Introduction

With climate legislation becoming more stringent, workers are increasingly susceptible to climate transition risks. For instance, coal miners may find themselves compelled to transition to alternative employment opportunities if downstream users follow the United Arab Emirates (UAE) agreement at the 28th meeting of the Conference of the Parties (COP28) and shift away from fossil fuels1. The occupations that are highly exposed to climate transition risks are usually characterised by greater occupational carbon footprints2, which are comprised of the direct and indirect carbon emissions stemming from the primary inputs of occupational labour3. Acknowledging occupational carbon footprints represents the first crucial step towards a green transformation in the labour market4 and identifying vulnerable workers. Nevertheless, global production fragmentation5 complicates the calculation of carbon footprints6,7, and there is still a notable gap in conducting a comprehensive assessment of occupational carbon footprints from a global value chain perspective8.

Occupational carbon footprints could be evaluated based on life cycle assessment, which is divided into the bottom-up processed life cycle analysis (LCA) and the top-down approach9. The former is based on detailed micro activity data of workers and could provide a more accurate reflection of the carbon footprints of certain activities of workers10. Yet, the difficulty of collecting the micro activity data of workers hinders its application at the global level11. The top-down approach, which is based on the input-output model12 and has the advantage of a complete system boundary, could provide an alternative method to trace occupational carbon footprints13,14. The input-output model has been widely used to trace the carbon footprints of different types of population groups classified by income15 and age16,17, yet the evaluation of occupation-specific carbon footprints is still lacking, although the classification of the population by occupations is widely accepted by the public18.

Climate change poses both physical and transition risks to workers. The physical risk refers to the impacts on occupational safety and health due to the physical effects of climate change, such as extreme temperature19,20 and sea level rising21. The transition risk means the potential costs to society of the shift towards a green economy on workers4,22,23,24. While the physical risk is more perceptible25, the transition risk tends to exert a greater impact on workers in the near future26. Previous studies have primarily determined the exposure of each occupation to climate transition risks based on the “brownness” or “greenness” of its tasks4,27,28,29,30, with alternative approaches remaining relatively rare. Consequently, it’s necessary to enrich the related literature by evaluating occupational exposure to climate transition risk from the value chain perspective.

In this work, we extend the literature by providing a time series and global-level analysis of carbon footprints of 13 occupations (Supplementary Information 1) across 45 sectors (Supplementary Information 2) in 67 economies (Supplementary Information 3) from 2000 to 2018. Based on the evaluation of occupational carbon footprints, the present study proposes a transition risk assessment framework to identify vulnerable occupations. Specifically, we evaluate the occupational exposure to climate transition risks from two perspectives: impact and likelihood. The risk impact of an occupation is measured by the ratio of its potential cost of mitigating its carbon footprint to its value added. The risk likelihood of an occupation is measured by the ratio of the volume of regulated carbon footprint to the gross volume of carbon footprint. Please refer to Supplementary Information 13 for the evaluation framework of the present study, and calculation equations are presented in the method section. Here, we first present the temporal trend and spatial pattern of occupational carbon footprints.

Results

Temporal trends of occupational carbon footprints

The gross volume of occupational carbon footprints at the global level increased from 6893.5 Mt in 2000 to 10018.2 Mt in 2018, accounting for 34.6% of global emissions (Supplementary Information 9). This indicates the significance of workforce transition to climate change mitigation. From the occupational perspective, craft workers and machine operators, a greater share of whom are mainly engaged in carbon-intensive energy and manufacturing sectors (Supplementary Information 4), rank first among the 13 occupations by the volume of occupational carbon footprints in 2018, constituting 28.7% of the total occupational carbon footprints (Fig. 1). Improving the production efficiency of craft workers and machine operators is essential for achieving the climate mitigation targets31. Over the studied period, craft workers and machine operators also correspond to the greatest growth in the volume of occupational carbon footprint, with a growth of 963.7 Mt.

