Transcript of Conference Call: “The Perils of Personalized Pricing” with Hal Singer and Gavin Sicard

On August 11, The Capitol Forum held a conference call with Hal Singer, Managing Director at EconOne, and Gavin Sicard, Analyst at EconOne, to discuss their recent article for The Sling, “The Perils of Personalized Pricing.” The full transcript, which has been modified slightly for accuracy, can be found below.

TEDDY DOWNEY: Hello, everyone. Welcome. I’m Teddy Downey, Executive Editor here at The Capitol Forum. Today, I’m pleased to be joined by Hal Singer, Managing Director at EconOne and longtime friend of The Capitol Forum, and Gavin Sicard, Analyst at EconOne, who we are super excited to welcome into The Capitol Forum community here.

We’ll be discussing their recent article for The Sling, “The Perils of Personalized Pricing,” which examines the growing use of consumer data and pricing algorithms to set individualized prices, the potential effects on consumer welfare and privacy, and recent efforts by state legislatures to restrict the practice. Hal and Gavin, thank you so much for doing this today.

HAL SINGER: Hey, thanks for having us on.

TEDDY DOWNEY: Really quickly, we’ll be taking listener questions towards the end of the call. But if you have questions, please put them in the chat or in the Q&A panel here and we’ll get to them later on in the call. But first, we’d love to hear how—we’ve talked about algorithmic price fixing in the past. Why did you and Gavin look at this issue now? How did it come up? And how did you think about writing this piece?

HAL SINGER: Sure. I’ll take it first stab and Gavin can weigh in. But we’ve noticed that there’s a bunch of legislation circulating at the state level to protect consumers from personalized pricing. And we thought it was important for economists to weigh in with an economically-based argument against personalized pricing.

It’s so typical to see economists kind of knee-jerk defending price discrimination. And they do it with appeal to an argument that isn’t even related to what we’re talking about here, which is they like to talk about group discounts and how great group discounts are. They’ll talk about student discounts and senior discounts. And that isn’t the issue that’s being debated here.

What’s being debated here is personalized pricing, that is setting an individual price for each buyer according to personal attributes of that buyer. It’s a completely different animal. And the typical defenses that economists tend to usher out often, either on behalf of corporate clients or chasing corporate clients, is to say that there’s winners, there’s newfound winners, that come about when you allow for this sort of price discrimination.

And we just think those arguments—which actually have merit in the old context of group discounts—just don’t apply in the current context of personalized pricing. And we can get into why it doesn’t. But I just felt that it was important for us to weigh in there.

GAVIN SICARD: Yeah, and if I can just start off, I think it’s good to start off with the definitions of what we’re actually talking about. So, personalized pricing, it goes by a bunch of different names, personalized pricing, surveillance pricing is pretty common in legislatures now. But in the economics textbooks, it’s probably going to be called perfect price discrimination or first-degree price discrimination.

And in the economic theory, this is essentially the idea that for each individual unit of a good, we can set [the price] to be personalized for that consumer consuming that individual unit of good. And the ultimate goal is to charge the consumer the maximum willingness to pay they have for that good.

That is different from what is known as group [based] price discrimination or third-degree price discrimination, which essentially you have a bunch of consumers and you put them into groups and you charge the groups different prices.

And we then have the topic of dynamic prices. And to us, that is more like a subcategory of group based price discrimination. And in that situation—you have during the time of day or during different demand conditions—you aggregate consumers based off when are they consuming. And, by those different discrete groups, then you set different prices.

So, these three things are related. But I think a lot of times people kind of muddy the waters behind what they’re talking about. And arguments that might be showing how group-based price discrimination is beneficial for certain consumers are not necessarily the same when we go to the perfect price discrimination case.

TEDDY DOWNEY: And it seems like you’re going to have some overlap between a dynamic and a personalized, in that they can be pretty close if you know about where they are. You can sort of fudge that line potentially a little bit.

GAVIN SICARD: I would just say it’s sort of like calculus-like group-based price discrimination. You have these big groups. But obviously, as the groups shrink, shrink, shrink, shrink, shrink down to the incremental unit, you get down to personalized price discrimination.

TEDDY DOWNEY: For the non-economists here, who think about these dynamic prices and personalized prices as just objectively totally unfair and unreasonable to a normal person. You ask a normal person on the street, do you think this is okay? They say, no, that’s unfair. Why should I get a different price from someone else? I just don’t like it. It feels wrong to me.

Where does this history of economists justifying this as a good thing come from? Sort of like, give me a sense of this history of like the disconnect between your average person looking at this and how economists are so thrilled by this sort of as a utopic kind of a world.

HAL SINGER: Yeah. Well, I would say that economists like to defend the firms. That’s kind of what we’re on earth to do, is that we’re defending bad practices and going against the popular will. So, I feel like there’s probably a lot of consulting money swirling around for economists to come up with just-so stories in defense of price discrimination.

And, like I said at the top, when you get into group pricing, you can make a very powerful argument that, relative to a uniform price, we could expand surplus by bringing in kind of these price sensitive types who otherwise wouldn’t be able to participate in the market.

So, of course, in that context, there was an opportunity for economists to make some fairly compelling arguments. We think those arguments go away now because we’re not talking about setting a price below anyone’s willingness to pay anymore. Under personalized pricing, where literally they’re doing everything they can, including biometric data, how you’re feeling, whatever they can get ahold of, right? They’re trying to come up with a price that is one penny below your willingness to pay.

