Economists usually think of labelling as a good thing – an increase in the information set. But what if the science behind the labelling is iffy? What if the organisations doing the labelling are responding primarily to incentives stemming from developed country markets? When it comes to the recent trend towards “carbon footprinting” everything, both of these issues arise.

There is a risk that carbon accounting and labelling instruments will not properly represent the complexity of production systems in developing countries and so adversely affect their export opportunities without a bona fide environmental justification.

Carbon labelling

Carbon accounting and labelling of products are new instruments that present information on carbon emissions in an attempt to identify major sources of emissions in supply chains. Information on emissions can allow producers, government, retailers, and consumers to take actions that will reduce emissions.

While the designers of these schemes are responding to public policy and corporate agendas to create new ways of responding to the challenge of climate change, they have to rely on very rudimentary knowledge about actual emission patterns of the varied production systems that occur around the globe.

This is because the underlying scientific understanding of greenhouse gas emissions is only partially developed. Knowledge of greenhouse gas emissions for production and processing activities is particularly low in developing countries and especially acute for agricultural products – often the principal export of these countries.

How to make labelling fairer and more effective

The key issues that need to be addressed if carbon accounting and labelling is to be an effective and fair mechanism for reducing greenhouse gas emissions are:

  • Land use change. To increase agricultural output, developing countries clear natural land for conversion to crop land and pasture. The conversion of forest to grassland and cropland results in the loss of carbon from the forest trees, and also from the soil. Similarly, the conversion of grassland to cropland results in a loss of carbon from the soil. The amount of carbon lost during any land use change depends upon the exact nature of the forest or grassland that is being converted. Typically the emissions from land use change are among the largest sources of emissions in the carbon footprint of crops produced in developing countries and so it is important that these emissions are calculated correctly.

Land use changes can be difficult in developing countries where relevant data on current and historical land uses are scarce or absent. Not only are there technical issues surrounding the calculation of emissions from land use change, but there is also a fairness issue. Developed countries typically do not need to include this source of emissions because they cleared their forests to create agricultural land decades or centuries ago. For example, the carbon accounting methodology developed by British Standards, PAS 2050, the most comprehensive methodology available today, requires that only emissions from land use change that occurred after 1990 be included in the carbon footprint of a product.

  • Carbon sequestration by tree crops. Many tropical developing countries export goods derived from tree crops, such as coffee, cocoa, tea, fruit and nuts. While trees themselves sequester carbon, the forest soils tend to sequester far more carbon than the above ground biomass. Currently, few carbon footprinting methodologies recognise the positive contribution made by carbon stored in trees used to produce food, or that sequestered in the soil. Thus many developing countries have to declare emissions from land use change, but cannot claim benefit from the management of tree crops.
  • There is a lack of necessary data and information for developing countries compared to developed economies. As a result, carbon accounting and footprinting analysts are required to use very imprecise and uncertain datasets that relate to very large geographic scales. Of particular importance are emission factors, which are used to calculate the amount of carbon emitted during particular parts of the production and consumption processes. Using aggregate data may mask important differences between different countries or regions within a country. For example, a country generating a large proportion of its electricity from renewable sources, such as hydro, should have a significantly smaller emission factor for electricity than one dependent on coal power generation. Relatively few farms and processing plants situated in developing countries will be visited by the analysts who calculate the carbon footprint of the food items they produce. Rather the farmers will be required to complete a questionnaire on their agricultural practices and the consultants will use these alongside standard databases of emission factors to calculate the carbon footprint. This brings two problems: first, the analyst may have an incomplete understanding of the system of analysis; and second, the databases may contain poor data on many developing countries.
  • Some of the commodities produced by developing countries, such as sugar, tea, and nuts, are sold to consumers as a blend of product derived from more than one country. In the absence of a clear country of origin, carbon accounting methodologies require that data relating to the worst case scenario be used. These can be quite different from the actual data for the different countries of origin. For some variables, such as land use change, the worst case scenario relates to the conversion of tropical forest in Malaysia, which is unlikely to be relevant for crops produced in Africa, South America, or Central Asia.
Recommendations for development-friendly carbon footprinting

Given this situation, the following would improve the efficacy of carbon footprints of food products and ensure that products from developing countries are not disadvantaged by weaknesses in the scheme:

1. Require all footprints to declare what the emissions from land use change would have been if the conversion occurred today. Thus products from Europe and North America would have to declare the greenhouse gas emissions that converting native forest to agriculture would emit if it happened today. This would allow for an equitable solution to the issue of including emissions from land use change in carbon footprints.

2. Include benefits derived from tree and bush crops in footprints. Tree crops can sequester carbon, and the soil under tree crops and agroforestry systems typically contain more carbon than other forms of cropland. It is important to provide credit for the carbon sequestered by these agroforestry systems

3. Develop better databases of land use change and emission factors for tropical and subtropical areas. If carbon footprints are to accurately reflect emissions from production activities in developing countries then the level of precision in the databases that provide emissions data and the historical and current distribution of land cover and land use must be increased. To prevent use of the global worst case scenario, which may be irrelevant, it is important for regional worst case situations to be identified and made publicly available.

4. All data and calculations of carbon footprints should be published in a public database. It should not be permissible for retailers or others to declare carbon footprints on consumer facing labels or Web sites unless the details of the calculation and the data used are publicly disclosed. These sites should also clearly state (i) all the assumptions made when calculating the carbon footprint, (ii) whether consultants actually visited the countries concerned or less accurate secondary data was used (iii) present information on the uncertainty that surrounds their calculations Such publication would allow governments, NGOs, journalists, producers, and the public to scrutinise the data and the methodology to judge the accuracy of the results.

5. Provide more disaggregated consumer information. Some carbon footprints require that emissions from the use phase of the product be included in the overall calculation. For food items, the main emissions relating to use are often from cooking and refrigeration. For products like coffee, the use phase is so large that it may mask carbon efficiencies in production. In this case, the footprint should be broken down to demonstrate the proportion of the overall emissions derived from the different phases of the life cycle (for example, on the farm, land use change, processing, transport, and use). This may enable consumers to realise that even though the footprint of a particular product is relatively high, it was not the farmers in developing countries who were responsible for the majority of emissions.

6. Encourage innovation in the food chain. Carbon footprints are often presented at an aggregate level, such as when multinational companies report the footprints for their final products from a region as if it were one uniform good (for example, beans from Kenya, grapes from Chile). There is then little incentive for the individual businesses who contribute to the production of these products to reduce their own emissions. If individual businesses could be provided with direct incentives for reducing their emissions, then innovation in the food chain would be encouraged.


Brenton, P, G Edwards-Jones, and M F Jensen (2009), “Carbon Labelling and Low Income Country Exports: A Review of the Development Issues”, Development Policy Review 27:243–65.

Brenton, P, G Edwards-Jones, MF Jensen, K Plassman, A Norton, and N Attarzadeh (2009), “Carbon Footprints and Food Systems: Do Current Accounting Methodologies disadvantage Developing Countries?” World Bank report.

BSI (2008), “PAS 2050:2008. Specification for the Assessment of the Life Cycle Greenhouse Gas Emissions of Goods and Services”, London, UK: British Standards.

Edwards-Jones, G, K Plassmann, E H York, B Hounsome, D L Jones, and L Milà i Canals (Forthcoming), “Vulnerability of Exporting Nations to the Development of a Carbon Label in the UK”, Environmental Science and Policy.

105 Reads