Climate-Resilient Industrialization Pathways: Sorting emerging-economy industry by the constraint that actually binds it

Hunter Hughes

Abstract

For most of the post-war period, an emerging economy's industrialization strategy could be judged against a single number: unit cost. Climate change and the policy response to it have converted that into a portfolio of four constraints — the price of electricity, the reliability of electricity, physical exposure to heat and flood, and compliance with the carbon rules of destination markets. The central argument of this report is that these four do not bind the same industries, and that a green industrial policy offering every sector the same package therefore misallocates. The useful question for a planning ministry is not whether an industry is green, but which of the four stops it first.

The evidence for each constraint is well documented and rarely assembled in one place. On price, electricity is the largest single operating cost in primary aluminium smelting and a substantial share in ammonia, cement and steel — the industries for which clean power is simultaneously a cost strategy and, under the EU's Carbon Border Adjustment Mechanism, a market-access strategy. On reliability, World Bank firm surveys covering more than 143,000 firms in 137 countries put sales losses from outages at about US$82 billion a year, exceeding ten per cent of sales in the countries with the least reliable grids; when drought and heat cut northern Vietnam's supply in mid-2023, the World Bank estimated losses of US$1.4 billion, around 0.3 per cent of GDP, with surveyed industrial firms reporting revenue losses of ten per cent. On exposure, a national panel of Indian factories shows annual plant output falling about two per cent per degree Celsius, with worker productivity down four to nine per cent per degree on hot days, and climate control substantially mitigating the loss; the 2011 Thailand floods inundated seven industrial estates and 904 factories, causing some US$45.7 billion of economic damage and halting production in plants that were never flooded.

From these, the report builds three archetypes and a screening test. Price-bound industries — the energy-intensive processors, which are also the six sectors CBAM covers — should be sited and sequenced around the cost of clean power, which is why the 'powershoring' proposition of relocating energy-intensive plants to clean-grid economies is a genuine and rare new source of comparative advantage. Reliability-bound industries — electronics, precision assembly, pharmaceuticals, cold chains — can absorb a higher tariff but not an outage, and their pathway runs through firm capacity, storage, and instruments such as Vietnam's 2024 direct power purchase agreement decree that let manufacturers contract clean supply directly. Exposure-bound industries — labour-intensive assembly and agro-processing — are constrained by heat and flood at the site and in the logistics chain, and their pathway is cooling, siting and drainage rather than tariffs.

Three implications follow. First, the constraints are ordered differently by sector but the fixes overlap: a firm, clean grid serves the price-bound and reliability-bound alike, and industrial-zone drainage and cooling serve every tenant. Second, some fixes are rival: cooling a factory is an electrical load, so the exposure fix raises the reliability requirement — an instance of the mutual dependency examined in a companion report. Third, the sequence matters more than the total, because these are capital stocks with forty-year lives being sited now; a plant placed in a flood plain on an unreliable grid to save capital today has locked in an operating cost that no later subsidy removes. The report closes with recommendations for planning ministries, industrial-zone operators, development finance and exporters.

Executive Summary

Green industrial strategy usually offers every sector the same package. The sectors are stopped by different things, and the cheapest strategy starts by asking which.

FINDING 01

One number became four

Industrial competitiveness used to be judged on unit cost. It is now a portfolio of four constraints: the price of electricity, its reliability, physical exposure to heat and flood, and compliance with destination-market carbon rules.

FINDING 02

They bind different industries

A smelter is stopped by the price of power. An electronics plant can pay a higher tariff but not survive a four-hour outage. A garment factory is stopped by a heatwave or a flood. Three problems, three pathways.

FINDING 03

Siting decisions are being made now

These are forty-year capital stocks. A plant sited in a flood plain on a weak grid to save capital today has locked in an operating cost that no later subsidy removes — which makes sequence more consequential than the total spent.

Industrialization remains the most reliable escalator out of low income, and the question for emerging economies pursuing it is no longer only whether they can produce at competitive cost. It is whether the plants they build in this decade will still be operable, financeable and admissible to their export markets in the next one.

This report assembles the evidence on each of the four constraints, and then does something the literature does surprisingly rarely: it sorts industries by which constraint binds them first. That sorting produces three archetypes with genuinely different pathways — and it explains why an industrial policy built around a single instrument, whether a tariff subsidy, a resilience fund or a green certification scheme, reliably helps one archetype and leaves the others where they were.

