Floods and the Duplication of Production: Evidence from the Global Car Industry

Michele Fioretti; Sophia Praetorius; Thierry Mayer

AI summary · arXiv: 2611.01329v1 · econ.GN

Floods and the Duplication of Production: Evidence from the Global Car Industry

Michele Fioretti, Sophia Praetorius, Thierry Mayer

Original paper Submitted to arXiv on 31 Oct. 2026 License: CC BY-NC-SA 4.0 Also in: fr

How to read this summary? Illustrated guide

The essentials

After documented floods, carmakers add second production sites for engine programs rather than relocating production, chiefly where their existing networks make duplication inexpensive.

Question

How do multinational carmakers adjust vehicle and component production after floods, and how do existing network costs, insurance, and public flood protection relate to those adjustments?

Method

The authors combine S&P Global Mobility plant-level production data for 2000–2024 with a manually collected, verified flood record, identifying 216 flooded engine programs, 121 transmission programs, and 421 vehicle model versions. They compare flooded products with not-yet-flooded controls matched on launch cohort and pre-flood plant count, and examine production patterns alongside investment-cost, insurance-market, flood-hazard, and flood-defence measures.

Findings

Result Flooded engine programs were 19.4 percentage points more likely than controls to be produced in two or more plants in years 5–10, with no persistent decline in output at their original plants.
See the passage in the paper
Flooded engine programs are made in two or more plants more often than their controls, by 0.085 (s.e. 0.027) in years 0–4 and by 0.194 (s.e. 0.061) in years 5–10, from a flat path before the flood (0.024, s.e. 0.024, in years $-4$ to $-2$).
Result The duplication response was concentrated in engine programs whose owners already made the same family at a plant in another country; programs without such a plant showed no increase.
See the passage in the paper
Engine programs with a plant abroad that makes the family get a second plant, by 0.146 (s.e. 0.069) in years 0–4 and by 0.359 (s.e. 0.105) in years 5–10. The others do not ($-0.030$, s.e. 0.018).
Result Vehicle model versions had a temporary increase in production across multiple plants three to four years after a flood, but this difference had disappeared by year five.
See the passage in the paper
Flooded versions are made in two or more plants more often than their controls three to four years after the flood, by 0.096 (s.e. 0.028), from a flat path before it ($-0.008$, s.e. 0.012, in years $-4$ to $-2$).
Authors’ interpretation The authors interpret the second engine plant as a potential backup that also serves nearby assembly markets in normal times, rather than as a replacement for the flooded site.
See the passage in the paper
Its value as insurance lies in a long disruption: the program then already runs at a second site, with tooling, suppliers and workers in place, where the firm could raise output for the customers of the flooded plant.
Authors’ interpretation Patterns in the insurance and flood-defence analyses are consistent with substitution between private duplication and other protection, but the authors characterize this evidence as fragile.
See the passage in the paper
Both signs fit a trade-off, since firms duplicate more where insurance is expensive and where no defence protects the plant. Neither result is robust, however, and both rest on satellite flood signals.

Limitations

Long-run estimates rely on a limited number of owners, the 2011 floods contribute substantially, and annual data cannot establish whether a second site covers production during short stoppages.
See the passage in the paper
Our evidence has limits. The long-run estimates rest on few owners, the floods of 2011 play a large role, and annual data cannot show whether the second site covers for the first during a stoppage of a few weeks.
The insurance and flood-defence tests depend on satellite flood signals, and mapped flood defences lack dates.
See the passage in the paper
The tests on insurance and flood defences rest on satellite flood signals, and the defences mapped in OpenStreetMap are not dated.
The authors cannot directly observe the value of redundancy during disruptions or firms’ beliefs about disruption risk.

Why it matters

The authors argue that flood-related output measures can miss firms’ main adjustment: restructuring production by adding sites rather than shifting output away from flooded plants. Because duplication depends on existing capacity and local demand, scale-intensive supply-chain segments may remain concentrated and exposed; public defences and policies encouraging second sources should therefore be considered together.

Written by AI from the paper. Passages are checked word for word against the text we analysed; the rest may contain errors.

