1 Introduction

The explicit inclusion of technical and vocational education and training (TVET) in the United Nations Sustainable Development Goals was greeted with excitement by advocates for this sector. For a sector that is generally neglected and has low status (Fooks, 1994), ensuring that the SDGs included a focus on TVET was a hard-fought accomplishment. Vocational education was once again a visible part of the international discourse on the role of education in development and would be taken seriously.

We know from the Education for All (EFA) campaign and the Millennium Development Goals (MDGs) that goals and their associated targets have real effects on education systems, with attention and resources following commitments to targets. There can be little doubt that the international targets for achieving universal primary education affected education systems across the globe. Arguably, many more young people are at school because of the focussed attention that these goals and targets generated. However, in this chapter we argue that the incorporation of TVET in SDG 4 and the explicit quantifiable targets associated with it may be something of a pyrrhic victory for the sector. We consider why this education sector in particular is not easily amenable to international targets, and why target-setting for TVET may have perverse or unintended consequences. While our argument applies across all contexts, we are particularly interested in the implications for TVET systems in developing countries where decisions to allocate scarce resources often have greater consequences.

We first outline in broad terms the shifts in the discourses and focus on TVET in international and national policy processes and provide an overview of the key targets and indicators of the SDGs and other regional and national strategies. We then consider why TVET is inherently not amenable to international target-setting – partly because of how heterogenous TVET systems are, and partly because of the ways in which they are embedded in specific contexts and structured by them. In the third section, we draw on experiences as researchers and in policy processes, primarily in South Africa. We problematise the use of targets and indicators by looking at three issues: the nature of quantifiable targets; the tendency for target-setting that results in policy posturing; and the difficulty of applying simple indicators to complex systems.

2 TVET in Development Discourses and the Emergence of Targets and Indicators

The status of TVET in development strategies has had a chequered history (King, 2009). In the decades after World War II, TVET and skills development were part of the orthodoxy of human capital development approaches that were championed by the World Bank and other international agencies. By the early 1990s, based on rates-of-return analyses and various other critiques, donors prioritised basic education, and the emphasis on TVET was reduced (Ashton et al., 1999; Bennell, 1996a, 1996b; Bennell & Segerstrom, 1998; Gill, Fluitman, & Dar, 2000; Wolf, 2004).

In recent years, as participation rates in schooling have improved, the World Bank, a long-time critic, has started advocating for building TVET systems. Developing vocational skills has, again, become the focus of national policies and donor agency agendas (King, 2013b). This reemergence of a focus on TVET promotion is attributed to a range of factors. Zeelen (2015) argues that as countries reached near universal primary enrolment, so attention shifted to the very high levels of drop-out, opt-out, or push-out from the schooling system that left large numbers of young adults neither in education nor work. Vocational education is viewed as providing a solution to a perceived mismatch between the school curriculum and the workplace. A second line of argument looked to vocational education as a mechanism for better aligning local economic skills needs with supply. This is particularly pressing in the context of the changing nature of work, with technological and social changes requiring more complex and specific skills.

The phenomenon of setting goals and targets at a global level emerged with the rise, in the 1970s, of project management approaches linked to performativity, measurability, and accountability, which continue to dominate management discourses (Boltanski & Chiapello, 2006). In order to accurately compare, report, and evaluate, these approaches hold that it is necessary to first define goals, and then set measurable targets and indicators. This marked a shift in the way in which interventions and policies were conceptualised and articulated, with increasing emphasis on restricted lists and clearly developed targets rather than complex, open-ended, and multifaceted processes, which had shaped many of the earlier approaches to development.

The EFA initiative adopted in 1990 at the Jomtien Conference set the pattern within the education sector. While the EFA movement advocated for quality education and improved outcomes, much of the focus remained on access and enrolment, catalysing governments in developing countries, aid agencies, and nongovernmental organisations (NGOs) around a single target. Schooling systems were expanded, teachers trained and employed, and infrastructure built in order to achieve the goal. Countries were compared and ‘shamed’ if they were not progressing. Significant strides toward meeting the targets were made. The Dakar World Education Forum of 2000 renewed the commitment to achieving EFA, albeit acknowledging that a greater focus on quality was needed, and advocating for a broader conception of basic education. While there were a handful of references to vocational education in the Dakar Framework, they largely emphasised skills development, career guidance, and vocationalisation of aspects of the schooling system. More than 10 years after the Dakar Framework was adopted, the 2012 Education for All Global Monitoring Report focussed for the first time on youth and skills, although even then the stress was mainly on school-and-skill (UNESCO, 2012b). In its summary, it argues, ‘Most advocates now see skills training not as separate from, but as integral to, general education, offering foundation and transferable skills at the same time as job skills’ (UNESCO, 2012b, p. 27). TVET as a distinct focus remained absent.

