Infrastructure in South Africa: Who is to finance and who is to pay?

Abstract

Against the backdrop of shifting views on the role of government in the provision of infrastructure, this paper distinguishes between the payment for and financing of the South African Government's infrastructure investment programme. The paper argues for a clear distinction between loan financing by the government for macroeconomic considerations and the benchmark approach to the financing of infrastructural projects. It presents a classification system that enables a systematic mapping of all prospective projects, with reference to considerations of efficiency and equity, and uses this system to question the government's financing strategy and identify alternatives. The Gautrain Rapid Rail project is used as a case study to demonstrate these alternatives.

Keywords:

• infrastructure financing
• government loans
• benefit taxation
• guarantees
• private–public partnerships
• Gautrain
• South Africa

JEL codes:

• H54
• H81
• H72

1. Introduction

In economic growth theory, investment in physical capital is one of the primary factors that enhance labour productivity and long-term economic growth (see neoclassical growth theories, e.g. Solow, 1956). Other factors that play a role are technology and human capital investment (see Romer's (1994) endogenous growth theory), and institutions (see North, 1990). These factors are particularly important in helping developing countries catch up with developed countries in per-capita income.

The role accorded to government involves, among other things, increased saving (e.g. running a ‘neoclassical’ budget surplus or incentivising private saving), thereby increasing per-capita growth en route to a new (higher) steady state; investment in physical and human capital, either directly or via incentives to the private sector; and creation or facilitation of the development of formal institutions (e.g. the constitution and the legal system) and informal institutions (e.g. norms and tradition), which create incentives for productive actions as opposed to ‘rent seeking’. Cyclically, enhanced Keynesian economic growth is pursued through increased aggregate demand, within which context increased public investment (in infrastructure) could be part of a government (deficit) strategy. Investment in infrastructure thus has an important place in both a short-term and a long-term growth strategy. In recent times, infrastructural investment has featured strongly in the development plans of high-performance East Asian countries and in the recovery strategies of industrial countries (e.g. the US in 2009).

Infrastructure is high on the South African Government's expenditure priority list, having been identified as one of the ‘binding constraints’ by an international panel, chaired by Hausmann (2008), advising the government on enhancing economic growth. Following the Accelerated and Shared Growth Initiative of South Africa (ASGISA) (RSA, 2006), the 2006 National Budget envisaged that R415.8 billion would be invested in infrastructure over a three-year period (National Treasury, 2007:45). In the 2009 Budget Speech, Minister of Finance Trevor Manuel noted the importance of continuing and expanding the expenditure on infrastructure, amidst fears of an impending international economic recession that would increase the costs of borrowing. This surge in investment followed the previous period of major cutbacks in public-sector investment, an important although not undisputed factor in the successful decade-long attempt at restoring fiscal sustainability (Calitz & Siebrits, 2005:255).

The pivotal role accorded to public investment in ASGISA raises questions that, during the past two decades, have generated increasing research interest in infrastructure and its proper financing (Estache, 2004:7). What are the financing needs of infrastructure? How much scope is there for private participation in infrastructure financing? How much fiscal space do governments have to meet infrastructure financing needs, and what should drive the new wave of financing mechanisms? In many respects these questions boil down to two important ones: how is this infrastructure expansion to be financed and who is eventually to pay? Important in this regard is to distinguish between structural financing choices, reflecting the basic or benchmark approach, and cyclical fluctuations in the appropriate source of finance, reflecting considerations of shorter-term macroeconomic adjustment.

This paper attempts to answer the above questions with reference to South Africa. Building on a theoretical treatment of infrastructure finance alternatives, reflecting the more structural or benchmark approach (Section 2) and the South African Government's existing financing arrangements (Section 3), the paper proposes an expansion of the alternatives for existing infrastructure projects (Section 4). A case study of the Gautrain Rapid Rail project is used to contextualise these proposals (Section 5).

2. Views on infrastructure financing: Towards a structural or benchmark approach

The financing of and payment for infrastructure are important not only for allocative and technical efficiency but also because of certain distinctive economic characteristics of infrastructure – a high capital intensity, externalities, elements of natural monopoly, and location specific investments (Fourie, 2006) – all of which affect the nature and extent of government involvement and private sector incentives to commit long-term capital (World Bank, 2006:150).

