Advertising freedom

Simon Wren-Lewis argues that tobacco plain-packaging mandates may be freedom-enhancing. He writes:

So we have individual freedom, interference, and crucially motivation. Advertising is usually portrayed as just providing information so that consumers can make informed choices. Sometimes it may do that. But advertising is often about suggesting associations, which provide no information at all. It is a mild form of brainwashing. Most of the time it is simply annoying.

For some, the information provided by some advertising might be useful. For most it is not. We can try and avoid advertising if we do not want its ‘information’, by turning the page, recording the programme and fast-forwarding through the adverts, averting our eyes, but this requires effort. Why should I have to make that effort? So for most people most of the time, it is advertising that mildly interferes with our freedom. (If I wanted to be clever, I could say that companies who advertise believe their product makes consumers better off, and therefore it is advertising that is paternalistic. However companies advertise to increase profits, not to increase consumer utility.)

So a government that prevents advertising can be seen as allowing individuals to make their own unencumbered choices. It is giving us a little more freedom and autonomy, rather than limiting it. The argument for advertising has to be that the benefits to the few in getting useful information outweighs the costs to the many in either avoiding it, or getting information they do not want. It is not paternalistic to ban advertising, just as it is not paternalistic to stop people being stalked.

Wren-Lewis is right about a few things, but I do rather disagree with him about the bigger picture here.

First off, we do indeed have plenty of models where advertising plays both informative and persuasive roles. And it is completely sane to model cigarettes as a Lancaterian good with a bundle of characteristics ranging from physical attributes like taste and burn intensity to more nebulous things like image and affiliation.

Brand affiliation for many goods is less about the physical merits of the product than it is about self-identification and self-expression, as Cass Sunstein nicely laid out a few years ago. Policies like plain packaging, that work to destroy the affiliative value of brand choice, could plausibly affect smoking rates. Similarly, a proponent of enhanced fuel-economy could insist that cars getting less than, say, 30 miles per gallon be subject to "plain car" regulations such that they had to be painted in beige, have horns that shouted "I hate the environment and probably am compensating for something" instead of beeping, and have some minimum designated asymmetry - for example, the left headlight would have to be twice the size of the right headlight. People would then be "nudged" into choosing the "right" car. On plain packaging, it's still not clear to me that the policy doesn't just push smokers to discount brands, but I suppose that the Australian data will tell us in a couple of years.

Tobacco branding and packaging is then akin to a Jaguar's sleek lines and beautiful rich colours. If Wren-Lewis wants to say that I'm brainwashed into thinking that an E-series is more beautiful than a Honda Insight, well, I'm not sure how I'd disprove that. I'm not sure he's offering a testable hypothesis there. If we stuck an Insight's body over a Jaguar's frame and engine and proved people liked it less, that's hardly evidence of Jaguar's brainwashing. We're doing more than just affecting irrelevant cosmetic or advertising details: the style is part of the product.

I disagree rather more substantially with the rest of his post. Wren-Lewis asks why the burden should be on him to avoid advertising in magazines by turning the page. The standard answer that any microeconomist would provide him is that the provision of advertising reduces the price of the magazine. Because most consumers prefer buying the bundle of "content plus advertising" at a lower price rather than content-only at a higher price, that's what's provided. At least for most publications. Switching to a "ban the advertising" equilibrium would have Wren-Lewis, and people like him, imposing costs on people like me, who are happier to have advertising and pay the lower price. How do we know which state of the world is better? Aggregate willingness to pay. Profit-seeking publishers weigh up how much they'll irritate Wren-Lewis and lose his custom against drops in profits where people like me are more sensitive to prices than to advertising. That publishers are profit-seeking rather than consumer-surplus-maximizers doesn't really wreck this result.

I'd thought this was pretty standard micro-theory. I don't get why Wren-Lewis gets to decide that the only permissable function of advertising is information-provision. In that world, we would discount to zero the utility that sports-fans get from seeing the affliative persuasive advertising for their preferred team - ads that focus on the team's logo, its chants, and perhaps some unrepresentative action shots.

Further, Wren-Lewis later argues that "the information content of the 'plain packages' [dominated by graphic ugly images] is clearly much greater than packages that were dominated by a logo. So this is one example where the information content of advertising does dominate any reduction in freedom that the advertising entails."

Suppose that smokers do find the packages dominated by big ugly warning pictures are less attractive. Is this evidence of the information content of the packaging having dominated? Hardly. Imagine that Australia had decided to run a different version of plain packaging in Tasmania. There, cigarette packages would have graphic images of slaughtered and injured animals instead of diseased lungs. There is no informational content in such images, but they are disturbing and work to break the affiliative brand utility. Like putting a Jaguar into an Insight's body.

For true paternalists, none of this matters: they know that less smoking is better for you even if you think you like it, so it doesn't matter what the mechanism is so long as it works. For those who at least still nod in the direction of revealed preference and of de gustibus, consider the uglification experiment I propose. Is it really more plausible that Oz's plain packaging regs work (if they do work) through an informational channel, or that they simply destroy part of the product's real value and thereby discourage consumption? And if you want to say that the value provided by beautiful packaging and brand affiliation is not real, think through the consequences of that position more consistently.

HT: Dick Puddlecote