Designers can create buildings which are 'zero carbon' or 'low carbon', but this doesn't mean that the predicted savings of greenhouse gas emissions will be generated. What does? The answers may surprise you.

BedZED, in South London, was an early example of a 'zero carbon development'. It stands for 'BEDdington Zero Energy Development'. It was constructed in 2002 by a consortium led by Bill Dunster Architects for social housing body The Peabody Trust, which included Arup, BioRegional Development, Ellis & Moore and Gardiner & Theobald.

It included plenty of state-of-the-art features such as:

• 100% renewable energy powered ventilation & heating systems
• South facing glazing for passive solar heating
• Bio-fuel CHP
• Heat recovery wind-cowl ventilation
• Water recycling
• PV & electric cars with a car share
• Super-insulation
• North facing workspaces with passive cooling
• Thermal mass energy store
• Ventilation wind-cowls with heat recovery
• Roof gardens provide family outdoor spaces
• Reused steel and timber
• IT connected E-community.

It sounds perfect. What could go wrong? But a post-occupancy evaluation by BioRegional found that the carbon footprints of occupants, who were predominantly public service workers, was not significantly less than anyone else living nearby. Certainly, the predicted savings of emissions were not being met.

This was puzzling. They realised that the residents were compensating themselves for their 'green' lifestyles by flying and other perks, perhaps bought with the money saved from not paying energy bills, which effectively wiped out the carbon savings provided by the building. Flying is the most carbon intensive thing you can do.

But this wasn't the whole picture. BioRegional dug deeper. The company's Nick James produced a report, called One Planet Living in the Thames Gateway, proposing a zero energy development east of London, which the UK Government was developing at the time.

The Thames Gateway covers over 700 square kilometres including over 1,000 hectares of brownfield (formerly industrial) land being redeveloped. As the largest regeneration site in Western Europe, it had, since the 1970s, been proposed as an area in which to meet new housing needs for London with some 160,000 to 200,000 new homes imagined.

While producing the report, the team found that "Only a small proportion of CO2 savings, and an even smaller proportion of ecological footprint, could be achieved by constructing very green buildings".

The proportion is just 4%. The team was shocked.

BioRegional's Pooran Desai, wrote later that "By targeting significant CO2 and water savings in the home by building to BRE EcoHomes 'Very Good' standard (with high insulation standards and fitted with energy-efficient appliances and lighting) rather than simply meeting Building Regulations, we were able to get a 32 per cent reduction in CO2 emissions from the homes in our model. This level of saving sounds good. However, it translates to an overall reduction in ecological footprint of only 4 per cent."

BedZED had claimed a 90% reduction in space-heating saving. But this was compared to existing homes. Compared with meeting current Building Regulations, whose rules were tighter, the savings were much less. BioRegional realised they needed to consider instead the marginal or additional benefit of going beyond Building Regulations.

The team went on to crunch more numbers. They wanted to find out how to achieve carbon savings the most cost-effectively. They soon discovered that a "capital cost saving of 1 tonne of CO2 per year varies between £265 for a green transport plan to £28,697 for the same savings from the building fabric (superinsulation and sun-spaces)".

It must be remembered that this is in the city, where it is relatively easy to get a green transport plan together. But even so, it was 100 times cheaper to make CO2 reductions by investing in green transport than in insulation and so on.

Furthermore, of the 'insulation and so on', it was glazing that took the lion's share of the cash. 80 per cent of the extra cost for the building envelope of BedZED was for south-facing windows on the sun-spaces (£282,200 out of the £352,750 for the six-plot terrace).

This meant that superinsulation was much more cost-effective than sun-spaces in cutting carbon dioxide emissions. Moreover, the crunched numbers also revealed that savings to residents of cash and CO2 in heating bills from both superinsulation and sun-spaces were comparatively small compared to the other measures.

PV panels for solar electricity were then over twice as expensive as nowadays and they too weren't cost-effective.

Far more effective was water-efficient appliances which, by saving hot water, turned out to save as much energy as energy-efficient appliances, not to mention cutting the water bill.

Finally the team found that the financial payback time for investing in a tonne of annual CO2 saving (dividing capital cost per tonne CO2 saved by bill saving for the 36 people in the terrace), worked out at almost 800 years for the sun-spaces, but only one week for the car club.

Surprised? I was - by the extent of the difference. Of course it's because the cost of running a car is high, and they produce a lot of emissions.

Pooran Desai, in his book on the subject One Planet Living, concludes: "This doesn't mean that I am saying don't bother creating energy-efficient homes. What I am saying is that there is a point of emphasis and of diminishing returns with building fabric."

Pooran adds: "We are getting there, but still have a way to go. If we stop at CO2 footprint we are not getting a full indication of our contribution to global warming."

But he goes further too. One Planet Living looks at the whole impact of our lifestyles. Greenhouse gas (GHG) emissions are measured as tonnes of carbon dioxide equivalent (CO2e), but we focus on this gas too much. The largest man-made contributors to global warming are from methane and nitrous oxide emissions and they come from agriculture. It's our agricultural practices which need to change.

We need to eat less meat and milk, because methane comes from cattle, farm more organically and labour-intensively, because nitrous oxides come from nitrogen fertilisers.

The UK government says that the individual carbon footprint of a UK citizen on average is 8.8 tonnes of the gas per year. But if you factor in GHG emissions from agriculture and other sources, it almost doubles to 16.34 tonnes of CO2e per person per year.

So when people quite statistics like "50 per cent of CO2 emissions come from buildings' it actually means "50 per cent of territorial CO2 emissions come from all existing buildings and the built-up space between buildings" but it ignores CO2 emissions from international transport and agriculture.

We will never reach sustainability in cities or elsewhere unless all these emissions are brought within safe limits, the limits of what our one, beautiful planet can provide for each of us and our descendants.