One night a man was walking around, stooping and intensively looking at the pavement under his feet. He moved in small circles in the close neighbourhood of the single street lamp that illuminated the otherwise dark alley. A passer-by greeted the man and asked what he was doing. ‘I am looking for my car keys’ the man explained, somewhat to the surprise of the passer-by since the alley was closed for cars. ‘Did you lose the keys here?’ the passer-by inquired cautiously. ‘Oh no’, the man replied, ‘Not here. In the street over there where cars are allowed.’ ‘But why are you looking here then?’ the passer-by asked, now obviously amazed. ‘Because there is no street lamp where I lost my keys’ was the reply.
Sometimes it seems as if legal researchers are like the man who looked for his car keys under the street lamp, rather than at the place where he could expect them to be. They search for answers where they are not to be found, because it is too difficult to search where the answer is to be expected. Or, possibly even worse, they look for what they do not need, because what they need is hard to find.
One example is the question what is the best level to create law. Most law is explicitly created, in the shape of statutes, by-laws, treaties, regulations, directives, contracts, last wills etc. This raises the question what the optimal level of regulation for particular topics is. For instance, should the EU determine which taxes are to be raised, or should this decision be taken by national states or by subnational levels such as the Länder of Germany, or even by municipalities? Should the rules that govern business to consumer contracts be established in the contracts themselves, by national legislators, or by the EU? Should traffic rules be given by municipal legislators, or by national legislators? Or should the traffic rules even be determined on a worldwide scale by a specialized body that only deals with traffic matters?
These are evaluative or normative questions, and to answer them it is necessary to apply a relevant standard to facts. Such a standard might for instance be that law should be created on the most efficient level, and the relevant facts then concern the efficiency attached to every possible level of regulation. However, efficiency is only one of the many possible standards. A serious discussion about the best level of regulation should also deal with the appropriate standard. If we look at legal research about the proper level of regulation within the European Union, we do not find discussions about the appropriate standards, but an appeal to standards which have as a matter of fact been adopted in the founding treaties, such as proportionality and subsidiarity. Moreover, these standards are not standards for the proper level of regulation, but for the factual question who is competent to make law. This recourse to standards which exist as a matter of legal fact and which do not even address the relevant question can well be explained from the situation that the search for standards that ought to be applied can be very hard, as hard as finding one’s car keys at night in a street without lighting. Explainable as it may be, this appeal to existing legal standards is not the right way to go about. Just as it is better to look for one’s car keys at the place where they can be expected rather than at the place where the search is easy, it is better to apply the standard that really answers the question what is the best level of regulation, rather than use the standards that happen to be the ones adopted by the positive law. We will find that the lack of relevant information is a serious limitation on our possibilities to answer the question on which level law should be regulated, but the discovery which knowledge we are lacking can at least suggest a research agenda for the future.
The question what is the best level for creating law asks for a comparison of alternative levels. Since it is not feasible to consider all possible levels of law-making (e.g. an individual person, the head of a family, the main inhabitant of a house, a delegation of inhabitants of a street, a neighbourhood, the city, a province, a state, or a federation, some functional organization such as the Red Cross or the WTO, etc.) it is necessary to single out a limited set of possible levels, and to determine which of these levels is the best one. The difficult question after the best possible level is then replaced by the easier question which level from a limited set is better than the others. This can only work under the presumption that the selected subset includes the best possible level, an assumption that may be false. How should this possibility be dealt with?
Often it is impossible to prove the correctness of some thesis, while it is still possible to adduce sufficient reasons to make it plausible that the thesis is correct. If one’s audience is not prepared to accept the thesis on the basis of the adduced reasons, the burden of proof shifts from the proponent of the thesis to the audience: the audience must produce convincing reasons why the thesis is incorrect despite the reasons that plead for it. If the audience does not succeed in producing these counter-reasons, the thesis counts as justified by default of sufficient counter-reasons.
This style of reasoning in which some thesis is adopted on the basis of convincing, but nevertheless inconclusive evidence is called ‘default reasoning’, because it is assumed by default that there are no exceptional circumstances which make that the thesis is wrong despite the convincing reasons pleading for it. In actual reasoning situations, default reasoning often goes hand in hand with shifts in the burden of proof. To justify a thesis, the proponent of the thesis has to a produce reasons which are convincing as long as no exceptional circumstances have been proven. The opponent of the thesis, including an audience that is not convinced yet, then has the burden of proof for the exceptional circumstances which would make the thesis incorrect.
If we apply this style of default reasoning to the selection of possible levels of regulation, and focus on the regulation of private law within Europe, it might be assumed that, in the absence of counterevidence, the best level of regulation is one of the following three: the level of the EU, of national states, or of private legal agents. That this might be assumed has more to do with the actual practice than with deep reasons. This reliance on actual practice may be justified by pointing out that the existing practice is the result of a kind of evolutionary process in which worse alternatives have been weeded out. Obviously this is not a decisive argument, but it may be strong enough to shift the burden of proof to those who want to introduce other potential legislators. Notice, moreover, that the likelihood of these three alternatives was brought about by narrowing down the research question to the regulation of private law within Europe. Other ways of narrowing the question down might have led to a different selection of alternatives, for instance by leaving out private agents.
On the presumption that private law in Europe should be regulated on one of the three selected levels, the issue at stake is which of these three levels is the best, that is: better than the other two. Notice that making a preliminary selection has changed the research question from an absolute one (‘What is the best level?’) to a comparative one (‘Which level is better than the two other levels?’).