From an efficiency point of view, the traditional view as to whether and to what extent government should assume responsibility for the provision (read: financing) of public goods has shifted substantially over the past few decades. Infrastructure provides public services and therefore assumes some kind of public sector involvement. Three types of market failure are relevant to infrastructure provision – public goods (goods with the usual characteristics of non-rivalry and non-excludability), externalities (marginal private benefit is smaller than the marginal social benefit) and incomplete markets (of which natural monopolies with decreasing average costs are the best example). Each type requires a different kind of government intervention: pure public goods require government to prompt the process and incur the cost of provision, externalities and incomplete markets may require government to take a more regulatory role, and incomplete markets may at times also require a financing role. It is important that a government does not assume an inappropriate role, or expect inappropriate involvement of private business. In infrastructural development it is not easy to divide the responsibilities and the associated incidence of risk between government and private business so as to achieve a proper balance.

Different kinds of market failures also necessitate different methods of budgeting and means of financing. Traditionally, pure public goods (such as street lights, highways and water storage) are financed by government through present or future taxation, with no or only partial user charge cost recovery. Positive externalities and natural monopolies create opportunities for benefit taxation through user charges or earmarked taxes, with or without government or private loan finance, while the use of private equity occurs where private-public partnerships are found in natural monopolies, as increasingly happens.

The respective roles of government and private business are shifting with changing technology and innovation. Figure 1 is an attempt to illustrate the way the traditional public finance view, in terms of which little if any infrastructure was seen to be provided by the market, has shifted to a view which accords a bigger responsibility to the market. The information and communications technology (ICT) revolution, the development of competitive markets and a better understanding of different ways in which market competition can be simulated, have all played a role. Two examples include the move towards alternative sources of power generation (given improved wind-powered, solar-powered, and wave-powered technology) and the growth of cellular and broadband technology in the telecommunications industry.

Figure 1: Shifting view on the ‘publicness’ of infrastructure

More importantly, the shift towards rival and excludable infrastructure goods has also reshaped the traditional view of infrastructure financing. Figure 1 shows that, because excludability has increased and rivalry can be simulated to a greater degree, the ability to determine a price (user charge) has gained ground over taxation. Consequently, a much larger range of goods and services (including infrastructure) traditionally classified as tax-financed public goods are now regarded as ‘priceable’ self-financing activities within or outside of government budgets. This is depicted by an expansion of the ‘infrastructure ellipse’ in Figure 1, towards the top (indicating a greater measure of excludability) and to the right (more rivalry); in combination this implies greater private sector involvement. Of course, a change in the finance mechanism also has distributional implications, to which we refer later.

New ideas about ‘pricing’ previously non-excludable goods have been developed to align beneficiaries of infrastructure closer to those responsible for bearing the cost. Vito Tanzi (2005) notes that earmarked taxes, fuel taxes for road construction and maintenance in particular, have become more important, especially in developing countries with fluctuating budgets where infrastructure is usually the first budget item to be cut in times of crisis. One downside of earmarked taxes, as opposed to user charges, is of course the rigidities with respect to prioritisation that they introduce into a budget system, with possible adverse consequences for allocative efficiency. There have also been suggestions internationally of imposing ‘betterment taxes’ to pay for infrastructure projects, such as taxes on the value of the land in the vicinity of the proposed new infrastructure. The idea is that the infrastructure project will increase the value of the land in the locality of the project, with a land tax ensuring that those who gain most from the infrastructure pay for it. Of course, there are many problems with such payment options, but they nevertheless suggest a move towards the benefit principle where economic infrastructure is concerned.

To sum up: from an efficiency point of view, recent views increasingly set the private supply and even financing of ‘public’ infrastructure as benchmark. Even support for taxation based on the benefit principle entails specific taxes on identified beneficiaries, rather than general taxation, as the appropriate method of cost incidence.

