Antitrust in the Digital Economy

(a) Google’s dominance of the market for internet search arose mainly from a combination of economies of scale and a quality algorithm.

(b) In light of Google’s dominance, its current operating practices could have a substantial negative effect on social welfare in the long run.

(c) The nature of the market dominance of technology giants in the digital economy warrants either the imposition of some kind of regulation or a fundamental change in antitrust policy.

Of our 43 US experts, 39 participated in this survey; of our 48 European experts, 41 participated – for a total of 80 expert reactions.

Google’s dominance of internet search

On the first statement about what underpins Google’s dominance of the market for internet search, a strong majority agreed that it arose mainly from a combination of economies of scale and a quality algorithm.

Weighted by each expert’s confidence in their response, 30% of the US panel strongly agreed, 65% agreed, 5% were uncertain, and 0% disagreed. Among the European panel (again weighted by each expert’s confidence in their response), 31% strongly agreed, 61% agreed, 5% were uncertain, and 3% disagreed. Overall, across both panels, 31% strongly agreed, 63% agreed, 5% were uncertain, and 1% disagreed.

More nuances among the experts’ views come through in the short comments that they are able to include when they participate in the survey. Nicholas Bloom at Stanford says: ‘Google’s search engine has been far better from the outset – this alone can explain why it dominated. It is simply a better search engine.’

Franklin Allen at Imperial adds: ‘They do have a good algorithm and this is a big part of their success, but increasing returns to scale due to network effects are large.’ And Daron Acemoglu at MIT notes: ‘Quality of algorithm likely played a role early on, but now it’s mostly network effects – dominance breeds dominance.’

Robert Hall at Stanford introduces an additional factor behind Google’s dominance: ‘Not to mention good timing.’ David Autor at MIT comments: ‘If Google hadn’t invented page rank, someone else would have. Google benefited from getting there first with a good idea.’

Pete Klenow at Stanford draws our attention to background information on Google’s cluster architecture. And John Van Reenen at LSE points to ‘General findings in literature including our work on superstar firms.’

Google’s operating practices

On the second statement about whether Google’s dominance and current operating practices could have a substantial negative effect on social welfare in the long run, a little over a half of the experts agreed, somewhat under a half were uncertain, and a small number disagreed.

Of the US panel (again weighted by each expert’s confidence in their response), 52% agreed, 40% were uncertain, and 7% disagreed. The results were fairly similar for the European panel: 16% strongly agreed, 42% agreed, 38% were uncertain, and 4% strongly disagreed. Overall, across both panels, 8% strongly agreed, 47% agreed, 39% were uncertain, 3% disagreed, and 2% strongly disagreed.

Among those who agreed with the statement, some refer to how potential negative effects might arise. Barry Eichengreen at Berkeley says: ‘Overwhelming market dominance creating formidable barriers to entry are not good.’ Bengt Holmstrom at MIT explains: ‘Difficult to assess. De facto monopoly due to superior algorithm. Worried about limited contestability and biased ad rankings.’ Franklin Allen adds: ‘As with many monopolies, they will have the wrong incentives with regard to pricing of advertising and to innovation.’

Others develop the theme of the possible impact of market dominance on a firm’s conduct. Christopher Pissarides at LSE notes simply: ‘Power corrupts.’ Larry Samuelson at Yale says: ‘Innovation brought Google to a dominant position, but Google bars no holds in preserving that position, with adverse consequences.’

Daron Acemoglu goes further: ‘It walks, swims, and quacks like a duck, it’s probably a duck. It looks, behaves, and dominates like a monopoly, it’ll probably harm welfare.’ And David Autor warns: ‘I fear Google is becoming the new *old* Microsoft, before the antitrust case. They may not be doing substantial harm now – but they could.’ Pete Klenow links to a 2010 study asking exactly that question: Is Google the next Microsoft?

A number of panelists pick up on the word ‘could’. Pinelopi Goldberg at Yale states: ‘They COULD; this does not mean they WILL. Regulators should be vigilant and scrutinize practices going forward.’ Robert Shimer at Chicago adds: ‘The key word in the question is “could”. It could also have a substantial positive effect, as it has in the past.’

Anil Kashyap at Chicago points to a mechanism by which negative effects could happen; a ‘kill zone’ in the space of startups, as described by venture capitalists, where the prospect of acquisition by an incumbent platform undermines early adoption and makes new entrants not worth funding. Kjetil Storesletten at Oslo worries about the effects on other industries: ‘Indirect costs: dominant tech giants soak up the advertisement revenue the free press/media depends on. This risks crowding out the press.’

Karl Whelan at University College Dublin expresses: ‘Weak agreement, but we’ve seen large tech companies rise and fall. This could be the “peak Google” era and something else replaces it.’ But Peter Neary at Oxford, who strongly agreed with the statement, responds: ‘A classic case of natural monopoly. Unfortunately, it is (almost) at the global level so countervailing action would have to be global too.’

Among the sizeable minority of panelists who said that they were uncertain, some find the question insufficiently specific. Jonathan Levin at Stanford replies: ‘Struggling with question framing. Some practices deserve scrutiny. On net, however, its products and services create enormous value.’

Kenneth Judd at Stanford says: ‘Fuzzy answer to a fuzzy question. “Could” have negative impact, yes. But current operations includes creation and sharing of powerful tools.’ Pol Antras at Harvard adds: ‘Ambiguous language here. Google will not have decreased steady-state welfare; but relative to a counterfactual with more competition, it may.’

