axis-weighted R10 surface, the year-three readings on the modelled scenario surface are approximately 0.42 under compound-shock, approximately 0.31 under PNRR-continuity, and approximately 0.24 under sustained-investment. Expressed relative to the 2025 baseline reading, the three central-tendency projections correspond to approximately +35% / 0 / −25% at year three. For comparison the 2017–2020 convergence-window minimum reads approximately 0.27 in the modelled backcast record. All forecast values are model-derived scenario outputs under the N = 5,000 Monte Carlo draws described in §1.5; they are not predictions and remain subject to canonical-register binding and sensitivity re-run prior to a v1.1 chapter-open amendment.

The three scenarios also separate on the compound-stressor cluster size: at year three the modelled cluster reads approximately 980 comuni under compound-shock, 770 under PNRR-continuity (unchanged by construction from the 2025 baseline), and approximately 680 under sustained-investment. At year five (2031) the modelled readings are on the order of 1,100–1,200 / 770 / 600–650. The near-stationary behaviour of the PNRR-continuity cluster size is the reference finding against which the two perturbation scenarios are measured: the baseline sub-system equilibrium absorbs its own operation on the modelled surface, and only a coordinated four-component perturbation (Scenario B) or a compound sub-system shock (Scenario C) moves the cluster count off its 770 near-equilibrium configuration.

Year three is where the three scenarios separate sharply enough to be distinguished statistically. It is also, by construction, the approximate drawdown horizon for the current PNRR Missione 5 envelope and the approximate implementation horizon for the proposed 2028–2034 EU Cohesion Policy cycle. The reader is invited to consider what the simultaneity of the statistical-separation window and the policy-decision window implies for intervention design.

§5 · Generalisability · four OECD cohort-jurisdictions

The Mezzogiorno pattern is a specific instance of a sub-national-divergence pattern that reads across OECD cohort-jurisdictions at per-LAU-2 granularity when the modifier surface is held constant. This section reads the pattern at four jurisdictional scales — Italian (the reference case), Spanish, French, German — rhyming not stacking.

§5.1 · Italy (reference case · re-read as cohort baseline)

Italy is the reference case the brief develops in §1–§4: a 7,901-comune cohort on the ISTAT LAU-2 register, a compound-stressor cluster of 770 comuni at threshold R10 ≥ 0.62, a backcast corridor divergent from 2003 through 2016 with a measurable 2017–2020 convergence window, a PNRR-active regime 2021–2025 extending partial convergence, and three forecast scenarios separating sharply at year three. The pattern reads as sub-national rather than national-aggregate: the per-NUTS-3 variation within the Mezzogiorno is large, the per-comune variation within a single NUTS-3 unit is larger, and the aggregate NUTS-2 statistic obscures both. The reference-case finding the brief carries forward to the three generalisability cases is that the pattern is a per-LAU-2 property of the four-component sub-system, not an artefact of NUTS-2 aggregation.

§5.2 · Spain · Andalucía + Extremadura

The Spanish reading is run on the INE LAU-2 register of approximately 8,100 municipios (the 8,131 figure is illustrative T2 · Modelled and is not a register-reconciled count; it stands pending reconciliation against the INE municipio register at Spanish cohort binding). On the modelled Spanish surface the equivalent-threshold compound-stressor cluster reads on the order of a few hundred municipios under threshold R10 ≥ 0.62 applied uniformly across the Spanish cohort; the Spanish cross-jurisdictional binding is pending and the quantitative value is reported as pattern-consistent-with-Italian-reading rather than as a bound measurement. The cluster concentrates in Andalucía (provincias of Jaén, Córdoba, Almería) and Extremadura (both provincias) with secondary concentration in the Serranía de Cuenca and parts of Castilla–La Mancha. The 2003–2025 backcast reads similarly to the Italian: divergent corridor 2003–2016 (overlapping the Spanish financial crisis cycle on a different fiscal-transmission substrate), a partial convergence window 2017–2020 associated with the Fondos Europeos Next Generation uptake, and a 2021–2025 PNRR-equivalent regime (the Spanish Plan de Recuperación, Transformación y Resiliencia). The sub-system reads consistently with the Italian case: political-economy regional asymmetry rooted in historical capacity differences, fiscal-transmission place-dependence, talent-flow out-migration to Madrid–Barcelona corridors, infrastructure-stock depreciation differentials. The W-axis records a different regulatory-mesh density profile (five instrument classes vs the Italian 5/1/1/3 distribution) but the three-modifier coupling operates with the same structural signature.

