Modes of Governance and the Limits of Policy

The preceding chapters describe the various efforts, initiatives and policy attempts to realise transformation in the South African higher education system from 1994 up to 2000. Chapter 12 (New South African Realities) is to some extent a summary assessment of the findings and discussions of chapters 4 to 11. As these chapters show, and as we will argue here, transformation in higher education was seen by politicians and laymen, policy specialists and ordinary people as an indissoluble part of moving away from apartheid as a state form to a more open, inclusive, equitable and democratic society. This book describes how the transformation project was launched, with great acclaim, consensus and fanfare – through a series of founding documents such as the National Education Policy Initiative (NEPI, 1993), the ANC Policy Document (Centre for Education Policy Development, 1994), the report of the National Commission on Higher Education (NCHE, 1996), and even the first Higher Education White Paper of 1997 – only to seem to veer off track. A loss of course was detected in at least three areas. First, it seemed as if, after 1997, the policy process gradually became less participatory and democratic, and it even seemed to some that we were returning to something of a top-down style of policy imposition reminiscent of an earlier era. Secondly, it seemed as if the state was forsaking the values of equity and social justice in favour of the values of efficiency, effectiveness and responsiveness. Thirdly, even where policy had been implemented as intended (in the National Student Financial Aid Scheme [NSFAS], for example), it was producing effects different from the ones expected and desired. What was producing these contrary effects? Why had the transformation project in higher education come to be so widely seen as an unpleasant surprise and distasteful disappointment?

political and politicised one -is rooted in the political reality of the form of transition from apartheid to democracy, one that places a premium on the efficacy of virtuous political will to effect virtuous transformation (Muller, 2000). The anti-apartheid struggle created the expectation of strong civil society participation in public affairs within a framework of government steering. To put it plainly, if perhaps crudely, the widely held popular assumption and firm expectation was that the achievement of transformation depended principally upon the exertion and application of concerted, collective political will in the form of participatory policy making. This form of the rational policy model is vulnerable to all the vagaries attending it elsewhere. But in this case a particular loading of political hope was added that made its alteration or non-attainment seem less bad luck than political betrayal.
The second class of explanation is a technical one which attributes policy surprise or change of course to one or other implementation or capacity deficit, an explanation resorted to at times by government itself (Department of Education, 2001). This class of explanation locates these deficits at either a national departmental level or at an institutional level. Examples attributed to the former include a lack of progress on finding a new funding formula and the lack of redress funding. An example attributed to the latter is the lack of ability of some institutions to raise external donor funding (Pilot Project Consortium, 2001).
These explanations most definitely explain a part of the reality, as we will go on to argue below. Both sets attempt to depict a greater or lesser derailing of the political project of transformation. On their own, however, they are not only partial but flawed. The principal reason for this is that they are both forms of explanation that rest on a deficit cause; the absence of virtuous political will in the case of the former, the absence of technical capacity in the case of the latter. What makes them misleading is that we are left with the expectation that, were the deficit to be remedied, the policy as expected would have produced the desired results. This is highly unlikely because all deficit explanations leave out of the reckoning at least three sets of social dynamics each of which will produce their own partly independent effects, each setting limits on what could have been achieved under even the most optimal of conditions. These are: • The effects of facing the necessity of trade-offs on both the process of policy formulation and the ranking of policy priorities. • The effects of the sociological characteristics of different institutional types on responsiveness. • The differential possibilities and limits of different governance regimes.
We will examine each of these in turn in the rest of this chapter. What we hope to show with this broader consideration of the policy environment than we have considered up to now in this book is that the vicissitudes of policy, the twists and turns of policies and their effects that this book describes, are to be understood not as some or other kind of policy or political failure, but indeed its opposite. By understanding better the broader context under the three rubrics of policy, governance and institutions, we hope to show that the main failure so far has been a disappointment of expectations.

TRADE-OFFS AND THE CONDITIONS FOR POLICY CHOICE
Traditionally, as we have said above, public policy as a driver of change has been viewed as a rational process consisting of causally linked phases: policy formation, policy implementation, policy evaluation, feedback, and policy adaptation. In this view, the effects of a policy are assumed to be 'measurable' in the sense that once a policy is implemented, one can assess how much of the policy intention was realised in practice. Since the seminal work by Pressman and Wildavsky (1973) on the implementation of policy, however, the conviction has grown that a policy process is far more complicated and irrational than is suggested by the linear view. As a consequence, in mature democracies, the focus of public policy analysis has moved from simple implementation analysis to studying the interactive dimension of the policy process as a whole (Gornitzka et al., 2002).
Apartheid was a comprehensive, state driven, 'top-down' system that consisted of policy formulated by government and implemented by the different government departments and state apparatuses. It was a fairly typical 'modern' central planning approach to the changing of society; an approach, if not an intent, shared with apartheid's great ideological foe, communist Eastern Europe. Despite their radically different goals and processes, what the two systems shared was a belief in the 'making' and 'remaking' of society, and that a key instrument in that 'making' would be policy formulated and implemented at the national level. Because both of them pursued an ideologically driven policy agenda, this similarity would have been invisible to both of them, as would a further similarity we have already touched upon -namely, that their best hopes and expectations were bound to be confuted as long as they viewed everything through an exclusively political lens.
Policy formulators preparing for a new democratic state started from the assumption that because apartheid had been such a pervasive, centrally driven state system, its undoing would likewise require a state driven, planned, policy process, with the key difference being that the goals and processes would be progressive and participatory. Thus, in addition to having different goals, the process would be democratised, but still with a strong, central planning (steering) component. There was widespread agreement that the market could not correct the injustices and imbalances caused by apartheid, and that individual institutional transformation, left to itself, would not result in a co-ordinated, equitable and efficient system. Along with Gornitzka et al. (2002) and Cerych and Sabatier (1992), we can divide a centralised, policy driven reform process into three distinctive phases:

Policy formulation
This involves identifying inadequacies in the existing system, defining or redefining the problem, followed by policy formulation that consists of setting new goals and objectives, targeting groups or institutions, examining one or more means of redressing the situation and suggesting instruments to achieve this. This process culminates in a formal (legal) decision by the cabinet or parliament to establish a new programme, institution or course of action. In South Africa the first (informal) phase started with the National Education Policy Investigation (NEPI, 1993), and the ANC 1994 policy statement (Centre for Education Policy Development, 1994). It begins formally with the appointment of the National Commission on Higher Education in 1995 (NCHE, 1996), the release of the government Green Paper and White Paper in 1997 and the publication of the Higher Education Act No. 101 of 1997. This process did not only involve problem identification and the prescription of certain remedies, but the development of a new framework of values and objectives that tried to embody the key principles of the new government, namely democracy, equity, responsiveness and development.

Policy implementation
In higher education, implementation is traditionally regarded as a matter between a national department of education and higher education institutions. This phase usually includes the elaboration of regulations, the creation of new structures necessary to translate government decisions into practice and day-to-day applications, with adjustments of initial decisions and regulations. In the South African context this started with the establishment of the Division of Higher Education within the Ministry of Education, followed by the abolition of the Advisory Committee for Universities and Technikons (AUT), the subsequent formation of the Council on Higher Education (CHE, 1998) and the associated Higher Education Quality Committee, as well as a plethora of implementation activities such as the passing of the National Students Financial Aid Scheme Act of 1999, the establishment of frameworks for accreditation of academic programmes, and the setting of conditions and criteria for the registration of private higher education institutions and programmes.

Policy evaluation
This phase is often, but not always, initiated by a new government or a new minister and involves a reformulation stage that revises programme goals, sets new objectives and may change or elaborate implementation instruments. In South Africa the appointment of a new minister after the second democratic election in 1999 led to what could be regarded, on the one hand, as a continuation of the implementation started by the previous minister, but on the other also contained certain elements of a more typical reformulation phase. Continuity of implementation activities consisted of the publication of a National Plan for Higher Education (Department of Education, 2001) that outlined priorities and set certain targets. Two further activities could be regarded as tending towards reformulation. The first of these was the establishment of a National Working Group (Department of Education, 2002) to make recommendations about restructuring the institutional landscape and the subsequent gazetting of a landscape that would reduce the number of institutions from 36 to 23, while simultaneously introducing a new type of institution (the 'comprehensive'). The second consists in a move from a consensus seeking orientation to a more top-down approach, an approach that can be seen in amendments to the Higher Education Act of 1997 that give the minister more powers, and in the manner in which institutional restructuring and the new funding system will be implemented. The first reform (reformulation) dealt with restructuring higher education, and the second with changing government-institutional relations, while both contribute to the ongoing reorganisation of governance relationships. It is this emerging re-alignment that requires explanation.

Symbolic policy?
The first higher education policy announcement made in South Africa after the transition to democracy in 1994 declared with fanfare that it was necessary for the new government to show that there would be a clean break with the past, a demonstrable break with the higher education system created by and inherited from the apartheid regime. However, this did not imply that such a break would be made through a fundamental, revolutionary process. This is clearly reflected in the following comment made by the first Minister of Education in the new democratic government: '… owing to the sophistication and fragile nature of higher education a radical approach to transformation was not adopted' (Bengu, 2002). What then would constitute such a clean break? Jansen (2001) has argued that the new policy had to be pre-eminently symbolic. It was more important for the government, in the first instance, to declare its intention of breaking with the past, than it was to develop policies that might have an immediate impact, for example, from a fiscal point of view. However, the fact that a policy is largely symbolic in nature does not mean that it will lack impact. On the one hand, as Scott states, 'symbolism, the mechanism by which meanings are shaped, exerts great social power ' (1995:129). The chapters in Section 2 (4 to 11) show, for example, that without the new South African government putting any explicit redress instruments in place with regard to student enrolment, institutions responded to the policy and acted in a variety of ways to change the composition of the student body. Jansen (2001), on the other hand, questions whether symbolic policy can be positive, arguing that in South Africa the intention was never to implement, but only to signal intent. While this may be true in certain instances, we do not think this was the case in 1994. Rather, the prime intention was to declare a break with the past and to signal a new direction. The need to declare a break with the past implied that the main items on the policy agenda had to reflect political priorities. This implied that the new policy issues with respect to higher education in 1994 were mainly concerned with the need to create more equity and democracy in the sector. But at this time the political imperative towards transformation acted to obscure both the nature of the necessary trade-offs that might have to be made to realise policy intentions, as well as their possible divergent effects. Indeed, we may say that a principal effect of symbolic policy, taken as 'actual' policy, is precisely to disguise this point.
The developments in Central and Eastern Europe after the changes of the late 1980s/ early 1990s show a similar pattern (Vlasceanu & Sadlak, 2001). First the new higher education policy signalled a break with the past. This implied an initial emphasis on the 'de-ideologising' of some of the curricula in higher education, as well as an attempt to strengthen institutional autonomy. Because of the difficult fiscal situation faced by the governments in those countries, neither of these two original policy aims improved the position and functioning of the battered public universities. What followed was a succession of symbolic policy attempts, none of which were co-ordinated with overall state policy or with efficient allocation instruments. The result was a constantly changing policy focus described as 'changing the changes' which led to policy fatigue and scepticism, especially among the academic staff.
In this case, as in South Africa, what was missing was an understanding of policy as requiring trade-offs, hence as requiring a supercession of the explicitly political phase of building unity through signalling political consensus. This requires further examination.