3. Towards an understanding of the South African Government's infrastructure financing strategy

Between 2005 and 2008, the increase in infrastructure expenditure by general government (i.e. national, provincial and local government combined) (Table 1) has been financed mostly from current revenue, i.e. general taxes. The growth in infrastructure expenditure by the public enterprises is particularly striking: while infrastructure expenditure by general government was more than double that of public enterprises in 2005/06, expenditure by public enterprises is projected to exceed that of general government by 2009/10.

Table 1: Public sector infrastructure expenditure and estimates, 2005/06–2011/12 (R million)

	2005/06	2006/07	2007/08	2008/09	2009/10	2010/11	2011/12
General government	48 559	57 869	73 426	100 442	118 288	128 098	144 219
Public enterprises	22 145	25 736	56 765	90 192	119 585	131 335	145 842
Total	70 704	83 605	130 191	190 634	237 873	259 433	290 061
Source: National Treasury (2009b).

Most of the expansion in infrastructure (both in general government and in public enterprises) was initially financed through current revenues. The national budgets of 2006/07 and 2007/08 predicted a budget surplus, with a small budget deficit predicted for 2008/09. Given that there was virtually no borrowing within provinces, this implies that the infrastructure was on balance financed by current revenue (tax income) (as implied, for example, by the negative borrowing requirement by general government from 2005 to 2008 shown in Table 2). The investment by public enterprises was financed mostly through non-tax income or user charges (profits), as implied by the negative borrowing requirement for 2005/06 in Table 2.

Table 2: Public sector borrowing requirement, 2005/06–2011/12 (R million)

	2005/06	2006/07	2007/08	2008/09	2009/10	2010/11	2011/12
General government	–207	–15 979	–23 351	33 097	94 312	83 405	67 660
Public enterprises	–6858	11 277	11 182	57 362	91 434	90 069	90 103
Total borrowing	–7065	–4702	–12 169	90 459	185 746	173 474	157 763
Source: National Treasury (2009b).

The 2009/10 budget highlights a significant shift in government policy regarding the financing of infrastructure. In particular, the 2009 Budget Review notes that ‘the long-term infrastructure programme of the public sector requires significant capital to be raised through debt’. More importantly, expenditure on infrastructure by non-financial public enterprises will exceed infrastructure expenditure by general government for the first time in 2009/10. Most of the additional expenditure by these institutions (notably Eskom and Telkom), will be financed through loans.1

This dramatic change in the financing strategy, apparently because of the pressures that weakening economic circumstances and prospects in 2009 brought to bear on tax revenue and, consequently, the national budget, raises a few questions about what the structural or benchmark approach to the financing of infrastructure should be. Should the basic approach be for a government to pay in cash (i.e. use current tax revenue) for infrastructural investment, which has a long economic lifetime and whose benefits arguably straddle more than one generation of tax payers? Why the use of current revenue between 2005 and 2008 by the South African Government? What other options are available and what are their advantages and disadvantages? And, consequently, why the shift from infrastructure financing through current revenues to loan financing in 2009?

For most of the first 10 years of the twenty-first century, the government's short-term macroeconomic strategy was aimed, inter alia, at softening the risks associated with the relatively high current account deficit over the period. Fiscal restraint in the form of a balanced budget or even a surplus was a key element. The aim was clearly to neutralise the inflationary consequences of high domestic spending and current account pressures on the external value of the home currency. The high import content of much of the envisaged infrastructural investment and the scarcity predicted for certain inputs and skills were likely to add further cost pressures (Sudeo International Business Consultants, 2007). The longer-term objective was apparently to increase the national propensity to save: at the time of the 2007 Budget the financing of the envisaged national and provincial infrastructural investment within a balanced budget would have directly increased the savings rate by about 2.6 percentage points.

