Have 'Big tech' platforms got too much power?

Introduction

Google has recently been fined 4.34 billion Euro by the European Commission for abusing its dominance over the Android mobile phone operating system.¹

In 2017, the European Commission fined Google 2.4 billion Euro for “the more favourable positioning and display by Google, in its general search results pages, of its own comparison shopping service compared to competing comparison shopping services”.²

These are not isolated incidents.

On July 5 this year, the Competition and Consumer Commission of Singapore provisionally ruled that the local merger of two ridesharing companies, Grab and Uber would substantially lessen competition.³

In December 2017, the German Federal Cartel Office made a preliminary finding that Facebook was abusing its market dominance in its collection and use of data.⁴

In May 2017, the European Commission fined Facebook 110 million Euro for providing misleading information about data matching as part of its acquisition of WhatsApp.⁵

Also in May 2017, the European Commission accepted undertakings from Amazon to modify its potentially anticompetitive e‑book contracts with publishers.⁶

This follows the 2013 decision by a district court in the US that Apple orchestrated a conspiracy among publishers to raise the price of e‑books.⁷

I could go on – but you get the idea.

So, have Internet platforms like Facebook, google, amazon, and others got too much power? Are they behaving in an anticompetitive manner?

I will let the cases speak for themselves.

However, it is clear that the growth of online business has raised a range of competition issues that, throughout the world, are being grappled with by governments, competition authorities and the courts.

In this talk I want to argue that these competition issues are intrinsically related to certain features of online business. These features are not unique to the internet, and we have ‘seen them before’ in other industries. But online businesses have moved issues that were at the fringe of competition analysis to centre stage.

In particular, I want to concentrate on the importance of consumer data.

What is new?

Having large businesses with market power attempting to abuse that power and undermine competition is nothing new.

Indeed, the potential for such behaviour is a key reason for the existence of competition laws. The original competition or ‘antitrust laws’, the 1890 Sherman Act in the United States, has two parts.

The first prohibits collusion between competitors.

The second prohibits ‘monopolisation’ – where a business tries to undermine competition to gain or protect a dominant market position. The law was motivated by the anticompetitive conduct of big businesses of the era – the large railroads, Standard Oil and American Tobacco.

So what is new?

There are four competitive features that, in my opinion, have been brought to the fore by global platform businesses.

The problem with platforms

The first feature relates to how these businesses are organised.⁸

Traditional competition laws were designed for traditional business structures. These involve a chain of production, with each stage adding value and the ownership of the products formally changing hands at each stage.

For example, an iron ore producer sells ore to a steel maker who then sells simple steel products to fabricators, before finished moulded steel products are sold to builders and other end users.

In contrast, Internet platforms often bring together different parties to enable direct transactions.

The platform facilitates the exchange and may charge for its services. However, the platform does not own the products being transacted. Uber does not own the vehicles passengers use and it argues it does not employ the drivers.

Apple and Google have operating systems, iOS and Android respectively, that facilitate trade between makers of applications and consumers. They set the rules for allowable applications on their stores. But generally, they do not own the applications or set the price for trade.

Other, older businesses are also organised as platforms. Traditional media businesses, such as newspapers and free‑to‑air television bring viewers and advertisers together. Physical shopping malls provide a platform for sellers to interact with buyers, just like their Internet equivalents.

However, the Internet has raised the size and reach of these platform businesses. They sit uncomfortably with competition laws that are designed for vertical production chains.

Why does this matter?

Well, the recent decision by the US Supreme Court regarding American Express has turned a century of competition logic on its head based on platforms. It suggests that a platform business can abuse one group of customers so long as the business passes some of the increased profits through to consumers on the other side of the platform. The dissenting opinion does a great job of explaining the flaws in the majority decision.

Positive feedbacks (network effects)

The second feature is the positive feedbacks that exist on these Internet platforms. These can entrench market dominance.

For example, people join Facebook because their friends are on Facebook. But the friends joined Facebook for the same reason, and with interlinked networks of friends soon all social networks end up on The Social Network.

Similarly, many Internet sellers use Amazon because it has lots of customers. But customers go to Amazon because it has lots of sellers.

Positive feedbacks exist in some traditional industries but physical constraints, like travel time, act against these feedbacks and limit size. Internet platforms can partially or fully overcome these physical costs.

Positive feedback effects without physical capacity constraints can result in winner‑take‑all markets. It makes entry and competition difficult. Entrants cannot slowly eat into the market share of the incumbent.