If it is possible, the easiest way to compare alternatives is to measure them on a single scale. If the three alternatives can be positioned on a single scale, running from worst to best, the best alternative can be identified by looking at its position on the scale. Whether this method is available depends on the standard(s) that are used to evaluate the alternatives. In section 3 the utilitarian standard of the maximization of happiness will be defended, and this standard allows, at least in theory, the comparison of the three levels of regulation on a single scale: which level of regulation produces the maximum amount of happiness?
When the relevant standards have been adopted (and justified), these standards determine which data are relevant for the evaluation of the alternatives. Assuming that the utilitarian standard of maximizing happiness is appropriate, the relevant data are easy to point out: for each level of regulation of private law in Europe it should be computed to how much happiness it leads. The level that produces the most happiness is the best one.
As can be seen immediately, it is practically impossible to ascertain which level of regulation produces the highest amount of happiness. There are several reasons for this. It is not easy to measure the happiness of a single person, let alone to quantify this happiness as is required by the demand for the highest total amount of happiness. Moreover, the happiness that will result from regulating law on a particular level may vary from one topic to another topic, depending on a broad and as yet unknown range of different circumstances. One part of these circumstances deserves separate mentioning, namely the development of the world in time. Measuring the happiness that results from a particular level of regulation involves prediction of the future, which is not easy.
In the light of these unsurmountable seeming difficulties, it is tempting to abandon the utilitarian standard and try to determine the best level of regulation in a different way. This temptation should be ignored. There may be reasons to replace the utilitarian standard by one or more different ones, but the difficulty of collecting the relevant data does not belong to these reasons. Remember the story of the man looking for his car keys under the street lamp. That it is difficult to gather the data required for taking a particular kind of decision does not make other kinds of data relevant.
The problem of collecting the relevant data for the application of the utilitarian standard may be huge, but it can be made manageable. One important tool to do so is default reasoning. The standard that happiness should be maximized is too abstract to be applied directly to concrete issues. It must be replaced by more concrete standards that are adapted to specific questions, such as the question on which level to regulate private law in Europe, or even more specific ones. Moreover, these more specific standards should require data that can effectively be collected. And, last but not least, the use of these more specific standards should contribute to the maximization of happiness.
The happiness standard is best seen as a standard to evaluate more concrete standards which are used to evaluate even more concrete standards, and so on until we have arrived at a standard which is sufficiently concrete to be practically applicable. An example of such a chain of ever more concrete standards would be:
- maximize happiness;
- maximize welfare;
- create a really free market;
- eliminate transaction costs;
- harmonize legal rules.
This chain of more and more concrete standards is only valid if all of the following claims are true:
a. the maximization of welfare leads to the maximization of happiness;
b. a truly free market leads to (is necessary for) the maximization of welfare;
c. the elimination of transaction costs leads to (is necessary for) a truly free market;
d. harmonization of legal rules leads to (is necessary for) the elimination of transaction costs.
Whether these claims are actually true should be determined by means of economic (c), empirical psychological (a) and sociological (b, d) research. In this connection, default reasoning plays a crucial rule, because the findings of these sciences will seldom be definitive. However, that does not exclude that it is possible to reach results on which one can with justification rely to develop a policy. These results are justified as long as counterevidence is lacking.
A large part of the discussion about the proper concrete standards on which value judgements should be based concerns issues that are amenable to, and should therefore be established by means of, empirical research. Such empirical research concerning the standard for the best level of regulation is to the present author’s knowledge almost completely lacking. It cannot be replaced by desktop research, unless this desktop research is meta-research about relevant empirical investigations. That does not exclude, however, that the study of literature and other desktop research are relevant for the formulation of hypotheses that can be tested empirically. The sections 4 and 5 deal with such heuristic considerations that can lead to the formulation of hypotheses which ultimately must be tested by empirical research. First, section 3 will explain the ultimate standard that is presupposed in this article, the utilitarian standard that the best act is the one which maximizes happiness.
Utilitarianism is an ethical theory which has become quite popular since the works of Jeremy Bentham (1748-1832) and John Stuart Mill (1806-1873). Many variants of Utilitarianism are in circulation, and there is no room here to discuss even an important subset of them. Therefore, inevitably, the following account of Utilitarianism is highly eclectic. According to the version of Utilitarianism that is presented here, the best act is the one which maximizes mean happiness.
Perhaps the most noteworthy characteristic of Utilitarianism is that it is a consequentialist theory. Acts are evaluated on the basis of their consequences only. The type of the act is in first instance not relevant. For example, an act is not good for the reason that it is polite, but it may be good because it makes some people a little bit more happy. An act is not bad because it is an instance of adultery, but because, and to the extent that, it makes some people less happy.
This does not mean that the kind of act is completely irrelevant. Polite acts tend to make people more happy and therefore politeness is relevant. By default it may be assumed that polite acts lead to more happiness than impolite ones. However, if some polite act under exceptional circumstances decreases the total amount of happiness, this act is wrong. Moreover, there is no reason to balance the politeness against the unhappy consequences, because politeness is only relevant to the extent that it causes happiness.
That Utilitarianism in last instance only has eye for the consequences of acts, and not for the action type under which the act falls is an advantage. It is unclear why the single fact that an act belongs to an action type, independent of the act’s consequences, might make the act good or bad.