−2%/°C

Annual output of an Indian factory per degree of warming; 4–9% per degree on hot days

Somanathan et al., JPE (2021)

$82bn

Annual sales lost to electricity outages across 143,000 firms in 137 countries

World Bank PRWP 8899

$45.7bn

Economic damage from Thailand's 2011 floods; 904 factories in seven industrial estates inundated

World Bank / GFDRR

2.2%

Of global working hours projected lost to heat stress in 2030 — 80 million full-time jobs

ILO, Working on a Warmer Planet

Sections 1 to 4 set out the four constraints and the evidence for each. Section 5 builds the three archetypes and a screening test. Section 6 gives the pathway and sequence for each, including where the fixes overlap and where they compete. Sections 7 and 8 give recommendations and conclusions.

The question is not whether an industry is green. It is which of the four constraints stops it first — because that is the one worth spending on. The binding-constraint screen

1. Four Constraints, Not One Cost

Each constraint has a different unit, a different instrument and a different timescale. Treating them as one "climate risk" line obscures all three differences.

The classic account of export-led industrialization is a cost story. A country with abundant labour and access to capital assembles goods below the prevailing world price, accumulates capability by doing so, and moves up the value chain. Nothing in that account has stopped being true. What has changed is the number of ways a plant can fail to be competitive that have nothing to do with wages.

Table 1 — The four constraints

Constraint	Measured in	Fails as	Instrument that relieves it	Timescale
01 Power price	Cost per kWh as a share of unit cost	A permanent margin disadvantage	Clean generation at scale; tariff design; siting near cheap resource	Years — set by the generation build
02 Power reliability	Outage hours; sales lost per outage	Scrapped batches, idle lines, lost orders	Firm capacity, storage, grid investment, direct contracting, on-site supply	Hours — but decisive for order retention
03 Physical exposure	Output lost per °C; flood recurrence at the site	Productivity decay, absenteeism, total stoppage	Cooling, building design, siting, drainage, logistics redundancy	Days for heat, decades for siting
04 Carbon compliance	Embedded emissions per tonne of product	Loss of market access or a border levy	Clean process energy; measurement and verification capability	Regulatory calendar — already live

Compiled by the author from the sources cited in Sections 2 to 4. The point of the table is the variation across columns: these are not four versions of the same problem.

Two features of that table drive the rest of the report. The first is that the instruments barely overlap: nothing a finance ministry can do about the price of electricity will keep a flooded industrial estate open, and no amount of drainage will satisfy a European customs declaration. The second is that the timescales differ by orders of magnitude, from the four hours that ruin a semiconductor batch to the forty years a factory sits where it was built.

2. Power: Price and Reliability Are Different Constraints

They are usually discussed together as "energy costs". They behave differently, they bind different industries, and conflating them is the most common error in industrial energy planning.

2.1 Price, for the industries where electricity is the input

For a narrow set of industries, electricity is not overhead but feedstock. Electricity is the largest single operating cost in primary aluminium smelting — on industry estimates around 40 per cent of production cost — and primary aluminium is more energy-intensive per tonne than either steel or cement (World Economic Forum). Ammonia, chlor-alkali, electric-arc steel and, prospectively, green hydrogen belong to the same class. For these, a two-cent difference in the power price is the entire investment case, and the location decision is effectively a decision about where clean electricity is cheapest.

This is the basis of the "powershoring" proposition — that energy-intensive plants under cost and compliance pressure will relocate to economies with clean, cheap grids. It deserves emphasis because genuinely new sources of comparative advantage are rare, and this is one: a country that has spent a decade building hydropower, solar or wind now holds an input that cannot be shipped and that a carbon-constrained industry must have.

2.2 Reliability, for the industries where an outage is the loss

A second set of industries is almost indifferent to the tariff and acutely sensitive to interruption. A pharmaceutical batch, a semiconductor wafer, a cold chain and a precision assembly line all lose far more from four unplanned hours than from a year of paying two cents more per kilowatt-hour.

The aggregate cost of that is large and well measured. Analysis of World Bank Enterprise Survey data covering more than 143,000 firms in 137 countries puts sales losses from electricity outages at about US$82 billion a year, with losses exceeding ten per cent of sales in the countries with the least reliable grids (Policy Research Working Paper 8899).

