I sometimes think I am the last of the hardcore anarchists or ... something. My first reaction was, "So what? It's their store, their prices, their goods. Buy or don't buy." My second reaction was, if it's gaming you, game it right back. Find out what it uses to discriminate, and use that to get lower prices.
Government is not the solution. Anyone who thinks it is, is afraid to shoulder any personal responsibility for making their life better.
Very much of the same mind. You’re not alone! I hate the presumption that Instacart owes anyone anything. People got along fine for centuries without Instacart. If you find out Instacart does something you don’t like, don’t use instacart. They don’t owe you a thing.
In a perfect economic vacuum 100%! Competition would/should solve any pricing discrimination as consumers migrate towards the lowest acceptable price given an adequate amount of options. Given enough options and enough connectivity, there would in aggregate be a bidding war for the consumers demand, driving down the price. Conveying the adage of lowest price a product is to be sold at and highest price willing to be accepted.
In a perfect world, I really don't see how prices or demand would change all that much.
This brings back fond memories of my childhood. My family moved, in 1957, from one neighborhood to another, and I acquired a new set of friends. One day after playing baseball we all went into the Mom 'n Pop grocery store for candy and soda pop, when they all pulled out coupons for a variety of products that were not for candy and soda pop. The owner readily accepted them. No hesitation whatsoever.
Doesn't perfect price discrimination eliminate consumer surplus entirely? This seems like a negative outcome that we should care about and build policy around.
But currently I’m free riding on more price-sensitive customers putting in the work to set those price ceilings and open space for competition if they get too high. There’s non-economic utility there, like there is in my ability to use Wikipedia despite not contributing.
Now I have to be vigilant about pricing, because the model allows for me to get overcharged.
My exact thought. Customers (and voters) seem to prefer a world with some deadweight loss and some consumer surplus versus perfect price discrimination and no consumer surplus even though the former is technically worse.
In your last paragraph, you state that randomization around the same mean price is better for consumers than stable pricing, which I agree with. But why is randomization around the same mean price the assumed counterfactual to stable pricing?
I think the "normie" intuition is that price discrimination leads to randomization around a higher mean price than stable pricing, hence the natural assumption that price discrimination is bad? Is there a mathematical basis for why price discrimination *doesn't* lead to a higher average price than pure stable pricing?
Yes. Assuming the only way to enforce a ban against price discrimination is to also ban price experiments. (Because otherwise companies can just claim any instance of the former is actually the latter.)
Price discrimination is bad for consumers. An obvious example is medical care. The medical profession (and yes, they're a cartel) should not be allowed to charge a rich person a million dollars for a thousand-dollar cataract surgery just because he'd be willing to pay it.
Other markets may be less cartel-dominated and may be more optional for the consumer. So the harm is less obvious. But it is still bad.
Why not? We've been through this before. Voluntary trades are beneficial and acceptable to both parties by definition. If the rich person is willing to pay a million bucks for a thousand dollar procedure, that's his business. It is nobody else's business what the "medical profession" charges or what the patient accepts, and the very wording shows an appalling lack of distinction between the collective "medical profession" and individual doctors or hospitals who actually do the charging, as if the entire "medical profession" or "cartel" operates like a single-minded Borg.
The problem with these strategies isn’t easily captured in Econ terms - there’s a labor tax on having to think about how the algorithms are trying to outsmart you. It turns a simple purchasing task into a complicated one. (Or you pay the tax on not doing the work.) I don’t want to have to think about what time of day to buy eggs.
There’s also consumer and producer surpluses baked into the economy (if there wasn’t, I think we’d have Marx’s zero-profit capitalism, no?) some technologies help reduced producer surplus (ie price comparison engines for airline tix), others can help reduce consumer surplus. As a grocery consumer, this is a bad thing!
What gets missed in these debates is that fairness is always measured against an imaginary uniform-price world. In most real markets, uniform pricing raises deadweight loss and excludes marginal buyers. Randomized testing and even basic segmentation usually increase output and shrink the inefficiency wedge. If critics want stable, non-experimental prices, they’re implicitly asking for less learning and slower adjustment. This hurts consumers far more than it helps them.
What I find interesting thing about discussions around price discrimination is the conceptual shifts around what prices are, or ought to be, over time. Foucault noted how in the last century the market was portrayed as some sort of truth seeking machine that through the laws of supply and demand would come to the 'correct' price (which we can see explicitly with Hayek, but also in equilibrium models), whereas, he said, in a more distant past prices were more often based on ideas of justice and fairness (therefore including price discrimination, such as charging more to people you don't know, or to the rich, etc.).