The above approach can be questioned from different angles. Economic theory suggests two reasons for financing infrastructure through loans rather than taxes. They are intergenerational equity and the associated application of the benefit principle for purposes of allocative efficiency, and Keynesian activism for purposes of macroeconomic stabilisation. The strongest arguments against loans are the neoclassical argument of crowding out and the public choice leviathan argument. In recent times fiscal rules have gained popularity as a reaction to the disillusion with fiscal discretion, even though it has been argued that fiscal rules are unlikely to add credibility benefits to fiscal discretion cum transparency-enhancing measures (Siebrits & Calitz, 2004; Calitz & Siebrits, 2009). In fact, standard fiscal rules adopted to ensure debt sustainability as part of macroeconomic adjustment programmes are increasingly being criticised as excessively binding constraints on appropriate countercyclical action (Estache, 2004:13). To the extent that the South African Government has reduced the budget balance below the levels required for macroeconomic stability, underutilised fiscal space exists (Heller, 2005). Macroeconomically speaking, this would allow for at least two other fiscal options or a combination thereof. The government could resort to loan finance that matches the total annual capital spending, which would enable the launching of more projects than under cash financing. Another option would be to reduce the tax burden.

The government's financing strategy implies the application of the ability-to-pay approach, which severs the link between the user of the infrastructure and the person or entity who pays for it. Moreover, the strategy implies that a major part of the infrastructure has to be financed via the budgets of the three tiers of government. This raises two related questions: should government just be the financier or should tax payers actually foot the bill? The first question has to do with the assignment of responsibilities for the financing of infrastructure in a market-based economy (i.e. indicating the government's view on the public-goods properties of infrastructural facilities); the second with the appropriate incidence of the cost.

In his 2009 Budget Speech, Minister Trevor Manuel announced that more borrowing will be required to finance the increased infrastructure expenditure. The reason for the shift in policy is the ‘countercyclical widening of the main budget balance in response to deterioration in the economic outlook’ which ‘requires national government to raise significantly more finance’ (National Treasury, 2009b:54). This statement implies that the shift from current revenue to loans is not the result of a change in financing policy, but rather a matter of necessity, i.e. the need to adjust to short-term economic conditions. In fact, Mr Manuel acknowledged as much by declaring that ‘the recovery in the national government balance will lead to an overall reduction of the public sector borrowing requirement from 2010/11’ (National Treasury, 2009b:47).

The above shift, while perhaps justified by short-term cyclical considerations, leaves unanswered the question as to what the basic approach to the financing of infrastructure should be, which would serve as a reference point (benchmark) for considering and managing short-term deviations. With this in mind, the next section makes a case for a reappraisal of loan financing.

4. A case for loan financing

We suggest that one way to identify the financing alternatives is with reference to the different organisational forms in which public infrastructure might be provided.

• If the activity forms part of the functions of a national, provincial or local government and is financed (in cash) within the annual budget, the financing will be effected either through reprioritisation, additional taxes or loan finance. In the case of loan finance, the loan is likely to be serviced through future tax revenue and/or user charges. Multilateral development finance may be a supplementary source.

• If the activity is the responsibility of a designated government agency other than a normal government department (e.g. a water board, a public enterprise such as Transnet or a public corporation such as Eskom), internal financing (if accumulated reserves are available) and external financing (loans or equity) – domestic or foreign – are the indicated sources. User charges will take care of the debt servicing and the operational cost. Government intervention may be in the form of interference with price setting (such as prescribing an internal subsidisation by high-income users of low-income users). Conversely, the entities may attempt to obtain government loans or guarantees to reduce private sector loan cost. Multilateral development finance may again be a supplementary source, whilst guarantees from an institution like the World Bank can play a role.

• If the activity is the responsibility of a private business entity (either on its own or in partnership with a government agency), loan finance may be complemented by equity finance. Business and political risk may again prompt the seeking of government guarantees or the allowance of sufficiently large profit margins. The latter is a particularly thorny issue when a public monopoly is replaced by a private one and price sensitivity is great, especially when there are a large number of low-income users. A government's inclination to regulate prices often results in the withdrawal or non-supply of private equity and even in a loss of private interest in running the business. The results of a study by Kirkpatrick et al. (2006:143) indicate:

that foreign direct investment in infrastructure responded positively to an effective domestic regulatory framework. By implication, where regulatory institutions are weak and vulnerable to “capture” by the government (or the private sector), foreign investors may be more reluctant to make a major commitment to large scale infrastructure projects in developing countries.