Fig. 1: Changes in occupational carbon footprints over the period 2000–2018.
figure 1

Each colour corresponds to a specific occupation, with workers classified into 13 occupational categories (Managers, Engineering professionals, Craft workers and machine operators, Agricultural workers, Drivers, Other professionals, Clerical support workers, Sales workers, Legislators, Health professionals, Teaching professionals, Personal service workers, Other, including armed forces).

Full size image

To provide a more detailed explanation on the changing trends in occupational carbon footprints and corresponding driving factors, we further divide the study period into three subperiods (Fig. 2). Over the first subperiod, the global occupational carbon footprints increased by 12.4%. The major driving factor to this increase was the growth in the scale of primary labour inputs (scale effect), which would cause the occupational carbon footprints to increase by 21.1% in the absence of other factors. The global carbon intensity (the amount of carbon dioxide emissions emitted per unit of GDP) decreased over the period 2000–200632, and the decrease in carbon intensity offset the global occupational carbon footprints by 12.8% (intensity effect). The change in occupation structure (structure effect, 0.04%) and industrial linkage effect (4.0%) played relatively modest roles. Although the financial crisis happened in the second subperiod, the volume of global carbon emissions increased sharply33. Meanwhile, the growth rate of the volume of global occupational carbon footprints reached as high as 19.5%, which is mainly driven by the scale effect (27.6%), industrial linkage effect (3.8%) and structure effect (0.49%). The decrease in carbon intensity contributes to a decrease in global occupational carbon footprints by 12.5%. Over the third subperiod, the growth rate of global occupational carbon footprints decreased to 8.3%. Although the scale effect was still positive, the intensity effect (−10.6%) and industrial linkage effect (−3.9%) contribute to a decline in global occupational carbon footprints. The change in occupational structure still played a relatively modest role. Policymakers should implement more proactive measures to promote the green transition of workforce structure.

Fig. 2: Driving factors of changes in occupational carbon footprints.
figure 2

Changing trends in the occupational carbon footprints over three sub-periods (2000–2006, 2006–2012, and 2012–2018) and four driving factors (Scale effect: variations in the carbon footprints of occupations due to the change in the scale of labour input; Structure effect: changes in occupational carbon footprints resulting from the change in occupational structure; Industrial linkage effect: the influence of industrial linkages on variations in occupational carbon footprints; Intensity effect: the role of carbon intensity in shaping trends in the occupational carbon footprints).

Full size image

Spatial pattern of occupational carbon footprints

Figure 3 presents the volume of carbon footprints of occupations in each economy in 2018. Workers (Fig. 3a) in China correspond to the greatest volume of carbon footprints (2988.7 Mt), followed by the United States (1773.8 Mt) and the European Union (1039.7 Mt). The ranking of occupational carbon footprints (please refer to Supplementary Information 5 for the volume of occupational carbon footprints at regional and sectoral level) is mainly determined by the positions of different economies in value chains, due to the uneven distribution of occupations and emissions along the value chains. Developed countries typically occupy upstream positions in value chains, where professionals constitute a larger proportion of the occupational structure. Figure 3b shows that the carbon footprints of managers in the United States reached 337.1 Mt, a figure notably higher than that of other economies, such as the European Union (133.0 Mt) and China (123.7 Mt). Similarly (Fig. 3c), the United States (218.4 Mt) is also ranked the first by the volume of carbon footprints of engineers, followed by the European Union (176.8 Mt) and China (163.7 Mt). This is attributed to the fact that a considerable number of engineers engage in manufacturing industries, primarily concentrated in developed countries and major emerging economies (as categorized by the International Monetary Fund34), such as China35. China, often referred to as the “world’s factory,” is home to a vast number of producers and production supporters. Therefore, these two occupations in China correspond to the top two positions in terms of the volume of carbon footprints. The volume of carbon footprints of producers and production supporters in China reached 1633.2 Mt and 832.2 Mt in 2018, which is notably greater than that of other economies (Fig. 3d–i). Workers in other occupations generally exhibit lower carbon footprints, with the spatial distribution of carbon footprints concentrated in the world’s two largest economies, the United States and China (Fig. 3j–n).

