Luka Andren\u0161ek, Katarina Sitar \u0160u\u0161tar, Senja Pollak and Matthew Purver<\/p>\n
ABSTRACT
In this paper, we examine the relationship between the discussion
of Environmental, Social and Governance (ESG) in companies\u2019
annual financial reports and their financial performance. Specifi-
cally, we analyse the companies\u2019 use of specific ESG terms along-
side the performance metric, sector-normalized Return on Assets
(ROA). Our motivation is to determine whether companies fre-
quently mentioning terms such as \u201cgender\u201d, \u201cequality\u201d, \u201ctalent\u201d,
and \u201cinnovation\u201d in their reports demonstrate a higher annual
ROA compared to those that rarely used these terms. To explore
this, we used existing datasets with reports and performance met-
rics from 348 companies, covering the years from 2009 to 2021. In
order to better examine differences, we then selected companies
whose ROA significantly differed from the average (either higher
or lower), allowing for a more pronounced examination of the
impact of ESG term usage on financial performance. The filtered
dataset consisted of 107 companies, with a total of 427 reports;
split into two sections representing higher and lower performing
companies. We then used an existing list of ESG terms derived
from a range of separate data sources, and applied a basic sta-
tistical n-gram language model to extract the probabilities of
each ESG term\u2019s occurrence in each of the higher- and lower-
performing dataset sections. Results show that while certain sets
of ESG concepts correlate with higher financial performance,
others do the opposite, and give some initial interpretation into
the light this sheds on company reporting behaviour.<\/p>\n
\u00a0<\/p>\n\n\n\n