§5.3 · France · Occitanie + Hauts-de-France

The French reading is run on the INSEE commune register of approximately 34,900 units, with the cohort restricted to metropolitan France (excluding overseas départements and territoires). On the modelled French surface the compound-stressor cluster reads in the low thousands of communes under threshold R10 ≥ 0.62 — larger in absolute terms than the Italian or Spanish cases, consistent with the finer per-commune granularity of the French LAU-2 register. The French cross-jurisdictional binding is pending; the quantitative value is reported as pattern-consistent-with-Italian-reading rather than as a bound measurement. The cluster distributes bimodally, with one concentration in the rural interior of Occitanie (Aveyron, Lozère, Gers départements) reading as a south-western rural compound-stressor pattern, and a second concentration in Hauts-de-France (Nord, Pas-de-Calais, Aisne, Somme) reading as a post-industrial compound-stressor pattern. The French case is the brief's clearest instance of two sub-national-divergence patterns operating in the same jurisdiction on structurally distinct substrate conditions. The sub-system reads consistently in each: the regulatory-mesh structure differs by région, the fiscal-transmission routes differ by sub-system path, but the compound-stressor configuration is the same.

§5.4 · Germany · East-West divide

The German reading is run on the DESTATIS Gemeinde register of approximately 10,790 units. On the modelled German surface the compound-stressor cluster reads on the order of around a thousand Gemeinden under threshold R10 ≥ 0.62. The German cross-jurisdictional binding is pending; the quantitative value is reported as pattern-consistent-with-Italian-reading rather than as a bound measurement. The cluster concentrates in Sachsen-Anhalt, Mecklenburg-Vorpommern, Brandenburg (eastern portions), and Thüringen, with a distinct western cluster in rural Nordrhein-Westfalen (parts of Sauerland, Eifel) and Saarland. The German case is the most pronounced instance of sub-system path-dependence in the four-jurisdiction atlas: the post-1990 reunification trajectory operates on a different time horizon than the Italian post-1861 unification trajectory, but the four-component sub-system reads structurally equivalently. The talent-flow component is the most distinctive German reading: the post-1990 west-ward out-migration cohort records a per-Gemeinde substrate effect that remains measurable on the R10 surface more than three decades later, consistent with the Italian Alto Sele valley finding on a different temporal substrate.

§5.5 · The generalisable pattern

FIGURE 9 · FOUR-JURISDICTION ATLAST2 · Modelled

Four-jurisdiction rhyming-scales atlas · per-LAU-2 R10 surface · IT · ES · FR · DE

Four-panel atlas · per-LAU-2 R10 surface for Italy + Spain + France + Germany on shared colour scale · compound-stressor cluster in each jurisdiction highlighted · cluster-count annotations

Figure 9 · Four-jurisdiction rhyming-scales atlas · per-LAU-2 R10 surface · IT · ES · FR · DE Four panels, one per jurisdiction, drawn from Eurostat GISCO NUTS 2024 at 1:1M. Country outline in grey; regions named in the brief as carrying the compound-stressor concentration filled in cayenne, with qualified regions in a lighter tone. Overseas territories are excluded so each mainland reads at a usable scale. Cluster counts outside Italy are modelled, not register-reconciled. LAU-2 R10 surface · real NUTS 2024 geometry (Eurostat GISCO 1:1M) · threshold R10 ≥ 0.62 applied uniformly · modelled atlas Italy · reference case ISTAT LAU-2 · 7,901 comuni · cluster n=770 Mezzogiorno · Campania · Puglia · Calabria · Sicilia Spain INE LAU-2 · 8,131 municipios · cluster n≈690 Andalucía · Extremadura · interior Castilla France INSEE LAU-2 · ~34,900 communes · cluster n≈2,840 Hauts-de-France (post-industrial) · Occitanie (rural) Germany DESTATIS LAU-2 · ~10,790 Gemeinden · cluster n≈1,050 Sachsen-Anhalt · MV · Brandenburg · Thüringen · rural NRW · Saarland LAU-2 base surface compound-stressor concentration high concentration · R10 ≥ 0.62 Geometry © EuroGeographics

The four-jurisdiction atlas reads the Mezzogiorno pattern not as an Italian pathology but as a European-cohort signature: the four-component sub-system §3.2 described operates with the same structural coupling in Italy, Spain, France, and Germany, each case anchored on the historical trajectory and political-economy substrate particular to its jurisdiction. The Italian compound-stressor cluster count at 770 is bound to the 2 November 2026 ISTAT LAU-2 reconciliation; the Spanish, French, and German cluster counts are reported on the modelled cross-jurisdictional surface pending independent register-reconciliation at each national cohort. The per-LAU-2 threshold R10 ≥ 0.62 is held constant across all four cohorts by construction. The sub-national-divergence pattern is not a national accident; it is a persistent configuration that the European sub-national-infrastructure substrate produces where the four-component sub-system reaches its fixed-point.

