George Serafeim    T. Robert Zochowski

Within this article we describe a new dataset that includes ~13,000 firm-year observations, from publicly listed firms globally for the years 2010-2018, that we are making available to researchers. This dataset has been used in our paper “Corporate Environmental Impact:  Measurement, Data and Information” and it is part of the Impact-Weighted Accounts Project at Harvard Business School.

The dataset can be accessed freely on the Impact-Weighted Accounts Project website. No account needs to be created other than submitting basic information (name, email, and organization) about the user of the data.  A version of the dataset that includes along with company name, an ISIN identifier, is provided by the authors to academic researchers upon request. No restrictions to use the data apply other than citing the original source when using the data.

The Problem

Since the industrial revolution, for-profit organizations have been incentivized to treat the environment as a resource to be exploited in advancing the best interest of their owners. The results have been entirely consistent with the economic problem known as the ‘tragedy of the commons,’ in which a shared resource is spoiled or destroyed by each agent acting in their own self-interest.

Widely available and consistent data on environmental impacts do not exist. Availability requires either mandatory disclosure regulation or companies voluntarily disclosing comprehensive environmental impacts, not just the most palatable ones. Both are emerging over time and still comparability and completeness of data are far from ideal.

The Solution

Impact-Weighted Accounts at Harvard Business School is working to produce a scalable methodology that increases transparency to impacts on the environment, employees, and customers and promotes comparability between organizations. Critical in this process is transforming the currently disclosed metrics through scientifically studied pathways into impacts to the end stakeholders and then converting those impacts into intuitively understood and comparable monetary values using a range of validated pricing methods including preferences or damage costs. This allows for the construction of accounting statements that include the impact organizations have on society.

Furthermore, our methodology can be used today within existing, albeit imperfect, publicly disclosed data to calculate operational environmental impact (data is still too sparse to include Scope 3 emissions). By supplementing disclosed data with industry-country input-output data from the Exiobase, a public good produced by numerous European research framework programs, we can impute missing data points, which increases comparability. We limit those imputations to avoid large errors in the data and we also disclose the % of environmental impact that is imputed to further transparency. Next, we use impact pathways and monetization frameworks from the Environmental Priority Strategies (Steen, 2019), the Available Water Remaining scarcity models, and the WaterFund production, delivery, and wastewater treatment estimates to produce monetized impact estimates which can then be compared to other organizations as a percentage of sales or operating income or incorporated into valuations and other financial management statistical techniques.

Our Environmental research and data, now available through the Impact-Weighted Accounts website and SSRN, provides a number of take-aways. First, our results demonstrate that our methodology distinguishes between both industry (approximately 60% of variation) and firm specific (approximately 30% of variation) factors driving environmental impact. When compared to environmental ratings from a number of leading providers, our calculated impacts exhibit a significant correlation when comparing firms across industries, but very little correlation when comparing competitors within industry, suggesting that within an industry environmental ratings are almost completely uncorrelated with estimates of environmental intensity.

Critical for investors, environmental impact is correlated with both risk and return. We find a significant relationship between higher (i.e. more damaging) environmental impact scaled by either sales or operating income, herein referred to as environmental intensity, with both lower Price-to-Equity and Tobin’s Q. Additionally, environmental intensity is significantly and negatively related to the Sharpe ratio with more environmentally intensive firms having lower stock returns and higher volatility. Such firms also exhibit higher systematic risk as reflected by higher market beta. Finally, environmental ratings do not exhibit a significant relationship with any of the financial characteristics other than volatility.

These results demonstrate that investors can calculate and compare organizations environmental impact even without an ideal state of disclosure. Furthermore, investors advocating for greater environmental responsibility are also aligned with shareholder interest in market returns and risk.

Author Bios:

George Serafeim is the Faculty Chair of the Impact-Weighted Accounts Project and the Charles M. Williams Professor of Business Administration at Harvard Business School.

Robert Zochowski is the Program Director and Senior Researcher for Multi-Faculty Sustainability and Impact Investing Special Projects, including The Impact-Weighted Accounts Project, the Social Impact Collaboratory, and the Project on Impact Investments at Harvard Business School.

Comments

One response to “An open access dataset of corporate environmental impacts”

1. Frederik de Coninck-Smith says:

   February 16, 2021 at 2:50 am

   Thank you for making this data public – it’s a great start to impact visibility. I’m curious to see the output and also plan to use it for investment research.

   Reply