Trade-offs and weak infrastructural power
In order to proceed from policy proposal to policy realisation, priorities have to be decided upon and strategic trade-offs made. In the South African case, GEAR (1996) is an attempt to develop and implement such a strategic framework with respect to macroeconomic policy. It is worth noting that when GEAR was announced, the Minister of Finance, to the great surprise of many ANC supporters, declared that it was 'not open to consultation or negotiation'. We will argue below that this move, ostensibly from democratic consensus to state-centred decision making, was determined by a particular feature of the state at the time.
After the divisions of apartheid, unity, even if illusory, was a discursive political priority. After all, the first post-1994 government was called the Government of National Unity (GNU). However, we will argue that, in Gelb's (2001) terms, the state was operating within a context of weak infrastructural power, meaning that, politically, it was not able to construct a framework within which explicit trade-offs could be negotiated and agreed upon across divergent interest communities without alienating potentially crucial allies. This problem did not only occur in higher education; it was observable in many other sectors and could be attributed to a variety of factors. The inestimable prize of the negotiated transition was the avoidance of another bloody civil war in Africa. The price, however, was the continuation of antagonistic interests that symbolic policy could unify only by simultaneously deferring making trade-offs between competing interests.
While the intentions outlined in the White Paper framework had broad support within the higher education community, it soon became clear that within and outside higher education there was no political consensus on the trade-offs that would be necessary to give effect to the policy. It can be argued that there were three types of disagreement.
The first type of disagreement can be described as differences within the higher education community. The Pilot Project Consortium Report claims that: 'some of the historically advantaged institutions were strongly opposed to redress funding. Many argued that it was wasteful to support apartheid institutions, that they should be closed. Others were concerned that this effort would cut into their own funding. There were also differences of opinion among the historically disadvantaged institutions about how redress funding should be allocated. Some wanted it allocated equally, rather than based on the quality of proposals for funding, demonstrated need, or ability to carry out proposed projects' (Pilot Project Consortium Report, 2001). This report asserts that a combination of opposition to redress and differences about goals and methods undermined the implementation of the policy on institutional redress.
The second type of disagreement was between powerful political actors within the new government. One group had studied at the historically disadvantaged universities and through this experience had been exposed to some of the worst faces of apartheid. As students they had opposed the management of these institutions as well as the apartheid government, but they had also developed strong emotional attachments to these institutions and the educational role they could play in socio-political development. After all, an institution such as the University of Fort Hare pre-dated apartheid and had in the past produced some of Africa's most eminent leaders.
In the new government, however, there was another group of powerful actors who had opposed the establishment of the historically disadvantaged universities, and who still regarded them as creatures of apartheid and 'as a product of the geo-political imagination of apartheid planners' (National Plan on Higher Education, 2001:Foreword). Many of these political actors were profoundly ambivalent about building up the historically black universities, particularly if that implied propping up all of them through state support.
The third, and arguably the most telling factor, was the government's macro-economic policy. When government formulated GEAR (1996) as an action plan intended to give effect to the Reconstruction and Development Programme of 1994 which was an inclusive all-things-to-all-people policy list, the trade-offs necessary for growth were rendered visible. GEAR is a package of mainly macro-economic measures that include faster fiscal deficit reduction, budget reform, consistent monetary policy, stable and coordinated policies, and a strong emphasis on efficiency and restraint on government spending (GEAR, 1996). The main aims were to stimulate growth through foreign investment and improved competitiveness. With GEAR, efficiency, effectiveness and responsiveness took precedence over redress.
We hope that the above makes clear that it is only when a strategic decision, or trade-off, is made, that the bouquet of policy values agreed upon in the consensus phase must be ranked and prioritised. Only at this point does it becomes clear what is chosen and what is not. It is hard to over-estimate the disconcerting effect this must have had on the politicians and civil servants involved who had not anticipated this dynamic, as can be seen in the words of two erstwhile officials and the former Minister of Education in the endnotes.

Towards differentiated policy
As is evident in many of the chapters in this book, comprehensive governmental policy intentions were articulated after 1994 to change the higher education system as a whole, but without the necessary trade-offs having been made to realise these policy intentions. 'Comprehensive policy' refers to a set of broad general principles and benchmarks for a whole sector (such as those in the 1997 White Paper). Both before and after 1994, South Africa traditionally concentrated on 'comprehensive' policy, thereby neglecting the difficult priority decisions and differential levers that have to be designed to implement it. Indeed, we may say that 'comprehensive' policy was the form that 'symbolic' policy took in the immediate post-1994 period. By the same logic, relinquishing comprehensive policy means relinquishing symbolic policy and the discursive facade of unity in the sector. 'Comprehensive' policy-making has to be distinguished from differentiated policymaking, which means identifying and agreeing upon particular institutional targets that prescribe the route each university and technikon is supposed to follow against broad systemic benchmarks, as well as the creation of an environment of pressure and support necessary to facilitate progress along the route.
Policy differentiation is not necessarily aimed at creating institutional differentiation. It can have quite the opposite intention, namely to reduce differentiation. In contrast, a comprehensive policy that is 'the same for all' often has highly differentiating effects, as the application of the funding formula based on the South African Post Secondary Education (SAPSE) information system has demonstrated in the South African case. One of the most important factors hindering the design of specific policy levers since 1994 has been the absence of an up-to-date information system and a common set of informational formats so that benchmarks can be constructed and each institution's performance can be compared -both with the performance of other institutions in the system and with their own performance over time. The National Working Group's greatest contribution to systemic governance may well turn out to be their commissioning of a benchmarking system which is robust and sophisticated enough to serve as the beginning for future systemic development.
A key aspect of differentiating policy is experimentation. Policy is a form of explicit and deliberate governmental intervention. It is very important that policy-makers design a policy in such a way (through experiments) that the effects of the intervention can be assessed; in other words, that knowledge about which measures and instruments work and which do not can be increased. It is risky to make any statements or come to any conclusions concerning interventions and their effects without using experimental or quasi-experimental research methods. Too much is assumed concerning the effects of policies and policy instruments and very little is actually known about these effects. No national system can prosper without continual monitoring and research that creates the kind of information and analysis around which collaboration most usefully occurs. This circulates information and allows all the actors -government, society/market and institutions -to be responsive.
The above discussion points to the fact that unidirectional comprehensive policy has not worked in South Africa in the post-1994 period. Instead, a different notion of higher education transformation, based on a more targeted, differentiated, information-rich policy interaction between government, institutions and society has to be developed. This is in many respects in line with Olsen's corporate-pluralist state model of state governance (Olsen, 1988) discussed in the first chapter of this book. This network, or corporate-pluralist state is entirely dependent on knowledge -the production of knowledge for purposes of competitiveness as well as the use of knowledge to better fulfil its steering and policy-making roles. It requires, in other words, greater infrastructural power than South Africa possessed at the time of transition to democracy in 1994. The shift from a comprehensive to a differentiated policy model is going to demand a more efficient government, together with a new approach to consultation. It will require that government is more sensitive to the self-regulating capacity of the higher education institutions and the consequences of the complex relations between higher education and society. It is to this area that we now turn.