Rather they need to leap‑frog and displace the incumbent, much as Facebook did with MySpace in the late 2000s. And leap frog competition is hard, as shown by Google Plus as a challenger for Facebook and Microsoft’s Bing as a challenger for Google Search.

Winner‑takes‑all competition creates strong incentives for incumbents to acquire potential threats. And often merger laws, which rely on a ‘substantial lessening of competition’ test, are impotent to stop this.

There are plenty of examples. As Richard Holden from UNSW and his colleagues document, “in the decade between 2008 and 2017, Google/Alphabet made 166 acquisitions, Amazon 51, Facebook 63, EBay 31, Twitter 54 and Apple 66”.⁹

These include prominent mergers such as Google/Waze and Facebook/WhatsApp. And they suggest that we need to think whether our current merger test needs strengthening.

The problem of pricing

Third, platform businesses have non‑traditional pricing. From the consumers’ perspective the platforms often involve a zero monetary price. Indeed, some platforms and some applications aim to pay consumers.

Think of the reward points that consumers ‘earn’ when they use a particular payment card at a merchant. Or consider Unlockd, an application service for Google’s Android platform that paid consumers to watch advertisements when they unlocked their mobile phones.

Some global platform businesses make their profits through setting positive monetary prices to only one group of participants, such as advertisers for Facebook and Google.

Of course, we have seen zero monetary prices before. The word ‘free’ in free‑to‑air television says it all.

Why does this matter?

Consider the Google Shopping case. The European Commission found that Google modified its search algorithms in a way that downgraded comparison shopping services that competed with Google’s own comparison service, Google Shopping. Google Shopping which is “prominently positioned, displayed in rich format is never demoted by those algorithms”.¹⁰

Why did Google do this?

After all, if Google wants to make more money from comparison shopping websites, why would it go to the bother of setting up its own website and destroying its rivals? An easier (and legal) solution would be to simply charge the existing comparison websites more – to get them to pay through the nose for using Google’s search engine to access customers. This would let Google keep all the profits at little cost.

This is where zero pricing comes in. Google cannot charge for different websites to get different prominence in search results without destroying the value of the Google search engine. Google’s search dominance comes from providing consumers with the best, unbiased links to a search query. If search results were based on price rather than relevance then consumers would quickly desert Google for an unbiased search engine, such as Microsoft’s Bing.

So zero pricing meant that Google was caught by its own business model and its anticompetitive conduct was the only way it could monetise comparison shopping websites. Needless to say, Google is appealing.

The importance of consumer data

The fourth feature is data.

While consumers may not pay a monetary price to platform businesses, they do provide data to these businesses. This consumer data can be used by the platform to create value. This may enhance the customer’s experience, for example when an e‑book or music streaming platform provides better recommendations to a customer. But it can also create more targeted advertising, raising the benefit to advertisers, and allowing the platform to raise advertising revenues.

Indeed, businesses can go further. They can use data to create personalised pricing, using your data to determine the most you will pay, and then charging you that price. Or they can encourage you to buy with personalised discounts. This is not a new practice. Bricks‑and‑mortar retailers have used data from loyalty programs and scanner data for years to target ‘coupons’ to buyers. However, on the Internet, the practice is opaque. You only see the price you are offered, not the alternative.

Data can also lock consumers in to a particular platform. This can be direct. For example, Facebook is valuable to users because it can connect an individual to a network of Facebook friends. It can be difficult for a consumer to recreate this network on a rival platform.

Data can also indirectly raise the value of a platform to a consumer. For example, Google’s Internet search platform uses the data from consumers’ searching to update and improve its search results. This feedback makes it difficult, if not impossible, for a new entrant in search to ‘catch up’ with Google.

Again, some traditional businesses, including banking and insurance, rely of consumer data for their business decisions and if consumers want to switch providers then they need to reproduce the data that they supplied to their current suppliers. And even then, consumers can be locked in simply because their current provider knows more about them than a competitor.

Data raises both competition and consumer protection issues. The recent German case that I previously referred to, involved Facebook, as part of its terms and conditions, requiring consumers to agree to allow not only their data from Facebook, but also their data from other applications, to be transferred to Facebook.

Basically, Facebook used its market power to ‘hoover up’ large amounts of consumer data – usually without the consumers being aware that this was happening. And trading consumer data is a major part of business for both Internet businesses and data brokers.

For some of us, the Facebook/Cambridge Analytica scandal from earlier this year was a shock – not because of the business behaviour but because so many consumers did not realise that online companies gathered and traded their data all the time.

So what needs to be done?

Consumer data raises at least three competition questions for platform businesses. Is consumer data used in the consumer’s favour, to offer bespoke products, or against the consumer, by allowing sellers to know the maximum they can get away with charging?