Figure 1 — Share of firms experiencing electrical outages, by region

Per cent of surveyed firms reporting electrical outages, latest available World Bank Enterprise Survey aggregates (indicator IC.ELC.OUTG.ZS, 2025 release). Regional aggregates cover firms of all sizes and sectors surveyed in each region.

US$1.4bn

Cost of the May–June 2023 power shortage in northern Vietnam

World Bank estimate

0.3%

Of Vietnam's GDP, lost in roughly six weeks

World Bank estimate

10%

Revenue loss reported by surveyed northern industrial firms

World Bank surveys

Vietnam in mid-2023 is the clearest recent illustration, and it is also the case that ties this section to the next. Record heat raised demand at the same moment drought emptied the hydropower reservoirs; the northern industrial provinces of Bac Ninh and Bac Giang — home to Samsung, Foxconn, Canon and Luxshare plants — went dark intermittently. The World Bank estimated losses of US$1.4 billion, about 0.3 per cent of GDP, with surveyed industrial firms reporting revenue losses of around ten per cent.

Note what that episode was. It was not a price shock, a fuel shortage or a policy failure in the ordinary sense. It was a climate event transmitted through the power system into the order books of some of the most sophisticated factories in Asia — which is to say, the reliability constraint and the exposure constraint are the same event seen from two places.

Why the distinction changes the policy

A government that reads "energy costs" as one problem will subsidise tariffs. That helps the smelter and does nothing for the electronics plant, which needed firmness — storage, reserve margin, a direct contract, on-site generation — and would happily have paid for it. Vietnam's Decree 80/2024 on direct power purchase agreements is an example of the right instrument for the second group: it lets a manufacturer contract clean supply directly from a generator, either through the grid or over a private wire, meeting both a reliability need and a corporate procurement mandate.

3. Physical Exposure: Heat Erodes, Flood Stops

The third constraint operates on two timescales. Heat lowers output continuously and invisibly; flooding removes it all at once. Both are site decisions before they are operating decisions.

3.1 Heat, measured inside the factory

The best evidence on heat and manufacturing is a national panel of Indian factories analysed by Somanathan, Somanathan, Sudarshan and Tewari in the Journal of Political Economy. Annual plant output falls by about two per cent per degree Celsius. On hot days, worker productivity declines by four to nine per cent per degree, with absenteeism rising as well. Critically for this report's argument, climate control substantially mitigates the loss — which converts a climate hazard into an equipment and electricity question.

Figure 2 — What a degree of heat costs a factory

Per cent loss per degree Celsius, from Somanathan et al. (2021). The annual figure is plant-level output across a national panel; the hot-day figure is worker productivity, shown as the reported 4–9 per cent range. The two measure different things on different timescales and are shown together to contrast the continuous loss with the acute one.

At the level of the labour force, the ILO's Working on a Warmer Planet projects that in 2030 heat stress will cost 2.2 per cent of total working hours worldwide — 80 million full-time jobs and US$2,400 billion. Agriculture takes 60 per cent of that loss and construction 19 per cent, with southern Asia and western Africa losing around five per cent of working hours. The projection assumes warming is held to 1.5 °C and that agricultural and construction work happens in shade, so it is a conservative floor rather than a central case.

Figure 3 — Working hours projected lost to heat stress in 2030

Per cent of total working hours lost, projected for 2030 under a 1.5 °C assumption, from the ILO. The regional figures are the report's approximate values for the worst-affected regions; the world figure is equivalent to 80 million full-time jobs.

3.2 Flood, measured across the supply chain

Flooding does not erode output; it ends it, and it does so beyond the perimeter of the flooded site. Thailand's 2011 floods inundated seven industrial estates and 904 factories, more than half of them Japanese-owned, and caused economic damage the World Bank put at about US$45.7 billion, of which roughly US$32 billion fell on manufacturing. Because Thailand then produced a quarter of the world's hard disk drives, plants on other continents that were never touched by water cut production for want of components.

Figure 4 — Thailand 2011: the anatomy of an industrial flood

US$ billions of economic damage, from World Bank and GFDRR assessment of the 2011 floods. Seven industrial estates and 904 factories were inundated; those counts are given in the text rather than plotted, since they are not in the same units.

