Shipping chokepoints: inventories and working capital

Observed global oil inventories fell by 10 million barrels in September. On its own, that figure suggests a relatively modest draw. The breakdown published by the International Energy Agency on 14 October tells a different story: on-land stocks dropped by 113 million barrels, while oil on water rose by 103 million. The small change in the total combines two pools of inventory whose immediate availability may differ substantially. ^[2]

That distinction gets to the heart of maritime supply risk. A cargo can exist, have a market value and appear on someone’s balance sheet long before it reaches a factory in a usable form. The IEA links the increase in oil on water to recovering Gulf exports. The additional 103 million barrels therefore cannot, by themselves, be treated as a measure of oil held up by longer routes. They do show why aggregate stocks are an incomplete guide to supply security. ^[2]

A bulletin published by the Bank for International Settlements on 15 October offers a useful framework. When a critical passage is disrupted, firms can draw down inventories and seek alternative suppliers. Those responses protect some production while adding costs. Their effectiveness depends on how long the disruption lasts and how readily inputs can be replaced. ^[1]

The economic question is what happens to the days lost at sea: ships remain occupied for longer, goods require financing before they can be used, and reserves are consumed while the next delivery approaches. This investigation follows those mechanisms to the conditions under which they begin to constrain output.

FIG. 01 The aggregate change combines two opposite movements. Oil on water and on-land inventory do not have the same immediate availability. Source: IEA, published 14 October 2026, September observations.[2]

Method and scope

Waterfall: −113 + 103 = −10 million barrels, September 2026 changes. Intermediate bars connect; they are not three inventory levels. The IEA links the rise in oil on water to recovering Gulf exports. This total measures neither immediately usable inventory nor the separate effect of diversions.

A fallback route can move the point of failure

A chokepoint concentrates maritime flows in a narrow passage. A map of maritime chokepoints can make exposure look deceptively easy to measure. Add up the shares of trade passing through each narrow passage and the total seems to describe the danger. Yet a single cargo may cross several chokepoints in succession. Counting its exposure to Malacca, Bab el-Mandeb and Suez separately would count the same goods several times. Network research published in Nature Communications in November 2025 makes these overlapping dependencies explicit. A transit share measures exposure, rather than an economic loss that follows automatically from disruption. ^[11]

The next question is practical: where can the ship go when its usual route becomes unavailable or too risky? On an Asia-Europe service, sailing around Africa provides an alternative to the Red Sea and Suez. Goods can continue to move, but spend longer in transit. The geography of the Gulf creates a different constraint: maritime access runs through Hormuz. Some oil flows have pipeline alternatives, which are unavailable to many producers and categories of cargo. ^[3]^[6]

Pipeline capacity therefore cannot be treated as an all-purpose substitute for shipping capacity. Oil must be able to reach the pipeline, spare space must exist, the terminal must operate and a ship must be available to load at the other end. Our analysis of the Yanbu oil route follows these onward connections. An outlet on the Red Sea also retains its own dependencies: a cargo loaded at Yanbu for Europe can head north towards Suez, whereas a sailing to the Indian Ocean passes through Bab el-Mandeb.

The October edition of the US Energy Information Administration’s outlook describes the temporary interruption of Saudi Arabia’s East-West pipeline after attacks in September. A fallback option has vulnerabilities of its own. Assessing its value requires looking at actual deliveries and infrastructure availability, rather than relying on nameplate capacity. ^[3]

FIG. 02 Lines show possible connections, not distance, flow volumes or the current operational status of routes. The pipeline is an option for some crude flows, not all Gulf cargo.[3][6][11]

Method and scope

Selected, non-exhaustive network. Line lengths encode no distance, flow, spare capacity or current operational status. Connections indicate possible routes. The East-West pipeline serves some crude flows; its terminals and onward routes retain dependencies of their own.

Longer voyages absorb capacity without removing ships

The Red Sea disruption in early 2024 provides an observable example. According to IMF estimates, the volume of trade passing through Suez fell by 50% in the first two months of the year compared with the same period of 2023. Transit around the Cape of Good Hope increased by 74%. At Panama, drought-related restrictions produced a decline of almost 32% over the same comparison period. These percentages use three separate bases. They describe a rearrangement of trade flows and cannot simply be added to estimate the change in global trade. ^[4]

A detour also consumes ship time. Consider a hypothetical service with a normal 70-day round trip, including loading, sailing, unloading and returning. If the cycle now takes 84 days, the ship cannot provide the same service frequency. Holding the fleet, load and operating conditions constant, transport capacity per unit of time falls to 70 divided by 84, or 83.3% of its previous level. The round trip becomes 20% longer, while throughput falls by 16.7%.

