The folly of breaking up Big Tech

Anupam Manur April 9, 2019

It’s about as likely as is Mexico paying for Trump’s wall,” commented an equity researcher in response to US Senator Elizabeth Warren’s proposal to break up big tech. Nevertheless, the idea is gaining support among several US presidential candidates.

The basic premise of Warren’s proposal is that big tech companies have too much power and influence that they wield over the economy, society, and the democratic process. To quote from her policy proposal that she posted on Medium: “They’ve bulldozed competition, used our private information for profit, and tilted the playing field against everyone else. And in the process, they have hurt small businesses and stifled innovation.”

She goes on to elaborate how big tech companies achieved this level of power, identifying two main reasons:

1. Big tech companies have used acquisitions to stifle competition — Facebook’s acquisition of WhatsApp and Instagram; Google’s purchase of DoubleClick (ad company) and Waze (maps company); and Amazon’s various acquisitions

2. Big tech companies indulge in vertical integration, where they own both the marketplace as well as participate in it. This can lead to a violation of the neutrality of the platform and create a conflict of interest.

While the reading of the problem has a few flaws, it is not entirely incorrect. Existing antitrust laws are not evolved enough to deal with some of the anticompetitive practices by new tech companies. Mergers and acquisitions, vertical integration, and common ownership of platforms indeed cause competition problems, as detailed here and here.

US senator Elizabeth Warren’s proposal to break up big tech has found several supporters but could prove disastrous to the tech and startup ecosystems. | Photo: Wikipedia

Seeking solutions from early 20th century

The dominant discourse around competition seems to have come a full circle. In the beginning of the 20th century, competition law existed predominantly to protect small businesses. At that time, Standard Oil went through a wave of expansion to make it one of the largest multinational corporations in the world. Standard Oil indulged in horizontal integration by aggressively acquiring smaller rivals and, thus, eliminating competition. It also indulged in vertical integration by having business interests in oil exploration, drilling and extracting crude oil, transportation (including railroads), refining crude oil into petroleum products such as petrol and gasoline, and distributing the fuel to company-owned retail stations. At one point, the Standard Oil Trust held stock in 41 other companies, which controlled other companies. These, in turn, controlled yet other companies. In 1904, Standard Oil Trust controlled 91% of oil production and 85% of final sales in the United States. The US economy had other such trusts in a variety of sectors. In fact, the term antitrust comes from the competition laws (Sherman Antitrust Laws) laid down in 1890 to rein in the market power of such powerful trusts.

All this changed in 1911, when the US Supreme Court, in a landmark judgement, ruled that Standard Oil was an illegal monopoly and that it should be treated as a public utility. The trust was broken down into 34 individual companies. Since then, other big companies were treated as public utilities and broken down under the Sherman Act, the most notable example being AT&T in the 1950s.

Antitrust standards took another big turn in  the 1970s when Robert Bork, an American judge and legal scholar, published a book called The Antitrust Paradox. In essence, the thrust of antitrust regulation changed from protecting competitors to ensuring consumer welfare. The consumer welfare standard did not explicitly care about how big firms were or the specifics of mergers and acquisitions. It only focused on whether any of these practices impacted consumer welfare. Thus, dominance by itself was not a problem but abuse of dominance was. This standard has been dominant in the way competition commissions around the world dealt with anticompetitive practices. Until now.

In the last few years, the consumer welfare standard has come into question across the world. The largely laissez-faire approach has been criticised as the reason for the tech companies becoming too big. There is a growing realisation that the traditional tools available with competition authorities are inadequate to detect and punish anti-competitive practices. The European competition commission has levied billion-dollar fines on Google on three separate occasions. There are growing calls in the United States to regulate the big tech companies more stringently.

Instead of finding solutions to modernise competition law to deal with the new economy, most solutions are cut from the 1911 vintage cloth

Unfortunately, instead of finding solutions to update and modernise competition law to deal effectively with the new economy, most solutions are cut from the 1911 vintage cloth. It is surprising when commentators who believe that the current competition law is outdated suggest a solution that was used a hundred years ago. Take Warren’s two main policy proposals to deal with big tech, which are most explicitly copied from the Standard Oil solution.

1.Large tech companies will be designated as platform utilities and broken apart from any participant on the platform. In essence, a company with an annual global revenue above $25 billion cannot be vertically integrated and those that already are would be broken up. Further, these companies “would be required to meet a standard of fair, reasonable, and non-discriminatory dealing with users. Platform utilities would not be allowed to transfer or share data with third parties. Smaller companies (annual revenue of less than $25 billion) will not be structurally broken up, but will still have to abide by the non-discriminatory clause.

2. Past mergers that are deemed to be anti-competitive will be reversed, including:

Amazon: Whole Foods; Zappos

Facebook: WhatsApp; Instagram

Google: Waze; Nest; DoubleClick

Impractical to implement

The Washington Post described Warren’s idea as a Washington problem — “a policy proposal with negligible impact on any real problem that nonetheless gains currency because it can be explained to voters in less than a minute”. This is a political message that voters can get behind, but can nearly be impossible to implement. The focus here is on the economic and technical difficulties in implementing a plan such as this, setting aside the political difficulties, which could be quite high as well.