Fig. 3: Spatial distribution of the occupational carbon footprints in 2018.
figure 3

a Presents the spatial distribution of the gross occupational carbon footprints. b–n Presents the spatial distribution of 13 occupations’ carbon footprints, respectively.

Full size image

Figure 4a shows that Australia’s workers correspond to the greatest volume of occupational carbon footprint per worker, being aligned with Australia’s high per capita carbon emissions36. Workers in the other countries with a higher level of fossil fuels dependence, such as Saudi Arabia, also exhibit a greater volume of per worker carbon footprint. A territorial comparison reveals that developed countries have greater carbon footprints per worker than most emerging countries (Fig. 4b–n). However, there is a growing trend in per worker carbon footprint in emerging countries (such as China), while that of the developed countries (such as the United States) is gradually decreasing (Supplementary Information 10). In developed countries, policymakers should impose stricter regulations to control the per worker carbon footprints of occupations. Simultaneously, efforts should be made to curb the growth of per worker carbon footprints in emerging countries. The carbon intensity and the value-added intensity (the amount of occupational value added created per worker) are the primary factors that define the volume of the carbon footprint per worker. Therefore, the occupations that are primarily involved in carbon-intensive (such as drivers, Fig. 4f) and value-added-intensive (such as managers, Fig. 4b) industries are associated with a considerable level of per worker carbon footprints. Yet, it should be noted that there exist considerable occupational differences in the carbon footprint intensity, indicating that some workers may be more vulnerable to climate transition risks.

Fig. 4: Territorial comparison of the occupational carbon footprints per worker in 2018.
figure 4

a Presents the occupational carbon footprints per worker in different economies. b–n Presents the spatial distribution of 13 occupations’ carbon footprints per worker, respectively.

Full size image

Occupational exposure to climate transition risk

Among the 13 occupations, drivers—who make up 3.0% of the workforce—are exposed to the greatest climate transition risk, which is measured from both risk likelihood and impact perspectives. The transformation of transportation is critical in addressing climate change37. Drivers are encountering escalating climate-related pressures, which increases their likelihood of experiencing risk. Climate mitigation mechanisms would raise the cost of fossil fuels use38, and drivers are particularly sensitive to these changes in cost. Compared with the other 12 groups of occupations, drivers have the largest share of the protentional carbon footprint mitigation costs in relation to the economic benefit they gain (Supplementary Information 11); therefore, they correspond to the greatest nominalized impact in Fig. 5. The other occupations that are exposed to great climate transition risk are craft workers and machine operators (such as extraction and building workers, labourers in mining, construction, manufacturing, and so on), which are typically characterised by their labour-intensive nature. The craft workers and machine operators rank second in terms of the number of workers among the 13 occupations studied, constituting 17.0% of the total workforce. In addition, the corresponding workers are usually engaged in carbon-intensive production stages, and they tend to encounter notably greater challenges in switching careers due to the limited scope of technical skills or the need for a considerable capital investment. Upskilling and retraining are essential for enhancing the risk adaptability of craft workers and machine operators. There also exist occupations accounting for a greater share of the total work force, such as agriculture workers (38.1%). Agricultural workers face relatively lower carbon pricing risks, because most of current carbon pricing policies have not yet extended to the agricultural sector.

Fig. 5: Occupational exposure to climate transition risk.
figure 5

The size of each circle represents the share of workers in the total workforce. The red colour represents high climate transition risk, and the blue colour represents low climate transition risk. Normalized impact (\(\hat{\alpha }\)) represents the share of potential carbon footprint mitigation cost of an occupation to its wage revenue. Normalized likelihood (\(\hat{\beta }\)) indicates the share of its carbon footprints that are sourced from the regulated sectors.

