Reusable Grocery Bags Can Kill (Unless Washed)


 

One recent local environmental cause, especially popular in California, has been to ban or tax plastic grocery bags. The expressed hope is that shoppers will instead carry reusable grocery bags back and forth to the grocery store, and that plastic bags will be less likely to end up in landfills, or blowing across hillsides, or floating in water. The problem is that almost no one ever washes their reusable grocery bags. Reusuable grocery bags often carry raw meat, unseparated from other foods, and are often stored for convenience in the trunk of cars that sit outside in the sun. In short, reusuable grocery bags can be a friendly breeding environment for E. coli bacteria, which can cause severe illness and even death.
Jonathan Klick and Joshua D. Wright tell this story in “Grocery Bag Bans and Foodborne Illness,” published as a research paper by the Institute for Law and Economics at the University of Pennsylvania Law School. As their primary example, they look at E. coli infections in the San Francisco County after it adopted an ordinance severely limiting the use of plastic bags by grocery stores.
As one piece of evidence, here’s a figure showing the number of emergency room visits in San Francisco County related to E. coli for the 10 quarters before and after the enactment of the ordinance, where zero on the horizontal axis is the date the ordinance went into effect. (The shaded area around the line is a 95% statistical confidence interval.)  Clearly, there is a discontinuous jump in the number of emergency room visits.

For comparison, here’s the same measure of emergency room visits related to E. coli infections for the other counties in the San Francisco Bay Area in the 10 quarters before and after the ordinance was enacted. These counties don’t see a jump.

Klick and Wright flesh out this finding in a variety of ways. They look at other measures of foodborne illness, like salmonella and campylobacter, and also find a rise associated with the ordinance limiting plastic bags. They look at other cities in California that have enacted such bans, and although it is harder for them to track health effects for individual cities–because the health statistics are collected at the county level–they find negative health effects as well. (The city of San Francisco is consolidated with the county of San Francisco.)
Overall, they find that San Francisco typically experiences about 12 deaths per year from intestinal infections, and that the restrictions on plastic bags probably let to another 5-6 deaths per year in that city–plus, of course, the personal and social costs of some dozens of additional hospitalizations. With these costs taken into account, restrictions on plastic bags stop looking like a good idea.
Of course, a possible response to this problem is to launch a public information campaign to encourage people to wash their reusable grocery bags. But that response then leads to two other issues.  First, if public information campaigns can be effective on the grocery bag issue, the campaign could simply focus on the need to dispose of plastic bags properly and recycle them where possible–without a need to ban them. The argument for an ordinance sharply limiting the use of plastic grocery bags is, in effect, based on an implicit assumption that public information campaigns about grocery bags don’t work well.  Second, if reusable grocery bags are washed after each use, then any cost-benefit analysis of their use would have to take into account the costs of water and detergent use. I have no idea if these costs alone would outweigh the benefits of reusable grocery bags, but it might be a near thing.
Like most economists, I have a mental file drawer of “good intentions aren’t enough” stories. The push to ban plastic grocery bags is one more example.

Reusable Grocery Bags Can Kill (Unless Washed)
Timothy Taylor
Thu, 07 Feb 2013 12:00:00 GMT

Advertisements

2 responses to “Reusable Grocery Bags Can Kill (Unless Washed)

  1. Almost all of the deaths counted in this study were from C. Difficile, a hospital or institutional acquired infection often the result of antibiotics killing beneficial intestinal bacteria. A careful analysis of the CDC data shows only1 E. Coli death during the time period of the study. Although the CDC public queries allow one to see only counts of 10 or greater, one can count the number of deaths caused by some condition “X” and also count the number of conditions caused by “X+Y” and take the difference. If “Y” is E. Coli, you find the difference is only 1. If “Y” is C. Difficile you will see many more deaths.

    See also http://blogs.berkeley.edu/2013/02/13/san-francisco-plastic-bag-ban-deaths/ for a more complete explanation of the other errors in the study, including the emergency room part.

    This is a good lesson in how what appears to be a “significant” statistical result can just be an anomaly.

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s