The ever-reliable Jamie Whyte has a superb piece in The Times in which he identifies quite precisely what’s wrong with ‘the precautionary principle’:
Suppose that, in return for an annual premium of £1, someone promises to pay you £1 million if you are abducted by aliens (such insurance exists). … You lack the information required to know if the insurance is a good deal. It is in such situations of uncertainty that the precautionary principle is supposed to apply. … [T]his principle tells you to buy the ticket. You should incur the £1 cost of the premium if there is any chance that it will save you from the greater cost of experiencing an uncompensated alien abduction. Whenever the prize is greater than the bet, and you do not know the odds, the principle says you should gamble. Bookmakers must dream of the day when punters bring such wisdom to the racetrack.
That’s a very illuminating parallel. What those who preach precaution are doing is secretly evaluating the likelihood of the Very Bad Thing we are supposed to be scared of as certainty, and their avoidance policy as perfect.
I would add, now Whyte has given me the right analytical start, that the way that the problem is usually posed should give this away directly. The precaution preacher says that: the Very Bad Thing (B) may be unlikely, but it is so very very bad, that however unlikely it is, it is too horrible to contemplate not doing onerous things P prevent it. It might as well be certain, but for P. That is implicitly a claim that both B is infinite in horribleness and that P is guaranteed to reduce its (unknown) likelihood.
Not only is it a bad bet, but the claim to the efficacy of P should be treated with skepticism. As well to remember that when dealing with Greens, securocrats and panic-mongers of all kinds.
That’s a perverted interpretation of a precautionary principle. A true precautionary principle would be that you don’t change what you’re doing unless you verify that the outcome will not be detrimental.
As such, the while global warming may be a concern, there’s no reason to change our activities unless it can be proven firstly that global warming is occuring, secondly that mankind is causing that change and thirdly that such change is actually detrimental.
A problem that arises is that do you regard economic growth as change or a constant? How about exponential economic growth?
Personally, I think we should carry on but keep an eye on things. These things have a way of correcting themselves anyway. Hopefully though that won’t be through something dramatic like rapid 25% population extinction or somesuch.
Rich
And fourth: that the proposed remedy will indeed prevent the detrimental outcome.
Fifth: That the proposed remedies will not be more detrimental than than the original perceived problem.
The precautionary principle says that a course of action should be rejected unless its consequences are well understood.
Accepting the precautionary principle is a course of action that may have consequences which are not well understood.
Therefore, the precautionary principle should be rejected.
Some background on the precautionary principle can be found here in Wikepedia. There is a usefully severe criticism of the currently prevalent (‘new’) interpretation under the title Beware the Precautionary Principle (which I have posted before on Samizdata).
For Erich (as an alternative):
From this arises the need to do everything material (or refrain from everything material) that you know nothing about.
Guy is, of course, correct that Jamie Whyte’s article is superb. This is particularly so in its separation of risk from uncertainty of knowledge (about the probabilities and good/bad effects of possible outcomes).
On this, and usefully positively in the mathematics of risk analysis, there are means of handling unknown costs/benefits and probabilities by subjecting those, themselves, to characterisation by Probability Density Functions (PDFs). This allows a further refinement of decision making; that is except for those who believe that so many things have infinite cost or benefit. This extension is usually referred to as maximum a posteriori (MAP), and is an extension to Bayesian statistical theory and Bayesian Risk Analysis.
Of course, most people will run a mile on seeing a few equations. But that just shows that most people should not be involved in making difficult decisions in fields where their experience is limited, their knowledge weak, and the extent of their ignorance a mystery to them.
Finally, a couple of examples, from history, of the ‘new’ precautionary principle in action.
Best regards
Isn’t the argument of the precautionary principle basically a nontheological Pascal’s Wager? It sure looks similar to it.
Nigel,
But that just shows that most people should not be involved in making difficult decisions in fields where their experience is limited, their knowledge weak, and the extent of their ignorance a mystery to them.
A dangerous statement, having the potential to put us all in the hands of assorted technocrats and self-appointed experts exercising precaution on our behalf.
I am all for attention to good analysis in making decisions. And I am prepared to concede there are certain bounded areas of knowledge that are well understood enough to be left to the experts with adequate understanding. But in matters of public policy there are no experts with certain knowledge, though accrued experience and self-justification tends to be held forth as such. The best thing experts can provide in public policy is rigorous counterexample.
