A Forest Per Worker: Quantifying the CO2 Reduction Contribution of the Marginal “Green” (Migrant) Worker

Abstract

Workforce constraints are a widespread bottleneck to decarbonisation, hindering implementation and investment. In a novel exercise, we model the decarbonisation impact of the marginal contribution of a “green-skilled” worker in contexts of labour shortage across six countries in two occupations: an electrician installing residential rooftop solar photovoltaic (PV) panels, and a heating technician installing residential heat pumps, during the period 2024 – 2032. We find that the additional (“marginal”) worker can contribute thousands of tonnes of CO₂ abatement, even when accounting for rapid grid decarbonisation. The marginal worker’s contribution can have a monetised social value of hundreds of thousands of dollars. It is the equivalent of planting thousands of trees: a forest per worker.

On this basis, we argue that labour shortages must not be allowed to constrain decarbonisation activities. Where domestic training cannot meet demand, labour migration is a valuable policy tool. Because workers may make larger contributions in countries of origin than in countries of destination, however, we note that care must be taken to avoid implementation gaps caused by brain drain in countries of origin. Partnerships that combine training and labour migration partnerships can mitigate these risks.

Introduction

Workforce constraints are widely reported to be a key bottleneck in decarbonisation. Shortages of skilled workers are already reported to be constraining the implementation of decarbonisation plans and limiting investment (IEA, 2022; IEA, 2024). The International Renewable Energy Agency notes the challenge of a “critical skills gap” (IRENA, 2025: 87). The problem is expected to grow in the coming years as green transition deadlines draw closer.

As a key limiting factor in decarbonisation, a shortage of workers has a meaningful impact on carbon emissions and the achievement of targets. One recent analysis (Hambrecht et al., 2025) suggests that globally, workforce shortages could increase power sector emissions to 12 percent above 2030 pledges and more than 100 percent above 2045 commitments.

In this paper, we model for the first time the marginal decarbonisation contribution of an additional worker in contexts of labour shortage across six countries in two occupations: an electrician installing residential rooftop solar photovoltaic (PV) panels, and a heating technician installing residential heat pumps. We model a period of work from the start of 2024 to the end of 2032. We find that individual workers, if filling a labour shortage-induced implementation gap, can lead to thousands of tonnes of CO₂ abatement, even when accounting for broader grid decarbonisation. This decarbonisation has a monetised social value of hundreds of thousands of dollars: a solar panel installer working in Italy, for example, will contribute abatement with a social value of over US$280,000. It is the equivalent of planting a forest of trees: more than 6,500 trees would need to grow for 50 years to sequester the equivalent amount of carbon.

On this basis, we argue that if domestic training pipelines are insufficient to meet workforce demand for decarbonisation within the urgent timeframes required, legal, skills-based labour migration programmes should be used to ensure that implementation gaps do not persist. Our chosen countries comprise three possible countries of destination facing shortages (the United Kingdom, Germany, and Italy) and three possible countries of origin interested in agreeing training and migration partnerships (India, the Philippines, and Kenya). All three countries of destination are already, to varying degrees, leaning on immigration to meet decarbonisation workforce needs.

However, migration of “green-skilled” workers can also pose a carbon risk. High-income countries typically have lower carbon emissions per kilowatt-hour of generated grid electricity than lower-income countries, and also have more ambitious grid decarbonisation plans. Moving a “green” worker—for example, an electrician installing solar PV—from a lower-income country to a higher-income country could lead to a large negative impact on net decarbonisation. For this reason, we argue, countries of destination must take care when recruiting to ensure that they are not leaving an implementation gap in countries of origin, harming a global public good. Instead, they should target underemployed workers or collaborate with countries of origin to train more workers, ensuring that the global stock of skilled workers rises and that tasks crucial to decarbonisation are not left undone due to labour shortages.

In the first part of this paper, we set the context. We summarise the decarbonisation targets to which countries have committed; the shortages of skilled labour currently experienced and anticipated; and current use of labour migration policy.

In the second part, we show the results of an exercise in modelling the carbon emissions reduction contribution of a marginal worker across several scenarios. We briefly explain modelling parameters; more detail is available in the associated methodology annex.

In the final parts of the paper we provide policy suggestions. We propose four models of international recruitment and partnership, depending on conditions in the countries of origin and destination. We also provide broader policy conclusions, noting that underemployed workers, such as refugees without work rights, should be targeted for recruitment, and that harmonisation of curricula or qualification recognition should be a priority to ensure that reallocation of workers into implementation gaps can be facilitated.

Context

Decarbonisation targets

This paper focuses on six countries, all of which have stated commitments to reducing emissions:

• The UK targets a figure of 95 percent clean power generation by 2030 and a fully decarbonised electricity system by 2035 (Bolton, 2025a), on the way to net zero emissions by 2050 (Burnett and Stewart, 2025).
• Germany aims for 80 percent renewable energy generation by 2030 and carbon neutrality by 2045 (European Commission, 2024a).
• Italy aims for 71 percent renewable energy generation by 2030 (European Commission, 2024b) and carbon neutrality by 2050 (Erbach, 2024).
• India targets clean power generation contributing around 39 percent in 2026/2027 and 44 percent by 2032, with a further ~4 percent contributed by nuclear to give a total of roughly 48 percent (Ministry of Power, 2023).
• The Philippines aims for 35 percent renewable energy generation by 2030 (Climate Action Tracker, 2023), and a 75 percent reduction in greenhouse gas emissions by 2030 versus business-as-usual during the period 2020 – 2030 (NDC Partnership, 2025).
• Kenya aims to generate 100 percent of its electricity from clean energy sources by 2030 (CIF, 2024) and to cut overall carbon emissions by 35 percent by 2035 versus a business-as-usual scenario (Government of Kenya, 2025).

These high-level targets are enacted through interventions in markets to encourage rapid adoptions of new technological standards: for example, transitions from gas boilers to heat pumps, internal combustion engines to batteries, or single-glazed to double-glazed windows. They translate to technology-specific installation goals.