INSTITUTIONAL DYNAMICS
In a wide ranging overview of institutional adaptation to demands for reform, Gumport and Sporn (1999) detect a global increase in the salience of management within higher education institutions. They attribute the expanding role of administration to three interdependent dynamics: resource dependency that is primarily motivated by organisational survival, institutional isomorphism that is motivated by legitimacy concerns, and professional authority that is motivated by a struggle for professional identity.
Primarily, management is responsible for maintaining the organisation's exchange relationships. Whether it be as a bridge or a buffer, it must transact in an increasingly complex environment that contains not only government and business agents, but also an escalating range of 'stakeholders' (Maassen, 2000;Neave, 2002). Management becomes more important because it is increasingly responsible for the development of strategies that increase existing sources of income, tap into new income sources and help to reduce existing dependency relationships.
A key resource is legitimacy, and it is increasingly important that institutions be seen to be responding to demands for reform in a business-like manner. By presenting a more unified and business-like front an expanded management core is much better positioned to represent the institution than the often fractious collegium singly and severally. Gumport and Sporn (1999) describe the effects of this as a shift in the authority structure within higher education organisations that entails an expanding domain for the administrators and the leadership, and a narrowing authority domain for the academic faculty.
The above mentioned factors come into play very forcibly, as has been the case in South Africa in the post-1994 period, when institutions are faced with a sudden increase in demands for reform from the government and the society, and an unleashing of market forces, by both the government and a burgeoning local and global private higher education sector. Secondly, through a process of mimetic and normative isomorphism, institutional management's notions of 'successful' institutions have been exchanged, mimicked and adopted through professional networks that are responding, on the one hand to resource and legitimacy demands and on the other to promoting their position (Cloete & Kulati, 2003). Under these circumstances, capacity to evolve managerially becomes a crucial resource variable.
Given the rudimentary stage of policy evolution discussed above, it was, in practice, up to the higher education institutions themselves to interpret the policy innovations announced in and after 1994. Institutions were, however, not only attending to national government policy; they were also casting an eye on global developments. This is well illustrated by the former vice-chancellor of the University of Potchefstroom: Simultaneous with the change in the new world order, the fall of the Berlin Wall and the totally new international dispensation, occurred also, of course, the change in South Africa itself…So, the university, belonging to the knowledge production part of the world as a university, had to take notice of this totally new dispensation ... (Reinecke, 2001) A number of institutions transformed rapidly in the direction of a predominantly outward orientation. Others decided to stick to an inward orientation of 'academic business as usual', opting to rely on their traditional academic strengths. A third group was left behind trying desperately to survive without being able to anticipate and respond to the intentions of the announced policy innovations, nor to survive on dwindling government subsidies only. From the viewpoint of resource dependency theory, it could be argued that the historically black institutions simply could not adapt if a change in the environment threatened critical resource relationships. Symbolic government policy generated unrealistic expectations about redress and at the same time, unexpectedly, these institutions faced intensified market competition for students. Most of them were located in impoverished rural areas without strong academic and management capacity, and as a consequence had virtually no resources on which to fall back in order to avoid a crisis.
In addition to management capacity, an especially important key to understanding the different higher educational institutional responses is the factor of academic capacity, which determines any institution's particular academic identity or niche. An institution's academic capacity resides not only in academics with reputable qualifications, but also in their ability to restructure programmes, to attract good undergraduate and postgraduate students, to engage with business, local communities and government in research and contract work, to be part of international academic networks, and to have effective relationships with funding agencies (Chapters 7, 8 and 9). Academic capacity thus enables an institution to establish extensive links with the larger society and with the possibility of increasing and diversifying its financial resources whenever necessary. In higher education the basic resources are not only financial and human, but also reputational, with finance following reputation, not necessarily the other way around. Chapter 4 on funding shows that most of the institutions with strong academic capacity and a diverse funding basis did well financially in the post-1994 period. In a much more precarious situation were those institutions with weak academic capacity that had to rely on a single source of income, in most cases government funding.
But strong academic capacity does not necessarily equal institutional homogeneity. Nor does strong leadership and managerial capacity mean uniformity of strategy and style.
Management may develop a strong, unitary sense of purpose and use academic capacity to drive an enterprising agenda. More often though high academic capacity fosters diversity. In such a diverse institution the established culture, or the power of the academics with their extensive societal links, may tend to resist the imposition of a unified purpose, making the institution much less centrally governable, and setting up a fluctuating tension between the central leadership and management and the faculties.
Every institution embodies a determinate mix of academic and managerial capacity. Below are three counter-intuitive conclusions about how this mix of capacity may affect institutional orientations with respect to reform:

Institutions may be unresponsive to reform or the market from positions of weakness or strength
• An institution may be unresponsive from a position of strength where domain consolidation can take precedence over active responsiveness to policy or the market. • An institution may be unresponsive as a consequence of managerial weakness.
• An institution may be unresponsive from both managerial and academic weakness. This is the most vulnerable place to be.

