This blog post is based on research done for Sustainable Energy Africa and the National Planning Commission. The full report, including references, can be downloaded here.

Investment in infrastructure in South Africa peaked in 1981, and then languished for two decades. From 2000 onwards, gross fixed capital formation increased dramatically, doubling in a period of 8 years. In the wake of the global financial crisis of 2008 investment levels dropped somewhat, but as a result of the FIFA 2010 World Cup and the related infrastructure projects, remained at a high level. South Africa’s infrastructure investment drive has been spearheaded by public corporations. Whereas the spending patterns of general government and private business have remained relatively stable since 2000, investment by public corporations has risen sharply, and this largely explains the recent rise in overall capital formation.

Figure 1 Trends in gross fixed capital formation, 2000-2011 (seasonally adjusted 2005 prices)

Source: DNA Economics based on SARB (2012)

Even after a decade of accelerating infrastructure investment, South Africa still faces an infrastructure backlog that is now acting as a major constraint on economic growth. It is no wonder then that a substantial expansion in infrastructure investment was one of the central themes of the 2012 Budget. Failing to consider future climate change policies when investing in infrastructure, however, may lead to suboptimal long-term outcomes.

Infrastructure investments have very long useful lives, and investments made today are likely to continue providing services for decades to come. Even infrastructure in place for relatively short periods of time can still have long-term impacts. The design life-span of road infrastructure in South Africa, for instance, is relatively short in infrastructure terms at 20 years (compared to the design lifespan for a new coal-fired power station at around 50 years, for instance). But the positioning of new roads has longer-term impacts through their influence on spatial development patterns. Moreover, the expected useful life of infrastructure is not always fixed, and the actual operational lives of assets can be longer than anticipated. Locomotives, for example, have an average life span of 16 years and wagons 20-25 years internationally. Transnet’s locomotives are on average 30 years old and its wagons 35 years. Capital is also typically bulky and supplied in large discrete components. Just before investments are undertaken, demand typically exceeds capacity, while the period directly after an infrastructure investment is typically characterised by excess supply. Consequently the expectation of future demand for infrastructure is a more important determinant of infrastructure investment than current demand. If the future demand for a type of infrastructure is overestimated, the value provided by investments will be reduced and the capital invested could have been deployed more efficiently elsewhere in the economy. This outcome is referred to as carbon lock-in.

Carbon lock-in can be viewed from two main perspectives. Emissions lock-in refers to the emissions implications of current and planned investments being largely irreversible for long periods under normal circumstances. While opportunities for retrofitting exist in some instances, the large technological and operational changes needed to significantly reduce greenhouse gas (GHG) emissions are usually only possible when capital stock is installed or replaced. Should conditions require it, however, the future emissions from current infrastructure can be avoided – but only at a high cost. Tough policy interventions can force infrastructure to be retired before the end of its useful life or to be only partially utilised. Infrastructure investments that do not consider the possibility of such future policy interventions run the risk of reduced returns. The useful life of assets may turn out to be much shorter than envisaged due to regulatory fiat, or the assets may simply become uneconomical due to the implications of policy (i.e. the financial cost of having to pay for the right to emit GHGs). An investor or public agency may thus end up locked into owning an asset (or worse, a portfolio of assets) that is worth considerably less than envisaged. This constitutes asset lock-in, and assets of which the financial value has been significantly reduced as a result of unforeseen policy, regulatory or legislative changes are referred to as stranded assets.[1]

A combination of economies of scale achieved in current infrastructure, and vested interests by stakeholders in developing, operating, financing and supplying the infrastructure, create strong incentives for maintaining the status quo. Users of existing infrastructure also often locate and organise themselves based on expectations that the current configuration will continue indefinitely. Additional forces that create inertia range from technical standards to the greater availability of financing for proven technologies, and even the existence of networks of private associations and educational institutions geared towards advancing existing technologies. When this combination of incentives and expectations becomes so entrenched that it delays the adoption of more efficient technologies, it is termed institutional lock-in. Institutional lock-in thus refers to an outcome where past investment decisions make it difficult to invest in areas or technologies that would increase overall economic efficiency.