And now we’re in this dynamic, no pun intended, that we can actually get to a world where all the surplus that consumers previously enjoyed, whether you were in the market or you were out of the market, right? All that surplus is going to be drained under a regime of personalized pricing. And that’s what we want to tap the brakes on. We’re saying, no, no, no. These economic arguments just don’t apply anymore. Gavin, do you want to add anything else?

GAVIN SICARD: Yeah, I think from the general textbook theory of economics [there] is a very neutral view about the distribution of surplus. So, from an economics perspective, the goal is to increase social welfare, which is covering both the producer surplus and the consumer surplus. A lot of economists will say the distribution about who’s actually getting that welfare is not necessary for us to decide.

So, the idea behind price discrimination is that price discrimination is possible when there’s market power in the economy. And when there’s market power, that usually means there’s some amount of deadweight loss. So, there’s some amount of inefficiency in the economy. There’s some amount of underproduction. Economists look to price discrimination as a way to sell more units and therefore eliminate that deadweight loss to increase output.

So, the whole idea is any strategy that increases output, any strategy that increases social welfare, is a beneficial one from an economic standpoint. What we’re trying to bring to this perspective is, is the trade-off of possibly eliminating a tiny deadweight loss worth consumers losing all their surplus?

So, we try to bring the “distribution of” question back in. And I think the “distribution of” question, when it’s absent from the discussion kind of ignores what consumers really have the big concerns about, [which] is that they don’t want to get screwed over when they go into the grocery store.

HAL SINGER: Teddy, is it okay if I just weigh in with one more point?

TEDDY DOWNEY: Of course.

HAL SINGER: Sorry, I don’t mean to hog the mic. But in the old defense of price discrimination, they could point to some winners. They knew they couldn’t point – [to] the firm profit, the total welfare guys might point at that, but they realized that wasn’t going to be a winning soundbite. But they could point to a group of winners, which were the high price sensitive types who were otherwise priced out of the market. They could look to their welfare and say, look, those guys are now getting to participate in the market and they’re earning surplus. And they didn’t get any of this before, okay?

And what we’re coming back with is, okay, that occurred because you were setting a price for the sensitive types in such a way as to allow them to earn surplus. But if the goal now is to literally set a price that’s one penny below everyone’s willingness to pay, even if you bring new people into the market, if you’re bringing them in just to drain 100 percent of their surplus, who are the winners? Who are the winners then?

Obviously, the firm who’s engaged in [price discrimination, and] their pricing consultants are the winners, but I don’t think that’s going to persuade any hearts and minds. But we don’t have any winners any longer that we can point to, at least for consumers.

I’d like to just make a concrete example if I could, and I promise I’ll pass this back to you. But you hear about the case about Uber conditioning your price on your battery life. And so, my CCIA friend and my Penn economist friend—hopefully they’ll still be friends after what I have to say—they’re going to say we want this sort of price discrimination. We want Uber to condition your price on your battery life.

And I just ask them, who is the winner? Who wins besides Uber? Who wins from that? Obviously, the people with low battery life who are now going to be exploited—this is also called desperation pricing—they clearly are the losers, right? I hope no one would ever, with a straight face, suggests that they’re better off, okay?

The idea that I think they’re trying to sell you on is that Uber is going to take the overcharges on these desperate customers with the low battery life and somehow plow them back into discounts for those lucky guys who happen to get on the ride with 100 percent battery still. And you just have to recognize that that’s not Uber’s goal. That’s not what the algorithm is designed to do. The algorithm is not taking overcharges on one class, the low battery guys, and going and finding savings. That’s not what it’s intended to do.

The algorithm is designed to extract maximal surplus from each and every buyer individually. There’s no subsidy, cross subsidy, going on. There’s no interlinkages going on between buyers. They want to find out your maximum willingness to pay. You’ve given up to them now—through no fault of your own that your battery is low.

And now it’s bing, bing, bing. Let’s hit this guy. Let’s hit this guy because if he doesn’t take the ride, he might spend the night on the street tonight. Okay, rest my case.

TEDDY DOWNEY: It’s interesting because they’re trying to transport that group discount argument that worked into the personalized pricing, but it just doesn’t make any sense anymore at that point. Yeah, yeah, we’re going to take this surplus and give it to poor people and veterans and discounts. And they’re just like, no. We just have this great scheme to maximize profit. But that argument completely goes away.

We have a listener question that’s relevant here. Maybe we can ask this. Can you provide a real world example of first degree discriminatory/personalized pricing that does not fall into the other categories of dynamic pricing, one that is present now and harms consumers?

HAL SINGER: Yeah, Uber. Gavin, I’m just going to go back to my Uber example, conditioning pricing on battery life. But there’s also—I don’t mean to pick on Uber, great company. But there’s also research that suggests —

TEDDY DOWNEY: Great company that exploits drivers and consumers on what would you call it, desperation?

HAL SINGER: I probably wasn’t going to be invited to Uber’s Hanukkah party this year anyway. But there’s also research that suggests that Uber is charging more depending upon whether the destination or pickup is in an area with more home owned by Black Americans. That’s in the literature. It suggests that either destination or pickup spot, percentage of Black households in that neighborhood cause the prices to go up, right? That’s another example, I think, that it’s personal. It’s depending on where you personally are going, right? It’s not something for the group.

Gavin, do you have another one? And we don’t mean to pick on Uber.

