Behavioral profiling allows advertisers to offer goods at different prices, what economists call price discrimination, to extract the maximum price from each

140 KB – 18 Pages

PAGE – 1 ============
How Big Data Enables Economic Harm to Consumers, Especially to Low-Income and Other Vulnerable Sectors of the Population The author of these comments, Nathan Newman, has been writing for twenty years about the impact of technology on society, including his 2002 book Net Loss: Internet Profits, Private Profits and the Costs to Community , based on doctoral research on rising regional economic inequality in Silicon Valley and the nation and the role of Internet policy in shaping economic opportunity. He has been writing extensively about big data as a research fellow at the New York University Information Law Institute for the last two years, including authoring two law reviews this year on the subject. These comments are adapted with some additional research from one of those law reviews, ” The Costs of Lost Privacy: Consumer Harm and Rising Economic Inequality in the Age of Google William Mitchell Law Review . He has a J.D. from Yale Law School and a Ph.D. from UC- Executive Summary: institutions organizing information not just about the world but about consumers themselves, thereby reshaping a range of markets based on empowering a narrow set of corporate advertisers and others to prey on consumers based on behavioral profiling. While big data can benefit consumers in certain instances, there are a range of new consumer harms to users from its unregulated use by increasingly centralized data platforms. A. These harms start with the individual surveillance of users by employers, financial institutions, the government and other players that these platforms allow, but also extend to to discriminate and exploit consumers as categorical groups. 1.Behavioral profiling allows advertisers to offer goods at different prices, what economists call price discrimination, to extract the maximum price from each individual consumer. Such online price discrimination raises prices overall for consumers, while often hurting lower-income and less technologically savvy households. 2.Behavioral profiling is used by especially seedy companies to target a variety of financial and economic scams at vulnerable populations most likely to fall prey to their offers 3.Examples include subprime mortgages targeting vulnerable consumers with worse deals based on racial and economic profiling. 4.Advertising- and payday lender advertisers exploiting financially distressed households in the wake of the financial crisis. B. User data is economically valuable, yet big data platforms manage to extract data from users with little financial compensation. 1. provided content on the web (search, videos, reviews on Amazon, shared social media content) and encouraging users to provide private data without compensation. 1

PAGE – 2 ============
2.U data is being shared with third parties. 3.Lack of competition means that there are not alternative services offering to share the those users, so the economic value of content & data flows largely for free to the big data platforms. Users are largely disempowered from demanding protection of their privacy, thereby increasing the flow of user data to the data platforms and advertisers. C. To deal with these consumer harms, regulators should implement a combination of strengthening individual user control of their data, structural changes in the market to encourage a more accountability to consumers in the marketplace, and public interest regulation of the larger big data platforms to ensure that they are held accountable, particularly in the realm of financial services, in areas where the market will not discipline their actions. D. Information asymmetry between big data companies and consumers is easily converted into economic inequality when one side of every transaction has so much more knowledge about the other during bargaining. The increasing information asymmetry in consumer markets, driven by data mining and facilitated by online services, may be an additional significant cause of this overall increase in economic inequality we have seen over the last four decades. Introduction t used by corporations to reshape markets and increase their market power and profits. 1 On the Internet, we see the accumulate ever increasing information on consumer behavior, interests and needs. While this data unquestionably increases the efficiency of the economy in numerous ways, what is in question is whether consumers are ultimately benefitting significantly from those productivity gains or whether that surplu leaving them vulnerable to economic exploitation by a range of corporate actors. These big data platforms, what Jaro Who Will Own the Future? 2, attract consumers with a variety of services that encourage those users to part with personal data, which in turn is analyzed and combined with private information from other users in massive networks of computers. These companies use that analysis to reshape markets and channel an ever greater share of economic wealth into the hands of these big data platforms. There is a particular concern t benefit not of those users but for third party corporate customers of those data platforms, particularly advertisers who drive a large portion of the revenue model of the online Internet economy. While much of that advertising no doubt serves traditional advertising goals of strengthening brand awareness or promoting new products to consumers, the rise 2

