2.1. Limiting Exit and Self-Selection

In the process of IRC, subjects should be randomly and equally distributed among the regulating contestants. IRC requires that the subjects of the developing and competing regulatory regimes are not able to opt out. The effectiveness of regulation is often highly influenced by the asset-specificity of the investment on the part of the regulated subjects. If the subject of the regulation is not attached to the regulated jurisdiction through specific investments that cannot be easily relocated to another jurisdiction, IRC processes might fail because subjects will prefer to migrate to another jurisdiction with seemingly more beneficial regulation. Preventing the exit of subjects from their assigned regulatory regime to another regime is a prerequisite for a successful IRC process.Footnote ²⁴

The main driver towards migration is the fact that IRC implicates the inherent risks of an unstable regulatory horizon. Instability emerges either because IRC contestants and regulatory subjects may naturally adapt to the changes in the regulated environment or as a consequence of exogenous forces – for example, policy choices in other jurisdictions. This is a drawback of IRC processes which must be considered and weighed in the final evaluation stage. However, this drawback should not be exaggerated. Even without using IRC as the tool for generating regulation, jurisdictions regulating under conditions of uncertainty are likely to produce different regulatory regimes and face the risks of subject migration. IRC can succeed only if subject participation is mandatory, but even if this is the case it would be difficult to completely prevent subjects from exiting to other jurisdictions.Footnote ²⁵

Another problem is that sophisticated subjects may try to strategically self-select into particular regimes. In order to avoid self-selection bias, the subjects themselves should not be able to decide which regulatory regime applies to them.Footnote ²⁶ Nevertheless, the self-selection bias in IRC is less troubling than self-selection in experimental law.Footnote ²⁷ The reason for this is that given uncertainty about the methodology and the efficiency of potential regulation, the subjects as well as the regulators may be hopeful but clueless about the consequences.Footnote ²⁸ It therefore follows that subjects are less likely to engage in strategic regime selection and, even if they do, there is a good chance that this effort will be ineffective.Footnote ²⁹ Nevertheless, random assignment should be imposed for the sake of ex ante equal treatment of the regulated subjects.Footnote ³⁰

2.2. Policy Objectives and Evaluation

Legislative mandates for regulatory agencies may be complex,Footnote ³¹ and IRC goals must encapsulate these mandates in order to guarantee performance. The objective of the policy that is being subjected to an IRC process matters. IRC would be better used with policies that are intended to facilitate positive results rather than negative results.Footnote ³² For example, it would be better to generate an IRC contest when the policy implementation sought should produce more activity rather than to completely stop or to stifle activities or occurrences. The reason is simple: increased or decreased activity levels are easier to monitor and evaluate compared with estimating the occurrence of events if regulation had not been in force.

IRC processes must include criteria for evaluating the effects of the competing regimes.Footnote ³³ It would be better to restrict IRC to fields in which the outcomes are likely to be at least identifiable, if not directly visible and measurable. Service industries tend to be better targets for competitive regulatory processes because the effects of regulation are visible – in either the products, the services or various disclosure and transparency requirements – making regulatory races to the top more likely.Footnote ³⁴

2.3. Duration

The duration of the IRC is one of the most challenging implementation issues. An IRC process is broadly composed of two time frames. In the first, IRC contestants need to come up with the proposed regime. The duration of this part matters only in so far as the regulation is urgent. Naturally, the more time the contestants have in which to develop and mould their proposed regulatory strategies, the better. The development of regulatory reforms, from the seed stage to the point at which they are ready for implementation, can take a short period of time, but may also last for years.Footnote ³⁵ The time required for moulding reforms into operational regulation or law depends on the complexity of the subject-matter, opposition from interest groups, public involvement in the process, and political pressures. Reform processes are prolonged by the mere fact that they are designed to change existing regulation and yet they do not or cannot guarantee that the change will be positive in terms of policy outcomes.

Similar obstacles may emerge in the process of setting the stage for IRC implementation. However, in the case of IRC these factors are less likely to stall the process. First, if IRC is called for in the absence of any regulation or after regulatory breakdown, the opposition is likely to be weaker or dispersed. Second, to the extent that interest groups exist, the multi-party nature of IRC can enable their incorporation into the developmental stage and dismantle objections and stalling tactics.Footnote ³⁶

Sunset provisions are perceived to be a fundamental and advantageous characteristic of experimental rules.Footnote ³⁷ The second time frame, in which the regimes are imposed for a limited period of time on the randomly chosen subjects followed by a sunset mechanism, requires delicate calibration. If the duration is too short, temporary regulatory regimes may fail to capture the long-term effects of the new regulation and the ability of subjects to react dynamically to a change in regulation.Footnote ³⁸ Therefore, an IRC process may just as well last several years.Footnote ³⁹ However, when policies have unknown pay-offs, it would be harder to justify their imposition over a long period of time without a guarantee that they will eventually expire.Footnote ⁴⁰ In addition, the more disparate the regulatory regimes and the burden they impose, the less that subjects will be willing to accept the legitimacy of the process for a long duration. The legitimacy of IRC processes will be inversely related to the degree of difference between the competing IRC regimes, assuming that the burden of the regimes will be felt by the subjects of the regulation at a relatively early stage of the process. On the other hand, the outcomes of the IRC may not be visible immediately. In such a case, compensation mechanisms should be installed to guarantee equal treatment at the end of the process and diffuse early-stage opposition.