GAVIN SICARD: Yeah, I would just say that firms are often are really evasive because they know this is a very unpopular thing, to try and come out and publicly admit, yeah, we’re using a perfect degree of price discrimination. So, whether or not firms are actually at the level of perfect price discrimination, I’m not sure we’re quite there yet. But there’s certainly firms that are starting to develop and push for more of this.

So, a lot of the state-level legislation spawned out of this Consumer Reports investigation into Instacart. And essentially, Instacart claims this was pricing experiments. So, just testing out how different consumers respond to different prices. But the analysis showed that Instacart was charging upwards of 23 percent price differences for certain foodstuff items. Upwards, according to the report, it claims that the average family could pay upwards of $1,200 more for goods. So, the idea was Instacart was using whatever consumer profile they’ve constructed to try and charge you more based off your willingness to pay.

And that was an experimental condition. But increasingly we’re seeing firms trying to propose this idea, face consumer backlash, kind of backpedal a little bit. But there’s certainly appetite for them to stop pushing and develop this where it becomes more and more a part of the real world.

So, famously, Delta’s CEO, I think, on an earnings call last year, talked about wanting to deploy personalized pricing to—I think it’s like 20 percent of flights. And then quickly Delta walked back and claimed that they weren’t going to use AI pricing algorithms to actually do that.

So, we’re kind of in like this interesting phase where there’s a lot of firms that want to move in this direction. But there’s also a lot of firms that recognize that publicly doing so is a big risk. So, I think as the technology continues to develop, we’re going to see more and more firms move in this direction. But whether or not we’re actually at that point there, I’m not sure.

TEDDY DOWNEY: We define personalized pricing, but some people call it surveillance pricing. You talk about how this actually works. Maybe walk our listeners through what’s really going on. You walk in the grocery store. How are companies going to use your data? Where are they getting the data from to personalize the pricing? I think that’s interesting context for us to understand what’s really going on here.

HAL SINGER: Yeah. Well, there’s a real example that’s happening now, and I want to give an example of what could happen, given the technology and some of the suggestions about their pricing consultants about where they’re going.

So, we know that in the case of Target—I think Lindsay Owens [president & CEO of Groundwork Collaborative] recognizes this in both her book [Gouged], but also in her congressional testimony last week—they’re conditioning their pricing based on how close you are to a physical Target location. Kroger, you’re finding out, is investing a lot of money in developing this ability to basically surveil you.

But just to make a concrete example—and this is a little sci-fi because I don’t know if it’s ready yet. But imagine if they could take your biometric data into account when you walk into a grocery store and you happen to have a headache. Now, if your body temperature is running hot, that would suggest that your willingness to pay for aspirin is going to be greater than someone whose body temperature is normal or just slightly elevated.

And so, a doomsday scenario in my mind is the notion that the poor guy who’s got the hottest head, who’s giving off the most heat, is going to be paying the highest price for the aspirin. Relative to someone who’s got a moderate headache. Anyway, I’ll leave it at that.

TEDDY DOWNEY: Yeah. Actually, I’ve got small kids. I’m walking in the drug store. I’m in the aspirin aisle. I’m in the Tylenol for kids aisle. I’m perfectly fine, right? I’m going to have a higher price than even that person buying the aspirin for himself, right? I’m the one that is going to get hit the most desperate, right? And it’s just because I have small children.

HAL SINGER: Teddy, there’s another one too. There are parking garages in Georgetown—I actually have regret. I’m going to do a self-deprecating joke. I drove my midlife crisis car into a parking garage in Georgetown and it just said dynamic prices are in effect and you’ll find out what you’re charged upon leaving. I was like, what the hell? I could have taken the Jeep, you know? But I took the midlife crisis car. Now I feel like, maybe like, God, I’m just inviting them just to pick a price to come after me. I know that one isn’t going to pull at anyone’s heartstrings. So, we should probably go back to bringing the baby.

TEDDY DOWNEY: I find it endlessly frustrating when they’re like we’re going to take a picture of your license plate and bill you later. Like, what are you talking about? How is it possible that I don’t know how much I’m paying to park in this garage? I mean, this is utter lunacy.

HAL SINGER: It is insane.

TEDDY DOWNEY: Utter lunacy. But look, we could spend a lot of time talking about the different ways it feels bad to be surveilled and be priced this way. You mentioned, obviously, there is public backlash when you disclose this. There are new laws being implemented at the state level. We’ve got historically no movement really in the federal legislature and Congress, unfortunately, to address these types of market changes. But at the states, the states are moving fairly quickly in response to some of this public backlash. Tell us what you’re seeing at the state level when it comes to legislation to address this and then maybe we can talk about law enforcement after that.

GAVIN SICARD: Yeah., So, just to answer that, firstly, to focus on the federal government—I think Hal referenced this—but just about a week ago, the Senate Judiciary Committee’s [Subcommittee on Crime and Counterterrorism] did hold a hearing on personalized pricing and how companies are using data to extract consumer welfare. And that was hosted by Hawley [R-MO]. So, a Republican. So, there is at least some bipartisanship.

Now, of course, Congress is famously deadlocked and it’s probably not going to actually pass anything. But there’s at least some federal appetite for analyzing these things. But again, yes, we’ve actually seen movement in the states.

So, right now we have three states that have passed and enacted laws. And we have one state that has passed the law, but is still waiting for a governor’s signature and one state that a governor vetoed.

So, Colorado is the state that the governor vetoed. Maryland passed the first law, Connecticut the second, New Jersey the third. And New York, we currently have the law passed by the legislature, but waiting for the governor’s signature.