We now explain the different options with reference to Table 3, which distinguishes between incidence of cost (columns) and source of finance (rows). In each cell the underlying economic consideration is given in brackets at the end of the brief description of the financing and payment position.

Table 3: Paying for versus financing of infrastructure spending

		Who pays
		Tax payer	User		Donor
			As per benefit	Cross-subsidisation
Who finances		Column 1	Column 2	Column 3	Column 4
Tax payers (cash)	a	Financed and paid for by tax payers (as if public consumption)	Financed by tax payers; users pay (benefit approach)	Financed by tax payers; certain users subsidise others (distributional)	Financed by tax payers; donors bear part of cost (distributional)
Lenders to government or government enterprises (loans or guarantees)	b	Loan financed by private investors; tax payers repay loans over time (inter-generational equity; allocative efficiency)	Financed by private investors; users repay loans (inter-generational equity; allocative efficiency)	Financed by private investors; users subsidise others (distributional)	Financed by private investors; sponsors pay (subsidise tax payers or users)
Private investors (equity, using internal or external funding)	c	Equity financed by private investors; tax payers pay to effect return on investment (public–private partnership) (equity; allocative efficiency)	Equity financed by private investors; users pay to effect return on investment (benefit approach; allocative efficiency)	Equity financed by private investors; users pay and subsidise others to effect return on investment (distributional)	Equity financed by private investors; donors pay (subsidise tax payers or users)
Development agencies (loans)	d	Loan finance by development agencies; tax payers pay to effect debt service	Loan finance by development agencies; users pay to effect debt service	Loan finance by development agencies; users subsidise others (distributional)	Financed and paid for by development agencies (equity)
Donors (grants)	e	Grant finance reduce cost to tax payers	Grant finance reduce cost to users	Grant finance reduce subsidisation by users to other users	Grants by sponsors (financed and paid for) (philanthropic; equity)

The cost of infrastructure is ultimately borne either by tax payers, users or donors (columns 1, 2, 3 and 4). The responsibility, if not obligation, of tax payers remains important to the present day, long after Adam Smith, in his Wealth of Nations (1776), noted that ‘the duty of erecting and maintaining certain public works and certain public institutions which it can never be for the interest of any individual, or small number of individuals, to erect and maintain’ falls to the state, ‘because the profit could never repay the expense to any individual or small number of individuals, though it may frequently do much more than repay it to a great society’ (Smith, 1776:IV.9.51).2

If the nature of the facility is such that the benefits can be attributed to identifiable users (i.e. low external benefits of costs), user charges are indicated. ‘Betterment taxes’ represent a variation on this theme. If low-income users are to be subsidised, certain users will pay more than the benefit received in using the infrastructure (column 3). When donors bear part of the cost (column 4), tax payers, users or a subgroup of users are the beneficiaries. Of course, as with all transfer payments, the ability to substitute non-subsidised goods or tax obligations for subsidised goods means that the intended beneficiaries may not be the eventual beneficiaries.3

A number of financing sources are available, as listed in the rows in Table 3. The infrastructure can be paid for in cash (row a), through the national, provincial or local budget (which basically represented the South African Government's pre-2009 position, as in cell 1a). Loan finance could be used by government or government business enterprises, the latter with or without government guarantees (row b). Debt financing will implicitly entail bulk lending for a composite set of projects, but ring-fenced project loans are also a possibility. In both cases, private individuals or companies are the financiers. The investment plans of public enterprises typically fall in this category. Equity investment is another source of finance (row c): private businesses buy a share (either as owners, co-owners or in partnership with government) in the activity, which needs to be properly structured for that purpose. Development agencies are another source of finance (row d). This normally entails cheaper loan finance, either because the borrower receives the benefit of cheaper rates that the bulk lender can negotiate in the financial market, or because deliberate subsidisation is provided. In the latter case, some or other sponsor bears a portion of the cost of the infrastructure (as in cell 4d). Finally, donors can be a source of finance (row e), in which case grants reduce the cost to whoever has to pay for the infrastructure.

