Renewable energy certificates threaten the integrity of corporate science-based targets

Abstract

Current greenhouse gas accounting standards allow companies to use renewable energy certificates (RECs) to report reductions in emissions from purchased electricity (scope 2) as progress towards meeting their science-based targets. However, previous analyses suggest that corporate REC purchases are unlikely to lead to additional renewable energy production. Here we show that the widespread use of RECs by companies with science-based targets has led to an inflated estimate of the effectiveness of mitigation efforts. When removing the emission reductions claimed through RECs, companies’ combined 2015–2019 scope 2 emission trajectories are no longer aligned with the 1.5 °C goal, and only barely with the well below 2 °C goal of the Paris Agreement. If this trend continues, 42% of committed scope 2 emission reductions will not result in real-world mitigation. Our findings suggest a need to revise accounting guidelines to require companies to report only real emission reductions as progress towards meeting their science-based targets.

Main

There is increasing focus on the role of the private sector in meeting global climate goals^1,2,3. To this end, science-based targets (SBTs) intend to align voluntary company-level emission reduction targets with the global temperature goal of the Paris Agreement^4,5,6. So far SBTs have been set by more than 1,000 companies, including many multinationals^7,8. When reporting scope 2 emissions, that is, emissions associated with the generation of purchased energy (primarily electricity, but also including steam, heat and cooling), the Science Based Targets initiative (SBTi) allows companies to use renewable energy certificates (RECs) to claim the use of renewably generated electricity. Companies can then report zero emissions for each unit of electricity consumption covered by purchased RECs⁹, regardless of the actual emissions produced by the electricity grid at their location; however, RECs do not reflect the physical electricity flow supplied to the companies purchasing them¹⁰, and there is evidence that RECs are unlikely to lead to additional renewable energy generation^{11,12,13,14,15,16,17,18}. Consequently, company-level emission reductions reported through RECs are unlikely to reflect real reductions of global emissions, which has the potential to compromise the alignment of SBTs with the Paris temperature goal^19,20. In recognition of this issue, several emission accounting standards^21,22 restrict or do not endorse the use of RECs, but RECs are nevertheless permitted by the Greenhouse Gas Protocol⁹ that forms the basis of SBTi’s requirements.

Here we combine information about SBTi-certified company emission reduction targets with data on REC purchases to assess the effect of RECs on the alignment of companies’ reported scope 2 emission trajectories with the Paris temperature goal. We use the climate change disclosures of 115 companies, which represents the subset of companies with SBTs that have also disclosed data which can be used to assess the contribution of RECs to their reported (2015–2019) and potential future emission reductions. Importantly, it was necessary that all companies in our sample reported their past emissions using both market- and location-based emission accounting methods (Box 1). We also distinguish here between RECs and power purchase agreements (PPAs), which represent a long-term commitment by a company to purchase power from a particular renewable energy project. Although empirical evidence is still needed, we have adopted here the common assumption that PPAs do lead to additional renewable energy production and real emission reductions, as the long-term power price de-risks new projects and allows access to project finance^14,15,17,18 (Box 1). By contrast, we assume that RECs and similar market-based instruments are non-additional, that is, not leading to additional renewable generation capacity or real emissions reductions, and we use the term RECs to refer to all non-PPA instruments for that reason (see Methods for details on these market-based instruments and terminology). Although existing literature suggest that RECs are non-additional due to their low and uncertain prices^{11,12,13,14,15,16,17,18}, some claim that RECs may still contribute to the generation of more renewable energy in the longer term by, in aggregation, signaling to the market that there is a demand for renewable energy^9,23. Analyses so far do not find evidence to support the existence of such an indirect market effect, and we consequently do not consider this potential effect here. We acknowledge, however, the possibility that such longer-term indirect effects may become evident in future analyses.

Box 1 Corporate scope 2 emissions accounting and the role of RECs and PPAs

Under the Greenhouse Gas Protocol⁹, companies are required to report their scope 2 emissions using both market- and location-based accounting. Market-based accounting allows companies to use market-based instruments, including RECs and PPAs, which give companies the right to claim the use of renewable electricity and report the emissions from its production (usually zero), rather than the emissions of the actual electricity mix they consume⁹. At a given site and year, a company calculates market-based (MB) scope 2 (S2) emissions (E) by, first, multiplying the part of their electricity consumption covered by the market-based instrument (C_MBI) by the emission factor of the instrument (EF_MBI) and, second, multiplying any uncovered electricity consumption (C – C_MBI) by a residual grid mix emission factor (EF_res), which represents the local grid without the electricity generation that has been claimed by RECs or PPAs:

$${{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2} = {{{{C}}}}_{{{{\mathrm{MBI}}}}} \times {{{\mathrm{EF}}}}_{{{{\mathrm{MBI}}}}} + ({{{{C}}}} - {{{{C}}}}_{{{{\mathrm{MBI}}}}}) \times {{{\mathrm{EF}}}}_{{{{\mathrm{res}}}}}$$

Under location-based (LB) accounting, all companies on a grid multiply their electricity consumption (C) by the average emission factor (EF_mix) for the grid mix, regardless of whether they have purchased RECs or entered into PPAs:

$${{{\mathrm{E}}}}_{{{{\mathrm{LB}}}},{{{\mathrm{S}}}}2} = {{{{C}}}} \times {{{\mathrm{EF}}}}_{{{{\mathrm{mix}}}}}$$

For companies with multiple sites and that consume other energy products from a grid (heat, steam or cooling) in addition to electricity, emissions (whether market- or location-based) are calculated for each energy product and site, followed by aggregation.

We use a simple example to illustrate the calculation of market- and location-based emissions for a company with one site that purchases electricity from the local grid, which also provides electricity to other consumers. The company uses no market-based instruments in Year 1, enters into a multiyear PPA in Year 2, and purchases RECs from an existing renewable energy generator in Year 3.

In Year 1, the company has the same market- and location-based emissions as it does not purchase any RECs or enter into any PPAs, and because the grid average and residual emission factors are the same (as no grid customers purchase RECs from or enter into PPAs with the local windfarm).