The UK, for example, has adopted the target of installing 600,000 heat pumps per year by 2028, rising to 1.6 million per year by 2035 (DESNZ & BEIS, 2023; Smalley and Sweeney, 2025). It has also set a target of increasing the total installation capacity of solar PV to 45 – 47 gigawatts (GW) by 2030, rising from 18.1 GW in 2025 (Hutton et al., 2025), and of nearly 80 GW of onshore and offshore wind (DESNZ, 2025). In Germany, the government has set itself a legally binding target of 215 GW of solar capacity by 2030 and around 400 GW by 2040, rising from 107.5 GW in July 2025 (Sternberg, 2025). It also targets 500,000 heat pump installations per year from 2024, with 6 million in operation by 2030 (BMWE, 2025). Italy aims for nearly 80 GW of solar capacity by 2030 (European Commission, 2024b), versus 40 GW in 2025 (Casey, 2025). It does not have a heat pump installation target, but does target sectoral renewables in heating increases expected to necessitate the installation of 8.6 million heat pumps by 2030 (Marchesini et al., 2025).

Lower-income countries have also adopted ambitious technology-specific targets. India targets 300 GW of solar capacity by 2030 (Rossi et al., 2025), and had installed 123 GW by August 2025 (MNRE, 2025a). Its PM Surya Ghar programme, a flagship federal initiative launched in 2024 and targeting the installation of 10 million residential rooftop systems by the end of 2027, will add 27 GW of new capacity (MNRE, 2025b). The Philippines targets 16.6 GW of solar capacity by 2029, versus a total 2.7 GW after a record annual solar deployment of 1.1 GW in 2024 (Rossi et al., 2025). Kenya targets 100 percent clean power by 2030; most renewable energy will be derived from geothermal and hydropower, but solar PV is also expected to make a major contribution (CIF, 2024), particularly through off-grid installations aiding the achievement of its universal energy access by 2030 (Kenya Power, 2025). Each installation of an off-grid household solar PV system is estimated to contribute emissions reductions of 431kg CO­­­­₂ per year (GOGLA, 2020).

The role of labour in achieving decarbonisation targets

Decarbonisation policies represent major government interventions to correct longstanding market failures (see e.g., Altenburg and Rodrik, 2017). Due to the urgency of the need to redress high emissions, the industrial policies deployed have extremely tight timeframes for success.

This requires a coherent approach to solve numerous challenges simultaneously. Governments must ensure that there is sufficient public finance available to entice private finance into clean technologies; disincentivise the use of high-emissions fossil-based technologies; overhaul processes that could slow take-up, including permitting systems; rapidly improve electricity grids; address communications issues and misconceptions of ‘green’ technologies; support further technological innovations; and solidify supply chains—among other key bottlenecks that must be addressed (IEA, 2024b; Addison et al., 2025).

This set of interventions encompasses changes to the policy environment, the adoption of new practices, securing of physical inputs, and provision of capital. Within that, supply of labour is a crucial input without which the desired tasks will not be completed and desired outputs will not be achieved. Figure 1 summarises the interrelationship of key elements in production.

Figure 1. Visualisation of the production function

Image

Source: Adapted from Granata and Posadas (2024).

As is evident in Figure 1, labour is not the only key input. The usefulness of other key inputs is, however, contingent on the supply of labour. In this paper, we focus on the decarbonisation impact of labour supply at the margin, and assume that other key variables in decarbonisation are addressed elsewhere in the policy system.

If these other variables fall away, labour demand will decline. This includes the presence of an enabling policy environment. Without policy interventions to drive consumer demand and ease the scaling of new technological take-up, there will not be a need for labour supply: decarbonisation-related tasks will not be viable within the timeframe needed, and emissions will not be lowered. For example, the UK faces an enormous need for a larger heat pump workforce to hit targets— but heat pump take-up is currently low, and incentivising greater rollout (and consequently higher labour demand) requires further policy assistance to rebalance price differentials between gas and electricity. Heat pumps need a price ratio of less than 2.5 to be perceived as attractive (Zackariat et al., 2025); in the UK and Germany, however, electricity costs over three times more than gas on a per unit basis (DESNZ, 2025b; SMARD, 2025). Levy reforms could increase the speed of the UK heat pump rollout by 60 percent (Sissons et al., 2025): when they occur, labour demand will spike.

Similarly, while solar PV is a cheap source of electricity across all countries of study, and its take-up is already happening at record scale (Rossi et al., 2025), further reforms are often still needed to accelerate it sufficiently. In Germany, for example, rooftop solar on residential buildings could provide up to 28 percent of the additional solar capacity needed to meet 2030 climate targets— but making this happen will require improved administrative processes and easier grid connections via smart meters (Fischer and Henger, 2025).

For labour supply to exist the policy environment must, furthermore, be predictable and reliable. Meeting decarbonisation targets requires policy that incentivises the completion of emissions-reducing targets, and in so doing creates sustained labour demand. Governments frequently state that this labour demand will create jobs to be filled by trained domestic workers. For this to happen, they must be trained to the level necessary in the required occupations. Training provision, however, frequently lags behind demand: training pipelines have multi-year lead times (apprenticeships typically take around four years to complete) and providers invest in training when demand looks durable (see e.g. Huckstep and Dempster, 2025). Where policy is perceived to be volatile or unpredictable, firms and training providers will rationally under-invest in training capacity, and potential trainees will doubt that credentials will be useful (see e.g. Barnes, 2025).

For this reason, reliable and predictable industrial policy is fundamental: companies, training institutions, and workforce entrants must anticipate sustained and predictable labour demand. The less reliable and predictable industrial policy is, the more likely it is that labour migration will ultimately be needed. Where there are fears of volatility of demand, the domestic training landscape will under-invest: investment and domestic supply might later catch up with demand, but there will be a period where migration is needed to act as a bridge. Sourcing workers to meet tight timelines will require international recruitment.

The scale of skilled labour shortages

As decarbonisation timeframes become tighter and the volume of tasks to be completed grows, the number of workers needed will rise and shortages will increase. This is challenging given that shortages are already widely reported to be hindering project implementation and increasing costs (see e.g. IEA, 2024; Huckstep and Dempster, 2025; Rossi et al., 2025).