Next we assess the merits and feasibility of the various options. We do not regard donor funds as a major source of finance for a planned long-term infrastructure expansion and therefore we focus only on the first four sources of finance in Table 3, namely tax payers (government), lenders, equity investors and development agencies.

From the perspective of economic efficiency we take the view that, where applicable, the cost of supplying infrastructure has to be borne by beneficiaries or proximate beneficiaries rather than tax payers in general. Where intra-generational equity is at stake, direct tax subsidies to the intended beneficiaries arguably allow for more accurate targeting (fewer opportunities for benefit shifting) than producer subsidies or cross-subsidisation between users. Inter-generational cost-sharing for equity purposes presents a strong case for long-term loan finance or even equity finance. We submit that unless there is clarity on these issues of cost incidence and financing sources in respect of each and every infrastructural project, there is no way of telling whether the macro budget strategies of the day present calculated deviations from a basic (‘default’) approach towards infrastructure financing or incoherent switching between different financing strategies that have little regard for the underlying efficiency considerations.

The above statement implies that the government should first decide whether or not an infrastructure project is a candidate for loan finance. By structuring this as a self-financing project, for example, the financing strategy would be isolated from the vicissitudes of macro fiscal borrowing strategies. In fact, the more users rather than tax payers are to bear the cost (as in column 2), and the more the activity is structured and financed as a self-financing project outside the national, provincial or municipal budget, the more additional fiscal space will be created in the budget, whatever the target for the budget balance. From an efficiency point of view, this benefit approach to financing the facility should be superior to a government investment programme financed from consolidated tax revenue and run administratively and possibly without having to meet strict financial performance criteria. Already government enterprises operate in this manner (line b), but with intelligent project design more infrastructural facilities could be offered in this way (see the list of key areas in Table 2). Given that the benefits of infrastructural projects at local and even provincial government level are often geographically confined, such projects may be good candidates. Should distributional considerations require a subsidy to ensure affordability, this could be either tax financed (as in cell 1a, although tax payers will then only bear part of the cost) or paid for by higher-income users (i.e. cross-subsidisation within the group of users, as in cell 3b).

It would seem that not all infrastructural projects lend themselves to payment and financing according to the benefit principle. Tanzi (2005) recognises that some types of infrastructure may justify financing through current revenue rather than loans. He argues that some infrastructure projects may be politically important, although the economic returns may not justify the investment if assessed purely in financial terms. Regional transport infrastructure, he argues, is a good example of projects where the immediate economic benefits may not be sufficient to justify the investment, although the positive impact on the political economy of regional integration may suffice. He suggests that if such projects (which include other ‘soft’ infrastructure, such as cultural monuments) are undertaken, they should be financed through current revenue rather than through loans, even though the benefits will mostly accrue to future generations. While he admits that this is a contradiction of the benefit principle, he argues that such projects (because of the exorbitant costs involved) may have serious macroeconomic implications, especially if foreign loans are used.

If the South African Government, therefore, decides to press ahead with infrastructure projects whose economic returns are insufficient to validate loan or equity financing, it may make more financial sense to finance them through current revenue, or at least the portion that the government will not be able to recover in the form of prospective future tax revenue. However, given the recent evidence of severe shortages in the South African economic infrastructure stock (Bogetic & Fedderke, 2006; Fourie, 2008), there should be a sufficient number of new projects yielding a positive net economic return (see also the Gautrain case study below).

As Table 3 suggests, a number of sources of loan financing may be available to governments. Given that South Africa is seen as a source of development finance rather than a destination, private capital will remain the primary source of loan and equity finance. In this regard local institutional investors might well be an important source of finance, provided such investment decision is taken without government intervention, except in so far as government needs to take responsibility for public-good characteristics or correct for clear market failures (Irving & Manroth, 2009).

South Africa's opportunity to mobilise private capital for infrastructure may in a number of respects be more favourable than for many other developing countries in general. Firstly, the country has relatively well-developed financial markets in which bonds and shares are well traded and in which institutional funds present an important source of finance. Bond financing is one of the most profitable and appropriate financing options. A notable current development that should be pursued in South Africa is the rapid expansion of international project bond markets, which provides developing countries with wider choices in bond financing (Kim, 2005:8). The South African municipal bond market has been in existence for many decades and now that the local government finance is steadily being placed on a better footing, the rating-based bond issues placed by Johannesburg and Cape Town (the latter forthcoming at the time of writing) open up new possibilities.4

Secondly, the government is a potential borrower of good standing, domestically and internationally, and the sovereign country rating benefits parastatals in the form of lower international debt cost, as and when foreign loans make economic sense.

