One of the international speakers at the Food Matters Aoteoroa conference was Giles-Eric Seralini, who has been a lightening rod for attention from industry-funded scientists for a few years now (his responses to those criticisms are here). So I was pretty interested in hearing what he had to say.
Here, in my own words, are three things I heard.
1. The components of herbicides and pesticides are often divided into two groups: active ingredients and adjuvants which are also referred to as "inert ingredients". This classification is done by the manufacturer and regulators are then given data from the manufacturer's trials that show the active ingredient to be safe when used as instructed. Regulators scrutinise that work and approve the poison if it looks OK.
2. The adjuvants are rarely if ever tested themselves, and neither are the whole products sold to farmers and gardeners (there can be many different formulations based on the same active ingredient).
3. Some products don't work or work much less effectively without the adjuvants. For example Glyphosate, which is the active ingredient in Roundup, is ineffective as a herbicide in its pure form. It is only effectively toxic when combined with adjuvants. He found that a commercial formulation was many thousands of times more toxic than glyphosate alone.
I should emphasise that these points are from my memory of what Seralini said. Hopefully someone will correct me if I mis-remembered, and I'm sure that someone will correct Seralini if he got any of this wrong. There are certainly differences in regulatory practices across countries, so some variation on these points is possible.
Actually I really hope these three points are seriously wrong, because the implications of them being right are shocking. I quite liked Seralini but I'd be happy to find that he's dead wrong on all of this, because otherwise I'll be forced to conclude that this regulatory system is deeply compromised.
Point 2 is devastating all by itself. If regulators have no information on the safety of the actual product being sold, how can they approve it? Imagine if the warrant of fitness test on cars only made sure the (active ingredient) engine worked OK and ignored the (adjuvant) brakes. Or testing just one ingredient in a cigarette. That would be crazy, which is also why the #1 demand for reform on this site makes a lot of sense to me.
Points 1 and 3 make point 2 worse. They open up an obvious loophole by allowing manufacturers to redefine the active ingredient list so that anything that might not pass a safety test is smuggled in as an adjuvant.
More generally, and stepping away now from Seralini's talk, there are also obvious incentive and oversight problems associated with regulators accepting confidential industry funded safety studies as evidence. Economists have long recognised a risk of "regulatory capture" whereby regulated firms use their superior information to gull the regulator into basically protecting them rather than constraining them.
Transparency of the information used to inform regulatory processes is one way to mitigate this capture risk. If all we're told is that 10 rats were tested, then it's pretty difficult to know whether the regulator has been captured.
My day job involves a fair bit of work on designing and operating regulatory systems. If the above is true then it's definitely the worst regime I've heard of. Please tell me it isn't so.