It is a great pleasure to be here. As you can imagine; the topic of this session, the frontiers of liberalisation, is absolutely central to the CMA’s work. As a newly formed competition authority we need to think carefully about how we can harness market forces to benefit consumers; and one of our five strategic goals is, precisely, to extend the frontiers of competition.
Traditionally, when we talk about the frontiers of competition, we tend to focus on the deregulation of monopoly industries and whether to shift the boundary between regulation and competition. The CMA’s perspective on this topic is slightly different because, with a few notable exceptions, we intervene in markets that are already subject to competition. In such cases, the question is not whether to introduce competition; but how to make competition work better. The changes that we introduce are perhaps less radical, but it does not mean that our task is easier.
Typically, the market arrangements that we review have evolved over long periods of time and they have their own internal coherence, even if it is not perfect from our point of view. So when we do find some problems, we need to balance the possible benefits of our intervention against the risk of disrupting existing processes that broadly work. It is quite a delicate exercise and I will give you some concrete examples of the difficulties we encounter later.
One of the main tools that we use to promote competition is the market investigation regime, and I would like to focus on this aspect of our work in this session. The regime has been in place for over a decade now, and most of you will be familiar with it, so I will not dwell too much on the process or the legal framework. I would like to focus instead on some of the new challenges that we are facing when using market investigations, and how we intend to meet these challenges. And I will try to do this by taking you through the ‘life cycle’ of an investigation, from the decision to refer a market (when to intervene) to the conduct of the investigation (how we intervene) and finally the choice of remedies and the follow-up (the effect of the intervention).
The decision to open a market investigation
Let me start with the question of when we should use market investigations. And by way of introduction, let me just point out that this is one of the few decisions that are reserved for the board of the CMA. This tells you how seriously we, and indeed the law-makers, take this decision. We are conscious that market inquiries are very costly, both for us and for the industry, so we will only open an investigation when there is at least a good chance that the benefits can outweigh the costs and the risks involved.
How do we make that assessment? Obviously there must be reasonably clear indications that competition is not working effectively. But very often, the most complicated part of the assessment is not to identify potential issues; it is to decide whether these issues are an enduring feature of the market and whether they can actually be remedied through regulatory intervention. This involves a lot of judgement, as you can imagine, and each case is unique. But in our recent decisions we have been confronted with two types of circumstances that make this evaluation particularly tricky: one is the presence of consumer biases in the market, and the other is the effect of regulatory and technological change. So let me say a few words about these two points, starting with consumer biases.
When we review markets, especially consumer-facing markets, one of the most common issues we encounter is that a large number of customers do not seem to respond to market signals. They do not switch when there are better products on offer, or, if they do switch, they appear to make choices that are not optimal for them. There is now a large body of literature that explains these decisions by reference to certain psychological biases: over-optimism, time inconsistency, unstable preferences, and so on.
Now, I am not saying that the existence of consumer biases in a market is necessarily a problem. There are many circumstances where markets can function reasonably well even if customers are affected by such biases. For one thing, customers are often a heterogeneous bunch. In any given market you might have a mix of ‘engaged’ and ‘disengaged’ customers, who suffer from biases to different extents. And if there is a sufficient number of engaged customers, and if firms cannot price-discriminate between the two groups, then the actions of engaged customers can indirectly protect disengaged customers.
There are also dynamic effects at work: customers may learn from their past decisions, or the market may provide tools that simplify comparisons between products; think of price comparison websites. There are many different effects at work here; but the thrust of these arguments is that, in certain circumstances, markets can be ‘self-healing’ and consumer biases can be part of a healthy process of discovery which can ultimately improve the way the market operates.
So when should we intervene? Put simply, I think we should focus on cases where this process of discovery and self-healing does not work; and where we have reasons to think that the market will remain stuck in a poor equilibrium without external intervention. Here I am drawing on the themes explored by Oren Bar-Gill in his new book, Seduction by Contract. The key point made by Bar-Gill is that there are many circumstances where competition does not alleviate market failures caused by behavioural biases. In fact, there are circumstances where market forces work to exacerbate such failures. If customers are biased, the dominant incentive facing suppliers can be to exploit these biases, not to correct them. They typically do this by designing complex products with opaque charging structures that maximise the perceived value of the products as opposed to their real value.