PAGE – 3 ============
of behavioral profiling of consumers using the private data extracted by these big data platforms increases the use of advertising for more exploitative practices. Big data platforms facilitate advertisers engaging in user profiling that aids those companies in extracting the maximum profit possible from consumers in the overall economy. Advertisers can deliver ads not just to the users most likely to be interested in the product but can tailor prices for individual consumers in ways that can maximize the revenue extracted from each purchaser. Consumers can be profiled and offered higher prices, unaware that other customers are getting better deals, while financially struggling houses are tagged as vulnerable and offered economically exploitative services such as payday and subprime loans. Since the rise of big data has coincided with the stagnation of incomes for average households, policy makers should be raising concerns that, alongside traditional explanations of rising inequality such as deunionization, globalization, and automation of unskilled jobs, the concentration of data into ever fewer corporate hands is helping to drive economic inequality in the broader economy. While big data can benefit consumers in certain instances, regulators need to take action to address new consumer harms to users from its unregulated use by increasingly centralized data platforms. The Federal Trade Commission has itself highlighted some of these problems in a number of recent reports, 3 as well as litigation against companies engaged i n deception in collecting personal data, 4 but it is clear that additional regulation and laws are needed to address the full scope of the harm to consumers. These harms start with the individual surveillance of users by employers, financial institutions, the government and other players that these platforms allow, including denial of employment or scholarships based on what people post to their personal social media sites. While a few states have taken action to restrict misuse of social media data to directly, there is a broader need for federal action. However, such individual surveillance is less of a danger to consumers than the broader aggregation of data so-called –and the ways it allows companies to discriminate and exploit consumers as categorical groups. Much of this profiling is invisible to consumers, making the need for public action all the more urgent and needed. Big data platforms collect so much information about so many people that correlations emerge that allow users to be slotted into marketing categories in unexpected and often unwelcome ways. Increasingly, every transaction, every website viewed, and every action online generates a data trail swept into the data platforms online. Most websites invite dozens of companies to track users on their site and follow them across the web. It is largely because of the ability to profile users and more precisely target ads that online advertising as a whole has exploded and become the largest advertising sector in the United States. In fact, 2013 was the year Internet advertising surpassed broadcast advertising revenues in the United States for the first time. 5 Online advertising amounted to $42.8 billion in the United States 6 and $117.2 billion globally 7 that year. 3

PAGE – 4 ============
What Advertisers Get for Their Money: Knowing What People Will Buy and the Price At Which They Will Buy It The question is what do advertisers get for their money? No doubt, user profiling helps advertisers more effectively identify the customers most likely to be interested in their products. However, the darker explanation is that such profiling also facilitates tailoring prices to individual consumers in ways that maximize revenue extracted from each transaction. This ability to charge different prices to different customers for the same good or service, different maximum prices they are willing to pay. And profiling consumers helps in point customer matching that maximum price they are willing to pay without them knowing that other deals are available. 8 Some economists argue that where consumers know all pricing options, they can potentially benefit from price discrimination, as when airline passengers choose between a cheap price at an inconvenient time to save money, which can fill seats, increase revenues for airlines and increase options for different customers. 9 But when peo such price discrimination is far more likely to hurt consumers. For example, a 2012 Wall Street Journal report found that major companies, including Staples, Home Depot, Discover Financial Services and Rosetta Stone, were systematically using information on user physical locations to display different online prices to different customers. 10 More disturbingly, contrary to any hope this might benefit low-income bargain hunters, the paper found that higher-income locations were offered better deals than low-income communities, because those poorer areas had fewer local retail outlets competing with the online stores. Credit card companies like Capitol One show different offers with different credit card deals based on view locations and guesses by the company about their income.11 In search advertising, this differential pricing overwhelmingly takes the form of web coupons offered to some people but not others based on their behavior and demographic data. As Ed Mierzwinski, consumer program director for the United States Public Interest desirable products than they offer me, or offer you the same product as they offer me but at 12 Economists like Nobel Prize Winner Joseph Stiglitz, who pioneered what has been called pricing. When consumers don 13 4