This is a calculation, not an estimate of current traffic. It isolates a mechanism: a vessel that still exists on a fleet register may already be occupied absorbing extra voyage time. Maintaining throughput then requires more ships, altered service patterns or longer waits for some cargo. Each response has a cost and faces constraints of its own.

UNCTAD’s global figures for 2024 illustrate the difference between the amount shipped and the transport effort required. Seaborne trade rose by 2.2% in volume, but by roughly 6% in tonne-miles, a measure multiplying cargo mass in metric tonnes by the distance travelled in nautical miles. Greater transport work places additional demands on the fleet without representing an equivalent increase in goods available to consumers. ^[5]

The price charged to cargo owners still depends on spare vessels, demand and contracts. A fleet with unused capacity can accommodate a diversion more easily than a market already operating close to its limits. The same additional sailing time can therefore have very different consequences depending on when it occurs.

FIG. 03 Left: historical transit volumes, each indexed to its own base. Right: a separate model in which a 20% longer cycle cuts throughput by 16.7% with a fixed fleet. On mobile these panels are stacked.[4]

Method and scope

Separate transit indices: January-February 2023 = 100 for each route. The IMF’s estimated changes yield 50, 174 and about 68 for January-February 2024; these bases cannot be added. The independent vessel-cycle model gives 70/84 = 83.3% of initial throughput, a 16.7% reduction, holding fleet, load per cycle and operations constant.

Twelve days of inventory can provide thirty days of protection

Inventory coverage creates another common misunderstanding. A warehouse holding twelve days of normal consumption will run dry after twelve days only if incoming deliveries stop completely. When supplies continue at a reduced rate, the reserve makes up the gap between consumption and arrivals.

Take a hypothetical factory with 1,200 units of an input and a normal requirement of 100 units a day. During the disruption it still receives 60 units daily, so its inventory declines by 40 units a day. The reserve can sustain normal output for thirty days under the model’s assumptions: a uniform input, regular deliveries, no waste and no required minimum safety stock. Using the net supply shortfall produces a very different result from assuming that all arrivals cease.

If deliveries recover after twenty days, 400 units remain. The factory has maintained production throughout the disruption, but its buffer has become thinner. Receiving 100 units while consuming 100 stabilises inventory; rebuilding the reserve requires higher deliveries or lower consumption. Improving traffic can therefore coexist with diminished protection against the next incident.

If normal arrivals resume only after thirty-five days, the reserve runs out at day thirty. For the next five days, the plant can use only the 60 units arriving daily. Assuming one unit of input is needed for each unit of output, cumulative production falls 200 units short of its normal pace. The model includes no subsequent catch-up. These dates are illustrative assumptions, unrelated to any predicted date of shortage in October 2026.

FIG. 04 Two illustrative recovery dates and the same initial reserve. Each unit of output requires one input unit. Recovery at day 20 stabilises inventory at 400; recovery at day 35 leaves 200 units of lost output. l0g calculations with hypothetical assumptions.

Method and scope

Source: l0g calculations; all inputs are hypothetical. Continuous model: 1,200 starting units, desired use of 100 per day, deliveries of 60 followed by 100 from day 20 or 35. One input unit produces one output unit, with no waste, minimum reserve or catch-up. The daily shortfall of 40 exhausts inventory at day 30; five days at reduced output create a cumulative shortfall of 200. Restored deliveries stabilise rather than rebuild inventory.

Our analysis of European gas between cargoes applies this distinction to seasonal supply.

Real operations add complications. Ships deliver cargo in batches; a plant may require a minimum inventory to keep running; a supplier may reallocate an order to another buyer. What matters is the timing and quantity of the next usable receipts. A monthly average offers little reassurance against a week without the component needed to operate an entire line.

A missing component can immobilise goods worth far more

A large total inventory can conceal a small reserve of the crucial input. A company may hold substantial finished goods while lacking a component needed for its next production run. The accounting value of inventory is then a poor guide to operational continuity. Its composition and the ease of replacing individual inputs matter more.