Going by its name only, Utilitarianism is in favour of maximizing utility. However, the variant proposed here (and defended by Bentham) recognizes only one kind of utility, namely happiness. Other things such as liberty, equality, wealth, friendship, and all human rights may be useful and valuable too, but they are only useful to the extent that they produce happiness. Happiness is the only intrinsic value; all other values are instrumental. Therefore Utilitarianism aims at the maximization of happiness.
Whether happiness is
the only intrinsic value can be tested by means of a thought experiment.
Suppose that apart from happiness some other intrinsic value exists, freedom
for instance. Then there should be circumstances, at least in theory, where one
should sacrifice some happiness in order to gain more freedom. Because
then it would be clear that freedom is wanted for its own sake, and not to become happier. It seems likely that those who say that they would sometimes prefer freedom above happiness still expect to be more happy with the freedom than with the ‘happiness’ that they sacrificed.
Not only human beings can be happy or unhappy; higher animals are, in varying degrees, also amenable to being more or less happy. Therefore the variant of Utilitarianism proposed here does not distinguish between human and non-human bearers of happiness.
In theory it is possible that the maximization of happiness requires a growth in the population which leads to less mean happiness. Then a choice needs to be made between maximum total happiness and maximum mean happiness. The variant of Utilitarianism proposed here opts for maximum mean happiness.
Another choice that needs to be made concerns the time span for which happiness is to be maximized. Given that Utilitarianism only has eye for consequences this time span must start in the present, but the question is how far it continues. The answer is: for eternity. The happiness of future generations should therefore also be taken into account. However, since the effects of acts in the far future are uncertain, their impact on the evaluation of acts in the present is less than the impact of immediate and highly certain consequences. The farther in the future, the more uncertain the effects and the less impact they have on the evaluation of present acts. Possible effects in the very far future, which are therefore highly uncertain, are to be discounted so strongly that they can be ignored for practical purposes.
Possibly the most controversial aspect of Utilitarianism is that it aims at the maximization of happiness as aggregated over all bearers of happiness. This means that less happiness in one individual can be ‘compensated’ by more happiness in another individual. In this connection each individual counts equally. Nobody’s happiness is more important than the happiness of somebody else.
Many object against this characteristic of Utilitarianism, stating that it is wrong to sacrifice (the interests, or rights of) one individual to (the interests of) another individual. The obvious reply asks whether the latter individual should then be sacrificed to the former. For instance, should the social security of the poor and disadvantaged be sacrificed to the property rights of the rich? Should the properties of innocent citizens be sacrificed to the liberty of the thieves? Obviously, these rhetoric questions cannot decide the issue about the desirability of Utilitarianism, but neither can the rhetoric questions whether we should allow torturing possible terrorists to increase society’s security, or whether (almost) everybody in a rich society should give more than half of his money to poor people in as yet underdeveloped countries. These are complicated issues, but if Utilitarianism is correct, the ultimate standard for answering them is the maximization of average happiness on the long run.
Utilitarianism is much too abstract to function as a standard by means of which actual acts can be evaluated. As pointed out above, we need concrete standards that are pinpointed to specific domains, which require data that can be obtained without too much trouble. The point of Utilitarianism is that in the end all these more concrete standards should be valuable in the sense that their application leads to more happiness. The value of Utilitarianism does not lie in its immediate practical applicability but in its function as basis for all concrete standards and as a pointer to the kind of data that we need to assess these more practical standards.
An important advantage of Utilitarianism that was not mentioned before is that it provides a common yardstick against which all standards, but in theory also all acts can be measured. Moreover, it allows comparison of alternatives, because Utilitarianism recognizes only one intrinsic value. There are no trade-offs to be made. In last instance only one question needs to be answered: what maximizes average happiness in the long run? Incommensurability does not play any role, because all concrete standards and all acts are commensurable. Of course, the lack of relevant data is an important problem in this connection, but that is a problem which can at least in theory be solved by means of empirical research.
Utilitarianism and concrete standards that are based upon it apply to all acts, including legislative acts. A legislative act is good if it increases the average amount of happiness, and the best legislative act is the one which produces most happiness. Three things should be noted in this connection. First, the legislative act that produces most happiness may still be suboptimal, because abstaining from legislation, perhaps in combination with alternative measures, may be still better.
Second, a legislative act should not be identified with the rules that are produced by means of it. An act of legislation may be good even it produces a bad rule, for instance because the legislation proves to the public at large that the legislator is concerned with a particular problem. (‘The government is doing something about this problem.’)
Third, there are good and bad
making characteristics of a legislative act which are not necessarily connected
to the rules that are produced by means of it. The central topic of this
article illustrates this point, because it concerns the level
of legislation, not the content. Level and content may be connected, but they
are different issues. It may, for example, be argued that contract law should
as much as possible be regulated by the contract parties, because this allows
for maximum autonomy in a field (contracts) in which autonomy is very valuable
(leads to more
happiness). This autonomy-based argument does not necessarily have anything to do with the quality of the contractual rules made by the parties.
After having adopted, at least for the purpose of this article, Utilitarianism as the ultimate standard at the hand of which more concrete standards must be evaluated, it is time to take a look at the reasons why private law in Europe should be created at the level of the EU, at the level of national states, or at the level of private legal subjects. There are at least two kind of reasons that are relevant for determining the best level of regulation. One kind has to do with the content of the rules that are being created. The assumption here is that if some level produces better rules than some other level, this is a reason why the former level should make the rules. The reasons based on the quality of the produced rules are discussed in the present section.