In many countries, the decarbonisation drive must compete with other recruitment pressures. Alongside emissions reduction commitments, numerous governments have also pledged to rapidly increase their housing stock and to rearm. In the United Kingdom, for example, the Labour government’s target of an additional 300,000 homes constructed per year has been estimated to require an additional 500,000 to 1 million additional construction workers— including in occupations required for achieving net zero emissions, such as electricians and heating technicians (Barnes, 2025). In Germany, the commitment to raise defence expenditure to over 2 percent of GDP has driven defence industry recruitment at a new scale; undersupply in key occupations shared with decarbonisation, such as welders, is already reported to be constraining expansion (Swaney, 2025).

Simultaneously, labour markets in countries of destination also face a structural demographic shift due to ageing populations (OECD, 2025). This is leading to large numbers of retirements within the period in which the labour force must instead be expanding.

In the UK, the qualified electrician workforce in England is estimated to have fallen by 26 percent during 2018 – 2024, with a further decline of around 32 percent anticipated if current trends persist, driven by falling apprenticeship numbers (a 10 percent year-on-year decline in 2024-25 alone) and rising retirements (JTL Training et al., 2025). There is a projected electrician shortfall of around 15,000 by 2030 (Eldred, 2025), with broader workforce gaps also anticipated across construction (CITB, 2025). The UK Government has highlighted shortages in multiple key occupations (DESNZ, 2025c), and numerous sector actors have warned of skilled workforce gaps in the tens or hundreds of thousands in the coming years (Huckstep and Dempster, 2025). These gaps are already reported to be driving up costs, leading to project delays, and disincentivising investment.

In Germany, energy and electrical trades are repeatedly registered as being in shortage in the Federal Employment Agency’s labour market assessments (Bundesagentur für Arbeit, 2024b). The scale of the energy and electrical trades gap has been estimated at nearly 60,000 workers in 2025, in the context of a broader shortage of nearly 165,000 STEM (science, technology, engineering and mathematics) workers (Anger et al., 2025). In the medium term, these shortages are not projected to abate (Zika et al., 2024). Acute shortages are concentrated in trade occupations essential for rooftop solar PV installation and heat pump deployment: construction electricians and heating and air conditioning technicians are among the most in-demand ‘bottleneck’ roles, with an estimated shortage of 18,300 electricians and 12,200 heat and air conditioning technicians in 2024 (Büchel et al., 2025).

In Italy, recruitment difficulties are reported to be increasing in the electrical sector: in 2024, 73.6 percent of companies reported shortages. Shortages were also reported in other green transition-relevant occupations, including welders and engineers (Unioncamere-Excelsior, 2025). Italy anticipates significant workforce challenges due to the retirement of 2.9 million workers by 2028; the construction and manufacturing sectors are projected to be among those with the most challenging labour shortages, due in part to a shortfall of between 15,000 – 17,000 young people entering construction and electrical apprenticeships each year (Unioncamere-Excelsior, 2024).

Shortages in green transition-relevant occupations are not limited to high-income countries. Despite enjoying a demographic dividend, lower-income countries also face challenges.

In India, acute shortages in trained technicians are identified across solar PV, grid integration, and electric vehicle charging. In the context of a government target of electric vehicles contributing 30 percent of new sales by 2030, shortages of tens or hundreds of thousands of electricians and technicians are projected (Singh et al., 2025). Renewable energy projects are already reported to be being delayed by shortages of qualified workers; in government projections of shortages to 2028, multiple electricity generation roles, including in solar, are highlighted, with shortages of tens of thousands of solar PV installers expected in 2025. Shortages are reported to be both quantitative and qualitative, with major upskilling support needed (MSDE, 2025; RenewableWatch, 2025). India has invested in the buildout of the Skill Council for Green Jobs (SCGJ), which has scaled rapidly. The SCGJ aims to have trained one million workers across clean energy sectors by 2030 (IEA, 2023), in a broad range of occupations (see SCGJ, 2024). A sharp rise in demand is anticipated, especially in rooftop solar installation following a policy push (IEA, 2024c).

Kenya is also reported to face shortages of green transition-relevant skilled labour. In the most recent survey of skill needs by the Federation of Kenya Employers (2023), electrical skills were among the TVET (technical and vocational education and training) skills most reported to be in shortage— although, with only 21.1 percent of firms reporting hiring difficulties, shortages may not currently be acute. However, there is relatively little labour market information available relevant to Kenya’s green transition skill gaps.

In the Philippines, manpower shortages and certification are highlighted as key challenges to its rapidly growing solar PV sector (Rossi et al., 2025). However, the pipeline of trainees is large: in 2024 the Philippines’ Technical Education and Skills Development Authority (TESDA) reported over 92,000 graduates from electrical and electronic TVET courses and nearly 3,000 in heating , and nearly 5,000 trainers serving electrical education streams (TESDA, 2024a). Industry stakeholders report a shortage of electrical technicians, in part because the distributed energy sector struggles to compete with the more established power sector (TESDA, 2024b).

The role of labour migration

The role of immigration

In the context of under-developed domestic training pipelines, migration is a valuable pressure valve. It allows employers to meet labour demand within the mandated timeframes, avoiding project delays and cost overruns and buying time for the buildout of the training of new entrants or worker transitions. The fact that labour migration will need to have a role in achieving decarbonisation workforce goals is increasingly recognised, despite the increasingly difficult place of migration policy in the political landscape.

In the UK, labour migration policy adjustments introduced in 2025 focus access to Skilled Worker Visas on occupations relevant to the eight priority sectors of the 2025 Industrial Strategy, one of which is Clean Energy (DBT, 2025). The Industrial Strategy recognises that “changes to the skills system will take time to come to fruition”, and that international recruitment will be needed in the interim. Numerous industry associations and companies have called for access to visas to avoid workforce bottlenecks (see e.g. Huckstep and Dempster, 2025; Barnes, 2025), and the House of Commons’ Energy Security and Net Zero Committee (2025) has recognised that labour migration will be key to achieving policy goals. In the 2025 review of shortage occupations undertaken by the UK’s Migration Advisory Committee, multiple roles relevant to the green transition are included, including electricians and heating and ventilation technicians (MAC, 2025). Several occupational pipelines are currently heavily supported by labour migration, including roofers and welders (Huckstep and Dempster, 2025).