Full size image

Figure 6a presents the exposure to climate transition risk of 13 occupations in different economies. The results show that most occupations in emerging countries are facing much greater carbon price rising risk than that in developed economies. Due to the greater carbon intensity in the emerging countries, occupations correspond to greater volume of carbon footprints relative to wage income; therefore, workers in the emerging countries face greater carbon footprint mitigation burdens. In addition, the emerging economies generally exhibit relatively modest climate regulations compared to those in developed economies. Therefore, workers in the emerging countries are more sensitive to the stringent regulations in the near future. Emerging countries face greater pressures to cope with the workforce transformation, particularly for drivers, craft workers and machine operators. In developed economies, there also exists territorial and occupational heterogeneity in the exposure to climate transition risks. For instance, the workers in the United States, which face a greater necessity for an enhancement of its climate regulations, are exposed to greater climate transition risk than those in the European Union. Among the studied occupations, drivers in developed countries, such as the United States and Australia, tend to be more vulnerable to climate transition risks than other occupations, and policymakers in developed countries should adopt more targeted measures to protect them. More details about climate transition risk impact can be found in Supplementary Information 12.

Fig. 6: Exposure to climate transition risk of occupations in different economies.
figure 6

There are 26 territories in figure. The first 13 territories are developed countries, and the last 13 territories are emerging countries. The colour of each circle in (a) represents the level of occupational exposure to climate transition risks, which is measured by the multiplication of \(\hat{\alpha }\) and \(\hat{\beta }\). The dark cyan bars in (b) and (c) represent the number of corresponding measures to protect workers in climate policies.

Full size image

Figure 6b, c further illustrate the occupational and regional distribution of measures outlined in climate policies to safeguard workers against climate transition risks. Figure 6b indicates that the highest number of protection measures are associated with craft workers and machine operators, given their large workforce and considerable exposure to carbon pricing risks, as highlighted in this study. Agricultural workers are the second most supported occupation by climate policies, due to their vulnerability and large workforce. With the rapid development of renewable energy and electric vehicles, engineering professionals and drivers are also receiving considerable attention in climate policies. Figure 6c shows that the EU climate policies correspond to the largest number of occupational protection measures, followed by the UK and Canada. In comparison, climate policies in emerging countries offer less protection for occupations facing climate risks. Yet, the results of the present study show that most occupations in emerging countries are exposed to notably higher carbon price risk than those in developed economies, highlighting a regional mismatch between the climate risks faced by occupations and the protection measures in place. With the gradual increase in carbon prices in the future39,40,41, we suggest that the climate policies in emerging economies should prioritize strengthening occupational protection measures, particularly for those occupations vulnerable to carbon pricing.

Discussion

Both climate action (SDG 13) and decent work (SDG 8) are integral components of the Sustainable Development Goals (SDGs). The present study reveals the close interaction between these two SDG goals by tracing carbon footprints of different occupations across the world. The results of the present study could help workers understand the environmental impacts linked to their economic activities, encourage their engagement in climate actions and foster a climate-aligned labour market. As workers serve as the ultimate implementers of most of the production-based climate movements, the engagement of workers is essential for building a climate-resilient society. The information on occupational carbon footprints can also serve as a guideline for job hunters. Some workers may choose a job with lower carbon footprints to control their exposure to climate transition risk. With climate-friendly job hunters accounting for an increasing share in labour market, businesses also need to adopt more active climate actions to make themselves attractive to prospective workers.

Governments have adopted several measures to control climate risks faced by workers, such as high-temperature subsidies in China42 and carbon tax rebate in Canada and Switzerland43. Yet, the existing measures focus on the direct climate risk exposure of workers in developed countries. In the present study, we hold the viewpoint that the occupational climate risk originates not only from the company’s direct production activities but also from its value chains, especially considering that the volume of value chain carbon footprints is notably greater than that of direct emissions13. For instance, the phasing out of fossil fuels in manufacturing industries would shrink the employment in coal mining sector. Against the background that businesses, for example, Apple company, are in collaboration with their value chain partners to control their carbon footprints44. The results of the present study could help in the determination of more targeted measures to protect vulnerable occupations from climate-related losses originating from the value chains.