Research by Boehm, Flaaen and Pandalai-Nayar on the 2011 Tōhoku earthquake provides an empirical example of this rigidity. Among the Japanese affiliates studied in the United States, disruptions to imported inputs translated strongly into lower production. The evidence concerns a particular industrial network, but illustrates why a replacement supplier does not become operational immediately when the usual source is interrupted. ^[8]

Even three separate suppliers may depend on the same port, maritime passage or upstream producer. In a simplified network of that kind, counting distinct contracts overstates diversification. Testing the protection they provide requires following the input beyond the first supplier and establishing whether the alternative relies on genuinely different infrastructure. Maritime network analysis helps identify this concentration of dependencies. ^[11]

Inventory is best understood here as a reserve of productive services: the appropriate material, available in time, at the place where it can be processed. That is more demanding than tracking a single balance-sheet figure. It also avoids assigning goods in transit the same immediate usefulness as goods that have already been unloaded and made available for production.

Extra days require extra financing

A moving cargo can continue to tie up capital while generating no cash receipts. Working capital bridges this operating gap, covering funds committed to inventories and customer receivables after allowing for financing provided by suppliers. Research by Kim and Shin, published by the BIS in 2023, puts the time between expenditure and sales receipts at the centre of supply-chain finance. ^[14]

Consider a hypothetical importer buying €50,000 of goods each day and paying for them at departure. If transit becomes permanently fourteen days longer, with purchase prices and throughput unchanged, the importer will eventually have an additional €700,000 tied up in goods still travelling. It must finance that amount for longer before the cargo can be used and sold.

The €700,000 is an additional capital requirement, backed by goods. Treating the whole amount as a loss would ignore the asset on the other side. Financing does, however, create an expense: at an illustrative annual interest rate of 8%, an additional €700,000 outstanding for a full year would cost €56,000 in interest, excluding fees. The bill for a temporary episode depends on the amount actually borrowed and how long it remains outstanding; a full year of interest cannot simply be assigned to a brief disruption.

FIG. 05 The financing requirement rises in the new transit regime. Tied-up capital remains backed by goods, while interest is a separate cost. l0g example with a hypothetical rate and the full balance held for a year.[14]

Method and scope

Amounts and rate are l0g assumptions, not company data. Payment at departure, constant purchases of €50,000 a day and a permanent 14-day transit extension imply €700,000 of additional in-transit goods at the new steady state. The €56,000 interest bill assumes the full balance is financed for a year at 8%, excluding fees. A temporary episode requires its actual financed balance and duration.

Our investigation into financing oil cargoes follows the advances, collateral and payment schedules.

Commercial terms determine who carries the timing gap. A supplier accepting later payment may finance part of it; an advance payment shifts more of the burden to the buyer. The relevant evidence is therefore payment schedules and expected cash receipts, rather than an assumption that every cargo at sea consumes the same party’s liquidity. Available credit facilities and their cost matter as well. ^[14]

A company that maintains production can still emerge from the disruption with more short-term debt or less cash. This provides a plausible channel to investment: funding current operations may delay spending that would have improved future capacity. Its importance must be established from the accounts and contracts of affected firms, rather than inferred from freight prices alone.

Credit can expand while conditions deteriorate

A Federal Reserve note published on 30 January 2026 offers a useful parallel. Around the US tariff announcements in the second quarter of 2025, banks more exposed to supply-chain risk experienced larger increases in credit utilisation and loan spreads. The authors interpret this combination as consistent with stronger credit demand to build inventories and bring forward imported equipment purchases before tariffs take effect. Their evidence concerns that trade-policy episode, rather than an estimate of Hormuz’s effects. ^[15]

The mechanism nevertheless calls for care when interpreting faster lending growth. A company facing longer lead times may need more financing to support the same sales. Additional borrowing need not indicate expanding activity; it may accompany an operating cycle that has become more expensive to carry.

Available credit can, conversely, be the very thing that prevents an interruption to production. Finance then acts as a buffer. Fragility emerges when additional needs reach a liquidity limit, or their cost absorbs too much of the operating margin. Extending payment terms to suppliers can also move the constraint upstream. These are channels to investigate, rather than evidence of a banking crisis caused by maritime chokepoints in October 2026.

FIG. 06 Possible transmission channels and buffers, not a measurement of their frequency or scale in the current crisis. The Federal Reserve evidence concerns the 2025 tariff episode.[1][14][15]

Method and scope

Qualitative channels, not measured probabilities, magnitudes or timelines. Inventories, credit and substitution can cushion the disruption. The Federal Reserve note studies the 2025 tariff announcements; it does not measure a banking crisis caused by chokepoints in October 2026.