The policy implication runs with the scaling. If the pattern were national, national-level fiscal instruments would be the appropriate response. Because the pattern is per-LAU-2 and sub-system-conditioned, the response that moves the fixed-point is place-based in the sense Barca (2009) framed, scaled across a four-cohort European policy space, and coordinated across the three modifier surfaces the §3.1 feedback loop measures. The brief's §6 implications read against this finding.

FIGURE 10 · PER-REGIONE CLUSTER-DENSITY INDEXT2 · Modelled

Per-jurisdiction cluster sizes + thresholds · four-cohort comparison

Horizontal bar chart · per-regione cluster-density index (share of cluster comuni within total comuni per regione) · Mezzogiorno regioni rendered in cayenne · Central/Northern regioni in steel · national-average reference line at 9.7%

Figure 10 · Per-NUTS-2 cluster-density index · Italian regioni · 2026 snapshot Horizontal bar chart of per-regione cluster-density index, defined as the share of compound-stressor cluster comuni to total comuni in each regione. Mezzogiorno regioni rendered in cayenne. Central/Northern regioni rendered in steel. Horizontal reference line at the national average cluster share. Cluster-density index = (cluster comuni / total comuni) per NUTS-2 · national average = 9.7% · modelled ordering 0% 10% 20% 30% 40% 50% cluster-density index national average 9.7% Calabria42% Basilicata37% Sicilia34% Campania30% Puglia27% Molise22% Abruzzo19% Sardegna18% Umbria8% Marche7% Lazio6% Liguria5% Piemonte5% Toscana4% Valle d'Aosta3% Veneto3% Lombardia2% Emilia-Romagna2% Friuli · Trentino1% Mezzogiorno regioni (above national average) Central/Northern regioni (at or below national average)

§6 · Where the measurement could be useful

Scope note. This section describes places where the per-LAU-2 surface could plausibly be read into decision-making frameworks that already exist. It is not a policy prescription. The brief does not recommend that any named institution adopt the SSI Index v4.2 methodology as an instrument of its own, does not propose that any specific programming envelope be sized differently, and does not advocate any intervention option over another. The institutional cross-walks that follow report where the measurement could be read into existing reporting cycles; the question of whether any of them should be is left to the institutions themselves, and the brief does not have standing to answer it. Readers who prefer the brief to stop at §5.5 may do so without loss of its substantive claim.

§6.1 · Where the measurement could be read by Italian national institutions

The per-LAU-2 surface permits investment-targeting and regulatory-audit functions to operate at the granularity at which the feedback-loop actually generates divergence. Five Italian national institutions carry the measurement load this brief reads against.

MASE (Ministero dell'Ambiente e della Sicurezza Energetica). The Strategia Nazionale di Adattamento ai Cambiamenti Climatici (SNAC) is the Italian national adaptation-planning anchor. The per-LAU-2 R6 surface reads directly against the SNAC's regional-vulnerability framework; the brief's three-scenario forecast fan maps onto the SNAC's adaptation-trajectory envelopes at per-NUTS-3 resolution. The SNAC regional-vulnerability clusters correspond at NUTS-2 to the brief's cluster-density index described in §6.2.

MEF (Ministero dell'Economia e delle Finanze). The Documento di Economia e Finanza (DEF) carries the fiscal-transmission framework that the compound-shock scenario (§4.3) operationalises. The per-NUTS-3 investment-rate trajectory in the backcast reads against DEF's historical fiscal-stance annotations through the four investment regimes identified in §2.2 (pre-crisis 2003–2008, austerity 2009–2016, PNRR-precursor 2017–2020, PNRR-active 2021–2025).

ARERA (Autorità di Regolazione per Energia Reti e Ambiente). The W-axis regulatory audit-state mesh anchors on ARERA's per-regione regulatory substrate. The Rapporto Annuale 2025 provides the baseline against which the brief's W-axis reading at per-comune granularity is calibrated — the 5/1/1/3 instrument-class distribution used across the brief rests on ARERA's per-regione regulatory-mesh density as its reference.

Cassa Depositi e Prestiti (CDP). The infrastructure-investment register maintained by CDP is the operational surface against which per-LAU-2 investment-targeting reads at the ground level. The brief's sustained-investment scenario (§4.2) maps onto CDP's identified pipeline plus the PNRR absorption trajectories reported in CDP's 2025 Piano Industriale; the convergence at year three described in §4.4 overlaps the CDP pipeline-drawdown horizon.

European Investment Bank (EIB). The EIB's Italian infrastructure portfolio operates adjacent to CDP, with specific weight in southern-regione project financing. Per-LAU-2 visibility supports ex-ante project screening against the feedback-loop mechanism this brief documents; the EIB southern-regione commitments over the 2021–2025 PNRR-active regime read against the §2.3 convergence-window observation on Napoli and Bari.

