IPS Fact Sheet on the World Bank's 
Carbon Back-Casting Study

This fact sheet is in response to a fact sheet circulated by the World Bank to the media, "World Bank Response to Changing the Earth's Climate for Business," which uses data derived from the Carbon Back Casting Study (CBC study), a study which the World Bank commissioned IIEC, Hagler Bailly, and the Stockholm Environment Institute to carry out, which was released on June 13, 1997.

Carbon Back-Casting Error #1: Not as comprehensive a review as IPS study

The objective of the CBC study was to "analyze the effect of placing a shadow price on carbon emissions" and to analyze "whether a shadow price would encourage investment in low-carbon alternatives." The objective of CBC study was distinct from the IPS study in that it did not attempt to calculate greenhouse gas emissions associated with World Bank projects in a comprehensive manner. The CBC study did not examine every project funded by the Bank between 1992 and 1997, but only took a "representative sample" of about 50 projects out of 154 loans between 1990 and 1996, and extrapolated from this limited pool.

The IPS study, on the other hand, reviewed all World Bank-financed fossil fuel projects listed in the Bank's own annual reports from 1992-1997. It is clear from this study that there is no such thing as "a random sample" of that set; some projects (such as investments made by the World Bank in oil fields in Russia, Kazakhstan and Azerbaijan, which together account for more than 50 percent of our grand total carbon emissions) clearly outweigh others in their significance; leaving out just one or two of such top emissions-producing projects will create an inadmissable statistical error.

CBC Study Clarification: Emissions figures are close to one-half of IPS figures

The Bank claims we use a "dishonest statistical trick" to arrive at our figures by comparing lifetime emissions of World Bank-financed projects with annual fossil fuel emissions, although we made clear we were making the comparison to place this figure in context. Ironically, the CBC study uses a number of "tricks" of their own (detailed below) which result in an underestimation of carbon emissions; nevertheless, their own figures are roughly one-half of ours--4.6 gigatons of carbon.

Table 3-2, page 3-3, estimates total project emissions for the main study sample--which is roughly one-third (50/154) of all projects funded by the World Bank--to be 1.413 billion tons of carbon. To arrive at the true total for these projects, we multiply 1.413 by 154/50 to arrive at the full carbon estimate for the portfolio, which is 4.6 billion tons of carbon equivalent. It seems someone misrepresented this for the executive summary, conveniently forgetting to mention that the annual figures they admit to are only 50/154th of the total. The IPS study estimate for total fossil fuel emissions associated with World Bank projects between 1992 and 1997 was 9.5 gigatons; the CBC Study figures were thus about 48% of our figure.

CBC Study error #2: Emissions data flawed

IPS provided a transparent methodology for our own calculations, based on state-of-the-art calculations compiled by the U.S. Department of Energy, Environment Protection Agency and others. We combined these calculations with objective data which is in the public domain on projects the Bank funds--namely, proven fossil fuel reserves for projects they admit they are assisting with financing.

The World Bank challenges our calculations, yet offers no transparent methodology for arriving at its own figures. Furthermore, the authors of the CBC study admit that their data was neither complete nor primary. They write:

We relied on several data sources to estimate emissions. The SARs (Staff Appraisal Reports) were the primary source, but they generally did not contain estimates of carbon emissions (emphasis added). However, the SAR data often included information such as fuel consumption forecasts or electricity sales and fuel mixtures that we were able to use to derive emissions. Additional sources of data for emissions calculations included other documents in the project file such as feasibility studies or demand forecasts, information provided by Bank task managers, and standard conversion factors in engineering data bases.

This is an extraordinarily vague, unscientific, and apparently ad hoc manner in which to gather hard data for something as scientifically complicated as carbon emissions.

CBC Study error #3: Report only looks at 20-25 years of projects, yet fossil fuel reserves the Bank has invested in commit the atmosphere to greenhouse gases beyond this time span.

There is no way of knowing with scientific certainty what amount of fossil fuel reserves the World Bank helps finance will be burned for domestic use, how much will be exported to repay World Bank loans, and how much will stay in the ground. The only figure one can rely on with anything close to scientific certainty is the quantity of proven reserves in the mines and oil and gas fields the World Bank is helping to finance, and what those reserves will translate into in terms of greenhouse gas emissions once burned. That is what our report calculated, while making sure not to include emissions from power projects that were already calculated in oil, gas, or coal fields.

However, the CBC study apparently ignores emissions from fossil fuel fields financed by the World Bank after an arbitrary cut-off point of 25 years; as a result, this study is vastly--and randomly-- underestimating greenhouse gases from projects it finances.

CBC study error #4: Emissions calculations are incomplete

On page 2-8 of the Carbon Back-casting Study, the authors write:

During the scoping study, upstream emissions such as methane emissions from fuel transport were quantified and found to be small relative to fuel consumption emissions (i.e., less than 20% of consumption emission). Consequently, we decided against quantifying them for thermal generation activities in the main study.

To admit that a systematic error margin in their measurements amounting to up to 20% is "negligible" is equivalent to saying that the authors think that the "true" error margin for their numbers is much higher than that. This means that the author's total calculations could underestimate the total by at least 20%, which, using the Bank's own emission numbers (at least 4.6 gigatons over 20-25 years), comes to at least 880 million tons of carbon per year, or an additional 1.76 gigatons over 20 years.

An additional underestimation occurs on page 2-9 of the report, where the authors write that they estimated "both fugitive emissions and combustion emissions for domestically consumed fuel only." Yet World Bank fossil fuel projects often are turned into tradable exports (especially in the case of oil and gas) in order to earn hard currency to repay these same loans. Thus, to not calculate these emissions simply because they did not get consumed domestically is to vastly underestimate the total by an unknown factor.

Back to the top