Price effects can extend beyond firms whose cargoes use the affected passage. A buyer seeking another source competes with customers already purchasing from that alternative supply. Costs may then rise for countries with little direct dependence on the disrupted route. An analysis by researchers published by ifo/EconPol in July 2026 draws attention to this distinction between physical route dependence and exposure to world prices. ^[10]

Transmission to consumer prices subsequently depends on transport’s share of total costs, the margin a firm can absorb, its contracts and the demand response. An IMF study published in 2022, using data from 1992 to 2021, examines how shipping-cost shocks spread over time. It does not provide a mechanical conversion from today’s freight rate to tomorrow’s inflation. ^[9]

The BIS working paper on Hormuz published in September 2026 also examines delayed responses in activity, prices and financial conditions. Its estimates use observations from 2019 to 2025. These historical associations are consistent with several transmission channels; they cannot establish a breaking point for the current episode. ^[7]

Adaptation can contain the lasting damage

The strongest counterargument to a general paralysis scenario is the evidence that supply networks adapt. In early 2024, the ECB observed that spare shipping capacity, subdued demand for goods and manufacturers’ inventories were limiting the impact of Red Sea disruption. Delivery times deteriorated without immediately reproducing all the difficulties seen during the pandemic. That assessment concerned the conditions prevailing then; assuming an identical buffer in 2026 would be unwarranted. ^[13]

Research published by the Banque de France in August 2026 goes further by examining adjustments in routes and maritime capacity. Under its long-run assumptions about substitution and canal tolls, the option of sailing around Africa substantially limits the global real-income loss from a permanent Red Sea closure. The distribution is much less favourable to Egypt, which loses some of the income associated with its canal. This is a model result for a specified disruption and horizon, distinct from a crisis affecting several infrastructures at once. ^[6]

The objection matters. Expensive detours can coexist with enough adaptation to contain lasting economic damage. Analysis must separate temporary additional costs, transfers between participants and actual lost production. A shipowner, an importer and a country collecting passage fees face different financial consequences; some additional income has its counterpart in higher expenditure elsewhere.

Adjustment can also run in reverse. UNCTAD warned in 2024 that a return to normal routes could expose overcapacity in container shipping after diversions had absorbed vessels. Freight costs need not follow a permanently rising path: the duration of disruption, vessels added in the meantime and the course of demand can reverse the balance of the market. ^[12]

The breaking point is found at the receiving end

Turning general exposure into a credible risk estimate requires reconstructing a specific chain. How much of the critical input is already usable? What volumes will arrive before it runs out? Can another material or supplier be brought into operation within that interval? Who has the liquidity to finance the wait? The thirty-day example shows why focusing exclusively on either inventories or traffic misses an essential variable.

Public indicators each answer part of the question. Vessel counts describe passages, subject to the limitations of positioning data; they do not constitute an inventory of delivered materials. Customs statistics record trade at another stage. Aggregate stocks sometimes combine locations and products that are poor substitutes. Research on maritime routes makes it important to preserve these distinctions when interpreting the numbers. ^[4]^[7]

The public evidence reviewed for this investigation does not establish a universal date for the exhaustion of industrial inventories, or a global financing requirement attributable to chokepoint disruption. Those estimates would require granular information on products, storage locations, delivery schedules and payment terms. The simulations presented here make the relationships calculable without filling gaps in the observations with invented data.

The conclusion that survives the counterarguments is more precise than a prediction of worldwide paralysis. Firms can buy time through inventories, alternative routes and credit. That time has a cost, distributed differently as disruption persists. Assessing when output will actually suffer requires comparing the arrival of the next usable input with the reserve still available and the means of financing the interval. At that level, geographical exposure becomes a measurable economic risk.

Sources and limits

Documentary research cut-off: 16 October 2026. Oil inventory changes refer to September 2026; the maritime route comparison covers January-February 2024 against January-February 2023. Inventory, vessel-cycle and financing scenarios are illustrative l0g calculations, with assumptions specified in the text and figures. No real-time traffic series was reconstructed for this article. The research papers cited express their authors’ views and do not necessarily represent the official position of their institutions.