§6.2 · Where the measurement could be read by Regioni-level leadership

The pattern reads differently at per-NUTS-2 (Regione), per-NUTS-3 (provincia), and per-LAU-2 (comune) granularities, and the three readings support distinct decision-classes at each level. The Regione-level reading is not simply a coarser version of the per-comune reading; it is a distinct view that is diagnostic for a specific decision-class — the coordination of the four sub-system components across the comuni within a regione's jurisdictional envelope.

At per-LAU-2 granularity, the compound-stressor cluster reads as 770 comuni with a geographic distribution that crosses nine regioni. At per-NUTS-3 granularity, the cluster reads as a concentration pattern — fifteen provincie across the nine regioni carry more than fifty cluster comuni, and five provincie carry more than one hundred. At per-NUTS-2 granularity, the cluster reads as a Regione-level cluster-density index that is directly comparable to the cohesion-policy regional-development instruments the regioni administer. The three readings are complementary: the per-LAU-2 reading identifies intervention targets, the per-NUTS-3 reading identifies coordination clusters, and the per-NUTS-2 reading identifies the policy-envelope level at which the four-component sub-system can be perturbed.

The Regione-level reading of the compound-stressor cluster places particular weight on Campania, Puglia, Calabria, and Sicilia as the primary cluster-density regioni, with Basilicata, Molise, and Abruzzo recording secondary density and Sardegna recording a distinct pattern conditioned on insular-infrastructure specificity. The brief does not prescribe per-regione intervention priorities — the four-component sub-system operates at per-comune resolution, and per-regione aggregation obscures the within-regione variance. The Regione-level reading supports the coordination function: the W-axis regulatory-mesh instruments that operate at regione level (ARERA regional regulatory activity, regione-level PNRR Missione implementation) can be calibrated against the per-regione cluster-density index, and the within-regione per-comune heterogeneity can inform the implementation sequence.

§6.3 · Where the measurement could be read by DG REGIO · Cohesion Policy 2028–2034

The Cohesion Policy 2028–2034 programming cycle is the next window in which the per-LAU-2 surface this brief reports can operate at scale across cohort-jurisdictions. Three observations follow from the §5 generalisability reading.

First, the per-LAU-2 surface restores the Barca 2009 place-based principle at the measurement level. The Barca Report on the reformed Cohesion Policy argued that place-based policy requires a granularity that matches the mechanisms producing regional divergence. The per-LAU-2 modifier surface this brief reports is that granularity; it operates at the administrative scale where fiscal transmission, climate-physical exposure, and regulatory audit-state coexist on a single surface. Cohesion Policy 2028–2034 instruments that read against per-LAU-2 can target the feedback-loop mechanism directly, rather than at aggregated proxies that compress it.

Second, the four-jurisdiction rhyming-scales reading (§5) is cohort-comparable at equivalent granularity. The Italian cohort reads at per-comune. The Spanish cohort reads at per-municipio. The French cohort reads at per-commune. The German cohort reads at per-Gemeinde. All four LAU-2 cohorts resolve to the same SSI Index modifier surface with cluster thresholds that scale consistently. This cross-jurisdiction comparability supports Cohesion-fund allocation decisions that respect place-based specificity while reading on a single analytical surface.

Third, the methodology is open-licensed. The SSI Index v4.2 methodology is published under CC BY-SA 4.0; the R6 climate-physical framework is peer-reviewed in JIPR 7(1):35 (doi:10.1186/s43065-026-00193-z) and ERE 3(3):035007 (doi:10.1088/2753-3751/ae87a5), while R10 and W-axis conventions remain operationalised in-brief pending independent peer-review anchors. DG REGIO's own analytical teams, Member State managing authorities, and the European Court of Auditors can replicate the per-LAU-2 reading independently against their own canonical registers. The brief does not propose that DG REGIO adopt the SSI Index as policy instrumentation; it reports that the per-LAU-2 surface exists as a public open-science artefact.

The specific 2028–2034 programming instruments on which the per-LAU-2 surface reads directly are the European Regional Development Fund (ERDF), the Cohesion Fund, the Just Transition Fund, and the Social Climate Fund. Each operates on envelopes that the Member State managing authorities program down to NUTS-2 or NUTS-3 units; none at present resolves to per-LAU-2. The brief observes, without prescribing, that the per-LAU-2 surface the SSI Index v4.2 methodology produces could be made available to the managing-authority programming cycle as input instrumentation, and that the four-jurisdiction cohort-comparability property would support ex-ante cross-Member-State calibration of cluster-density indicators prior to envelope allocation. Whether DG REGIO, the Member States, or the European Parliament would find this instrumentation useful is a reader-inference question that the brief declines to answer on the readers' behalf.

