Public Event Reconstruction
Bab-el-Mandeb: How a Chokepoint Shock Propagates
A reconstruction of how disruption at a strategic maritime chokepoint propagates through routing, transit time, fleet utilisation, freight capacity, inventory requirements, working capital, and downstream availability.
Methodological disclaimer
This reconstruction is an analytical demonstration using publicly available information. It is not a client engagement and not a record of contemporaneous advice. The event chronology is drawn from public reporting; the analytical outputs are a reconstruction of method, not a historical observation. Where a value is illustrative it is labelled as such.
Decision support
The decision this reconstruction bears on
The decision this serves
- Big bet · Argue the case against before committing: alternatives, what must be true, and what would make this wrong.
- Complex · Probe, sense, respond: define the probe, its stopping condition, and what would change the reading.
- Who faces it
- Importers, manufacturers and the lenders exposed to the Europe–Asia container string.
- The choice
- Whether to hold additional inventory at destination, diversify the routing mix, or re-sequence production before the next disruption.
- Standing
- The reading is available on public evidence. The exposure it implies is not quantified here, and no figure is asserted that the public record does not carry.
The case against this reading
If carriers restore effective capacity faster than this reconstruction assumes — through reallocation, faster steaming or new tonnage entering the string — the schedule shock is shorter than modelled, and buffer held against it is cost without return. If destination cover is deeper than the public record shows, these propagation paths describe a system the reader is not in. Both cases are stated rather than argued away.
The question
What happens to downstream inventory, transit time, freight economics, and exposed industries when traffic through the Bab-el-Mandeb is materially disrupted?
Period
The reconstruction window runs from the onset of sustained disruption in late 2023 through the following twelve months.
Geography
The Bab-el-Mandeb and southern Red Sea, the Cape of Good Hope routing, European and Mediterranean destination nodes, and Asian origin ports.
Industries considered
- ·Container shipping and freight forwarding
- ·Retail and consumer goods importers
- ·Automotive and industrial manufacturing
- ·Energy and refined products
- ·Port and terminal operations
Inside the model
- Vessel routing and effective capacity
- Transit time and schedule reliability
- Freight pricing and war-risk insurance
- Inventory in transit and destination cover
- Working-capital timing
Outside the model
- Air freight as a full substitution path (treated as an indicator only)
- Domestic road and rail distribution inside destination markets
- Contractual and legal remedies between counterparties
- Second-order political consequences beyond trade flow
A boundary that is never written down cannot be challenged. These exclusions are a design decision, not an oversight.
The system
How the shock travels
Select any node to inspect its definition, its upstream and downstream relationships, the evidence behind it, and its limitations.
The modelled sequence
- Bab-el-Mandeb
- Red Sea Routing
- Transit Time
- Vessel Utilisation
- Available Freight Capacity
- Freight Economics
- Inventory Requirements
- Working Capital
- Regional Availability
- Downstream Business Effects
The chain is the structure of the model, not an observation of traffic. Each node carries its own evidence, confidence and limitations, shown in full below.
Schematic · Indicative geometry · Not a map · Not live
No vessel positions, current volumes or present conditions are shown. The corridor is drawn to make the model's structure legible, not to represent geography at scale.
The strait between the Horn of Africa and the Arabian Peninsula, carrying a large share of Europe–Asia container and energy traffic.
Upstream
— none —
Downstream
routing
Indicators
Transit volume against trailing baseline · Vessel targeting reports
Confidence
High
Evidence
Public maritime traffic reporting and published trade route data.
Limitation
Volume shares vary by source and by quarter; the model uses a range, not a point value.
The shock
What was observed, inferred, assumed, and unknown
These categories are never blurred. An inference presented as an observation is the most common failure in this kind of analysis.
Observed
- Sustained targeting of commercial shipping in the southern Red Sea from late 2023.
- Major container lines suspending Red Sea transits and routing via the Cape of Good Hope.
- Naval coalition escorts announced, with transits continuing for some operators.
Inferred
- That diversion would become the default routing for a sustained period rather than a short interruption.
- That effective capacity, not just transit time, would be the binding constraint.
Assumed
- That freight economics would transmit to inventory policy within one planning cycle.
- That destination markets would absorb variance through buffer rather than through service reduction.
Uncertain
- Duration of the disruption.
- Whether a durable escort regime would restore carrier confidence.