Germany has explicitly linked the need for immigration reforms to decarbonisation-related workforce pressures (BMAS, 2024). Germany conducted significant reforms from 2023, relaxing access to immigration for applicants with vocational experience and reducing German language requirements while easing recognition of qualifications (Schneider, 2023). They also introduced a points-based ‘Opportunity Card’ allowing one-year (renewable) residence for jobseekers (Tollenaere et al., 2024). So far, increased visa grant rates have not fully translated into increased migration rates, possibly due to long visa processing times (Schultz and Mecke, 2024). Germany is also facilitating multiple pilot projects supporting both training and migration of workers with green transition-relevant skills, and has concluded several labour migration agreements (Sanderson, 2024; ZDH et al., 2024). Achieving decarbonisation is expected to require continued high levels of immigration (Büchel et al., 2025).

Italy has recently relaxed its immigration restrictions. The state labour-market intelligence system (Unioncamere-Excelsior, 2024) projects that between 2024 and 2028 around 640,000 immigrant workers will be needed. It has also moved to a programme of three-year labour migration quotas, setting access to non-seasonal visas according to a shortage list. For 2025, 70,720 non-seasonal visas were available (Ministero Dell’Interno, 2024). The 2025 list of occupations included several relevant to decarbonisation, including electricians and heating and ventilation technicians (Confindustria, 2024). Visas are also available outside the quota system for workers trained in ministry-approved vocational and civic-language programmes in non-EU countries (currently a list of 23) who also possess a job offer from an Italian employer (Ministero del Lavoro e delle Politiche Sociali, 2023). Confindustria, the national association of enterprises with 113,000 members, has suggested that Italy must prioritise training and integrating international skilled workers before they enter (Confindustria, 2025). Since 2023, Italy has also had a Migration and Mobility Agreement and joint working group with India (Ministero Degli Affair Esteri, 2024); in 2025 7,000 annual non-seasonal visas were allocated to India above the general quota (MEA, 2023).

The role of emigration

Just as all three high-income countries are increasingly recognising the importance of labour migration in the green transition, the three selected lower-middle-income countries are all prioritising facilitated labour emigration.

India is actively seeking labour migration agreements to promote the mobility of “green-skilled” workers (see e.g. Sarkar et al., 2025; Economic Times, 2024). Its SCGJ has worked closely with several other governments, notably Germany and Australia, on training for international labour markets (see e.g. SCGJ, 2024). Both the SCGJ and the National Skill Development Corporation (NSDC) have arms aimed at training workers to international standards, in collaboration with partner countries. In late 2025, India sought to reform its overseas mobility legislation to enhance cross-government policy coherence in the management of emigration (MEA, 2025).

Kenya has a large youth population, with high rates of youth unemployment (World Bank, 2024). Facilitated emigration has become a prioritised policy tool. In late 2024 Kenya agreed a new migration partnership with Germany (ILO, 2024). The agreement was warmly welcomed by Kenyan President Ruto, who suggested that he hoped 250,000 workers would move annually (Fox, 2024). By contrast, only around 1.25 million Kenyans are estimated to hold qualifications eligible for ‘skilled worker’ classification under Germany’s migration system (Kaltner, 2024), suggesting that support in training is likely to be important. Kenya is now seeking to establish multiple further agreements, including with Canada (Mwangura, 2025), with a stated aim of exporting one million workers per year until 2028 (Ross and Martinez, 2025). In 2023 it established a National Policy on Labour Migration, seeking to better coordinate skills mappings, training provision, and international placements, and to establish bilateral labour migration agreements (Parliament of Kenya, 2023), approved in 2025 (Parliament of Kenya, 2025).

The Philippines has long been a major country of origin, with a longstanding deliberate focus on training workers for international labour markets as part of its national development strategy. Between 2006 and 2019, over 1 million workers were deployed internationally each year (Opiniano and Ang, 2024). The Department of Migrant Workers, created in 2021, provides coordination across agencies, streamlining the bureaucracy of emigration management. In the year to January 2025, nearly 40,000 ‘Craft and Related Trades Workers’ (a group including electricians, welders, and other occupations relevant to decarbonisation) emigrated (DMW, 2025).

Quantifying the marginal decarbonisation contribution of an additional (migrant) worker

What is the marginal decarbonisation contribution of an additional worker in contexts of shortage? We estimate the marginal contribution of labour and report results (the change in avoided emissions from one additional installer) on a per-additional-worker basis. For brevity we refer to this as the “‘marginal worker” effect. A marginal worker is understood to, in the first instance, be contributing to technology installations that would not happen if they were not present; or to be bringing forward installations that would otherwise only happen later, as labour supply catches up to green transition-driven demand.

We select two occupations: electricians installing residential rooftop solar PV, and HVAC (heating, ventilation, and air conditioning) technicians installing heat pumps in households. We choose these two occupation niches because they are: (a) contributing the installation of key technologies to the green transition; (b) recognised to be in widespread and growing shortage; and (c) relatively easily modelled: the technologies installed make discrete abatement contributions. Many other key occupations (including welders constructing offshore wind turbines, retrofit insulation installers, and linesworkers building out the grid) are also reported to be in critical shortage but are less easily modelled; future research could assess their per-worker contributions.

We model the marginal worker’s contributions in three high-income countries facing labour shortages: the UK, Germany, and Italy. We select these countries of destination because they are anticipated to face significant difficulties in sourcing sufficient workers for their decarbonisation goals, and because they are currently taking proactive policy approaches to try to connect immigration policy to their green industrial policy goals. We also model the marginal worker’s contributions in three lower-middle-income countries: India, the Philippines, and Kenya. These are also facing skilled labour shortages, but are seeking to position themselves as potential suppliers of skilled labour to international markets. The marginal worker could be a domestic worker, but given recruitment pressures, we discuss them as though they are a migrant worker.

Our modelling covers the period 2024 to 2032. This is a key period for achieving decarbonisation targets, and a plausible period for labour migration. We take 2024 as the base year due to the fact that the most recent data for several key inputs, including grid emissions factors, is from that year. If applying this analysis to a migrant worker, we assume they arrived at the start of 2024 and work until the end of 2032. They may subsequently remain in the country of destination, and thus contribute to continued abatement, or may return to their country of origin to support what is likely by then to be an accelerating clean energy market.

