Earlier this month, Greenpeace released a report slamming McKinsey’s work on REDD – in particular the McKinsey cost curve. On 14 April 2011, David Ritter, a Biodiversity Campaigner at Greenpeace UK, gave a presentation about McKinsey’s role in promoting deforestation (pdf file 85.4 KB) at the Civil Society Policy Forum of the Spring Meeting of the World Bank in Washington DC. The Bank’s reaction was fascinating.
One of the most extraordinary aspects of McKinsey’s cost curves is that no one outside McKinsey appears to know what data or assumptions lie behind the apparently precise graphs and figures that company produces. One of Greenpeace’s recommendations to McKinsey is that the company must:
Immediately publish all the data, assumptions and analysis underlying the international and national versions of its cost curve and include such disclosures in all future publications.
McKinsey has not contacted Greenpeace since the report was published. But Benoît Bosquet, Coordinator of the Forest Carbon Partnership Facility (FCPF) at the World Bank, who was in the audience listening to Ritter’s presentation, confirmed that McKinsey’s secrecy is problematic. Here’s how Ritter described Bosquet’s response:
At the end of the presentations, Mr Bosquet admitted that the concerns over McKinsey’s secrecy about their method were widely shared, noting that ‘the blackbox is a problem for everybody’. Mr Bosquet also emphasised that forest plans should be informed by good economics – presumably in contrast to the false assumptions and mathematical errors that characterise the ones McKinsey have a hand in.
The World Bank Institute is currently carrying out a series of workshops titled “Estimating the Opportunity Costs and Implementation Costs of REDD+ for the National Planning Process.” The third meeting will take place in Colombia from 16-20 May 2011. All of which raises an interesting question: Will the World Bank be using McKinsey-type cost curves to estimate the costs of REDD?
The World Bank has produced a 262-page training manual to accompany its workshops: “Estimating the Opportunity Costs of REDD+ A training manual” (pdf file 8.8 MB).
In a section titled “What is an abatement cost curve?” the report looks at a “supposed example of a national abatement cost curve, for Indonesia,” which was produced by McKinsey in 2009. Having noted that “this ‘abatement cost curve’ only considers direct, on-site opportunity costs,” the report comments as follows:
The fact that such a widely shared and well-publicized analysis is not actually of REDD+ abatement costs highlights the importance of reviewing methodological assumptions.
But having critiqued McKinsey’s cost curve for Indonesia, the Bank decides to use it anyway:
Despite actual abatement costs being higher, arising from implementation and transactions costs, the graph is useful for illustrative purposes.
Two paragraphs follow explaining what the cost curve shows and the cost curve itself is illustrated.
58. Reduction options associated with REDD+ are highlighted by red boxes. Their relative contribution is measured by the width of the respective bars. For example, abatement of forest conversion to smallholder agriculture would reduce emissions by approximately 250 MtCO2e per year, whereas avoiding timber extraction would reduce about 90 Mt CO2e per year. Reforestation could reduce emissions by approximately 100 MtCO2e per year (Dewan Nasional Perubahan Iklim and McKinsey & Co., 2009).
59. The differences in opportunity costs can be substantial. The vertical height of each bar represents the cost of each option. While reducing forest conversions to low productivity slash-and-burn agriculture is estimated to cost less than €2 per tCO2e, the opportunity cost of reforestation is approximately €10 per tCO2e and reduced forest conversion to intensive agricultural production can cost over €20 per tCO2e. Such cost difference affect feasibility of abatement options within national REDD+ programs.
Note the precision of the figures used. And note that the Bank is using these figures fully aware that they are not accurate. Even worse, the additional costs (that the Bank acknowledges are important) will not be the same for each of the options being considered. The figures are thus inaccurate and not in proportion to each other. The graph is only useful to demonstrate how not to estimate the costs of REDD.
The Bank’s training manual “focuses on estimating direct, on-site opportunity costs,” excluding, “social-cultural costs” (which could include, for example, loss of livelihood and knowledge as indigenous peoples and local communities are forced to stop swidden agriculture) and indirect, off-site costs (increases in prices of food, for example, as less land is available for agriculture and increases in prices of timber as less forest is logged). Having acknowledged that these could “represent significant opportunity costs,” the Bank’s training manual ignores them:
For the sake of brevity, the term opportunity cost will refer to direct, on-site opportunity costs throughout this manual.
In Chapter 7 of the training manual, the Bank explains how to “Generate an opportunity cost curve of REDD.” The first (and only) example in the chapter is McKinsey’s cost curve for Indonesia.