Thirdly, a programme of gradual and phased privatisation of public enterprises has for the past 20 years been running in one form or another, which has entailed private equity investment and the competitive restructuring of the relevant economic sectors. Some unresolved issues remain, however, such as the deregulation of Telkom and securing the financial viability of public transport businesses (rail and air travel).

Fourthly, public–private partnerships have developed steadily over the past 10 years as a method of partnering public and private institutions in delivering services in which the public and private goods components are difficult to untangle when it comes to who operates, owns and finances them, such as water supply and toll roads. Hence, as shown in Table 1, the private finance that is envisaged over the period from 2008/09 onwards. Despite all this, the fact that the investment surge in South African infrastructure has occurred at a time when internationally private sector interest in this kind of investment appears to have waned may well mean that the potential for private financing could be less than what other developing countries have experienced during the past 10–15 years.

Nonetheless, there are more opportunities for private investment in infrastructural development than the government's tax financing choice offers. The Gautrain Rapid Rail project may offer some guidance.

5. The Gautrain Rapid Rail project as case study

The R25.5 billion Gautrain Rapid Rail Link is an 80 km mass-transit system in Gauteng that will ultimately link Johannesburg, Pretoria and the OR Tambo International Airport. The project has been structured as a public–private partnership, with Bombardier Transportation (25 per cent), Bouygues Travaux Publics (25 per cent), Murray & Roberts (25 per cent), and SPG (25 per cent) as shareholders in the Bombela Concession Company.

According to Gauteng's 2008/09 budget, the public sector has committed R22.5 billion to the project, with the contribution of the private sector set at an additional R3 billion. The private contribution is financed through loans from a joint agreement between Bombela Concession Company and First National Bank and Standard Bank. Furthermore, SPG (Strategic Partners Group), a broad-based black economic empowerment company, is funded by the Industrial Development Corporation and the Development Bank of Southern Africa.

Private owners only invest if they estimate they will obtain an appropriate return. Such revenue originates from user charges, with no or little projected cross-subsidisation. The above-mentioned private consortium will use external funds (i.e. loans from the private sector) to finance construction and these loans will be repaid from the projected income after 2011. In Table 3, this financing type is located in cell 2c.

Should the private sector play a more important role? While the trend over the last decades has shifted to public–private partnerships, the international evidence suggests that government, both at a regional, national and international (for example, the European Union) level, is still the primary source of funds for rapid rail developments (Nijkamp & Rienstra, 1995; Debande, 2002). In particular, Roll & Verbeke (1998) argue that the socioeconomic benefits of high-speed rail networks are usually underestimated, implying an important role for government involvement. Given this argument, the private sector involvement in the Gautrain seems to be in line with the international experience.

Between 2006/07 and 2011/12 the national and provincial governments plan to contribute between R22.5 billion and R24.3 billion to the construction of the Gautrain. While exact numbers differ across sources (between the budgets of Gauteng and National Treasury, and from the financial forecasts of the Gautrain Rapid Rail Link), it is expected that the provincial government will contribute R11.6 billion of the total cost (Gauteng Treasury, 2008), with the national government responsible for the remainder of the public contribution of R12.7 billion (National Treasury, 2009a:818).

After 2011/12 the project will move into the operational phase and costs will be covered by operating revenue (which will position this on cell 2a). This suggests that the initial R22.5 billion will not be financed through future user charge repayments, with the incidence of costs falling fully on existing and future tax payers, in Gauteng but also elsewhere in South Africa. The construction costs (R11.6 billion) are bundled together with existing expenditure categories in the Gauteng budget, which suggests that no special loan facility was created for the Gautrain at provincial level. This implies that current tax payers are footing the bill when the benefits will only accrue to the taxpaying public after 2011. Returning to Table 3, 88 per cent of the cost is borne by tax payers, locating this part of the financing in cells 1a and 1b.