Others were unsure that they knew enough for anything more than a response of ‘uncertain’. Christian Leuz at Chicago notes: ‘Given Google’s dominance in the market for search, there clearly is potential for such harm. Don’t know enough about its practices to answer.’ Darrell Duffie at Stanford comments: ‘Depends on how Google exercises its market power, and on the quality of potentially superior entrants. Those are beyond my expertise.’

Patrick Honohan of Trinity College Dublin, who voted ‘no opinion’, responds: ‘Future dynamics of the sector too uncertain for me to assess.’ And Richard Schmalensee at MIT says: ‘I don’t know enough about its current operating practices to be very confident.’

Among those who disagreed or strongly disagreed with the statement, Robert Hall states: ‘I take it that would mean that regulators are currently passing up a constructive intervention, which is not the case.’ And Nicholas Bloom concludes: ‘Search is a natural monopoly. If better searches exist they will win – Google beat out Alta Vista on product quality.’

Technology giants

On the third statement, which broadened out from the specific case of Google to ask whether the nature of the market dominance of technology giants warrants regulation or a fundamental change in antitrust policy, there were greater differences in the results across the two panels than with the two previous questions. A considerably larger proportion of experts on the European panel agreed or strongly agreed with the statement than the US panel (88% compared with 53%); and just over a fifth of US experts disagreed.

Of the US panel (as usual, weighted by each expert’s confidence in their response), 7% strongly agreed, 46% agreed, 26% were uncertain, and 21% disagreed. Among the European panel, 23% strongly agreed, 65% agreed, 6% were uncertain, 3% disagreed, and 4% strongly disagreed. Overall, across both panels, 16% strongly agreed, 57% agreed, 15% were uncertain, 11% disagreed, and 2% strongly disagreed.

Among the comments, Christopher Pissarides, who strongly agreed with the statement, says: ‘They are becoming monopolies in very sensitive areas. Human nature cannot be trusted in such circumstances.’ Bengt Holmstrom, who agreed, responds: ‘Market values strongly suggest a need for review. Especially now that competition from China curtailed.’ And Barry Eichengreen, who also agreed, concludes: ‘Absent public sector intervention, this problem won’t solve itself.’

Others who agreed with the statement go into a little more detail on what should be done. Larry Samuelson says: ‘The tech industry is rife with natural monopolies, which are routinely regulated in other sectors.’ Franklin Allen comments: ‘We do need new antitrust policies and new regulation as well as new taxation to deal with the issues raised by the tech firms.’ Daron Acemoglu adds: ‘This should probably involve more than light-touch regulation. We should also deal with the effects of Big Tech on direction of innovation.’ And Jonathan Levin notes: ‘To take one example – thoughtful, informed regulation is needed in areas like privacy and data rights.’

Austan Goolsbee at Chicago goes back to history for perspective: ‘Go look at what happened the last time there was a massive disruption to technology and increase in corporate power 1880-1930.’ David Autor adds: The Sherman Act was not set up for the networked world. Case in point: Facebook should never have been allowed to buy WhatsApp.’

Jose Scheinkman concurs: ‘Need to increase scrutiny of acquisitions of related businesses, for example, WhatsApp by Facebook.’ Richard Thaler, who said he was uncertain, also reflects on policy on mergers and acquisitions: ‘It is hard to favor unspecified changes in the rules. I don’t think Google should be able to buy Waze, nor Facebook buy Instagram.’

Among others who said they were uncertain, Darrell Duffie at Stanford asks: ‘Are existing antitrust laws enough? DOJ seems to think so, and they seemed to work for Bell and Microsoft. For Google: stay tuned.’ Richard Schmalensee warns: ‘Those sorts of changes surely deserve serious consideration, but I’m not confident that we can find changes that are net beneficial.’

Among those who disagreed, several comment on the adequacy of existing antitrust laws. Judith Chevalier at Yale says: ‘Enforcement of merger policy, for example, would ideally be stronger, but I don’t think that derives from a fundamental deficiency in the laws.’ Similarly, Michael Greenstone at Chicago remarks: ‘There are always legitimate questions about enforcement but I think the laws are up to the task.’

Kenneth Judd at Stanford adds: ‘A key point in the current discussion is the Apple-Google arrangement. I guess that current law can handle this; no need for new rules.’ And Robert Hall, who said he was uncertain, notes: ‘Merger regulation is in place and a good idea. Regulators should understand that there is not competition in the market, but for the market.’

Others who disagreed doubt that there are grounds as yet for changes in regulation or antitrust policy. Robert Shimer says: ‘We are yet to see evidence of significant damage caused by these firms, but do see substantial social benefits.’ Pinelopi Goldberg adds: ‘It is not clear what problem the regulator is asked to solve. Being worried about future abuse of power is no justification for regulation.’

Nicholas Bloom, who strongly disagreed with the statement, is concerned about the dangers: ‘The history of government regulation is poor, and particularly with governments like Trump in the US, I do not want them to have more control.’ And Anil Kashyap, who voted ‘no opinion’, asks: ‘How do we know what kind of changes would result – and probably they would not just be designed based on economic principles.’

Finally, Pete Klenow directs us to further reading on the economics of data property rights. And Karl Whelan reminds us of a book co-authored more than 20 years ago by Google’s chief economist, offering ‘a strategic guide to the network economy’: ‘I learned lots from reading Varian-Shapiro’s “Information Rules”. Markets for information goods are innately imperfectly competitive.’

All comments made by the experts are in the full survey results. Combined US and European survey results are available here.

Romesh Vaitilingam
@econromesh
November 2020