In Year 2, a new windfarm becomes operational due to a multiyear PPA that the company has entered into with the project developer. The PPA achieves additionality as the project would not have happened in the absence of the PPA, which allowed the project developer to secure the necessary loans. Moreover, as total electricity demand does not change, the coal power plant reduces its production. The PPA gives the company the right to apply an emission factor of zero to 100 MWh of its electricity consumption, and the residual grid factor (which decreases slightly due to the reduced generation from coal) is applied to its remaining electricity consumption. The PPA allows the company to report a substantial market-based emission reduction. The company’s location-based emissions are also reduced due to the lower grid average emission factor from the reduced generation from coal. However, the reported emission reduction is less than with the market-based approach, as the emission benefit of the new windfarm is shared equally by all consumers on the grid under location-based accounting.

In Year 3, the company, whose PPA is still active, purchases RECs from the pre-existing windfarm. The RECs are non-additional as they do not lead to the generation of additional renewable energy, as the windfarm was operational in years 1 and 2 and would have continued to operate in Year 3 in the absence of the company’s RECs purchase. The RECs (covering 50 MWh) allow the company to report a reduction in market-based emissions compared with Year 2, even though total grid emissions did not change (889 tCO₂e). The reported reduction is therefore not real and the RECs effectively increase the market-based emissions of other energy consumers on the grid through an increase in the residual emission factor. By contrast, the company’s location-based emissions are the same as in Year 2 as the grid average emission factor did not change.

Effect of RECs on reported historical emission reductions

The sample of companies reported a combined 30.7% reduction in market-based scope 2 emissions between 2015 and 2019 (from 68.9 to 47.8 Mt CO₂e per year, see solid black line in Fig. 1e). This corresponds to a substantially higher reduction than the annual 4.2% of base year emissions required by SBTi’s linear 1.5 °C global mitigation pathway (purple line). However, most of this reported emission reduction is caused by the companies’ use of RECs (Fig. 1d), which increased from covering 8% of their purchased energy in 2015 to 27% in 2019. Based on the existing empirical evidence, we assume that this part of the reported reduction does not reflect actual reductions of emissions from the energy grid. Without the RECs contribution, market-based emissions would have reduced by only 9.9% between 2015 and 2019, with the resulting emission trajectory closely aligning with the required annual 2.5% base year emission reduction of SBTi’s linear well below 2 °C pathway (compare the dashed black and turquoise lines in Fig. 1e). This real emission reduction is driven by decarbonization of the residual electricity grid (representing the grid mix without the energy production that has been claimed by RECs or PPAs; see Box 1) (Fig. 1b) and to a lesser extent by a net-increase in company use of PPAs (Fig. 1c). By contrast, the increase in energy consumption had a net-positive contribution to the change in emissions (Fig. 1a). Hence, the appearance that the combined historical scope 2 emission trajectory of companies with SBTs easily aligned with the 1.5 °C goal is strongly misleading and a consequence of heavy reliance on RECs, which are not associated with real emission reductions.

Fig. 1: Combined scope 2 emissions of sample companies and the factors contributing to their reductions since 2015.

a–e, Market-based accounting. f–h, Location-based accounting. a–h, The contributions of energy consumption (a,f), residual grid mix (b), grid mix (g), PPAs (c) and RECs (d) to the reported market- (e) and location-based (h) emission trajectories, with the contribution of RECs to reported market-based emission reductions highlighted in orange.

Reported location-based emissions (black line in Fig. 1h) reduced from 77.1 to 69.2 MtCO₂e per year (10.3%) over the period, far less than the reported market-based emissions. This is as RECs and PPAs are not considered in location-based accounting (Box 1), which is also why location-based emissions were higher than market-based emissions in 2015. Instead, decarbonization of the electricity grid was the main contributor to the reported emission reduction (Fig. 1g), with the increase in energy consumption again having a net-positive contribution to the change in emissions (Fig. 1f). The location-based emissions trajectory overall barely complied with the well below 2 °C goal (turquoise line) and is similar to the market-based trajectory adjusted to exclude RECs (dotted black line in Fig. 1e).

Companies headquartered in Europe and North America (88% of the sample) reported larger combined market-based emission reductions in 2015–2019 than companies based in Asia (11% of the sample), but also relied more on RECs in their reporting. When removing the contribution from RECs, North American companies’ combined market-based emission trajectory merely aligned with the well below 2 °C goal, whereas European companies in aggregate did not align with any Paris goal (see Supplementary Fig. 7). At the industry level, only companies in ‘materials’, ‘hospitality and biotech’, ‘health care and pharma’ (together accounting for 21% of the sample companies) aligned with the 1.5 °C goal in aggregate after adjusting the emission trajectories for RECs (see Supplementary Fig. 8).

Renewable energy certificates also had a substantial influence on the reported emission trajectories of individual companies; 89% of sample companies purchased RECs in the 2015–2019 period and the sample companies’ median market-based emission reduction changed from 30.2% to 8.5% when removing the contribution from RECs. Likewise, the share of companies aligned with the 1.5 °C goal, aligned with the well below 2 °C goal, and not aligned with either goal changes from 68%, 8% and 24% to 36%, 12% and 52%, respectively, when excluding RECs (Fig. 2); 40% of the companies whose reported market-based emissions aligned with the 1.5 °C goal did not align with any Paris goal after adjusting for RECs (see the grey flows between first and second column of Fig. 2). Hence, market-based emission disclosures give the impression that three-quarters of the companies were in alignment with one of the Paris goals. However, when removing contributions from RECs, only half of companies still demonstrated such alignment and, of these, alignment with the less ambitious goal became more common. Much of the difference in the distribution of company alignment with the Paris goals between the market- and location-based accounting disappears after removing the contribution of RECs (Fig. 2). However, for one-third of the companies the temperature goal alignments is different for adjusted market- and location-based accounting. For example, for companies that increased their use of PPAs over the period adjusted market-based emissions commonly aligned with a more ambitious temperature goal than location-based emissions.

Fig. 2: Share of sample companies whose 2015–2019 scope 2 emission trajectories aligned with temperature goals for three emission accounting approaches.

