A quick summary of the paper's findings is given by:
- Politicians and lobbyists who promote new regulations and taxes typically claim to have science on their side. Scientific evidence shows that the actions they wish to discourage are harmful and that government intervention would reduce this harm. Yet much ‘evidence-based policy’ is grounded on poor scientific reasoning and even worse economics.
- Recent examples from the U.K. of flawed evidence-based policy include the proposal to introduce a minimum alcohol price, the ban on smoking in enclosed public spaces, measures to reduce greenhouse gas emissions and attempts to increase gross national happiness.
- A frequent error is to ignore the costs resulting from the policy. For example, minimum alcohol price plans do not consider the welfare losses associated with reduced consumption among recreational drinkers. The benefits of alcohol consumption, and hence the cost of reducing it, are simply ignored in the analysis.
- Evidence-based policy typically also fails to account for substitution effects, such as the way a minimum alcohol price would encourage consumers to purchase drinks in the shadow economy or adopt intoxicating alternatives to alcohol.
- The external costs of harmful activities are central to the arguments for state intervention but often cannot be calculated with any certainty. To estimate the external cost of carbon emissions, for example, we would need to know the subjective preferences of people around the world, and somehow weigh them against each other. We would also need to make assumptions about the preferences of people living many decades in the future.
- The predictions of theories that have not been tested, and are not entailed by well-known facts, do not warrant high levels of certainty. Those who insist on this are not ‘anti-science’, as they are often claimed to be. On the contrary, it is those who are willing to be convinced in the absence of predictive success who display an unscientific cast of mind.
- High levels of scientific doubt are often concealed as a result of ‘noble-cause corruption’. Scientists may exaggerate levels of confidence in their findings if it promotes actions they happen to support. This problem is particularly acute in fields that have long been policy battlegrounds, such as climate, health and education. Many scientists entered such fields because they were already committed to a particular policy agenda.
- Scientists are also interested parties. They stand to gain from policy taking one direction rather than another and will be tempted to support the personally profitable policy direction. Public policy can create demand for their skills and hence drive up the rewards accruing to them. Scientists are natural supporters of policies that draw on their expertise and thus inclined to overstate the credibility and importance of their ideas.
- Expert practitioners in one field may be quite ignorant of other fields, knowing little about either their theory or methods. ‘Expertise slippage’ is the tendency to defer to experts on matters which fall outside their area of expertise. Climate scientists, for example, are experts on hardly any of the issues that determine which climate polices are best. They have no special knowledge of how businesses will respond to taxes or the relative welfare costs of reduced growth.
- Paternalist policies promoted by experts and politicians show contempt for the actual preferences of the general public. People are forced to live according to values that they reject. For example, supporters of ‘happiness policy’ believe the state should coerce people to act against their preferences in ways that policymakers think will increase their wellbeing.
In any policy making both costs and benefits have to be taken into account, that is we have to consider both the winners and the losers from a change in policy. Any policy analysis which considers only the bad is not worth the paper it is written on. In Whyte's view the failure to acknowledge trade-offs and competing preferences is endemic in much of the so-called evidence-based policy. By looking at only the "bad side" of a situation you run the very real risk of causing more harm than good. Why not, for example, have a $500 minimum price of alcohol or a 5 mph speed limit. I'm sure that the evidence would show that both policies would reduce harm and thus would pass as "evidence-based policy". But it comes at the cost of denying people the pleases of drinking and the benefits of driving. Would it be a trade-off worth making?