And there’s a little bit—and again, I reference that, I think a lot of this movement started with this Consumer Reports [article] on Instacart. So, a lot of these bills have been focused on grocery stores specifically or grocery delivery services specifically. That includes Maryland’s and New Jersey’s are both grocery store specific. But New York’s and Connecticut’s bills are more broadly applicable to personalized pricing and surveillance pricing as a whole. There’s also a lot of pressure, obviously, legislative-level, I think, for more carve outs.

So, the complaint that you constantly see from firms is that they fear that a personalized pricing ban will lead to restrictions on their ability to offer loyalty discounts and couponing. And I think going back to the idea that personalized pricing and group-based discrimination is different, I think they are kind of relying on the idea of conflating these two. Like, the ability to enter loyalty programs and then to uniformly offer discounts to those who are part of that loyalty program is different from the ability to use all the personal information they collect to try and specifically target discounts towards you.

And then just one point of clarification I want to also focus on. In Colorado, the governor vetoed the bill claiming that he was worried that it would block lower prices or the idea that [conusmers] can discount off list price. It’s important to keep in mind that a discount doesn’t necessarily mean you’re getting a deal on a product.

So, the idea is that you can inflate your list price so that it’s beyond what you would charge in a normal market condition and then use discounting to personalize the price. So, it might be the case that eggs normally would be $3, but I could inflate the price to be $5 so I can more readily discount off that $5 list price to charge consumers higher prices than they otherwise would in a normally competitive markets. So, that’s my spiel. I don’t know, Hal, if you have anything to add.

HAL SINGER: Well, you raise the federal stuff and I just thought it was worth—I went through the testimony. I got real excited. I saw this argument from Lee Hepner [Senior Legal Counsel of the American Economic Liberties Project] in his filed testimony last night. And it’s one that we didn’t contemplate in our piece. And I just think, Teddy, I’d like to get it out because I thought it was so potent.

We made this public policy argument that surveillance pricing is bad because it induces firms to invest more heavily in privacy invasions. The winner of this competition is going to be the firm that can most ably and exactly predict what your willingness to pay is and charge a penny below that and gather that information. So, the gathering of the information is going to distinguish the winners and the losers on the supply side of the market.

And we just feel like it creates a perverse and distorted set of playing rules that says the person who steals the most and gathers the most and invades your privacy the most is going to be the winner. That’s basically the competition that we’re unleashing.

But Lee made this great point—and it’s a related but different one—and it says that small firms who don’t have the ability to gather this sort of information at the customer level are going to turn over their pricing authority to a third-party agent who is likely doing the same for other companies in the same space, including horizontal rivals.

So, again, if what we’re unleashing here are these market forces whereby everyone wants to invade your privacy and the smalls who don’t have the means to gather this personalized pricing as efficiently are going to turn over their pricing authority to third parties. Well, we know what happens then—when we go to the RealPage example—is that common pricing algorithms engender non-competitive outcomes. They facilitate collusion. They allow for prices to move in the direction of monopoly levels.

So, it’s just another public policy argument that I kind of wish I had made. I’m jealous of Lee. I have Lee envy. I’m channeling my inner Lee. And I just think that it’s so compelling. It’s another public policy argument about the perverse incentives that we are unleashing if we don’t constrain firms and their ability to compete on the basis of personalized pricing.

TEDDY DOWNEY: It’s interesting. When we were writing about Instacart’s new product where they would go to the grocery stores and say, hey, we’ll do this for you. We’ve got these shoppers. We can get all this information. We can take your information and get you better prices along these lines. They pushed back so hard. Obviously, subsequently, we have the Consumer Reports investigation. But that’s exactly what you’re talking about. If you’re a small store, there’s no way you’re going to be able to have that type of information available. And you’re going to be partnering with the Instacarts of the world to do that.

I want to spend a little bit of time talking about, so what is the reality for your state legislature? Obviously, I think that we’re going to get a new governor in Colorado who’s very likely to vote for, not veto, this legislation the next time around. You’ve getting pushback in Maryland that they have too many loopholes in their bill. What should—if you are a rigorous, thoughtful state legislator—what types of arguments should you be mindful of?

You’re pointing out, hey, really these are different things. The privacy invasion, the discounting, it doesn’t mean that you can’t have lower prices. What’s the reality for these state legislatures so that they’re not being—they can assess these laws and implement them with intentionality to address the problem that we’re talking about? There is a public dislike for this stuff. And how do you solve that from a legislative standpoint?

GAVIN SICARD: Yeah, it’s tricky because they’re going to be facing—obviously, the biggest lobbies are going to be the grocery stores or the firms fighting against it. And the ability for consumers to organize against those lobbying groups is troublesome. But I do think the idea is that they have to be explicit about what they’re actually targeting.

So, Maryland is the first mover, but it’s an often criticized law. And there’s a lot of confusion in the law itself. Dynamic pricing is defined as surveillance pricing. While, in New York’s law, they clearly define the two as separate things.

So, the ability to be clear and upfront about the fact that I think you have to specifically target the activity of the data collection. So, I think New York’s law does this more specifically than Maryland. But the actual principle of banning the collection of data gets you to part of the way there of solving the personalized pricing dilemma.

And I think a lot of the times the ideas that they’ll [proponents] present is that competition will save us. And I think what we’re kind of giving the perspective is that increasingly in this day of age—and there’s still legal disputes about whether or not algorithm price collusion is illegal. The Third Circuit Court and the Ninth Circuit Court currently disagree on that.