The lesson repeatedly drawn from 2011 was about inventory and supplier diversification. The lesson relevant here is about siting: industrial estates are typically placed on flat, cheap land near rivers and ports, which is a description of a flood plain. That decision is taken once, by an estate developer optimising land cost, and every tenant inherits it for the life of the plant.

The fix for one constraint is a load on another

Climate control is what makes the heat constraint tractable inside a factory — and it is electricity demand, arriving in the hours when the grid is already strained by everyone else's cooling. The exposure fix therefore raises the reliability requirement, and the reliability fix has to be built with that demand in it. A companion report works through this mutual dependency between protective services and clean supply in general terms; in industrial planning it appears as a simple sequencing rule: size the power system for a factory that is cooled, not for the factory as it operates today.

4. Market Access: The Carbon Rules of the Destination

The fourth constraint is unlike the other three: it is not a physical fact about the plant but a regulatory fact about its customers, and it is already in force.

The EU's Carbon Border Adjustment Mechanism entered its definitive phase on 1 January 2026. Importers must now buy and surrender certificates against the verified embedded emissions of covered goods, with proposals to extend coverage to steel- and aluminium-intensive downstream products from 2028. The mechanism covers six sectors: cement, iron and steel, aluminium, fertilisers, electricity and hydrogen.

Read that list against Section 2.1 and something useful appears. Every CBAM sector is a power-price-bound industry. This is not a coincidence — both facts follow from energy intensity — but it has a practical consequence that is easy to miss: for these industries, cheap clean power is simultaneously the cost strategy and the market-access strategy. There is no trade-off to manage, only a single investment that resolves two constraints. No other archetype in this report gets that.

What these industries do need in addition is unglamorous and often missing: the capability to measure and verify embedded emissions to the standard a foreign customs authority accepts. A producer with genuinely clean power and no verified emissions data faces the same levy as a dirty one. That is a statistical and institutional investment, not an energy one, and it is cheap relative to what it protects.

A caution on scope

CBAM is the most developed instance of this constraint, not the only one, and this report uses it as the worked example rather than the whole category. Buyer-side requirements — brand commitments on renewable electricity, supply-chain emissions reporting, product rules in other jurisdictions — apply the same pressure to sectors CBAM does not cover, including the labour-intensive exporters in the third archetype, whose customers increasingly ask for clean power that their grids may not supply.

5. Three Archetypes and a Screening Test

Sorting industries by their binding constraint produces three groups with different pathways — and makes visible which policy instruments will do nothing for which sector.

Table 2 — Which constraint binds which industry

Archetype	Examples	Binds first	Also material	Screening question
Price-bound	Aluminium, ammonia and fertiliser, cement, electric-arc steel, green hydrogen, chlor-alkali	Power price	Carbon compliance — the same six sectors CBAM covers	Would a two-cent change in the power price change the investment decision?
Reliability-bound	Electronics and precision assembly, pharmaceuticals, cold chain and agro-processing, data centres	Power reliability	Exposure, through the cooling load and the logistics chain	What does one unplanned four-hour outage cost, and how often does it happen?
Exposure-bound	Apparel and footwear, light assembly, food processing, furniture	Physical exposure — heat inside, flood outside	Buyer-side carbon requirements; reliability where cooling is installed	What is lost per hot day and per flood year at this specific site?

The archetypes and screening questions are the author's analytical framework, constructed with the cited evidence already known; the fit is therefore not evidence for the framework. Industries are placed by their dominant constraint, and many firms sit across two.

Figure 5 — The binding-constraint screen

A conceptual schematic of the screen described in Table 2, not a quantitative model. Most real firms sit across two archetypes; the screen asks which constraint binds first, because that is where the marginal money belongs.

The screen's practical value is negative as much as positive. It says what will not work: a power-tariff subsidy aimed at an electronics cluster that needed firmness; a resilience grant for a smelter whose problem is the price per kilowatt-hour; a certification programme for an apparel exporter whose plant loses a month of output to a flood. Each of those is a real policy that has been tried somewhere, and each addresses a constraint that was not binding.

6. Pathways and Sequence

Three pathways, the shared infrastructure that serves more than one, and the reason sequence matters more than the total.