Institutions may be responsive to reform from positions of weakness or strength
• An institution may be responsive in order to consolidate, to be enterprising or domainseeking from a position of relative strength. In this case strength can be compounded. • An institution may be responsive from a position of weakness, using either the market or higher education policy to act as a gyroscope for seeking a domain. This is a precarious strategy because here applied research or relevant curricula are sought in advance of consolidated disciplinary capacity. It is a strategy that is not sustainable in the long term.

Institutions may be responsive to the state, but unresponsive to the market and vice versa, from positions of weakness or strength
• For institutions with stronger academic and managerial capacity, the state and the market are both exogenous forces with variable possibilities, and both need to be treated with caution, but both also provide opportunities (Muller, 2001:23-24). • Among institutions that are weak in academic and managerial capacity, state support in the form of input-based subsidies is a lifeline, whereas the demands of the market are most likely to generate a crisis. When the state moves to output-based subsidies, however, the difference dissolves.
Institutional theory suggests that strong organisations are easier to influence from without when the external signals correspond to their internal criteria of, and learnt capacities for, relevance. When the external signals go against these, the institutions become highly resistant: they are able to 'ignore control signals, to forego incentives, and to absorb sanctions, without changing their ways in the direction desired by government policy makers' (Scharpf, 1987:105). Higher education institutions are an especially good example of this.
This section makes it clear that a realistic assessment of what institutions may do in the face of changing societal conditions or changing policy, or both, requires an understanding of their constituent sets of capacities, simplified above as being of two key kinds, managerial and academic. It also tries to make clear that a determinate mix of capacities might translate into stable dispositional strategies which might be highly labile and adaptable, or which can harden into entrenched habits that are very difficult to change. Institutional case studies, of which there are so far all too few, will show, we predict, that both these dispositional tracks have their advantages and disadvantages.

MODES OF GOVERNANCE
One of the central tasks for the post-1994 period was to change the relationship between government and higher education. Democratising the mode of governance was seen as the key process in developing a new value framework for the whole society and for higher education in particular. New policy, and its implementation, was to occur through a process of democratic participation at national and institutional levels. In Chapter 12 it is argued that: 'It seems that the consultative and participatory process of policy formulation and the first cycles of the three-year planning dialogue gave way to a much stronger state-steering approach, driven by government's frustration at the lack of progress made towards achieving transformation goals'. We will also show, however, that, government rhetoric to the contrary, this seeming return to state steering was overlaid by a number of market driven features that yielded in effect a hybrid market/state-steering mode of governance. To many observers, this hybrid seemed both contrary to accepted policy as well as incoherent. The rest of this section will argue that this governance hybrid was not only possible, but probably inevitable, given a series of key structural features of the state form at the time. In what follows, we will briefly review the contemporary literature on governance, developing a conceptual language for discussing the apparent U-turn.