Our modelling relies on a number of key assumptions and proxies. Further detail is available in the associated methodology document. Importantly, our approach:

• Assumes that grid decarbonisation rates meet stated government targets (also modelling 75 percent achievement of targets);
• Uses average grid emissions factors (carbon intensity of grid-drawn kWh), in the absence of reliable projections of marginal emissions factors; and
• Assumes varying and changing rates of labour utilisation, installation efficiency, and other key installation-related factors, specific to focus countries.

We assume that solar PV installations generate electricity that displaces grid electricity, and that heat pumps replace new high-efficiency gas condensing boilers. In general, we use conservative assumptions: the estimates generated are more likely to understate an installer’s impact than to overstate.

A thousand tonnes of CO₂: The contribution of the marginal (migrant) worker

Solar photovoltaic panels and heat pumps are extremely carbon-efficient means of generating electricity and converting electricity to heat. A single installer, operating where otherwise installations would not take place, can make a major cumulative impact over the nine-year period studied. This is in large part because the emissions reduction effect is cumulative: PV systems and heat pumps continue to abate emissions beyond their year of installation.

In Figure 2 we show emissions reductions, in tonnes of carbon dioxide, achieved by a marginal electrician facilitating the installation of rooftop residential solar systems (with a rising proportion of installations also including batteries). The anticipated number of installations over the time period varies, and is particularly affected by labour utilisation rates, the prevalence of batteries, the size of PV systems, and the prevalence of retrofit installations versus new-builds. In Germany, an electrician is projected to install 202 solar systems over the 2024 – 2032 period; in Italy and the UK, more than 360 are projected to be installed.

Large decarbonisation gains are projected in all countries: nearly 700 tonnes CO₂ in the UK, rising to over 1,000 tonnes in Germany and 1,200 tonnes in Italy. These figures assume that countries’ stated grid decarbonisation targets are achieved, and thus that the carbon intensity of electricity displaced by solar PV additions is falling. (For this reason, abatement figures for the UK, which has highly ambitious decarbonisation goals, are relatively low for solar PV.) In the scenario modelled with goals only 75 percent achieved, grid carbon intensity is higher, and the abatement effect is therefore also higher. In this scenario over 1,400tCO₂ is abated in Italy.

Figure 2.

Figure 3 shows emissions reductions achieved through heat pump installations. We conservatively assume the installation of new high-efficiency condensing gas boilers (rather than continued use of older low-efficiency boilers) as our counterfactual: despite this, the high carbon emissions of gas as a heat source, and the much higher efficiency of heat pumps, mean that heat pumps are still better for abatement than solar PV systems. In every country, the marginal worker’s labour contributes over 1,500tCO₂ of abatement. Because they draw electricity from the grid, they become more efficient as system-wide decarbonisation continues. In the scenario in which grid decarbonisation targets are only 75 percent achieved, contrasting with solar panels, heat pumps’ abatement contributions fall.

Figure 3.

At a certain point, there will be a plateau effect as installed units come offline and must be replaced; for both solar PV and heat pumps, this occurs well after the period of study.

Why speed matters: 'Additional' vs. 'acceleratory' workers

Our base modelling scenario assumes that the marginal worker is fully additional throughout the period of study: the installations are assumed not to have otherwise been undertaken at all in the timeframe studied in the worker’s absence. In this scenario, there is a persistent workforce bottleneck from 2024 – 2032. This seems a reasonable expectation given how much demand for workers is expected to grow, and given that labour supply is already widely reported to be a bottleneck. The width of the margin within which a worker can maximally contribute, without detrimentally competing with other workers, is projected to be large.

This marginal worker could be a migrant, but could also be a domestic worker. For the purposes of migration policy-setting coherent with decarbonisation policy, it will be important to assess whether migrant workers are needed to fill gaps, and how important a contribution they might make in complementing the domestic workforce in different labour supply scenarios.

If domestic labour supply can rapidly ramp up in response to demand stimulated by the green transition, the contribution of a marginal migrant worker might be reduced. In a scenario in which demand is outstripped by supply, they could end up competing with domestic workers, rather than complementing them. In this case, their carbon emissions reduction contribution would be limited. As noted previously, this is far from the current situation and far from what is expected.

In a second scenario, we therefore model for an “acceleratory” worker. We assume that the (migrant) marginal worker’s presence brings forward—accelerates—installations that would otherwise have taken place later, after domestic workforce supply responded successfully, but still with an unavoidable delay, to demand. We do this conservatively and simplistically: we assume governments initiated large-scale apprenticeship intakes in 2024 with high throughput, such that all the workers that will be needed begin to be trained, completing their apprenticeship the typical amount of time later. (In Germany an apprenticeship typically takes 3.5 years; in Italy and the UK it generally takes around four.)

In the “acceleratory” scenario, we show the contribution if the introduction of a migrant marginal worker brings installations forward by at 3.5 years or slightly more. We “cap” emissions reductions from installations at the end of this period, recognising that by then the new domestic pipeline would have caught up. For the years of the apprenticeship periods the “additional” (base) and “acceleratory” workers are the same, diverging after.

In Figure 4, we show this scenario for electricians installing solar PV systems. Despite the cap on the effectiveness of installations, the abatement contribution is still significant even if broader decarbonisation goals are achieved. In the UK, an “acceleratory” worker still abates nearly 600 tCO₂, in Germany over 800 tCO₂, and in Italy more than 1,000 tCO₂ over the period.

Figure 4.

In Figure 5, we show the “acceleratory” scenario for heat pump installers. For all countries, the ‘acceleratory’ worker still contributes over 1,200tCO₂ of abatement.

Figure 5.

These results suggest that even where migrant workers are only anticipated to be needed in the short term while the domestic labour supply scales up to meet policy-driven demand, their contributions are sufficiently significant that the avoidance of workforce-related bottlenecks should be a priority.

