The logic of the slippery slope

Are you sure you want to cross the Rubicon?

Reports that the British government is considering banning alcohol advertising yesterday led to much rejoicing from the, er, anti-smoking lobby.

The anti-smoking group Fresh shares an office with the anti-alcohol group Balance North East where they relieve taxpayers of £2 million a year. They have the same director and they officially merged to become Fresh Balance a few years ago, so it is not entirely surprising that they agree with each other.

Action on Smoking and Health (ASH), however, are entirely dedicated to the prohibition of tobacco (thanks to the Tobacco & Vapes Bill, we can finally say this without being accused of hyperbole). Alcohol is not within their remit and yet they are over the moon about the prospect of alcohol being treated like tobacco.

The thing about ASH is that they used to swear on a stack of Bibles that there was no slippery slope. In 2012, I warned that if the government set the precedent of plain packaging for tobacco, puritan activists were likely to call for it to be extended to other risky products. Sure enough, in the years that followed there were demands for plain packaging for ‘junk food’, soft drinks, alcohol and gambling products, and ASH themselves have called for plain packaging for vapes. But at the time, ASH were outraged by my suggestion and responded with an indignant press release saying…

…the “domino theory” i.e. that once a measure has been applied to tobacco it will be applied to other products is patently false.

To illustrate why it is “patently false”, they mentioned a specific example of a domino that was still standing…

The same argument was used against the ban on tobacco advertising, but 9 years after the tobacco ban in the UK, alcohol advertising is still permitted with no sign of it being prohibited.

Well, this is awkward, isn’t it? Not only is there every sign of alcohol advertising being prohibited, but ASH are actively supporting such a prohibition. And the ban on alcohol advertising will go alongside a ban on adverts for ‘less healthy’ food which even the most wary libertarian could have barely imagined in 2003.

So the domino theory is not patently false, is it? It is patently true. It is perfectly obvious that if you set a precedent it becomes more likely that similar things happen in the future than if you don’t set the precedent.

Why, exactly, does this happen so often in politics despite the solemn promise of all involved that it will not? It is, after all, not inevitable. The slippery slope is often described as a fallacy because A does not inevitably lead to B and C. Nevertheless there is a strong tendency for A to lead to C and in ‘public health’ it happens more often than not. What are the mechanisms and can it be prevented?

Mario Rizzo and Glen Whitman have done some fine work on this, such as their splendidly titled article ‘The camel’s nose is in the tent: rules, theories and slippery slopes’. In their masterpiece Escaping Paternalism, they list a number of reasons why regulation snowballs in ways that may or may not have been intended by those who supported the precedent-setting policy. The basic mechanism is as follows:

An effective means by which a proponent can argue for a new policy is to show that the accepted justification for an existing policy also provides the foundation for a new one. This is rational in most circumstances because policymakers, especially voters, lack the ability, time, energy, and motivation to analyze each policy proposal on its own merits. They will be rationally ignorant about most proposals. Therefore, if a new policy is seen as a small extension of an existing policy, decision-makers will tend to defer to the perceived rationales behind existing policies, or will at least not strongly resist new policies based on them.

The key word here is perceived rationales. As Rizzo and Whitman say, “the rationale for the old policy must be reconstructed”, but policymakers and voters may have a poor memory and bad actors might deliberately misrepresent the original rationale in order to justify a very different policy.

They see three ways in which this can happen.

1. Substituting broader for narrower justifications. For example, a behavioural nudge could be justified on the grounds that it will make an individual’s actions more closely align with their ‘true preferences’, but this is “easily replaced with the simpler and broader argument that government can help advance the target agents’ welfare according to some exogenous or ‘objective’ definition of the good.”

2. Paring multiple justifications down to one. Policies are often justified on the basis of several different reasons. In ‘public health’, it is common to hear a policy being justified on the basis of a combination of the following: that it will save lives, save taxpayers money, correct a market failure, reduce inequality, reduce health inequalities and boost the economy. When combined, these reasons may be persuasive, but taken in isolation they would not be seen as important enough to justify the costs of the policy. If the policy is misremembered as being sufficiently justified by only one of these reasons, it becomes possible to justify a much broader range of actions. The number of draconian policies that could be justified on the basis that it would save taxpayers money, for example, is vast.

3. Stripping justifications of their qualifications. A policy might be introduced on the condition that it does not apply in certain circumstances, with the understanding that it would be unacceptable if it were introduced without qualification. The qualification is later forgotten about, leaving only The Rule which has by that time become an unquestioned principle. Reasonable exemptions that were deliberately created are then portrayed as loopholes to be closed.

All of this might happen accidentally, but it often happens because “policymakers, experts, moralists, and rent-seekers interact to alter the terms of the public debate.” It happens all the time in ‘public health’ because most of the activists are mendacious charlatans who can never be appeased.

The smoking ban, for example, was sold to the public as a health and safety measure to protect people who work in hospitality from ‘secondhand smoke’, but it was subsequently portrayed as a way to get smokers to quit. To a large extent, the ‘success’ of the ban was measured by how many people (supposedly) stopped smoking as a result. This was a fundamental change of rationale. It turned a policy that was ostensibly about protecting third parties from harm into a policy that was overtly paternalistic. When the time came to campaign for smoking bans outdoors, the paternalism was made explicit and you had politicians openly admitting that the aim was to have “fewer places where you actually can smoke”.