- The speed at which shippers would substitute sourcing rather than absorb cost.
Observation timeline
Targeting of commercial shipping in the southern Red Sea begins and is publicly reported.
- Source
- Public reporting
- Relation to prior assessment
- Shock onset — the initial condition of the reconstruction.
Major container lines suspend Red Sea transits; Cape of Good Hope routing is adopted on affected strings.
- Source
- Carrier service advisories
- Relation to prior assessment
- Confirms the routing layer of the model.
Sustained diversion persists; effective capacity absorption becomes the dominant commercial effect.
- Source
- Published freight indices and schedule data
- Relation to prior assessment
- Confirms the capacity layer; the model had ranked this second to transit time.
Air freight substitution is more limited than the model's substitution path implied.
- Source
- Published air cargo data
- Relation to prior assessment
- Challenges the substitution pathway.
Working-capital effects are visible in some sectors but not traceable at the resolution the model implied.
- Source
- Sector reporting
- Relation to prior assessment
- Unresolved — the layer could not be tested with public data.
Propagation paths
Four routes through the system
Each pathway states its evidence, its confidence, the indicator that would confirm it, and the condition under which it fails.
Capacity absorption
- Routing disruption
- Longer transit
- Reduced effective capacity
- Higher freight pressure
- Evidence
- Routing distance, service rotations, published freight indices.
- Confidence
- High
- Observable indicator
- Capacity absorbed as a share of deployed fleet.
- Possible failure mode
- New-build deliveries offset absorption faster than modelled.
Inventory in transit
- Longer transit
- More inventory in transit
- Higher working-capital requirement
- Inventory planning change
- Evidence
- Transit extension, stated payment terms (inferred).
- Confidence
- Low to moderate
- Observable indicator
- Days of goods in transit; destination cover by category.
- Possible failure mode
- Firms absorb the extension inside existing buffer without changing policy.
Lead-time variance
- Freight disruption
- Supplier lead-time variance
- Production planning uncertainty
- Downstream availability pressure
- Evidence
- Sector reporting on lead times.
- Confidence
- Low to moderate
- Observable indicator
- Lead-time variance on affected supplier lanes.
- Possible failure mode
- Substitution at the sourcing layer removes the pressure before it reaches production.
Sourcing substitution
- Regional exposure
- Alternative sourcing
- Supplier substitution
- Cost and lead-time change
- Evidence
- Trade flow data; published sourcing announcements.
- Confidence
- Low
- Observable indicator
- Share of affected volume re-sourced.
- Possible failure mode
- Substitution is slower and narrower than assumed — this is what occurred.
Indicators to watch
What would confirm or challenge the assessment
No live values are shown. Where an indicator cannot be observed with public data, it is marked as such rather than estimated.
Vessel routing
Rising- Current
- Elevated diversion
- Baseline
- Trailing 12-month norm
- Confidence
- High
- Timestamp
- Reconstruction window
Primary observable for the routing layer.
Transit time
Rising- Current
- Extended
- Baseline
- Pre-disruption schedule
- Confidence
- High
- Timestamp
- Reconstruction window
Drives capacity absorption.
Freight capacity
Tightening- Current
- Absorbed
- Baseline
- Deployed fleet norm
- Confidence
- Moderate
- Timestamp
- Reconstruction window
Binding constraint in the model.
Freight pricing
Rising- Current
- Elevated
- Baseline
- Trailing index
- Confidence
- Moderate
- Timestamp
- Reconstruction window
Transmits cost into landed cost.
War-risk insurance
Rising- Current
- Elevated
- Baseline
- Pre-disruption premium
- Confidence
- Moderate
- Timestamp
- Reconstruction window
A cost layer the model under-weighted.
Inventory cover
Unresolved- Current
- Not observed at firm level
- Baseline
- Seasonal norm
- Confidence
- Low
- Timestamp
- —
Cannot be tested with public data.
Supplier lead times
Unresolved- Current
- Variable
- Baseline
- Contracted lead time
- Confidence
- Low
- Timestamp
- Reconstruction window
Partially observable through sector reporting.
Air freight substitution
Weaker than modelled- Current
- Limited
- Baseline
- Pre-disruption share
- Confidence
- Moderate
- Timestamp
- Reconstruction window
Challenged pathway 04.
Consequence map
What the model identified
Consequences are not presented as equally certain. Confidence degrades down the chain, and the lowest tier is stated plainly.