The need for care: Ensuring net emissions reduction

Grid carbon intensities (the amount of carbon emitted with every kilowatt-hour of electricity generation) vary significantly across countries. This is potentially significant when seeking to maximise the net decarbonisation impact of labour reallocation. A migrant worker contributing marginal labour can evidently make a large impact for decarbonisation in a country of destination; but if that impact would have been larger in their country of origin, their relocation will have a net negative effect on carbon reduction.

To explore this we compare the decarbonisation impact of marginal labour across countries of destination and origin. We use grid carbon intensity figures from Ember (2025). These are average emissions factors, used as a proxy for marginal emissions factors in the absence of adequate data. (This may either overstate or understate decarbonisation contributions depending on the technology and country.) We project future emissions factors on the basis of declared country decarbonisation goals, also projecting partial (75 percent) achievement of stated goals.

Of the six countries, Kenya has the lowest grid carbon intensity in 2024, in large part due to high generation from geothermal, hydropower, and other renewable sources. India has the highest, followed by the Philippines, each emitting two or more times the gCO₂/kWh of the UK, Germany, and Italy.

Figure 6 shows projected average grid carbon intensity from 2024 – 2032.

Figure 6.

The wide variance in grid carbon intensity results in corresponding variance in the abatement impact of technologies lowering carbon emissions. Each solar panel installed in India will, if it displaces grid-drawn electricity, have a much greater carbon emissions reduction effect than one installed in the UK.

Countries of origin, as noted earlier, also face workforce shortages. This can potentially pose a problem. If a migrant worker is equally marginal in both the country of origin and the country of destination, the discrepancy in grid carbon intensity may mean that, even if we assume lower productivity and installation rates in the country of origin, their abatement contribution would be much larger remaining in place rather than moving to fill a bottleneck role in a high-income country facing a workforce crisis.

In Figure 7, we show the difference between the carbon emissions reduction contribution made by a marginal worker in countries of destination and countries of origin. In Kenya, where solar installations are predominantly off-grid and are estimated to displace around 430kgCO₂ per year (GOGLA, 2025), there is little risk of a net carbon deficit (although care still needs to be taken to avoid harming energy access goals). In India and the Philippines, by contrast, the departure of a migrant worker that is equally marginal in both country of origin and country of destination could leave an abatement gap of 4-5,000tCO₂ by 2032.

Figure 7.

This does not mean that labour migration of electricians and other workers from these countries should be ruled out. It does, however, mean that attention must be paid to ensuring that countries of origin are not left without necessary workers.

Figure 8 shows grid carbon intensity and energy demand per capita in 2023. Many potential migrant countries of origin, such as Morocco, Egypt, India, or the Philippines, have grid carbon intensity multiple times higher than most—but not all—potential countries of destination.

Figure 8.

Recruitment from countries of origin with high carbon intensity of energy generation should be handled with particular care. They may have surplus or underemployed workers, but where they don’t, training support by the country of destination may be needed to ensure that recruitment does not leave an abatement deficit. Equally, recruitment from countries with low energy provision per capita, in need of expanded clean electricity access, should also be undertaken carefully to avoid negative development outcomes.

A forest per worker: Comparing the marginal worker's contribution to other benchmarks

To place the contribution of a marginal “bottleneck” worker in context, we compare their abatement contribution to the number of trees that would need to be planted to capture the equivalent amount of carbon emissions avoided. We use figures from the UK’s Woodland Carbon Code (UK Forestry Commission, 2025), and assume the planting of new native woodland for a 50-year lifespan (West, 2024)—the accounting lifespan most relevant to decarbonisation timeframes. Because sequestration is vulnerable to reversal (a tree might later be burnt or rot), this comparison is only illustrative: direct abatement should be prioritised over offsetting (Axelsson et al., 2024).

Figure 9 shows the equivalent value in tree-planting of a marginal solar PV electrician’s work. In the UK, installations undertaken during the period 2024 – 2032 would have the equivalent decarbonisation contribution of planting over 3,500 trees; in Italy, it would be the equivalent of more than 6,500. In India and the Philippines, high grid carbon intensity pushes the equivalent number of trees to nearly 30,000. (Note that as the carbon intensity of displaced grid electricity falls, the annual abatement of the stock of installed solar PV systems also plateaus and falls, despite continuing additions.)

Figure 9.

The equivalent figure for heat pumps is larger still, ranging from over 8,000 in Italy to over 9,200 in Germany (Figure 10). (We don’t model the country of origin scenario for heat pumps, given that heating technicians in the country of origin are likely to be installing air conditioning units. These may make a valuable adaptation contribution, but would cause emissions rather than contribute to abatement.)

Figure 10.

The marginal worker’s abatement contribution can also be expressed in monetary terms, using the social cost of carbon (an estimate of the global social value of avoiding the long-term damages from one additional tonne of CO₂). We use the social cost of carbon values used by the US Environmental Protection Agency under the Biden Administration, which calculates a US$190 value in 2020, rising to US$230 in 2030 and US$270 in 2050 (EPA, 2023). These figures suggest significant social value returns from facilitated labour migration to fill bottleneck roles. The work of a marginal electrician installing rooftop solar PV is estimated to have a value of between approximately US$150,000 and US$280,000 in countries of destination, and in excess of US$1.2 million in India and the Philippines (Figure 11). In Kenya, where off-grid solar installations abate roughly 430 kgCO₂/year, values are lower.

Figure 11.

For heat pumps, in line with higher emissions reduction benefits, monetised values are higher. In Figure 12 we show the social value of the marginal heating technician’s work, showing how social value varies with decarbonisation and ‘acceleratory’ versus ‘additional’ worker scenarios. In a scenario in which decarbonisation goals are achieved, the marginal ‘additional’ worker’s decarbonisation contribution would be valued at more than US$330,000 in each country, rising to over US$400,000 in Germany; the ‘acceleratory’ worker’s contribution would be valued at around US$270,000.

Figure 12.

Gas imports reduced, costs avoided: Co-benefits of filling workforce gaps

Carbon abatement contributions are a global public good. However, marginal workers also contribute more local goods, benefiting households directly or contributing to national treasuries. Firstly, there is the benefit of greater energy generation or efficiency. A marginal electrician is estimated to install more than 200kWp of rooftop solar capacity each year: this generates a total of around 10 million kWh of electricity by 2032, reducing generation needs from other sources.