So, if customers are myopic or overoptimistic; the charging structure will involve low payments upfront and higher charges later, or low payments for salient features of the product and higher payments for contingent features (ie, features of the product that are only used under certain conditions). The key point here is that suppliers have no incentive to depart from this equilibrium; if they were to use a different approach unilaterally, they would just lose market share and make losses. In that type of context, the competitive process rewards those who most effectively exploit the biases.
Bar-Gill gives many examples of these situations; in mortgages, credit cards, and mobile phone contracts. John Kay, who wrote a column on this subject last year, gives additional examples in the hotel and car rental sectors. The example that comes to my mind is the market for PPI (Payment Protection Insurance). When the Competition Commission investigated the market in 2008, they found that PPI was mostly sold at the point where customers took out a loan. There was very limited competition or shopping around at that stage, and providers were making huge profits on PPI; with returns on equity of around 500%. Many providers argued that they were using these profits to cross-subsidize the price of loans, because this is what customers focused on. So they were essentially ‘baiting’ customers with low interest rates on loans, and then recouping their losses by cross-selling low-value PPI at high prices.
What is most interesting is that some suppliers recognized that this business model was somewhat flawed, but they stressed that they had no choice; any attempt to offer more cost-reflective prices for credit and PPI would only result in a lower market share and lower profits. This is the key point to me; the fact that the market appeared to be trapped in a poor equilibrium.
When we suspect that this type of dynamics is at work, then I think it is worth launching a market investigation to at least consider whether we can intervene to restore a healthier dynamic. Of course, there is always the risk that we do not get the intervention right and we do more harm than good as a result. So even the existence of this type of dynamics is not a sufficient reason for intervening. But it is at least a good reason for opening an investigation to examine potential remedies more carefully.
Markets in flux
When we consider the case for launching an inquiry, we are almost always told that the moment is wrong because the market is in a state of flux. There can be a range of reasons why this is the case, but the most common arguments revolve around the effects of regulatory change and technological change.
On regulatory change, we often hear that the industry has embarked on a process of self-reform which is about to bear fruit; or that the government is already looking into the market and that it may intervene separately. In general, I do not see any difficulties in reviewing a market alongside other regulatory bodies; be they trade associations, or government, or sectoral regulators. If an industry is already subject to this amount of scrutiny, then it is possibly because the issues are particularly acute. If we were to accept the argument that we cannot intervene alongside other bodies, then we might actually end up excluding ourselves from some of the most important sectors with the most complicated issues; precisely the type of setting where we can make a difference.
On technological change, the specific nature of the claim varies, but typically this argument has two strands. One is that the internet will empower customers by providing them with more information and better tools to analyse this information; the other is that digital or other innovation will facilitate new entry by reducing fixed costs, especially in distribution. In this context, the argument goes, a market investigation will either be a waste of time and resources, or, worse, it might actually disrupt the very process of change that is about to solve the issues that we have identified.
This is only a very quick summary. I recognize that I am not doing justice to the subtlety of the arguments that are put to us in this area, but for our purpose today I think this is sufficient.
Let me just point out in passing that there seems to be a common theme running between this argument and the anti-intervention argument on consumer biases. In both cases, there is a sense that regulatory intervention is particularly risky when we are faced with a market environment that is complex and evolving. In both cases, there is an underlying confidence that the market will eventually ‘self-regulate’ and that suitable solutions will emerge in a decentralized, bottom-up manner.
I have been in business as well as in regulation, and I know full well that real markets are always in a state of change. I am perfectly conscious of the risk of intervening in complex market environments. But equally I do not share this confidence that markets always self-regulate in the long-run, and conversely I strongly believe that we can mitigate the risk of inappropriate regulatory intervention with a suitable process. I have explained how we approach this trade off when it comes to dealing with consumers biases, so let me know explain how we look at technological change.
I should say upfront that the technology ‘defence’ is an argument that we take very seriously. There have been examples in the past where the CMA’s predecessor bodies arguably misjudged the pace of technological change and its impact. One example that comes to mind is the Yellow Pages inquiry in 2006. At the time when the CC started the investigation, Yell had been subject to a price control for about 10 years, but the company was arguing that the market was about to be completely transformed by the entry of internet search engines. The CC examined the evidence for this argument but in the end it decided to maintain the price control. Very quickly after the inquiry concluded, search engines started to attract more traffic, Yell started pricing below the cap, and between 2006 and 2012 the revenues of the company declined by over 60%, to the point where its financial viability was at risk. Now, I do not think that the group was wrong to be sceptical about the potential for new entry given the evidence available at the time. I also do not think that the intervention was particularly harmful, even if it was unnecessary in retrospect. But this example certainly illustrates the difficulties of anticipating the pace of change in this type of market.