Investors are primarily concerned with avoiding asset lock-in and stranded assets when considering carbon lock-in. This risk of stranded assets increases as the useful lives of assets, payback periods and lead times of projects increase, the higher up the cost curve the technology employed sits, the more fluid the policy environment is, and the larger the GHG emissions associated with the project are. These factors are highly project-specific. To reduce the risk of asset lock-in, project assessment frameworks should take a long-term perspective and explicitly focus on asset lock-in. Ideally a number of different climate change policy scenarios should be evaluated. A necessary (but not sufficient) component of such assessments is understanding the GHG emissions associated with the use of infrastructure over its lifetime. The general cost-effectiveness of the technology employed is also important as the more efficient the technology is compared to its competitors, the better able a project will be to cope with the shock of unanticipated costs. Being more efficient than competitors also reduces the probability of infrastructure becoming stranded as a result of more direct regulatory interventions, like production or efficiency standards. It is also paramount that infrastructure be developed in a way that is compatible with current climate change policies and strategies. Rather than viewing the National GHG Emissions Trajectory Range as an absolute cap on infrastructure emissions (since it is subject to change), however, it should be seen as signalling the expected future extent of climate change policy interventions.[2] It is worth noting that a fluid climate change policy framework increases the risk of asset lock-in by making future policies less predictable.

Policymakers are primarily concerned institutional lock-in and its wider systematic impacts. Focussing on individual infrastructure investments (even when they are lowest-cost and efficiency enhancing) can mean that sight is lost of the overarching emissions reduction path of an economy. A focus on marginal efficiency improvements may mean that more significant trigger points to switch to new technologies or processes are missed. A focus on the impact of investment decisions over an entire transition period is thus desirable. In the absence of this long-term perspective, the risk increases that infrastructure plans may not be consistent with the long-term mitigation objectives of an economy – potentially leading to stranded assets and/or missed mitigation targets. The compatibility of infrastructure investments with the economy’s overall emissions reduction trajectory (not just in terms of emissions outcomes, but also whether certain types of investments will be possible in future, given current technology choices) is a key factor in determining whether investments avoid institutional lock-in. The difference between the useful life and the payback period of infrastructure is also an important consideration in avoiding institutional lock-in. Given the need for coordination and adherence to existing planning frameworks to reduce institutional lock-in, it is worrying that local public corporations largely invest based on internal imperatives, rather than a broader set of strategic, integrated infrastructure development goals. Even with greater coordination, it is questionable whether the Peak, Plateau and Decline (PPD) trajectory currently provides sufficient detail to guide infrastructure investments that avoid institutional lock-in.

Key to answering the question of whether infrastructure investments are locking in a high emissions path is good quality data sources on actual infrastructure spend. However, current sources of infrastructure investment data are not ideal for investigating the impact of infrastructure investment on GHG emissions patterns. They are too aggregated, and more granularity in the reporting of infrastructure spending (ideally splitting out individual technologies or sub-sectors) is required. Furthermore, the carbon intensity of infrastructure investments, used on its own, is a highly imperfect indicator of the risk of carbon lock-in. More important is understanding how individual infrastructure investments impact on the future efficiency of South Africa’s total infrastructure network under different climate change policy scenarios. This necessitates a review of not only individual project assessment criteria, but a detailed comparison of current and planned infrastructure projects with the PPD trajectory assumptions.

Given the complexities of determining whether or not infrastructure investments are vulnerable to lock-in effects and the lack of detailed data on investment in South Africa, it is not possible to provide a definitive conclusion on the current risk to South Africa’s infrastructure portfolio. That said, it would appear that most current major infrastructure projects come with relatively high GHG emissions, while for planned projects, there is a sharp shift towards relatively low carbon projects.

Figure 2 Current committed and planned major infrastructure projects

Source: DNA Economics based on National Treasury (2012)

Despite this change, the relatively short-term focus used to evaluate infrastructure projects (by placing undue emphasis on capital costs at the expense of other life cycle costs), and the lack of attention paid to the possibility of lock-in in project assessments (although it is about to start, the DBSA does not yet evaluate GHG emissions as part of its standard project assessments – and neither does the Department of Public Works), raises the possibility that some infrastructure asset lock-in will occur. Similarly, the lack of coordination with respect to public infrastructure investment, and the lack of a sufficiently detailed local emissions trajectory to guide infrastructure spending, seems to indicate that institutional lock-in should be a serious concern in South Africa.