Discriminating anti-discrimination requirements

Requiring a platform to adjust their algorithms to cease discriminating in how they rank suppliers is an increasingly used behavioural remedy. In a recent post I discussed why finding the counterfactual in these cases is not a trivial exercise. This post returns to the topic of anti-discrimination requirements in relation to ranking algorithms, to illustrate, with a simple example, how ill-considered remedies can make the algorithms continue to exhibit the same anti-competitive effects, even where the platform is demonstrably clear of any misbehaviour. With this post, I’d like to reiterate my more general earlier point, that information asymmetry is one of the main obstacles to designing effective remedies for anti-competitive outcomes emerging from autonomous systems, therefore there should be an increased requirement on the operator of the system to demonstrate not just that its design was changed in accordance with the wording of the remedy/commitment, but also that the system continues to exhibit the desired outcome.

Damage quantification in algorithmic abuse of dominance cases – the elusive counterfactual

(post also appeared on the Centre for Competition Policy blog)

On the 5^th of  February 2024 the Competition Appeal Tribunal published a ruling to determine the carriage dispute in relation to two applications to commence collective proceedings against Amazon regarding Amazon’s Buy Box.[1] Interestingly, the ruling (which went unanimously in favour of Mr Hammond’s application) included a short discussion of the counterfactual.[2] The proposed method for the determining the counterfactual in Mr Hammond’s opt-out class action included re-running Amazon’s original algorithm without the abuse (and logging the resulting outcome as the counterfactual). The CAT seemed sympathetic to this solution. I personally think the proposal in Mr Hammond’s application is a viable approach, and I do believe the CAT must be commended for understanding the novelty of the issue and being open to a method untested in courts. But this ruling draws attention to a much more general question: how to find the counterfactual in a case where the offence is related to conduct by an algorithm?

The Legal Definition of Self-Preferencing: Too Narrow, Too Broad, or Both?

Self-preferencing is usually defined as when a large digital platform treats its own (vertically integrated) products more favourably than those of its competitors.  Such behaviour may be harmful to consumers.[2]  Versions of this definition have now been embodied in EU and UK legislation which also makes such behaviour illegal for gatekeepers or businesses of strategic market status. However, is this definition of self-preferencing too narrow in that vertical integration may not be a prerequisite for essentially similar strategies to be profitable and similarly harmful? Or is it too broad because it does not specify consumer harm as a condition of illegality, even though self-preferencing often does not harm consumers?

Public service broadcasting – a renewed need in the time of recommender systems?

Digitisation has entirely rewritten the way we consume audio and audviovisual content. Some of the market failures that characterised the analogue broadcasting era, have seemingly disappeared in a world where evaporating fixed costs mean that the most niche content is now available online. In this world, there are increasingly vocal arguments that there is no longer need for Public Service Broadcasting (PSB) if the market can also provide for even the most diverse audiences. But the large platforms which host vast amounts of digitalised content, may not quite be the land of milk and honey some had thought for consumers, and certainly not for the producers of content. The reason is to do with how the content is provided on the platform, through algorithmic recommendations, which are likely to suffer from biases. These biases have the tendency to create a world, that suffers from the same market failures (including the disproportionate underrepresentation of diverse, and niche content) that justified PSB in the analogue era. Because of these market failures, a new generation of PSB could have a critical role in ensuring the provision of online content that adheres to objectives, such as the provision of creative, high quality and distinctive output and services in a way that reflects, represents, and serves diverse communities.

When market concentration increases, poorer households are the worst hit

An intensively researched theoretical and empirical question in industrial organisation is the relationship between market concentration and prices. There is ample evidence showing that increasing concentration is associated with increased prices, and that more competition lowers prices. This effect has often been linked more directly to market exit and entry. But is the change in prices the same for every consumer?

Our work offers evidence of the distributional effect of changing market concentration. Areas with lower income experience a larger increase in petrol stations’ price margin as a result of market exit. On the other hand, entry does not benefit the same low-income areas with a larger reduction in the margin than in high-income areas. We argue that these findings are due to differences in how much consumers in different demographic groups engage with the market. Our findings give support to the argument that antitrust could help address inequality while staying true to its mission of promoting competition, provided that priorities are given to not only fixing supply-side problems but also to exploring demand-side remedies.

How much do universities spend on lobbying in Brussels?

Why would universities lobby in Brussels? Well, the most obvious reason is to ensure that the EU continues to spend on academic research, or to be involved in the negotiations for upcoming research framework programmes (mainly Horizon 2020). I decided to look at the amount of money UK institutions spend on lobbying in Brussels. Among others, I was wondering if there was an increase in the amount spent after the Brexit referendum (for example to lobby for the post-Brexit continuation of research funding paid to UK universities). The data does not offer a lot of support to this. On the other hand, the main story is two big lobby spenders, The University of Nottingham and the University College Dublin.

The flying-related carbon impact of academic conferences

How large is the carbon footprint of academic conferences? We collected data from a sample of 263 economics conferences. Speakers traveled around 417 million kilometres to get to these conferences. This, even under very conservative assumptions, amounts to around 50,000 metric tonnes of CO2 emission. It is a staggering figure, given that this is only a tiny fraction of all academic conferences. When linking this data with the number of citations researchers get on their presented papers, the paper offers evidence that the number of trips matters for more citations, but longer distances are not associated with higher citation numbers. This supports the argument that in most cases a preference for localised conferences, rather than long-distance trips can deliver the same benefits of conferences but at lower social costs. Paper and data available here.

How fast are computer storage prices falling?

Short answer: faster than price indices suggest. PPI is likely to underestimate the price fall, as it mis-measures improvements in simple characteristics, such as areal density. We found that the retail unit price of storage falls almost 4 times as fast as PPI would suggest. And that is before taking into account improvements in speed, power consumption, or failure rates.

Chopping Board Economics

This morning I activated what I have been working on in my free time for a while (besides running after three pre-school kids). Chopping Board Economics is a venture to explain simple economics through topical subjects, using an everyday motivator: cooking. I started this venture by drafting a book manuscript but after 6 chapters I realised that the concept would work better on video. So here’s the first 5 episodes, discussing the economics of food waste, inequality, trade, collusion, and the IKEA effect. Two more episodes, on the price of food, and on the sugar tax are also ready to be released but I decided to drip feed them, releasing a new episode every few weeks.

Brexit uncertainty: fewer mergers, big business winners

While it is widely recognised that last year’s EU referendum caused significant uncertainty for markets, some early indications were that it had not reduced the level of business confidence. Our results show that the uncertainty surrounding the referendum triggered a fall in the number of mergers and acquisitions, and M&A activity has not recovered. The mergers that have suffered the largest drop have been the ones that were large in comparison to the size of the acquiring firms. Finally, as a result of the post-referendum uncertainty, the businesses that have remained more M&A active have been typically the largest ones. Our results – that are based on a careful study design of a treatment and control group – contradict some of the earlier, more optimistic discussions, which were based on simple before-after comparisons.