The MDGs, adopted in 2000, ran in parallel to the EFA initiative but had a wider development remit. Goal 2, devoted to education, was essentially a narrower version of the EFA commitment, focussing again primarily on access to basic education. Adult, higher, and vocational education could at best be indirectly inferred in some of the five other goals, but were not targeted. Aid money and technical assistance committed to TVET from a variety of international organisations have subsequently increased – although at the same time, policy focus has shifted from traditional notions of building technical skills to entrepreneurship and transferable skills, and to competence-based training and national qualifications frameworks (Allais, 2014).

The debates about the SDGs marked a significant moment for advocates for TVET. They won agreement on the inclusion of a right to all forms and levels of education, explicitly including vocational education. In the Education 2030 Framework for Action, vocational education is seen as an important part of broader educational goals, as well as key to supporting equitable and sustainable economic and social development, contributing to the realisation of human rights, and developing the productive capacity of people, their societies, and their economies (WEF, 2015). These broad goals are translated into two specific targets: 4.3, which focusses on equal access to affordable quality technical, vocational, and tertiary education; and 4.4, which says that by 2030, we need to have substantially increased the number of youth and adults who have ‘relevant skills, including technical and vocational skills, for employment, decent jobs, and entrepreneurship’. However, the agreed indicators, which measure the achievement of the target, are very narrowly and inappropriately framed: Indicator 4.4.1 reads: ‘Proportion of youth and adults with information and communications technology (ICT) skills, by type of skill’ (see Appendix 2).¹

The approach of setting goals as a policy tool is not restricted to the United Nations and its agencies. For example, the African Union (AU) has followed up with its own commitments. On 31 January 2016, the AU adopted the Continental Education Strategy for Africa as the framework for a transformative education and training system in Africa (AU, 2016). The strategy aims to shape the policies of African countries along similar lines to the SDGs, including expanding TVET opportunities at both secondary and tertiary levels and strengthening linkages between the world of work and education and training systems. This follows an earlier 2014 AU resolution that adopted a continental TVET Strategy, calling on member states to (i) enhance support and investment for TVET as it is fundamental for skill development for the youth and to promote employability and entrepreneurship through innovation; and (ii) to align their national TVET strategies to the AU Continental TVET Strategy for effectiveness. Clearly at the level of the African continent, TVET is seen as being key to education and development (AU, 2018). The expansion of TVET is seen as solving apparent skills shortages, enhancing productivity, and absorbing a major youth unemployment problem, among many other goals.

At the national level, governments throughout the world have adopted similar approaches in national development plans and national education and skills strategies. Vocational and skills education is highlighted as an important strand of development agendas, and goals and targets are specified to concretise the commitments made under the plans.

This section has shown that there is a confluence of both a new commitment to TVET globally, regionally, and nationally, and that this commitment is expressed at various levels through the process of goal setting, usually coupled with targets and indicators that need to be measurable and achievable. In the next section, we explore three issues that separately and in combination convince us that, for TVET systems at least, there are real dangers in this new TVET gospel, as well as in the attempt to set goals and targets at such high levels. The issues are interlinked but are separated for the purpose of presenting the arguments. We start by considering the inherent heterogeneity of TVET internationally, which is problematic for target setting. We then reflect on our experiences working in the South African policy terrain, and the ways in which goals and targets took away policy attention from the important task of building institutional capacity. Finally, we argue that goals like SDG 4 can have the effect of focussing attention on reform only within education and training systems, while TVET is shaped by economic, political, and social factors to a greater degree than the rest of the education and training system.