6.1 The price-bound pathway

Build clean generation at a scale and price that makes energy-intensive processing viable, and treat the emissions measurement and verification system as part of the same investment. This pathway has an unusual property: it is the only one where the cost constraint and the compliance constraint are relieved by the same expenditure, which makes the internal case for it far stronger than a carbon argument alone would be. Where a country's clean resource is genuinely cheap, the powershoring proposition is available — an opportunity to attract relocating industry rather than merely to retain existing plants.

6.2 The reliability-bound pathway

Buy firmness, and let the firms pay for what they value. The instruments are storage, reserve margin, distribution investment where the losses actually occur, and contractual mechanisms such as direct power purchase agreements that allow a manufacturer to procure clean, firm supply without waiting for the whole system to improve. The reason to prefer clean firmness over diesel is not only emissions: diesel backup is an operating cost that scales with the frequency of failure, which is rising.

6.3 The exposure-bound pathway

Cool the building, site the estate above the flood line, and drain what is already built. The heat evidence says climate control substantially recovers the lost output, which converts an intractable-sounding hazard into a capital expenditure with a measurable return — and, as Section 3 noted, an electrical load that the power pathway has to anticipate. The siting decisions are the ones with no second chance, and they are made by estate developers and permitting authorities rather than by tenants.

6.4 Where the pathways overlap — and where they compete

The overlaps are what make an integrated plan cheaper than three separate ones. A firm, clean grid serves the price-bound and reliability-bound archetypes simultaneously. Industrial-zone drainage and raised siting serve every tenant regardless of archetype, which is why zone-level provision beats firm-level provision for that constraint. Emissions measurement capability, built once at national level, serves every exporter.

The competition is equally real. Cooling raises demand at the system peak. Capital spent on relocation is capital not spent on protecting what exists. And the political economy of industrial policy favours the visible new plant over the drainage channel that prevents an invisible future loss — the same asymmetry an earlier report traced in prevention spending generally.

6.5 Why sequence dominates

An industrial plant is a forty-year asset whose exposure is fixed on the day the site is chosen and whose energy costs are largely fixed by the grid it connects to. A plant built this decade in a flood plain on a weak grid, to save capital at the moment of construction, carries an operating penalty that no later subsidy fully removes: the drainage can be retrofitted at several times the cost, the grid may improve, but the site cannot be moved. That is why, for economies industrializing now, the order of investments matters more than the total — and why the screen in Section 5 is most valuable before the first spade goes in.

7. Recommendations

What follows for the actors who plan, finance and host industrial capacity.

7.1 Planning and industry ministries

Run the screen before the instrument. Classify target sectors by binding constraint — price, reliability, exposure, compliance — and check that every proposed measure relieves a constraint that actually binds the sector it is aimed at.
Separate price policy from reliability policy. Tariff subsidies and firmness are different products with different beneficiaries; conflating them under "energy costs" reliably overspends on one and underspends on the other.
Treat emissions measurement and verification as trade infrastructure. With CBAM's definitive phase live since January 2026, an exporter with clean power and no verified data pays as though it had neither.
Make flood-line siting a permitting condition for new industrial estates, not a recommendation. It is the one decision in this report that cannot be revisited.

7.2 Industrial zone operators and utilities

Provide protection at zone level. Drainage, raised platforms and shared cooling and power infrastructure are cheaper per tenant than firm-by-firm provision, and they are the reason a tenant chooses one estate over another.
Size the power system for a cooled factory. Cooling load is not an add-on to the industrial demand forecast; in hot climates it is a large and growing part of it.
Offer firmness as a product. Reliability-bound tenants will pay for guaranteed supply — through storage, on-site generation or contracted clean power — and pricing it explicitly funds the investment.

7.3 Development finance and investors

Underwrite the shared layer. Grid firmness, zone drainage and national measurement systems are public-good investments whose returns accrue across every tenant and exporter, and which no single firm will fund.
Price resilience into project appraisal, so that a plant sited above the flood line on a firm grid is financed on better terms than one that is not — as it should be, since it is a better credit.
Back the powershoring proposition with real instruments in countries whose clean resource is cheap: long-tenor finance, guarantees against offtake and currency risk, and early anchor tenants.

7.4 Exporters and industry associations

Measure what the constraint costs you. Output lost per hot day, sales lost per outage hour, embedded emissions per tonne: firms that have these numbers can make the investment case; firms that do not will be told their problem is wages.
Buy clean firm power contractually where the framework allows it, rather than waiting for the system average to improve.

