More ‘Secret Science’ at EPA

More ‘Secret Science’ at EPA
July 1, 1999



The U.S. Environmental Protection Agency (EPA) is once again using “secret science” to push for more regulation.

On May 1, EPA proposed that sports utility vehicles (SUVs) be required to meet more stringent tailpipe emissions standards, and that gasoline contain less sulfur. The agency claims the rules will prevent as many as 2,400 premature deaths and 8,000 cases of bronchitis per year, and reduce the number of hospital visits, lost work days, and “an assortment of respiratory ailments.”

EPA claims the rules would produce between $200 million and $16 billion in net annual benefits. But up to $14.3 billion of those projected benefits are based on the agency’s assertion that the rules will save 2,400 lives per year: Each life is valued at $5.9 million. No one’s been able to verify the statistical analysis underlying EPA's claims.

The agency’s estimate of lives saved is based on a single, controversial statistical analysis. Many suspect the analysis (known as the “Pope study”) is junk science. The study’s data may be of such poor quality that the statistical analysis is akin to “garbage in, garbage out.” No one knows for sure . . . because EPA has so far blocked all attempts--even by Congress--to have the study independently examined.

This is not the first time EPA has tried to shield the Pope study from scrutiny. In 1996, the agency claimed that tiny particles released into the air from industrial processes, driving, and even fireplaces, caused 15,000 premature deaths. The number of deaths was calculated based on the Pope study.

Faced with regulatory costs potentially exceeding $100 billion a year, industry, local governments, and Congress asked to see the data underlying the Pope study and EPA’s projections for particulate matter. The agency balked, saying there was no use in releasing the data since the study was published in peer-reviewed journals--akin to assuming something is true because it's in writing.

Independent verification of results--including after publication-- is essential if scientific studies are to form the foundation of sound law and policy. In June 1996, Tulane University researchers claimed combinations of PCBs and pesticides were potent disrupters of hormonal systems. Their study drove a panicky Congress to require EPA to develop a multi-billion dollar program for testing chemicals. It later turned out the study could not be replicated by independent laboratories; after a year, the authors retracted the study. But the law remained, and a monster regulatory program lives on.

When EPA finally requested the air pollution data, it was the Harvard University-based researchers' turn to stonewall--even though they had been paid with taxpayer money. They claimed releasing the data would infringe on the privacy of study subjects. This is a red herring argument, however; other researchers have released their data without compromising anyone’s privacy. The U.S. Air Force's study of exposures to Agent Orange during the Vietnam War, for example, won't be complete until 2002. But the data--minus information that would allow individual subjects to be identified--are already available to the public. To its credit, the Air Force released its data before anyone asked for it. Imagine the uproar from veterans groups and Congress had the Air Force or its contractors refused to release the data.

The second reason offered by the Harvard researchers amounted to little more than character assassination. One of the investigators dismissed the requesters--who included, among others, industry scientists--as "industry thugs." Did that make House Commerce Committee Chairman Thomas Bliley (R-Virginia), who also requested the data, a "congressional thug?"

The Harvard researchers ultimately agreed to turn over their data to the Health Effects Institute, a private organization jointly funded by EPA and the auto industry. The institute isn't permitted to release the data to outside parties. Not only does this hardly constitute release to the public, which paid for the studies, but the results of the HEI review won't be available for years--well after the proverbial horse has left the barn.

“Withholding scientific data paid for by taxpayers--especially in the context of public policy--is outrageous,” noted Michael Gough, a former government scientist and now an adjunct scholar with the Cato Institute. “The scientific method requires data be available for review to see if the conclusions are valid. Without validation, the results cannot be considered scientific.”

Fortunately, a law signed by President Clinton in 1998 requires that federally funded data used to support federal policy be made available to the public through the Freedom of Information Act.

EPA’s SUV proposal looks to be a good test case for the new data access law.