Ideas to Action:

Independent research for global prosperity

New from CGD

World Bank Power Projects: Crossroads on Renewable Energy (Slide Show Commentary)

May 10, 2008

The attached slides illustrate my forthcoming paper with Kevin Ummel on a dynamic strategy for developing solar thermal power as a cost-competitive alternative to coal-fired power. We have presented versions of these slides at a number of conferences and policy workshops during the past two months. Here is a summary of the forthcoming paper and its conclusions.

The World Bank and other multilateral financial institutions have begun touting low-carbon investment strategies, but they have not adopted any form of carbon accounting. A low-carbon strategy without carbon accounting is not really a strategy, since it can neither establish clear guidelines for project choice nor assess progress toward its objective. In a recent paper (Wheeler, 2008), I illustrate the potential consequences of this shortfall by assessing proposed World Bank support for a huge coal-fired power plant at Mmamabula, Botswana. Despite the project’s enormous projected CO2 emissions, advocates of Bank support label it “climate-friendly” because, they assert, the Bank’s support would enable the use of efficient (supercritical) combustion technology that an unconstrained private developer would shun for a putatively cheaper, and certainly dirtier (subcritical), technology. In the paper, I reframe the choice by introducing carbon accounting charges and a low-carbon option: concentrating solar power (CSP), also known as solar thermal power. My analysis indicates that CSP is preferable to the proposed coal-fired project when its projected emissions are subjected to carbon accounting charges that are conservative by current standards.1

The obvious criticism of the paper is that low-carbon options remain more costly than coal-fired options, so clients will reject World Bank loans for renewable power unless the Bank depletes its capital base by financing the extra cost. However, this criticism has become obsolete for three reasons. First (as always), the World Bank lends money at below-market rates, and this subsidy would be sufficient to close a moderate cost gap between coal-fired and renewable power. Second, low-carbon projects are eligible for carbon credits under the Clean Development Mechanism (CDM). At current prices (around $US 15/ton for CO2 emissions), potential CDM credits are enormous for large renewable power projects that will avert millions of tons of CO2 emissions. Third, the U.S., UK, and Japan have recently established a large Clean Technology Fund (CTF), initially administered by the World Bank, that is intended to target the cost gap between dirty and clean technologies.

With the addition of CDM and CTF funds, project choice based on carbon accounting has now become financially feasible as well as environmentally desirable.2 But even carbon accounting at the project level cannot resolve the real strategic question in this context. The needs of power-sector investment in poor countries are enormous, and the aggregate cost gap between dirty and clean power will inevitably dwarf the resources in a clean technology fund. For this reason, the CTF only makes sense if it provides strategic leverage by promoting industry learning and production cost reduction in the renewable power sector. If the learning curve is steep enough and the initial cost gap narrow enough, then CTF financing can drive renewable power to cost-competitiveness with dirty power within a few years and catalyze a rapid transition to renewable power in the private sector.

In a forthcoming paper, Kevin Ummel and I extend the Mmamabula analysis to address this strategic question for the choice between supercritical coal-fired power and the least-cost solar thermal technology (compact linear Fresnel reflector [CLFR]). We begin with my static cost comparison, which is based on data from Mills and Morgan (2007) and World Bank/ESMAP (2007). We conclude that an accounting charge of about $35 per ton of CO2 emitted is sufficient to warrant switching from supercritical coal combustion to CLFR. We then extend the analysis in several ways. First, we account for input-cost inflation in coal-fired power since 2005, when the cost estimates for the benchmark ESMAP study were obtained. Second, we draw on current solar industry information to update the input cost estimates for CLFR. Third, we drop the assumption that the Clean Technology Fund automatically covers the cost gap between the two technologies. The Mmamabula paper relies on a simple comparison of levelized unit production costs (in cents/kWh) for power production.3 In this paper, we develop full cost estimates for large-scale (500 MW) power facilities and compute the actual cost gap that the CTF would have to cover.

Fourth--and most critically from a strategic perspective--we specify a multi-year, CTF-financed program whose objective is to drive the cost of CLFR solar thermal power to parity with the cost of coal-fired power. In this program, the CTF promotes private CLFR projects by paying their incremental cost, which falls along the learning curve as solar power production grows (supercritical coal is already a mature technology). The program terminates when the unit cost of CLFR reaches parity with the cost of supercritical coal. We assess the solar thermal investment program under a variety of assumptions about initial costs, learning curves, discount rates, accounting charges for CO2, emissions and post-program diffusion of CLFR technology in the private sector.

We find that the solar thermal investment program is cost-competitive with coal-fired power at extremely low carbon accounting charges—so low, in fact, that there is no reasonable case for delaying the switch to solar thermal power (and, by implication, other renewable technologies with similar learning curves and cost differentials). Using the best available evidence on costs and solar learning curves, we conclude that a focused investment program can drive solar thermal power to cost parity with coal-fired power in five to ten years, for a total cost of four to eight billion dollars. This is well within the current resources available to the Clean Technology Fund. If the World Bank and other aid organizations want to match their “transformational” rhetoric with action, our results provide a clear roadmap for rapid progress.


1 Systematic cost analysis also reveals that higher coal prices have made supercritical combustion technology more profitable than subcritical technology. This eliminates the rationale for subsidized World Bank credit in any case, since an unconstrained private developer will now prefer supercritical technology.

2 Since 1991, the Global Environment Facility (GEF) has also financed the incremental cost of clean power projects implemented by the World Bank and other multilateral organizations. The GEF has helped finance many pilot power projects, but their scale has been insufficient to produce baseload power at competitive costs.

3 Levelized cost is the constant-dollar electricity price required to recover capital costs, O&M and fuel costs over the life of a power plant.


ESMAP (Energy Sector Management Assistance Program, World Bank). 2007.
Technical and Economic Assessment of Off-grid, Mini-grid and Grid Electrification Technologies, ESMAP Technical Paper 121/07, December. Available online at

Mills, David and Rob Morgan. 2007. “Solar Thermal Power as the Plausible Basis of Grid Supply,” Presented to the International Solar Energy Society World Congress. Beijing, China. Sept. 19. Available online at

Wheeler, David. 2008. “Crossroads at Mmamabula: Will the World Bank Choose the Clean Energy Path?” CGD Working Paper No. 140. Available online at