Musings on the Food Safety Modernization Act

We all want safe food.  Right?  It's hard to argue with the intentions behind the new Food Safety Modernization Act (FSMA).  Indeed, when I've asked people on surveys, they almost always say that food safety is one of the most important food issues to them and that they support stricter government rules and government spending on food safety.  

I haven't weighed in much on the FSMA because before now I haven't taken the time to really dig into the meat of the rules.  However, the FDA recently released drafts of the proposed rules, so now is probably as good a time as any to start digging in.

My first thought after reading through the text of the law and the proposed rules is how overwhelming it all is.  Its like trying to read the IRS documents on how to fill out your taxes but even worse.  In this regard, its not any different than many other federal rules, but I suspect most people who say they want more food safety on surveys have never taken the time to read through these things (here is the text of the law; proposed rules are here; or see the 547 page document here).  I agree with a lot of the spirit of what the FDA is trying to do here (in short, firms have to create a "plan" for managing food safety that it is risk based), but if I put myself in the shoes of a food company owner, and tried to read these documents, I think I'd faint.  

Just to given one simple example, one of the proposed rules begins by indicating who the rules apply to, and answering this question is no easy task (we haven't even gotten to a description of what the rules are).  Here is but one exception to the rule: 

The proposed rule would provide a qualified exemption and modified requirements for farms that meet two requirements: (1) the farm must have food sales averaging less than $500,000 per year during the last three years; and (2) the farm’s sales to qualified end users must exceed sales to others. A qualified end-user is either (a) the consumer of the food or (b) a restaurant or retail food establishment that is located in the same State as the farm or not more than 275 miles away. Instead, these farms would be required to include their name and complete business address either on the label of the produce that would otherwise be covered . . . or at the point of purchase.  This exemption may be withdrawn in the event of an active investigation of an outbreak that is directly linked to the farm, or if it is necessary to protect the public health and prevent or mitigate an outbreak based on conduct or conditions on the farm that are material to the safety of the produce.

In other words, if you sell at farmers' markets or to nearby restaurants, you don't have to comply (because apparently you are incapable of selling unsafe food??).  Oh, but we want your name on the produce anyway.  And, if we think it's necessary, we will make you comply anyway, so basically you're not exempt.

At the end of the day, the question is: will the costs of the rule generate more benefits than costs.  And, while I appreciate the FDA's work in trying to answer this question (I've done similar calculations on other topics myself), we have to say: we really don't have a clue.  The agency conducted a regulatory impact analysis (see document 194 here).  In their own words, they:

lack sufficient information to fully estimate the proposed rule’s likely benefits

and they later say that:

. . . quantification of the human health benefits derived from this rule is difficult and complex . . .

One thing that struck me is the actual hard data on how many people are affected by the food safety outbreaks the FSMA is supposed to prevent.  Here is my re-working of one of their tables:


So, we have a law that will create roughly $1 billion annually (for the first few years) in costs.  And the actual hard evidence we have is that there were 37.5 annual hospitalizations and 1.7 annual deaths from the foods covered by the FSMA.  That's about $27 million in cost for every confirmed hospitalization or $588 million for every confirmed death.  

The authors rightly point out that these illness/death numbers are probably under-estimates of the true number of cases.  But, the extrapolations are - to put it bluntly - wild guesses based on the evidence of a single study.  They do ask:

We seek comment on our assumption that the share of illnesses caused by unknown pathogens that are attributable to food covered by this rulemaking is equal to the share of illnesses caused by known pathogens that are attributable to food covered by this rulemaking.

But I wonder if this is even knowable.  Even more difficult to ascertain is how implementation of FSMA will change the number of annual illnesses.  The impact report does the best it can to cobble together the available evidence, but it is simply sparse, and it often amounts to comparing companies who have already adopted plans like those mandated by FSMA to those who haven't.  But, surely selection is playing a big role here (i.e., the firms adopting food safety plans were already the kinds of companies that were going to have higher food safety than companies who weren't) and comparing these two types of firms does not give a causal impact of FSMA.    

As I said at the onset, the public seems to favor more food safety regulation.  But it sure is messy seeing the sausage being made.  And, I suspect fewer people would want the sausage if they took the time to see its proverbial production process.