The decision about the length of the process can be relegated to a predetermined meta-regulatory agency or an internal IRC referee that will supervise the IRC process; in such a case there is no need to set a duration limit in advance, and the cut-off decision will be at the sole discretion of the referee. Considerations that may guide such a cut-off decision include, inter alia, the failure of a competing regulatory regime or the strict and uncontroversial superiority of one regime over the other, or the accumulation of sufficient information for analysing the long-term implications of the developing policies vis-à-vis the goals they are intended to meet.

2.4. Incentives

IRC must produce sufficient motivation for the contestants to invest in innovation; otherwise, they may shirk or collude.Footnote ⁴¹ One possible incentive for the contestants could be promising the developers of the superior regulatory regime long-term benefits such as a career role as implementers of the policy.

The assumption that career considerations could motivate better regulation is an acceptable premise in public choice theory.Footnote ⁴² This type of carrot is analogous to the monopoly granted to the developer of an invention through a patent regime. A patent guarantees the inventor a monopoly over the revenue stream from the invention for a given period, provided that the inventor actually moves to implement the invention. In addition to this incentive, granting the ‘winner’ of the IRC a career role of implementing the superior policy increases the incentive to develop a policy that has sustainable effectiveness, because the victorious regulator will have to continue to implement this regime throughout the next stage of his or her career.

Professional career incentives may reduce the problem of politicisation of regulation, which is one of the main concerns raised by Stephen Breyer,Footnote ⁴³ because political gains will be substituted by non-political benefits that will motivate professional rather than political actors to participate in innovative rulemaking.

The design of career-based incentives should take into account at least two significant factors. The first is the length of the average career of a regulator in similar roles. Guaranteeing the winner a long-term career would be futile if the average regulator prefers to serve a short term in that position and then advance to another position or switch to the private sector. Another factor that must be considered is the material and reputational benefits associated with the career prize. The lower the relative benefits, the lower the motivation to invest in innovation.

Finally, the assumption that the developers of an innovative regulatory regime are necessarily its optimal implementers is by no means always correct. Some authors suggest that the opposite is true,Footnote ⁴⁴ namely that creativity and management skills are negatively correlated. Hence, a superior regulatory regime may fail if it is executed by individuals who lack the appropriate capabilities for implementation. Agencies should take this risk into account and, where relevant, award an alternative prize to the developers of the superior regulation and entrust the execution of the regime to more capable hands.

4.1. Random and Experimental Law

In recent years arguments have been made in the literature in support of randomising within the law and experimenting in law.Footnote ⁷¹ These two theories are closely related and have some common elements with IRC. In this section I explore the similarities and differences between IRC and these two theories.

Randomised experiments, involving a treated group and a control group, are a good way to resolve a debate on whether a proposed regulation may work better than an existing regulation by either (a) subjecting a group of subjects to a proposed new regulation, or (b) by imposing the new regulation on the market and then exempting a group of subjects so that it becomes the control group.Footnote ⁷² Randomising law focuses on the imposition of a regulation on randomly selected subjects as opposed to the entire subject population, whereas experimental law stresses the sunsetting element – namely, that regulation is imposed on all subjects for a defined time span, which alleviates the apprehension of lawmakers with respect to costly regulation with unknown results.Footnote ⁷³

IRC is distinguishable from both regulation methods, although it incorporates and builds on some of their fundamental traits. It builds on randomising regulation in that it is fundamentally a real-life experiment in regulatory solutions. However, the existing literature on experimental random regulation focuses on testing already proposed but contested regulatory solutions. IRC aims to tackle regulatory voids that are so fraught with uncertainty that both regulators and lawmakers prefer to avoid their regulation altogether in spite of the fact that there are strong reasons to apply a regulatory regime to the pertinent field.

Using the analogy between IRC and laboratory experimentation, existing literature on experimental and random law deals with the stage at which the treatment has already passed laboratory trials and is now ready for human testing. In essence, IRC deals with a situation where the patient cannot wait for laboratory testing and animal testing is unacceptable. Of course, IRC would also be useful when regulatory policies do exist, but it is impossible to rank them in terms of superiority and efficiency because of the intensity of regulatory unknowns.Footnote ⁷⁴

Zachary Gubler claims that only a miniscule part of regulation is experimental in nature in spite of the fact that using experimental mechanisms makes sense.Footnote ⁷⁵ This scarcity of experimental legislation is explained, inter alia, by the fact that such regulation may be less favoured by well-organised interest groups.Footnote ⁷⁶ IRC, as it is proposed here, can dissipate some of the objections of such interest groups or succeed in spite of potential opposition. For example, in emerging fields of regulation, coalitions and interest groups may not be fully organised; thus, IRC is likely to encounter less opposition. On the other hand, in fields in which regulation has become obsolete, the introduction of IRC may benefit existing interest groups because existing market actors sometimes suffer from the costs of obsolete and ineffective regulation, whereas new entrants may have found ways to avoid this regulation and supply their products at lower than the pre-existing competitive prices.Footnote ⁷⁷

The mechanism of regulatory experimentation proposed by Romano in the context of regulating global financial markets is similar to IRC in some of its stages. Romano suggests that regulators start with a regime backed by econometric or formal modelling; include a review process, ongoing monitoring and periodic reassessment; and finally make all the information collected in the process available for public scrutiny.Footnote ⁷⁸ IRC follows this model with the fundamental exception that the suggested regulation is likely to be far more loosely backed by formal or empirical data, given the uncertainty and novelty of the regulation. IRC in itself is a method of generating and collecting information. The periodic review process would focus primarily on whether sufficient information about the competing regimes has been accumulated in order to evaluate which is the superior regime. Eventually, in IRC the law should ensure that the superior regime is actually adopted and that knowhow from the overall IRC process is implemented in the adopted regime.