But the idea that firms are increasingly moving in a way where they might collaboratively set prices and the idea that with these data collection tools and with the third-party services, they might, again, if you’re relying on a third-party source to collect this data, to aggregate it, to give you pricing recommendations, there’s nothing stopping the grocery store down the street from using that same source.

So, I think they should be mindful of the principle that, yes, in certain cases, competition may be able to resolve some of the concerns with personalized pricing. Is that really going to be the world in which we’re moving with the increasing advances of technology? Hal, I’m not sure if you have any other thoughts.

HAL SINGER: I would just again point to Lee Hepner’s testimony from last week where he lays out like a template for what the ideal legislation would look like. I know an important thing that Lee says at the top is you want to be explicit about the carve-outs to immediately eliminate this fake argument that they’re bringing up about group discounts or senior discounts, that you’re going to make it clear that certain activity is exempted. And you have to be very clear as to what those exempt activities are.

I think I would just also add, Teddy, no one’s asked me to write a bill yet. But if I were to write one, I think I would want to spell out the types of information that we don’t want firms to be conditioning on in setting prices. So, in the case of Uber, I think we should be explicit about saying that we don’t want the user’s battery life to be an input that goes into the price that one gets quoted.

And similarly, we could go into other examples. An airline. Should an airline be able to condition your price on knowledge that they found out based on your prior web history? Who knows that the reason why you’re going somewhere is because of a funeral. I just feel like that sort of personalized information should be off limits.

So, to wrap up, I would say that we want to spell out explicitly what sort of information cannot be conditioned on when setting someone’s individual price.

TEDDY DOWNEY: I think about this sometimes from a little bit of a philosophical standpoint of like what are prices? And like, what’s the point of this? And having a price that everyone can see and that is consistent and right in front of you seems like pretty what people want, right? I think if you just think about this from a separate standpoint of like, what do people actually want? They want to be treated fairly and on the same terms as everyone else. And that’s what you get when you have a price on something. And you can see the price and the label is there.

You can change the price, but it’s like the card in front of the thing and it’s a deliberate thing and it’s not personalized. It doesn’t mean you can’t have coupons and things like that. But the price here is in front of you. I know there have been articles about, oh, well, that’s inefficient because that means you have to have people in the store to go out and change the price. Well, that’s not the point here. Who cares? The point is this, having a price that everyone can see in front of them, that is something a lot of people want, right?

And so, if you just think about it from the standpoint of like, well, what do people want? They want to be treated fairly. And they don’t want to get ripped off. And they don’t want to be feeling like they’re being surveilled. These are all obviously negative things for the citizen.

Certainly, if you think about it just from the standpoint of the company, they want to do something else. But if you’re thinking about it from the voter, from the citizen, standpoint, what about just this era where we had laws about like having the price in front of you? Maybe this is too sort of neo-Luddite for you, Hal and Gavin, but I’m curious to get your thoughts on like, what about when we had just a price in front of everyone on a card?

GAVIN SICARD: Yeah, I think just my thoughts on that is you get to the point where like those are costs, those are economic costs. The discomfort of me not liking the idea that I’m getting an unfair price, that decreases my utility. That decreases my satisfaction of life. It’s similar to the idea of privacy.

The problem that we have as economists is that these things are very hard to put a number on. So, often they get minimized and pushed to the side. So, the fact that there are privacy costs to the increased use of surveillance in society doesn’t really get as much word count as like the idea that you’re going to nominally increase the efficiency of people getting the goods that they want.

The fact that people are going to be upset or the fact that it’s going to cause discomfort, distrust in society, all these things are like economic costs. The problem is it’s really hard to quantify them. So, it’s really hard for people to consider the trade-offs. And I think oftentimes we get the situation where we say, oh, this is something really hard to quantify. Let’s quantify it as zero. And then we ignore it completely. And I think that is a flawed perspective.

So, I think the idea is that legislators—and it’s a hard task—legislators have to consider these as real trade-offs. And simply because there’s the potential for—in the general economic equilibrium that they’re looking at, you maybe increase this market’s efficiency slightly. Is that really worth the other, maybe not as quantified, impacts that adopting this technology will allow?

TEDDY DOWNEY: I want to ask a little bit of the same question about you’ve mentioned surveillance in the piece as something negative people don’t want. You don’t want to create competition over how much and how many different ways you’re surveilled.

To me, this paints a picture of going to the store, the supermarket or your bodega or your drug store. And to me, I think about it a little bit differently in that these are third spaces where people historically have gone to run into each other in their community. And you’re turning this place that is for community, is for running into people, is for getting to know your neighbors, into a dystopic surveillance experience where you have this negative experience when you walk in. You’re not sure if you’re going to be getting ripped off. You’re not sure if you’re going to be exploited for having an irregular heartbeat or a high temperature or having a sick kid or whatnot.

This just becomes such a negative experience for you. And you have an incentive to not go in there, right? Like you’re going to have such an uncomfortable experience, you might be better off sitting at home, shopping on the internet with a VPN, pretending you’re in a poor country or whatever, right? This is sort of the opposite of what we have a society that is defined by increasing loneliness, increasing isolation, and you’re just exacerbating that potentially.

And I’m curious if the goal of these laws is to get you to compete, to create a better experience for people, or if it’s to let them, like you said, create this dystopic environment where they get the best price, but you get a bad experience. I’m curious to get your reaction on sort of these broader problems if you allow for this type of stuff to happen.