So we do take this argument seriously, but we also approach it with a degree of scepticism. Yes, there is a risk of underestimating the pace and the impact of technological change, but in my experience, the risk of over-estimating the impact of digital innovation is at least as great. At the risk of using a rather tired anecdote, recall that when BT was privatised in 1984 the expectation was that facilities-based competition would develop very rapidly and that price-cap regulation was only needed to ‘hold the fort’ for a little while. 30 years on, there are still economic bottlenecks in the UK telecoms industry that are subject to detailed, intensive ex ante regulation. And this is in electronic communications – an industry that one would expect to be particularly prone to rapid, disruptive technological change.
Also, I do not think that we can simply assume that technological change always works to empower customers and facilitate new entry. To give just one example, there is a lot of work on how companies can use digital technologies to better exploit consumer biases; or how price comparison websites can limit competition between them by using clauses that constrain the suppliers’ pricing policies.
So I hope this has given you a glimpse of our thinking on consumer biases and technological change and how we take these factors into account when we decide to open an investigation. I should stress that this does not mean that we are no longer interested in more ‘conventional’ cases of oligopolies and collusion. I think our recent record shows that this is not the case. But I thought these would be some interesting themes to discuss here because they are more recent developments in competition analysis and they are relevant to recent cases.
Before moving to my next topic, I should emphasise that we are also conscious of the cost of our interventions and the risk of getting them wrong. There is a large body of literature that shows that some institutions and business arrangements that look ‘second best’ at first glance may actually be the best option, once we acknowledge all the costs associated with regulation in a ‘real world’ setting; such as information costs, enforcement costs, resource constraints and the risk of political interference.
For example, Joskow shows that regulation often persists despite the absence of economic rent; because it has so distorted industry structures that changes would lead to large losses to incumbents. As he says (I paraphrase) if we had a wise, impartial and omniscient regulator, we could safely intervene a lot more often in markets. We do need to take these risks into account, both in our decisions to launch an investigation and in our follow-up on remedies; and I will say a word or two more about this later.
The conduct of a market investigation
I am now turning to the second stage in the ‘life-cycle’ of a market investigation; namely the conduct of the investigation itself. As I mentioned earlier, the regime has been in place for over a decade; resulting in 16 inquiries. Most of you will be familiar with the process, so I just want to take a few minutes to highlight a couple of new challenges we are facing.
The new challenges we are facing
The first of these challenges is simply that the markets that we are investigating, or proposing to investigate, tend to be bigger and more complex than before. Before 2010, almost all the markets that we investigated had turnovers below £2bn. Since 2010, there have been five inquiries in markets with turnovers greater than £4bn. The groceries market had a turnover of £110bn, and the energy market has a turnover of roughly £35bn. Some of these new markets that we are investigating are very complex; the energy market is a case in point.
I think there are two reasons why the markets we are investigating tend to be bigger and more complex: one is simply that our experience of using the regime has grown over the years, and so we now feel more confident to take on these larger markets; the other reason is that given the cost of running these inquiries we recognise that they are most obviously justified to tackle big, knotty problems with a large value at stake.
The second challenge we are facing is that a number of the markets that we are investigating, or proposing to investigate, have been under regulatory scrutiny for a while. The energy market has been the subject of successive probes and reviews by Ofgem since 2008 and before. The banking market has attracted no less than 10 different reports from the OFT, the CC, the FCA and the PRA (the Vickers commission) all the way back to the Cruickshank report in 2000.
On the one hand, this can be helpful because there is a lot of existing material, and the issues are perhaps more clearly delineated than when we start from scratch. On the other hand it is a challenge because the problems that are left to resolve are, by nature, the most complicated and difficult ones. In general, the most obvious remedies have already been implemented; the ones left to discuss are more controversial in their principle, more costly to develop, and with higher risks of unintended consequences. In other words, these are circumstances where it is more difficult to ‘add value’.
Take Banking Current Accounts, for example. As I have just mentioned, there have been a number of competition reviews over the past decade; and these have led to the implementation of a series of remedies designed primarily to improve transparency and make switching easier. These initiatives were certainly not cost-free, but they were not particularly controversial in their principle. Few people would argue that increasing transparency or reducing switching costs could be a bad thing. Now it appears that these remedies have not delivered all the benefits that were expected. Customers have a lot of information at their disposal, but often they do not seem to ‘process’ that information or act upon it.
