It's not new, but I just recently stumbled on this very insightful lecture, examining some of the fundamental differences between how economists and non-economists think about government regulation: http://www.aei-brookings.org/admin/authorpdfs/page.php?id=1144

In case you haven't heard of the speaker:

Sam Peltzman is one of the few economists, and probably the only regulatory economist, to have an effect named after him -- the "Peltzman effect." The Peltzman effect arises when people adjust their behavior to a regulation in ways that counteract the intended effect of the regulation. So, for example, when the government passes a seatbelt law, some drivers may respond by driving less safely.

Say what? Don't worry, Peltzman has the facts to back it up. His research found that after government-mandated introduction of numerous car safety features, "occupant deaths per accident did indeed fall substantially compared to what might otherwise have been expected. But this was entirely offset by a combination of more accidents and more fatalities to nonoccupants such as pedestrians, bicyclists, and motorcyclists, whom the devices did not protect."

Still skeptical? You aren't alone. As Peltzman explains, "This finding met with considerable skepticism from economists and noneconomists. Unlike most noneconomists, most of my colleagues accepted the underlying logic and reserved their skepticism for whether the facts fit the logic." Further research, however, was kind to his hypothesis. For example:

A recent article by Alma Cohen and Liran Einav (2003) on the effects of mandatory seatbelt use laws is reasonably representative. This particular study tdoes not find increased nonoccupant deaths to be an important part of the story. But it shares with most such studies the crucial bottom line: The real-world effect of these laws on highway mortality is substantially less than it should be if there was no offsetting behavior. They conclude that the increased belt usage occasioned by these laws should, in the absence of any behavioral response, have saved more than three times as many lives as were in fact saved.

Peltzman goes on to give unrelated examples -- the Americans with Disabilities Act may actually harm the ability of the disabled to find jobs, while the Endangered Species Act may actually promote more rapid destruction of endangered animals' habitats. He then insightfully notes:

My examples of offsetting behavior drip with irony: The highway safety act promotes reckless driving; the disabilities act disemploys the disabled; the endangered species act promotes clear cutting and suburban sprawl. This ironic aspect, I think, accounts for some of hte appeal of the concept to economists as well as some of the controversy about it. We economists tend to prefer unexpected results to expected, especially if they illustrate the power of the simple logic of our discipline. But we are also skeptical of unexpected new results, and the rest of the polity has resisted them at least since Adam Smith was pushing free trade more than two hundred years ago.

And you should be skeptical. Here I am telling you how offsetting behavior has compromised three significant attempts to improve our lives. Yet each of them is a sacred political cow. Not one is in any serious danger of substantial alteration. How can that be? [...] It has become routine, especially in this area of social regulation, for economists to find regulation with compromised benefits or with negative net benefits. There must be a very powerful force that protects such regulation.

He argues that in a growing economy, most aspects of our lives gradually improve over time, with or without regulation. We see regulations being passed, and we see our lives improving, so what's the problem? Simple: we're better off than we used to be, sure, but with smarter government policy we could have been better off still. "The economic analysis begins by posing a counterfactual: What would the world have been like without the regulation? Then the analyst compares the actual world to the counterfactual. Regulation is deemed successful only if its intended effects are realized to a greater degree than could reasonably been expected according to the counterfactual benchmark." On the other hand, non-economists don't pay attention to such counterfactual scenarios -- "the political process seems to require evidence of malfunction in some absolute sense before there is serious pressure for change."

He gives examples from his personal history, including his work on freight regulation (trucking, rail, and air freight). Economists widely recognized in the 70s that the regulations currently in place were completely ridiculous -- the Interstate Commerce Commission had endless rules on what truck drivers could and couldn't do. Not just safety regulations -- these regulations were truly ridiculous, serving no purpose but to stifle competition on behalf of the incumbent monopolists: (http://www.econlib.org/library/Enc/TruckingDeregulation.html )

From 1940 to 1980, new or expanded authority to transport goods was almost impossible to secure unless no one opposed an application. Even if the proposed service was not being offered by existing carriers, the ICC held that a certificated trucker who expressed a desire to carry the goods should be given the opportunity to do so; the new applicant was denied. The effect was to stifle competition from new carriers.

Purchasing the rights of an existing trucker became the only practical approach to entering a particular market. By the seventies the authority to carry certain goods on certain routes was selling for hundreds of thousands of dollars. Because the commission disapproved of "trafficking" in rights, it was hostile to mergers and purchases and attempted to restrict authority as much as possible. The result was often bizarre. For example, a motor carrier with authority to travel from Cleveland to Buffalo that purchased another carrier or the carrier's rights to go from Buffalo to Pittsburgh was required to carry goods destined for Pittsburgh through Buffalo, even though the direct route was considerably shorter. In some cases carriers had to go hundreds of miles out of their way, adding many hours or even days to the transport.

Even so, it took the complete collapse of the railroad industry to prod government into action to end this absurdity.

A more controversial example Peltzman gives is that of the FDA. His research has found that "the proof-of-efficacy requirement [for new drugs] was a public health disaster, promoting much more sickness and death than it prevented. Nothing I have seen since has moved me to change that conclusion -- the disaster is ongoing. But there is no realistic chance that this regulation will share the well-deserved fate of the Interstate Commerce Commission anytime soon."

The argument is simple: the efficacy requirement delays the introduction of new drugs, costing many times more lives than might be wasted by the introduction of ineffective drugs.

In some cases, it is a cost well spent. Some ineffective drugs are screened out, and the extra testing catches some that are dangerous as well. But every effective drug that ultimately makes it to market also incurs the time cost [...] the extra time means that some potential beneficiaries [...] will die or suffer while the FDA sifts the test results. My reading of the available evidence was simply that this latter cost far outweighs any benefit. [...] I found that the unregulated market was very quickly weeding out ineffective drugs prior to 1962. Their sales declined rapidly within a few months of introduction, and there was thus little room for the regulation to improve on market forces.

So why do we put up with such a foolish policy? Again, simple:

Medical progress of all kinds continues. Beneficial new drugs, however delayed, ultimately do make it to market. Mortality rates decline on the order of 1 percent per year [...] And, perhaps more important, the deaths of which I speak are counterfactual deaths, not deaths that can be directly connected to any regulatory malfeasance. [...] the actual victims of the regulation did not swallow a bad FDA-approved pill. They merely failed to swallow a good one in time and never knew what they had missed. As long as medical progress continues, a case for reform built on anonymous, unknown victims of regulation is going to be difficult politically. Indeed, the defenders of the status quo make good use of the progress. They will tell you how much regulation has contributed to the progress and what dire consequences would follow from any meaningful regulatory change.

Very interesting stuff, and well reasoned. Hope it helps elucidate how we economics geeks think about the world...