There is an element of the bait and switch about the new alcohol advertising proposals too. Health officials are saying that they want to bring the regulation of alcohol adverts “closer in line with advertising of unhealthy food”, i.e. banned entirely on the internet and on TV before 9pm. But the food advertising ban was officially designed to combat childhood obesity. This framing was important because the government wasn’t yet ready to tell adults that they are too stupid to handle seeing ice cream commercials. Since children cannot buy alcohol, it made no sense to include alcohol in the regulations, but once the ban became law, the children were quickly forgotten about and the ‘public health’ lobby demanded a ‘level playing field’.¹

Rizzo and Whitman implicitly assume that policymakers are interested in a range of costs and benefits and will take all sorts of considerations into account. They assume that policies are not be introduced purely on paternalistic grounds - not officially, at least - and that paternalists have to rewrite the history of the original policy in order to extend their reach. Historically, that has generally been the case, but there is one simple way by which policy A leads inexorably to policy B and that is when Policy A, without any misrepresentation or distortion, really does logically lead to Policy B.

We are getting close to that point in the UK. Once you have policies being justified solely on the grounds that the government knows best and that the law should be used to protect people from themselves, you need no jiggery-pokery to extend the principle almost without limit. Aside from some economic gibberish about saving the NHS money, that is essentially the justification for the generational ban in the Tobacco & Vapes Bill. It is unadulterated paternalism. As already mentioned, outdoor smoking bans have been proposed on the same wholly paternalistic basis. Once “it’s for your own good” becomes an established principle of law, all bets are off.

The slippery slope is not inevitable but it is so common now that its risks should be counted among the costs of any precedent-setting legislation. Curiously enough, there is an analogy with smoking. The government is keen to stop young people trying their first cigarette, not because one cigarette is particularly bad for them but because it could lead to a lifetime’s addiction. A lifetime’s addiction is not inevitable (most smokers quit), but it is a real risk that may not be fully understood - or at least may not be sufficiently salient - to the individual at that moment.

The same applies to policymakers who, like the young person, are susceptible to present bias and hyperbolic discounting (but also, unfortunately, to action bias). Assuming that the policymaker is opposed to Policy B, he should factor in the increased risk of Policy B being introduced if he supports Policy A. It is not enough for him to say that he would never support Policy B, because he may not be in office by the time it is voted on. If he believes that introducing Policy A means there is a 50% chance of Policy B being introduced at some point in the future (and that is lowballing it in ‘public health’), he should count 50% of the costs of Policy B among the costs of Policy A. He may still think Policy A is worth it, but those costs should be baked into his decision.

This is why I say ‘no’ when people ask me whether I agree with laws forcing people to wear seat-belts (libertarians get asked this a lot). Seat-belt laws were quite explicitly paternalistic and were never justified on any other grounds.² Paternalists love talking about them because wearing a seat-belt is eminently sensible and the restriction on liberty from seat-belt mandates is very minor. Few people disagree with such laws these days and you risk looking like a kook if you oppose them.

But the main reason they love talking about them is because they set an overtly paternalistic precedent³ which they have been citing ever since as a justification for further draconian regulation of self-regarding behaviour. Although difficult to quantify, I believe that they have been successful in this and that the trivial loss of liberty that resulted directly from the seat-belt law has been dwarfed by loss of liberty that has resulted indirectly from setting that precedent.⁴

One of the few MPs who had to foresight to factor in the future risks of the slippery slope was Ivan Lawrence who said during the parliamentary debate in 1979:

Why should anyone be forced by criminal sanction not to hurt himself? That was never, at least until the crash helmet legislation, a principle of our criminal law. Where will it end? Why make driving without a seat belt a crime because it could save a thousand lives, when we could stop cigarette smoking by the criminal law and save 20,000 lives a year? Why not stop by making it criminal the drinking of alcohol, which would save hundreds of thousands of lives?

… In the end we shall find that our liberties have all but disappeared. It might be possible to save more lives in Britain by this measure—and by countless other measures. But I do not see the virtue in saving more lives by legislation which will produce in the end a Britain where nobody wants to live.

The domino effect is patently real. Resist it while you can.

Update

This is ASH’s latest blog post. Drinkers should be nervous. I know I am.

1. In both these examples it was the original framing that was dishonest. The smoking ban was always primarily about making it harder for people to smoke and childhood obesity is always used as a cover to control what adults eat. When I was on the radio with former public health minister Steve Brine a few years ago, he openly admitted than whenever he talked about childhood obesity he was really thinking about adults.

2. Some nerd will probably point out that seat-belt mandates in the back seats could prevent a passenger flying forward and killing someone in the front seats, so I will confine my comments here to seat-belts in the front.

3. Alongside motorcycle helmet laws.

4. This doesn’t mean I would necessarily repeal the law. There is no point. The damage has been done.