AI-generated keywords: Flood risk multinational production supply-chain resilience production redundancy flood insurance

Read further

Comprehensive Summary
Layman's Summary
Blog article
Original abstract

The paper examines how multinational automakers adapt their production networks after floods and how their responses relate to insurance and public flood protection. It asks whether firms relocate production away from flooded sites or establish backup production elsewhere, and why they duplicate some products but not others. The authors compile annual, plant-level data on vehicle, engine, and transmission production from 2000 to 2024, including links between component plants and vehicle assembly plants. They combine these data with a manually verified record of floods gathered from news reports and company statements. The record distinguishes direct hits to production sites from supply disruptions caused by floods elsewhere. The main analysis uses a staggered difference-in-differences design. For each product affected by a documented flood, the authors compare production changes before and after the event with changes for products not yet affected by a flood. Control products are matched to the affected product’s launch cohort and to the number of plants producing it before the flood. The largest control plant must also be located in another country. This approach addresses the fact that products made at more plants have more opportunities to be affected and may already be expanding. The study examines engine programs, transmission programs, and vehicle model generations separately. The central finding is that floods lead to additional production sites for engine programs, rather than lasting relocation or contraction at the original plant. Five to ten years after a documented flood, an engine program is about 19 percentage points more likely to be produced at two or more plants than comparable control programs. This is substantial: roughly half of the affected programs were produced at a single plant before the flood. Original plants generally retain the program, and their output shows no persistent decline. Output falls by about 11% in the flood year but recovers the following year. New plants are usually located abroad, and their production mainly supplies vehicles assembled near the new site rather than serving the original plant’s customers. This pattern is consistent with firms creating backup capacity to help manage a longer future disruption, although the annual data cannot directly reveal how that capacity would perform during one. Duplication depends strongly on a firm’s existing network. For single-plant engine programs whose owner already produces the same engine family at a plant abroad, the likelihood of multi-plant production rises by about 36 percentage points in years five to ten after a flood. The authors find no comparable increase when such a plant is absent. They estimate the average investment cost of a post-flood addition at about $26.8 million, or roughly $500 per engine of annual capacity. Under their assumptions, this investment is difficult to justify as protection against the short interruptions observed in the data. It would pay off only if disruptions lasted longer or local supply offered additional value. Responses vary across products. Vehicle model versions show a temporary increase in multi-plant production three to four years after a flood, of about 10 percentage points relative to controls, but the increase disappears by year five. The additional plant generally serves new markets, while the flooded version expands into additional countries more slowly. Transmission programs are not duplicated overall. The paper suggests that automatic transmission lines have a large minimum efficient scale relative to demand near most plants, whereas manual gearboxes appear more similar to engines. The authors also find suggestive evidence that firms add plants more often where insurance may be expensive and less often near mapped flood defences. This pattern is consistent with redundancy substituting for insurance and public protection, but the results are not robust enough to establish the relationship firmly. Limitations include the small number of owners underlying the long-run estimates, the influence of the major 2011 flood episode, incomplete and imperfect flood records, and the use of annual production data. The paper concludes that flood adaptation can change how production is organized even when average output quickly recovers. However, the costs and scale requirements of a second source leave some parts of supply chains concentrated and exposed.

How Floods Reshape the Global Car Industry

When a flood interrupts a factory, a company can wait for production to resume, move work elsewhere, or create a second source for the same product. Each option has costs, and the best choice depends on the product and the company’s wider manufacturing network. In “Floods and the Duplication of Production: Evidence from the Global Car Industry,” Michele Fioretti, Sophia Praetorius, and Thierry Mayer examine how multinational automakers make these choices.

The study follows vehicle models and generations, engines, and transmissions at the plant level from 2000 to 2024. The authors connect production sites to the assembly plants they supply and compile a verified record of floods from news coverage and company statements. They distinguish floods that directly hit production sites from events that disrupt suppliers elsewhere.

Following production before and after floods

To estimate how floods affect production networks, the researchers compare products affected by a flood with similar products that have not yet experienced one. They match comparison products to the affected product’s launch cohort and its number of production plants before the flood. They also require the largest control plant to be in another country. This helps account for the fact that products made at many sites have more chances of being exposed to floods and may already be growing.