Is consumer data locking consumers into particular platforms, reducing competition?

And is consumer data being gathered and traded without the consumer’s knowledge and potentially being used against the consumer’s interests?

And the final question of course is – what needs to be done?

Consumer data was a key part of the Productivity Commission’s 2017 Report on Data Availability and Use.¹¹

Following this report, the government has decided to legislate a new Consumer Data Right, in part by amendments to the Competition and Consumer Act 2010.¹²

This right will first be applied in banking before being extended to other parts of the economy.¹³

The Consumer Data Right approach being taken in Australia is appropriately balanced and likely to be pro‑competitive. It addresses all three data related issues.

For example, under Open Banking, banks are required to release information about a client’s accounts to either other financial institutions or to the client themselves, at the request of the client.

This solves the basic competition problem relating to data – to ensure that a consumer’s data is being used in their interest, not against their interest. The consumer data right allows the consumer to both access their relevant data and to transfer that data to competitors that they choose. It also addresses the issue of ‘lock in’. The consumer is no longer locked in to their current provider simply because that provider has their data.

It will also help deal with issues of data abuse and trust. The consumer data right will include a range of privacy protections including extensions to the Privacy Act.

It will require that consumer data is only transferred at the direction of the consumer. Through this, it will help rebuild consumer trust that their data is not being abused.

While the consumer data right addresses the key data related competition issues for platform businesses, it also avoids over‑reach.

Knowing that data can be shared alters the incentives of businesses to collect the data in the first place. A data right that is too broad may undermine the creation of value added data.

For example, in open banking, the focus is on account and transaction data. Financial institutions need this data to provide the relevant products so there is little risk that the data will not be collected.

Importantly, open banking does not apply to value added data, where a bank has worked with customer data creating larger data sets that, for example, allow it to better manage its whole‑of‑business risk.

As the broader consumer data right spreads, the clear distinction between customer data and value‑added data must be maintained to avoid creating perverse incentives.

The consumer data right also avoids over‑reach by creating a shared data right. Consumers will not own their data. For example, under Open Banking, consumers will have no right to require a business delete their data.

Again, this is important to make sure that businesses who legitimately compete with consumer data are not undermined. For example, in banking, consumer data is aggregated for use in broad credit modelling. If consumers could remove specific data from these models, even though the data in the models is de‑identified, then the validity of these models could be undermined.

Consumers have a right to the data they need to encourage competition but firms retain the right to use aggregated data sets to make better business decisions.

When applied more broadly, the consumer data right developed here in Australia will help ensure competition on the Internet, and will reduce consumer lock‑in.

Interaction with overseas standards

So, in my opinion, the Consumer Data Right will bring real benefits to customers, including but not limited to dealings with Internet‑based platform businesses.

However, there is a longer term question about whether individual country data rules, like our consumer data right and the open banking rules, can remain country specific over the longer term.

The European Union (EU) has developed open banking reforms, called PSD2, and a broader General Data Protection Regulation or GDPR. Unlike Australia, however, these reforms have developed in parallel and it is not clear that they are consistent.¹⁴

It is clear, however, that there are inconsistencies between the Australian approach and the EU. For example, the Australian approach to Open Banking does not give consumers a right to delete data on them that is held by a bank. In contrast, the GDPR has an explicit ‘right to erasure’ or ‘right to be forgotten’ in certain situations.

As different countries roll out their data regimes, these inconsistencies will grow. It is likely that regimes that are adopted by larger trading areas, such as the EU, and that are more conservative, will dominate.

Businesses will not want to support multiple compliance systems. And they will not want to be isolated from large trading blocs.

In other areas where standards are required for consistent operations of businesses in multiple jurisdictions, such as telecommunications, international standards bodies are established.

At present, this doesn’t exist for data. The result may be a patch work of inconsistent international standards with the Australian standards being largely usurped by more conservative overseas rules, such as the GDPR.

Conclusion

So what are the takeaways?

First, the competition issues being highlighted by ‘big tech’ platforms are not new. We have seen them before in other settings. And they have been dealt with in other industries. It is not a new brew. Rather, it is ‘old wine in new bottles’.

Second, recognising that the core competition issues are not new, we need evolution, not revolution, in competition policy. There will be key touch points: zero pricing and merger tests for example. But, to continue the analogy, we do not need a new glass for old wine.

Finally, data and the new Consumer Data Right will be a key part of this evolution. Australia is leading the way with a well‑constructed approach to data rights. And in this area, we should encourage other jurisdictions to follow our lead.

End notes