Pages

7 April 2018

How to fix competition in a market owned by Google, Facebook, Amazon and other digital platforms

Anupam Manur

Almost every digital major out there – from Google to Amazon, Ola to Swiggy – plays both platform and player to expand their reach and market. This itself is not anti-competitive but it becomes unfair practice when the platform takes benefit of data of others on the platform to its benefit. India does not have an adequate legal framework to deal with this yet but it can explore different options even if it means moving from textbook regulation. 


Whether we like it or not, there is a certain inevitability about the dominance of technology companies in our lives. Apps for ecommerce, social media, ride-hailing services, search engines, food delivery, payments, and several others permeate our lives. They are also highly successful companies for the ease of use and the convenience they bring into millions and millions of lives. However, apart from being ubiquitous and indispensable in most of our lives, some of like Amazon, Google, Facebook, and Uber share something else in common: they are accused of anti-competitive behaviour. Regulators regularly fine them, the CEOs and top management are summoned by lawmakers for questioning, and even technology enthusiasts and one-time supporters are now alarmed at the power they hold, the latest being the Facebook-Cambridge Analytica breach in the broader context of the 2016 US elections.

Regulators across the world are often two steps behind the rapid pace of innovation set by the market players. Enter the digital economy and the gap is even wider. The trusted method of checking for price and output to determine market dominance and concentration of market power, however, will not yield the complete picture in this space. The term anti-competitive behaviour can come to mean an entire range of practices. For example, the European Commission for Competition recently released a detailed reportelaborating on the different ways in which online platforms might abuse their market power: the use of price restrictions by online platforms; asymmetries of bargaining power; vertical integration and leveraging; mergers and acquisitions; and the role of data in these markets and its impact on competition enforcement. While each of these concerns has to be studied and addressed in India, this article will focus on the anti-competitive nature of “vertical integration” in platform markets.

Platform business models, which use technology to connect an otherwise disparate set of buyers and sellers, are increasingly common and lucrative in the digital economy. Uber connects consumers and drivers, Airbnb connects travellers with home and lodge owners, Amazon links buyers with multiple retailers, and even Facebook, Twitter and Apple open their ecosystem and potential customers to many third-party software developers. The concern here is that many of these tech platforms could use their market power to engage in “vertical integration”, whereby the platform which acts as a marketplace also acts as a competitor on the very platform. E.g.: Amazon, the marketplace, and Amazon, the retailer, on the very same marketplace.

Just to be clear, the platform acting as a competitor by itself is not a case of anti-competitive behaviour. However, the issue is whether the platform abuses its market power to favour itself as a competitor on that very platform.

Just to be clear, the platform acting as a competitor by itself is not a case of anti-competitive behaviour. However, the issue is whether the platform abuses its market power to favour itself as a competitor on that very platform. Does Amazon regularly display products sold by Amazon itself on top of search results?

To illustrate this: Take a typical smartphone, where third-party developers who did not produce the smartphone often provide apps. The phone acts as a platform where users and app developers interact within sub-markets. Now, imagine if the next smartphone you bought only had access to apps that were proprietary software developed by the manufacturers of the smartphone and it was made difficult to search for and install third-party apps. Surely, that will limit your choice and diminish the value that you derive from the smartphone.

Microsoft: Roll back the clock 20 years. It is 1998 and Bill Gates, the CEO of Microsoft, the biggest software company then, is on trial in a US District Court for anti-competitive practices. The United States versus Microsoft Corporation is a landmark case in anti-competitive practices on platforms. The Department of Justice sued Microsoft for illegally thwarting competition in order to protect and extend its software monopoly. Of particular interest here is that Microsoft forced computer makers to include Internet Explorer as the default web browser in the host Windows operating system. The Windows OS was seen as a platform where third-party software developers could compete but the bundling of Explorer was responsible for Microsoft’s early victories in the browser wars, as every Windows user had it as the default web browser.

Remember the context – in 1998, it was expensive and slow to download a rival web browser, Netscape or Opera, over a modem. It was also alleged that Microsoft manipulated or altered its APIs to favour Internet Explorer over third-party web browsers. The court ruled against Microsoft.