8. Conclusions

Climate has not made industrialization impossible for emerging economies. It has made it a four-variable problem in which the variables bind unevenly.

The evidence for each constraint is unambiguous. Electricity is the largest operating cost in the industries that the EU's border mechanism now covers. Outages cost surveyed firms around US$82 billion a year in lost sales, and a single climate-driven shortage cost Vietnam US$1.4 billion in six weeks. A degree of heat takes about two per cent of an Indian factory's annual output and four to nine per cent of worker productivity on hot days. A flood took 904 factories offline in Thailand in one season and interrupted supply chains that never saw the water.

What is less often said is that these four facts do not describe one problem. They describe four, with different units, different instruments, different timescales and different victims. The smelter and the semiconductor plant and the garment factory are not three instances of climate-exposed industry; they are three different industries with three different binding constraints, and the policy that helps one will pass the others by.

Sorting by binding constraint also identifies where the fortunate cases are. For the energy-intensive processors, cheap clean power resolves the cost problem and the market-access problem in a single investment — the only place in this report where two constraints fall to one expenditure, and the foundation of a genuinely new comparative advantage for countries with clean grids. For everyone else, the work is less glamorous: firmness bought and priced, buildings cooled, estates drained and sited above the water, emissions measured to a standard a foreign customs officer will accept.

The reason to do this now rather than later is that industrial capital is being sited now. A plant placed this decade will still be there in the 2060s, in whatever climate arrives, connected to whatever grid was built around it. The constraints will tighten. The sequence chosen in the next few years determines which of them a country's industry is still able to clear.

References

Every quantitative claim above is attributed inline. The principal sources are collected here.

World BankUnderutilized Potential: The Business Costs of Unreliable Infrastructure in Developing Countries, Policy Research Working Paper 8899 — 143,000 firms in 137 countries; about US$82 billion a year of sales losses; more than 10 per cent of sales where grids are least reliable. See also the Enterprise Surveys indicator on value lost to outages.
World Bank, via contemporary reportingPower outages cost Vietnam $1.4B — the May–June 2023 northern Vietnam shortage, 0.3 per cent of GDP, and 10 per cent revenue losses reported by surveyed industrial firms. Context on affected industrial hubs in Vietnam Briefing.
Somanathan, Somanathan, Sudarshan & TewariThe Impact of Temperature on Productivity and Labor Supply: Evidence from Indian Manufacturing, Journal of Political Economy 129(6), 2021 — about 2 per cent of annual plant output per °C; 4–9 per cent productivity loss per °C on hot days; climate control mitigates. Working paper version here.
ILOWorking on a Warmer Planet: The Effect of Heat Stress on Productivity and Decent Work — 2.2 per cent of working hours lost in 2030, 80 million full-time jobs, US$2,400 billion; agriculture 60 per cent and construction 19 per cent of the loss. Summary here.
World Bank / GFDRRThai Flood 2011: Rapid Assessment for Resilient Recovery and Reconstruction Planning — about US$45.7 billion of economic damage, seven industrial estates and 904 factories inundated, roughly US$32 billion of manufacturing damage.
European Commission implementation guidanceCBAM's definitive phase from 1 January 2026 — certificate obligations, the six covered sectors, and the proposed extension to downstream steel and aluminium products from 2028.
World Economic ForumNet-Zero Industry Tracker 2023: Aluminium — electricity as the dominant cost and emissions driver in primary aluminium.
CAFPowershoring — relocating energy-intensive plants to economies with clean, competitive electricity; see also powershoring and the green neoindustrialization of Brazil.
IEEFAVietnam's direct power purchase agreement decree could catalyse a new era for renewable energy — Decree 80/2024/ND-CP, its two models, and its significance for manufacturers with clean-power mandates.
H HeuristicsRelated reports in this series: integrating adaptation and mitigation, on the mutual dependency between protection and clean supply; the economics of prevention, on why protective spending loses to visible alternatives; and adaptation and convergence.

