This implies:

• specifying responsibilities of process- and product-managers;
• developing solid and balanced patterns of interaction;
• using fewer directives and incentives to implement policies;
• having a procedure to filter out sub-optimal or irrational decisions.

If these preconditions are not met, even the best steering models remain only partly viable. As the theoretical juridical problem lies deeper, it would be missing the point to read this article along traditional classifications of steering models. We would still be making cars without wheels. To do their work, and not to create impediments, jurists need their own 'shadow-model' to handle new steering concepts that have been developed in the relevant literature. Otherwise, the concepts that are supposed to be applied, can only be seen, they cannot be touched.

3. Characteristics with a view to regulating

3.1 Freedom, autonomy, responsibility: modern steering and legal confusion

It is no accident that we find institutional autonomy only explicitly protected in some new laws in central and eastern Europe. Its precise meaning is rarely defined; how could it be? In the western world, autonomy is at most implicitly regulated by the answers of the legislature to the question: 'who is in charge of specific issues.' Only academic freedom is explicitly protected.
In some countries of central and eastern Europe, the pendulum of the newly won institutional autonomy, (individual professors enjoyed considerable academic freedom, especially in the natural sciences) swung after the political changes in 89-90 so rapidly, that hardly anything was left of political responsibility and thus of political presenee at government level. For some time (Baltic) countries even had no ministry responsible for higher education, thereby effectively jeopardising a fair share of funding through the national budget. To protect regained autonomy and academic freedom, the Czech and Slovak Republic(s) introduced the legal personality of the faculty, until then only known in (the countries of the former) Yugoslavia. lt seemed a reasonable demand of the oppressed faculty, but at the risk of making the university a half-empty shell. The faculties were granted de iure as well as de facto autonomy, illustrated by, for example, direct negotiations between faculties and the ministry. This change creates impediments to internal synergy, national and international co-operation, and deterioration of the institutions' economy of scale. In Latvia, units below faculty level may have legal personality. The reasoning was the facilitation of reintegration of teaching and research inside the university. It is now realised that reform of this kind actually hampers work at the lower levels of the organisation. In other countries, autonomy is also repeatedly claimed below university level.
Non-owned or private authorities can act in the general interest, but democratic governments and parliaments cannot be kept out. Legitimate State intervention (see 5.2) must be ensured. Final political responsibility cannot be borne by the managers of public or publicly funded universities. This is not always wholeheartedly accepted, and there are even reasons to be sceptical. The State has also difficulties in taking a position. In Albania for example, Article 7b of the 1994 Law provides that the elected rector automatically gets the status of delegated civil servant who represents the government as well. Such a classic 'double cap', leads to difficult legal disputes as to which interests the rector has to pursue first. The Albanian rector, having two bosses, his senate and his minister, easily finds himself paralysed by 'double binds' (contradictory impulses or orders for decisions). It makes rectors vulnerable to undue political pressure. The far reaching re-establishment of autonomy in central and eastern Europe necessitated a division of autonomy into responsibilities. Then it can be properly defined, becomes a workable juridical concept, and can be (re-)distributed. This is particularly important in countries which are redrafting their legislation, hut also useful under stable conditions.

The most important normative legal question is where responsibility starts (or ends, depending on the point of view). This is not just a matter of laws jurisdiction, or jurisprudence, but to a large extent a matter Of political taste and culture, maturity of the administration, and mutual faith. To avoid the balance of power becoming a matter of power or fortune determining the distribution of power, the constituent elements Of autonomy must be spelled out, but treated as interdependent. The stakeholders should basically agree with the distribution of responsibilities. Where the anatomy of autonomy is unclear, the drafting of regulations on governance is at its best a game.

For such an important right, there is remarkably little case-law on academic freedom. In literature and political discussions, academic freedom is always mentioned, but in fact entirely overshadowed by discussions en institutional autonomy. If steering should reach the work floor rather than being stuck at the level of administrators (McDaniel, 1997), decentralisation should be taken one step further: that is to faculty or department level. Decentralisation has consequences for the relationship between autonomy and academic freedom that should not be disregarded in new steering models.