= 27d567bc2d48d9f40e9ccc20ea370b15)
Salemicus comments on Open thread, Oct. 13 - Oct. 19, 2014 - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (355)
I definitely agree with the line of argument that advertising is a public good, because it leads to more informed consumers, and I am highly sceptical of knee-jerk claims that it is a negative-sum arms race. So at least we're both consistent!
However, I don't think that advertising (whether commercial or political) should be subsidised, because I think the government is very bad at encouraging public goods. My point was merely that HonoreDB's charity, although no doubt well intentioned, appears to me to be destroying value, rather than creating it...
Consider your own analogy to commercial advertising. Suppose Coke and Pepsi signed a compact to reduce their advertising expenditures by a specified amount; would you suppose that to be good or bad for the consumer?
I don't think that agitprop and/or advertising leads to more informed voters/consumers because its purpose is not to inform. Its purpose is to manipulate, to force the subject to a certain opinion by all means necessary. Any "informing" that happens is entirely coincidental and, depending on the circumstances, could be considered a feature or a bug.
In local terminology, advertising tries to change the map in your head and the main feature of the one it wants to install is that it shows all paths leading to the same place, the one conclusion that it wants you to make. An accurate map is bad from the advertising point of view and needs to be replaced. In the service of this goal the advertisers can and do use biases and fallacies, they spin, mislead, and obfuscate, and on occasion just lie.
Economically -- good. Psychologically -- I don't know. People like to be told what to prefer :-/
Wow. Let's just say we're very far apart on this.
There's a wealth of law and economics literature about the effect of advertising, which demonstrates that advertising bans hurt consumers and help producers - see for example this classic. An agreement within a cartel isn't the same as a legal ban, but we should surely expect it to have a similar effect - especially given that many real-world advertising bans were lobbied for by major incumbents. Do you have any rationale for why you think consumers would actually benefit?
I was inviting you to consider what I considered an obvious cartel behaviour aimed at suppressing consumer ability to get the best deal. But bravo on biting the bullet!
First, we were not talking about legal bans (which I am generally not in favor of).
Second, you have to be quite careful here not to confuse "advertising" and "intensity of competition". I have no doubts that reducing the competition hurts consumers, but I am not convinced that reducing advertising expenditures necessarily leads to reduced competition. I suspect that these two things are often conflated (and the causation flipped).
In this particular case, do you think that if both Coke and Pepsi reduce their advertising budgets by, say, $10m each, the consumer will be hurt economically? What is the mechanism for that?
Third, are you implicitly claiming that the current level of advertising expenses is optimal? If we accept your thesis and start to increase advertising, will there be some point when the curve bends -- the advertising becomes excessive? Presumably so. Where are we with respect to this point? How do you know?
Plain-vanilla cost savings some which will be passed on to consumers.
Huh? I walk into a supermarket and look at the prices of Coke and Pepsi which are there side by side. I know from experience to which degree I prefer one over another. How will advertising help me get the best deal?
Glad to hear it. Do you agree with the wealth of literature showing that bans on advertising are bad for the consumer? And do you agree that a binding agreement within a duopoly would have a similar effect to a legal ban?
Yes, I think the consumer would be hurt. Advertising alerts us to new products, changes to existing products, and changes in the terms (eg price) under which those products are sold. Let me give you two examples of Coke/Pepsi advertising and how it affects me.
Where I live, Coke produces a wide variety of products, and is constantly adding more. Currently, they are heavily advertising their new "Coke Life" product, which has a different kind of sweetener, and a slightly different taste. If Coke had a smaller advertising budget, fewer consumers would be aware of this new product and what it's about, resulting in loss of the potential consumer surplus from drinking the new product among those who prefer it to other Coke or Pepsi products.
In addition, Coke frequently has promotional offers on. Just walking into the supermarket and look at the prices is inadequate, I specifically go there to buy Coke because of the promotional offer. Otherwise I might miss out. And I know about the promotional offer because of advertising. In the absence of this, consumers would have to go to the supermarket on a much-more-frequent basis, just to check the price of Coke. This would be a loss.
I am claiming that, given that the current level of Coke vs Pepsi advertising is the result of adversarial competition in a free market, I think there's a very heavy burden on people who claim it's "too high" (or "too low"). I am not claiming that it's "optimal" by everyone's idiosyncratic criteria.
Why on earth would the cost savings be passed on to consumers? Do you think Coke or Pepsi is sold at marginal cost? This is a market with unique products and partial substitution, so these companies are price-setters, not price-takers. This saving would just increase their profits.