3 Complicated and Complex Systems

TVET systems are not easily amenable to international targets for the same reason that there is no strong comparative education research tradition in this sector: national TVET systems are very heterogeneous (Cedefop, 2017) and idiosyncratic (Bosch, 2017). The heterogeneity arises from the fact that how TVET is defined and bounded within each country varies, and so there is little agreement on what is included or excluded under the term. Vocational education straddles formal, informal, and nonformal education as well as the public and private sectors. It can take place in school-like institutions such as colleges, in workplaces, or in hybrid spaces. It can be undertaken before entry into an occupation or afterward (or in a liminal time/space such as an internship or apprenticeship where one is neither fully student nor fully worker). It can occur at various levels of education from very basic or primary through to higher levels, and the points and manner in which it splits from ‘general’ education differ across countries. There are differences in terms of when specialisation starts, whether learning is primarily in workplaces or education institutions, and what the pathways are to other education and training programmes (Bosch, 2017). Different types of labour markets (internal labour markets, occupational labour markets, tournaments) all have very different implications for training and skills because education pays off very differently in all of them (Marsden, 2009; van de Werfhorst, 2011). All of these factors mean that the configuration of TVET varies across and within contexts and is idiosyncratic because the internal and external factors shaping the system (the nature of the economy; the regulatory context, including regulation of the labour market; the cultural and historical antecedents) vary greatly.

This alone would suggest that setting common goals and objectives for TVET should be approached with caution. How can there be targets and indicators that make sense in such heterogeneous systems?² Furthermore, TVET systems are also complicated (they have many and diverse component parts) and, more importantly, complex (the parts are interconnected in ways that feedback within the system unpredictably).³

This complexity is in no small part a consequence of the fact that vocational programmes emerged in different ways in different economic sectors within countries. Programmes are often related to the very specific needs of particular industrial sectors in smoothing the transition from school to work and attempting to improve the skills of job applicants, or ensuring that the particular needs of employers are met. All of this means that within countries, let alone between countries, a wide range of different types of programmes, providers, and award systems exist. This is in sharp contrast with schooling and university systems, which may differ from country to country, but which have had stronger traditions of nationally recognised certificates and award bodies, and share many similarities across national boundaries. In many countries today TVET systems are described as fragmented, and as having a proliferation of qualifications, which is one reason why qualifications frameworks have been such a focus of reform for TVET systems (Allais, 2017a). Countries that have adopted the British qualification model have often seen further fragmentation and proliferation of TVET qualifications through the very policies aimed at streamlining them (Allais, 2014).

The strongest comparative$research tradition focussed on TVET comes from politicql science, and what has been labelled as comparative capitalism, following Hall and Soskice’s Varieties of Capitalism (Hall & Soskice, 2001). There is much debate within and about comparative capitalism: Does it remain empirically useful? Is it merely additive (Ashman & Fine, 2013)? Is it of any use in thinking about developing countries and development and, in particular, in African countries (Breckenridge, 2018)? Researchers have drawn attention to its inadequate account of power relations (Streeck, 2012) and the inherent limits of methodological nationalism (Lauder, Brown, & Ashton, 2017). But it is clear from the body of literature that has developed in this tradition that TVET systems are shaped by their social, political, and economic context in ways that are different from school and university systems. Institutional political economy demonstrates that factors such as labour market regulation; unionisation; the nature and extent of employer organisation; the role of industry peak bodies; the broader political, institutional, and cultural context; the degree of federalism in a country; and the relative powers of national governments and states/provinces all affect how people are educated for different occupations, and how the relationships between education and training systems and labour markets function (Iverson & Stephens, 2008; Martin, 2017; Streeck, 2012; Thelen & Busemeyer, 2012). These factors interact with each other in complex ways, such as how incentives to learn are shaped by labour market opportunity structures (Keep, 2012).

The changing structure of labour markets adds further pressure. In wealthy countries there is increased casualisation, rising youth unemployment, and a trend toward ‘hourglass’ economies – where the middle-level skills section of the labour market for which traditional TVET systems predominantly prepare learners is shrinking, and employment opportunities are mainly for those with high- and low-level skills and unskilled labour. In African countries, the numbers of people employed in formal, well-paid jobs have always been very small. Stagnant economies and deindustrialisation, with some exceptions, make it increasingly difficult to build TVET systems. In the regulated sector of the labour market, professions are sometimes protected, but there are few well-remunerated, protected, and stable employment opportunities beneath them, other than in the civil service. It is hard to improve vocational education because it does not, in fact, lead to many good labour market opportunities (Allais, 2018b).

