The Separation of Scientific Powers

Rep. Edward Markey (D-Mass.) has flagged an Office of Management and Budget (OMB) decision urging the Environmental Protection Agency (EPA) to use test data from companies and other existing data in judging whether chemicals found in pesticides and plastics could disrupt the human endocrine system, which releases hormones.[source]

I suspect that the OMB is actually suggesting that the EPA use all scientific data regardless of source. US regulatory agencies have long suffered from “not invented here” syndrome in that they rather routinely reject valid scientific data that originates from any source outside themselves. For example, they often refuse to use data generated by European regulatory agencies even though the standards of those agencies are just as good as ours.

However, I think he has a point (echoing Adam Smith) that data is suspect when it comes from companies that have an economic incentive to downplay the dangers of their products. However, Markey and others like him do not see that the EPA and other regulatory agencies have the same kind of built-in conflicts of interest as companies do.

Since the EPA’s budget and scope of powers depend on the degree of environmental hazard believed to exist, the EPA has a built-in institutional bias towards exaggerating dangers that is every bit as strong as the bias of profit-driven companies to underplay dangers. Worse, as a political creation, the EPA is subject to ideological pressure from politicians, lobbyists and activists across the political spectrum.

We should divide the EPA, and every other regulatory agency, into two departments that are entirely separated all the way up to and including the level of cabinet secretary. One department would be responsible for regulation while the other will be responsible for doing the science upon which the regulation depends. We might even go so far as to create a Department of Science which would perform scientific research for all regulatory agencies. Only then could we be assured that the science upon which we base our regulations would not be contaminated by institutional biases one way or the other.

As the scope and power of government have grown, we have neglected the concept of the division of powers, even though the Founders and history judge this concept absolutely vital to maintaining freedom. Today, each regulatory agency is its own little fiefdom and many are far larger than the entire Federal government was even 80 years ago. Yet we’ve never applied the Founders’ wisdom about the separation of powers to these vast and powerful organizations.

We need to start doing so.

Lies, Damn Lies and…

One of my academic advisors used to say that any argument without numbers is a religious one. And we all know how productive they are.

Being a numbers jock and P-Chemist, that statement resonated with me. It still does.

But then I went into business, and for a while my job involved the quantitative prediction of consumer behavior. Entering into the social sciences like that, where there is no ideological bias, just a financial incentive to get the model right, was good for me. It trained me to look at the instrument that was used to derive the numbers. To ask if the questioner was asking the right questions.

So my brain perked up when I saw this article on the decline of newspapers:

Big whoop. After several statistical triple back-flips, we now know that 96 percent of newspaper reading is done in the printed product. That’s like talking about modern transportation by pointing out that 96 percent of buggy drivers use buggy whips. Hello? We switched to cars 100 years ago.

Writing on the Nieman Journalism Lab Web site, Martin Langveld made some valid statistical conclusions about newspaper readership. The problem is that he was asking the wrong questions. It isn’t about newspapers; it’s about news.

Read more