The other kind of reasons has to do with the authority, or – maybe better – autonomy, of the rule producing agent. The assumption underlying this kind of reasons is that some agents are somehow entitled to make rules, by and large independent of the quality of the rules they produce. The most familiar candidate in this connection is nowadays that agents should legislate for themselves (autonomy), and that nobody should be subjected to somebody else’s rules. These reasons are the subject of section 5.
If we only want to consider the possibility that rules should be created by the agent who produces the best rules, we should have some clarity about what the best rules are. Does the question what the best rules are have an objective answer, if only objective in the sense that it reflects the will or the interests (not necessarily the same!) of the people who are affected by the rules? Or is it merely a matter of taste what the best rules are, and even whether some particular rule is better than another rule? If it is merely a matter of taste, it makes little sense to point out a rule-making agent for the reason that this agent would make the best rules.
Having adopted the happiness criterion as the ultimate standard for evaluating acts, including decisions with which rules to organize society, the answer to the question what the best rules are is objective. It is not merely a matter of taste which rules create most happiness, although the objective answer to the question what these rules are may depend, amongst others, on the tastes or preferences of people.
Even if the question after the best rules has an objective answer, this answer is hard to establish. Therefore an important factor in determining what the right level of regulation is, is which agent has the best knowledge. This factor will be discussed in section 4.2. Even if an agent knows what the best rules are, it may be unwilling to create those rules, for instance because it has an interest in creating other rules. Here the phenomenon of ‘externalities’ plays a role. Externalities and the way to deal with them are discussed briefly in section 4.3. A third important factor has to do with the phenomenon that rules are not individually good or bad, but as a set. A rule is a good rule if it belongs to a set of rules which as a whole is good. The implications of this phenomenon are the subject of section 4.4. The final factor that will be discussed (in section 4.5) is that the quality of rules may depend on the scope of their application. Some issues should be regulated for large territories; for other issues this is not necessary or even desirable.
In order to make the best possible rules, it is important to know what the best possible rules are. Given the standard of happiness, a rule is the better, the more happiness it produces. The quality of a rule depends therefore on its effects. These effects are not only determined by the content of the rule, but also by the circumstances under which the rule must function. Part of these circumstances are factual ones, such as demographic, geographical and economic circumstances. It is well imaginable that if these circumstances are different in different regions, the best rules for one region, the rules which maximize happiness there, may differ from the best rules for another region.
At first sight the phenomenon that the best rules for one region may differ from the best rules for another region suggests that the rules for which this holds should be made regionally, with as extreme case rules that only apply to contract partners and which should therefore be made by the partners themselves. However, it would be too hasty to draw this conclusion in general, because there is no reason why the relevant knowledge should only be available to local law-makers. Take for instance rules about marriage. Perhaps the Italian culture differs sufficiently from the Swedish culture to justify different rules for Italy and Sweden. Does this mean that the rules for Italy should be made in Italy and those for Sweden in Sweden because the relevant information about a country is only available in that country? Not at all! There is no reason why the reports that are used for the preparation of legislation in some European country should not be made available to for instance the European Commission. There may be some lower boundaries on the available legislative staff to allow for the necessary diffusion of knowledge, but as long as those are respected, the availability of knowledge is no reason to regulate a subject on one level, rather than another level.
On the contrary, from the perspective of available expertise there are good reasons to regulate many subjects as central as possible because the prediction of the effects of regulation requires highly specialized knowledge. The more and the better the staff that is available to collect this knowledge, the better the prospects for good rules. If the EU is sufficiently funded, it seems more likely that the required expertise can be developed on a central level, eventually assisted by local representatives, than on the level of national states.
The above plea for centralized legislation does not apply to rules which only affect a few agents, such as rules in many contracts. For most individually negotiated contracts it holds that the rules affect mostly or even only the contract partners and that these partners have better knowledge of the involved interests than anybody else. From the perspective of expertise, this pleads for leaving the determination of the consequences of contracts to the contract partners. However, if contracts are not individually negotiated, there may be reason to have their content – or relevant mandatory law – determined on a more central level.
Even if an agent knows what the best rules are, there is no guarantee that it will create these rules. If we ignore the possibility of malevolence, the creation of ‘wrong’ rules may be the result of externalities. Externalities are effects of acts on interests which are not the interests of the agent, and which therefore run the risk of not being taken into account by the agent. The best rules promote overall happiness, but not necessarily the maximum amount of happiness for the law-creating agent or the persons it feels responsible for. Even if a national legislator promotes the happiness of the inhabitants of the state, it may ignore the from an utilitarian point of view equally important happiness of non-inhabitants. Contract partners are prone to ignore other interests than their own. Functional legislators, such as for instance water boards, may not take other interests than those connected to their functions sufficiently into account.
One solution to this problem of externalities is to make sure that they are ‘internalized’, that they also become interests of the agent and are consequently being taken into account in deciding which acts to perform. The fact that an agent has internalized possible externalities may be a reason, not necessarily a conclusive one, to point out this agent as the proper level of legislation.
Another solution is to make sure that there are as few externalities as possible by having a rule-making agent who feels responsible for as many as possible potential bearers of happiness. Although it is theoretically possible that national legislators or even private persons take everybody’s happiness into account, it is more likely that central legislators, such as the EU, have eye for everybody’s happiness. The desire to limit the amount of externalities provides a contributory (prima facie) reason why the creation of law should be centralized.