§6.4 · Where the measurement could be read by civic-society and academic researchers

Three data registers support independent replication of the readings this brief reports. All three are published under CC BY-SA 4.0 at the publication URL ikenga.eu/reports/mezzogiorno-pattern-2026-11.html.

  1. The 770-comune compound-stressor cluster. Per-comune aggregates of R10 and R6 modifiers for the cluster, as of the 2026 snapshot. Source: SSI Index v4.2 canonical register.
  2. The four-jurisdiction per-LAU-2 R10 surface. Aggregated at per-NUTS-3 for Italy, Spain, France, Germany. Supports independent replication of the generalisability claim in §5.
  3. The forecast-fan trajectory. Per-NUTS-3 per-scenario trajectory 2026–2031 under PNRR-continuity, sustained-investment, and compound-shock parameterisations. Supports independent replication of the §4 reading.

Full per-comune raw trajectories across the 7,901-comune Italian cohort are held in the SSI Index canonical data register. That register is not currently published at a public URL, so the trajectories are available for academic research use on request at c.berard@ikenga.eu rather than by download. The ISTAT LAU-2 2026 register, Eurostat regional accounts, ISTAT BES indicators, and OIPE energy-poverty indicators used as socioeconomic substrate are public under their respective issuing-body licences.

What this brief does not claim

  1. It does not claim that R10 is measured at comune level. The surface is modelled from socioeconomic substrate whose resolution is recorded, and disputed, in the register note at §1.2. Every figure resting on it carries a T2 · Modelled badge.
  2. It does not claim peer review for R10 or for the W-axis. Both are first operationalised in this brief and in SB-01 and SB-02. The two peer-reviewed anchors cited (JIPR 7(1):35 and Environmental Research: Energy 3(3):035007) anchor the R6 climate-physical modifier and the v4.0.2 canonical register, not R10 and not the W-axis.
  3. It does not claim the Spanish, French and German cluster counts are register-reconciled. Only the Italian cohort is bound to a national register reconciliation. The other three are read on the modelled cross-jurisdictional surface.
  4. It does not claim the backcast or the forecast fan are measurements of outcomes. The 2003–2025 corridor is a reconstruction over the modifier surface and the 2026–2031 fan is a scenario projection; neither is an observed outage, cost or investment record.
  5. It does not claim that the SSI Index's published per-substation distributive-justice field is the surface used here. That is a separate quantity, published at a different resolution, and it is not interchangeable with this brief's surface.

Conclusions

"Il Mezzogiorno è l'Italia, ma l'Italia non è tutto il Mezzogiorno." — Gaetano Salvemini
"La questione meridionale è la questione nazionale in uno dei suoi aspetti più antichi." — Antonio Gramsci, Alcuni temi della quistione meridionale, 1926

The brief reports what the per-LAU-2 surface records across the 7,901-comune Italian cohort: a structural divergence pattern with a 770-comune compound-stressor cluster at the R10 ≥ 0.62 threshold, a 22-year backcast of widening per-NUTS-3 divergence under three of four investment regimes, a 5-year forecast fan that separates sharply at year three under three scenarios, and a structural feedback-loop mechanism coupling R10, R6, and the W-axis at per-comune granularity.

The brief also reports what the per-LAU-2 surface records across three further OECD cohort-jurisdictions: the same pattern, at equivalent granularity, with the modifier surface held constant.

The brief does not conclude whether and how Italy, Spain, France, or Germany should intervene. The per-LAU-2 surface is instrumentation. What it measures is the pattern. The conclusions are the reader's to draw.

Back matter

Chapter-open invitation

The §5 generalisability chapter is structured to admit a post-ship v1.1 chapter amendment co-authored with peer-research institutions approached for that role. On the cross-jurisdictional side the invitation is extended to ISPI (Istituto per gli Studi di Politica Internazionale, Milan), Real Instituto Elcano (Madrid), IFRI (Institut français des relations internationales, Paris), and SWP (Stiftung Wissenschaft und Politik, Berlin). On the Mezzogiorno side — equally weighted in the amendment mechanism — the invitation is extended to SVIMEZ (Associazione per lo Sviluppo dell'Industria nel Mezzogiorno, Roma), Università degli Studi di Napoli Federico II, CNR-IRISS (Istituto di Ricerca su Innovazione e Servizi per lo Sviluppo, Napoli), and Università della Calabria (Rende). The §5 cross-jurisdictional amendment and the §3 Mezzogiorno case-anchor amendment are tracked as symmetric mechanisms; institutional response on either axis is acknowledged in the v1.1 record. The brief treats the Southern-institution and Northern-institution amendment paths as co-equal in weight, not as a courtesy addition.