You seem to be making a fundamental assumption which I disagree with. You are assuming that what is best for the producer is best for the consumer and that increased consumption is a public good. You are assuming that we are dealing with homo economus who decides correctly and for whom more information is always a good thing. We are dealing however with homo sapiens, who can be easily led into things against his best interest. I do not think your basic assumption holds and I point to the massive increases in obesity which have benefited producers (more demand) but not consumers (die sooner) as evidence. To use the specific example you've been using, coke is rather unhealthy, being mainly simple sugars which have been proven to lead to obesity in sufficient quantities. Its consumption is kept well above the normal set point by advertising and I think this is a negative thing on the whole.
However, the issue has become sidetracked in economic minutiae. The real question is this: Is campaign funding a greater or lesser good than effective altruism. $1000 to Malaria Foundation provides 20-100 DALY, as Yvain said higher up. I find it spectacularly unlikely that $1000 spent on TV adverts extolling the virtues of a candidate and lawn signs showing his face can provide a similar benefit or even one within the same order of magnitude. This is especially relevant when half comes from each candidate.
So, the ball is in your court.
Who decides what the "best interest" is?
I honestly couldn't say. In the borderline cases you would presumably need some kind of impartial observer with sufficient specialist knowledge. Luckily, I don't have to worry about borderline cases because the three cases we have here are fairly obvious. For an example of an obvious case of homo sapiens being led into things against his best interest consider smoking. It is extremely rare that smoking is in anyone's best interest given the high cost in both money and years of life such a habit entails.
I feel that's a major issue you'll have to face.
Sure. People have smoked a variety of dried plants (including but not limited to tobacco and marijuana) for a very long time. Much, much longer than advertising has been around. So, what's the "best interest" here, who decides what it is, and who "leads" people into something against their best interest?
Note, by the way, that if you honestly can't say who decides what's in a person's best interest (other than herself, of course) then the phrase "It is extremely rare that smoking is in anyone's best interest" doesn't mean anything.
Ideally I would measure best interest compared to the human utility function, but we do not have the luxury of a fully unpacked utility function. In the mean time I'll just go with (length of life x happiness) - (very large number x atrocities committed). As to who leads someone against their best interest, that would be advertisers as the agents of companies who wish to sell people things. Some advertisers also move people towards their best interest. The point is that the best interest of the buyer and the best interest of the seller are rather disconnected and intersect rather randomly. The best interests of the people being sold to are far less relevant to the people doing the selling than the amount of money a person can be persuaded to spend.
Edit: I feel like this would be a good place to put a chart of cigarette usage by % of population over the years. At current time, 42% of people smoking have tried to quit over the last year. I feel like this is fairly conclusive: these people are acting against their own self interest
Have you read any behavioural economics? These are rather central things to the theory and there are books out there that can explain this a lot better than me.
Also, we're getting sidetracked again. I thought the whole advertising thing was just a useful example to talk about the original disagreement about charity vs giving to political parties.
I am not sufficiently familiar with it and, frankly, I don't care enough about the topic to go read a bunch of economics papers and then fisk them. My data-less suspicion is that bans on advertising are a consequence of reduced competition and/or near-monopoly behavior by incumbents, just a harm to consumers is also a consequence of the same thing, and people misinterpret the correlation between "less advertising" and "harm to consumers".
So, you pointed out the benefits. What about costs? Why do you believe the benefits are higher than costs?
Also, you're ignoring the advertising for established products, as well as for failed products (e.g. the New Coke).
Note that I'm not saying that all advertising is harmful and that zero advertising is the desired state. I am saying that my best guess at the "optimal" point (which balances costs against public benefits) is such that I think the current levels of commercial advertising are above that point. Reducing advertising would get us closer to that optimum -- though, obviously, I don't know where exactly it is.
Of course the optimal point which balances costs against public benefits is different from the optimal point which balances costs against the firm's benefits.
Let's not get quite this ridiculous X-)
That's a cop-out :-) Besides, adversarial competition in a free market optimizes for the firm's benefits from advertising, not for the public benefits.
Because, as you mentioned, there is "adversarial competition in a free market". That includes price wars, promotional coupons, etc. By your logic, there should never be promotions for a product -- why lessen your profits for no good reason?
But I don't see costs to consumers here. Savings would not be passed on to consumers (see below), so what is the problem? That some people find the adverts annoying? Sure, but others find them entertaining. Coke in particular has had many adverts that have entered public consciousness.
This is exactly the kind of claim that I think should have to face a very heavy burden. Your WAG (which you cheerfully admit is not based on a careful reading of the literature) is that the public costs of advertising (which you do not specify) are greater than the (equally unspecified) public benefits. Because you can't specify or quantify any costs or benefits, you can't say how much you'd like to reduce advertising by, but it's just got to be reduced, dag nabbit!