Uber, Ola: There are instances of platforms abusing their market power to favour their own retail products or companies right around us today. E.g.: the recent indefinite protest in India by drivers partnered with Uber and Ola, the cab aggregator companies. One of the main issues, apart from incentive systems and commission fees, is that the cab aggregator apps are in violation of a newly coined term – “cab neutrality”. The two companies charge around 26% commission or service charge for drivers acquiring customers through their apps. However, they are seen to give first preference to cabs owned or leased by the aggregators themselves as opposed to cabs owned by the drivers.

Uber and Ola bought the cars themselves and leased it out to drivers to increase the number of cabs on their platforms. However, the drivers did not earn enough on a consistent basis to pay back the companies. One can readily imagine how the aggregators would be incentivised to tweak their algorithm in a way that favours a cab that they own over a third-party partner.

Google: Tomes have been written about search giant Google’s market share and dominance. One aspect to focus here is the violation of the neutrality of platforms. In 2017, the European Commission for Competition slapped a $2.7 billion fine on Google for this very reason.

Before 2007, if one were to search for a particular product, say a mobile phone, Google would show results from other sites that listed these products. The sites with lower prices and the more relevant listings got ranked higher and were placed at the top of the search page. However, as the “comparison shopping sites” got more successful, things changed. Google makes money by placing ads next to its search results and it would be difficult to place an ad – say, from a mobile phone maker – next to a listing that displayed significantly lower prices for the same product.


Then, Google came up with its plan, which was later detailed in a Federal Trade Commission report. First, it worked on lowering placement of rival comparison shopping sites, which was estimated to reduce traffic to them by about 20%. Then, Google started displaying prices from its own shopping service, known as Google Product Search, at the top of the page. Then, Google shut down its product search service and introduced Google Shopping. Google Shopping did not display the lowest price for the product — rather, it displayed ads at the top of the search results page in response to the user’s search term. According to Gary Reback, a Silicon Valley antitrust lawyer in this Washington Post article, “the ads were carefully placed by Google’s algorithms to minimize price competition among merchants, by, for example, showing ads next to each other that featured different product models at different price points”.

And the last part of the plan was to allow merchants to buy these ad spots. Thus, merchants were now promoted due to their spending on ads rather than their price and relevance. A study by Consumer Watchdog reveals, “consumers buying a product located through Google Shopping could pay as much as 67 percent more than if they had made the purchase using information from a competing comparison shopping engine.”

Swiggy: “It definitely bothers a restaurant owner as on the one hand you call us a partner and on the other hand you start something which is in direct competition with us,” quipped a restaurateur in Bengaluru against the online food ordering and delivery platform Swiggy in a FactorDaily report. The food delivery start-up allows a wide variety of restaurants on its platform and takes care of delivery, thereby increasing the choice and comfort of its customers. In return, it charges 15%-30% commission from the restaurants and a delivery fee from the customers.

But with its own ‘cloud kitchen’ called The Bowl Company, things have changed — Swiggy is in competition with its platform members. Then, follows the typical abuse of platform power by the company. Users suddenly found The Bowl Company in the top of their recommended restaurants. Also, as published in a purported blog of ex- and current Swiggy employees in July 2017, “Instead of growing our restaurant partner’s businesses, we recently took the best business zone in Bangalore (Koramangala) and started intentionally routing all the users to order from Bowl Company… This just directly hits at the heart of restaurants we “partnered” with to grow our business in the first place.”

Swiggy CEO Sriharsha Majety responded that all the platform aimed to do was address market gaps. “We believe everything starts with the consumer and as long as we don’t solve a consumer problem — we don’t embark on any of these initiatives with a cannibalization approach. We do not intend to create properties that compete with existing restaurants,” he said in a company blogpost.

In Bengaluru, The Bowl Company serves the well-selling items such as idlis, pongal, butter chicken, rajma chawal, chicken biriyani etc. (as also items not easily available like peri peri chicken and granola). Swiggy’s access to historical user data and preferences would make it easy to customise its menu based on the highest-ranking dishes. The same would apply to Zomato or other foodtech platforms.

Amazon: In 1994, Amazon was incorporated in Seattle, Washington as an online bookstore. Twenty-four years later, it is the world’s largest online retailer in terms of revenue and market capitalisation. There are few products that one can’t order off Amazon – it currently sells nearly 120 million different items. Many of these are sold by third-party retailers, who pay Amazon a fee to sell on its platform. Other goods are sold by Amazon itself as a traditional retailer. Amazon’s amazing internet reach puts retailers in touch with millions of customers worldwide that they could not have hoped to reach on their own. However, this exposure comes at a cost – they are competing with Amazon itself.