Methodology version pin

SSI Index v4.2. Peer-reviewed R6 anchors: Bérard (2026) Journal of Infrastructure Preservation and Resilience 7(1):35, doi:10.1186/s43065-026-00193-z; Bérard (2026) Environmental Research: Energy 3(3):035007, doi:10.1088/2753-3751/ae87a5. R10 and W-axis conventions operationalised in-brief; peer-review anchors in preparation. Any v4.3 refinements Q4 2026 absorbed at ship cut.

Limitations

The brief carries a number of limitations the reader should hold in view. First, the R10 distributive-justice modifier and the W-axis regulatory audit-state mesh are first operationalised in this brief and in the prior SB-01/SB-02 Strategic Briefs; independent peer-review anchors for both are in preparation but do not yet exist. The brief's substantive claim rests on an R6-anchored peer-reviewed framework extended by two in-brief constructs, and the reader is invited to weight the extensions accordingly. Second, the illustrative numerical values in Figures 1–10 and in the §2.1, §2.2, §4.4, and §5.2–§5.4 prose are current modelled-surface readings, not canonical v4.2 register binds; the final bind is scheduled for the v1.1 chapter-open amendment. Third, the three-modifier coupling (§3.1) and the four-component sub-system (§3.2) are conceptual framings supported by the empirical surface but are not themselves measurement instruments; the Buldyrev et al. (2010) cross-network cascade framework informs the coupling-structure specification rather than supplying it as a closed-form. Fourth, the three-scenario forecast (§4) is scenario-analytic, not predictive; the compound-shock cascade specification in §4.3 is informed by reference events (2022 Po Valley drought, 2023 Emilia-Romagna floods, IPCC AR6 Mediterranean projection under SSP2-4.5) but the per-step amplification coefficients are sensitivity-dependent and are reported with the Monte Carlo band rather than as point estimates. Fifth, the four-jurisdiction generalisability (§5) is reported on the modelled cross-jurisdictional surface; quantitative binding against the INE (Spanish), INSEE (French), and DESTATIS (German) canonical registers is pending. Sixth, no consultation or review was sought with the three case-anchor comuni (Scanzano Jonico, Laviano, Partanna) or with their regioni prior to v1.0 publication. The per-comune characterisations rest on public-register data and the published historical record. Concurrent with the v1.0 Zenodo deposit the author is extending a pre-circulation courtesy note to the three Sindaci and to the analytical offices of Regione Basilicata, Regione Campania, and Regione Siciliana, with an open invitation to lodge observations, corrections, or objections through the v1.1 chapter-open amendment mechanism; the Mezzogiorno-institution axis of that mechanism (SVIMEZ · Federico II · CNR-IRISS · Università della Calabria) is activated on the same cycle. Any reply, including a reply declining to engage, will be acknowledged in the v1.1 record. A fuller pre-publication consultation process is adopted as the standing protocol for future Strategic Briefs in the series. Seventh, the brief is single-author lead; the post-ship v1.1 chapter-open amendment mechanism is the brief's mode of inviting independent scrutiny.

Conflict of interest

The author is SSI Index Coordinator at Ikenga SL and leads the SSI Index methodology that this brief operationalises. Ikenga SL maintains no commercial contract conditioned on any outcome of this brief and retains no proprietary interest in any policy outcome the brief may influence. The brief is self-funded; no public grant, consultancy, or client relationship conditioned the content. The Fondazione SSI Index is in-establishment (Naples DPR 361/2000) and the brief does not generate revenue for Ikenga SL or the Foundation. The two peer-reviewed anchor papers are authored by the same person; the reader is invited to weight that fact in interpreting the brief's methodological independence.

Review process

The brief was author-reviewed under a five-pass internal protocol (structural coherence, factual integrity, register discipline, acknowledgements and citation integrity, externalisation scrub), with an independent critical-reviewer pass applied post-draft; the findings of that critical-reviewer pass are reflected in the current cut. Post-ship, the v1.1 chapter-open amendment mechanism remains open on two co-equal axes: the cross-jurisdictional axis — ISPI (Milan), Real Instituto Elcano (Madrid), IFRI (Paris), SWP (Berlin) — and the Mezzogiorno axis — SVIMEZ (Roma), Università Federico II (Napoli), CNR-IRISS (Napoli), Università della Calabria (Rende). Amendments landed under either axis will be reflected in a v1.1 Zenodo deposit.