Currently pricing in stores is often still the same for everyone (excepting coupons), but online there likely is more price discrimination than we're aware of, which sparks these fears of algorithmically driven price discrimination (or diversification if you wish to use a more neutral term). The more price diversification, the less tenable the idea of a markets as driven by the 'natural laws' of supply & demand coming to the 'correct' prices. Instead we become more and more aware of the fact that companies charge whatever they can (still moderated by things like costs, power, information & competition though).
Our conceptualization of what markets are, or how they function, thus moves further and further from the idealized theoretical (equilibrium) models used in the previous century. We now recognize there is more freedom for companies to set prices than was acknowledged before (when 'market prices' would shut down every discussion), albeit still moderated by external forces. But, this recognition that there IS some level of control (for companies to set their own prices), also reopens the discussion whether there ought to be additional limits to that control based on other moral/political/economic ideals such as efficiency, justice/fairness or limiting (market) power.
Price differences dont amount to price discrimination if they simply reflect differing marginal cost between customers (eg different costs to serve). As long as the ratio of price to marginal cost is the same to different clases of buyers, that's not price discrimination. And in antitrust/competition law price differences dont matter if they dont significantly affect competition between buyers.
If you're going to ban price discrimination, you're going to have to make arguments why government action should be the default. You're arguing for guilty until proven innocent.
Oh, so you like meddling intrusive know-it-all government?
Your basic premise is that government knows best. "If you want to say XXX should be allowed ..." You have surrendered your personal responsibility to the government.
The current status quo is that price discrimination is NOT illegal. Brian gave numerous examples of perfectly lawful price discrimination. Like all conduct that can have positive or negative welfare effects in theory (though, as Brian points out, perfect price discrimination by a monopolist has only positive welfare effects in theory), you must explain how we can separate the good from the bad if you want to make it illegal.
I sometimes think I am the last of the hardcore anarchists or ... something. My first reaction was, "So what? It's their store, their prices, their goods. Buy or don't buy." My second reaction was, if it's gaming you, game it right back. Find out what it uses to discriminate, and use that to get lower prices.
Government is not the solution. Anyone who thinks it is, is afraid to shoulder any personal responsibility for making their life better.
Very much of the same mind. You’re not alone! I hate the presumption that Instacart owes anyone anything. People got along fine for centuries without Instacart. If you find out Instacart does something you don’t like, don’t use instacart. They don’t owe you a thing.
In a perfect economic vacuum 100%! Competition would/should solve any pricing discrimination as consumers migrate towards the lowest acceptable price given an adequate amount of options. Given enough options and enough connectivity, there would in aggregate be a bidding war for the consumers demand, driving down the price. Conveying the adage of lowest price a product is to be sold at and highest price willing to be accepted.
In a perfect world, I really don't see how prices or demand would change all that much.
This brings back fond memories of my childhood. My family moved, in 1957, from one neighborhood to another, and I acquired a new set of friends. One day after playing baseball we all went into the Mom 'n Pop grocery store for candy and soda pop, when they all pulled out coupons for a variety of products that were not for candy and soda pop. The owner readily accepted them. No hesitation whatsoever.
My first lesson in price discrimination.
My reply https://open.substack.com/pub/dwmackenzie/p/is-ai-a-tool-for-monopolistic-waste?r=ltb8d&utm_campaign=post&utm_medium=web
Awesome article Brian, 2 questions:
Doesn't perfect price discrimination eliminate consumer surplus entirely? This seems like a negative outcome that we should care about and build policy around.
That is the case with monopoly. As I say in the piece, competition changes that. I wasn't explicit but perfect price discrimination is consumer welfare maximizing in my paper that I link. I also talk about it here https://www.economicforces.xyz/p/what-the-heck-is-welfare?utm_source=publication-search
But currently I’m free riding on more price-sensitive customers putting in the work to set those price ceilings and open space for competition if they get too high. There’s non-economic utility there, like there is in my ability to use Wikipedia despite not contributing.
Now I have to be vigilant about pricing, because the model allows for me to get overcharged.
My exact thought. Customers (and voters) seem to prefer a world with some deadweight loss and some consumer surplus versus perfect price discrimination and no consumer surplus even though the former is technically worse.
In your last paragraph, you state that randomization around the same mean price is better for consumers than stable pricing, which I agree with. But why is randomization around the same mean price the assumed counterfactual to stable pricing?
I think the "normie" intuition is that price discrimination leads to randomization around a higher mean price than stable pricing, hence the natural assumption that price discrimination is bad? Is there a mathematical basis for why price discrimination *doesn't* lead to a higher average price than pure stable pricing?
That’s just to make a fair comparison and isolate only the randomization part
Would Banning Price Experiments Help Consumers?