The grey flows between the first and second column indicate the shares of companies for which the temperature goal alignment changed when removing RECs from their market-based emission trajectories. See the Supplementary Spreadsheet for individual company results.

Scenario of future REC usage in pursuit of SBTs

We now turn from the sample companies’ past reported emissions to their future commitments to reducing emissions through SBTs (Fig. 3a). Few companies (6%) report scope 2 SBTs independent from other emission scopes. Most companies (82%) report scope 2 SBTs in combination with SBTs for scope 1 (covering direct emissions²⁴) and the rest (12%) in combination with scope 1 and scope 3 SBTs (covering value chain emissions beyond scope 2; ref. ²⁴). Most companies (89%) state that their SBT refers to market-based accounting, whereas the SBTs of the remaining 11% refer to location-based accounting (in short, market- and location-based SBTs). This indicates that most companies aim to use RECs and PPAs in pursuing their targets.

Fig. 3: Annual future emission reductions based on the SBTs of the sample companies.

All SBTs have been annualized by dividing their targeted percentage emission reduction by the time span (target year minus base year). a, SBTs as reported by companies, covering different emissions scopes and emission accounting approaches. b, scope 2 SBTs estimated on the basis of companies’ reported SBTs (from a) and past scope 2 to total emission reductions. c, Market-based scope 2 SBTs (from b) adjusted by removing the estimated contribution from RECs based on past REC-related to total scope 2 emission reductions. The number of targets covered is indicated above each box. The non-outlier ranges (whiskers) are defined as 1.5-times the 25–75th percentile range. Outliers are not displayed. See the Supplementary Spreadsheet for individual company estimates.

Companies with market-based SBTs tend to commit to more ambitious emission reductions (compare the orange and blue boxes in Fig. 3a). The trend is especially pronounced for SBTs covering scopes 1 and 2, where the median annual reduction is 1.7 percentage points higher for market- over location-based SBTs (4.2% versus 2.5% of base year emissions). This may be because market-based accounting offers a relatively low-cost means of appearing to reduce emissions¹⁴, and therefore companies using this approach are willing to set more ambitious reduction targets; 58% of market-based SBTs align with the 1.5 °C goal (below purple line in Fig. 3a) and 28% with the well below 2 °C goal (between the turquoise and purple line). The corresponding shares for location-based SBTs are 8% and 54%, respectively. Some SBTs do not align with either temperature goal as SBTi, until 2019, approved targets aligning with a less ambitious 2 °C goal⁸ and one target-setting method (the sectoral decarbonization approach^4,25) allows companies with certain characteristics to reduce emissions at a lower rate than is required globally.

We next estimated the SBTs specifically for scope 2 (Fig. 3b) from SBTs that cover scopes 1 and 2, or scopes 1, 2 and 3, assuming future scope 2 emission reduction will have the same relative contribution to total emission reductions as in the past (see Methods). These estimates illustrate the implications of a continuation of current trends, rather than an explicit prediction of future emissions pathways. The estimated market-based scope 2 SBTs are generally more ambitious than the reported market-based SBTs for overarching emissions scopes and are closer to the reported SBTs specifically for scope 2 (Fig. 3a). Note, however, that 3% of these estimated market-based SBTs involve emission increases (above the 0% line in Fig. 3), as the companies in question increased scope 2 emissions during the past reference period; 75% of the estimated market-based scope 2 SBTs comply with the 1.5 °C goal and 12% comply with the well below 2 °C goal. Across companies, the estimated combined market-based scope 2 SBTs involve an annual reduction of 7.2% of base year emissions (weighted average by base year emissions; see Fig. 3b) and would thereby seem to easily comply with the 1.5 °C goal (purple line). By contrast, the estimated location-based scope 2 SBTs only barely comply with the well below 2 °C goal collectively (the weighted average reduction rate is 2.7% of base year emissions). However, the potential use of RECs may overstate the apparent Paris goal alignment of the market-based SBTs.

To investigate further, we estimated and removed the future contribution of RECs from the estimated market-based scope 2 SBTs (Fig. 3c) on the basis of the relative contributions of RECs to past scope 2 emissions reductions (see Methods). This exploratory scenario results in a combined real reduction rate of 3.6% of base year emissions across companies (see the weighted average in Fig. 3c), which is markedly lower than the 7.2% reduction rate for unadjusted market-based scope 2 SBTs (Fig. 3b) and merely complies with the well below 2 °C goal. Moreover, although nearly all (77 of 102) companies pursuing market-based scope 2 SBTs seem to align with the 1.5 °C goal (Fig. 3b), far fewer (38 of 102) companies will in fact align with the 1.5 °C goal if they continue their past pattern of REC usage (Fig. 3c). In that scenario, companies will most commonly not align with either temperature goal (45 of 102), whereas a minority (19 of 102) will align with the less ambitious well below 2 °C goal. The 25–75th percentile range of estimated market-based emission reduction rates adjusted for RECs (Fig. 3c) is similar to the corresponding range for estimated location-based SBTs (Fig. 3b), which challenges the notion that market-based SBTs tend to be more ambitious than location-based SBTs (Fig. 3a,b). Taken together, the implication of this future emission scenario is that an estimated 42% of the committed cumulative scope 2 emission reductions (101 MtCO₂e) from base year to target year will come from REC purchases and hence may not reflect actual reductions of global emissions (Fig. 4).

Fig. 4: Estimated cumulative scope 2 emission reduction from base to target year on the basis of the sample companies’ SBTs and the estimated role of RECs in achieving these reductions.

The inner circle represents the estimated emission reduction commitments by companies whose SBTs refer to market- and location-based accounting. The outer ring represents the estimated contributions of RECs and other factors (see Fig. 1) to these reduction commitments. The median base year and target year of the SBTs are 2017 and 2030, respectively.

At the regional level, European and North American companies generally targeted higher future emission reductions than Asian companies. However, these regional differences substantially reduce after adjusting the targets for estimated future RECs purchases (see Supplementary Fig. 9). Companies in half of the ten industries (together accounting for 45% of the sample companies) collectively align with the 1.5 °C goal after adjusting the SBTs for RECs, whereas companies in two industries align with the well below 2 °C goal and companies in three industries do not align with either (see Supplementary Fig. 10).