Legal rules do not operate in a vacuum. They are part of a system of rules and the effects of individual rules combine with the effects of other rules, and it is this combined effect that affects the happiness of persons. Therefore it is undesirable to study the effects of isolated rules, unless there are good reasons to assume that these effects are not influenced by the effects of other rules. Legal rules should cohere, work together, and the best rules are those which are elements of the best sets of rules.
The desirability of coherent rules impacts the best level of regulation, at least if it is assumed that rules created by a single legislator are prone to be more coherent than rules stemming from different legislators. An example is the regulation of security property rights. These rules function within a setting of other rules on legal procedure, obligations, and redress. If these other rules are created on the national level and the rules about security property rights were to be created on the European level, the chances are real that at least in some national states the total set of rules becomes incoherent.
The desirability of coherent law may therefore provide us with a reason for having related rules made on the same level. In the case of property security rights that might be the national level, but it might also be the level of the EU.
Rules have scopes that determine, in combination with the rule conditions, on which territory, to which persons, and during which time frame the rules can be applied. International private law determines the scope of rules of private law. The different national systems of criminal law determine to which crimes and to which suspects the rules of these systems apply. Most rules of constitutional law apply to the governmental agencies of a particular country only, but the human rights provisions of constitutional law tend to have a wider scope of application. Laws that are connected to a religion, such as Talmudic law and the Shari’a apply normally to adherents of these religions, while canon law applies to cases which somehow concern the Roman Catholic church.
For our present purposes, the territorial scope of a rule is the most relevant type of scope. It is usually, but not necessarily, determined by the body that created the rule. A national legislator makes rules that can apply to cases within the territory of the national state; rules created by provinces and municipalities can apply to cases within the provinces, respectively the municipalities. Regulations created by the EU apply within the EU member states (Art. 288 TFEU) and there is little reason to assume that this is different for directives which are applied directly to citizens and non-state organizations within the EU. In general, rules with a wider territorial scope prevail in case of a conflict over rules with a smaller scope. EU-law prevails over the law of the EU Member States (CJEU C-6/64; Costa v Enel), national law normally prevails over the law of subnational government agencies such as provinces and municipalities, and the mandatory rules of state law normally prevail over law made by private law makers such as contracts and by-laws, unless the state law has determined otherwise.
As a consequence the level on which law has been be created is determines to a large extent the territorial scope of rules, and therefore also the degree of uniformity of the way in which a topic is regulated in a larger territory. If the rules on a particular topic are created on a central level, with a large territorial scope, the way in which this topic is regulated will in principle be the same over the broad territorial scope of the regulation. If, on the contrary, rules on a topic are created on a highly decentralized level (e.g. contract partners) then the way in which the topic is regulated will normally be highly diverse.
Because circumstances vary from one person to another, from one municipality to another, and from one country to another, the best rules may be different from one unit to another. This may be a reason to make rules that are adapted to the local circumstances and, assuming an automatic ‘translation’ from the scope of the rule to the level of regulation, therefore also a reason to make rules on decentralized level.
Obvious as this may be, there is still a major complication, namely that the size of the units is not something given. What should we take as the units when it comes to deciding whether the rules of one unit should differ from those of another unit? It is tempting to think in terms of states, because for historical reasons those are seen as the bearers of sovereignty. If we set aside these historical contingencies, states are not a priori in an advantaged position relative to sub-states (Länder; gewesten; Kantons), districts, provinces, municipalities, families, or individual persons. One street differs from another street. Should this be a reason why every street should have its own rules, and that therefore at least some rules should be made on the level of streets? There are no two persons who are in exactly the same position. Is this a reason to let every person make his own rules? Obviously not, but the difference between persons is at least a reason for personal autonomy which must be outweighed by other reasons if autonomy is not to decide the issue about the level of regulation.
The main reason for having uniform rules is also a reason for having rules in the first place. Following Westerman it is possible to distinguish between the guiding, the coordinative, the predictive, the evaluative and the institutional function of rules. In their guiding function, rules assist people in deciding what to do. They facilitate the decision making procedure by limiting the amount of possible choices.
In their coordinative function, rules make it possible for agents to coordinate their behaviour. For instance, the owner of a good expects that others will not damage it, while the others will not damage the good because as non-owners they are not allowed to do so.
In their predictive function, rules enable people to predict the behaviour of others. This makes it possible for a landlord to invest money in a house, because he can predict that the lessor will pay the rent.
In their evaluative function, rules make it possible to evaluate or justify behaviour after it was performed. Damaging a good that belongs to somebody else will be evaluated as unlawful, while a landowner will normally be justified in removing buildings that were erected on her land by somebody else without permission.
In their institutional function, finally, rules aim to secure that a particular state of affairs will be accomplished. For instance, car producers must make sure that their cars satisfy a particular safety standard.
In particular in their coordinative, predictive and evaluative functions, rules deal with the relations between different agents; these three functions concern the interaction between persons or organizations. Let us therefore call these three functions the interaction functions of rules.
Rules cannot fulfil their interaction functions well if the group of people within which they function is too small. Take as an extreme example a rule that tells car drivers to drive on the right. If this rule only holds for two car drivers, or only for the drivers in a tiny village, it will hardly function unless there are no other car drivers, or the village is not visited by drivers from outside. The bigger the group of people who use this rule to guide their own behaviour, but also to predict the behaviour of others, the better the rule can fulfil its interaction functions. Since a wider territorial scope normally means a bigger set of people who use the rule, the interaction functions of rules plead for a scope that is as wide as possible. And, assuming again a translation from the scope of rules to the level of regulation, the interaction function also pleads for central regulation.