First reconceptualisations
The first point to note about the concept of 'governance' is that currently renewed academic discussion of governance has to do with the development of alternatives to hierarchical government control, i.e. to the traditional mode of state-dominated coordination (Mayntz, 1998;Peters, 2001). Despite variations between countries it can be argued in line with Peters (2001:4-13) that the traditional governance model was based on the following common principles: • The civil service has to be apolitical, in other words 'neutrally competent' (Kaufman, 1956). In addition, politics and administration have to be seen as separate elements of governance. • Public management has to be based on hierarchical principles and ruleboundedness. • The governmental organisations have to be permanent and stable.
• Civil service should be institutionalised and governed as a corporate body.
• The civil service should be strictly controlled and regulated in detail.
• Finally, equality should be an important principle in governance, with respect to outcomes as well as organisation.
The importance and appropriateness of these principles for modern day governance has been questioned at the very least, if not wholly rejected (Kersbergen & Waarden, 2001;Peters, 2001). Many authors have identified possible causes for a decrease in the effectiveness of traditional governance arrangements and hence the rationale for introducing new governance modes. A cause referred to by many authors is the economisation of societies. It is argued that the economic crises of the 1980s and early 1990s have forced governments to adapt their governance arrangements and to put economic considerations at the forefront of their governance approach (Savoie, 1995). However, other authors have suggested that this explanation alone cannot account for the fundamental and far-reaching nature of the shifts in governance. Peters (2001:14-15), for example, points to demographic factors, including the ageing of Western societies and the decreasing social and political homogeneity among individuals and groups in society.
There is a growing 'lack of common ground' with respect to many issues. As a consequence, the traditional pattern of government-led negotiations between various interest groups has become problematic, and arriving at social and political compromises has become more difficult. In addition, traditionally stable governance arrangements and organisations have become destabilised, making it more complicated for government to intervene in society (Cohen & Rogers, 1994). These general developments are observable worldwide, even though many variations can be found at the national level.
With respect to governmental strategies for dealing with the 'governance crisis', higher education offers interesting examples. Amongst other things, the high level of public expenditure on higher education and the growing acceptance of an economically instrumental interpretation of the role of higher education in society (Gumport, 2000) makes the sector an obvious target for governance reform.
Since the late-1980s, a number of higher education scholars have used state or steering models developed by other social scientists to analyse changes in the relationship between the state and higher education. Van Vught (1989), for example, introduced a central planning and a self-regulation model of government steering, later elaborated into state control and state supervision models (Neave & Van Vught, 1991;Maassen & Van Vught, 1994;Maassen, 1996). These models were based on the classic work of social science authors such as Meyerson and Banfield (1955), Ashby (1956), Lindblom (1959Lindblom ( , 1965, Steinbrunner (1974), and Beer (1975). The implicit assumption in the state control and state supervision models was that a development from state control to state supervision should be promoted because if the state had a supervisory role it would lead to the better performance of higher education than if it had a controlling role. From this perspective, state steering in the form of state supervision was the preferred alternative to the by now widely discredited, traditional, 'top-down' form of co-ordination. In this period, however, these were not the only models to emerge that proved to be useful for analysing and understanding the relationship between the state and higher education.
During the 1990s a number of European higher education researchers (see, for example, Heffen et al., 1999;Gornitzka & Maassen, 2000) began to work with the four so-called state models introduced by Olsen (1988): the sovereign (or unicentric) state, the institutional state, the corporatist (or segmented) state, and the market state. Of these four models the first two, the sovereign state model and the institutional state model, can be regarded as variations of the traditional governance model discussed above. The other two models are alternatives to the traditional governance approach. These four models are not necessarily mutually exclusive, nor are they normative. They represent different ways of organising the relationship between the state and society that correspond more or less to state dominance and control (the unicentric state), state protection of specific social values and norms (the institutional state), the state as one of the involved interest groups (the corporatist state), and the minimal state (the market state, that intervenes minimally and allows the dominance of market forces).
The steering models introduced by Van Vught (1989) and the state models developed by Olsen (1988) reflected the real governance shifts of the 1980s. This was a transition period in which states were experimenting with new governance approaches without the old ones having been rejected completely. Researchers studying the shifts during this period, such as Van Vught (1988, 1989), talk about the Janus-headed character of state governance with respect to higher education. (See also Amaral & Magelhães, 2001).
Fifteen years down the line, it is clear that the transition period is over and that the traditional governance model in its basic form has been 'left behind' in practice and as a model advanced by governance theory. As a consequence, instead of comparing 'old' and 'new' models it is now assumed that the traditional model is no longer acceptable and that various alternatives have been developed to replace the traditional model in practice. Peters (2001) makes a distinction between two waves of reform to traditional approaches to governance. The first wave consisted of ideologically driven reforms that were enacted in the 1980s and early 1990s, and the second was a more recent, pragmatic set of reforms that combines further 'repair work' to the traditional model with attempts to deal with some of the flaws of the earlier ideological reforms. In the first wave, four alternative approaches to governance emerged, namely, governance through applying market mechanisms, governance through increased participation, governance through greater flexibility, or governance through deregulation. The ideological nature of the reforms was especially clear in the case of the market approach that was introduced in many countries as an unquestioned improvement to, and advance on, the traditional governance approach.
Even though there is some overlap between various aspects of these models, the four approaches can be distinguished on the basis of their different problem diagnosis with respect to the functioning of the traditional governance model and their ideas about the nature of the reforms necessary to address these problems. In addition, even though in governance reforms one can observe various elements of different approaches implemented at the same time, these combinations are not always compatible in practice. In general one can argue that the market approach appears to be most compatible with the flexible governance approach, just as the participation approach is most compatible with the deregulation approach (Peters, 2001:95).
In the second, current governance reform wave, two basic types of change can be observed. First there is change that is a continuation of the reforms introduced in the first wave, although now it is more pragmatic, less ideological and employs different instruments. The use of the market mechanism in public governance, notably extolled by Margaret Thatcher in the UK, for example, was ideological and fairly controversial in most countries in the 1980s and 1990s. In the meantime it has become more generally accepted and embraced by governments of all political shades, including traditional anti-market parties such as the Labour parties in Germany and the United Kingdom. Secondly, there are changes that are responses to problems created by the first wave of reforms. This concerns, for example, the perception by governments that the first reform wave has led to excessive autonomy of public sector institutions, leading to a loss of governmental control over their functioning. Related to this is the feeling that the reforms resulted in a decrease in public accountability of public sector institutions, leading to the need to set up formal performance evaluating bodies (Peters, 2001:119-121).
For a more elaborate discussion about hybrid modes of governance, see Gornitzka and Maassen (2000), while Maassen (2003) has discussed the relevance of Peters' governance models for higher education. For a more detailed application of Peters' models to the South African governance shift in higher education, see Cloete, Maassen and Muller (2004). For the purposes of this chapter, we can summarise the discussions and conceptualisations included in these publications by saying that, if one reads solely through the policy documents and avowals of educational politicians, it seems that the Department of Education started off with a participation approach, only to change course half way to a market approach overlaid with deregulation and flexible mode features. That these approaches are indeed radically different can be discerned from examining their five constituent features (table 1, after Peters (2001). Applying insights derived from these to the South African situation we can say that, if one reads solely through the policy documents and avowals of educational politicians, it seems that the Department of Education started off with a participation approach, only to change course half way to a market approach overlaid with deregulation and flexible mode features. That these approaches are indeed radically different can be discerned from examining their five constituent features (Table 1, after Peters, 2001).
With these conceptualisations of the governance of higher education, we have a clearer grip on the anatomy of the governance reform process. What remains to be examined is why it took the form it did in South Africa.