Clearly the ‘problem’ with skills is more complex than the deficit-based approaches within supply-side policy that blame educational institutions for being inflexible and unresponsive, and for not producing the right kind of graduate with the right skill set for work. Governments in wealthy liberal market economies have been trying to ‘fix’ TVET for decades, without paying attention to the structure of the labour market, the way in which demand for skills is articulated, and the role that workplaces need to play in supporting the development of skills (Keep, 2005; Raffe, 2015; Wheelahan, Buchanan, & Yu, 2015; Wolf, 2002).

The massification of higher education in many countries aggravates the difficulties faced by TVET systems, because anyone who can access higher education will rather do so than enrol for a TVET alternative. All of this reinforces the ways in which TVET is expected to play a highly compensatory role, which in turn makes it particularly difficult to introduce specialised knowledge, and very complex to take decisions about what should be prioritised in the sequencing and selection of knowledge in the curriculum.

This is exacerbated by other factors that are more common in Africa and other post-colonial contexts than other parts of the world. In many African countries, formal TVET systems are tiny, low in status, and fraught with crises:

Although some countries have invested in the development of greater access to technical and vocational training (Egypt, South Africa, Morocco and Tunisia, for example), most still have training systems that cater for only a very small minority (between 1% and 6%) of the young people in education. (ICQN/TVSD Ministerial Conference, 2014)

A study of southern African countries’ TVET systems commissioned by UNESCO found that while Southern African Development Community (SADC) countries placed significant emphasis on TVET in their public commitments, the (rather limited) data available highlighted a number of weaknesses in the systems. These included poor management information systems, a lack of a common definition of what is meant by TVET between and within countries, uneven or weak quality assurance systems, policy incoherence, complex governance and funding arrangements, and generally low levels of employer participation (McGrath et al., 2013).

The UNESCO SADC report highlights a problem we face as researchers interested in TVET – lack of system level data:

It is evident that TVET Management Information Systems (MIS) are often absent or weak. This was by far the worst ordinal indicator in the national monitoring reports. There are huge data gaps and weaknesses for several indicators. … Even getting basic data and ensuring their accuracy is [sic] beyond some systems, and the ability to disaggregate for target group, to compare public and private provision or to do any forecasting are beyond the horizon for several countries. It is difficult to see how TVET systems can be successfully transformed when there is a lack of feedback data at both institutional and national levels. (McGrath et al., 2013, p. 17)

Historical factors have worked against building strong formal TVET systems in many developing countries. Despite the variations (shaped to some extent by different colonial legacies), most countries place a high value on TVET in policy pronouncements as a vehicle for addressing economic and other social challenges, but the societies generally do not place a high value on TVET, with academic routes being viewed as the preferred educational pathway by the general public (Zeelen, 2015). The reasons for this are many and complex. Foster (1965) advanced the classic thesis that young people’s preference for general education is rational despite policymakers’ attempts to guide them into vocational education. Nherera (2000) pointed out that the state has always been the major employer in African countries, and that the state has always favoured general education. Another explanation is that African countries have not been able to catch up with industrialisation (Amsden, 2007). Even those countries in Africa with stronger TVET systems tend to be primarily centre- or school-based rather than apprenticeship- or internship-based. TVET programmes thus often have quite weak linkages to employers, and the data around youth transitions from TVET to work is often poor. Thus, there are high expectations for one of the least respected and resourced parts of the education system.

None of this means that there are no successes for TVET in Africa. There are certainly examples of good practice as well as innovation in many institutions, and some pockets of excellence exist. And international policy focussed on building access to high quality technical education is important. But the capacity to respond to the demands of high-level goals or targets imposed from outside the education system is limited. International and national targets tend to lead to an emphasis on policy reform within the sector for which the targets are set. These high-level targets do not take account of the complex relationships the education system has to factors outside itself. This could be very damaging for TVET systems, especially in poor countries, and especially those where formal TVET systems have historically been very weak. It places high expectations on weak institutions, whose weaknesses are in many instances shaped by their context: what Busemeyer and Iverson (2014, p. 242) call the ‘institutional context which shapes the level and composition of skills’.

The second problem with targets, which is more acute for poor countries and for TVET than the rest of the education and training system, is that targets focus too narrowly on one factor, such as access, which can be dangerous and destructive. This is widely argued to have been one of the problems with EFA – simply flooding schools with learners without changing anything else in the system and claiming that the targets have been met (UNESCO, 2015b). In countries as diverse as South Africa, Uganda, and Ethiopia, one of the things that has undermined quality and completion rates in both university and TVET systems over the past decade has been dramatic expansion in order to meet targets, but with insufficient accompanying financial expansion. Educational expansion implies not just scaling up (which in many cases has not happened) but also the adding of more resources per student than was the case before, because the expansion of numbers often means expansion to student populations with weaker educational backgrounds.