Transaction costs provide still another reason for a wide scope of rule application and therefore for central regulation. The smaller the territorial scope of rules, the more often legal actors will have to deal with other rules than the ones they are accustomed too. This leads to transaction costs which make that transactions which would have been economically viable without these costs lose their viability. This is an economic loss and in the end normally also a loss in welfare, and – on the assumption that welfare increases happiness – a loss of happiness.
In section 4 we looked at the desirable level of rule production in the light of the quality of the rules produced. A completely different approach is possible, though. This approach does not focus on the quality of law, but on the authority of the law creating agent. The question is not so much who will make the best law, but who is entitled to make law: who is to decide? This question who is entitled to decide seemingly does not make much sense from the utilitarian point of view, because it apparently ignores the consequences of the level on which law is created. The focus is completely on the type of the act, rule creation by agent A, instead of its consequences.
Completely disregarding this authority-based approach would be a bridge too far, however. Several centuries ago Rousseau pleaded for autonomy of the people which would be free if it obeyed its own rules. The underlying assumption here is that it is good for people to be free, or – cast in utilitarian terminology – that being free makes people happy. There is certainly something to be said for this view: most people like to be free, and lack of freedom may cause unhappiness.
However, in its extreme version this view leads to autonomy for every single person. This extreme version is unattractive if only because it completely ignores the interaction functions of rules. Also the less extreme version that pleads for autonomy of a people, rather than of individual persons, has its drawbacks. The fact that a people wants something is insufficient proof that it is also good for the people. Because of false beliefs and irrational attitudes, people often want things that are not good for them. In those cases, following the will of people is not the best way to further their interests, or – to stay in the utilitarian framework – to promote happiness. Giving people what they want is a good idea if people tend to want what makes them happy in the long run, or if they are at least better judges of what will make them happy than the agent that would otherwise make the rules for them. At this point, the argument from autonomy has merged with the argument from expertise that was discussed above (section 4.2).
The argument thus far may have stricken some readers as highly theoretical. It hardly mentioned the EU-oriented setting within which the actual discussion about the best level of regulation takes place, and neither does it pay attention to the practical implications if some field of law will have to be regulated by a rule creating agent which is allegedly ‘better’, but which is not the same as the agent that actually creates the law in this field. Can such a theoretical argument have any practical relevance?
It can, because it is useful to know which ideal to strive for, even if there is no guarantee that it can actually be achieved. Even unachievable ideals can indicate into which direction the actual situation should be developed. However, the relevance of ideals is heightened if attention is paid to side-constraints and to transition. Side-constraints are limits on the number of alternatives that are considered in deciding what is the best course of action, or – in our present discussion – what is the best level of rule creation. In a sense, the above argument made use of side-constraints by limiting the alternatives to three levels of rule creation: the EU, national states and private agents. The reason behind this limitation was that these three are the most likely alternatives given the actual legal practice. However, the actual practice was not taken at face-value, but merely as an indication of the alternatives that should by default be taken as the relevant ones. It is still possible to present reasons for making additions to the list of alternatives that should be taken into account.
Whether this approach, in which the number of alternatives is limited by default, is the right one depends on the kind of enquiry one is engaged in. If this enquiry is from the beginning limited to the alternatives that are available within the framework of actual positive law, a different list should be considered than if the enquiry aims at finding the best level of rule creation in general. For scientists the latter enquiry seems more appropriate than the former; for politicians it is probably the other way round.
There are at least two ways to approach a society in which the average amount of happiness is maximal. The one, implicitly adopted above, is to sketch an ideal situation, or the ideal answer to a particular question, from scratch, and then try to change the current situation into the direction of the ideal. The other approach takes the current situation as its starting point and aims at making continuous improvements that increase the amount of happiness. Both approaches to deal with the transition from the current situation to a hopefully better one have their advantages and disadvantages. Starting from scratch opens one’s eyes for what is theoretical possible and allows to take a step backward in order to take a different road which is better suited to reach the ideal (avoidance of local optima). It has the disadvantage that it easily leads to ‘daydreaming’ and underestimating the constraints that are imposed by what is actually the case. Starting from the actual situation has the corresponding advantages and disadvantages: the risk of being unrealistic concerning what is practically manageable is minimized, but at the cost of possibly losing sight on what might be reached because the complications of the actual situation blind one’s view.
Utilitarianism offers, at least in theory, a solution for this dilemma, because it aims at maximizing happiness in the long run, with a probability discount for future happiness. Any project that moves from one way to organize society to a different way brings costs with it, costs which can in the end be expressed in terms of (a loss of) happiness. These costs are justified by the increase in happiness that results from having the better organization, but only if the costs of change are not too high in comparison to what is to be gained. The happiness calculus does not only provide us with a tool to compare future states of affairs, but also to compare courses of change. As sketched above (section 2), comparative reasoning demands that some alternative courses of change are preliminary selected and that the happiness effects of these courses are calculated. The usual problem of lack of relevant data will present itself, and the usual way to deal with this problem – search for default solutions – will be appropriate.
Which courses of possible change in the current situation concerning the level of rule creation should by default be considered, and which possible solution should be default be adopted are questions beyond the scope of this contribution. The important finding that remains is that the question what is the best level to create law should be replaced by the question what is the best course of transition from the current situation to some better situation.