A drawback that proponents of experimental law face is the impossibility of applying the ‘double-blind’ standard of experiments. The double-blind standard is the purest form of unbiased experimentation in that it prevents both the researcher and the subjects from knowing which treatment is being administered.Footnote ⁷⁹ Indeed, whereas double-blindness is a major component of experimentation, it is not present in IRC. However, IRC is not experimental but rather is developmental in its nature; therefore, the fact that the subjects know and may react to the regulation imposed on them is not necessarily a drawback. Nevertheless, at least at the initial stages, the subjects may be as blind to the implications of developing regulatory regimes as are the competing regulators in applying the regimes. Conversely, the subjects of IRC that do have private information about the efficacy of a certain regulatory regime have an incentive to share at least some of it – information that will lower the imposed costs of regulation – with the regulator, thereby improving the cost-effectiveness of the regulation.

IRC has an advantage over randomising law in that it can be applied on both large and small scales. Whereas randomising law requires the identification of similar large-scale groups of subjects,Footnote ⁸⁰ IRC can be applied to significantly smaller constituencies. Even in markets in which there are only a few competing firms, as is sometimes the case in highly regulated network industries, IRC may be applicable. Of course, applying IRC in small-scale markets is not optimal in terms of drawing meaningful results about the efficacy of innovative regulation, but it is important to bear in mind that the primary role of IRC is to create a process for the development of non-existent regulation.

Compared with experimental law, an IRC initiative to generate a race towards the most efficient regulation has the additional benefit of creating positive externalities.Footnote ⁸¹ Knowledge acquired through the experience of one regulatory regime may be later implemented to improve the prevailing regime, in a process of cumulative knowledge generation.Footnote ⁸² IRC has an advantage in that it produces information about more than one regulatory regime at the same time, and the information about the competing regimes is simultaneously comparable. On the down side, particularly if dynamic and reactive regulation is allowed, the information is likely to be less accurate and less conclusive compared with random experimental regulation.

4.2. Inter-jurisdictional Competition

Another strand of regulation literature deals with inter-jurisdictional competition – sometimes referred to as ‘international regulatory competition’ or ‘federalism’ – as a mechanism that creates state-level laboratories of innovative regulation.Footnote ⁸³ In this context, the term ‘regulatory competition’ is often used to define a situation in which states or other smaller jurisdictions ‘vie with each other to retain or attract investment within their jurisdictions by adjusting their regulatory regimes, and firms engage in regulatory arbitrage by moving capital or relocating accordingly’.Footnote ⁸⁴ This type of competition has been heralded as an important mechanism for producing innovative regulation.Footnote ⁸⁵ On the other hand, some scholars argue that global harmonisation and integrated regulation is better than competition.Footnote ⁸⁶ While this may hold true in some areas of regulation, particularly when there is at least some degree of agreement on what are the ‘correct’ methods of regulation, it is not a viable approach when the regulation is obsolete or non-existent. Harmonisation and stability become less relevant the more the market becomes fraught with uncertainties and unknowns.

Inter-jurisdictional competitive rulemaking is an imperfect solution for regulating new fields, to say the least.Footnote ⁸⁷ Inter-jurisdictional regulatory competition often generates regulatory arbitrage: subjects choose the privately beneficial but not necessarily the social welfare-enhancing regulatory regime by exiting from the regulating jurisdiction and migrating to the desired jurisdiction.Footnote ⁸⁸ In fact, as opposed to IRC, for regulatory competition to work subjects must be able to exit from the regulated jurisdiction and migrate to a competing jurisdiction fairly easily.Footnote ⁸⁹ However, the jury is still out as to whether inter-jurisdictional regulatory competition creates a race to the bottom or a race to the top.Footnote ⁹⁰ A finer observation asserts that regulatory competition yields convergence towards either a less strict common denominator or towards the most stringent regulation – if any convergence happens at all. Arguably, regulation affecting production costs converges to the least strict common denominator because market participants exert pressure on regulators to reduce the cost of regulation. On the other hand, regulation affecting access to the market tends to be stringent because it insulates domestic competitors against foreign competitors.Footnote ⁹¹ Hence, there is indecisive evidence as to whether inter-jurisdictional regulatory competition will improve regulation and, either way, it may fit only particular types of policy.

Disallowing exit options forces the subjects to interact with the regulators and share at least some information if they want to affect the calibration of the regulation. This is a significant advantage for IRC over competition between jurisdictions in which subjects of regulation may exit from the jurisdiction without having to share information with the previous regulator about the reasons for the exit and the possible changes that could have prevented that result.Footnote ⁹²

There are additional reasons to suspect that inter-jurisdictional regulatory competition is not always an optimal technique for producing innovative regulation. States may not invest in innovating regulation, but rather prefer to free-ride on the investment of other states.Footnote ⁹³ The free-rider problem means that innovative regulation overall will be suboptimally produced. Furthermore, small countries may be particularly prone to under-investment because of the relatively limited resources and the seemingly cheaper solution of imitating regulation developed by others.Footnote ⁹⁴ This also means that innovative regulation is likely to be produced by larger jurisdictions and consequently be more compatible for larger countries than smaller ones.