The analysis considers engine programs, transmission programs, and vehicle model generations separately. That distinction matters: the cost, scale, and flexibility of adding a second production site are not the same for every part of a car.

Engines gain additional production sites

The clearest long-run response appears in engine production. Five to ten years after a documented flood, an affected engine program is about 19 percentage points more likely than a comparable control program to be made at two or more plants. The shift is notable because roughly half of the affected engine programs had been produced at just one site before the flood.

This is generally not a story of companies abandoning the flooded factory. The original plant usually keeps producing the engine, and its output shows no lasting decline. Production falls by about 11% in the year of the flood, then recovers the following year. The longer-run change is an additional site, often abroad. That new plant tends to supply vehicle assembly nearby rather than serve the customers previously supplied by the original plant.

The pattern is consistent with building backup capacity for a future, longer disruption—not simply shifting output after a brief interruption. However, the annual production data cannot show directly how much protection the added capacity would provide during a particular crisis.

Existing networks shape the response

Firms are more likely to add a second site when they already have a relevant foothold abroad. For single-plant engine programs whose owner produces the same engine family at another country’s plant, the probability of multi-plant production rises by about 36 percentage points in years five to ten after a flood. The researchers find no comparable increase when the owner lacks such a plant.

A post-flood addition costs an estimated $26.8 million on average, or roughly $500 per engine of annual capacity. Under the study’s assumptions, that investment is hard to justify as insurance against the short interruptions observed in the data. It may make more sense if future disruptions last longer or if local supply brings other benefits.

Different products, different strategies

Vehicle model generations show a more temporary response. Three to four years after a flood, the chance that a model version is produced at multiple plants rises by about 10 percentage points relative to controls. By year five, the increase has disappeared. The extra plant generally serves new markets, while the flooded version expands into additional countries more slowly.

Transmission programs, by contrast, are not duplicated overall. The authors suggest that automatic transmission production may require a large minimum scale relative to demand around most plants, making a second source uneconomical. Manual gearboxes appear more like engines in this respect, but the overall results do not show a broad transmission-duplication response.

Insurance, flood defences, and limits to resilience

The study offers suggestive evidence that firms add plants more often where insurance may be costly and less often near mapped flood defences. This is consistent with production redundancy serving as a substitute for insurance or public protection. But the relationship is not robust enough to establish firmly.

The findings also come with important caveats. Long-run estimates rely on a relatively small number of company owners, and the major 2011 flood episode has considerable influence. Flood records are incomplete, while annual data cannot capture the details of short-term operational responses.

Overall, the research shows that a quick recovery in average output does not mean a flood leaves production networks unchanged. Some firms respond by adding distant production sites, particularly for engines and when their existing networks make expansion practical. Yet the cost and scale required to create a second source can leave other products concentrated—and exposed to the next disruption.

Created on 09 Nov. 2026

Was this synthesis useful?

Score: 0

Ask questions about this paper to our AI assistant

How to chat with this paper? Illustrated guide

Compare with the PDF

You can also chat with multiple papers at once here.

AI assistant instructions?

Tools for this paper

Explain it differently 1 free rewrite left today

Save to a collection Group papers in collections to compare them, chat with them and export them. Continue with Google Create a free account Log in

Export BibTeX · RIS (Zotero, Mendeley, EndNote) · Obsidian (Markdown) · Notion (Markdown)

Advanced tools Premium

For researchers who follow a field closely: citation alerts, journal-club slides, version comparison and Zotero export. Discover Premium

Zotero

Citation alerts Citation alerts guide

Who works on it? The people and institutions who publish on its topic

Weekly letter Presented in "Economics and finance this week" of 9 November 2026. Economics and finance this week: When prices, protection and ownership change the rules

Journal club Premium+ Slides to present it Propose for the journal club

Versions What changed between versions

Talk about it in public Premium+ Press kit: release, questions, hype check Press kit guide

Papers around this one

See them on a map

Ranked by how close their title and abstract are to this paper’s. Select some to read them together.