Figure 13.

In the context of persistently higher electricity prices due to supply chain and geopolitical disruptions (Bolton, 2025b; Eurostat, 2025), rooftop solar PV installations can, with appropriate financing support, save household consumers money (e.g. DESNZ, 2025d).

Heat pumps can also save consumers money, if installed within the right policy setting (see e.g. Sissons et al., 2025; Harrison, 2025). They will also reduce the amount of gas needed to be produced or imported, a valuable factor in the context of a continuing fragile global liquid natural gas market (see e.g. IEA, 2025). A marginal heating technician is estimated to install a total of around 240 household heat pumps during the period 2024 – 2032: cumulatively, this obviates the need for around 1 million m³ of gas in the UK and Italy, and more than 1.2 million m³ in Germany.

Figure 14.

For Germany and Italy, which are under the EU’s legally binding Effort Sharing Regulation, there is potentially a direct financial incentive to ensure that a shortage of “bottleneck” workers does not lead to breached emissions pledges. For each tonne of CO₂ emitted over an agreed ceiling set under the Effort Sharing Regulation, EU Member States must purchase an Annual Emissions Allocation (AEAs) from another Member State that has a surplus, within a restricted market (European Commission, 2023). (Funds from sales are expected to be used nationally for climate-related purposes; it is not clear what will happen if no other Member States have sufficient sellable AEAs.) Both Italy and Germany are projected to heavily overshoot their limits under the Climate Change Regulation, with an estimated combined 246 Mt CO₂e deficit for 2021 – 2030 (Transport and Environment, 2024). (Germany alone is expected to have a deficit of 226 Mt CO₂e (Umwelt Bundesamt, 2025).)

There is no transparent market price for AEAs. Proxies derived from prices in related EU carbon markets are often used, and suggest that an AEA could be priced at around EUR 122 (US$143) by 2030, with an anticipated average value of EUR 99 (US$116) for the period 2027 – 2030 (BloombergNEF, 2025). Germany and Italy would need to buy AEAs in or ahead of the 2032/33 compliance window for excess carbon emitted during the mechanism’s 2026 – 2030 accounting period.

During the 2026 – 2030 accounting period, the emissions abatements enabled by a marginal worker would, at an EUR 99 AEA value, total well over EUR 60,000 (US$69,000) in both Germany and Italy (Figure 15). This is the cost both countries would need to pay for additional emissions that could, if labour shortages are a limiting factor, be abated with the addition of workers. Where labour bottlenecks are preventing carbon-reducing installations, both countries therefore have a clear direct fiscal incentive to fill roles, including through the use of targeted labour migration.

Figure 15.

The amounts calculated to be saved are likely to be many times greater than the costs of international recruitment, even where training or similar support is provided to the country of origin. If this is the case, coherency across policy areas and cost centres becomes still more important to realise both abatement opportunities and savings.

Carbon emissions resulting from migration do not pose a risk

Migration is not without carbon costs in its own right. International air travel is estimated to emit around 150gCO₂e per kilometre (Ritchie, 2023). We compare workers’ decarbonisation contributions with the carbon cost of their relocation, using estimates of travel-related CO₂ emissions produced by the International Civil Aviation Organization. We adapt these CO₂ figures to account for additional heating effects caused by other factors, prominently radiative forcing, following a methodology developed by the UK’s Department for Transport (DfT, 2024). We assume that a migrant worker visits their country of origin every two years.

We also calculate the increased carbon emissions caused by lifestyle differences between countries of origin and countries of destination. To do this, we take adjusted net differences between per capita emissions, also incorporating emissions embodied in trade.

The carbon costs of migration are not insignificant. A migrant from the Philippines to the UK, for example, would cause emissions of over 1.6tCO₂e in their initial outbound flight, while lifestyle emissions would also be estimated to go up by around 5tCO₂ per year.

In the context of the marginal worker’s decarbonisation contribution, however, the carbon costs of migration are far outweighed. Even in the UK, where the targeted grid decarbonisation trajectory limits the abatement contribution of the marginal solar installation, the carbon costs of migration still equal less than 10 percent of a marginal migrant installer’s abatement. In Italy, against 1,255tCO₂ abated by a solar PV installer, the net abatement across the period of an installer moving from India would be 1,216 tonnes (Figure 16).

Figure 16.

Unsurprisingly, the abatement contribution of a heat pump installer net of migration carbon costs is still greater. In the context of 1,500 – 1,700 tCO₂ abated, 30 to 60 tonnes due to migration are almost a rounding error (Figure 17).

Figure 17.

Implications for labour migration policy

Given the decarbonisation contribution that can be made by qualified workers in contexts of labour shortage, better mechanisms for increasing supply and effectively allocating it to areas of high demand are crucial. Labour migration will have a crucial role to play in supplying the workforce needed at the margins. At the same time, as noted, the possibility of contributing to a global public bad means that multiple migration models need to be considered. We provide four non-mutually exclusive options.

Option 1: Recruit as normal

Some countries of origin may have surplus workers who, if remaining in situ, would be underemployed or unemployed. In these cases, international labour migration can allow them to be reallocated to locations in which they can contribute more productively. This may be because a country of origin has an overly-scaled training pipeline; because it suffers an economic downturn, leading projects to be cancelled and workers to be freed up; because it hosts a skilled refugee population unable to contribute due to lack of work rights; or because its fossil sectors scale down before green sectors scale up, leaving workers with transferable skills underemployed. Where this is the case, international recruitment can take place as normal without the need for fears of a decarbonisation gap being left in the country of origin. This is not expected, however, to be the normal scenario.

These workers could be recruited into temporary visas, in which they remain in the country of destination for a fixed period of time (for example, 3 years) before returning to the country of origin (Figure 18). In some cases, where jobs are not in demand year-round, a circular model (in which workers repeatedly migrate to the country of destination before returning) may be suitable, but in decarbonisation-related sectors this is likely to be rare. Fixed-term migration can usefully respond to limited-duration surges in tasks due to policy interventions, or can fill non-structural workforce gaps while domestic training ramps up. It may have the advantage of providing workers with increased skills or employment to high-demand decarbonisation in countries of destination before they return to assist growing clean growth in countries of origin. In some contexts, it may also be more politically palatable than longer-term migration. On the other hand, shorter-term migration may also be less attractive to employers in the country of destination. In some contexts, migration with a path to permanency will be both politically acceptable and structurally necessary.