5.1. Innovation Costs

Reducing costs is always a dominant issue in discussing regulation. IRC seems costly in comparison with other alternatives, primarily because of the duplicate regulatory processes. The organisational duplication is perceived intuitively to be the most wasteful element of IRC.Footnote ⁹⁸ Duplicity and regulatory overlap are often the reasons behind efforts to merge existing regulatory agencies, although merging crippled or obsolete regulators does not guarantee that the new merged regulator will necessarily be more efficient.Footnote ⁹⁹ In fact, efficiency arguments often ignore the fact that regulatory duplicity can increase the desired effects of regulation.Footnote ¹⁰⁰ The strongest counter-argument against the cost criticism, however, is the short-term versus long-term perspective. While IRC may be a costly measure in the short term, this cost would be outweighed in the long term by the increase in the efficiency and efficacy of the developed regulation.

Other cost-related arguments focus on the competitors and their behaviour during the IRC process, for example, by trying to window-dress their results, thereby leading to bias in the final outcome. IRC may also turn out to be costly and inefficient if competing regulators engage in predatory behaviour in an attempt to undermine the policy outcomes of their opponents. Such behaviour is typical in market competition and is therefore likely to emerge in IRC processes. Contestants may also have an incentive to collude, for example, by trying to prolong the process if they receive rents throughout the duration of the process. They may also independently develop similar regimes, in which case the justification for a costly competitive implementation process disappears. However, the underlying conclusion in the latter case may be that the identical or similar proposed regulatory solution is consensually the best remedy for the issue at hand, and that all IRC competitors should share the reward for its development.

Turf wars may also emerge. If the IRC process has already started, part of the process may require deciding within the ambit of which IRC competitor a new market entrant will fall. If the additional entrant increases the chance of regulatory success, the competitors are likely to engage in a turf war over the inclusion of the actor within their jurisdiction. If the additional entrant is seen as a cost, the actors will engage in a wasteful fight to push the participant towards the opponent's jurisdiction. A possible solution would be to arrive at these decisions by lottery.

The costs inflicted by competitors acting strategically can be mitigated in two ways: (i) by choosing observable evaluation standards, so that monitoring the IRC process throughout its duration becomes relatively easy, and (ii) by appointing a meta-regulator or referee to supervise the IRC process, with the authority to intervene in the event of foul play.

Finally, there are costs imposed on regulated subjects. A drawback of IRC is that potential market entrants may be deterred by the existence of an IRC process. This is an expected consequence of regulation and is especially aggravated when there is uncertainty as to the type and nature of the potential regulation. Hence, IRC would be most justified in markets that lack effective regulation but already have a sufficient number of actors, and where the risk of exit by one or more of these actors as a result of new regulation is low. Having noted that, the risk of deterring potential entrants should not be exaggerated. First of all, market entrants always expect some form of regulation. Second, the risk of deterrence could be allayed by structuring the aims of the IRC to include reducing barriers to entry as one of the evaluated components of the competing regulations.

In the short term, IRC may result in harm to some of the subjects of the regulation. This does not necessarily imply that the harmed subjects are those that were subjected to the less successful regime. It might very well be that the regime that turns out to be superior from a social welfare perspective at the end of the IRC process is the regime that imposed more costs on the regulated subjects during the process; this means conversely that the subjects under the inferior regime were better off during the IRC period. It would be sensible to include some form of compensatory mechanism within the framework of the IRC to equalise losses and gains by subjects of competing regulatory regimes after the IRC process is over. Equalisation payments should also take into account the fact that the subjects of the inferior (and therefore unadopted) regime will have to invest additional costs in order to shift compliance to the superior, adopted, regime.Footnote ¹⁰¹

The structure of the compensation mechanism and its financial sources can take many forms depending on the regulated market, the expected length of the IRC, the expected harm to regulated subjects and the risk of attrition. So where the risk of attrition is high the mechanism could be announced in advance. Where no risk of attrition exists, perhaps a compensation mechanism is superfluous. Funds for compensation payments could be accumulated by imposing taxes or fees on the regulated subjects. The distribution of payments should take place at the end of the IRC process. This type of funding is suitable when the benefits of better regulation accrue particularly to market participants, such as in the case of regulation of financial intermediaries. In other cases, compensation would have to come out of the public budget. No compensation mechanism would be able to perfectly restore the equilibrium between the regulated subjects to the market's pre-IRC starting point. Some regulated subjects would probably lose their share of market dominance following the conclusion of the IRC process without any possibility of recovering this position regardless of any compensation.Footnote ¹⁰² However, a compensation/equalisation mechanism is important because it ameliorates many of the costs associated with potential strategic behaviour by the regulated subjects.