A shift towards greener jobs is crucial for workers in occupations at higher climate transition risk, yet the ease of job shifting varies notably by occupations. For instance, switching from traditional fuel vehicles to electric vehicles is relatively straightforward for drivers, whereas coal miners often face more substantial challenges to find a new job45. For occupations in the first category with readily transferrable skills, career counselling and guidance services are essential to help workers plan and navigate job transitions. For occupations facing greater barriers to job shifting, policymakers should prioritise providing upskilling, retraining opportunities, or subsidies to support workers in these fields46. Finally, occupations that are resilient to climate transition risks may still face the threat of job displacement from other vulnerable occupations. To address this issue, policymakers should prioritise supporting the development of labour-intensive green industries, as generating green jobs is fundamental to resolving the problem at its core.

There are several directions that are worth future studies to further pursue. First, the input-output model adopts the sectoral homogeneity assumption; therefore, we evaluate occupational carbon footprints at the sectoral level. As an increasing number of businesses have been reporting their value chain carbon footprints47, the future studies could further compare the carbon footprints of workers at the firm level. Second, the economy groups in the basic input-output table, which is obtained from the Organization for Economic Co-operation and Development (OECD)48, are mainly made up of developed economics and big emerging economics. The future studies could further extend the research scope to the small emerging economics49, where workers are more sensitive to climate change. Third, in this study, we present the occupational comparison of climate transition risk, especially focussing on the carbon pricing risk. The future studies could provide a more comprehensive analysis to identify the major sources of climate risks for different occupations. Fourth, the occupational classification of the present study is based on the skill types, the future study could use the type of authority as the primary criteria for job classification50. Additionally, while this study analyses employment from a macro perspective, occupational choice is also shaped by various micro-level factors, such as individual preferences for work location. Future studies are expected to incorporate these micro-level determinants to provide a more nuanced understanding of employment patterns. Finally, future research may prolong the duration of this study, particularly considering the recent changes to the occupational structure.

Methods

Calculation of occupational carbon footprints

We suppose the world is composed of \({{{\rm{H}}}}\) countries and that each country has \({{{\rm{N}}}}\) sectors. \({{{\bf{y}}}}\,=\left[\begin{array}{c}{\sum }_{j}{y}_{1j}\\ {\sum }_{j}{y}_{2j}\\ \vdots \\ {\sum }_{j}{y}_{{hj}}\end{array}\right]\) is a \({{{\rm{HN}}}}\times 1\) vector of final demand, where \({{{{\boldsymbol{y}}}}}_{{ij}}\) represents the final product of country \(i\) required to meet the final demand of country \(j\). \({{{\bf{A}}}}\) is a \({{{\rm{HN}}}}\times {{{\rm{HN}}}}\) matrix of intermediate input coefficients, representing the number of intermediate inputs used in the production of one unit of output. According to the Leontief input-output model12, we obtain

$${{{\bf{x}}}}={({{{\bf{I}}}}-{{{\bf{A}}}})}^{-1}{{{\bf{y}}}}={{{\bf{By}}}}$$
(1)

where \({{{\bf{I}}}}\) is an \({{{\rm{HN}}}}\times {{{\rm{HN}}}}\) identity matrix, and \({\left({{{\bf{I}}}}-{{{\bf{A}}}}\right)}^{-1}\) is the Leontief inverse matrix51,52. \({{{\bf{x}}}}\) is the output vector. Let \({{{\bf{V}}}}\) be a \({{{\rm{HN}}}}\times {{{\rm{HN}}}}\) diagonal matrix, and the diagonal elements represent the value-added coefficient of each sector. The value-added vector \({{{\bf{v}}}}\) can be expressed as

$${{{\bf{v}}}}={{{\bf{VBy}}}}={{{\bf{V}}}}({{{\bf{I}}}}+{{{\bf{A}}}}+{{{{\bf{A}}}}}^{2}+\cdots ){{{\bf{y}}}}$$
(2)