What is the best level to create law? Ideally we should now be able to answer this question, but our investigations so far have mainly made it clear how much information that is required to answer this question is still lacking. And – perhaps more surprising – it has also made clear that the research question that underlies the present contribution was wrongly formulated from the beginning. It should not have been what is the best level to create law, but what is the best course of transition toward a situation in which the level of law creation is handled appropriately.
That the relevant information is still lacking could only be established in the light of an evaluative standard at the hand of which the question after the appropriate level of rule creation is to be answered. In this contribution the utilitarian standard that happiness should be maximized was more or less taken for granted. Given this standard, the relevant information concerns in last instance what makes people happy and what are the consequences, including the future consequences, of creating law on a particular level. Noteworthy in this connection is that for practical reasoning purposes the happiness standard should be replaced by more concrete standards that are pinpointed towards specific issues and which require data that can relatively easily be obtained. The role of the utilitarian standard is then confined to the justification of these more concrete standards.
Since the lack of much relevant information made it impossible to deal with very concrete issues in connection with rule creation, the main line of the argument had to remain rather abstract. Five factors were identified as relevant for the best level of regulation: the availability of expertise, the willingness to use this expertise to make the rules that maximize overall happiness rather than further the interests of special groups (no externalities), the coherence of the rules that result from the rule creating process, the scope of the rules that are being created, and the actual situation concerning the level on which law is being created. The relevance of the actual situation occasioned the change in the question that was to be answered.
It should be emphasized that these findings are only a first step on the road to a satisfactory theory on the best level of creating law. The more important findings are that our present insights are no more than a starting point for the formulation of hypotheses that can be tested empirically and that the agenda for legally relevant research should contain questions that can to a large extent be answered by empirical research.
It may be worthwhile to compare the results of the present contribution with those of Smits, who raises the question ‘how viable criteria can be found for the optimal distribution of competences among the EU and the member states’. In this connection Smits limits the number of alternative levels of law creation to two levels that were also taken into account in the present contribution. The level of law creation by private agents is ignored, presumably because that level does not play a role in the present discussion about the division of competences within Europe.
It is noteworthy that Smits does not focus on the best distribution of competences, but on ‘viable’ criteria for the optimal distribution of competences. This formulation suggests that what Smits is after consists of the ‘concrete’, or ‘more specific’ standards mentioned above (section 2.3). The present contribution mentioned in this connection three factors, namely whether the standard
- is sufficiently pinpointed to the rule creating act;
- requires data that can relatively easily by acquired;
- is justified in the light of the utilitarian demand that happiness should be maximized.
Smits’ introduction is mainly devoted to stage setting, to illustrate that the central research question does not only have scientific significance, but also addresses a ‘hot’ issue. After this introduction Smits first points out that the existing literature on competences hardly deals with the normative question of which competences belong at which level of government. He points out that political scientists aim to describe and to explain the emergence of authority beyond the state, and that legal scientists seek to classify the competences and try to answer the question how conflicts between the different levels of government should be solved, and arrives at the conclusion that these sciences have little to say about his research question.
This is different for the economic approach. Smits points out that the theory of (fiscal) federalism provides a general normative framework for the assignment of functions to different levels of government. The finding is that ‘things must be dealt with at a higher level if they create cross-border effects … or if economics of scale can be achieved’. This result seems related to the finding of the present contribution that the avoidance of externalities and the coordination function of rules (may) plead for central regulation.
A further result would be that ‘Lower levels of government must provide goods and services of which the consumption is limited to the own jurisdiction because the efficient level of output of a public good will vary across jurisdictions.’ This argument was discussed in section 4.2, where the outcome was that it would be too hasty to draw the conclusion that rules should be made regionally if the best rule for one region might differ from the best rule for another region. There is no automatic translation from a difference in rules to a difference in rule makers.
Another ‘economic’ argument pleading for local government would be that local governments have better knowledge of the preferences of local people and of the local conditions. Apart from the fact that local preferences should according to Utilitarianism not automatically determine what the best rules are, it was found above (section 4.2) that there is no deep reason why knowledge about local circumstances should be better available to local governments.
The economic argument for local government that a central government cannot vary levels of public output across jurisdictions because it is expected to treat everyone as equal is not convincing. Unequal cases should be treated unequally. The issue is which inequalities should be taken into account and in what manner.
The economic argument that local government leads to competition between governments and thereby to limitation of monopolist behaviour by the central government may be true, but local government may also lead to competition which in extreme cases may amount to war. It depends on the circumstances which reason is the strongest. What seems to be lacking in this discussion is a clear higher level criterion at the hand of which sub-criteria such as balance of powers and avoidance of war can be measured. Utilitarianism provides such a higher level criterion.
Having mentioned the economic arguments, Smits concludes that the general insights do not offer much guidance because the optimal solution depends completely on the specific circumstances.
Having found that existing political, legal and economic science is not very helpful for answering his research question. Smits continues to develop his own research agenda. He distinguishes two elements: mapping the existing distribution of authority among the EU and the member states, and developing normative criteria for the optimal assignment of competences for various policy fields. This first element is important because knowledge about the legal status quo is relevant for the comparison of transition courses (cf. section 6 above).
The development of normative criteria is exactly what Smits’ research question asked for and the section of his contribution devoted to this topic (most of section 3.2) may therefore be branded as the core of his paper. Smits identifies two methodological challenges in this connection. The one is the dependence of what is optimal on ‘… valuation of the relevant policy arguments that one aims to achieve in a certain field’. And, Smith continues, one can reasonably argue about which goals are more important in the circumstances of the case. Actually, there are two issues at stake here:
- it must be determined which goals should play a role in determining the best level of regulation; and
- the relative importance of these goals must be established.
