Analyzing oil and gas firm responses to climate change using NLP

-
How often do oil and gas companies talk about climate change? How do shareholders exert pressure on these issues? This project analyzes the text from 1,656 earnings calls-- over 16,000,000 words in total-- to provide answers.

All available earnings calls for the 10 largest publicly traded oil and gas companies were concatenated into a dataframe. Speech was classified into three categories depending on whether the speaker was a shareholder, a firm representative speaking unprompted, or a firm representative answering a shareholder question. A list of climate-related keywords was then used to extract paragraphs discussing climate change. Finally, depending on the keyword(s) contained therein, the sentiment of the paragraph is classified to reflect whether the speaker appears to accept/reject climate science, environmental regulations, carbon capture and storage, etc. 

The output is visualized below. Blue indicates unprompted firm speech, light blue indicates firm responses, and orange denotes shareholder questions. The search bar limits the view to paragraphs containing a certain word. The dropdown filters on the side indicate sentiment to the questions mentioned; for example, to compare relative rates of climate science denial among oil and gas majors, tick the "rejects climate science" box. The time filter allows one to compare how attitudes towards climate change have evolved over time. Finally, the grey tabs at the top toggle between raw word count and total percentages. 

 The overall picture is bleak. Take Exxon Mobil, for example:
Total word count of earnings calls: 1,922,756 (100%)
Word count of paragraphs containing climate-related keywords: 41,201 (2.14%)
Of those paragraphs, WC of those which were brought up unprompted by the firm: 24,984 (1.29%)
WC of those displaying at least a partial acceptance of climate science: 1,612 (0.08%)
WC of those paragraphs displaying an unambiguous acceptance of climate science: 0 (0%)

 Some key takeaways: 
  1. Out of over 16 million words, the combined word count of paragraphs across all firms which displayed an unambiguous acceptance of climate science was less than 1,500. 
  2. American firms tend reject climate science outright, while European firms combine equivocal support for climate science with an emphasis on carbon capture and storage; this is conveniently not mutually exclusive with increased production and exploration
  3. The amount of shareholder pressure varies considerably between firms.

Comments

Post a Comment

Popular posts from this blog

U.S. Foreign Military Sales Tracker

-
Image
Type the name of a country into the search box to view U.S. arms sales between 2001-2015 What are Foreign Military Sales (FMS) ? FMS is the primary process through which the U.S. Government sells weapons to the governments of other countries. In conducting FMS, the Department of Defense (DoD) acts largely as a brok er between foreign governments and U.S. arms manufacturers. The defense ministries of foreign governments submit a Letter of Request (LoR), which the U.S. State Department reviews. If approved, a DoD agency called the Defense Security Cooperation Agency (DSCA) mediates negotiations between the applicant and the contractor, levying a commission on the sale to finance its own operations. As such, the FMS program is operated at no net cost to taxpayers, but is slightly more expensive than private arms sales(Ibid). Once a request is approved by the DoD and DoS, it is sent to Congress for final approval. If the sale is successful, U.S. military personnel
Read more

Did Saudi Arabia Know an Attack was Coming?

-
Image
On September 14th 2019, a combination of drone and missile attacks targeted Saudi Arabia’s oil infrastructure at the Abqaiq and Khurais Aramco facilities, cutting the country’s oil production by over 50% and leading to the largest spike in the price of oil ever recorded . The attack was particularly embarrassing for Riyadh because billions of dollars' worth of air-defense systems were thwarted by drones that may have cost as little as $15,000 . A year later, the expert consensus seems to be that the attack was successful because the Saudis were caught off guard, with the element of surprise giving the assailants a decisive edge.  An open source analysis of radar interference from Saudi Arabian air defenses suggests that the Kingdom might actually have been expecting an Iranian attack on its oil infrastructure: five days before the attack, a missile defense system pointed at Iran was switched on for the first time at one of the largest refineries in the country. This analysis also
Read more

Calculating High-Resolution Standardized Precipitation-Evapotranspiration Index

-
Image
This post outlines a workflow for the calculation of a high-resolution version of the Stardardized Precipitation-Evapotranspiration Index (SPEI). This is accomplished in four steps, utilizing Python, R, and JavaScript. Quantifying the precise onset, duration, and end of droughts is difficult. Most approaches have historically measured drought as a prolonged deviation from the mean rainfall in an area. The Standardized Precipitation-Evapotranspiration Index (SPEI) builds on previous approaches by incorporating temperature data to account for the amount of water lost to evaporation. This is important because in an arid climate, a six week shortage in rainfall during the summer would have a much greater impact on soil moisture than the same shortage during the winter. This is especially important for drought forecasting that accounts for anthropogenic climate change. Indices that do not incorporate temperature data will persistently underestimate drought incidence as global t
Read more