Big Data, Differential Pricing and Consumer Empowerment
Last Wednesday, the White House released a report on “Big Data and Differential Pricing.” The reason you have probably not heard much about it is that it was about as exciting as an Economics 201 lecture. If you have heard about it (and you are not tenured faculty in an economics department), you may have meandered across an article such as this one, whose headline seems to imply that there is a problem that regulators need to solve.
However, if you take time to read the report, the conclusions (to the extent there are conclusions) are far more balanced and restrained. In fact, the report notes that price differentiation is not widespread, usually beneficial, and certainly not unique to the online world. Where differentiation is potentially problematic (not necessarily for economic reasons, but for matters of “fairness”) is in risk-based pricing markets (such as insurance, employment or credit-issuance). But in risk-based markets, the takeaways are not unidirectional either. In these scenarios, differential pricing is often a good thing, as it can discourage risky behavior upfront (and guard against adverse selection). Indeed, the report recognized these types of benefits. For example, if health insurance companies offer lower rates to non-smokers or generally healthy individuals who eat-well or exercise regularly, this can encourage healthy behavior upfront. The report recognizes other benefits as well, such as expanding output. To again use the health insurance market as an example, if health insurers couldn’t differentiate pricing and had to offer uniform prices, those who engage in risky behavior (e.g. smoking) would be forced to subsidize those who engage in non-risky behavior, causing fewer low-risk individuals to purchase health care plans. Even in non-risked-based markets, differential pricing often leads to more efficient outcomes (particularly in high-fixed cost markets).
Real problems can arise when factors outside one’s control (genetic predispositions) or protected classes (race, sex, age, sexual orientation, etc.) are used to directly disadvantage consumers. However, these problems are not unique to the online/big data world, and — as the report points out — antidiscrimination provisions of existing laws (such as the Fair Credit Reporting Act and the Civil Rights Act) already apply in these situations.
Nevertheless, some voice Orwellian fears that Internet-enabled big data might allow sellers to more precisely gauge individuals’ willingness to pay, thus transferring wealth (i.e. “surpluses” in economic terms) to producers. This is “first degree” or “perfect” price discrimination in economic textbooks, which is hardly elaborated upon outside of theoretical models because it is virtually impossible in the real world, and — despite overblown concerns — still nearly impossible in the online world as well.
This stylized view of the Internet economy is highly problematic and proves entirely unrealistic, even in a world where big data technology is rapidly advancing. For one, no matter how good the “big data” is, it still amounts to the use of proxies to determine a customer’s willingness to pay, and these are extremely imprecise. Even if a seller knows where I live based on my IP address (which can easily be circumvented by using free VPN services), my purchase history, what articles I like to read, etc., it is still a big leap to figure out what an individual consumer is willing to pay for a particular product. This explains why, as the White House report notes, such extreme differential pricing strategies are rarely used, and to the extent they are, still largely focused on estimating the demand curve for products — which is a common practice in the offline world as well.
For argument’s sake, assuming that it is feasible to know exactly how much each consumer is willing to pay for a product, the “perfect” price discrimination concern belies the fact that consumers have brains and use them to think. In fact, some consumers are very sophisticated, and the report recognizes their strong ability to mitigate potential price differentiation — think bulk eBay sellers that find abnormally low prices and resell the products online. This process, known as arbitrage, involves consumers who receive low prices reselling the product, thus making this type of discrimination strategy infeasible. [In fact, there was even a Supreme Court case about this a few years ago. The case revolved around intellectual property issues but supported the position of the “arbitrageur.”] Furthermore, developing and implementing sophisticated price discrimination schemes, even in the age of big data, is costly and thus undercuts much of the benefit of heavy reliance on such schemes. And, moving away from purely economic models, if a company is exposed as charging wildly different rates to consumers for non-differentiated products, the negative press it would receive would not only reflect badly on the company and drive customers away, but would push the customers that stay to employ more sophisticated schemes to circumvent price discrimination strategies that require them to pay more. And this is precisely where Internet innovation can help.
What the report also takes important note of (here) is that technology, particularly Internet technology, also gives consumers better tools to circumvent sophisticated price discrimination strategies. Specifically, Internet platforms lower the transaction costs in acquiring information and exchanging goods, costs that have traditionally paved the way for price discrimination. Because, in a pre-Internet world, it was costly to compare prices across different geographic markets and to conduct sophisticated searches to determine the range of different pricing options available, sellers could more easily charge different prices to different consumers based on where they lived (or target other cross-sections of consumers with higher prices). Besides eBay and other online auction sites, the phenomenon is most easily observable in travel markets.
Airlines are perhaps the most obvious example of companies that have been using price discrimination for decades. Since business travelers are often less cost sensitive than recreational travelers, airlines uses proxies to attempt to charge these travellers more, such as last-minute ticket purchases, extremely expensive business class options, frequent flyer programs or whether you are traveling over a weekend or not. Furthermore, the “search costs” associated with comparing the pricing of different flight options used to mean that all travelers were often in the dark about cheaper fares that involved only a slight alteration of their original travel plans. Despite advances in airline algorithms, online innovation in travel search has greatly empowered consumers to fight back. Now with a few clicks of a mouse, on sites like Kayak and Skyscanner, consumers can see nearly all the palatable options and choose the best (and often cheapest) option for them. Besides making the travel process way more consumer friendly, it transfers much of the “surplus” back to the consumer. In fact, entire Internet companies are built around empowering consumers to circumvent long-existing sophisticated differential pricing strategies.
If there is a meta-narrative about the rise of internet technology and its relationship to economic surplus, it’s not the transfer of wealth away from the consumer, but the huge surplus transferred to the consumer (e.g. online media and electronic bookstores). In fact, search technology alone, which allows consumers to easily navigate to the lowest prices offerings, accounted for nearly a quarter of a billion dollars in additional consumer surplus 5 years ago. (That’s $20 a month for the average U.S. or European consumer.) It is highly likely that that number has increased drastically since then. So, if you zoom out and look at the big picture, not only are most instances of differential pricing economically efficient, but the Internet does far more to empower the average consumer than it does to harm her.