Pressure to expand access to TVET invariably means that the expansion focusses on the formal state system and takes the form of increasing the number of colleges or expanding their capacity to enrol more students. This college model of TVET is driven by a school logic and does not mean that students graduating from those colleges will have access to the world of work or genuine labour market opportunities upon completion. Instead, often this type of expansion is about signalling certain messages that have little to do with TVET. In addition, because of the particular colonial, neocolonial, and Bretton Woods institutions’ influences, the organisational forms and the policies that regulate educational processes (curriculum, assessment, pedagogy) tend to mimic systems that have been developed for different contexts. This mimicry takes various forms, either through symbolic policies that signal high-level intentions without clear procedural or resource specification, borrowing of policies and strategies that are inappropriate to the context, or focussing on form over function and scale over quality (Fuller, 1991; Harley & Wedekind, 2004; Jansen, 2002; Mattson & Harley, 2003; Phillips & Ochs, 2003; Steiner-Khamsi & Waldow, 2012).⁴ This has resulted in education systems with uneven or generally fragile institutions with limited capacity to respond to the pressures to address the high-level intentions and targets that are imposed through international political processes and the influence of international development agencies.

The SDG indicator for TVET (4.4.1) currently reduces the whole of TVET to ICT skills. While such a narrow target is problematic, another problem is created when targets are too broad. Within TVET reform, there is a tendency for academic scholars, policymakers, and civil society actors to emphasise a broad role for TVET. We are not arguing for narrow and overspecified TVET; in fact TVET, particularly in today’s labour market and particularly in contexts where many learners have weak school education, should contain substantial components of general education (Gamble, 2013). So UNESCO, for example, argues for the importance of a wider lifelong learning perspective informed by a human development perspective (UNESCO, 2016c). The SDGs oblige TVET to include education for human rights, sustainable development, and citizenship. This sounds desirable at face value. But in practice, it places yet another obstacle in front of weak institutions trying to offer courses that are not highly valued, often to poorly prepared students. Actual policy modalities are difficult in TVET; it is really not easy for governments to figure out what to do. And, as we discuss in the following pages, this often leads to government posturing.

In summary, we have discussed a number of ways in which targets can be particularly problematic for TVET. If they are overspecified by focussing on, for example, expanding enrolment or measuring access to information technology without recognising the constraints on the system to do so, they can result in unintended consequences such as pockets of excellence being weakened or destroyed, and resources being allocated for one part of the curriculum rather than another. If they are too broad, they can result in TVET systems being expected to take on more functions than their capacity allows. Because goals are almost always divorced from the TVET system’s internal dynamics, they can potentially damage the system that they are supposed to be strengthening. And because goals are also focussed on TVET as a system, they ignore the ways in which the broader economic and social context shapes the possibilities of improving TVET.

4 Reform and Policy Posturing

In this section, we draw on our joint experience working as academic researchers focussed on knowledge, curriculum, pedagogy, and political economy, but also as participants involved with national policy at the highest levels since the democratic transition in South Africa, with a considerable focus on policy for TVET.⁵

In our experience, the SDGs have not been a major factor driving policy development or implementation in South Africa, and there is little evidence that they have been anything more than additional reporting obligations for people in the Ministry of Foreign Affairs and statistical offices. The same can be said for national target-setting. Our experience of working in and with government in education policy development and implementation, particularly in TVET, is that there is enormous pressure on government officials to engage in posturing rather than actually changing things. There are many forces that push governments toward ‘playing at doing government’ – being engaged in activities that are essentially posturing, or busywork.

Michael Fielding playfully suggested, ‘[Target-setting] is, in one sense, the viagra of economic and educational under-performance: set some targets and you’ll feel better, be seen to get something done and satisfy the prurience of an increasingly promiscuous accountability’ (1999, p. 277). Spreen captured a similar idea, arguing that stating something in a policy document tends to create the impression that it will happen:

My own experience in educational reform in the United States has made me susceptible to the argument that when policies are put on paper with a coherent logic we are deluded into thinking fundamental change is taking place. Observers of policy reform often do not take into account what it takes to truly alter the structure of society or its institutions nor do they consider important distinctions made on the ground by those implementing the policies. (Spreen, 2001, p. 17)

This is even more likely to be the case with a goal such as SDG 4, which has a number of systemic implications that may be at odds with each other. The inclusion of education and skills at all levels – at the survivalist level, at the mid-level of occupations, and at the level of complex professional work – may sound straightforward but it is unhelpful for policymakers who have limited resources.