Breaking up technology companies would mean the government or regulators having to split up the companies’ underlying teams and technology. It would require a great deal of technological competence on the part of the regulators to keep each of the targeted companies separate from the other’s markets.

Given that most of the technology companies are highly integrated firms, such a breakup would be a death sentence. As the Wall Street Journal observes, “Google and Facebook rely on flexible teams that cross the normal divisional boundaries to solve problems.” Splitting up these firms would require government officials to go cubicle by cubicle — a difficult and draconian move.

Given that most of the biggest technology companies are highly integrated, breaking them up would be a death sentence

An even more difficult task is to break up the companies’ technologies. “Google’s back end is also tightly integrated,” adds the WSJ article. “The Google File System stores massive data sets; Spanner distributes stored data across multiple servers; and Dremel enables effective queries.” It would be nearly impossible for the regulator or even the company itself to decide which of the broken-up firms would get what technology.

Further, breaking up these companies would significantly reduce the value consumers get due to the high interconnectedness of the products. A lot of the value that Google has seen in the Maps platform, for instance, comes from all the data that they have from Search. Customers also receive a lot of value from other Google products that are cross-subsidised from revenue earned in other products. YouTube, for instance, is widely believed to be non-profitable but is supported by revenues earned by other products.

We would also have to stop and wonder how is it that one of the most integral parts of our lives — Google Search — is provided free of cost. Google can give the service for free because it can monetise it with advertising. If Google is broken up, this would no longer be possible. Breaking up any one of these services would give us substantially less valuable services.  

Breaking up these technology companies would also have a severe impact on innovation in the sector. As an article in Politico points out, “The top five spenders in research and development in 2017 were all tech companies. Amazon alone spent more than $22 billion. The development of autonomous vehicles, artificial intelligence and voice recognition wouldn’t be nearly as advanced as they are now if it weren’t for the work of Google and Amazon”. Investing in R&D and finally introducing them into the market is an expensive ordeal. However, big tech companies can afford to do so because of the nature of interconnectedness that exist within their products.

The other charge that is often levied against the big tech companies is that they destroy their competition through aggressive acquisitions. The Economist, a magazine that stands for economic liberty, ran an alarmist cover story about how the big American technology companies are creating a ‘kill zone’ — where young firms entering into the domains of big tech either get swallowed whole or their tech features are blatantly copied by one of the big players, thus forcing them out of the market. The article further elaborates that venture capitalists were reluctant to fund any startup that was potentially entering into the kill zone.

Any move to enhance competition by disallowing acquisitions of startups could end up making it harder to finance startups in the first place

While there is some truth to this, there is enough evidence to the contrary as well. Venture capitalists also view an exit via the tech companies as a favourable opportunity. They fund startups with the expectation of recouping their investments and one of the established ways of doing that is by selling out to one of the FAANGs — Facebook, Apple, Amazon, Netflix or Google. A probability of acquisition by these companies raises the startups’ initial funding chances. Thus, ironically, an antitrust move to enhance competition by disallowing acquisitions of startups could end up making it harder to finance startups in the first place.

A scalpel, not a sledgehammer

Most countries have approached the problems posed by the technology companies with a sledgehammer instead of a scalpel, which is more appropriate. Vertical integration and other anti-competitive practices require surgically precise policy intervention instead of a broad-based sweeping policy that destroys the entire edifice due to a couple of cracks. This is equally true in Europe, which has already fined Google thrice, and in India, where foreign platforms have been effectively broken up through changes in the foreign direct investment rules.

The first problem with this approach is the assumption that being big is bad. The consumer welfare standard does not bother with the size of the company. Instead, what matters is whether the size is being used to do harm. Indeed, the head of the US Justice Department’s Antitrust Division, Makan Delrahim, has said, “Big is not bad. Big behaving badly is bad.”

Big is not bad. Big behaving badly is bad: Makan Delrahim, Assistant Attorney General for the US Department of Justice’s antitrust division. | Photo: Wikipedia

The first task of antitrust laws is to establish dominance and consequently prove that the dominance is used for anti-competitive practices that cause consumer harm. Unfortunately, Warren’s proposal misses this nuanced process completely. Establishing dominance can also prove to be problematic in some cases. Amazon is a massive ecommerce player but still has a small market share in overall retail. Sure, Google has done amazingly well and captured a large share of the market in Search, but most other Google products, with all the money power behind it, has failed to establish dominance. Remember, Google+ anyone?

The direct and indirect network effects that exist in the platform markets inevitably lead to concentration of market power in the hands of a few big companies. The equilibrium in platform markets tends towards oligopoly or a handful of big firms, as Jean Tirole had shown in his seminal paper on two-sided markets. However, an oligopoly does not imply a lack of competition in these markets. Apple and Google are at a constant war with each other to develop the latest products. Amazon has been bleeding losses in order to increase its market share.