Implications and outlook

In this study we assessed the use of RECs by companies with SBTs and the implications for their Paris alignment claims. The widespread use of RECs raises doubt on companies’ apparent historic Paris-aligned emission reductions, as it allows companies to report emission reductions that are not real. Moreover, a continuation of recent trends would mean that nearly half of future scope 2 emission reductions reported by companies with SBTs would not be real. Overall, our results confirm earlier suppositions^19,20 and suggest that corporate use of RECs in the pursuit of SBTs is the norm rather than the exception. This is consistent with recent findings for a smaller sample of companies with net-zero targets¹⁸. SBTs must cover both scope 1 and scope 2 emissions⁶, and for many industries scope 2 emissions are the larger of the two (see Supplementary Figs. 11–13 for a contribution analysis of our company sample). Consequently, companies’ use of RECs threatens to undermine the integrity of SBTs as a whole. Together with recent findings of widespread accounting and reporting issues for scope 3 emissions^26,27, our findings should inform future work scrutinizing SBTs and companies’ progress against them^5,8,28.

Although our sample only covers 115 (14%) of the 813 companies with SBTs at the time (due to data availability and the exclusion of energy generators, utilities and companies with intensity targets), it is fairly representative in terms of regions and industries (see Supplementary Figs. 5 and 6). We encountered several difficulties in interpreting companies’ disclosure of RECs and PPAs and our main results (Figs. 1–4) are based on conservative interpretations (see Supplementary Section 1 for details). A set of parallel results (Supplementary Figs. 1–4) show that reported emission reductions could be even more inflated by RECs than is the case for our main results. Our evaluation of the Paris alignment of corporate scope 2 emission trajectories is based on the same global mitigation pathways used in SBTi’s target progress assessment⁸ (4.2% and 2.5% base year emission reductions annually for the 1.5 °C and well below 2 °C goals, respectively). However, Paris-aligned mitigation scenarios commonly involve substantially higher reductions in direct emissions for the power sector than for the rest of the economy (for example, 7.2% and 3.7% reduction of 2020 emissions annually between 2020 and 2030 for the 1.5 °C and well below 2 °C goals, respectively, in the mitigation scenario underlying SBTi’s sectoral decarbonization approach^25,29). This would translate to a requirement for companies to reduce scope 2 emissions at a higher rate than the global mitigation pathways, meaning that fewer companies would be Paris-aligned than suggested by our study.

Our findings have implications for SBTi’s current approach of allowing companies to choose between market- and location-based accounting when setting SBTs and reporting target progress⁶. In addition to the problem that market-based accounting allows reporting emission reductions that are not real, there is a risk of double counting the emission benefits of renewable energy generation if one company claims the use of specific renewable energy generation using market-based accounting, whereas other companies count that same renewable energy using the grid average emission factor in their location-based accounting. There are at least two alternatives that would make it more likely that all reported scope 2 emission reductions are real and renewable energy generation is only counted once (see Table 1). First, SBTi could require all companies to use only location-based accounting. A potential drawback with this option is that it would disincentivize companies from using PPAs or other market-based instruments that can lead to additional renewable energy generation. In the second alternative, all companies could be required to use a restrictive version of market-based accounting, involving mandatory demonstration of the additionality of market-based instruments (whether PPAs or RECs), that is, evidence that the renewable energy generation would likely not have occurred without the instrument^30,31. The Net Zero Carbon Buildings Framework of the UK Green Buildings Council includes such a requirement^22,32. However, experience from carbon-offsetting markets shows that demonstration of additionality is complicated and often contested³³ and more research and method development may therefore be needed for this option to be viable. Some stakeholders might also reject the use of market-based instruments altogether, whether additional or not, as this way of claiming to be supplied by renewable energy does not reflect how renewable energy supply relies on a broader system of grid-balancing, back-up capacity and transmission services, which are often supported by tax payers or other energy consumers¹⁰.

Table 1 The current emission accounting requirement of SBTi and two (not exhaustive) alternatives that could potentially prevent double counting and the reporting of emission reductions that are not real

Our study highlights issues with the use of RECs in the context of SBTs. However, the Greenhouse Gas Protocol scope 2 guidance argues that the use of non-additional market-based instruments is not a problem as the main goal is to allocate total grid emissions to individual consumers⁹. From that perspective, individual companies can legitimately use RECs to report emissions reductions that do not reflect a global emission reduction, as the market-based emissions of all energy users on the grid sum to total grid emissions. Based on our analysis, we encourage the Greenhouse Gas Protocol to reconsider its stance on additionality for two reasons. First, as many corporate energy users do not report emissions (nor do residential households), a company’s purchase of non-additional RECs effectively increases the market-based emissions of other actors that do not report emissions (through an increase in the residual emission factor; see Box 1). This means that total reported market-based emissions will always overstate the actual grid-total emissions reduction due to incomplete reporting. Second, as companies increasingly disclose emissions in the context of targets informed by the need to reduce global emissions (SBTs⁵ and net-zero targets^34,35), it is clearly misleading to stakeholders if companies can meet these targets without reducing global emissions. This would also be the case in a situation with complete reporting. We acknowledge that this second issue raises a related and broader limitation with the use of standard corporate emissions accounting (whether market- or location-based), in that changes in emissions outside of the scopes 1, 2 and 3 accounting boundary are not shown. Apparent reductions within the boundary may therefore not reflect total reductions in global emissions. For example, a company’s decision to use bioenergy may cause emissions outside its value chain through indirect land use change³⁶. Accordingly, SBTi should consider options for complementing standard scopes 1, 2 and 3 accounts with consequential emission accounting methods³⁶ to ensure that actions taken to achieve STBs do not unintentionally increase emissions outside the accounting boundary.