Target-setting has a particular allure when the modalities of actually making things happen, and making things work, are not obvious or easy to create; and TVET systems are examples where policy decisions are complicated and difficult. Some examples of poorly planned policy modalities include a recent decision, in both Zimbabwe and South Africa, to link vocational programmes more closely to employers by making it compulsory that students are attached to work places. The net effect in both countries has been either to reduce enrolment dramatically, or to allow people to enrol, but then never be awarded their qualification, because they are never able to complete the workplace component. This problem was discussed in many national planning meetings in which we participated.

Another example relates to the introduction of a new TVET certificate, which was formally phased into South African TVET colleges in 2007. This was a qualification that was far broader than its predecessors, such that it resembled the broad kind of TVET for which academic researchers have been advocating (McGrath, 2012; Wheelahan et al., 2015). On paper it looked good. In practice, employers were unfamiliar with it, and graduates found that it had little purchase in the market place. Colleges did not have the capacity (both in terms of infrastructure and teachers) to deliver the programme as the curriculum designers had intended, and the students who entered the programme were not eager young people interested in exploring a particular occupational field, but in the main were either high school drop-outs/push-outs or high school graduates with no access to other study or work opportunities. These students more often than not had experienced difficulties studying mathematics at school and were hoping for something more practical. The curriculum, however, was demanding. So, the system did not have the capacity, and the society did not have the understanding, to make this qualification worthwhile or productive (Wedekind, 2018).

South Africa is a good example of a country with a penchant for high-level plans and targets. The country has a National Development Plan (NDP); a Human Resource Development Council with its own strategy; a National Skills Agency with a National Skills Development Strategy; and national plans for parts of the system such as the post-school system, industrial development strategies and plans, and economic and sector-specific plans and strategies. Each strategy or plan has targets or goals or initiatives that have a direct impact on the institutions that need to deliver the education and skills part of that plan.⁶ For example, the NDP proposes expanding the system of training artisans to 30,000 graduates per annum by the target year of 2030. This aspirational target, set with no clear justification, has become something of a mantra in the skills system, with much of that system working toward achieving a target that is not based on any explicit rationale. In South Africa the term artisan refers to a very limited set of trades in the engineering field and does not take account of the shifts in South Africa’s economy (the financial services sector is now the largest) or any clear analysis of what the mining and manufacturing sectors actually require. While we do not take issue with the potential benefits of having greater numbers of qualified artisans, the point is rather that a relatively arbitrary target has been elevated to a holy grail. The achievement of this target is one of the primary drivers shaping the TVET system, potentially to the detriment of training provision for other sectors and occupations. Indeed, the recently announced vocational stream within the schooling system is justified primarily in terms of meeting this 30,000-artisan target, despite the fact that there has been no discussion about how this school track actually articulates with apprenticeships and the qualification of artisans. Once again, the target seems to develop a life of its own without account being taken of the complexity of the institutional arrangements.

An example of perverse outcomes can be seen at the level of policy and strategy. Because of a concern about TVET and skills development funding being channelled into short skills courses (of varying quality), and a wider critique of narrow training, the third iteration of the South African National Skills Development Strategy placed high priority on the funding of full qualifications that would supposedly ensure that workers would be able to progress into higher or other forms of work. While not referring directly to international goals, the rationale for this in the strategy aligns with the notion of lifelong learning – that education should not just focus on efficiencies and productivity training, but empower people to develop and grow within the occupation and articulate into other careers or higher learning.

This high-level goal was then translated through the funding mechanisms of the various Sector Education and Training Authorities and resulted in a major shift away from short course training to supporting students in full-time qualifications at public TVET colleges. The problem with this focus on qualifications was that existing employees tended to be excluded as neither they nor the employers were willing or able to commit to extended periods of study in order to get qualifications that were not highly regarded in the industry. Furthermore, in some sectors (such as ICT for example) the majority of employees already had appropriate qualifications, and short courses to update knowledge or introduce new technologies were the most appropriate intervention; yet, the commitment to a national strategy meant that they could not be prioritised. The consequence was that some employers funded their own training, and there was a weakening of the link between the public provision of training and employers.

