There's a "relatively low correlation" between language patterns found in prospectuses of mutual funds and exchange-traded funds with "ESG" in their names and a fund's environmental, social and governance rating, a study released Tuesday found.
The results of the four-month analysis by a team of University of California San Diego graduate students of 94 mutual funds and ETFs with ESG in their names came as little surprise to Andrew Behar, CEO of As You Sow, a nonprofit shareholder advocacy organization, which asked UCSD to conduct the analysis.
"We wanted to do a deeper analysis, and that's why we got involved with UCSD data scientists to actually take the prospectus language, tear it apart, look at all the words …," Mr. Behar said in an interview Wednesday. "And, basically, what they came back with is there is no correlation between the prospectus language and what's inside the fund."
As You Sow approached UCSD to oversee the data analysis after discovering that 94 of the 3,000 mutual funds and ETFs in its "Invest Your Values" scorecard had "ESG" in their names, but 60 of those had earned a "D" or an "F" grade on one or more ESG criteria.
The UCSD researchers used natural language processing, a type of artificial intelligence that analyzes human language, according to the report, "MGTF 490 Capstone Project: Identify 'Greenwashing' Funds using NLP in Firms' Prospectuses."
The report shows that what his organization knew anecdotally now has "actual data science around it," Mr. Behar said.
He said As You Sow presented the study to the Securities and Exchange Commission Jan. 6 "and said, 'Look, there's a couple of things you have to do. One is, you have to make the prospectus language machine-readable.'"
Automating the process would make it faster and easier to evaluate prospectuses — including any updates — to determine whether holdings are consistent with the ESG language those prospectuses contain, the CEO said.
Mr. Behar also wants to see a standardized glossary of ESG terms.
"We believe there should be a glossary of words that have actual meaning and that are enforced," he said. "If somebody says something is low-carbon, it should mean the same thing for everybody, and the SEC should enforce it."
An SEC spokeswoman declined to comment.