Critics have also often contested that being big does not give permanent dominance. There have been lots of instances in the tech world where an incumbent seemed infallible only for an upstart to embarrassingly reach greater heights. As a Harvard Business Review paper sounds out: “In the early years of the internet age, a half-dozen companies were serially crowned the victor in search, only to be unseated by more innovative technology soon after. Yahoo and others gave way to Google, just as Blackberry faded in response to the iPhone. MySpace (remember them?) collapsed at the introduction of Facebook, which, at the time, was little more than a bit of software from a college student. Napster lost in court (no new laws were needed for that), leaving Apple to define a working market for digital music.”

It seems that the best regulator of technology is simply more technology.

Fixing the cracks

As mentioned before, the reading of the problem statement by Warren’s team is not entirely wrong. Vertical integration can lead to violation of platform neutrality that can hurt small businesses that trade on the platform. Amazon has been known to use market data about product performance from the businesses that sell on its platform and then introduce the products at lower prices. Google has been found guilty of promoting its own brand and services ahead of others in its Search services.

Surprisingly, Warren alludes to the famous Microsoft antitrust case of the 1990s to justify her plan of breaking up big technology companies. That is completely missing the point. The main takeaway of the Microsoft case was to not break up Microsoft but to eliminate those practices that led to abuse of dominance and weakened competition.

Vertical integration and other anti-competitive practices require surgically precise policy intervention instead of a broad-based sweeping policy that destroys the entire edifice

Microsoft was forced to disclose details about the workings of its Windows operating system to the competition authorities and prove that the algorithm did not have a systematic bias against its competitors. Similarly, to ensure non-discrimination, regulatory bodies can appoint a special panel, in consultation with the companies in question, to look into the relevant proprietary code, data, and algorithms to test for discrimination. Amazon, for example, can be asked to prove that the data collected by the platform on consumer preferences are not shared with its retail arm or that the algorithm is not designed to favour its own products.

Another possible solution is to create an arbitration body, possibly within the competition authorities (the Competition Commission in Europe and India, and Federal Trade Commission in the US). Just as there are consumer courts to dispel justice to consumers who have been hard done by companies, such an arbitration body can adjudicate on business to business transactions. This neutral arbitrator could determine whether there is a violation of the platform neutrality or non-discrimination requirement. If a company deems that Google has deliberately pushed its products or services to the second page of its search results, while promoting Google’s own brand, it can approach such an arbitrator for mediation with proof of such an event occurring.

In such a case, the burden of proof will fall on the complainant and that proof would consist of three parts, as outlined by Hal Singer in this interview: First, you would have to prove that your content is similarly situated to the content that is being favoured by the platform. Second, that the reason you are getting unfavourable treatment is your lack of affiliation with the platform. And third, as a result of one and two, you are materially impaired in your ability to compete effectively.

It would make sense for competition authorities to automatically review any deal that involves an exchange of certain forms or a certain quantity of data

Finally, mergers and acquisitions have been under the purview of the antitrust authorities for a long time and they normally make the decision to investigate cases according to a monetary threshold and market share. However, the monetary value of a deal may not always highlight its anti-competitive effects. Stories of big technology firms acquiring small startups at bargain prices are quite common. The value of these startups could lie in the networks (user bases) or the data they possess. This is what transpired when Facebook acquired Instagram and WhatsApp. Nevertheless, a retrospective breakup can do more harm than good, as has been previously discussed. Going forward, it would make sense for competition authorities to automatically review any deal that involves an exchange of certain forms (or a certain quantity) of data. This is not to say that such mergers and acquisitions are disallowed, but a greater form of scrutiny with more relevant tools can lead to much better results.

What is required is for constant innovation in competition law to build newer tools to detect and answer anti-competitive practices that might emerge in a dynamic environment, not a wrecking ball solution that is a throwback to the start of the 20th century.

Constant innovation is required in competition law to detect and answer anti-competitive practices, not a wrecking ball solution that is a throwback to the start of the 20th century

None of the solutions outlined above answers the real problems with regard to big tech — that they have gained an enormous control over our social and political life. Big tech wields an inordinate amount of political power with its ability to sway elections and enough lobbying power to shape favourable laws. This is outside the ambit of competition law. They have gained this power as a result of a Faustian pact that we, as a society, have made with them in exchange for life-enriching services. Breaking them up will not solve this problem, as it does not reduce the political power that Amazon or Facebook has acquired over time.

This is not unique to the United States or to big technology companies, for that matter. Several big companies across the world hold a disproportionate amount of influence in the political realm. There is no denying the amount of political clout held by Reliance in India. The company broke up early this decade, albeit of its own accord. Yet, its political influence has grown. There are no clear technocratic answers to these questions, at least not in the realm of competition law. Solutions might yet emerge, but for now, it is partly a price that we are paying for choosing to live in a free society and a dynamic economy.  


Disclosure: FactorDaily is owned by SourceCode Media, which counts Accel Partners, Blume Ventures, Vijay Shekhar Sharma, Jay Vijayan and Girish Mathrubootham among its investors. Accel Partners and Blume Ventures are venture capital firms with investments in several companies. Vijay Shekhar Sharma is the founder of Paytm. Jay Vijayan and Girish Mathrubootham are entrepreneurs and angel investors. None of FactorDaily’s investors has any influence on its reporting about India’s technology and startup ecosystem.