PAGE – 5 ============
Big Data Platform Price Discrimination Increases Prices Overall for Consumers Many had a vision of an online economy where consumers could quickly compare prices, but studies have shown that hidden discounts, the posting of multiple versions of the same product, an keep prices up. 14 Where prices are obscured and sellers impose price discrimination, economic models generally show that overall prices in the economy will end up higher than any model where consumers knew all prices. 15 This argument is not one initially made by critics of the online economy but has actually been made by boosters of the opportunity for companies to profit from it. Academic Hal Varian has a long history of examining various models of price discrimination and in 2005, he was appointed Chief Economist for Google. That same year, he co-authored an article in the industry-based academic journal Marketing Science touting the gains for companies engaging in online price discrimination, particularly against what the authors labeled prices for them. Varian and his coauthor argued technol personal information to make price discrimination profitable. 16 In a foreshadowing of both lock-in users to particular services, block anonymous participation, and seek out the coveted 17 While various economic models yielded – that in many cases, any economic value added to the economy due to increased efficiencies actually falls overall. 18 s generally lose out financially under price discrimination using targeted consumer profiling. Recent research on online advertising reinforces this analysis of consumer loss due price discrimination combined with consumer profiling. Comparing traditional regimes of mass- market advertising to online advertising, researchers Rosa-Branc Esteves and Joana Resende found that average prices with mass advertising were lower than with targeted online advertising. 19 Similarly, Benjamin Reed Shiller found that where advertisers know consumers willingness to pay different prices, companies can use price discrimination to increase profits and raise prices overall, with many consumers paying twice as much as others for the same product. 20 Big Data Platforms Enable Racial Profiling and the Exploitation of the Most Economically Vulnerable Groups in Society Once upon a time, people celebrated the Internet as promising a new era where shoppers invisible on the web could not be judged based on their race or otherwise discriminated against. However, online behavioral targeting can combine a home address and a few more 5

PAGE – 6 ============
characteristics to create an almost perfect proxy for race. If anything, such online discrimination can be more vicious for its subtlety and invisibility since c even know what prices are being offered to other people of different races or such harms if they could be made visible, since as George Mason University professor 21 discrimination based on location discussed above, companies like Wells Fargo listing houses for sale have collected zip codes of online browsers and directed those buyers towards neighborhoods of similar racial makeup. 22 This online discrimination parallels the broader reality of companies like Wells Fargo illegally steering an estimated 30,000 black and Hispanic lenders from 2004 to 2009 into more costly subprime mortgages or charging them higher fees than comparable white borrowers. 23 As ColorLines magazine has not (including US census, education, population, STD stats, and state financial data) presumably could also be folded into the personalized search algorithm to surmise a lot more than your race.” 24 Latanya Sweeney in an academic article describes how on sites detailing legal information about individuals, when people searched for a name “on the more ad trafficked website, a black-identifying name was 25% more likely to get an ad suggestive of an arrest 25 What is disturbing is that people online can find themselves losing opportunity as their p in the algorithms of big data platforms. For example, Kevin Johnson, a condo owner and businessman, found that after returning from his honeymoon, his credit limit had been lowered from $10,800 to $3800. The change was not based on anything he had done but, according to a letter from the credit card company, he had shopped at stores whose 26 If your habits associate you with particular categories or groups, you will invisibly find opportunities opening up or closing down based on how data algorithms choose to place you. Similarly, whether you get a refund when making a complaint to a company will often be heavily influenced by the categories in which data analysis places a caller. For less ethical companies, big data gives them the ability to seek out the most vulnerable prospects to exploit and entice them with scams and misleading offers. Such niche scams and economically exploitive relationships can be focused on those most vulnerable to the while remaining essentially invisible to everyone else, including reporters and researchers trying to evaluate the harms from online advertising methods. The data broker industry even has a term for the poor, old and less educated groups that they compile for such unethical marketers. For example, people who reply to sweepstakes offers are put onto a list by one data broker company and offered to 6

