Engine number 1 uses independent sources and estimates to assess the firm level costs of emissions, resource use, waste, social practices, and a host of other ESG factors. The framework includes a range of factors including climate change, air quality, water consumption, land use, waste, human rights, corruption, bribery and fraud, community relations, customer privacy and data security, product quality and safety, employee health and safety, training and development, diversity, equity inclusion among others. The evolving framework is comprehensive across the supply chain, capturing the financial costs of upstream and downstream impacts, as well as those stemming from the firm's own operations, what we called scope 1, 2, and 3 emissions earlier. As we refine the investment strategies by sector and incorporate more measures, the power of this framework will become even stronger. Right now, some factors are easier to measure than others. Tracking the social cost of greenhouse gases is relatively easy. Techniques are increasingly available for converting companies emissions into tons of carbon dioxide equivalents and calculating the effect on global welfare. Similarly, it's possible now to compute the social cost of other types of air pollution, the negative effects of land and biodiversity waste, water consumption, the adverse health consequences of tobacco and alcohol, workplace injuries, fatalities. The social consequences of wages that fall below a minimum standard or the financial benefits of worker training and voluntary community donations, all of this can be monetized at present. Of all the factors in the framework, GHG data is the most widely reported, either by the companies themselves or a host of specialist data providers. But a lack of objective audited data still remains the biggest obstacle to assessing a company's ESG performance, including its carbon emissions. Further, we believe a lack of visibility or accountability, as it concerns scope 3 emissions, a lack of objective audited data remains the biggest obstacle to assessing a company's ESG performance , including on GHGs. A particular concern are scope 3 emissions estimates. As recently as the start of 2020, for example, neither BP nor ExxonMobil had ever published estimates of its downstream emissions impacts. These are the impacts of the people who buy the oil and gas from BP or Exxon Mobil and burn it in their cars, boats, planes, or other uses. Where companies like ExxonMobil and BP failed to report the relevant information, we and others have to impute the data from models that draw upon a wide variety of sources including input-output tables developed by the United Nations, or the International Labor Organization, or the OECD, or the European Union. In the absence of reliable primary data from the companies, we have to settle for adjusted averages or estimates of GHG and other ESG factors. In all cases, we go through great efforts to validate these results, looking for outliers, looking for discontinuous jumps, always trying to improve the algorithms we're using to fill in the gaps. But at the end, these are still estimates. The magnitude of these net externalities vary significantly across sectors. Some sector showed 10 times that of another sector's total externalities per dollar of revenue. That makes it paramount when we're looking for correlations between externalities and shareholder value to look within the industries to compare apples to apples, not try to compare ExxonMobil to, say, Facebook. Our initial analysis strongly supports a far more consistent and powerful association that found via traditional ESG data. It shows that the difference between a firm's total value, and its shareholder value, and changes in those net externality relative to its industry peers are strongly indicative of future changes in financial outcomes. Between 2010 and 2019, for example, this figure shows that the 10 firms in the S&P 500, with the largest negative impacts as ranked in 2019, substantially underperformed the market in terms of share price. The 10 companies with the smallest negative impacts meanwhile significantly outperformed the market. By contrast, when we use traditional ESG measures to analyze the same groups of top 10 and top bottom firms in the S&P 500, we see no consistent pattern or correlation with share price. The pattern is similar when the engine number 1 methodology is applied to the best and worst performers in each sector. As this figure shows, in all but 1 of the 10 sectors, the worst performing firm on total value, those with the absolute highest negative externality as measured by the total value framework, dramatically underperformed the median firm in their sector. In the majority of cases, by more than 50 percent over a decade. If we follow this logic, we can look at the total portfolio returns based on the performance of companies in the total value framework. In this figure, for example, we analyze the performance of S&P 500 firms between December 2nd, 2011 and August 9, 2021. Using the total value framework and separating firms into five quintiles, the top quintile, number 1, represents the set of firms with the lowest total value score or the largest negative externalities. These substantially underperformed the S&P 500 benchmark. Five represents the highest quintile of total value score or the smallest negative impacts, which dramatically outperformed the benchmark. While this doesn't account for the evolution to come in the data framework and it holds assumptions about the time period constant, we can still see in the chart that the framework can already be an important and impactful methodology to deploy towards generating favorable financial returns while at the same time quantifying ESG impacts, not just providing ratings and rankings, but providing dollar estimates. Subsequent versions of this framework are going to seek to employ a deeper sector-specific analysis, better to find weighting schemes, more direct attribution statistics. We're seeking to