HAL SINGER: Yeah, I think it’s plausible that someone, if they understood that they were being surveilled inside of a physical store, their body temperature, for example, that might discourage them from wanting to go out. Now, my hunch is that people don’t realize the extent of surveillance pricing yet and how much they are being surveilled. I mean, this would be something that they’d have to come to learn. That would be a bad effect if it caused us to stop socializing.

But the real problem, Teddy, is that there’s really not a safe place for you to turn when you go back to your computer. I know you mentioned VPN. But these guys are monitoring all your prior purchases, what type of computer you’re coming with. I don’t know how many of these things that you can cloak just by going back into your office. So, there really is no escape from this world. And that’s why I feel like we need to tap the brakes on this experiment.

TEDDY DOWNEY: Let’s go to some listener questions here. We have one. What specific illegal conduct are the state laws trying to prevent data collection and use, raising prices based on that data? In other words, are these deception or unfairness statutes?

GAVIN SICARD: Yeah. So, generally, it depends on the state and the specific law. But generally, it’s targeted to the actual data collection. So, the Massachusetts law that didn’t pass, that we were writing about, was specifically trying to prevent the actual biometric collection of data in grocery stores.

So, grocery stores right now, they’ve started moving in this direction. I’m not sure how much have actually deployed this technology. But grocery stores are already tracking how long you walk through the store, how long you’re staying in the store. They might not quite be at the point of scanning your face, but that’s certainly in the direction that they’re moving. So, the first thing is that a lot of state laws are crafted in the idea of this is a type of data that is being collected, we are going to prevent that.

Secondly, a lot of states also then include provisions of just explicitly saying, you cannot use this collected data to personalize your price. I think New York says it’s pretty explicitly just like that. So, it’s kind of a double whammy. They both have laws that are restricting the collection of the data. And then they have laws that are restricting the ability for [firms] to use that data to then set prices.

I’m not quite familiar with the actual legalese of if it’s a deception statute or unfairness statute. I’m not sure. Hal might have better insight to that. But that’s like the overarching idea of what they’re actually blocking.

HAL SINGER: I’ll just make this point, Teddy, that even if the bills only target the pricing, conditioning pricing on certain personalized information, that by itself would undermine or at least lessen the incentive to want to collect data at a personal level and to invade one’s privacy. So, I feel like you could get at it indirectly by just banning surveillance pricing, that is the pricing itself, the activity of pricing.

TEDDY DOWNEY: And to your mind, being really specific about that being about collecting types of data to give an individual a specific price based on that data, as opposed to what we were talking about with dynamic and group pricing, it makes it a little bit more targeted, a little bit easier to enforce.

Another listener question here. If companies can use algorithms and personal data to estimate consumers’ willingness to pay, could employers similarly use worker data to estimate an individual’s willingness to accept a wage? And do you think the same legal or economic concerns around surveillance pricing would apply? We’ve been writing about this already, we already mentioned Uber with their desperation pricing for workers. Hal, Gavin, thoughts on using these tactics to set wages?

HAL SINGER: Yeah, it should similarly be banned on the wage side. And there’s a nice—it’s nice, but it’s also dystopic—story of Uber using personal information about drivers to set their wages. This is insane, and it ought to come to an end, both on the labor side and on the consumer side.

Look what the firms are doing here—not to pick on Uber, but they’re just such a good example—is they’re trying to extract as much surplus as they can on the consumer side and on the labor side, right? And so, if they can get a sense that because of your schedule or your pay history or how much money is in your Uber checking account, that you’re desperate to take a ride, that could cause them to pay you less than they would otherwise. And that sort of stuff should be banned as well.

GAVIN SICARD: Yeah, my thinking of this is that quintessentially, the ability to defeat price discrimination is the ability to have competition in a market. When you’re looking at it from the employer/employee relationship perspective, the labor literature is increasingly moving to the perspective that most employers have some amount of market power over their employees. It’s not very easy for an employee to say, oh, you’re screwing me over for my income. I’m going to leave and go to a job that better represents and better pays me. A lot of employees are constrained in their ability to do that. So, essentially the ability for a firm to pay their workers below what they are actually worth, it’s pretty readily available for a lot of employers because they have this market power. The only problem is that [employers] probably have a lot of unclarity about like how low they can actually drive that bargain. And I think allowing more of this technology maybe would allow them to push that [wage] level even lower.

Just to bring in some of the labor econ stuff, there is this idea of internal equity in labor economics, which is generally the idea that workers of similar caliber will be paid the same amount regardless of some of the individualized differences in their willingness to [accept a lower wage]. Which that might, the idea of internal equity, the idea that if I look over to my coworkers sitting next to me doing the exact same job and I realize they’re getting paid $15,000 more than me, I’m probably going to be wicked upset and that’s going to affect my worker productivity.

I think that maybe constrains the ability to individualize prices to the degree that a retailer might, but I’m not quite sure. I think there are at least some different principles that an employer has to deal with as opposed to a retailer.

TEDDY DOWNEY: We wrote an article recently on a company called PAVE, a compensation benchmarking service. I’ll just read you a quick clip from this. “It gives over 9,000 companies access to sensitive non-public salary and bonus data, platforms features, including monthly data refreshes and the ability to benchmark against customer peer groups could facilitate coordination among competing employers.”