PAGE – 8 ============
specific revenue from particular companies, but reports at the time showed that mortgage loan companies were paying to prices going for as much as $20 to $30 each time a user clicked on a search ad. 34 Online companies would then sell information about the users identified as likely prospects to mortgage companies, which in turn would contact them. Customers targeted through these online leads for subprime mortgages were disproportionately low income, black and Latino. Usually unaware that better deals existed, studies showed that people of color offered these subprime mortgages were 30% more likely to be charged higher interest rates compared to white borrowers with similar credit ratings. 35 Burdened with unrealistic in later years. 36 This was the most toxic version of price discrimination possible and led to one of the largest scale destructions of wealth among low-income and minority communities in the modern era 37, even as the data platforms that helped facilitate this process continued to explode in revenue and profitability. Even today, the financial industry remains bedrock of revenue for advertising-driven big data platforms. According to WordStream, a company specializing in helping companies bid effectively on Google Ads, the three most expensive categories of keyword searches as measured by cost per click are in financial services (insurance, loans and mortgages), with 45.6% of the top 10,000 advertising keywords falling in those categories. 38 Depressingly, bottom-feeding subprime mortgage offers were replaced in the aftermath of the financial crisis by companies exploiting the financial distress of families, particularly by payday loan lenders who offer extremely high-interest loans in exchange for a commitment for repayme 39 Such loans have been banned or severely restricted as exploitative in multiple states and the Consumer Financial Protection Bureau (CFPB) has held hearings specifically on abuses in the industry, with CFPB head Richard Co 40 Their ubiquitous presence in online ads is not an accident; in fact, data platforms have actively solicited ads from the industry, including trade group made up primarily of payday lenders. Industry observers like Robert X. Cringely, who has covered Silicon Valley for over twenty-five years, argue that Google 41 Whether or not, as Cringely argues, data platforms do hide negative information about the evils of many of their online financial advertisers, what is true is that they proliferate in the feeds of low-income Internet surfers. As many families saw their mortgages balloon above promising to help homeowners in advertisements appearing when people searched for helping them at all. Despite scathing reports highlighting the problem by consumer group Consumer Watchdog in 2011, 42 Google refused to stop until ordered by the Treasury 8

PAGE – 9 ============
homeowners who fall prey to these scams, initially do so through these Web banners and other We Asset Relief Program (TARP), said in an interview. 43 Similarly, the data broker and credit score company Equifax kept selling lists of people late in paying their mortgages t o fraudulent marketers until the FTC fined Equifax $1.6 million in 2012 for the practice based on companies bilking those customers of millions of dollars. 44 In this way, the data and privacy lost by consumers has translated into tens of billions of dollars in profits for the data platforms and the enabling of exploitation by predatory companies using that data for an even larger scale of economic losses by consumers. Consumers Lose Financially as the Value of their Personal Information Flows to Big Data Platforms Beyond losses from price discrimination and from direct scams using targeted data, consumers lose out online as the value of their personal data is coopted for the profits of the big data platforms. In a broad sense, users lose out doubly since the data platforms not only sell their data to advertisers but also use the free labor and data provided by all users collectively to attract users to their sites in the first place. While users may vaguely feel that they are giving up some control of their data in exchange for services provided by big data platforms like Google, Facebook, those companies actually depend on the free labor of individuals posting their updates to Facebook, reviews on Amazon, edited stories on Wikipedia, and their links on blogs to make their services valuable. diffuse labor of people across the Internet. Its original Page Rank algorithm used the links to other websites created by web site creators across the Internet as a tool to assess and rank the likely value of websites containing similar information or keywords, an algorithm which has only been strengthened by tracking the sites for which people search. 45 Each click adds to the algorithm that can direct users with similar searches and interests to see the same link highly ranked as well. its search engine vis a vis any challenger search technology which would lack access to the network of users and the information they generate on search preferences. 46 Similarly, social networks like Facebook and LinkedIn depend on the daily infusion of writing and links by their users to provide value to other users and use experiments on the behavior of those users to strengthen their algorithms. 47 These big data platforms have positioned themselves to take advantage of what media studies professor Clay Shirk aggregated. 48 That so much seems free on the Internet is just the flip side of people providing so much free labor without being paid themselves and then accessing it on big 9