Data availability

Open-licence (CC BY-SA 4.0) data subset available in data/ at ikenga.eu/reports/mezzogiorno-pattern-2026-11.html: (i) 770-comune compound-stressor cluster per-comune aggregates; (ii) four-jurisdiction per-LAU-2 R10 surface at NUTS-3 aggregation; (iii) forecast-fan trajectory per-NUTS-3 under three scenarios. Full per-comune raw trajectories are held in the SSI Index canonical register and are available on request; they are not currently published at a public URL.

Citations

Full citation list at citations/sources_v1.0.md and footnoted inline throughout the brief. Reference-integrity instrument machine-check output at citations/integrity_check_v1.md .

Author

Cedric Bérard is SSI Index Coordinator at Ikenga SL, Madrid. He leads the SSI Index methodology development and the open-science publication programme that frames this brief, in partnership with the Fondazione SSI Index (in establishment, Naples DPR 361/2000). His prior reports for the SSI Index Strategic Brief series include SB-01 State of OECD Grid Adaptation Intelligence 2026 (July 2026), SB-02 The Verification Standard for Asset-Level Adaptation Intelligence (October 2026), and the X1 themed analysis Compute Sovereignty and the 4IR Civil-Society Surface (October 2026). Contact: c.berard@ikenga.eu.

References

Thirty-four entries across six source classes, reproduced from the brief's own citation register (citations/sources_v1.0.md). In-text citations use author–date short form; this list resolves them. Two further register entries record the R10 modifier and the W-axis mesh as having no peer-review anchor and are stated as such in §1.2, §1.6 and in What this brief does not claim.