In the present paper, the adoption of the utilitarian standard determines how these two questions should be answered. Smits does not adopt that approach, but nevertheless mentions five ‘parameters’ that are relevant for the optimal assignment of competences. One parameter, called ‘dimension’, is whether the competence concerns rule setting, implementation or enforcement.
The second parameter actually consists of nine sub-parameters which mention factors that are relevant for the assignment of competences. These factors are local preferences, responsiveness, the desire to internalize externalities, the capacity of governments at different levels to effectively set, implement and enforce rules, achieving economies of scale, the desire to experiment with different solutions, the risk of a race to the bottom, legitimacy, and legal-systematic arguments. These factors sound plausible, but what is lacking is a general view of what makes them relevant. Lacking such a more general view, it is difficult to determine whether the list is complete, and what might be possible other candidates if the list is not complete. Moreover, it is unlikely that all these sub-parameters are relevant for all different cases, and lacking a more general view, there is no criterion for determining what makes a factor relevant. Smits’ list is useful to get an impression of factors that might be relevant, but it also makes clear why a general view, such as the one offered by Utilitarianism, is crucial.
The third parameter mentioned by Smits is the need to balance the sub-parameters mentioned under two. Smits states that it may be hard to find a common standard on which these factors can be measured, but correctly assumes that rational reasoning about this is still possible. He is correct in this last view, but if Utilitarianism is adopted even a common standard becomes available.
The fourth parameter would concern the goals of European integration, which would influence the balancing of the factors. That these goals are relevant is far from obvious, certainly if one has adopted Utilitarianism. But also if one does not embrace this ethical theory, it is not immediately clear that the balancing of the sub-parameters depends on the goals of European integration.
The fifth parameter rightly mentioned by Smits ‘concerns the extent to which the existing division of competences stands in the way of achieving an optimal distribution’. This is the transition issue, discussed in section 6. The ‘solution’ proposed there (adapted to the terminology of Smits) was to replace the question after the optimal assignment of competences by the question after the optimal course of transition.
Perhaps the most interesting conclusion to be drawn from the comparison between Smits’ approach of the question after the optimal assignment of competences and the approach of the present contribution is that Smits does not adopt an abstract criterion for the evaluation of levels of regulation. In the absence of such a choice it becomes hard, if not outright impossible, to justify any concrete criteria for the assignment of regulatory tasks or competences. Smits mentions nine factors, but many questions (is this list complete; are all factors relevant under all circumstances; how should they be balanced in case of a conflict?) are hard to answer in the absence of an abstract criterion.
Formulating an ultimate evaluative standard such as the utilitarian one cannot be uncontroversial. However, a legal science which sees it as one of its tasks to answer normative or evaluative questions must deal with a standard, and it also must deal with empirical research to acquire the data to which this standard and its specifications refer. It may be difficult to make the proper choice and to obtain the relevant data, but these difficulties should not withhold us from leading legal science in the right direction. Legal scientists should not become like the man who looked for his car keys under the lamp, rather than where he lost them.
Dolan & Peasgood 2010
Dolan, P. & Peasgood, T., ‘Measuring Well-Being for Public Policy: Preferences or Experiences?’, in: E.A. Posner & C.R. Sunstein (eds.), Law & Happiness, Chicago: The University of Chicago Press 2010, p. 5-31.
Hage, J., ‘The Justification of Value Judgments’, in this volume.
Layard, R., Happiness. Lessons from a New Science, London: Penguin Books 2003.
Rawls, J., A Theory of Justice, Oxford: Oxford University Press 1972.
Rousseau, J.-J., Du contrat social, many editions.
Smits, J.M., ‘Who Does What? On the Distribution of Competences among the European Union and the Member States’, in this volume.
Westerman, P., ‘Ruimte door regels’, 61 Ars Aequi, 2012, p. 720-728.
 Hage 2015, section 6.1.
 For examples, see Smits 2015.
 Hage 2015, section 9.
 Hage 2015, sections 10.3 and 12.
 See Hage 2015, the sections 4 and 5.
 However, see Layard 2003, chapter 2, and Dolan & Peasgood 2010, p. 5-31.
 See also the negative findings of Smits 2015.
 This likelihood was not established by me in any scientific fashion, but it is the repeated finding of informally querying students who follow a course on ethics, including Utilitarianism, both in Hasselt (Belgium) and in Maastricht (Netherlands).
 A prominent example is Rawls 1972 that can be read as an extended argument why Utilitarianism is wrong for this very reason.
 Cf. Hage 2015, section 8.3.
 Remember that we are considering only one reason (expertise) to legislate locally, rather than globally. There may be other reasons, for instance based on autonomy. These reasons will be discussed in section 5.
 These boundaries would not be respected if private persons would be the law-makers, as they are in contracts.
 This may be different for open-ended rules, because their effect may depend on local circumstances, such as customs and legitimate expectations. Moreover, although it is not usual, it is not in principal impossible that a national law maker creates rules that only apply in a particular region of a country, or that a supranational body creates rules that only apply in one particular state.
 Westerman 2012, p. 720-728. In translating Westerman’s expressions into English the terminology was slightly changed.
 Rousseau 1797.
 Smits 2015.
 See the discussion of comparative value judgments in Hage 2015, section 9 and the literature mentioned there.