Finally, there is some probability that splitting the regulated subject population between competing regulators may stifle market innovation in the underlying field.Footnote ¹⁰³ In other words, there is a trade-off between innovation in regulation and innovation in the underlying regulated field. This loss of market innovation may occur if a regulated entity comes up with a welfare-increasing market innovation that is not permissible under its pertinent regime but would have been acceptable under the competing regime.Footnote ¹⁰⁴ In such a case the innovation may not be taken up during the IRC period. Assuming that the IRC process will produce better welfare-enhancing regulation, the trade-off is likely to justify the means because the social and private loss from the stifling of some market innovation is limited. The reason is simple: if the particular market innovation is permissible under another competing regime and that other regime prevails at the end of the IRC process, the innovation is likely to be taken up after the IRC concludes, thereby mitigating some of the social loss. However, if the prevailing regime ends up to be that under which the stifled innovation is not permissible, then stifling the development of that innovation imposes no social or private welfare loss.Footnote ¹⁰⁵

Is IRC ill suited for markets that require a boost in terms of competitiveness? Not necessarily. Theoretically, the goal of increasing the level of competition in a particular market either by attracting competitors or by reducing the anti-competitive behaviour of existing market actors can be integrated into the objective function of the IRC process. In practice, markets with low competition levels are usually already heavily regulated and the possibility of introducing costly uncertain regulatory measures will probably invoke strong opposition from the market's population.

At the bottom line, the importance of cost–benefit analysis should not be overstated.Footnote ¹⁰⁶ It rarely captures the costs of voids or regulatory inaction.Footnote ¹⁰⁷ Regulating without stifling productivity or creativity in the market is not an impossible mission.Footnote ¹⁰⁸

5.2. Evaluation

There are two evaluation phases in IRC. During the first phase, the agency conducting the IRC process needs to evaluate the newness of the regulatory regime that each of the contestants proposes to test. Determining whether an innovation is new and novel may be difficult in some cases but this challenge should be less problematic when the process of IRC is applied in a field in which existing regulatory technologies have failed. Even if the regulatory regime proposed by the IRC contestant is an emulation of a tool already in use in another field, it might be innovative to apply it to the new field, given that this has never been done before.Footnote ¹⁰⁹

The second and more challenging evaluation takes place at the end of the process, after the implementation phase of the competing regimes. The evaluation of the effects of the different IRC regimes, for the purpose of determining which is the prevailing regulatory regime, depends on the complexity of the regulated subject, the visibility of the effect and the level of expertise required for the assessment. Evaluation by a group of well-informed objective experts is a good solution, but an evaluation by the community at large should also be considered, where possible, because it reinforces the community's support for the chosen regulation and increases its faith in the process.Footnote ¹¹⁰

One possible critique of IRC is that it resembles current empirical policy evaluations in that it tests policies without comparing them with a population with neutral treatment.Footnote ¹¹¹ This criticism holds water only if IRC is perceived as a tool for merely evaluating policies rather than as a method of developing innovative regulation. Indeed, IRC is not a mechanism for evaluating a particular policy; it is rather a mechanism for creating better regulation.

5.3. Equality

The issue of equality is a challenging legal obstacle for adopting IRC as a regulatory development tool,Footnote ¹¹² because IRC deliberately applies different regulation to identical subjects. However, IRC is founded on ex ante equal treatment. Although IRC means that identical or similar subjects will be regulated by different regimes, the ex ante pairing of subjects with regulatory regimes is random,Footnote ¹¹³ the efficacy of the imposed regimes is unknown, but the aspired outcome is identical. In spite of the fact that randomness entails arbitrariness, legal systems do employ ex ante randomness quite often,Footnote ¹¹⁴ although it is generally more acceptable in administrative procedures.Footnote ¹¹⁵ Essentially, IRC is a case of regulating under a veil of ignorance with ex ante equal and benevolent intentions.

Judicial review of each of the IRC regimes being developed according to the ‘arbitrary and capricious’ standard should not be prevented.Footnote ¹¹⁶ However, a court should not strike out one regime simply because another regime has been proposed and implemented by another IRC contestant. Courts should give some latitude to IRC because it concerns fields in which no effective regulation exists, and the lack of regulation is acknowledged by lawmakers and regulators as a void that requires filling, in the public interest.

Judicial review of IRC should focus on ensuring that:

• • the initial terms of the IRC are balanced and fair;

• • resources are fairly allocated to the IRC competitors;

• • no conflict of interest or ‘capture’ exist in the role of the IRC meta-regulator or referee; and

• • the learning process of the IRC is implemented in the final stage of applying the superior regime.Footnote ¹¹⁷

It should also guarantee appropriate compensation mechanisms to the extent that costs and harm may be unevenly spread among subjects of IRC.

The superior regime will, of course, be subject to more scrutinising judicial review,Footnote ¹¹⁸ but should be given some deference by courts in view of the fact that it has been tried and succeeded ‘in the field’.

Assuming that equality concerns are addressed properly by the designers of the IRC process, sophisticated referees/meta-regulators of the process can rank proposed regulatory regimes according to their estimated risk or probability of success. If the assessment of risk is sufficiently grounded, it can justify a departure from the ‘equal size of subject population’ principle. Meta-regulators can then set the size of the subject population allocated to each IRC contestant at an inverse relation to the risk associated with the regulatory regime applied to that group. Such an allocation limits the dangers involved in extremely innovative regulatory ideas without sacrificing the opportunity of letting the creativity of regulators run wild. It follows that in such sophisticated IRC processes, compensation to subjects should also be adjusted in relation to the harm to which they were exposed.

5.4. Entrenchment and Opposition

Path dependence is an obstacle to developing and implementing innovative regulation.Footnote ¹¹⁹ Even if there is no existing regulation, or there is a consensus that the existing regulation is obsolete, innovation is likely to encounter some opposition;Footnote ¹²⁰ interest groups, for example, may attempt to block innovation because they enjoy the benefits of the current regime or lack thereof.