Closest papers

Close Another approach to the problem
Technology, Licensing and Multinational Production
Sophia Praetorius · 2026 · CESifo working papers · Not summarized yet

Technology, Licensing and Multinational Production studies how automakers respond to local shocks, focusing on platform constraints and technology licensing rather than flood-driven production duplication.

Abstract
How flexible are multinational production networks when assembly technologies are specialized and costly to reconfigure? Using data on the global automotive industry from 2000-2023, I document that production is organized around discrete process technologies (platforms) that constrain where each product can be assembled. Exploiting the 2011 Tohoku earthquake as a quasi-natural experiment, I show that platform compatibility limits firms’ ability to reallocate production after local shocks. Firms overcome this constraint by licensing technologies to each other. I develop and estimate a structural model of multinational production with discrete technology choice and inter-firm licensing. Models…
Related Broader context
Deep uncertainties in sea-level rise and storm surge projections: Implications for coastal flood risk management
Perry C. Oddo, Ben Seiyon Lee, Gregory G. Garner et al. · 2017 · Applications · cited 61 times · AI summary available

Deep uncertainties in sea-level rise and storm surge projections examines infrastructure investment and coastal flood-risk decisions, not firms’ production responses to floods.

Abstract
Sea-levels are rising in many areas around the world, posing risks to coastal communities and infrastructures. Strategies for managing these flood risks present decision challenges that require a combination of geophysical, economic, and infrastructure models. Previous studies have broken important new ground on the considerable tensions between the costs of upgrading infrastructure and the damages that could result from extreme flood events. However, many risk-based adaptation strategies remain silent on certain potentially important uncertainties, as well as the trade-offs between competing objectives. Here, we implement and improve on a classic decision-analytical model (van Dantzig 1956)…
Related Same question
The influence of natural disasters on supply chain concentration: the moderating role of digital capability
Qian-Qian Han, Shen-Yang Jiang, Wen-Lan Zhang et al. · 2026 · International Journal of Operations & Production Management · Not summarized yet

The influence of natural disasters on supply chain concentration estimates how disasters affect firms’ supply-chain structures, using a staggered difference-in-differences design.

Abstract
Natural disasters impose substantial operational and strategic challenges on firms. Drawing on Resource Dependence Theory (RDT), this study examines how natural disasters (specifically, typhoons and earthquakes) affect firms' supply chain concentration and investigates the moderating role of digital capability in this relationship. Using a panel of A-share-listed firms on the Shanghai and Shenzhen stock exchanges from 2010 to 2023, comprising 27,578 firm-year observations, we exploit a staggered difference-in-differences (DID) design to identify the causal effect of natural disasters on firms' supply chain concentration. We find that firms exposed to natural disasters significantly reduce bo…
Related Broader context
From Maritime Disruption to Firm-Level Resilience: Risk Transmission Mechanisms in Supply Chains
Ying-Feng Zou · 2026 · Finance & Economics · Not summarized yet

From Maritime Disruption to Firm-Level Resilience explains how shipping disruptions transmit into firm operations, rather than studying firms’ responses to floods at production plants.

Abstract
Global production and trade are structurally dependent on maritime transport, while geopolitical conflict, extreme weather, port congestion, and restrictions at major chokepoints have increased uncertainty in transport services. Maritime route disruptions are initially reflected in rerouting, schedule revisions, changes in port calls, and higher transport costs. Their impacts enter firm-level supply chains when these transport changes create procurement fulfilment deviations that interfere with purchasing and production plans. This study applies integrative literature analysis and inductive reasoning to explain how risks are transmitted from the maritime network to firm operations. The analy…
Related Another approach to the problem
Research on cross-industry production based on system dynamics under major emergencies
Yi-Can Yuan · 2026 · Journal of Modelling in Management · Not summarized yet

Research on cross-industry production models how firms can repurpose capacity during emergencies, whereas this paper measures automakers’ actual responses to floods.