Conclusion

Our study shows that the common voluntary corporate practice of using RECs that are unlikely to drive additional renewable energy production casts serious doubt on the veracity of reported corporate emission trajectories and their apparent alignment with the most ambitious Paris Agreement temperature goal. More broadly, there is a need to critically consider the extent to which voluntary corporate actions can be relied on for achieving a Paris-aligned transition⁵.

Methods

Terminology for market-based instruments and their potential effects on renewable energy generation

There is no formally agreed terminology for the market-based instruments that companies can use to obtain renewable energy attributes (that is, rights to claim usage of renewable energy) under the market-based accounting approach. A large number of instruments with similar characteristics exist and the same instrument may go by different names in different regions⁹. For example, the Bloomberg database³⁷ defines RECs broadly as the “…amount of bundled and unbundled energy attributes purchased in the form of tradable certificates from an official registry, in thousands of megawatt hours. This field includes voluntary certificates such as Guarantees of Origin sold within the Renewable Energy Certificate System and European Energy Certificate System as well as Renewable Energy Certificates commonly used in the United States”. SBTi identifies RECs and PPAs as the two available market-based instruments for meeting renewable electricity targets within its net-zero standard³⁸, although other contractual arrangements exist, such as supplier-specific emission factors, and green power products or tariffs. It should also be noted that PPAs themselves generally include the renewable energy attributes associated with the power that is purchased (though it is also possible to structure a PPA for the power only, without the associated attributes). For simplicity in the terminology used we use the term REC for all contractual arrangements for renewable energy attributes other than PPAs and we use the term PPA for PPAs that include renewable energy attributes, unless otherwise stated.

We use the term additionality in relation to the effect of RECs and other market-based instruments on the generation of renewable energy. The additionality term originated in the carbon-offsetting literature^31,33. In the context of using market-based instruments for renewable energy to report emission reductions, we here follow the definition of the term from the ISO 14064-2 standard³⁹: “Additionality, as a concept, describes the relationship between cause and effect. For any cause and effect, the effect can be described as being additional if it would not have occurred in the absence of the cause”. In this study, the effect can both refer to the generation of renewable energy and the development of new capacity (for example, windmill or solar panels) for renewable energy generation.

Company data

Our initial sample was composed of the 813 companies with approved SBTs as per the online SBTi database⁷ on July 20th, 2021. The SBTs in the SBTi database generally contain the following data of relevance to this study: sector, emission metric (absolute or intensity-based), emission scope (1, 2, 3 or a combination), base year, target year and targeted percentage reduction in the emission metric. We first excluded 28 energy generators and utilities (3% of the initial sample) due to our focus on companies that purchase energy. We excluded another 101 companies (12% of the initial sample) that only have intensity-based SBTs covering scope 2 to avoid the additional uncertainty associated with converting intensity targets to absolute emission targets. For companies with multiple SBTs covering different emissions scopes and target years, we selected a single SBT, prioritizing targets specifically for scope 2, when available (otherwise, we prioritized targets for scopes 1 and 2 combined over targets for scopes 1, 2 and 3 combined), followed by prioritization of the shortest target time span (that is, the difference between the base year and target year). In addition to the target data sourced from SBTi, we collected information on the scope 2 accounting approach that each SBT refers to (market- or location-based) from company disclosures to CDP (formerly the Carbon Disclosure Project)⁴⁰ (this information is not provided by SBTi; see Supplementary Section 2.4 for more details). Note, however, that only around half of companies with approved SBTs reported to CDP. We were therefore left with 338 of the 813 companies after removing energy generators and utilities (28), intensity targets (101) and companies that did not report to CDP (346).

For these 338 remaining companies, we analyzed their past emissions for 2015–2019. The start year (2015) aligns with the publication of the Greenhouse Gas Protocol standard related to market-based accounting (see Box 1)⁹, the introduction of a distinction between market- and location-based scope 2 emissions in the CDP questionnaire⁴⁰ and the approval of the first SBTs by SBTi⁷. We ended the period in 2019 to avoid abnormalities caused by COVID-19 and as 2019 is the most recent year for which complete data was available for many companies. We used companies’ 2015–2019 emissions disclosed to CDP and, in cases of missing or apparently erroneous data, complemented with emission data from the Bloomberg database³⁷. We also used CDP data on companies’ purchased energy (electricity, heat, steam and cooling) and purchased RECs and PPAs (details in Supplementary Section 1). We adjusted ambiguous or apparently erroneous datapoints (details in Supplementary Section 2). We excluded 223 companies (27% of the initial sample and 66% of the restricted sample of 338 companies) due to missing or poor-quality data, leaving 115 companies (14% of the initial sample) that are included in our analysis (details in Supplementary Section 2).

The resulting sample of 115 SBTs have been approved by SBTi between 2016 and 2021. We used SBTi’s classification of headquarter regions and industries⁷ to assess the representativeness of the final company sample (Supplementary Section 4) and estimate combined results at the region level (Supplementary Figs. 7 and 9). For combined results at the industry level (Supplementary Figs. 8 and 10) and the industry-level assessment of scope 2 emissions relative to other emission scopes (Supplementary Figs. 11–13), we used the industry classifications of CDP⁴⁰ as it involves a more manageable number of industries (10 versus the 49 of SBTi’s classification system⁷).

Reference mitigation pathways for the Paris temperature goal

We used SBTi’s linear 1.5 °C and well below 2 °C global mitigation pathways⁴¹, involving annual reductions of 4.2% and 2.5% of base year emissions, respectively. The SBTi developed these pathways from a subset of the pathways described in the Special Report on Global Warming of 1.5 °C of the Intergovernmental Panel on Climate Change⁴². SBTi determined this subset by applying criteria related to temperature limit probability, temporary overshoot of emission budget, year of peak emissions and near-term emission reduction rate, with the aim of isolating pathways conforming with principles of plausibility, responsibility, objectivity, and consistency⁴¹. The SBTi notes that linearization of emission pathways over long timespans can result in substantial deviations of the pathways’ cumulative emissions and therefore recommends the use of the derived reduction rates (4.2% and 2.5%) for the shorter time span of 2020–2035. However, the SBTi also advises companies to apply these reductions rates to set SBTs for base years before 2020⁶, and SBTi applied the requirement of the 1.5 °C pathway (4.2% reduction in base year emissions per year) as a benchmark for the combined emission trajectory of companies with SBTs in the 2015–2019 period in its latest target progress report⁸. Following SBTi, we here apply the annual emission reduction rates of the two SBTi pathways (4.2% and 2.5%, respectively) as references to evaluate the Paris alignment of past corporate emission trajectories (2015–2019) and future targeted trajectories (median values 2017–2030) for the 115 companies.