Does this dual role as platform and player create a conflict of interest for Amazon? Yes. Just like the aforementioned cases, Amazon the marketplace has been consistently accused of unfairly favouring Amazon the retail company. A point to note: unlike Uber or Swiggy, which started off as a marketplace and then became a competitor, retail was the primary business of Amazon before becoming a marketplace.

The third-party retailers felt that Amazon used the marketplace as a lab for its own retail company. It would let retailers innovate and compete against one another, and then cherry-pick the best products for themselves and capture the value. This was confirmed in a study by Harvard Business School Feng Zhu and Qihong Liu of the University of Oklahoma, where they identified 164,000 products sold exclusively by third-party retailers and not by Amazon. Ten months later, it was found that Amazon had started to sell around 3% (some 5,000) of these items. Five thousand new items in a span of ten months are significant. Almost expectedly, the items sold by Amazon were now the default option on the results page.

What’s interesting is the nature of products chosen by Amazon. The study revealed that Amazon “was more likely to offer products directly when those products had strong demand relative to other items in the same category; when there were numerous third-party sellers (signaling that it was easy to source the goods from manufacturers); when shipping costs were low (especially important because Amazon often offers free shipping); and when prices were relatively high”.

Another study found that in the case of women’s clothing, Amazon “began selling 25 percent of the top items first sold through marketplace vendors.” In summary, Amazon, the retailer, is most likely to offer products that are doing well on Amazon, the platform.

Even if independent retailers can think of the impossible and compete with Amazon on price (most retailers who have tried this have failed miserably), they cannot beat Amazon on ad spends. Amazon can use its own platform for the most accurate and impactful ad placement. Further, Amazon has the kind of data about consumer preferences that most retailers will trade a limb for. (Read this The Guardian piece to get an insight from independent British retailers selling on Amazon marketplace.)

To be sure, similar dynamics, constraints, and practices would hold true in the case of platforms such as Flipkart and Paytm, too.

What can be done about it?

Approach to regulating platforms: As is evident by the various cases mentioned above, platform economies which abuse their dominant power and favour their retail segments has the potential to cause harm to both producers and other consumers. Smaller retailers, cab drivers or restaurateurs will find it difficult to compete against the might of the internet giants. Many have in fact shut shop and had to lay off its employees. Consumers tend to lose out as well, either in the form of increased prices, as with Google’s case; reduced choice with Google or Amazon making it harder to find results from rival companies; or with the fear of future monopoly pricing.

To be sure, when attempting to solve the problem, it is essential to recognise that any of the harmful effects of the gig economy pale in comparison to the enormous gains that these platforms have bestowed upon the average consumer and producer alike. If, by attempting to fix the problem, the tech ecosystem gets damaged, we will be doing ourselves a massive disservice. The millions of people who have used these services will attest to the utility they have derived from them and how their lives have been enriched by these services. Thus, any regulation that attempts to fix this problem must be mindful of the potential harm it can do to the goose that lays the golden egg.

Take the typical solution that is preferred by the left – breaking up of the internet giants into separate and different entities. This is not without precedent. The Sherman Antitrust Act (1890) was invoked to break up Standard Oil in 1911 and again when AT&T was regulated as a public utility in 1913. However, the efficacy of such a solution in today’s economy is seriously questionable. It has the potential to do far more harm than the problem it is trying to solve. As The Economist argues: “a full-scale break-up would cripple the platforms’ economies of scale, worsening the service they offer consumers”. Further, a typical price-cap based regulation carried out on public utilities cannot work in an environment where most of the services offered are for free.

India does not presently have a comprehensive competition law framework to look at the specific problems caused by the new age digital two-sided markets. In its absence, the potential for harm is as great or even bigger than what is observed in the US or Europe. It also serves an opportunity, however, to approach the problem in a fresh light.

India does not presently have a comprehensive competition law framework to look at the specific problems caused by the new age digital two-sided markets. In its absence, the potential for harm is as great or even bigger than what is observed in the US or Europe. It also serves an opportunity, however, to approach the problem in a fresh light.