So? That only implies that firms pass on cost savings if we have perfect competition (driving price down to marginal cost). If you have imperfect competition, firms are (at least partially) price-setters, not price-takers, and set price based on demand, to maximise profits. For an extreme example of imperfect competition, a Damien Hirst artwork that was cheap to make doesn't necessarily sell for any less than one that was expensive to make. For a standard example of (pseudo-)perfect competition, see petrol - it's an essentially indistinguishable commodity, so all petrol stations sell it at basically the same price (small changes based on location), and cost rises/falls are passed on to the consumer.
On the contrary, promotions for a product are an excellent sign that you don't have perfect competition - that's why you never see a sale on petrol (see above). Companies run promotions because they are selling well above marginal cost, but they want to be able to price-discriminate to make additional profits. For example, suppose I am selling Coke, and there are two people, Alan and Bob. Coke costs me 10p per litre to produce. Alan values Coke at £2.50 per litre, and would like to buy 1 litre per day. Bob values Coke at 60p per litre, and would like to buy 1 litre per day. So I price the Coke at £2 per litre, and make £1.90 (Alan pockets 50p consumer surplus, Bob doesn't buy). That's more profitable to me than pricing the Coke at 50p per litre, because although I'd make an additional profitable sale to Bob, it would reduce my profits from the Alan transaction by more. What I really want to do is sell Coke to Alan at £2 per litre, but Bob at 50p per litre, and that is where promotions etc come into play. Ideally I will find a way to advertise my promotion to Bob, without letting Alan find out.
I don't know what you mean by ridiculous. I buy a lot of Coke, but I am a thrifty shopper. I carefully collect coupons etc for discounted products (not just Coke) to take advantage of the cost savings, and buy staple products like Coke only when they are on promotion. In the absence of advertising for discounts etc, I definitely would go to the supermarket more often to check for offers. This would be an annoying waste of my time. There are plenty of other people like me.
I feel there's some disconnect here. Advertising costs are effectively paid for by consumers so of course the magnitude of these costs impacts the consumers. Imagine them doubled or quadrupled -- you don't think this would result in higher prices? Or do you believe the prices to be a ratchet going one way only so that reduced costs never lead to reduced prices?
Since it's the consumers who pay for advertising, the direct public costs of advertising are pretty easy to estimate: that's the revenue of the advertising industry. You can, of course, then start adjusting this number is a variety of ways.
Public benefits, I have no good estimate for.
My WAG is, of course, a WAG, but I don't see why your position that the level of advertising expenses happens to be optimal for the public benefit should enjoy the advantages of being the default baseline.
The reasoning behind my estimate is pretty simple. Firms set the levels of advertising expenditure based on their estimates of the benefits to the firm. My assumption is that any advertisement brings more benefit to the firm which places it than to the public at large. Given this, the market-determined level of advertising is going to be too high from the public benefit point of view.
Yes, of course, and it's a very complicated process which depends a lot on the particular details of the industry. However I find the blank assertion that the firms will not pass any cost savings onto the consumer (especially in a highly competitive industry like soft drinks) to be not tenable. There is the market force pushing prices towards the average (not marginal) cost and while it may be counterbalanced by many things it's still there. Look at, say, electronics -- as the costs drop so do prices.
I understand price discrimination, but that's irrelevant for the subject under discussion which is whether the consumer will ever see part of the cost savings.
Going to the supermarket "on a much-more-frequent basis, just to check the price of Coke" implies a ridiculously low value of your time. I don't know of anyone who goes to the supermarket just to check prices on soft drinks. And in this particular example we're talking much more about information rather than advertising. To illustrate the difference, if you subscribe to receive emails about Coke promotions that would get you all you need. Expensive video clips showing attractive women orgasming as Coke touches their lips are pretty useless for your purposes.
You keep asserting this. You provide no evidence or argument that it's true. I agree that advertising costs are likely to be paid for by consumers in (say) the petrol market, although given that market is complicated by franchises, even there it may not be true. I think they are very unlikely to be paid for by consumers in markets featuring (partial) monopolies, such as Coke/Pepsi. So no, I don't think that if Coke quadrupled its advertising budget it would be able to pass on the cost to consumers.
You seem to think the soft drink market is "highly competitive." And you're right, in the sense that everyone is trying to bring the best products to market, to make a profit. But you're wrong, in the sense that the products are not direct substitutes in terms of consumer experience. Coke does not taste the same as Pepsi, and only the Coca-Cola Corporation knows how to make Coke. This is why Coke can sell their product for twice the price of some supermarket own-brand cola; they are earning rents on their intellectual property. The same goes, to a lesser extent, for Pepsi. This is a partially-monopolistic market, very different from the market for electronics, where the products are functional substitutes, and so are close to commodities, and indeed, cost savings are passed on.
And note that we see the most advertising precisely in partially monopolistic markets, and very little in commodity markets, precisely because of the effect on prices.
I am sorry, but what other options are there? The advertising costs are paid out of interest on the firm's bank balances? Out of tax subsidies? Out of charity donations?