Estimating the contribution of scope 2 to total emission changes for 2015–2019

Based on the equation for calculating market-based scope 2 emissions (Box 1), we used the following set of equations to calculate the contribution of changes in energy consumption, use of PPAs, use of RECs (collectively referred to as market-based instruments or MBI) and residual grid mix emission factor to reported changes in market-based scope 2 emissions for each year (that is, 2016, 2017, 2018 and 2019, noted with ‘t’ below) relative to 2015 (Figs. 1 and 2).

Contribution to reported emission changes (ΔE) from change in energy consumption (C) relative to 2015 (in tCO₂e per year):

$${\Delta}{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{C}}}},{{{\mathrm{t}}}}} = {{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,2015} \times \frac{{{{{{C}}}}_{2015} - {{{{C}}}}_{{{\mathrm{t}}}}}}{{{{{{C}}}}_{2015} - {{{{C}}}}_{{{{\mathrm{MBI}}}},\,2015}}}$$

(1)

Contribution to reported emission changes from change in energy consumption that is covered by PPAs (C_PPA) relative to 2015 (in tCO₂e per year):

$${\Delta}{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{PPA}}}},{{{\mathrm{t}}}}} = {{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,2015} \times \frac{{{{{{C}}}}_{{{{\mathrm{PPA}}}},\,{{{\mathrm{t}}}}} - {{{{C}}}}_{{{{\mathrm{PPA}}}},\,2015}}}{{{{{{C}}}}_{2015} - {{{{C}}}}_{{{{\mathrm{MBI}}}},\,2015}}}$$

(2)

Contribution to reported emission changes from change in energy consumption that is covered by RECs (C_REC) relative to 2015 (in tCO₂e per year):

$${\Delta}{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{REC}}}},{{{\mathrm{t}}}}} = {{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,2015} \times \frac{{{{{{C}}}}_{{{{\mathrm{REC}}}},\,{{{\mathrm{t}}}}} - {{{{C}}}}_{{{{\mathrm{REC}}}},\,2015}}}{{{{{{C}}}}_{2015} - {{{{C}}}}_{{{{\mathrm{MBI}}}},\,2015}}}$$

(3)

Contribution to reported emission changes from change in the residual grid mix emission factor (EF_res) relative to 2015 (in tCO₂e per year):

$$\begin{array}{l}{\Delta}{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{EF}}}}\_{{{\mathrm{res}}}},{{{\mathrm{t}}}}}\\ = \left( {{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,2015} - {{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{{{\mathrm{t}}}}}} \right) - ({\Delta}{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{C}}}},{{{\mathrm{t}}}}} + {\Delta}{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{PPA}}}},{{{\mathrm{t}}}}} + {\Delta}{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{REC}}}}})\end{array}$$

(4)

Note that equation (4) estimates the contribution from a change in the residual grid mix emission factors as the difference between the change in reported emissions (the first parenthesis) and the sum of the contributions to emission changes from changes in energy consumption, PPAs and RECs (the second parenthesis) as companies do not directly disclose the values of EF_res behind their reported emissions to CDP. Note also that equations (1–4) are based on the approximation that EF_MBI is zero for all RECs and PPAs, which allows treating RECs and PPAs as if they virtually eliminate the energy consumption they cover in equation (1–3). This approximation is needed as companies only reported EF_MBI values to CDP for part of the study’s time period (2016, 2017 and 2018). We consider the approximation that EF_MBI is zero in all cases sound, as biomass and biogas incineration is the only type of renewable energy generation known to emit (small amounts of) GHGs during operation (in addition to emission of biogenic CO₂, which is considered climate neutral in corporate accounting²⁴), and because this energy source accounted for a negligible share of the combined energy consumption claimed by RECs or PPAs (C_MBI) across the company sample (less than 1% in 2019). We note that some companies reported non-zero values of EF_MBI for other types of renewable energy generation such as hydropower, wind and solar, and we assume that this reflects a misunderstanding of the CDP questionnaire (for example, reporting life-cycle emission factors instead of operational emission factors).

We next calculated adjusted market-based scope 2 emissions for each year following 2015 (t) (aE_MB,S2, in tCO₂e per year) by removing the contribution from RECs to reported emission changes (see equation (3)):

$${{{{{\mathrm{a}}E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,\,{{{\mathrm{t}}}}} = {{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{{{\mathrm{t}}}}} + {\Delta}{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{REC}}}},{{{\mathrm{t}}}}}$$

(5)

For location-based accounting, we used a similar set of equations, based on the equation for calculating location-based scope 2 emissions (Box 1), to calculate the contribution of changes in energy consumption and the average grid mix emission factor to reported changes in scope 2 emissions between 2015 and each following year (t):

Contribution to reported emission changes from change in energy consumption (C) relative to 2015 (in tCO₂e per year):

$${\Delta}{{{{E}}}}_{{{{\mathrm{LB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{{C}}}},{{{\mathrm{t}}}}} = {{{{E}}}}_{{{{\mathrm{LB}}}},{{{\mathrm{S}}}}2,2015} \times\frac{{{{{{C}}}}_{2015} - {{{{C}}}}_{{{\mathrm{t}}}}}}{{{{{{C}}}}_{2015}}}$$

(6)

Contribution to reported emission changes from change in the average grid mix emission factor (EF_mix) relative to 2015 (in ton CO₂e per year):

$${{{{E}}}}_{{{{\mathrm{LB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{EF}}}}\_{{{\mathrm{mix}}}},{{{\mathrm{t}}}}} = \left( {{{{{E}}}}_{{{{\mathrm{LB}}}},{{{\mathrm{S}}}}2,2015} - {{{{E}}}}_{{{{\mathrm{LB}}}},{{{\mathrm{S}}}}2,{{{\mathrm{t}}}}}} \right) - {\Delta}{{{{E}}}}_{{{{\mathrm{LB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{{C}}}},{{{\mathrm{t}}}}}$$

