Carbon intensity — scaling GHG emissions by a company’s revenue — is a common measure used to build low-carbon portfolios. It bridges the gap between non-financial ESG metrics and financial performance, offering a sorely needed integrated point of evaluation. Scaling by net income or some other profit metric might provide even more meaningful comparisons, but this is nit picking. The real issue with carbon intensity measures is the way the GHG Protocol treats electricity.
Purchased electricity, the largest source of operational emissions for many companies, falls under Scope 2 of the GHG Protocol. Companies have a choice of two different approaches for measuring emissions from this electricity. First, they can use emissions information related to the electricity grid where facilities are located. This is known as the “location-based” approach. As an alternative, they can take into account specific contractual relationships with electricity suppliers, including purchases of renewable energy. This is the “market-based” approach.
For many companies, the measured Scope 2 emissions under each approach are similar. But as companies are under increasing pressure to address their carbon emissions and achieve “net-zero” at some point in the next generation, divergence is likely. This is already evident with some of the largest U.S. companies that have made a major push to power their operations with 100% renewable energy.
Take Apple. Their Scope 2 GHG emissions under the location-based approach are nearly 100x what they report under the market-based method. Those figures are 16x for Microsoft and Mastercard, 6x for Alphabet (Google), and nearly 2x for Visa. For the largest 25 companies in the S&P 500, the median level of market-based Scope 2 emissions is almost 20% smaller than its location-based counterpart. Naturally, the lower market-based figures are what these companies use in their communications about their carbon footprint.
Lower market-based Scope 2 measures are a consequence of a well-intended system to incentivize the construction of renewable generation. Renewable developers are able to bundle the sale of renewable electricity with separately marketable renewable energy certificates in the U.S. (RECs) or guarantees of origin (GoO) in Europe. These instruments certify that the bearer contributed to the generation of 1 MWh of electricity through its purchase.
Not all RECs are created equal. Location matters. The GHG emissions reduction from incremental renewable energy sources depends upon how much fossil-fuel generation is displaced. Callaway et al. (2015) suggest renewable energy that displaces coal-fired generation in the upper midwest during the winter may have twice the impact of displacing natural-gas power in California. Although large corporate renewable energy purchasers like Google make an effort to match RECs with the grids they operate on, this is still not practical for companies with global footprints. As a result, the emissions reductions computed from the application of RECs may bear little resemblance to the physical reality of a company’s electricity consumption.
Crude accounting issues are problematic, but can be overcome. The bigger issue is the potential lack of additionality. Gillenwater (2013) finds that RECs are “insufficient to even have a marginal impact on wind power investment in the United States.” This conclusion was drawn seven years ago, when the cost of new wind generation was higher than it is today. The latest Energy Information Administration (EIA) report on the costs of renewables suggests wind and solar power are competitive with new natural-gas powered generation — even after subsidies (which are being wound down) are excluded. The EIA is particularly bullish on utility-scale solar in most parts of the U.S., given its evaluation of the cost of generation displaced (i.e., the value created) as well as the cost of the renewable source itself — implying that more solar energy will be added to the grid than combined-cycle natural gas simply on existing economics.
If economic parity between new fossil-fuel and renewable energy sources has been achieved, the key additionality condition on which RECs rest is shaky. If projects would have gone ahead anyway without non-utility purchasers signing contracts, the credibility of a corporate purchaser claiming that their electricity is clean is impaired.
Electricity, like money, is fungible. The physical GHG emissions of the electricity drawn from the grid can’t be partitioned by intentions. The only fair allocation is the average emissions from whatever sources were active at the time the electricity was consumed. That takes us back to the location-based alternative for computing Scope 2 emissions. Socially responsible investors are advised to ignore market-based measures for computing a company’s total GHG output. In addition to altering comparisons of carbon intensity, a focus on location-based Scope 2 emissions will improve company models that incorporate a social cost of carbon.
Should companies get any credit for their efforts to source electricity from renewable sources? Absolutely. Corporate purchases of RECs indicate a commitment to sustainability that may manifest itself in several ways, such as through greater success with efficiency improvement. When projecting the path of a company’s emissions intensity, we might expect companies that demonstrate sustainability commitments to shrink their resource use at a faster rate than other companies. So while the delta between GHG emissions computed with location- and market-based approaches makes little difference for the level of emissions used in our analyses, it may prove indicative of the right negative growth rate to use.