The answer to regulation in these cases should first start with taking a completely fresh approach towards competition law in the 21st-century marketplace. Competition in marketplaces cannot be judged using a narrow set of outcomes such as pricing, output and “consumer harm” that is currently used in most other sectors. Most of the above cases would fail these tests. Pricing would become a redundant criterion when judging Facebook or Google, who mostly provide their services for free (consumers pay through watching ads). Instead, the law should include “producer harm” as a criterion when judging marketplaces.

A paper in the Yale Law Journal scopes out this approach:

“Applying this idea involves assessing whether a company’s structure creates certain anticompetitive conflicts of interest; whether it can cross-leverage market advantages across distinct lines of business; and whether the structure of the market incentivizes and permits predatory conduct”.

More precisely, when looking at vertical integration in marketplaces, regulators should look at:

– Neutrality of the platform: Does the platform favour its own services over others? Platforms essentially enable a search – the best seller for your needs. Thus, it would be crucial to check the neutrality of the search results. Does the platform consistently display its own services at the top of the search result or engage in other preferential placement of its own services compared to others? What is the default option: is it the service provided by the platform?

– Leveraging data: Platforms amass large swathes of data, which is then used to favour its own services against others. Studies, such as the one conducted by Zhu and Liu mentioned above, should be regularly conducted to check for misuse of the collected data.

– Rules of the game: In cricketing analogy, the platform acts as the ICC (setting the rules of the game), an umpire and a player. Regulators should check whether the rules of the game are written in a fashion that is meant to privilege the player.

Regulatory solutions:

First, a preventive measure: regulators should scrutinise mergers and acquisitions that create anticompetitive conflict of interest, vertical integration, and the cross-leveraging of data. There should be close monitoring of M&As where insights generated through data acquired in one sector can be used to undermine rivals in another. Currently, M&As over a specific monetary threshold attracts the regulators’ attention. However, the monetary value of a deal does not always capture the scope and scale of data that is at stake. Thus, it would make sense for regulators to automatically review any deal that involves transfer of a certain form or quantity of data. If this rule were applied, Facebook’s acquisition of Whatsapp and Instagram would have received greater scrutiny, for example.

A stricter approach to take and one that should be used sparingly is to place a preventive ban on vertical integration. This would mean banning a dominant firm from entering any market that it already serves as a platform – in other words, from competing directly with the businesses that depend on it.

Ex-post-facto regulation should impose a non-discrimination policy. This would allow the platforms to be vertically integrated, but a strong non-discrimination policy would prevent the firm from unfairly preferring its own business or unfairly discriminate among platform users to gain leverage or market power.

One way to ensure non-discrimination would be to go back to the Microsoft solution. Back in 2001 when the court found Microsoft guilty of anticompetitive practices, the Department of Justice reached an agreement with Microsoft to settle the case. The proposed settlement required Microsoft to share its APIs with third-party companies and appoint a panel of three people who will have full access to Microsoft’s systems, records, and source code for five years in order to ensure compliance. This ensured that third-party vendors had the same advantage as the dominant firm. In today’s day and age, Uber could selectively share its algorithm to regulators to prove that they do not favour company-owned cars. Regulators could ask Amazon to prove that the data it has collected on its sellers and consumer preferences as a platform is not being shared with Amazon’s retail business. The very threat of scrutiny that such rules hold will act as a deterrent to anti-competitive behaviour by the platforms.

Regulators could also ask the platforms to conduct regular internal audits to check specifically for these anti-competitive practices. The firms could create an auditing body comprising of experts, with access to relevant proprietary code, data and algorithms. This body would report to the Board of Directors of the company but would have the powers to make its findings public, if it was felt needed. For example, this body would be asked to check if the platform Swiggy consistently recommends The Bowl Company to its users, even though it does not meet the regular criteria. This worked reasonably well in the Microsoft case.

Finally, in terms of judicial remedy, there should be a business-to-business dispute resolution mechanism. Just like the consumer court, businesses that have disputes with platforms could seek redressal from these new outfits. Uber drivers who feel cheated or retailers who feel that their product is placed too low, despite meeting all the criteria, on Amazon or Google’s product search page, could approach the B2B court.

Only if these methods consistently fail should the regulator in India look at adopting more drastic measures. It is vital to remember that many government attempts to correct market failures have ended up causing a bigger mess.

No comments:

Post a Comment