Metadata

Keywords: industrializationgreen industrial policyclimate resilienceemerging economieselectricity reliabilityheat and productivityflood riskCBAMcarbon border adjustmentpowershoringindustrial zonesmanufacturing competitivenessbinding constraintgreen convergence
Topics: Emerging Markets Energy Transition Infrastructure Climate Adaptation
JEL classification: O14, Q54, L60, F18, O25 — industrialization and manufacturing; climate and natural disasters; industry studies: manufacturing; trade and environment; industrial policy
Data and method: This report synthesises published institutional and peer-reviewed evidence rather than producing new modelling. Firm-level outage costs are from World Bank Enterprise Survey analysis in Policy Research Working Paper 8899; Vietnam's 2023 shortage from World Bank estimates as reported at the time; heat effects on manufacturing output from Somanathan, Somanathan, Sudarshan and Tewari in the Journal of Political Economy (2021); working-hour losses from the ILO's Working on a Warmer Planet; the 2011 Thailand flood damages from World Bank and GFDRR assessments; CBAM's definitive phase and covered sectors from EU implementation guidance; the powershoring proposition from CAF and IDB analysis; and Vietnam's direct power purchase agreement framework from Decree 80/2024/ND-CP as reported by IEEFA and legal analyses. All figures were verified against their sources in September 2026. The three archetypes in Section 5, Table 1's constraint matrix and the screening questions derived from them are the author's analytical framework, constructed with the cited cases already known; the framework's fit to them is therefore not evidence for it, and it is offered as a planning checklist rather than a predictive model. Figure 5 is a conceptual schematic. Figures drawn from different studies use different years, methods and country samples and are not additive. The report is analytical rather than predictive.
Report: H Heuristics Digital Report № 2026-13 · Published 17 September 2026
Licence: CC BY-NC-ND 4.0
Cite as: Hunter Hughes (2026). Climate-Resilient Industrialization Pathways: Sorting emerging-economy industry by the constraint that actually binds it. H Heuristics Digital Report 2026-13. https://digitalreports.hheuristics.com/reports/climate-resilient-industrialization-pathways/

↑ Back to top

[1] World BankUnderutilized Potential: The Business Costs of Unreliable Infrastructure in Developing Countries, Policy Research Working Paper 8899 — 143,000 firms in 137 countries; about US$82 billion a year of sales losses; more than 10 per cent of sales where grids are least reliable. See also the Enterprise Surveys indicator on value lost to outages.

[2] World Bank, via contemporary reportingPower outages cost Vietnam $1.4B — the May–June 2023 northern Vietnam shortage, 0.3 per cent of GDP, and 10 per cent revenue losses reported by surveyed industrial firms. Context on affected industrial hubs in Vietnam Briefing.

[3] Somanathan, Somanathan, Sudarshan & TewariThe Impact of Temperature on Productivity and Labor Supply: Evidence from Indian Manufacturing, Journal of Political Economy 129(6), 2021 — about 2 per cent of annual plant output per °C; 4–9 per cent productivity loss per °C on hot days; climate control mitigates. Working paper version here.

[4] ILOWorking on a Warmer Planet: The Effect of Heat Stress on Productivity and Decent Work — 2.2 per cent of working hours lost in 2030, 80 million full-time jobs, US$2,400 billion; agriculture 60 per cent and construction 19 per cent of the loss. Summary here.

[5] World Bank / GFDRRThai Flood 2011: Rapid Assessment for Resilient Recovery and Reconstruction Planning — about US$45.7 billion of economic damage, seven industrial estates and 904 factories inundated, roughly US$32 billion of manufacturing damage.

[6] European Commission implementation guidanceCBAM's definitive phase from 1 January 2026 — certificate obligations, the six covered sectors, and the proposed extension to downstream steel and aluminium products from 2028.

[7] World Economic ForumNet-Zero Industry Tracker 2023: Aluminium — electricity as the dominant cost and emissions driver in primary aluminium.

[8] CAFPowershoring — relocating energy-intensive plants to economies with clean, competitive electricity; see also powershoring and the green neoindustrialization of Brazil.

[9] IEEFAVietnam's direct power purchase agreement decree could catalyse a new era for renewable energy — Decree 80/2024/ND-CP, its two models, and its significance for manufacturers with clean-power mandates.

[10] H HeuristicsRelated reports in this series: integrating adaptation and mitigation, on the mutual dependency between protection and clean supply; the economics of prevention, on why protective spending loses to visible alternatives; and adaptation and convergence.