The conditions for IRC should ameliorate some of the concerns. When existing regulation is obsolete or when the regulated field is terra incognita, path dependence opposition is weaker because the path is no longer worth depending on, at least from the regulator's perspective.

One might argue that incumbents in the market may oppose new regulation through IRC processes because they impose additional costs. It should first be noted that these costs would probably be incurred anyway because regulatory voids are not likely to persist forever. Therefore, the level of objection from subjects in unregulated fields should be no different from the normal opposition to any other regulation. Secondly, incumbents that are currently subjected to obsolete regulation are actually more likely to support IRC than object to it. This is because their competitors are likely to be new entrants enjoying the benefits of regulatory arbitrage or circumventing old regulation with new technology.Footnote ¹²¹ Incumbents would therefore prefer either complete deregulation or measures that will equalise the regulatory burden in the market.

Path dependence may occur within the IRC process itself. There is a danger of entrenchment by the subjects and the competing regulators as a result of network effects. That may cause a preference towards a less than optimal regime bcause of path dependence created by the IRC process, if parties that made asset-specific investments on the basis of an inferior regulatory regime stand to lose these investments and resist the adoption of the superior regime. This difficulty is not insurmountable. Unlike the danger of entrenchment or resistance in market-wide experimental laws,Footnote ¹²² in the case of IRC the network effects encompass at most only half of the market and they are therefore less powerful in comparison with market-wide path dependence. This concern can be further mitigated by integrating compensatory mechanisms, which will guarantee that subjects will engage in desirable asset-specific investments during the IRC process in spite of the uncertain horizon of the particular regime to which they are subjected.

5.5. Attrition and Spillover Effects

The attrition of regulated subjects is a concern in experimental law and might be an alarming problem in IRC,Footnote ¹²³ particularly when the IRC process is applied to markets with a small number of subjects. A possible solution is to define the prevention of attrition as one of the guidelines or objectives of the process, such that if significant attrition occurs, the regulatory regime will be deemed to have lost the competition. Obviously, if the risk of attrition is particularly high, IRC may not be a suitable mechanism for innovation in that field.

Proponents of experimental law are concerned with spillover effects from the treated group of subjects to the control group.Footnote ¹²⁴ IRC can prevent arbitrage-related migration of subjects from one regime to another, but it would be hard to prevent spillover effects; nor would it be justified. Assume, for example, that one regulator of media content imposes a regime which causes subjects to develop content that generates significant revenue and, as a result, the subjects of the competing regulator voluntarily generate similar content, thereby making it difficult to distinguish between the two IRC players. A meta-regulator or referee could be granted the power to prevent subjects from voluntarily conforming to regulation developed by a competing regulator on top of the regulation imposed by ‘their’ regulator, but an equally possible conclusion would be that the outcome is an indicator of the superiority of the first regime and therefore the IRC process should be stopped and evaluated.

6.1. Regulating for Cultural Pluralism in (Small) Media Markets

6.2. Example: Labelling versus Educating

How would IRC work in practice? This section does not pretend to provide a comprehensive illustration of an IRC process in action but rather a rough outline of how IRC would look and a sample of the challenges of its design.

The waning power of current regulation is often accompanied by calls for deregulation. Supporters of deregulation argue that in the media industry deregulation allows a thousand flowers to bloom and that a plethora of sources also means that the media consumer is exposed to diverse content.Footnote ¹⁶⁴ This is not necessarily true.Footnote ¹⁶⁵ Hence, the challenge for regulatory agencies is in regulating a market that is severely constrained financially, with strong foreign influence and new technologies, some of which may be located outside the territorial jurisdiction.

How would a media regulator set up an IRC process?

In the first stage, a referee will be appointed to select the competing regimes, randomly divide the market between the selected competing teams, and determine the length of the process.Footnote ¹⁶⁶ For example, the market may include all media platforms and sources providing audiovisual content regardless of length, method of production, country of origin and language, including internet websites. The referee must also outline the criteria for determining the prevailing regime at the end of the process. The criteria in our example may include elements such as viewer attendance on desired content, changes in the production of desired content, growth or attrition of local content providers, share of the budget spent by the regulator in the course of implementing the regime, and so forth.

At this point potential competitors propose regulatory regimes. One competitor may assert that, given the large number of sources available on the internet, the challenge no longer lies in facilitating content but rather in generating attention to various desirable sources.Footnote ¹⁶⁷ This competitor may suggest that transparency is sufficient and focus on generating a system of classifying recommended content on the web, labelling it, and perhaps also promoting it on popular internet-based networks with the use of public opinion influencers. I shall refer to the regulator of this regime as ‘the labeller’. Another competitor may assert that viewers would be incentivised to access desirable sources and content if they are educated on media consumption at a young age. This regime would therefore focus on promoting media education in schools by, inter alia, introducing quotas for the production of local online content appropriate for children. I will refer to the regulator of this regime as ‘the educator’. There may be other proposed types of regulation, but let us assume that the latter two are found by the referee to be the most convincing.