Each of the above examples is more complex than presented here, but the overarching point is this: whether at global level such as the MDGs and SDGs, or at regional or continental level such as AU and SADC, or at national level in plans and strategies, setting well-intentioned and ambitious targets has two effects. Either they force compliance with the form without achieving the intended outcome, or more often than not, they create distortions that have unintended but serious consequences, particularly in further weakening the TVET system by expecting it to do things that it was not designed to do. This creates a vicious cycle in which institutions that have low status in society are further undermined, and thus the prejudices are confirmed.

A major problem with targets is fake target-meeting or gaming. For example, governments commit to creating jobs and then release lists of ‘job opportunities’ that have been created. There is much research on the fact that targets are proxies, and focus on meeting the proxy often displaces the meaningful activity to which it was supposed to lead (Elton, 2004; McNay, 1999; Campbell, 1979). Of course this is a general problem, as Goodhart’s Law⁷ suggests, not a TVET- or even education-related one.

Our experiences in South Africa suggest that policy posturing and target-setting do little to strengthen and develop the TVET system and in many instances have weakened it. We suggest that target-setting and internationally comparative measures should be treated with great caution.

5 Conclusions

TVET has drawn increasing attention from international donors and national policymakers; its explicit incorporation into SDG 4 can be seen as part of that. We argue that a general problem with target-setting for social policy is that setting goals for a specific system creates the illusion that the system itself can achieve those goals – that policies and actions endogenous to the system under consideration are required, even when some lip-service is paid to context or interconnectedness.

TVET offers a particularly stark reminder about why that is not the case. Economic factors, labour markets, and the nature of school and university systems largely shape the possibilities and difficulties experienced in TVET systems. This reinforces the tendency of governments toward posturing, which is most likely in areas where it is hard to figure out the actual policy modalities required to change something or make something happen. There are many forces that push governments toward activities that are essentially posturing, or busywork. This has manifested in serial reforms in all aspects of education systems, but particularly, in TVET. Where progress has been made – and despite the extreme fragility of TVET there are pockets of progress – it is often in spite of, not because of, grand target-setting.

We suggest that the focus on TVET and skills in the SDGs has done little to improve TVET, and may weaken it, because vocational education in developing countries can only make a limited contribution to equity in the context of substantial labour market reform and industrial development. National and international targets in the main get in the way of building TVET systems in Africa that can make this limited contribution.

Target-setting for TVET tends to lead to the designing of TVET systems for labour markets that do not exist and for student populations that countries do not have. There are vastly inflated ideas of what TVET can and should do, and by expecting too much from TVET we make it very difficult for it to do almost anything. Governments take fragile institutions and give them impossible tasks, and then create the illusion that the reasons for the failure of TVET systems are reformable by changing TVET policy.

Does this mean that there should be no international commitments and goals that shape aspirations for national systems? We believe that the SDGs can serve a useful purpose in focussing attention on a sector that has been poorly understood. However, we would suggest that narrow indicators such as enrolment, throughput, or provision of ICT should be avoided. Instead, countries could be asked to account for the extent to which there are policies in place to support the development of a TVET system in its full complexity. These policies must be related to TVET, and not subsumed within a general education framework. Such policies should address the specific contextual conditions of the TVET system, the labour market, and social and political interactions, because this complexity has to be tackled by governments, which want to improve skills levels and productivity. Governments should be asked to report on what policies are in place, what resources are committed to the implementation of those policies, and how they will be monitored and evaluated. Because of this context-oriented approach, a qualitative process of peer review rather than a technical evaluation may then be a more appropriate mechanism, as substantive, contextually sensitive judgements need to be made.

We noted earlier that there is a paucity of data and research on TVET in many developing contexts. In part, this is related to weaknesses in administrative capacity, but also the fact that vocational education has a very weak research base in universities. In order to develop and monitor the policies proposed above, and deepen the insights into the dynamics of the system, this research capacity needs to be strengthened at universities. A concrete measure on which governments could report would therefore be the degree to which they are supporting the development of this critical capacity.

TVET work has always been, and will continue to be, extremely difficult in poor countries. What is needed is dedicated energy, resources, time, creativity, and experimentation. TVET systems are never going to solve large-scale educational or social problems in the absence of major structural economic and labour market reform. The corollary of this is that the more pressure there is on TVET to solve such problems, the less it will succeed even on a modest scale.