Yes. Assuming the only way to enforce a ban against price discrimination is to also ban price experiments. (Because otherwise companies can just claim any instance of the former is actually the latter.)
Price discrimination is bad for consumers. An obvious example is medical care. The medical profession (and yes, they're a cartel) should not be allowed to charge a rich person a million dollars for a thousand-dollar cataract surgery just because he'd be willing to pay it.
Other markets may be less cartel-dominated and may be more optional for the consumer. So the harm is less obvious. But it is still bad.
Why not? We've been through this before. Voluntary trades are beneficial and acceptable to both parties by definition. If the rich person is willing to pay a million bucks for a thousand dollar procedure, that's his business. It is nobody else's business what the "medical profession" charges or what the patient accepts, and the very wording shows an appalling lack of distinction between the collective "medical profession" and individual doctors or hospitals who actually do the charging, as if the entire "medical profession" or "cartel" operates like a single-minded Borg.
The problem with these strategies isn’t easily captured in Econ terms - there’s a labor tax on having to think about how the algorithms are trying to outsmart you. It turns a simple purchasing task into a complicated one. (Or you pay the tax on not doing the work.) I don’t want to have to think about what time of day to buy eggs.
There’s also consumer and producer surpluses baked into the economy (if there wasn’t, I think we’d have Marx’s zero-profit capitalism, no?) some technologies help reduced producer surplus (ie price comparison engines for airline tix), others can help reduce consumer surplus. As a grocery consumer, this is a bad thing!
What gets missed in these debates is that fairness is always measured against an imaginary uniform-price world. In most real markets, uniform pricing raises deadweight loss and excludes marginal buyers. Randomized testing and even basic segmentation usually increase output and shrink the inefficiency wedge. If critics want stable, non-experimental prices, they’re implicitly asking for less learning and slower adjustment. This hurts consumers far more than it helps them.
I’m not sure what cases have consumers hurt more but it is definitely true about the learning and excluding marginal buyers
What I find interesting thing about discussions around price discrimination is the conceptual shifts around what prices are, or ought to be, over time. Foucault noted how in the last century the market was portrayed as some sort of truth seeking machine that through the laws of supply and demand would come to the 'correct' price (which we can see explicitly with Hayek, but also in equilibrium models), whereas, he said, in a more distant past prices were more often based on ideas of justice and fairness (therefore including price discrimination, such as charging more to people you don't know, or to the rich, etc.).
Currently pricing in stores is often still the same for everyone (excepting coupons), but online there likely is more price discrimination than we're aware of, which sparks these fears of algorithmically driven price discrimination (or diversification if you wish to use a more neutral term). The more price diversification, the less tenable the idea of a markets as driven by the 'natural laws' of supply & demand coming to the 'correct' prices. Instead we become more and more aware of the fact that companies charge whatever they can (still moderated by things like costs, power, information & competition though).
Our conceptualization of what markets are, or how they function, thus moves further and further from the idealized theoretical (equilibrium) models used in the previous century. We now recognize there is more freedom for companies to set prices than was acknowledged before (when 'market prices' would shut down every discussion), albeit still moderated by external forces. But, this recognition that there IS some level of control (for companies to set their own prices), also reopens the discussion whether there ought to be additional limits to that control based on other moral/political/economic ideals such as efficiency, justice/fairness or limiting (market) power.
Great article! Thanks for this. Are you aware of any similar analysis with Amazon’s algorithmic pricing ?
Price differences dont amount to price discrimination if they simply reflect differing marginal cost between customers (eg different costs to serve). As long as the ratio of price to marginal cost is the same to different clases of buyers, that's not price discrimination. And in antitrust/competition law price differences dont matter if they dont significantly affect competition between buyers.
If you're going to ban price discrimination, you're going to have to make arguments why government action should be the default. You're arguing for guilty until proven innocent.
Oh, so you like meddling intrusive know-it-all government?
Your basic premise is that government knows best. "If you want to say XXX should be allowed ..." You have surrendered your personal responsibility to the government.
The current status quo is that price discrimination is NOT illegal. Brian gave numerous examples of perfectly lawful price discrimination. Like all conduct that can have positive or negative welfare effects in theory (though, as Brian points out, perfect price discrimination by a monopolist has only positive welfare effects in theory), you must explain how we can separate the good from the bad if you want to make it illegal.
Looks like he has left the building. But I learned something. However these comments were deleted, they no longer show up in my "Activity" log.
Substack really needs a user manual. But it is kinda fun treating it as a black box and trying to figure out what's inside from how it behaves.
What a stupid argument. “Government’s on my side so I don’t have to justify its actions. Nyah nyah.”