Dividing up market actors between the competitors would be random, but it would be reasonable to ensure that competitors are each assigned a variety of technological platforms. In the past, European regulation of media content focused on the platform through which the consumer received the content. This focus enabled regulators to treat commercial terrestrial television broadcasters differently compared with cable or satellite television broadcasters. Technological developments and convergence have eroded the ability for a straightforward implementation of separate regulatory regimes for different platforms that supply similar services. In Europe, for example, in response to platform convergence, the new Audio Visual Media Services Directive defined ‘audiovisual media service’ in a broad way,Footnote ¹⁶⁸ which aggregates all broadcasters of audiovisual content, including all of the above and probably more. While this allows one regulatory agency in each national jurisdiction to impose a content policy on all audiovisual media service providers within the jurisdiction, it undermines the possibility of having regulators of one platform compete against regulators of another platform.Footnote ¹⁶⁹

The labeller imposes relatively low costs on the regulated sources and its regime may easily apply to content produced locally and internationally. Indeed, foreign content suppliers may find it beneficial to produce high-ranking content if the labelling system turns out to yield more viewers and more attention. The educator, on the other hand, imposes production costs on content suppliers by requiring them to meet quotas. This might lead to some attrition in the market for content production. This regime is also likely to apply only to local content providers unless the educator finds an innovative method of applying the regime to foreign sources. Furthermore, the effectiveness of education needs to be observed over a longer period of time. The differences between the regimes mean that the design of the IRC must allow for a sufficient period for comparative evaluation because the labeller's results may be observable sooner than the educator's results. Oversight throughout the IRC process is also critical. For example, the referee must make sure that the labeller does not label content produced by those foreign content suppliers that fall within the regulatory ambit of the educator. Such oversight will also prevent arbitrage exits, such as content suppliers regulated by the educator migrating to another state in order to be labelled by the labeller.

One cannot ignore the costs. In the production of content, an uncertain regulatory horizon may cause subjects to produce content that aims for the lowest common regulatory denominator in order to guarantee its suitability for any future regulation.Footnote ¹⁷⁰ This is an inefficient result that IRC designers must bear in mind during the formative stages of the process.

Whether imposing stringent regulation on media market actors creates incentives for regulatory arbitrage is debated. Gibbons argues that regulatory arbitrage in the media is influenced by a variety of factors and not by regulation in and of itself.Footnote ¹⁷¹ On the other hand, EU case law indicates that domestic media content regulation does motivate some arbitrage. In the Veronica and TV10 cases, broadcasters chose to locate their transmissions in Luxembourg, outside their destination country, in order to enjoy less stringent programming standards, but they lost their cases in the European Court of Justice (ECJ).Footnote ¹⁷² In the Belgium and VT4 cases, the ECJ delivered an opposite result.Footnote ¹⁷³ Incorporating exit restrictions can mitigate part of this risk. Small markets can also rely in part on the asset specific investments that existing actors in the media market have invested and on other barriers to exit such as unique languages. Nevertheless, IRC processes in media markets that are often characterised by a small number of actors will have to integrate the prevention of attrition into their objectives.

Eventually, and this may require a few years, the regimes’ performance will be measured and the prevailing regime will be identified. Regulating for cultural pluralism in media content can be measured and evaluated.Footnote ¹⁷⁴ As argued earlier, the case for IRC is most compelling when the process is applied to the facilitation of positive rather than negative results.Footnote ¹⁷⁵ In the context of audiovisual content, it would be better to apply IRC to finding ways to generate more domestic content by content producers than to apply it to finding ways to block the flow of foreign content in the domestic audiovisual content market. Either way, in the audiovisual market, the outcome of the regulatory regime should be relatively observable and therefore fairly easy to evaluate.

After the selection of the prevailing regime, a compensation or equalisation mechanism should be activated. This could be organised, for example, through an administrative tribunal that would accept compensation claims. Assume that the labeller prevails. In such a case it would make sense to compensate some of the subjects in the losing regime for costs incurred in the process of complying with production quotas. Content providers that were shut down as a result of attrition should also be compensated in this case. The compensation might resurrect the latter. On the other hand, if the educator prevails, it seems that there would be no need for compensation payments.

7.1. The Regulatory Void

Securities markets are yet another field characterised by dynamic innovations, technological advances such as the use of artificial intelligence, externalities caused by international or inter-jurisdictional competitionFootnote ¹⁷⁶ and, consequently, regulatory voids. A seemingly endless chain of regulatory failures have been unravelled during recent decades, including Enron and its progeny of accounting scandals, the subprime mortgage debacle, and so on. All of these support the argument that current securities regulation needs fundamental changes.Footnote ¹⁷⁷

The introduction of technology into securities trading has fundamentally changed the scene by introducing superior investors or traders – also known as ‘computerised traders’, ‘algorithmic traders’ or ‘high frequency traders’Footnote ¹⁷⁸ – into the market.Footnote ¹⁷⁹ Traders or investors with superior artificial intelligence technology, stronger computing power, and the ability to locate their computers in close proximity to the stock exchange enjoy an advantage over other market participants, thereby undermining the underlying assumption in securities regulation that investors play on a level playing field.Footnote ¹⁸⁰ Although this phenomenon contributes to a dangerous erosion of public trust in the integrity of the securities market, securities regulation is still founded on the premise that investors are all more or less the same and that investor protection should thus focus on disclosure obligations imposed on issuers of securities.Footnote ¹⁸¹