Primary consequence
- ·Effective container capacity absorbed by longer voyage rotations.
- ·Elevated freight and insurance cost on affected strings.
Secondary consequence
- ·Schedule variance at destination nodes, absorbed initially by buffer.
- ·Working-capital extension where goods remain in transit longer.
Conditional consequence
- ·Category-level availability pressure — conditional on buffer depth and service-level policy.
- ·Production planning change — conditional on supplier concentration and substitution options.
Low-confidence / uncertain
- ·Attribution of margin effects specifically to corridor cost versus other concurrent inputs.
- ·Duration of the disruption and the point at which routing normalises.
Scenario framing
Scenario framing
Short interruption
Assumption
Diversion reverses within one planning cycle.
Exposed systems
Freight and insurance cost only.
Sustained disruption
Assumption
Diversion persists across multiple planning cycles.
Exposed systems
Capacity absorption, inventory policy, working capital.
Persistent disruption
Assumption
Diversion becomes structural, with network reconfiguration.
Exposed systems
Sourcing structure, network design, capital allocation.
Scenario outputs are distributions and conditional pathways, not predictions. Confidence is conditional on the stated boundary, the available evidence, and the model assumptions.
Assessment record
The full record
- Assessment date
- Reconstruction window — analytical demonstration
- System boundary
- See system boundary above.
- Initial conditions
- Established Europe–Asia container routing through the Bab-el-Mandeb, with destination buffer held at seasonal norms.
- Shock / scenario
- Sustained disruption to commercial transit through the strait.
- Propagation pathways
- Four propagation pathways — capacity absorption, inventory in transit, lead-time variance, sourcing substitution.
- Expected observable indicators
- Diversion share, transit days, capacity absorption, freight pricing, air-freight substitution share.
- Scenario distribution
- Three scenario bands rather than a point forecast: short, sustained, persistent.
- Confidence
- High at the routing and transit layers; low at the inventory, capital, and availability layers.
- Known limitations
- Firm-level inventory and payment terms are not public. Substitution behaviour was under-modelled.
- Subsequent observations
- See the observation timeline above.
- Model revision
- Capacity layer promoted; insurance layer added. See the revision sequence below.
- What was missed
- Duration, insurance transmission speed, and the weakness of substitution.
Where the assessment failed
Publishing the failure modes is the point, not an admission against interest. An assessment that cannot show where it broke cannot be trusted where it held.
Duration
The model treated the disruption as bounded by a short escort-and-normalise cycle. It persisted far longer, which moved the binding constraint from transit time to capacity absorption.
Insurance as a cost layer
War-risk premiums were modelled as a minor input. They proved material and moved faster than freight indices, so the cost layer was under-weighted.
Substitution speed
Pathway 04 assumed sourcing and modal substitution would relieve pressure. Substitution was slower and narrower than assumed, so the relief did not arrive.
Resolution mismatch
Working-capital effects were asserted at a resolution the available data cannot support. The layer should have been marked unresolved from the outset rather than modelled.
What changed
Initial assessment
Assessed as a transit-time shock, with capacity and inventory effects following at one remove.
New observation
Sustained diversion made capacity absorption the dominant commercial effect, and insurance moved ahead of freight pricing.
Model revision
The capacity layer was promoted above transit time in the causal ordering; insurance was promoted from a cost line to a distinct layer.
Updated consequence map
The revised ordering is reflected in the node sequence on this page: routing → transit → utilisation → capacity → freight economics.
Limitations
What this reconstruction cannot establish
- ·Firm-level inventory, payment terms, and margin data are not public, so the lower half of the model is inferred rather than observed.
- ·Attribution is contested: corridor cost is one input among several concurrent macro factors.
- ·The model is calibrated to public sources only; no proprietary data was used.
- ·Scenario outputs describe conditional consequences under stated assumptions — they are not predictions.
- ·Confidence is materially lower for the inventory, working-capital, and availability layers.
Model output is not a single-point forecast
The Institute models relationships and conditional consequences. Outputs describe distributions, pathways, indicators, and uncertainty rather than asserting that one predetermined future outcome will occur.
Model
≠
Certainty
Scenario
≠
Prediction
Indicator
≠
Outcome
Bring us a system.
The same method applies to a defined decision, a shock, an exposure, or an operating question. We begin by writing down the system boundary and stating what cannot be established.
Start with the question