Their webpage reads “benchmark pay against the exact companies chasing your people. Unlike traditional salary surveys, which typically rely on static periodically collected data, PAVE integrates directly with HR systems, aggregating real-time comprehensive compensation information. Customers can filter that data by location, industry, company headcount, job level and customer peer groups, which provides a granular targeted picture of competitors’ pay practices.”

To your point about it being personalized, I mean, it’s just a different—yeah, like you said, that the difference between the group and the personalized is a little bit less clear, in terms of you can have a law against personalized pricing, but they may not necessarily be doing that with personalized wages. It’s a little bit more general than that. It might not cover it perfectly.

GAVIN SICARD: Yeah, and obviously, like, beyond the personalized pricing topic, that type of conduct raises concerns about how aggregate collusion leads us to just monopoly-level or monopsony-level pricing. So, its competition that allows you to get paid what you’re worth, the more that firms collaborate to set their pricing together, set their wages together, the lower amount they can afford to pay you.

TEDDY DOWNEY: I want to ask actually to the extent that I actually have a hard time in terms of having my reporters work on these different things, privacy, algorithmic price fixing, personalized pricing. They tend to have a lot of overlap in terms of like when you’re investigating what’s going on. Do you think there’s anything to having an expertise and thinking about these issues as intertwined?

You mentioned privacy and personalized pricing as totally intertwined. It’s hard to even imagine enforcement of these laws without having an appreciation for them at some level. How do you think states, municipalities, Congress should be collaborating when it comes to understanding the interconnection between these three things, personalized pricing, privacy, and collusion?

HAL SINGER: Well, some of the practices could be challenged today under current consumer protection law. So, we don’t have to wait for a new law to challenge it.

Gavin and I are involved in a case right now. I shouldn’t disclose the defendant, but it involves telling one’s customers that they weren’t going to use a certain type of data of yours in a way to monetize and to make money. But, in fact, they did. And our job is to figure out what kind of damages the consumers felt as a result. The problem with litigating things, besides enriching testifying experts, is that it takes a long time. And it’s a random chance of whether or not the party is actually going to be made whole.

So, if this sort of conduct violates your sense of fairness—Teddy, you keep coming back to fairness, yes, which is an important thing. It’s, of course, lost on economists. But it’s an important concept that non-economists, i.e. humans, care about. We should just ban it, right? There’s no reason why we need to litigate each one of these things under antitrust laws for five years or under consumer protection law, which is slightly faster at two years, unless the goal is just to maximize billings for expert witnesses and law firms, which it shouldn’t be, right? It should be maximizing the welfare of society.

So, I feel if there’s something that’s happening today and you don’t like it and it’s a question mark as to whether consumer protection law would stop it or antitrust law would stop it, write a new law that directly makes it illegal. There’s no reason why we have to tolerate this.

GAVIN SICARD: I was just going to say that, at its core, personalized pricing is possible because of data. And the less amount of data that you have, the more infeasible it becomes. So, the idea of why firms have not been able to personalize price for most of their history is that you don’t really have the data to create an algorithm that can then determine what each individual’s willingness to pay is.

If you have very limited data on every individual you’re facing, you’re simply not able to determine the maximum price that they are willing to pay. The more and more data that you collect, the greater the ability of the firm to specify whatever price they believe will be maximizing your willingness to pay.

So, I think certainly people who are proponents of personalized pricing will rely on earlier literature and literature that is often ten years old or more to say that, look, personalized pricing is not perfect because they cannot specifically target how much this consumer is willing to pay for a thing.

I think, as technology has advanced, and as our intrusions on data and privacy have advanced, we’re increasingly going to be moving towards a world where perfect price discrimination becomes more the reality and there’s less error that the firms are making when determining the willingness to pay of consumers.

TEDDY DOWNEY: One of the things that I think is interesting about all of this is, when we do the investigating, a lot of the data that’s out there is bad data. A lot of the data that they’re accumulating and buying is actually not right. And so, to me, it just feels like an excuse to just be able to experiment with price and just charge people basically willy-nilly. If it’s accurate or not, I’m getting more money in the end, one way or another. I just need a little bit of it to be right.

I’m curious what you think like what kind of data—we’ve got two listener questions here that are related. What information do companies rely on for personalized pricing? How often are they buying data to be able to personalize the price? And I think inherent in those questions is how good is this data really? You’re talking about this perfect world, but it seems like we’re pretty far from that.

HAL SINGER: Well, I just want to say—and Gavin raised this issue with Delta. Delta, of course, did walk it back, but they still use the personalized pricing—this Israeli outfit, I think, called Fetcherr for a certain portion. And their CEO then takes to the airwaves and says that, as a result on that portion of flights for which we use this personalized pricing, our margins were much, much higher than they’ve ever been and higher than in a group of flights where we didn’t use this personalized pricing.

So, it’s very black box, Teddy, in the sense that we don’t know what Fetcherr I hope I’m saying their name right—is conditioning on. But whatever they’re conditioning on, they are bragging to investors and to would-be clients that this allows you—if you adopt our algorithm—it allows you to extract greater surplus and earn higher margins than you otherwise, would.

So, I mean, that’s also part of the problem is we don’t know what they’re getting into, what Fetcher is getting. But I think that the test of whether it’s a good thing from the firm’s perspective is whether their margins go up. And they’re boasting to us, they’re boasting to investors, that indeed their margins are going up when they’re surveilling us. And so, that, to me, sends a signal that something is wrong.

TEDDY DOWNEY: Gavin, any thoughts about the data being bought and used?

