A study of 579 publicly listed South Korean companies found that a stronger environmental rating was associated with roughly 3% higher sales, while there was no statistically significant direct relationship with return on assets. The researchers identified a positive indirect route from environmental responsibility to financial performance through sales, but the evidence does not show that greener practices caused customers to buy more.
The research by Arturo Garcia of Universidad Autónoma de Nuevo León and Sang-Ho Lee of Chonnam National University was published in Corporate Social Responsibility and Environmental Management on 22 May 2026.
It offers a more useful question than the familiar debate over whether environmental, social and governance policies are good or bad for profits. If environmental responsibility has financial consequences, through which part of the business do they appear?
For the Korean companies in this study, sales were the proposed connecting link.
Researchers followed 579 companies for four years
Garcia and Lee combined Korean ESG ratings with company accounts covering 2019 to 2022. After excluding financial institutions, companies without December financial year-ends and observations with missing variables, they had a balanced panel of 2,316 firm-year observations from 579 companies.
The environmental information came from the Korea Corporate Governance Service, now called the Korea Institute of Corporate Governance and Sustainability. Financial information came from the Kis-Value database.
A company was classified as environmentally responsible when it received a B+ rating or higher for the environmental part of the Korean ESG assessment. The measure was binary. A B+ company and an A+ company therefore received the same value in the analysis, while a company immediately below the threshold was placed in the other group.
That is an important constraint. The study cannot tell us whether each additional improvement in emissions, waste, energy use or environmental management produces a predictable increase in revenue.
Financial performance was measured mainly by return on assets, or ROA. This compares net income with the assets a company uses to run its business. The researchers also repeated the analysis using return on equity.
Sales provided the statistical connection
In the main fixed-effects model, the environmental classification had a coefficient of 0.03 when sales were measured using the natural logarithm of total revenue. In practical terms, that represents an association of about 3%.
Sales were in turn positively associated with ROA. The direct relationship between the environmental classification and ROA was negative but not statistically significant.
When the researchers tested the full indirect path, the estimated effect was 0.128 percentage points of ROA. Its 95% bootstrap confidence interval ran from 0.011 to 0.335 percentage points, remaining above zero. The indirect result also held when return on equity replaced ROA.
The model therefore supports a statistical sequence: observations in which companies were in the stronger environmental category also had higher sales, which were associated with better financial performance.
The study did not observe consumers
The authors interpret sales as a market-based channel linked to green consumerism. That is plausible, but the study did not follow shoppers, measure brand trust or record whether customers knew about a company’s environmental rating.
Its sales variable was total company revenue. Revenue can rise because a business sells more units, raises prices, buys another company, expands overseas or changes its product mix. The figures do not isolate purchases made for environmental reasons.
This is perhaps the most useful distinction in the research. Sales may be a better place to look for the commercial effects of environmental policies than profit alone, but total revenue is still several steps removed from a customer’s motivation.
A stronger test would connect particular environmental actions with product-level volumes, prices, repeat purchases or customer retention. That was outside the scope of this company-level study.
Subgroup results need careful reading
The researchers also compared companies affiliated with South Korea’s large family-controlled business groups, commonly called chaebols, with other companies.
The estimated indirect effect was positive and statistically significant for the chaebol group. Its confidence interval ran from 0.003 to 0.396. For non-chaebol firms, the interval ran from minus 0.033 to 0.390 and included zero.
It would be tempting to conclude that large business groups are better at converting environmental performance into sales. The formal comparison does not justify that conclusion.
The study’s index of moderated mediation, which tests whether the two indirect effects actually differed, had a confidence interval from minus 0.239 to 0.306. Because it crossed zero, the difference between chaebol and non-chaebol companies was not statistically significant.
Visibility and reputation may still matter, as the authors propose. This dataset did not establish a reliable chaebol advantage.
The sales-mediated effect was also stronger in the period the authors labelled post-COVID. The indirect effect for that period was 0.212, with a 95% confidence interval from 0.063 to 0.498.
The time categories were unusual. The researchers treated 2019 as before COVID-19, 2020 as during the pandemic, and both 2021 and 2022 as after it. Their formal comparison found the later indirect effect was stronger than the 2020 effect.
That does not demonstrate that the pandemic made consumers greener. Only four years were included, and the study did not measure consumer attitudes. Changes in regulation, company finances, prices and the economic reopening could also have influenced the pattern.
Could profitable companies simply afford to be greener?
Cause and effect remain the central difficulty. A green rating might help a company attract customers, but a successful company may also have more money and management capacity to invest in environmental programmes.
Garcia and Lee ran additional strict-exogeneity and cross-lagged tests and reported no evidence that reverse causality explained their main result. Their underlying research data are publicly available through Figshare.
Even so, those checks do not turn the analysis into a controlled experiment. The main mediation model uses observational company data and contemporaneous measures, so unmeasured factors that vary over time could still affect both environmental ratings and sales.
A separate 2026 study of 119 Korean listed companies reached a different conclusion using panel vector autoregression and data from 2013 to 2021. Jiyeon Kim and Wooyoung Yang found that profitability predicted subsequent environmental ESG investment, while ESG improvements did not predict later profitability.
The samples, periods and methods were different, so one paper does not cancel the other. Together they show why positive ESG and financial results should not automatically be read as evidence that environmental action caused the improvement.
What companies can reasonably take from the findings
The Garcia and Lee study does not show that environmental spending automatically pays for itself. Nor does it show that a publicity campaign around green policies will increase demand.
It does suggest a practical way to examine the business case. Instead of waiting for an environmental programme to appear in bottom-line profit, companies can test whether it changes sales volume, pricing, customer retention or the mix of products bought.
That measurement should distinguish genuine environmental performance from marketing claims. It should also separate consumer-facing sales from business-to-business contracts and exports, where purchasing decisions may follow very different incentives.
The Korean findings are best treated as evidence of a possible route, not a universal return on going green. Sales may be the missing link, but the study leaves open the most difficult part of the question: whether customers caused that link to operate.