Computerised trading has turned markets into a volatile and dangerous trading scene.Footnote ¹⁸² The so-called ‘Flash Crash’ is the event most associated with this risk. On 6 May 2010 the Dow Jones Industrial Average dropped an unimaginable 9.16 per cent within less than an hour and then recovered most of that fall.Footnote ¹⁸³ The response of US market regulators to this crash was to introduce ‘circuit breaker’ programmes, which pause trade for five minutes whenever market fluctuations exceed 10 per cent during the preceding five minutes.Footnote ¹⁸⁴ Clearly, this measure focuses on mitigating the effects rather than addressing the cause of the problem, which still remains blurred.Footnote ¹⁸⁵ Consequently, in spite of measures taken by regulators to prevent future flash crashes, numerous such crashes have been recorded in the US, as well in other international stock exchanges, which later adopted precautions similar to those adopted by US regulators.Footnote ¹⁸⁶ Experts assess that the phenomenon is far from waning.Footnote ¹⁸⁷ Hence, the production of additional information is required in order to understand the phenomenon and propose innovative measures to address it.

7.2. The Potential for Intra-Regulatory Competition

In the US, securities regulation is dominated by two particularly prominent regulators: the Securities Exchange Commission (SEC) and the Commodity Futures Trading Commission (CFTC).Footnote ¹⁸⁸ This institutional duality can be exploited to generate a meaningful IRC process in order to overcome the current state of regulatory void.

Historically, the dual regulatory structure made some sense. Securities are broadly regarded as means of raising capital or debt whereas the trading in futures contracts is perceived to be primarily a mechanism for pricing commodities and hedging the risk of price volatility.Footnote ¹⁸⁹ However, the two regulators discussed here have always had and still have an identical mission, which is essential for the functioning of both the capital and the commodities markets: to protect the public from fraud and manipulation in transactions of commodity futuresFootnote ¹⁹⁰ and in securities markets.Footnote ¹⁹¹ Nevertheless, the implementation of this identical policy by the two regulatory agencies differs in terms of both jurisdiction and regulatory culture. To exemplify the jurisdictional difference, commodity futures must be traded on a CFTC-registered exchange according to the Commodities Exchange Act,Footnote ¹⁹² which means no trading in commodity futures can take place outside the purview of the CFTC, unlike securities which may be traded on or off a registered stock exchange.Footnote ¹⁹³ As for cultural differences, such as rules-based versus standards-based regulation, these are cited as an obstacle to institutional merger between the SEC and the CFTC.Footnote ¹⁹⁴ On the other hand, the regulatory differences make the two authorities the perfect candidates for developing competing innovative regulatory regimes.

Today, the dichotomy between securities and futures regulation is a fallacy. Securities, derivatives and other financial instruments have changed and evolved, and the original delineation between securities markets and commodity futures markets in the US has disintegrated. Investors may now buy and sell instruments with the same underlying financial logic on any stock exchange. Similar to media markets, the markets for financial instruments have converged.Footnote ¹⁹⁵

In this case, it is not only that the regulated market has converged and is ready for IRC; the institutional foundations are also available. Ever since the creation of the CFTC in 1974 a debate has been ongoing as to whether it should continue to exist separately from the SEC or whether the twain should meet, and be merged.Footnote ¹⁹⁶ I take no part in this debate. However, the existence of two specialised professional regulators means that the institutional infrastructure for IRC is available. Furthermore, the natural tendency of regulators to compete with each other is evidenced by the historic fact that these two agencies, with identical objectives and closely related markets, immediately engaged in turf wars.Footnote ¹⁹⁷ Hence, some competitive power struggle already exists between the agencies. If the two agencies are merged, then an IRC process could be designed within the merged agency. If the two agencies remain separated, they can be pitted one against the other, on the condition that a neutral referee is appointed to oversee the process.

To clarify, some countries adopt a ‘twin peaks’ system of regulation in financial markets in which one regulator is in charge of prudential aspects whereas another regulator is in charge of consumer or investor protection.Footnote ¹⁹⁸ This is not equivalent to the IRC process suggested here. This is because, first, the jurisdiction of the two regulators (according to the twin peaks institutional regulatory structure) is likely to overlap from the point of view of regulated subjects, as opposed to IRC in which regulated subjects must each be clearly assigned to one regulator only. Second, in IRC the competing regulators have an identical objective, whereas in the twin peaks model regulators have different objectives which need to be implemented simultaneously, and they are separated because of concern that a merged regulator may give undue preference to one objective over the other.

An IRC aimed at developing innovative securities market regulation would require a further segmentation of the regulated population. The current division of regulated subjects, traders and investors, between the SEC and the CFTC, is imperfect and would have to be altered. Currently, an investor buying stock in a SEC-regulated exchange is within the SEC's jurisdiction, while the same investor buying futures falls within the CFTC's jurisdiction; this means that those investors operating in both markets are subject to overlapping regulation. Therefore, IRC requires that all investors are pooled and then allocated to one competing regulator or the other, regardless of whether they trade in stocks or futures. This is because traders in commodities tend to be of two types: risk hedgers and speculators.Footnote ¹⁹⁹ Traders in the regular stock exchange are more diverse and include institutional investors and long-term private investors. However, the competitors must develop a regulatory regime that takes into account all types of investor.

The challenges of guaranteeing fairness and regenerating investor trust in securities markets call for the implementation of an IRC process, at least with respect to generating more informed regulation and developing smart ways to regulate trading technology.