Determinate features of the global and local context
Governance theory tells us that governance-mode change in stable, mature democracies happens incrementally and on the basis of the prior mode. Where a clean break is targeted, albeit only symbolically as we argued above, in the desire to move decisively away from a dual system with two, parallel, racially-crafted versions of the traditional approach, it is never as clean as the aspiration inscribed in the rhetoric would have it. In fact, as this case shows, there was a symbolic shift in much of the comprehensive policy rhetoric, but at the level of policy instruments (for example, the retention of the SAPSE-based funding system) it remained a largely traditional approach, albeit now consolidated from two versions into one. This traditional governance mode, however, proved to be highly unstable as a framework for the statehigher education relationship. This was partly because it not only maintained, but actually amplified the disadvantage of the poorer institutions, a situation visibly at odds with the symbolic rhetoric. But a confluence of other circumstances also contributed to its instability. Some of these, identified by higher education analysts, are reported above. Others, specific to South Africa, will be discussed below. As the decade passed the halfway mark, the governance mode had shifted into a stalled, unstable and ultimately contradictory position: it had created a stalemate between a participation approach advanced in rhetoric and an unstable, comprehensive, traditional approach exercised in practice. The situation was ripe for a stalemate breaker.
The following key pressures are among many contributing to the instability of the comprehensive traditional mode: • Rapid globalisation of higher education, i.e. global diversification of finance, global competition for students further stimulating international student mobility, research collaborations, policy borrowing and the escalating costs of higher education put a premium on managerial adaptability ('responsiveness'). As we showed earlier in the chapter, this increases pressure on managerialism (see Savoie, 1995;Peters, 2001;Gumport, 2000), especially at the institutional level. However, this is a specific kind of managerialism, in the sense that it pushes towards the flexible governance mode, towards multiple clients, including the market, and away from the traditional mode. As we saw above, however, successfully deploying such a mode at the institutional level depends upon a certain level of academic and managerial capital in the institution, which the disadvantaged institutions, by definition, did not have, and it therefore increased institutional heterogeneity still further.
• Macro-state policy, in the form of GEAR, set the policy template for all other state portfolios. As we stated earlier, this drives a higher education system to both fiscal austerity and towards market types of governance.
• Finally, the state had weak infrastructural power (Gelb, 2001), which meant that policy trade-offs, which would advantage some stakeholders and not others, could not be negotiated but rather had to be imposed and consolidated through symbolic legitimisation. GEAR again is the paradigm case. The added difficulty in higher education was that there were real competing interests, and the ones with least cachet for the sector (i.e. the historically black universities) were also the ones with most political clout, at least as far as national politics was concerned.
These imperatives bearing on higher education were partly contradictory. But together, they all drove against the tenability of a participatory approach. That the stalemate described above retarded this mode from realistically getting going (moving from symbolic to real policy) leads to the conclusion that we never managed to get beyond the rhetorical promotion of a participatory governance mode. In this first phase of governance reform (Peters, 2001; see also Table 1), most South African commentators were led by the rhetoric, i.e. they imagined that consultative agreements would compel both policy and the mode of governance to move in this direction. In practice, structural features of the national and global environment simply superseded the rhetoric of participation. What we had, in effect, was a partial shift from an authoritarian, bureaucratic version of the traditional mode, to a racially reformed version. However, because of the increased reliance on the market mechanism, the intensifying effects of globalisation and the growing requirements for responsiveness, this implied in practice a rapidly expanding managerial power in those institutions that had the capacity to handle it, and continued steering together with some interventions to compensate for the worst depredations of the market for the weak institutions. That the strongest of these weak institutions decided to resist steering, and to fight the centre, only confirms the managerial crisis from which they suffered.
This policy shift, interpreted by some as an aberration, was really only an inevitable move: from a dual traditional governance model to a differentiated, market-driven model with dual versions -a flexible version for the strong institutions and a deregulated, bureaucratically steered version for the weak. This is partially disguised by the symbolic political necessity for the government and bureaucracy (a necessity for national politics, not for the sector) to be seen as treating all institutions in 'the same' way. This implies that some residual remnants of comprehensive rhetoric are likely to remain a part of policy in order to justify public intervention in not only the weak but also the stronger institutions, intervention aimed, for example, at the 'lack of transformation' of their staff racial profile. The decisive factor here though, is the continuation and amplification of an extremely heterogeneous sector. While this heterogeneity could be expected to force the government to develop a differentiated governance mode for the sector, the differentiated governance mode used in practice is unlikely to alter the stratification in the sector along the lines of colour and class. This is because it resembles an arrested market model, with some unconstrained features (e.g. student mobility via NSFAS), and some highly constrained features (e.g. private universities). And the implications of the new subsidy formula remain as yet unclear. Thus, South Africa appears to confirm rather than confute global trends.