GAVIN SICARD: Yeah, I mean, it is a big business for a lot of these companies. So, obviously, some of the biggest companies in the world, like Google, their bread and butter is collecting data to sell to advertisers. The idea that you can then use that data to personalize price is going to be something I think that becomes more and more prominent.

I think it’s fair to say that, yes, we are not in the world in which perfect price discrimination is possible right now, because there is issues with the data. But the incentive of firms is, of course, the idea that personalized pricing maximizes your profit. The incentive is to get better and better data.

And I’m pretty confident that, as our technology advances, as the amount of surveillance advances in society, that data is going to improve over time. And this is a major business for even grocers. I think Kroger gets a third of its net profits off its data collection arm. So, the ability of Kroger to track consumers and to track the consumers online spending is a pretty big portion of the actual profit that they take home.

And then again, what actual data is being used? It is like a black box. But we’ve already had some examples. So, there’s a big case against Target that was settled in California and Target was using a GPS location. So, it’s using how close you were to a store to determine what your willingness to pay would be. I think we see, as we mentioned in the article, there’s this example of JetBlue. The customer service representative responding to a person complaining about the price of an airline ticket being spiked and saying that you should probably clear your web browser history and clear your cache.

I think there is the, instance there that, even though JetBlue walks that back, is that they are using a web browsing history. They even have the technology to see how long your mouse cursor is hovering over a certain portion of the page. I think they increasingly are trying to use any sort of data that they can to try and figure out what your willingness to pay is.

And I’m not an expert on data. So, I can’t tell you what that will ultimately be. But I do think that as we get more and more data and as these customer profiles get more comprehensive and then more correct, the ability to determine what you’re willing to pay for something will improve over time.

TEDDY DOWNEY: I want to push back on just this notion that it’s willingness to pay. Because you’re talking about concentrated markets. You’re talking about supermarkets. You’re talking about airlines. It’s how much rent can they extract? These are not willingness to pay. You know what I’m saying? Like, this is not a competitive market. And so, really what you’re just testing and getting, you don’t need the data. Like you know what I’m saying?

The data is almost secondary. It’s like, you can just test higher prices. And, yeah, I guess people like—it’s not like they can have another option. And so, because they have a lack of choice, a lack of competition, really what you’re doing here is just testing the upper limits of how much rent you can extract in many of these concentrated markets. I get it how it’s like, there’s some personalized—I get how it could be like willingness to pay in some respect if it’s a competitive market and you can go other places maybe. But like, because it’s a concentrated market, it just feels like I’m just finding the pain threshold for this person to otherwise just stay at home or what have you.

HAL SINGER: But Teddy, the best you can do, the best that a monopolist or any firm facing a downward sloping demand curve can do, is to line up everyone according to their willingness to pay and then charge each and every buyer one penny below. That’s the best they can do. That’s the most rent extraction or surplus extraction you can get away with.

And it raises an important point. I’m glad you did this. Because the CCIA guy and the Penn guy who testified in front of Congress last [week], they love to say that competition—markets are competitive and competition is going to save us. But, I mean, anyone who’s been paying attention—and Lindsay Owens, by the way, begins her book this way, pointing out how uncompetitive our industries have become. I think 75 percent of our industries have become more concentrated over the last few decades.

And the notion that—this is an important concept too—is that you can’t even begin to price discriminate, right? Unless you face a downward sloping demand curve. If you’re in a competitive market and you’re facing a horizontal demand curve, there’s no opportunities to price discriminate, right?

So, the very fact that these people have the ability to price discriminate is what we would call direct evidence of their market power. That is the very definition of market power, facing a downward sloping demand curve. So, the real problem is that we have too much market power in the economy.

And once that opens up, when you have the advent of these technologies, you really are in a position where you’re going to have this giant sucking sound, this consumer surplus is all going to be drained away by the consumers who are using this surveillance technology.

GAVIN SICARD: I would just say that when we say willingness to pay, that is in economics—it’s like the tangency condition. It’s at my willingness to pay, I will buy it. A cent above, I will not buy it. So, it’s my tangency condition of when I will go into a market and walk away from the market without buying anything.

And to answer your point about like, why isn’t it just like increasingly hiking up the price? Firms can do that. But as they get above certain individuals’ willingness to pay, they drop out of the market. And [firms] are constrained in the ability to do that because consumers dropping out can affect ultimately effect their profits.

So, in a profit maximization setting, they can’t just constantly raise prices because they’ll lose consumers. Personalized pricing allows them to charge each consumer that tangency condition, thereby extracting the maximum amount from every consumer in the market.

TEDDY DOWNEY: It’s like when Uber, getting back to Uber, it’s like when you have an Uber and you also have Lyft on your phone and you also have the Curb app on your phone, they’re giving you a lower price because you can just go on and check the other two.

Well, Hal, Gavin, this was an incredible conversation. Thank you so much for doing it. I’m looking forward to reading your work as you continue to cover these issues.

HAL SINGER: Thank you so much, Teddy.

TEDDY DOWNEY: And thanks to everyone for joining the call today. You can find that article I mentioned on PAVE in our platform. We’ll have this podcast up along with the other podcasts that we’ve done on personalized pricing and algorithmic price fixing. So, thank you to everyone for joining us today. This concludes the call.

HAL SINGER: Thanks for having us.

TEDDY DOWNEY: Thanks, Hal.

GAVIN SICARD: Thanks so much.

TEDDY DOWNEY: Thanks, Gavin.