Although the finance of and ultimate payment for these services from existing government revenues is not unique in the international experience, it is out of tune with new trends in financing arrangements for transport projects. In the US, the growing costs of maintaining and repairing the existing infrastructure network have shifted the fiscal responsibility away from national and state governments to local governments through earmarked taxes imposed outside the traditional budget process (Goldman & Wachs, 2003).

In the case of the Gautrain, a loan facility (combined with an extra-budgetary earmarked tax, for example) would have opened up considerable fiscal space in either the provincial or the national budget. As an example, Gautrain construction costs in the Gauteng budget of 2008/09 amount to R4.7 billion, which is considerably higher than the revenue of the Gauteng province of R2.8 billion (excluding the transfers from National Government). Had a loan finance facility been available, the extra fiscal space in the budget could have financed more urgent current expenditures, or maybe even meant money back in the pockets of existing tax payers.

To be sure, an argument can be made for using current tax revenues. Following Tanzi (2005), one may argue that the economic benefits derived from the Gautrain may not be sufficiently significant as to ensure long-run, sustainable returns, as in the case of regional rail links. Yet there is ample evidence that the Gautrain will create significant positive externalities for future generations (Bohlweki Environmental, 2002), another argument for using loan finance. Unlike regional rail networks (and monuments, another Tanzi example, with little immediate economic benefit), the Gautrain would have an immediate impact with supposedly large returns (for both the private and public purse). One relevant South African example for financing infrastructure through current revenue would be the stadiums constructed for the 2010 Soccer World Cup. These buildings may not yield immediate economic benefits that government can use to repay the costs of borrowing.

Two conclusions can be drawn from this analysis. Firstly, the international literature suggests a role for private sector participation and government involvement. Although the nature and size of private sector participation may vary across projects, the Gautrain corresponds with international practice. Secondly, where immediate economic returns can be realised, such projects may best be structured and managed as extra-budgetary activities, financed in a ring-fenced manner through loans and paid for by users or through earmarked taxes. In this regard there is thus little economic justification for using current revenue rather than loan financing to foot the construction bill for the Gautrain.

6. Conclusions

The surge in infrastructure expenditure by the South African Government, together with the growing ‘priceability’ of infrastructure, necessitates a re-evaluation of the standard or benchmark approach to the financing of and payment for such expenditures. This paper has reviewed the government's current stance from an efficiency and equity point of view, suggesting that there is an excessive focus on financing through current revenue (taxes), without a proper distinction between the financing strategy of the national budget for macroeconomic considerations and the financing strategy of infrastructural projects. The paper also highlights the importance of distinguishing between loan financing by national government and by public enterprises. It seems clear from the South African budget documents that public enterprises will continue to depend on loan financing, while the national government, in better economic times, will revert to financing through current revenues. This also has implications for the case study of the Gautrain, which is financed not by public enterprise but by the national government. Public enterprise in effect does what this paper suggests should be done – by removing infrastructure expenditure from the general budget to a ‘special project’ to be financed by loans and paid for by user charges. Too many infrastructure projects (including the Gautrain) are, however, positioned in the budgets of national government, which contradicts the standard principles of public finance. The wider application of the benefit principle of allocative efficiency should lead to higher levels of loan financing and consequently increase the fiscal scope at national level, all things being equal.

Acknowledgements

The authors would like to thank two anonymous referees and participants at the Economic Society of South Africa conference in Johannesburg, 2007 and at Stellenbosch University Seminar series for valuable comments and suggestions. Any remaining errors are the authors' alone.

Notes

¹This includes loans and guarantees by national government.

²Notation refers to Book IV, chapter 9, paragraph 51.

³See in this regard the standard fiscal literature on the incidence of intergovernmental grants (e.g. Rosen, 2005:532–6).

⁴The Develpoment Bank of South Africa (DBSA) is another important domestic source of infrastructural finance (grantor, lender, investor and underwriter of guarantees), especially for local governments (DBSA, 2007).