Do carbon labels change shopping behaviour?

If everyday items were labelled according to the carbon emissions embodied in them, would shoppers change what they buy? And if they did, would it make a difference in the grand scheme of things?

Voluntary reductions in emissions by households do play an important role, and groceries account for a substantial proportion of the environmental burden of the average household through production, transport, and consumption. Indeed, American researchers found that individual household purchasing generates 30-40% of US emissions. But despite much enthusiasm for greenhouse incentives (carbon tax, solar rebates, ceiling insulation), there has been little consideration of the likely consumer response to carbon labelling of groceries, so my students set out to gauge their response.

Their study was completed in a suburban grocery store in the NSW north coast town of Ballina during a 12-week period during mid-2008. They identified five categories of grocery items with a relatively high-turnover and wide product choice, including milk, spreadable butter, canned tomatoes, bottled water, and non-perishable pet foods. These 37 products were categorised according to their CO2 emissions generated right up to the point of sale, and labelled with one of three symbols indicating average (yellow footprint), lower than average (green footprint) and higher than average (black footprint).

The classification was strongly influenced by the energy used during transport (for example, some bottled water was freighted long distances by road) and by the energy embodied in packaging. But distance did not tell the whole story, because some water bottled abroad and shipped by sea had less embodied carbon than Australian water that had been trucked interstate. Thus “food miles” may be an over-simplistic concept, and it was this realization that led us to use three different coloured footprints.

Labels were in place for 8 weeks and their introduction was communicated via leaflets and media coverage. The media interest was high and, in the first fortnight following the placement of the labels, gross store turnover increased by 12%.

Over the eight weeks, the share of black-labelled sales fell from 32% to 26%, whilst the share of green-labelled sales increased from 53% to 57%. Further analysis revealed three different trends of customer response. When green-labelled products were also the cheapest in the category, there was a strong consumer response and sales increased by about 20%. However, if the green-labelled goods were not the cheapest then the response to labelling was weaker.

Our study also found that other factors may dominate over carbon footprint and price, such as with perishable goods, where labelling appears to have little impact. For example, with fresh milk it appeared consumers had a strong preference for a particular size of package, which appeared to be more important than the footprint associated with this package.

Our research generated strong media interest, and this could have affected consumer choice, if for instance, publicity attracted strongly greenhouse-conscious people to come and shop at this store. However, the changes observed remained evident until the end of the study, even after the media attention diminished. And there was no long-term influence on store turnover, so it seems that any impact of the publicity was short-lived.

Our study provided an interesting insight into carbon labelling in a real-life context and indicated the potential for carbon labelling in a broader context. The strong response of consumers when price difference and carbon footprint are both low indicates that combining a carbon label and price incentive (via a carbon tax or emissions trading system) could be effective in encouraging sustainable consumption.