Abstract
This study aims to investigate how supply chain disruptions caused by major public emergencies, such as geopolitical conflicts, affect firm performance and examines whether cross-industry production (CIP) can mitigate losses by repurposing capacity to produce contingency survival materials. A three-echelon supply chain model – supplier, plant, distribution center – was developed using system dynamics. It simulates simultaneous disruptions in supply, capacity and demand. Performance is measured by cumulative profit and service level under different disruption scenarios and CIP capacity allocation rates. Demand disruption can reduce the negative impact of capacity disruption, as lower downstre…
Related Broader context
Global Value Chains, Resilience and Strategic Behaviour
Sabyasachi Saha, Ripudaman Bhardwaj · 2026 · Economic & Political Weekly · Not summarized yet

Global Value Chains, Resilience and Strategic Behaviour examines import dependencies and industrial-policy choices as ways to strengthen value-chain resilience.

Abstract
In addressing risks to value chains, the choice of policy instrument matters and depends on institutional context. The role of fiscal policy in strengthening supply chain resilience has received increasing attention in recent years. We present empirical insights from India by estimating the extent of (critical) import dependencies for sectors like electrical machinery and equipment; machinery and mechanical appliances; organic chemicals; inorganic chemicals among others. In response to disruption risks affecting value chains, fiscal instruments could support resilience in multiple ways. In this paper we develop a game theoretic model to capture the industrial policy choices of the Government…
Related Broader context
Ranking the wrong places: vulnerability assessment, flood losses and risk sharing in Italy and Europe
Stefano Blando · 2026 · Not summarized yet

Ranking the wrong places evaluates flood vulnerability rankings against realized losses and risk sharing, rather than firms’ production adjustments after floods.

Abstract
Before a flood, governments decide where to invest in protection; after it, how much of the loss to compensate. The European Union informs the first decision with a composite index that ranks every subnational unit by vulnerability. This paper tests the ranking by comparing Italian provinces struck by the same flood between 2005 and 2024. The published index does not order provinces by the losses they suffer, with or without controls for hazard, exposure and income ($\rho$ between -0.06 and 0.09). The failure lies in what it weights. As published, the index mostly measures economic development, while losses follow water: they arrive in recurring groups of provinces that share a river basin, …
Related Broader context
AI-Guided Diversification of Green Technology Supply Chains
D. Erokhin · 2026 · Sustainability · Not summarized yet

AI-Guided Diversification of Green Technology Supply Chains uses trade-network analysis and optimization to diversify suppliers against disruption risks in green technologies.

Abstract
Green technology supply chains combine rapid demand growth with concentrated production, uneven environmental performance, and exposure to trade disruption. This article links trade network analysis, next-year bilateral trade link prediction, multi-objective portfolio optimization, and disruption stress testing across 14 importing markets and six technology and material groups from 2015 to 2024. The cleaned sample contains 15,186 supplier-year observations and USD 696.1 billion in foreign supplier trade. Concentration rose most clearly in photovoltaic equipment, lithium-ion batteries, and wind equipment. In the chronological 2022–2023 test, gradient boosting achieved ROC AUC 0.911 and a Brie…

Go further

How are these papers found?

We compare the title and abstract of this paper with those of every paper summarized on this site, and with the papers Semantic Scholar recommends, those it cites and those citing it (for a much-cited paper, its most recent citations and, from OpenAlex, the most cited ones). Among the prior work and the follow-ups, the papers that use each other in depth, a method it uses and the much-cited follow-ups come a little higher. The levels (very close, close, related) come from the similarity of the texts, not from a judgment of quality; * marks a paper without an abstract, compared by its title. The sentence under a paper is written by an AI from the two abstracts: check it in the papers. A paper marked not relevant by three readers leaves the list.

Updated on 9 November 2026.

Disclaimer: The AI-based summarization tool and virtual assistant provided on this website may not always provide accurate and complete summaries or responses. We encourage you to carefully review and evaluate the generated content to ensure its quality and relevance to your needs.

Floods and the Duplication of Production: Evidence from the Global Car Industry

The essentials

Question

Method

Findings

Limitations

Why it matters

Read further

Definitions

How Floods Reshape the Global Car Industry

Following production before and after floods

Engines gain additional production sites

Existing networks shape the response

Different products, different strategies

Insurance, flood defences, and limits to resilience

Ask questions about this paper to our AI assistant

Tools for this paper

Advanced tools Premium

Papers around this one

Closest papers

Go further

Cite this paper (BibTeX)