Analysis Run Amok

By Michael Gough and Sanford L. Weiner
This article originally appeared in the Chicago Tribune.

The new theme of the Food and Drug Administration is regulatory flexibility. Yet one of its latest regulations can only be described as analysis run amok. The agency has banned imported wine in bottles that have lead foil capsules (the brightly colored foil at the top of the bottle). Why? To protect young, wine-drinking children from ingesting lead.

Can this really be a problem? Even the FDA concedes that the minuscule amount of lead involved is no threat to adults. Furthermore, since the FDA first proposed the ban, all the domestic wine producers have phased out the lead foil, as have those in the European Union. Less than 5 percent of the wine still on the U.S. market is involved.

In 1992 the agency said that the potential exposure was so small that the benefits of a ban would be “impossible to quantify.” But now, through the magic of cost-benefit analysis, the FDA has done precisely that. Its analysts estimate that the remaining producers can switch to other coverings for $90,000 per year. That’s the cost. But it takes a remarkable chain of logic to reveal the benefits.

First, some lead is left behind on the bottle when the foil is removed, and it gets into the wine as it is poured. Okay, so far. However, research findings that remain controversial except in federal agencies suggest that very low lead levels can decrease children’s IQs. The FDA therefore estimates that 10 percent of all children aged 3 to 6 have “elevated” lead levels (even though exposure has dropped to a 50-year low).

But how many of those 1.4 million young children drink wine? Agriculture Department surveys report that 0.1 percent drink at least once glass every three days, averaging two ounces per day. The FDA further assumes that children drink imported wine in the same proportion as adults. Thus the number of children with elevated blood levels of lead, who also enjoy, say, Australian chardonnay, works out to 84. Exactly 84 children in the entire United States.

If the survey numbers are not just the statistical fluke they appear to be, the FDA may have uncovered a small but real problem. Consider one of those children who twice a week gets a buzz from drinking six ounces of wine: is the child’s ingestion of lead the thing that we should worry about, or might there be other risk factors in that family setting? Clearly alcohol is the issue here, not lead.

But the FDA analysis cannot deviate from its intended path. Using estimates from the Centers for Disease Control, the FDA concludes that the 84 children will ingest enough lead to lower their IQs by one-fourth of a point, and that deficit will reduce their lifetime earnings, accounting for inflation, by $512 each. Multiplying $512 by 84 yields an annual benefit of $43,000. Using similar methods, the FDA estimates an additional $16,000 annual benefit from reducing the lead exposure of fetuses carried by wine-drinking women who have elevated lead levels. The annual benefits of $16,000 plus $43,000 are less than the cost.

One more intellectual leap, though, carries the day. The FDA shifts its focus from the few children with elevated lead levels to all the 3-to-6 year olds and all the pregnant women who drink imported wine. Although those children, born or unborn, have lead levels well below the government’s safety thresholds, the FDA assumes that they will all be harmed by ingested lead. In other words, anything above zero exposure is assumed to cause measurable lifetime damage, despite the lack of evidence. In the FDA’s final analysis, eliminating that risk brings the benefits up to $6.2 million per year — finally higher than the costs.

The calculated figure should be $2.1 million. In its lengthy review the FDA never noticed that two-thirds of the expected benefits stemmed from mistakes in arithmetic.

The FDA’s breathtaking piling of assumption upon assumption is matched only by the spurious precision of the estimates employed. The FDA analysts presume to know the drinking habits of small children, down to their choice of imports, and the impact of minute amounts of lead on their IQs and of IQ on their lifetime incomes. When all that failed to generate the desired positive number, the assumption was made that children never thought to be at risk from lead impairment would nevertheless benefit from even less lead exposure.

This is regulatory flexibility.

Michael Gough is director of science and risk studies at the Cato Institute and Sanford L. Weiner is a professor at the Center for International Studies, Massachusetts Institute of Technology.