As the article says, there are actually laws that require internal research that shows a product may be harmful to be made available to the EPA. That just doesn't do anything if the company either lies, or the EPA doesn't end up taking it seriously. As is often the case, the mechanism is there. The politics are not.
Bad mechanism design. Publication bias is bad even in academia, but within a corporate context there are overwhelming incentives to bury research that shows harms. Making safety requirements that work, for anything, requires independent auditors and researchers, and allowing the value of corporate secrecy to override laws ensuring safety mean that safety will never be a priority in a meaningful sense. A requirement about internal research is nearly useless, mostly enabling issuing some trifling fines way after the fact of something going horribly wrong
