Helen Jackson

Analytic and Research Support

Interpretation of data: Sensitivity analysis of renewable technology costs

Rationale: Quantitative estimates can be subject to large uncertainties. There is often a need to understand how they change under a range of different scenarios. A standard approach to uncertainty is to present estimates as mid-points within a range. However, this does not allow decision-makers to explore and fully understand how material outcomes will be affected if an estimate is at the high or low end of its range.

Computation and data visualisation can help. Programming makes it straightforward to calculate and compare thousands of different estimates, randomly changing variables through their ranges. This can give decision-makers a better understanding of how the world looks away from the midpoint scenario.


Overview: Figure 1 presents the levelised cost of electricity (LCE) for various renewable technologies in the UK, i.e. the price at which electricity would have to be sold in order to break even. The LCE is split into capital (green) and operational (orange) components. The different technologies are ranked by cost, with 1 indicating the lowest and 12 indicating the highest. The green shading shows the range of prices at which electricity is currently sold to domestic and industrial customers. Click Play Fig 1 to see the effect on the ranking of changing the discount rate (with a technology-specific hurdle rate at the end).

Figure 2 presents the results of a more sophisticated sensitivity analysis changing all of the variables used in the cost estimates. 100,000 different sets of LCE estimates were computed, by randomly selecting component variables from within their range of estimates. The technologies were ranked by cost for each set. Click Play Fig 2 to see how frequently each technology is placed under each rank.

Interpretation: According to this data, in a UK context, solar PV and tidal are consistently the most expensive technologies; biomass and larger hydro are consistently the cheapest; while the others are generally medium-cost and vary more in their ranking. Larger installations are, unsurprisingly, cheaper than smaller ones due to economies of scale. The exception to this is offshore wind, but this is due to particular historical factors during the period of data collection.

At a low discount rate, the LCE of most technologies is comparable with or lower than electricity prices. As the discount rate increases, only biomass, larger hydro and large onshore wind stay within the price-competitive range. At high discount rates, the present value of future revenue streams is low, therefore the technology needs to recoup any capital costs quickly through higher electricity prices. The actual appropriate discount rate is likely to vary by technology, due to differing risks and minimal acceptable rates of return for investors (hurdle rate).

Caveats:

  1. The frequencies shown in Figure 2 should not be interpreted as probabilities, as there are no probability distributions associated with any of the variables.
  2. The economics of solar energy depends on irradiance, which varies by location; what holds for the UK need not hold for other countries.

About the data: Electricity prices are taken from the Digest of UK Energy Statistics.

Data on renewable technology capital and operating cost ranges, load factors, typical project lifetimes and hurdle rates were taken from a 2011 study by the engineering firm Arup for the UK Department of Energy and Climate Change: Review of the generation costs and deployment potential of renewable electricity technologies in the UK. Discount rates within the sensitivity analyses are taken from the range 5-20%. The lower bound represents a commonly-used social discount rate, while the upper bound represents a high hurdle rate. Hurdle rates used in the Arup report were in the range of 7.5% (for hydro-electric) to 22.7% (for geothermal).