How do you claim if you have been abducted?
dearieme,
It must be possible to buy insurance against the possibility that no-one believes you in the event that you have been abducted. if not, I smell a niche…
Guy writes, on my view:
I’m happy to take this as not totally unjustified, but …
Firstly, in so far as I implied anything in favour of technology, it was to use mathematics to understand better the extent of one’s ignorance, and how this affects the relevance of one’s thought processes and conclusions.
Concerning whether such use of numbers (eg in statistical measurement) is technology does, itself, raise interesting questions. I rather doubt that many people would claim the rules of logical deduction should be viewed as technology; attaching probabilities goes only into a bit more detail.
Those thought processes do not have to be about technology, but most thought processes concerning government and other group action should have a rational and evidential basis. Where the evidence is thin, we are ignorant; where the rationality is thin, we are stupid. To believe that action is justified beyond the evidence supporting it (in preference to possible alternatives) is, itself, irrational.
I am a technologist (I hope Guy does not think I am truly a technocrat) and value very highly its contribution to our wellbeing (higher than just about everything else except the golden rule, and perhaps the parts of legal and political systems that support it). I agree with Guy that, in the decision on whether and how to use any particular technology, wider aspects should be considered than the technology alone. I see technology as but a tool that can be used for good or bad, right or wrong. It’s how it is used that is important, not its existence or the possibility of its misuse.
It is, at least in my view, a sad fact that UK government is increasingly bereft of those who understand the meaning, importance and utility of numbers. Nevertheless, they quote statistics (often meaningless or wrong) left right and centre in support of their ‘solutions’. This does little for either the efficiency or the overall effectiveness of government.
Secondly, I am instinctively suspicious of anyone who appoints themselves as expert, unless I can see some basis (evidential and rational) for their claim. This is what first put me on to the IPCC and AGW – claims that were immediately and obviously not self-consistent and, on further investigation (painfully time-consuming), not supported by the available evidence (any better than alternative hypotheses).
If, on (untrollish) request for explanation of his/her views or argument against competing hypotheses, an expert is unable to offer useful information that is, step-wise, understandable, even to those new to the field, then he/she is no expert in anything but, perhaps, rhetoric.
Lastly, the ‘true’ precautionary principle has been hijacked and inverted by those busy busy people who search (often desperately in my view) for support for their prejudices or power-crazed aspirations. It used to be something good (like prudence) and has been turned around completely.
There is, I strongly suspect, a relationship with the hijacking of the precautionary principle and the increasing reluctance to make political and societal decision-making sufficiently grounded in numerate thought. It is formed from the ascendency of indiscriminate political advertising over political judgement (which favours established politicians and parties who have greater access to and control of the media), the dumbing down of our educational system (which downplays knowledge and sceptical thought in the demos, and their politicians), and the increasing centralisation of decision-making on issues that should be localised or outside of any government control (which renders much less effective the better routes, ie those less violent, to political and social change).
It’s worth remembering that it took well over half a century for the economic and other failings of Soviet communism to become widely accepted by those on the inside. And the response of the West to this is to go even further in centralisation and forced wealth redistribution. Interesting.
Best regards
Nigel,
A very thoughtful post.
Political decisions are not made on basis of “numerate thouhgt”, unless the numbers involved are public opinion polls -).
Even when technically savy aides forward a sensible plan, the decision making politician has additional factors to consider: is it popular?, can it be passed in parliament?, will it please my backers?, will it enhance my career?, will it improve my, and the party’s chances in the next election ? etc.
“Numerate thought” is scarce in the population in general, and in particular among opinion makers – the media, the intellectuals and the politicians.
CFM almost brings up my point, but fortunately for me, doesn’t.
And that point is, that even if P is a perfect remedy for B, do we know if P’s opportunity cost will be less than the harm of B, or some other remedy P’ such that P’ + the difference in opportunity costs between P and P’ – the difference in mitigation of B, results in greater benefit?
…and the proponents of P should be scourged.
Actually, offering insurance against alien abduction isn’t necessarily bad — I can think of far more ridiculous scenarios.
Where this becomes iniquitous is when Parliament (or Congress) passes the Alien Abduction Protection Act, making said insurance mandatory.
One must be cautious in applying the precautionary principle.