bring a new framework to ESG investing, which we hope carries significant upside potential. Our analysis further shows that the strength of the correlation increases across each decile of total value performance with the strongest relationship observed among the top-performing firms. The lesson is clear. Mitigating negative externalities or contributing positive ones is a key differentiator of top-performing firms for over half the companies in the S&P 500. How does the data offered by the total value framework improve the ESG investing strategies? Because it monetizes the impact of ESG factors. Through this monetization, investors are able to take this data and incorporate it within their traditional financial analysis, without having to carry out bespoke and time-consuming assessments required by current ESG data to link emojis to dollars. The objective nature of the data will also allow total value accounts to be properly audited, standing in stark contrast to the voluntary or custom third-party reporting that is common at the moment. While engine number 1's total value framework represents a major breakthrough, there's still substantial room for improvement. Of all the possible pathways linking shareholder to stakeholder value, not all are currently integrated into the framework because we can't get the data necessary. Looking ahead, we're going to keep looking for new sources of data, procuring these from third-party providers or where necessary working independently at engine number 1 or at the Wharton ESG Analytics Lab to develop them ourselves. In this way, we can realize the full potential of this new way of seeing value. More advanced techniques based on AI will provide further opportunities for improvement. We'll continue to experiment with data imputation to plug the holes. It's important to note that the missing information about ESG is far from random. High-performing firms are generally among the first to report, while laggards like ExxonMobil were the last. We're trying to better track the speed with which companies internalize these externalities, and we furthermore hope to identify new factors that'll drive the internalization of externalities and give us better signal as to the timeline along which this is going to occur. The current data structure already allows us to capture the time-varying stakeholder salience of different ESG issues. One way to illustrate this is to look at how over the last decade, which includes the period of the BP oil spill at the Gulf, stakeholders in the upstream oil and gas industry were heavily focused on accidents and short-term environmental damage. Such news comprised 24.3 percent of all the stories on ESG for the upstream oil and gas industry in 2010. By the time we get to the end of the period though, 2020, however, GHG, greenhouse gas emissions, and climate risk dominated the news. Stories on GHG rose from 19.4 percent of all ESG news in 2010 to 28 percent in 2020, whereas news about accidents fell from 24 percent to 7.8 percent. The total value framework methodology allows these shifts to influence the likelihood that the impact of total value from GHGs will be going up on shareholders and the impact of accidents will be going down. The upstream oil and gas industry also provides an example of the impact of technological innovation. In 2010, the overall ESG performance of ExxonMobil, Chevron, ConocoPhillips was basically on par with the rest of the sector. But as the 10 years unfolded, these companies started to lag behind their peers on GHG emission reduction. As a result, that lagging performance on ESG factors stood out more to investors who started downgrading and downweighting their stocks, leading to a stronger impact on financial and shareholder value. The consistently weak link between ESG performance and the share price of the so called FAANG companies, I mean Facebook, Amazon, Apple, Netflix, and Google, demonstrates the role of anti-competitive barriers. In these cases, customers are locked in. Network scale economies make it difficult for them to shift to another technology provider. That limits the fallout from ESG underperformance. Activist pressure such as the campaign by [inaudible] to force Apple to adopt tools enabling better parental controls of their child's iPhone have occasionally resulted in ESG performance improvements. But for the most part, the negative social behavior by FAANG companies related to customer privacy, data security, customer welfare hasn't yet impacted their share price. Even optimists believe that the impact of the recent Facebook whistleblower is going to take years to generate concrete government regulatory change. Looking ahead, we hope to add to these signals, new information on impending legislative regulatory or legal shifts, as well as indications of pressure from within a company's supply chain or the specific geographic communities in which it operates to address the costs imposed on stakeholders. More fundamentally, we want to turn the industry away from separating ESG analysis from other financial and operational analysis. We want to encourage the integration of the two, shifting the focus to the dollar value of a firm's total value creation or destruction and the likelihood of its internalization by shareholders. Reframing the ESG integration process in this way is consistent with the recent emphasis on materiality, but that transformation is still going to require a deep cultural shift. The market requires just such a shift to its approach to ESG measurement. The monetization of ESG performance as provided by engine number 1's total value framework is an important step in this evolution. We anticipate that the strong, albeit preliminary evidence, that links between value and value use will spur interest among a wide range of investors, academics, and policymakers. We together with other researchers and practitioners look forward to working together to continue to advance the frontier of academically rigorous but practically relevant ESG data.