where \({{{\bf{V}}}}{{{{\bf{A}}}}}^{n}{{{\bf{y}}}}\) represents the value added induced by final products through a supply chain with \(n+1\) production stages. Value added will be counted in the output for each production stage, and a sector’s output is the accumulation of value added generated by many sectors53. We define the cumulative output of a sector as the direct and indirect output that is enabled by this sector’s value added. Then the cumulative output matrix could also be expressed as

$$\widetilde{{{{\bf{x}}}}}={{{\bf{V}}}}({{{\bf{I}}}}+2{{{\bf{A}}}}+3{{{{\bf{A}}}}}^{2}+\cdots ){{{\bf{y}}}}={{{\bf{VBBy}}}}$$
(3)

It should be noted that the elements of \({{{\bf{x}}}}\) and \(\widetilde{{{{\bf{x}}}}}\) have different economic meanings. The former represents a sector’s output that is cumulative regarding different sources of value added, while the latter represents the cumulative output enabled by a sector’s value added. We define a diagonalized matrix of carbon intensity \({{{\bf{F}}}}\), the diagonalized element of which represents the emissions emitted per unit of a sector’s output. The gross carbon emissions generated in the production of cumulative output is

$${{{\bf{g}}}}={{{\bf{V}}}}({{{\bf{F}}}}+{{{\bf{FA}}}}+{{{\bf{AF}}}}+{{{\bf{F}}}}{{{{\bf{A}}}}}^{2}+{{{\bf{AFA}}}}+{{{{\bf{A}}}}}^{2}{{{\bf{F}}}}+\cdots ){{{\bf{y}}}}={{{\bf{VBFBy}}}}$$
(4)

The value added is made up of three parts in the input-output table (Supplementary Information 16): labour income, capital income, and production tax. Labour income is further divided into different parts that are obtained by workers belonging to different occupations. Based on Timmer et al. (2019)18, we define \({{{\bf{O}}}}=\left[\begin{array}{cccc}{{{{\rm{O}}}}}^{1} & 0 & \cdots & 0\\ 0 & {{{{\rm{O}}}}}^{2} & \cdots & 0\\ \vdots & \vdots & \ddots & \vdots \\ 0 & 0 & \cdots & {{{{\rm{O}}}}}^{h}\end{array}\right]\) as a \({{{\rm{HK}}}}\times {{{\rm{HN}}}}\) matrix, where \({{{\rm{K}}}}\) is the number of different occupations. The elements of the matrix \({{{{\bf{O}}}}}^{h}\) represent the shares of labour income that are obtained by workers belonging to different occupations in country \(h\). The matrix \({{{\bf{O}}}}\) is obtained from Kruse et al. (2023)54. Let \({{{\bf{L}}}}\) is a \({{{\rm{HN}}}}\times {{{\rm{HN}}}}\) diagonal matrix, and the diagonal elements represent the ratios of labour income to value added at the country and sector level. We define the carbon footprint of an occupation as the amount of carbon emissions enabled by the value added generated by this occupation to produce final products. Then, occupational carbon footprint could be calculated by

$${{{\bf{e}}}}={{\bf{OL}}}{{{\bf{VBFBy}}}}$$
(5)

The elements of \({{{\bf{e}}}}\) represent each occupational carbon footprint at the country level. Then occupational carbon footprints per worker are obtained by dividing \({{{\bf{e}}}}\) by the number of workers. In order to clarify the analysis framework of this study, we provide a further explanation using a simplified economy consisting of only two industries and two occupation categories (Supplementary Information 14).