CONCLUSION: OUR OWN HYBRID JANUS HEAD?
South Africa had the good fortune (in retrospect, a mixed blessing) to initiate and undergo a period of radical reform during what has come to be known globally as the 'Roaring Nineties', roughly bracketed by the fall of the Berlin Wall and 9/11. This was the decade, for nearly all of the developed countries and many of the developing countries, of unprecedented economic growth and new levels of prosperity. This rapid growth, requiring rapid adaptability, fuelled the gathering rejection of the 'iron cage' of the traditional governance approach in the private and public sectors alike. Leading the charge were the apostles of growth, dubbed the 'deficit-reducers' by Joseph Stiglitz, leading to what he has called the 'misguided "ascendancy of finance"' (quoted in the New York Review of Books, 15/1/04, 28). What this meant in general terms was that criteria of efficiency and productivity came to trump values of participation, equity and justice, at least in the short term. What this meant in the South African case was that once GEAR, a deficit-reducer par excellence, had been installed as the premier instrument of finance policy in South Africa, it was only a matter of time before its hegemonic effects would be felt in all other domains of policy and governance. Under these circumstances, the Department of Education's participation approach to governance and policy in higher education, the antithesis of GEAR's market-led approach, stood no chance of being implemented. Its sole function, during the middle years of the decade, was to buy symbolic legitimisation and consensus, whatever had been the undoubted good intentions of its proponents.
This chapter has also sought to show, however, that on its own, such a conclusion can all too easily regress to political pessimism, underlain by a naively rationalist conception of policy. The chapter consequently explored three key domains that shed light on, clarify and complicate the picture.
• The weak infrastructural power of the state in the early years of the decade, up to the decision on GEAR, saw to it that the participation approach was frozen in a symbolic phase and never proceeded beyond tokens and good intentions. This was because the trade-offs necessary to implement it could not be successfully negotiated in a sector that was driven by powerful interests and different kinds of political backing. This log jam was only broken when global currents tipped the balance towards the 'ascendancy of finance' and the decision on GEAR, significantly not through any participatory process. Following that super-ordinate decision, taken at the highest level of the state, came the slow process of conforming sectoral policies and approaches, to a greater or lesser extent, to the hegemony of finance. Starting as it did at the opposite end of the ideological spectrum, there should be no surprise that the resultant mode was a hybrid (Cloete et al., 2004).
• Having to conform to national macro-policy pushed higher education into the market, while it partly still clung to remnants of traditional governance steering and at the same time clung ideologically to the rhetoric of participation/equity/social justice. This curious composite mode, with a benign mythical superstructure and a more destructive, but invisible base, was understandably represented and responded to quite differently by different constituencies, expressing different interests: -The Minister of Education and his Department expressed themselves in philosophical opposition to the market, while technically implementing market features, or at least features that were to have market effects, like the NSFAS funding scheme, the SAPSE-based formula, etc. In addition they were intervening to control the worst depredations of the market, such as the collapse of some historically disadvantaged universities, or to dampen the damaging effects, for example, of private higher education.
-The policy analysts allied to the ANC, by the same logic but drawing diametrically opposite conclusions, lamented the betrayal of participation, thereby curiously fulfilling the party's necessary ideological role of unity-construction in a time of market-fuelled dispersal of interests.
-The weak institutions complained that it left them dangling in the wind of the market (true), and about the lack of state intervention (not entirely true).
-The strong institutions invoked university autonomy (a questionable proposition under conditions of such exaggerated institutional heterogeneity), and complained of too much state intervention (also not entirely true).
Little wonder then that there was, and remains, such confusion about the real direction of the modes of governance with respect to higher education. In this chapter we have characterised this direction, following Peters (2001), as a market-led governance approach with two distinct subtypes: a flexible/managerial sub-type with minor interventionist features, and a deregulatory/interventionist sub-type with substantial market features that may continue to pose threats to the weaker institutions. It would not be adequate to describe this approach as a political U-turn, nor would it be adequate to attribute it solely or principally to a technical deficit. It comes about as a result of a broader structural set of forces shaping the field.
We can conclude, with some confidence, that South Africa has reached the end of the traditional governance era, and in this respect conforms more nearly to the global trend than is usually conceded. We may also conclude, with reasonable confidence, that Peters' (2001) first phase of ideological tinkering is by and large over, and that, in higher education at least, hard-nosed efforts to stabilise the new governance mode will become increasingly sophisticated and differentiated. A cardinal feature which will continue to have determinate effects on the future direction of policy is the extreme heterogeneity in the system, which has increased, rather than decreased since 1994. The symbolic rhetoric of the now ending first phase may yet linger awhile since it may have some use, but it cannot last long. It is probably fair to say that South African higher education policy is in that Janus-headed phase described by Maassen and Van Vught (1988;1989) for the Netherlands and Amaral and Magelhães (2001) for Portugal. In all of this, South Africa follows in well-trodden reform footsteps -albeit not with the effects expected from the 1994 'revolution'. ****************************************** NOTES 1 This identification of two types of explanation can also be found in Bovens et al. (2002) who have discussed success and failure of public policies by making a distinction between the programmatic and political mode of assessment of policy success or failure. The former is based on a view of policy making as social problem solving, related to questions such as: do governmental policies contribute to the solving of social problems, and do they work in a sensible way? The latter refers to the way in which policies and policy makers 'become represented in the political arena. It is not the social consequences of policies that count, but the political construction of these consequences, which might be driven by considerations of wholly different kinds.' (Bovens et al., 2002:20) 2 Trevor Coombe, Deputy-Director General in the Department of Education in charge of systems and finance between 1994 and 1999, commented on the tension between the two approaches (Coombe, 2001): 'The Department of Finance had to be convinced that a redress fund or allocation made sense. And there was a counter-argument, namely that the whole country has been screwed up by apartheid. If a redress principle had to be applied consistently -and a budget policy looks for consistency -it would have to be applied across the board. Now what would that actually mean? Would it mean that historic deprivation would have to be compensated for through the budget by special grants to all institutions that have been deprived? Consider the implications at school level, for instance, or in the hospital set-up? It is an intolerable proposition, especially if you are running a tight fiscal ship, if you are attempting to bring down a budget deficit and there are certain limits on your expenditure ceiling. And, of course, that was the prevailing policy. So there was really no encouragement on the part of the Department of Finance for a major fund or allocation under the name of redress and it was a very difficult matter to argue, because we were not in a position to argue a similar case with respect to the school system.' Chabani Manganyi, Director General of the Department of Education from 1994 to 1999 reflects: '… the most difficult thing was to persuade the new democratic government that you could put something called "redress" into your budget -as strange as it sounds.' (Manganyi, 2001).
The different positions of the actors involved in the institutional redress policy are clearly reflected in the following comments made about the outcomes of the policy process by the democratic government's first Minister of Education (Bengu, 2002): 'I honestly believed in institutional redress. I fought valiantly for the operationalisation of institutional redress and I was fully supported and partly pushed by the historically black institutions. When I left office I was, however, not a winner because my Department was not as committed to redress as I was. The question of institutional redress was also a bone of contention within the Universities' and Technikons' Principals Forums.'