BRI / BIS : Chokepoint disruptions and their macroeconomic consequences.
15 October 2026. Bulletin 138; text and relevant PDF pages consulted. Inventories, substitution and costs condition transmission. Its simulations are not an October 2026 global-output forecast; the research expresses the authors’ views.
AIE / IEA : Oil Market Report – October 2026.
14 October 2026; September observed changes. Public overview consulted, subscription report not used. On-land and on-water inventory differ in availability. Rising Gulf exports accompany the increase on water, which does not isolate diversions.
EIA : Short-Term Energy Outlook – Global oil markets.
6 October 2026 edition, forecasts completed on 1 October. Public HTML consulted. It describes a temporary East-West pipeline interruption and partial resumption after repairs. Rolling URL; no PDF-table values are used.
FMI / IMF : Red Sea Attacks Disrupt Global Trade.
7 March 2024. January-February 2024 against the corresponding 2023 period: estimated transit volumes with separate bases. l0g calculates the indices 50, 174 and about 68. No AIS series or global volume is reconstructed.
CNUCED / UNCTAD : Review of Maritime Transport 2025: press conference ; original 2025 report.
24 September 2025. Official transcript of UNCTAD’s Geneva launch, with relevant statements read directly: 2024 volume grew by 2.2% and tonne-miles by about 6%. The original report specifies 5.9%; its indexed text was consulted but direct download failed. Units are metric tonnes and nautical miles. No 2025 forecast is treated as an outcome.
Banque de France : Navigating Shocks: The Ripple Effects of Shipping Route Closures.
Published 6 August 2026; page updated 24 September. Abstract, nontechnical summary and PDF sections available as text consulted; direct PDF download denied. Long-run model conditional on substitution and toll assumptions. The Red Sea closure differs from a Hormuz transit-fee scenario and from simultaneous infrastructure failures.
BRI / BIS : Maritime chokepoints and the global economy: evidence from the Strait of Hormuz.
17 September 2026, Working Paper 1378. PDF and methodological pages consulted. Historical estimates over 2019-2025 are separate from descriptive 2026 charts. Associations with activity, prices and finance do not establish a current shortage deadline. AIS counts may omit vessels.
Federal Reserve Board : Input Linkages and the Transmission of Shocks: Firm-Level Evidence from the 2011 Tōhoku Earthquake.
October 2015; record updated 19 June 2020. FEDS 2015-094; abstract and empirical PDF section consulted. Evidence concerns the US affiliates of Japanese firms studied after the 2011 earthquake. No coefficient is generalised to all companies or chokepoints.
FMI / IMF : Shipping Costs and Inflation.
25 March 2022, Working Paper 2022/061. Record and PDF passages available as text consulted; direct download denied. Observations run from February 1992 to December 2021. The Baltic Dry Index covers dry bulk, not container or tanker rates. No coefficient is turned into a 2026 inflation forecast.
ifo / EconPol : Mapping International Trade Exposure to Critical Maritime Chokepoints.
31 July 2026. Researchers’ analysis published by ifo/EconPol; estimated routes include German imports in 2023. It separates physical exposure from world-price transmission. No German share is converted into a global current observation.
Nature Communications : Systemic impacts of disruptions at maritime chokepoints.
24 November 2025, Nature Communications 16, article 10421. Original HTML consulted. Modelled transport network with flows including a 2022 trade baseline. A cargo may cross several passages; exposure differs from realised loss. No modelled monetary-risk amount is reused.
CNUCED / UNCTAD : Vulnerability of supply chains exposed as global maritime chokepoints come under pressure.
22 October 2024. Indexed institutional text consulted; direct page retrieval failed. The conditional overcapacity warning concerns container vessels returning to shorter routes, not an observed October 2026 surplus.
BCE / ECB : ECB staff macroeconomic projections for the euro area, March 2024.
7 March 2024. Relevant boxes in the March projections consulted. Spare shipping capacity, goods demand and inventories then cushioned Red Sea disruption. These 2024 conditions are not presented as the state of the economy in 2026.
BRI / BIS : Theory of supply chains: a working capital approach.
25 January 2023, revised October 2024. Working Paper 1070; presentation and relevant PDF passages consulted. A theoretical model of financing production time. €50,000, €700,000, 8% and €56,000 are l0g assumptions and calculations, not observations from the study.
Federal Reserve Board : Supply Chain Risk and Bank Lending Amid Trade Policy Uncertainty.
30 January 2026. Full text, methods, results and notes consulted. Sample of 23 banks over the first half of 2025, including the second-quarter tariff announcements. The credit-demand interpretation of utilisation and spread changes belongs to the authors. No 2026 Hormuz banking effect is estimated.