Figure 18. Fixed-Term Migration model

Image

Source: Dempster and Huckstep (2024)

Option 2: Recruit into apprenticeships

Many countries, such as the UK (Huckstep and Dempster, 2025), are struggling to increase the provision of domestic training: they may even have excess demand for training unmatched by supply. In Germany, however, a confluence of factors means that there are tens of thousands of apprenticeships unfilled by domestic workforce entrants (Bundesagentur für Arbeit, 2024a).

Where country of destination training capacity persistently outstrips local demand, unfilled apprenticeships can be filled by international youth assisted in accessing migration pathways. This is a model that is already being tested in Germany by organisations including the employer association Bauverbande (Schneider, 2023), the Nepal Secretariat of Skills and Training (NSST, 2025), an educational body, and the non-government organisation Malengo (Malengo, 2025).

This approach fills labour market gaps in the country of destination, and does not risk leaving the country of origin with skill gaps: participants moving internationally are not yet skilled, and receive training in the country of destination. It is likely that Germany, with an apprenticeship system with excess capacity, is a rare case, although similar dynamics on a much smaller scale are also reported in Austria (BMWET, 2024) and Switzerland (News Service Bund, 2024).

Option 3: Global Skill Partnerships

Where both the country of destination and a potential country of origin have predictable labour shortages in the same occupations, programmes that pair training and migration—including models such as the Global Skill Partnership—can be valuable. The Global Skill Partnership model (Figure 19) sees the country of destination provide technology and finance to train a cohort of workers in the country of origin, before part of the cohort is supported in migrating. The model allows the country of destination access to a reliable and predictable pipeline of workers trained to the standards needed, reducing risks related to credential recognition problems.

The model also obviates the risk of skill gaps left in the country of origin causing a decarbonisation deficit. Instead, because of the partial-cohort migration outcome, the total stock of skilled workers is increased in both countries. This “dual track” system is the defining characteristic of the Global Skill Partnership model.

Figure 19. Global Skills Partnership model

Image

Source: Dempster and Huckstep (2024)

Option 4: Recruitment with parallel investments

In some cases, countries of origin may not face labour-related bottlenecks –they may have sufficient workers already trained, or have adequate training pipelines in place to replace emigrating workers— but may require other forms of support. In a final migration model, the country of destination expands the migration of pre-trained workers while also providing investments to the country of origin to support future training or the development of broader systems. These investments could, for example, be in labour market intelligence, the quality of training facilities, or necessary equipment for post-training activities.

This model (Figure 20) could be applicable for contexts in which a Global Skill Partnership is not suitable, possibly because establishing one would be too expensive or time-consuming, or because skill needs are unaligned. It has the greatest potential when the country of destination has pressing shorter-term skill shortages, and where the country of origin has non-workforce needs with which the country of destination can support. In this way a partnership could benefit both countries in their clean energy transition.

Figure 20. Parallel Investments model

Image

Source: Dempster and Huckstep (2024)

Policy conclusions

The absence of a ‘green-skilled’ worker can have a large carbon emissions implication at the margin. Ensuring that workforce supply is adequate to the task should therefore be a major policy priority. The domestic workforce should provide the vast majority of the skilled workers needed. But for most countries, training capacity will not scale to the extent needed at the speed needed (see e.g. Hambrecht et al., 2025). Where this is the case, the use of labour migration policy can ensure that workforce gaps do not lead to implementation gaps.

While labour migration is an increasingly difficult area of policy politically, these findings support the use of targeted labour migration to supplement domestic workforce supply. Across high-income countries seeking to decarbonise rapidly, the thousands of migrant workers needed to fill labour gaps are likely, on the basis of this paper’s modelling, to contribute to the abatement of millions of tonnes of CO₂.

Several policy conclusions can be derived.

Where domestic labour pipelines alone cannot meet needs, international recruitment should be facilitated. From a carbon emissions reduction standpoint, labour migration is not a tool that can be set aside. This requires coherency between reliable green industrial policy, workforce development policy, and immigration policy, informed by strong labour market intelligence.

International recruitment must be conducted with care to ensure that it does not leave decarbonisation gaps in countries of origin. Because of differences in grid decarbonisation levels, a worker can often make a greater contribution at the margin in a country of origin than a country of destination. Because the climate is a global public good, this recruitment would therefore be harmful to the country of destination.

Underemployed workers should be targeted for recruitment. Many workers will not be equally marginal in countries of origin and destination: international recruitment will often not leave a gap. Partnerships can be agreed with countries with surplus pre-trained populations. Equally, some populations, such as refugees unable to work in host countries, may have necessary skills but be unable to use them (see Dempster et al., 2025). Recruiting under- or unemployed workers will maximise the net decarbonisation gain from workforce reallocation.

Training and migration partnerships with countries of origin can mitigate risks. Partnerships that increase the total stock of workers before helping some to move to where they’re needed can benefit both destination and origin countries. For the country of destination, such a partnership can serve as a form of insurance and as a contribution to broader decarbonisation. If they do need workers, they have a known supply whose training they have ensured is to the standard required. If they do not, the workers can go where they are needed: in the green transition they are unlikely to go unemployed. This is a model being tested by Australia in partnership with India.

Given the value of a marginal skilled worker, training and migration partnerships are a good use of climate finance. Without skilled workers, decarbonisation cannot happen. An individual’s abatement contributions can run into thousands of tonnes of CO₂: preventing workforce bottlenecks is a good use of climate finance.

Harmonise curricula and qualification recognition procedures to facilitate workforce mobility. International reallocation of workers to countries with skill bottlenecks is crucial for decarbonisation. Difficulties in recognising training standards, however, frequently delay or derail mobility. Harmonisation of curricula to international standards and improvements to credential recognition should be priorities for standard-setting and recognition bodies at the international and national levels. Mutual recognition or service level agreements and overseas assessment partnerships could help at the bilateral level.