(7)

Annualization of SBTs

As shown in Fig. 3a, to allow for comparison of the targeted emission reductions across SBTs with varying timespans, we annualized the reported SBTs, formulated as a percentage reduction between the company-specific base year (B) and target year (Y), following Wang et al⁴³:

$${{{\mathrm{SBT}}}}^\prime = \frac{{{{{\mathrm{SBT}}}}}}{{{{{{Y}}}} - {{{{B}}}}}}$$

(8)

The resulting annualized SBT (SBT′) indicates an annual percent reduction of base year emissions and assumes a linear emission trajectory between base year and target year.

Estimation of SBTs specific for scope 2

As shown in Fig. 3b, most companies in our sample (94%) do not have SBTs specifically for scope 2. Instead they have targets that combine scope 1 and 2 (82%), or scope 1, 2 and 3 emissions (12%). To derive SBTs specifically for scope 2 in these cases, we first calculated the past contribution of scope 2 emission changes (S2C_past, dimensionless) to the emission reductions for the emission scopes covered by a company’s reported SBT:

$${{{\mathrm{S}}}}2{{{\mathrm{C}}}}_{{{{\mathrm{past}}}}} = \frac{{{{{{E}}}}_{{{{\mathrm{S}}}}2,{{{{b}}}}} - {{{{E}}}}_{{{{\mathrm{S}}}}2,{{{{y}}}}}}}{{{{{{E}}}}_{{{{\mathrm{SR}}}},{{{{b}}}}} - {{{{E}}}}_{{{{\mathrm{SR}}}},{{{{y}}}}}}}$$

(9)

Here, E_S2 is the reported historic scope 2 emissions and E_SR is the reported historic emissions for the scopes covered by the company’s reported SBT, both referring to market-based accounting for companies with market-based SBTs and location-based accounting for companies with location-based SBTs; b and y are company-specific start- and end-years within the 2015–2019 period, selected to represent the longest consecutive decrease in E_SR. This approach (as opposed to a common start- and end-year) avoids past increases in E_SR, which is desirable given that S2C_past forms the basis for projecting a company’s future scope 2 emissions in the context of a targeted emission decrease across the emission scopes covered by its reported SBT. On average across our sample, the b to y period spans 2.5 years. For one of the sample companies, E_SR increased consistently in the 2015–2019 period, meaning there was no company-specific basis for projecting scope 2 emissions. For that company we instead assigned S2C_past the average value calculated for the other sample companies with reported SBTs covering the same emission scopes (1–3, market-based), which was 13.4%.

Next we extrapolated from the past reference period (b to y) by assuming that S2C_past applies to the full period between the specified base year, B, and future target year, Y, of the reported company SBT and calculated the corresponding targeted annual reductions in scope 2 emissions for that period ($\Delta E^{\prime}_{\mathrm{S2}}$) (again, market- or location-based, depending on the reported SBT):

$${{{\mathrm{{\Delta}}}E}}_{{{{\mathrm{S}}}}2}^\prime = {{{{E}}}}_{{{{\mathrm{SR}}}},{{{\mathrm{B}}}}} \times{\Delta}{{{\mathrm{SBT}}}}^\prime \times {\Delta}{{{\mathrm{S}}}}2{{{\mathrm{C}}}}_{{{{\mathrm{past}}}}}$$

(10)

Finally, we estimated the annualized SBT specifically for scope 2:

$${{{\mathrm{SBT}}}}_{{{{\mathrm{S}}}}2}^\prime = \frac{{{{{{{\Delta}}}E}}_{{{{\mathrm{S}}}}2}^\prime }}{{{{{{E}}}}_{{{{\mathrm{S}}}}2,{{{{B}}}}}}}$$

(11)

Adjustment of scope 2 SBTs for RECs

Given that many companies in our sample used RECs to report market-based emission reductions in the 2015–2019 period, it is likely that they will continue to use RECs as a contribution to meeting their scope 2-specific SBTs. To take this into account, we modified equation (9) to remove the contribution of RECs (Fig. 3c) to past changes in scope 2 emissions, by drawing on equations (3) and (5), which were modified to cover the b to y reference period (same as for equation (9)) instead of the 2015 to t period:

$$\begin{array}{l}{{{\mathrm{aS}}}}2{{{\mathrm{C}}}}_{{{{\mathrm{past}}}}} = \frac{{{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{{{{b}}}}} - {{{\mathrm{aE}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{{{{y}}}}}}}{{{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{SR}}}},{{{{b}}}}} - {{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{SR}}}},{{{{y}}}}}}}\\\qquad\qquad = \frac{{{{{{E}}}}_{{{{\mathrm{S}}}}2,{{{\mathrm{mark}}}},{{{{b}}}}} - \left( {{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{{{{y}}}}} + {\Delta}{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{\Delta}{{{\mathrm{REC}}}},{{{{y}}}}}} \right)}}{{{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{SR}}}},{{{{b}}}}} - {{{\mathrm{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{SR}}}},{{{{y}}}}}}}\\\qquad\qquad = \frac{{{{{{E}}}}_{{{{\mathrm{S}}}}2,{{{\mathrm{mark}}}},{{{{b}}}}} - \left( {{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,{{{{y}}}}} + {{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{S}}}}2,2015} \times \frac{{{{{{C}}}}_{{{{\mathrm{REC}}}},\,{{{{y}}}}} - {{{{C}}}}_{{{{\mathrm{REC}}}},\,{{{{b}}}}}}}{{{{{{C}}}}_{{{{b}}}} - {{{{C}}}}_{{{{\mathrm{MBI}}}},\,{{{{b}}}}}}}} \right)}}{{{{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{SR}}}},{{{{b}}}}} - {{{{E}}}}_{{{{\mathrm{MB}}}},{{{\mathrm{SR}}}},{{{{y}}}}}}}\end{array}$$

(12)

aS2C_past can be interpreted as the contribution of a real change in scope 2 emissions (that is, without the contribution of RECs) to a reported change in emissions for the scopes covered by the reported SBT. Following this, we calculated the adjusted scope 2-specific annualized SBTs (${\mathrm{aSBT}}^{\prime}_{{\mathrm{S2}}}$) by using equations (10) and (11) with aS2C_past as input instead of S2C_past. For the sample company with a consistent increase in E_SR in the 2015–2019 period, we (similarly to above) assigned aS2C_past the average value calculated for the other sample companies that have SBTs that cover the same emission scopes (1–3, market-based), which was 11.9%.