According to Supplementary Information 15, we obtain \({{{\bf{e}}}}={{{\bf{OLV}}}}{{{\bf{aGf}}}}\), where \({{{\bf{Va}}}}\) is the diagonal matrix of value added, \({{{\bf{f}}}}={{{\bf{F}}}}{{{\boldsymbol{\tau }}}}\) and \({{{\boldsymbol{\tau }}}}\) is a \({{{\rm{HN}}}}\times 1\) column vector with each element equating to 1, which is employed to sum the rows of the matrix on the left. \({{{\bf{G}}}}\) is the Ghosh inverse matrix55,56. \({{{\bf{O}}}}\) represents the occupational structure, \({{{\bf{LV}}}}{{{\bf{a}}}}={{{\bf{V}}}}{{{{\bf{a}}}}}_{L}\) represents the labour income at the country and sector level, \({{{\bf{G}}}}\) and \({{{\bf{f}}}}\) represents the industrial linkage and carbon intensity, respectively. The change in occupational carbon footprints over a period is \(\Delta {{{\bf{e}}}}={{{{\bf{O}}}}}_{t}{{{{\bf{Va}}}}}_{L,t}{{{{\bf{G}}}}}_{t}{{{{\bf{f}}}}}_{t}-{{{{\bf{O}}}}}_{0}{{{{\bf{Va}}}}}_{L,0}{{{{\bf{G}}}}}_{0}{{{{\bf{f}}}}}_{0}\). We first deflate the input-output tables to the constant price of the year 200057,58 and further adopt the structural decomposition analysis to analyse the driving factors of the change in occupational carbon footprints.

$$\Delta {{{\bf{e}}}}= \frac{1}{2}\left(\Delta {{{\bf{O}}}}{{{{\bf{Va}}}}}_{L,t}{{{{\bf{G}}}}}_{t}{{{{\bf{f}}}}}_{t}-\Delta {{{\bf{O}}}}{{{{\bf{Va}}}}}_{L,0}{{{{\bf{G}}}}}_{0}{{{{\bf{f}}}}}_{0}\right) \\ +\frac{1}{2}\left({{{{\bf{O}}}}}_{0}\Delta {{{{\bf{Va}}}}}_{L}{{{{\bf{G}}}}}_{t}{{{{\bf{f}}}}}_{t}-{{{{\bf{O}}}}}_{t}\Delta {{{{\bf{Va}}}}}_{L}{{{{\bf{G}}}}}_{0}{{{{\bf{f}}}}}_{0}\right) \\ +\frac{1}{2}\left({{{{\bf{O}}}}}_{0}{{{{\bf{Va}}}}}_{L,0}\Delta {{{\bf{G}}}}{{{{\bf{f}}}}}_{t}-{{{{\bf{O}}}}}_{t}{{{{\bf{Va}}}}}_{L,t}\Delta {{{\bf{G}}}}{{{{\bf{f}}}}}_{0}\right) \\ +\frac{1}{2}\left({{{{\bf{O}}}}}_{0}{{{{\bf{Va}}}}}_{L,0}{{{{\bf{G}}}}}_{0}\Delta {{{\bf{f}}}}-{{{{\bf{O}}}}}_{t}{{{{\bf{Va}}}}}_{L,t}{{{{\bf{G}}}}}_{t}\Delta {{{\bf{f}}}}\right)$$
(6)

The four parts in Eq. (6) represent the occupational structural effect, scale effect, industrial linkage effect, and intensity effect.

Evaluation of exposure to climate transition risks

The cost of controlling unit carbon emissions could be reflected by the carbon price level. Right now, only a small share of global emissions are covered by carbon pricing regulations with a relatively lower carbon price level (Supplementary Information 8)59,60. In the near future, increasing the carbon price while expanding the coverage from a global level is essential for limiting climate change61, and the mean carbon price estimate will be reaching $185 per tonne of CO262. We define the carbon price rising potential matrix \({{{\bf{P}}}}\), which is a diagonal matrix and the diagonal elements represent the gap between the current and future carbon price level ($185 per tonne of CO2 in the present study62) of the mining, manufacturing, electricity, and transportation sectors, as most of the existing carbon market covers these sectors. Then, the scale of occupational carbon footprint mitigation cost is measured by

$${{{\bf{c}}}}={{\bf{OL}}}{{{\bf{VBPFBy}}}}$$
(7)