We are unsure why this is the case; it could be because customers tend to under-estimate their propensity to use overdrafts, and therefore they do not pay much attention to charges that are actually relevant to them. Or it could be because the complexity of charging structures makes it impossible for them to compute the overall cost of using their accounts. The root causes of the problem are still unclear, but you can already see that if we do decide to investigate further; this could take us towards a more complex territory in terms of remedies.
If customers under-estimate their propensity to use overdrafts, then it is not enough to give them information about the product they are using, you also need to make them more aware of their usage of the product. And if customers cannot understand complex charging structures, then it might be necessary to simplify these structures. There is a somewhat uncomfortable paradox here, that to extend the frontiers of competition we may need to intervene deeper into the competitive processes in the first place.
The third challenge has to do with the use of behavioural economics in competition assessments. Now, it may seem odd to have behavioural economics mentioned as part of the ‘challenges’ faced by a competition authority. Clearly, behavioural economics is a source of great insights for our work. But the development of behavioural economics has also raised expectations about the amount of work we need to do on the demand side. To understand how customers make decisions we need to do surveys, interviews, field studies, and lab experiments. And then we may need to do further tests to develop remedies for the problems that we have found.
Our experience in this domain is still limited, but I suspect that this process will sometime involve a degree of experimentation and trial-and-error. And that is the essence of this new challenge that we are facing. For, as you know, market inquiries have a fairly ‘regimented’ process with tight deadlines; and this process does not lend itself easily to experimentation and trial-and-error. So we might need to find different way of operating, and I will say a word about some of our ideas in a moment.
This leads me to my final point concerning the challenges we are facing. As you know, we now generally have only 18 months to reach a final decision, instead of 24 previously. So we have less time to tackle more complicated issues in bigger markets, with higher expectations regarding our work on consumer decisions! So these are some of the challenges that we are facing. It may sound like we are facing a rather daunting task, so let me explain some of the changes that we are introducing in our processes.
The changes we are introducing to the process
Obviously we intend to make the most of the synergies created by the establishment of the CMA. Concretely, this means that some of the staff working on phase 1 market studies will sometimes work on phase 2 inquiries. Hopefully this will enable the phase 2 teams to capitalise on their sector knowledge and data to make a quicker start in the phase 2 work
Clearly, we need to achieve these synergies while maintaining the independence of our phase 2 work and avoiding confirmation bias. We achieve this by making sure that phase 2 inquiries are directed by, and decisions are taken by, panels of independent members that are not involved in the phase 1 decision. To my knowledge this feature of the process is quite unique in Europe.
While we are on the subject of synergies, I should say that we also intend to make the most of our relationships with sectoral regulators. This is obviously relevant to the energy investigation, and it would be relevant for retail banking if we do decide to refer the market.
These relationships can help us in two ways. At the early stages of the investigation, we expect that the regulators will give us a ‘leg up’ by sharing their knowledge of the market and their understanding of the issues. At a later stage in the investigation, we expect that our relationships with regulators could also influence the way we think about remedies. In some circumstances, we might ask the regulators to monitor the implementation of the remedies; or even to take over some aspects of their detailed design.
Now, I am perfectly conscious that working with the regulators also involves certain risks. The purpose of our intervention is partly to provide a fresh outlook and an independent opinion on certain issues, and this could possibly be compromised if we were allowing ourselves to work too closely with the existing regulatory bodies. But I think that this risk can be avoided by concentrating the interaction at the beginning and the end of the inquiries, while forcing ourselves to ‘withdraw’ from this exchange at the points where we develop our analysis and we reach our conclusions. What we are trying to do, fundamentally, is to combine our strengths with those of the regulators. We are trying to conduct these inquiries so that they complement and ‘invigorate’ the overall regulatory process. We are not trying to replace this process, unless the need for regulation has passed away completely; which is unlikely in most contexts.
The conclusion of a market investigation and the follow-up
So I am now turning to my last point; the choice of remedies and the follow-up on the remedies. Whether remedies are necessary, and the identification of the right remedy, is highly dependent on the facts and context of a particular investigation and requires the exercise of judgement by the inquiry group.
