Design flaws impede policy implementation

Lawrence Schlemmer thinks that delivery problems may be more complex than orthodox opinion realises.

Summary - It is common these days to hear political observers talk about ‘gaps’ between policy and performance. For example, they may say that local development policy is sound but that its implementation is deficient or needs improvement. However, it may be that they have the wrong end of the stick, and that the problems originate, not in delivery, but in the policy itself. In a perceptive analysis published last year, professor Louwrens Pretorius of Unisa outlines two different views on public policy. The linear view sees policy-making as a distinct stage in a sequence; policies are self-contained and can be evaluated without reference to subsequent performance. This view is prevalent in South Africa today. Policies are admired for their intrinsic ideological merits, using a checklist of virtuous criteria such as equity, empowerment and representivity (terms that are often written into the policy texts to underscore the point). Linear adherents then look beyond the policy to identify causes of failure, hence the consensus in government circles, reinforced by ‘experts’, that their policies are excellent, and that therefore failures in implementation must be blamed on other problems such as the legacy of apartheid or lack of capacity in the delivery machinery. The second view Pretorius identifies, which is now accepted by analysts overseas, is the interactive view, which holds that policy is continuously remade as a result of interaction between the domains of policy-making and policy implementation. Thus a policy is defective if it fails to make provision for lack of capacity at the implementation stage. For example, the reforms in school funding and pupil-teacher ratios, the imposition of training levies on business, the upgrading of the taxi industry and the provision of dual-medium education all illustrate (or will illustrate) the iron law of policy in practice, which is that desirable planned consequences are often overshadowed by disastrous unintended consequences. Thus the education reforms have led to an exodus of the country’s best teachers, the training levy has created countless opportunities for self-aggrandisement and inefficiency, the taxi policy is being implemented in a nearly ungovernable industry, many elements of which are hostile to the reforms, and parallel-medium education has become a formula for the destruction of Afrikaans. Many of these policies are the work of highly paid consultants who are not accountable for the results. This must change. All new policy should be drafted by experienced senior officials who must submit a rating of the likelihood that the policy can be implemented, as well as an assessment of its possible unintended consequences.

It has become conventional wisdom that although government policies are frequently very sound and sophisticated, their implementation is equally frequently disappointing or deficient. Hence observers across the political spectrum talk of “gaps” between policy and performance.

A key example is seen in local development policy. Municipal re-demarcation has dramatically extended the boundaries of local authorities, integrated urban and rural administration and shared local tax bases between urban middle class settlements and rural subsistence communities. Transfer payments from the central treasury are dedicated to expanding under-developed infrastructure. The plan for uplifting the poor is comprehensive, if not heroic.

Yet in the Eastern Cape there are rural roads that have deteriorated to such an extent that small farmers without tractors or 4X4s can no longer transport their potatoes to their erstwhile markets. Often the roads need no more than a thorough grading, but funds, equipment and manpower are not available for the task.

We read in ThisDay (October 17, 2003) that the Eastern Cape provincial government spent R277 million on advisors and consultants over the past financial year, an increase over the previous financial year of 207 per cent. The cause of the increased expenditure is said to be “gaping staff vacancies” and the consequence is crippled delivery capacity. The situation in the Eastern Cape is more serious than most but these problems arise in all areas and provinces.

The general view in government and among most media commentators seems to be that the policies are fine but that delivery must be given an injection of adrenalin. Unfortunately it is not as simple as this. We might be trying to beat down the barriers to delivery in vain. The problems lie with the initial policies.

In a perceptive analysis of gaps between policy and delivery published last year, professor Louwrens Pretorius of Unisa distinguishes between two views in the literature on public policy.1 One view sees policy as the first step in a “linear” process or sequence. Policymaking is a distinct stage before implementation and therefore policies are self-contained and can be evaluated separately from the subsequent performance.

As Pretorius emphasises, in South Africa today we are particularly prone to evaluate these policy products in terms of rather virtuous but repetitive checklists of criteria like “equity”, “empowerment”, “participation”, “integration”, “inclusiveness”, “representivity”, etc. To make doubly sure that these virtues will be recognised, our policy makers usually write these terms into the policy texts, with the result that many policy documents have a kind of inbuilt self-praise and ideological smugness. Professor Jonathan Jansen has observed that our policymakers regard their published policies as products to be admired on their intrinsic ideological merits, with cavalier disregard of what the actual consequences may be. Policy consultants, particularly those with progressive academic backgrounds, are admirable in helping politicians to produce these works of “ideological artistry”.

If one has a linear view of public policy one looks automatically at the stage beyond policy-making to identify causes of failure. Hence the running consensus in government and among its circles of adherents that our policies are excellent and that we have to blame failures in implementation on such post-policy demons as the “legacy of apartheid”, the reluctance of the old middle classes and businesses to lend help and above all the lack of programme management and professional capacity in the delivery machinery. This leaves the policies themselves unsullied by the subsequent failures, so much so that any opinion poll in Parliament would probably find that most MPs would judge the old failed “RDP” to be a great policy wrecked by conditions beyond its control.

“Experts” reinforce this pattern. At an Aginfo/AMT Agricultural Outlook Conference in Pretoria on 30th September 2003, for example, the influential policy analyst professor Willie Esterhuyse reviewed the entire field of South African policy and came to the resounding conclusion that it was all sound and well directed. This conclusion was protected by the fact that he qualified his presentation with the statement that he would not consider implementation — linear segmentation in action.