  1. Bérard, C. (2026). Markov degradation modelling for fleet-scale substation preservation. Journal of Infrastructure Preservation and Resilience 7(1):35 · doi:10.1186/s43065-026-00193-z. [R6 climate-physical modifier anchor; v4.0.2 canonical register] Cited at §1.2, §1.6 definition pin · Exec summary.
  2. Bérard, C. (2026). Multi-Hazard Environmental Risk Assessment for Electricity Substations. Environmental Research: Energy 3(3):035007 · doi:10.1088/2753-3751/ae87a5. [R6 environmental-hazard anchor; v4.0.2 canonical register; Pearson ρ 0.72 against documented failure record] Cited at §1.2, §1.6 definition pin · §4 forecast.
  3. European Union · Regulation (EU) 2021/1119 establishing the framework for achieving climate neutrality (European Climate Law) · Article 5 (adaptation) Cited at §6.3 DG REGIO policy implications.
  4. European Commission · EU Adaptation Strategy 2050 · COM(2021) 82 final · adaptation-trajectory framework Cited at §6 policy implications.
  5. European Union · Regulation (EU) 2021/1058 (ERDF) + 2021/1057 (ESF+) · Cohesion Policy 2021-2027 · place-based adaptation funding Cited at §6.3 DG REGIO.
  6. European Union · Directive (EU) 2022/2557 on the resilience of critical entities (CER Directive) Cited at §3.3 case anchors · W-axis regulatory substrate.
  7. European Sustainability Reporting Standards (ESRS) E1 · Climate change · adaptation disclosure framework Cited at §6.1 Italian national institutions · SSI-ENN cross-reference.
  8. Repubblica Italiana · Piano Nazionale di Ripresa e Resilienza (PNRR) · 2021 adoption · Decreto 77/2021 · Missione 2 (Rivoluzione verde e transizione ecologica) Cited at §2.2 investment regime · §4.1 PNRR-continuity scenario.
  9. MASE (Ministero dell'Ambiente e della Sicurezza Energetica) · Strategia Nazionale di Adattamento ai Cambiamenti Climatici (SNAC) · 2023 adoption Cited at §6.1 Italian national institutions.
  10. MEF (Ministero dell'Economia e delle Finanze) · Documento di Economia e Finanza (DEF) · fiscal-transmission framework Cited at §4.3 compound-shock scenario.
  11. ARERA (Autorità di Regolazione per Energia Reti e Ambiente) · Rapporto Annuale 2025 · W-axis regulatory substrate at per-regione granularity Cited at §1.2 definition · §3.3 Partanna case anchor.
  12. ISTAT (Istituto Nazionale di Statistica) · Benessere Equo e Sostenibile (BES) · 2025 edition · per-territorio substrate indicators Cited at §1.1 cohort construction · §3.2 feedback-loop substrate.
  13. OIPE (Osservatorio Italiano della Povertà Energetica) · Rapporto 2025 · per-comune energy-poverty indicators Cited at §3.2 feedback-loop · §6.4 civic-society reference.
  14. Cassa Depositi e Prestiti (CDP) · Rapporto Annuale 2025 · infrastructure-investment register Cited at §6.1 Italian national institutions.
  15. ISTAT LAU-2 2026 register · Italian comuni canonical register Cited at §1.1 cohort construction.
  16. Buldyrev, S.V., Parshani, R., Paul, G., Stanley, H.E., Havlin, S. (2010). Catastrophic cascade of failures in interdependent networks. Nature 464, 1025–1028. doi:10.1038/nature08932 Cited at §4.3 compound-shock cross-shock amplification.
  17. Pastor, L., Veronesi, P. (2013). Political uncertainty and risk premia. Journal of Financial Economics 110(3), 520-545 Cited at §3.2 feedback-loop (cost-of-capital substrate · lightly) · prepares C1 Dec 2026 report anchor.
  18. Reckien, D. et al. (2023). Maladaptation typology · Nature Climate Change doi:10.1038/s41558-023-01774-6 Cited at §1.2 definition pin · §3.1 feedback-loop coupling.
  19. Payo, A. et al. (2016). Adaptation planning · Nature Climate Change · the adaptation-literature baseline Cited at §1.2 methodology novelty claim.
  20. Barca, F. (2009). An Agenda for a Reformed Cohesion Policy: A place-based approach to meeting European Union challenges and expectations (Barca Report) · European Commission Cited at §5 generalisability · §6.3 DG REGIO policy.
  21. Bagnasco, A. (1977). Tre Italie: la problematica territoriale dello sviluppo italiano · Il Mulino · the structural-divergence frame Cited at Introduction framing · §3.2 feedback-loop.
  22. Krugman, P. (1991). Geography and Trade · the new-economic-geography foundation for sub-national-divergence patterns Cited at §5 generalisability.
  23. Rodríguez-Pose, A. (2018). The revenge of the places that don't matter. Cambridge Journal of Regions, Economy and Society 11(1), 189-209 Cited at §5 generalisability · §6 policy implications.
  24. SSI Index · SB-01 · State of OECD Grid Adaptation Intelligence 2026 · July 2026 · §2.1 Mezzogiorno-trap diagnostic Cited at Introduction framing · §3 full-depth development.
  25. SSI Index · F-01 · The 2026 European heatwave reads on the v4.2 modifier surface · July 2026 Cited at §1.4 backcast + §3.1 R6 coupling.
  26. SSI Index · F-02 · Heatwave second edition · September 2026 Cited at §3.1 R6 coupling.
  27. SSI Index · P5 ERE Fiscal Maladaptation · July 2026 · peer-reviewed R10 anchor Cited at §1.2 definition · §3.2 feedback-loop · §4.3 compound-shock.
  28. SSI Index · P7 ERE · August 2026 · peer-reviewed methodology anchor Cited at §1.2 definition pin.
  29. SSI Index · TA-01 · Nodal Locational Price Discovery · July + November 2026 editions Cited at §1 methodology pattern reference.
  30. SSI Index · SB-02 · The Verification Standard for Asset-Level Adaptation Intelligence · October 2026 Cited at §1 methodology · controls-register register inheritance.
  31. ISPI (Istituto per gli Studi di Politica Internazionale) · Sub-national-divergence + Italian political-economy publication series · 2024-2026 Cited at §5.2 ES case · §5.3 FR case · §5.4 DE case — chapter-open invitation target.
  32. Real Instituto Elcano · Sub-national-divergence + Iberian political-economy publications · 2024-2026 Cited at §5.2 ES case — chapter-open invitation target.
  33. IFRI (Institut français des relations internationales) · Centre d'études géopolitiques des régions · sub-national divergence France publications · 2024-2026 Cited at §5.3 FR case — chapter-open invitation target.
  34. SWP (Stiftung Wissenschaft und Politik) · East-German-reunification + sub-national-infrastructure publications · 2024-2026 Cited at §5.4 DE case — chapter-open invitation target.

Disclaimer

This brief is published by Ikenga SL in partnership with Fondazione SSI Index (in establishment, Naples DPR 361/2000). The content reports the per-LAU-2 surface the SSI Index v4.2 methodology produces on the modelled working surface as of November 2026; quantitative values flagged illustrative in Figures 1–10 and in the relevant prose (see §1.6) are pending final binding to the canonical v4.2 register at v1.1 chapter-open amendment. The brief does not constitute investment advice, legal opinion, financial recommendation, or guarantee of outcome. The three forecast scenarios in §4 are scenario-analytic projections conditioned on the parameterisations stated in §1.5, not predictions, and separate sharply at year three (2029) by construction. Readers acting on the brief do so under their own responsibility. The methodology is CC BY-SA 4.0 open-licensed, and independent replication is invited — replication material is supplied on request at c.berard@ikenga.eu, the public deposit being still to come. Ikenga SL retains no proprietary interest in any policy outcome the brief may influence. For correspondence on factual errors, methodological questions, or replication requests, c.berard@ikenga.eu.