The firm's costs are paid out of the firm's revenues. If the firm's revenues come from selling things to consumers, the consumers are paying for the firm's costs -- all of them, including production, distribution, advertising, office space, janitors, and executives' membership in the golf club. The consumers get the product in exchange, of course.
As you mentioned, "You keep asserting this. You provide no evidence or argument that it's true." Let me provide a counterexample.
Many fast-food chains have exclusive contracts with Coke or Pepsi. McDonalds, for example, serves only Coke. Given this, you can directly observe whether Coke is accepted as a substitute for Pepsi: often enough at the counter you can hear the following exchange:
-- What's your drink?
-- Pepsi (automatic answer as that's what the person is used to drinking)
-- Sorry, we have only Coke.
And at this point the customer can either accept the substitution (and say "Coke is fine") or decline it (and say "I'll have X instead"). I don't have actual data, but I've seen this case happen many times and the number of people who will accept Coke is much higher than the number of people who will refuse it.
Coke and Pepsi are functional substitutes. They don't taste exactly the same, but then Samsung's and HTC's phones don't look and behave exactly the same either.
Citation needed. Advertising is basically buying market share. I would argue that we see most advertising in highly competitive markets where you can buy market share. That means that you can differentiate your product and convince part of the public that the product is better than the other guy's and not just because it's cheaper. And I'm not willing to call all markets with differentiable products "partially monopolistic".
Your ability to persuade an average bloke that petrol of brand X is better than petrol of brand Y is limited. Therefore your ability to buy market share is limited. Therefore you don't spend much money on advertising. But you ability to persuade the same bloke that beer X is better than beer Y is much higher. Thus you can buy market share and advertising beer is worth it (for the firm, of course).
This doesn't jibe with my intuition - I think virtually no one would be upset if there were fewer soda advertisements.
Do you think the same is true for iPhone advertisements?
Yeah, I think so. Maybe this is a culture-bubble thing, but I don't think I know anyone who would notice, much less care, if there were more or fewer advertisements for one particular product or another (ad space, keep in mind, is fungible).
But how will they know which product is cool, that is, is efficient at signaling status?
If it becomes effective at signaling status by being advertised, then the only people who would be disadvantaged by reducing the advertisements would be people who already had iPhones. People with the money to consistently own the newest iPhone surely have other ways of purchasing status, anyway, and would continue to do so even if that status wasn't assigned specifically through the medium of advertising.
Even if political advertising produces a little more informat voters, I find it unlikely that the money is as well spent as money on a GiveWell recommended charity.
Furthermore a lot of TV ads don't really inform and aren't completely honest. Watching a news show is more likely to inform than watching a campaign ad.
Polling that interrupts people also steals them valuable time and many people are too polite to simply put down the telephone. Less money spent on pollsters that optimize advertising messages is a net gain.
GiveWell's top recommended charity is giving direct aid to poor Africans. This may make their lives more pleasant, but is very unlikely to have any long-term effect - Africa is poor because it has bad institutions, not inadequate consumption. In 30 years time, GiveWell will still be trying to find ways to alleviate African "poverty," but will that word mean near-starvation, or something akin to the lives of poor Westerners today? That will be determined by the rates of economic and technological growth for the world as a whole, which in turn are critically influenced by public policy in the First World. Public policy in (broadly-defined) Western countries is the most important issue facing mankind today, and even small improvements are therefore worth vast sums. My own altruistic giving is entirely to a domestic political party for just this reason.
But a lot of news shows don't really inform and aren't completely honest, so your conclusion doesn't follow. Campaign adverts allow politicians to get their message out unfiltered by the news media - which has its own agenda. This is particularly important for anti-incumbent politicians. Advertising turns information presentation around elections into a properly adversarial process. If information only goes through the news media, that crucial element is often lost, and with it much of the accountability of elections.
Oh come on, this is marginal at best. Did you object to the census on the same grounds, or is this just mood affiliation?
Well yes, ceteris paribus. But presenting election information in a way that doesn't speak to the electorate is a net loss, ceteris paribus. I complained the other day that you can make anything look good under "benefit analysis" - here we have the converse, a "cost analysis." We do both sides of the cost-benefit analysis for a reason.
Have you looked at the actual arguments put forth by GiveWell? The money isn't mainly used for consumption but often used by people to start businesses that they otherwise couldn't start.
Empowering individuals to start businesses has advantages over funneling money into bad existing institutions.
I do value checks and balance and I don't want unfiltered lies.
The problem is that the value of the time of the person answering the phone isn't priced into the calculations of the person running the query.
I think the census does provide valuable data. More targeted political ads don't provide much value.
I'm not sure whether all the advertising is just about choosing between the two brands. A costumer might drink many different beverages besides Coke and Pepsi.