The second view that Pretorius identifies is the concept of policy as an interactive process: “An alternative to the linear model is one that depicts policy as being relatively incomplete and as being continuously (re)made… (involving) more or less conscious, more or less reflexive, more or less coherent interactions between processes (in the) domains of policy-making and policy implementation respectively … policy-makers remain co-responsible for success or failure. (p. 17-18).

This view is now accepted by experienced analysts abroad. In South Africa it would insist that any policy is defective to start with if it fails to assume and predict in advance that there will be serious problems of implementing capacity. It would be tantamount to designing a fine racing car without wheels. Self-evident as this may seem, South Africans have become so wrapped-up in large goals and commitments to transformation that they miss the obvious. Without building in, and thereafter working through, all the easily predictable devils in the detail, policy-making degenerates into ideologically elegant political masturbation.

Many high level policies have been subjected to informed criticism but the symbolic power of their virtuous goals has wrapped these policies in self-congratulation and hence in political armour. The radical reforms in school funding and pupil-teacher ratios in education made perfect sense against a background of apartheid inequalities in education. South Africa needs to upgrade its level of human resource development and training and in this light the imposition of training levies on business to fund Sector Education and Training Authorities (Setas) seemed an obvious need in a situation where business has traditionally short-changed training. The new legislation providing for the upgrading and modernisation of the taxi industry is justified on dozens of grounds — safety, cost-effectiveness, control over standards and the like. Pressure for parallel or dual medium education to improve access to all schools by all learners seems logical and utterly fair, given that the continued use of the original language (usually Afrikaans) is nominally accepted.

Each one of these few examples, however, illustrates or will yet illustrate an iron law of policy in practice — the nemesis of unintended consequences. They often overshadow the consequences planned in the policy. We know now that the educational reforms led to an exodus from the profession of South Africa’s best teachers and a rapid growth of private education that has robbed the state system of significant income flows. The Sector Education and Training Authorities (Setas) while both “inclusive” and honed to the specific human resource needs within sectors, have led to 25 boards of often inexperienced stakeholders and with that a proliferation of opportunities for self aggrandizement and inefficiency. A steady flow of examples of poor management seems to have convinced the minister that he will have to exercise more central control over Setas, but central bureaucratic control has its own well-known hazards.

The new taxi policies will be implemented in a nearly ungovernable industry (that already includes nodes of fierce resistance to the policy) and may simply lead to the return of completely illegal broken down vehicles to undermine the profitability of the new larger and approved taxis. Parallel medium education sounds like a very reasonable compromise between culture and access, but where it has increased teacher loads and brought the indigenous language into competition with the international medium of English, it has become a formula for the covert (possibly even intended) destruction of indigenous Afrikaans.

To return to the initial example of the expanded and integrated local authorities, the longer lines of reticulation of urban services has increased service loads and complexity for the majority of local authorities that are financially marginal and debt-ridden, at a time when they are rapidly losing experienced officials. These complexity-capacity ratios may yet cripple the brave new experiment.

These elevated policies that are so estranged from problems of implementation are more often than not the products of consultants. There is a place for consultants but surely not as many as the R277 millions worth in the Eastern Cape. One concern is the inevitably higher expenditure for the same product that should have been provided by the professionals and senior administrators in the civil services themselves. A second concern is that where consultants fill vacancies that cripple the effectiveness of administrations, it exposes a critical lack of depth in the established professional echelons. A third is the ever-present danger of “near-corruption” in the selection and engagement of consultants, for whom the cultivation of political and other connections to smooth appointments must be recurring temptations.

Perhaps the most serious concern is the lack of ongoing accountability that fixed-term or project-based consultancies can imply. Often the contracts do not provide for the consultants to participate in implementing their plans or to evaluate their results. They are external to the unwritten contract that is supposed to exist between the public administrations and the communities that they serve.

This situation leads one to contemplate the need for an over-riding rule in the public administration to govern all new policy. Ministers obviously must set the end goals, and consultants and advisors may help with the technicalities, but all policy and strategy should be “authored” by experienced senior officials who have to submit to the Auditor General a rating of the likelihood of the policy being implemented and an assessment of its possible unintended consequences. If senior officials with proven experience are not available, then policy implementation in each relevant location must simply be deferred until they are. Such a rule is arguably as necessary as public policy itself.

1 Louwrens Pretorius, “Six contributions to understanding ‘gaps between policy and implementation’: an overview and comments”, Politeia, vol 22, no 1, 2003, pp 6-21.


Flaws in the policy-design process apply to some extent even to the most acclaimed policy of all among audiences of centre-right persuasion, the government’s macro-economic “Gear” strategy of 1996 onwards. The goals were comfortingly aligned to the international economic consensus and even inspiring in that they committed the country to international competitiveness. Since then a soft warble of endorsement and gratitude in the business community has been background music for the policy. Most policy analysts were utterly captivated by the sight of former Soviet-endorsed socialists embracing the market so fulsomely. Although they do not do the same when policies are counter-establishment, it was mainly left wing observers who expressed alarm that the policy was not even cleared in advance by Parliament, let alone the labour movement and spokespeople for the unemployed.

But the main problem was that the policy seemed to be based on an economic model that, aside from some obvious technical contradictions, took little account of the subtle but powerful factors external to the macro-economic variables themselves that inhibit savings, fixed investment and the expansion of production capacity. Even now that GDP growth and employment creation have lagged badly behind early Gear predictions for six long years, the coterie of business and bank economists still pay homage to the wonders of Gear. Aside from predictable left wing ideological criticisms of the “neo-liberal” fiscal and monetary regime, virtually no one has reached the obvious conclusion that any economic policy that does not fully allow for predictable externalities cannot be good policy.