Contribution of RECs to committed cumulative scope 2 emission reductions

Based on the above estimations of ${\mathrm{SBT}}^{\prime}_{{\mathrm{S2}}}$ for market- and location-based targets, and ${\mathrm{aSBT}}^{\prime}_{{\mathrm{S2}}}$ for market-based targets, we estimated the contribution of RECs to the cumulative emission reductions from the base year, B, to the target year, Y, for each SBT, again assuming a linear pathway (Fig. 4). First, we calculated the reductions in scope 2 emissions (market- or location-based, depending on the reported SBT) between Y and B:

$${{{\mathrm{{\Delta}}}E}}_{{{{\mathrm{S}}}}2} = {{{{E}}}}_{{{{\mathrm{S}}}}2,{{{{B}}}}} \times{{{\mathrm{SBT}}}}_{{{{\mathrm{S}}}}2}^\prime \times \left( {{{{{Y}}}} - {{{{B}}}}} \right)$$

(13)

For 33 companies, equation (13) resulted in a ΔE_S2 larger than E_S2,B, which would imply negative scope 2 emissions in Y. This is unlikely to occur (the median Y is 2030) and we therefore assumed that these companies will stop reducing scope 2 emissions when they reach a value of zero. We therefore corrected the value of ΔE_S2 to E_S2,B in these cases and corrected Y to the time of zero scope 2 emissions:

$${{{{Y}}}} = {{{{B}}}} + \frac{1}{{{{{\mathrm{SBT}}}}_{{{{\mathrm{S}}}}2}^\prime }}$$

(14)

For each SBT, we then calculated the cumulative reduction of scope 2 emissions between B and Y:

$${{{\mathrm{cum}}}}{\Delta}{{{{E}}}}_{{{{\mathrm{S}}}}2} = {\Delta}{{{{E}}}}_{{{{\mathrm{S}}}}2} \times \left( {{{{{Y}}}} - {{{{B}}}}} \right) \times \frac{1}{2}$$

(15)

For market-based SBTs, we further calculated the annual scope 2 emission reduction between B and Y adjusted to remove the REC contribution (aΔE_S2) by using ${\mathrm{aSBT}}^{\prime}_{{\mathrm{S2}}}$ instead of ${\mathrm{SBT}}^{\prime}_{{\mathrm{S2}}}$ as input in equation (13). In cases where ΔE_S2 is larger than E_S2,B, we downscaled aΔE_S2 on the basis of the correction of ΔE_S2 (described above):

$${{{\mathrm{a}}}}{\Delta}{{{{E}}}}_{{{{\mathrm{S}}}}2} = {{{{E}}}}_{{{{\mathrm{S}}}}2,{{{{B}}}}} \times \frac{{{{{\mathrm{aSBT}}}}_{{{{\mathrm{S}}}}2}^\prime }}{{{{{\mathrm{SBT}}}}_{{{{\mathrm{S}}}}2}^\prime }}$$

(16)

We then calculated the adjusted cumulative emission reductions with the contribution of RECs removed (a.cumΔE_S2) by using aΔE_S2 instead of ΔE_S2 as input in equation (15). We then estimated the cumulative contribution from RECs to emission reductions (REC.cumΔE_S2) as the difference between cumΔE_S2 and a.cumΔE_S2. Finally, we aggregated REC.cumΔE_S2 for all market-based SBTs and compared it with the aggregated cumΔE_S2 for all market-based SBTs and the aggregated cumΔE_S2 for all location-based SBTs.

Note that five of the sample companies reduced their purchases of RECs between b and y (the start- and end-years in equation (9)) to such an extent that a continuation would lead to a projected negative use of RECs in Y or earlier, which would not be possible. For the estimation of REC.cumΔE_S2 of these five companies, we therefore corrected ${\mathrm{aSBT}}^{\prime}_{{\mathrm{S2}}}$, assuming that the use of RECs would decrease linearly from B until reaching a value of zero in Y:

$${{{\mathrm{corrected}}}}.{{{\mathrm{aSBT}}}}_{{{{\mathrm{S}}}}2}^\prime = {{{\mathrm{SBT}}}}_{{{{\mathrm{S}}}}2}^\prime + ({{{\mathrm{aSBT}}}}_{{{{\mathrm{S}}}}2}^\prime - {{{\mathrm{SBT}}}}_{{{{\mathrm{S}}}}2}^\prime ) \times \frac{{{{{{C}}}}_{{{{\mathrm{REC}}}},{{{{B}}}}}}}{{({{{{C}}}}_{{{{\mathrm{REC}}}},\,{{{{b}}}}} - {{{{C}}}}_{{{{\mathrm{REC}}}},\,{{{{y}}}}}) \times \frac{{{{{{Y}}}} - {{{{B}}}}}}{{{{{{y}}}} - {{{{b}}}}}}}}$$

(17)

Note that the SBT of one of the five companies has 2011 as B, which is before the Greenhouse Gas Protocol began permitting the use of RECs through the market-based accounting approach⁹. To account for the reported use of RECs between b and y, we used b instead of B in equation (17) for this company. The corrections of ${\mathrm{aSBT}}^{\prime}_{{\mathrm{S2}}}$ for these five companies had a modest influence on results, causing the aggregated REC.cumΔE_S2 (illustrated in Fig. 4) to change from 97 MtCO₂e to 101 MtCO₂e.