We define the variable \({\alpha }_{{ho}}\), which represents the share of potential carbon footprint mitigation cost of occupation \(o\) in country \(h\) to its wage revenue.

$${\alpha }_{{ho}}=\frac{{{{\bf{w}}}}{{{{\bf{O}}}}}_{{{{\boldsymbol{ho}}}}}{{{\bf{LVBPFBy}}}}}{{{{\bf{w}}}}{{{{\bf{O}}}}}_{{{{\boldsymbol{ho}}}}}{{{\bf{LVBy}}}}}$$
(8)

where \({{{{\bf{O}}}}}_{{ho}}\) represents the shares of labour income that are obtained by workers belonging to occupation \(o\) in country \(h\). \({{{\bf{w}}}}\) represents the column sum vector. A greater \({\alpha }_{{ho}}\) means the transition to a low-carbon economy would pose a greater impact on workers of this occupation, which are exposed to greater climate transition risks from the impact perspective.

Occupational carbon footprints consist of emissions from various sectors, and the coverage of these sectors by climate regulations is uneven. For practical reasons, the existing carbon market primarily encompasses the energy-intensive sectors. We designate the likelihood of climate regulations targeting the mining, manufacturing, electricity, and transportation sectors as 1, and 0 for other sectors. We define a diagonal matrix \({{{\bf{D}}}}\), with the diagonal elements representing the likelihood of climate regulations targeted at corresponding sectors. Then, the likelihood \({\beta }_{{ho}}\) that an occupation will be impacted by climate transition is measured by the share of its carbon footprints that are sourced from these regulated sectors.

$${\beta }_{{ho}}=\frac{{{{\bf{w}}}}{{{{\bf{O}}}}}_{{ho}}{{{\bf{LVBDFBy}}}}}{{{{\bf{w}}}}{{{{\bf{O}}}}}_{{ho}}{{{\bf{LVBFBy}}}}}$$
(9)

In order to provide a comprehensive evaluation of the occupational exposure to climate transition risks from both likelihood and impact perspectives, we normalize \({\alpha }_{ho}\) and \({\beta }_{ho}\) to range between 0 and 1 using the following equation: \({\hat{\alpha }}_{{ho}}=\frac{{\alpha }_{{ho}}-{\alpha }_{\min }}{{\alpha }_{\max }-{\alpha }_{\min }}\) and \({\hat{\beta }}_{{ho}}=\frac{{\beta }_{{ho}}-{\beta }_{\min }}{{\beta }_{\max }-{\beta }_{\min }}\). \({\alpha }_{\min }\) and \({\alpha }_{\max }\) represent the minimum and maximum share of potential carbon footprint mitigation cost to wage revenue, respectively. Similarly, \({\beta }_{\min }\) and \({\beta }_{\max }\) represent the minimum and maximum share of carbon footprints sourced from sectors with stringent climate regulations. An occupation vulnerable to climate transition risks corresponds to great \(\hat{\alpha }\) and \(\hat{\beta }\) simultaneously.

To link the results with actual policies, we further assess the extent to which vulnerable occupations are effectively safeguarded in actual climate policies. We first constructed a correspondence table linking the 13 occupations to the various terms that may be used to describe them in climate policies (Supplementary Information 6). The terms in the vocabulary are extracted from the detailed job titles that are related to 13 occupations in ISCO-88 categories63. Based on a strict matching method, we identify a sub-group of policies from the Global Climate Change Mitigation Policy Dataset (GCCMPD)64 that are related to 13 occupations. However, some of these policies are not specifically focused on occupational protection within the context of climate mitigation. Therefore, we conduct a case-by-case review of the selected policies, ultimately identifying 351 policies that mention measures to protect workers from climate transition risks after year 2000. These measures include subsidies for farmers, training programs for coal miners, and the creation of green jobs for craft workers, among others. The number of measures outlined in climate policies to protect 13 occupations across 26 regions are presented in Supplementary Information 7.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.