PAGE – 10 ============
free labo 49 Recognizing the economic gain from user production of information, companies like Facebook go out of their way to encourage the maximum sharing of data possible, using what m -like mechanisms to reward family when they post. 50 The incredibly outsized stock valuations of web-based firms such as Amazon (with its user- generated product reviews), Facebook (with its user-generated content and social links) -generated YouTube videos etc.) can best be understood in terms of the free labor and data each is harvesting. For example, when Facebook went public with its initial public offering (IPO), one analyst estimated that users 2011, which translated into about $100 billion of the value of its stock market capitalization –with each Facebook user contributing around $100 of user labor to the stock wealth created for Mark Zuckerberg and his fellow shareholders. 51 Analysts like Michael Fertik, CEO of the company, which helps keep their information anonymous online, estimates that data can be worth in the thousands of dollars each year to all the data platforms a consumer may use. 52 One other measure of the value of use data is the fight Apple had with publishers over terms for sales of subscriptions iTunes; most were willing to pay 30% of their subscription price to Apple but balked at Apple retaining control of data on subscribers, indicating that publishers valued the user data at more than 30% of the cost of any purchase online. 53 Yet while publishers negotiate hard over control of that personal data with Apple, the consumers themselves generally give their personal data away for free without a thought. Consumers underestimate the value of their data and lose out continually in these online transactions. De facto they are in a barter relationship with big data platforms, trading data and the economic history of barter is that less sophisticated partners in such exchanges inevitably lose financially. 54 information used by big data platforms are also being shared with third parties to assist in marketing advertising. 55 Users rarely read the fine print when they click acceptance of the terms of service on these sites and receive little information about the consequences of preferences: a 2012 Pew survey found 73 percent of the American public were opposed to search engines even tracking their search history even to improve search results and 68 percent opposed using user data to help advertisers target ads. 56 Users who understand that such sharing is happening express frustration that they lack the capacity to stop it 57, even though the desire to stop such tracking, aggregating and sharing of data has been increasing. 58 Notably, the former Federal Trade Commissioner J. Thomas Rosch expressed concern in his 2013 opinion for example, that its gathering of information about the characteristics of a 10

PAGE – 11 ============
or near-m 59 any user for whatever purpose the companies chooses to use their data. The problem is that if few users know how the data is actually being used, such consent is meaningless. extremely unclear how users can be in a position to effectively negotiate a fair economic Industry Concentration Decreases Market Pressure on Data Platforms to Prevent Economic Harm to or to Share Economic Value of Data with Users In too many data platform services, one company is so dominant that consumers have little leverage to demand greater control of their data and less harmful use of that data. Whether Google in search adverting, Facebook in social networks, Amazon in online retail, Netflix in video streaming, the dynamics of control of user data strengthen concentration in particular sectors. Part of this are network effects that mean the more people participating on a service, the more valuable it is to other users of that service. Part of the drive to concentration is that as companies collect user data, they gain competitive advantage against any potential challenger who will lack that user data in setting up any rival service. Such data can be redeployed by dominant players not just to strengthen their position in existing services but used in related new services to expand their economic reach. In this way, you see Google, Amazon, Apple and Facebook expanding rapidly into a multiplicity of emergin g data-related fields, making it extremely hard for upstart companies to get a toehold except in very specific niches. 60 The upshot of this dynamic is that the marketplace is doing little to create options for consumers that might alleviate the misuse of consumer data, better protect user privacy or encourage big data platforms to better compensate users who are willing to share their data. There has recently been a flurry of political interest in abusive practices by data brokers who buy and sell personal data, with major reports released by both the Senate 61 and the Federal Trade Commission. 62 While the consumer harm detailed in those reports are important, it is worth noting that the companies involved are relative minnows in the big data ecosystem compared to the major big data platforms and are likely to be even more marginal over time. Experian is one of the largest at $4.8 billion in sales per year 63 while Acxiom, a data broker often cited as having one of the largest datasets on consumers, has only about $1 billion per year in revenue. 64 Even collectively, these data brokers are dwarfed by a company like Google with over $60 billion in annual revenue. 11

140 KB – 18 Pages