The use of structural remedies
In terms of their potential choice of remedies, I want to focus on the use of structural remedies, not because they are our preferred option (they are not) but because they have been a particularly controversial aspect of interventions by the Competition Commission. As you know, the Commission required divestments in a number of inquiries (airports is perhaps the most well-known) but divestments have also been required in relation to aggregates and private healthcare; and such remedies have been sometimes criticized as being ‘too draconian’. As a rule, I do not think that we should have any qualms about using tough measures if we can demonstrate that they are necessary, effective and proportionate. But my main issue with this criticism is that the precise nature of the problem is often left unclear; what do people mean when they say that divestments are ‘too draconian’, and what are the implications for competition? I think we need to ‘unpack’ this claim if we want to have a useful debate.
In what I am about to say, I do not want to suggest that we do not regard requiring divestments as an intrusive measure; or that we should not exercise particular care in our analysis of the competition problem, and the range of potential remedies for the problem, before satisfying ourselves as to the necessity and proportionality of such a remedy in a particular case. And as the Courts have recognised in airports, it is important that we allow assets to be sold for a fair market value.
However, I would like to try to remove some of the hyperbole that can enter the debate by making some comments from an economic perspective.
When people say that divestments are draconian I think they mean one of two things: one is that divestments impose a very high cost on the firm that has to divest the assets, and this is unlikely to outweigh the benefits of the intervention; the other concern is that divestments represent a particularly intrusive form of intervention in markets, and this could somehow result in unintended consequences or perverse incentives.
Let me look briefly at the first claim. If you ignore for a moment the effect of synergies and transaction costs, in principle a divestment should not harm the selling firm. In economics jargon, the firm should be indifferent between holding this asset and selling it at a fair price.
In this context, the only reason why it may incur a loss is if its previous valuation incorporated the effect of market power, and the divestment reduces this market power; but that is exactly the effect intended by the measure! Now, I am conscious that this is a somewhat theoretical argument, and I will highlight some of the complications in a minute. I am just using this argument to point out that a reduction in the value of the selling firm is actually something that we would expect, and that it is not necessarily a sign that we got the remedy wrong. Indeed, it can be a sign that we actually got the remedy right.
The second claim, that divestments are a particularly intrusive form of intervention, also needs to be seen in context. A divestment is a one-off intervention. The fundamental premise is that once we have corrected an initial flaw in the market structure, we can trust the market to respond to consumer needs, and we do not need to further constrain the behaviour of market participants. If you think about it for a moment, this approach reflects a higher degree of faith in market processes than other forms of remedies; which generally imply a direct and sustained constraint on the actions of market participants. I would certainly argue that price caps or product bans are much more interventionist than divestments.
Now, I need to pause here to emphasize that I am conscious that this is a rather simplistic analysis. In general, we would expect market structures to evolve organically to maximise efficiency. Firms experiment continuously to achieve the most efficient scale and the most efficient split between functions that are internalized and functions that are externalized.
So by mandating divestments, we potentially disrupt that process of experimentation; and we potentially lose the resulting efficiencies. I am perfectly conscious of these risks, and as a rule I would expect divestments to remain relatively rare.
I should also stress that we recognise that there is no straightforward relationship between concentration and competition. Our analysis of competition is much more subtle than this. In the recent past we have cleared a number of mergers that resulted in significant increases in concentration because we thought that the market arrangements implied that this higher concentration would not translate into market power. So you can be reassured that we are not obsessed about concentration, and we are not embarking on a new crusade again big firms.
Following-up on remedies
I am now turning to my final point about how we intend to follow-up on remedies. What we are trying to do, essentially, is to build a sense of ongoing responsibility around the measures for which we are responsible. At the moment we are monitoring 64 undertakings and orders covering 27 markets. Not all of these measures were adopted following market inquiries, some were taken following phase 1 studies, and others date back to older regimes that existed before market inquiries. We have a statutory duty to keep these undertakings and orders under review, and we conduct a number of reviews per year.
Evaluating the effect of remedies is not easy. We need to wait for a reasonable period of time before assessing the impact. We need to gather sufficient information, even though we do not have formal information-gathering powers at that stage. And we need to form an opinion on what would happen without the remedy, which can be difficult conceptually. But I think that it is important that we continue to devote resources to this exercise.
This concludes my remarks. I have tried to give you an overview of our thinking on market investigations. We are ambitious, but also careful and pragmatic. We never start with a preconceived idea of the outcomes, and we certainly do not aspire to set up idealized, perfectly competitive markets. We try to be as transparent as possible about our objectives and our reasoning, so that they can be scrutinized and challenged; not just in formal inquiry settings but also in more informal forums like this one. So we are always grateful for comments and I am very much looking forward to the discussion.