Data localisation: Do spend-based CO2 estimates really change that much between countries?
Calculating emissions data between different localities is like comparing apples to apples — in that an apple in New Zealand costs twice as much as an apple in the UK.
As with pricing, the emission rates between like-for-like purchasing categories can change a great deal depending on where you are in the world. Plus, it’s not the only variable involved; inflation rates also have a significant impact at a country level, among other factors. The localised differences become more and more apparent as these factors stack up.
Meaningful estimates are a snapshot in time and place. People need to know the carbon costs of their transactions based on the most up-to-date data available for their local area.
Taking a global view is not accurate enough — data localisation matters. Here are three examples to prove it.
Example 1: The Energy Mix
2021 Carbon intensity of electricity, generation only. Source for UK. Source for France. Source for US.
In the US, around 20% of the energy mix is attributed to renewables, with reliance on fossil fuels hovering at about 60% in 2024.
By comparison, in the UK, about 30% of electricity comes from wind. Another 30% comes from natural gas, which is considered to be lower emitting than other fossil fuels.
In fact, even within this small island, there are major differences in the energy mix, i.e. wind generated 78% of all renewable electricity output in Scotland in 2022. Given Scottish energy feeds into the national grid, UK-level data is considered acceptable at this time. However, if the national grid were ever devolved regionally, carbon calculation providers may be expected to re-evaluate their approach.
It’s clear even neighbouring countries that are geographically and economically similar can be at very different stages in their use of renewable energy. Reasons such as policy priorities, infrastructure requirements and resource availability all have an impact. France is considered a global leader in the use of nuclear power as a share of total energy production, deriving about 70% of its electricity from nuclear energy. This is a totally different picture to the UK, just across the Channel, which will inevitably need to be reflected in data localisation calculations.
The energy mix of renewables, nuclear, fossil vs. natural gas and other sources often changes at the country border. So, in most instances, taking a broad regional average is not going to provide a clear enough picture. You can see this play out here:
Poland is a heavy coal user with 73% of electricity derived from fossil fuels and 61% derived from coal in 2023. Iceland, on the other end of the scale, derives almost all of its electricity from renewables, mostly geothermal and wind power. Both countries have a completely different emissions rate in kilograms of CO2e per kWh compared to the average in Europe.
Example 2: Product Categorisation
Let’s return to our shopping basket. Looking at food and grocery categories illustrates the regional differences for similar types of produce:
Derived from: source.
Take rice as an example in this chart. Between Germany and Poland, the difference between the CO2 emissions factors (kg CO2e/EUR) of this crop is considerable.
The reason for country-level differences here are many, from how much of the produce is imported to different production processes or levels of efficiency along the supply chain.
It can also be down to sociological shifts in the demand for certain foodstuffs within the country or for exports, and the positive and negative impacts of scaling crop growth. A classic example to illustrate this point is the meteoric rise of quinoa from Bolivia and Peru. As vegan diets in the west drove interest in this “superfood”, it quickly became so in demand that pressure has since driven land-use for diverse crops towards an ecologically damaging monoculture. Emissions associated with this grain are further compounded by the fact that most quinoa is then exported overseas.
“While the majority of agricultural systems remain the same over time, sometimes there are sudden drastic changes and databases have to be updated, so the numbers can fluctuate dramatically for certain base products or whole product categories.”
— Josh Couchman, Head of Data at Connect Earth
Carbon Intensity of a weighted basket of groceries. Derived from: source.
Beyond specific products, when calculating a consumer’s grocery shop as a “basket of goods” emissions factors have to be further weighted to account for local consumption habits. Some cultures may consume more or less meat or dairy, for example. Accounting in part for the results in the chart above, Germany consumes 51.6 kilograms of meat per capita while in Poland the average person eats 64.2 kilograms.
At Connect Earth, we know these consumer behaviour trends are ever-evolving, so we are always seeking ways to improve the accuracy and granularity of our models.
Example 3: Inflation
Finally, let’s consider the impact of inflation on the ability to accurately estimate emissions from spend, and how that might impact data localisation considerations. Consumers may spend twice as much on fuel, for example, from one month to the next not because they are using twice as much of the fuel, but because the price has doubled due to inflation. It’s crucial to account for this in estimating the impact of these types of purchases.
Here, we’ve charted inflation rates across three such product categories from 2020 to today: electricity, gas and petrol.
Electricity
Gas
Petrol
Across Europe, we know electricity prices have risen considerably over the last two years. However, even in regions that you might anticipate being homogenous, like the EU, there are massive changes between countries due to economic situations or policy approaches such as price caps. Looking at the data, the differences are obvious not only between regions but over different time periods.
The clearest outlier, Belgian gas prices saw a massive spike in 2022 following the release of lockdown and the outbreak of the war in Ukraine. In 2023 those prices then fell below what people in Austria, Germany, France and Portugal were paying. Again, we would not see a complete picture if we looked at European averages. It’s only by drilling down to the country level that we see these trends and differences.
Our Gold Standard for Data Localisation: Country-level Data
The data experts at Connect Earth have found that over-reliance on broad regional data can produce unacceptable inaccuracies that don’t appropriately reflect reality. On top of this, the available data sources are dictated by official reporting guidelines from different countries, showing not only differences in the actual emissions, but also differences in calculation methodologies.
As a result, wherever possible, we strive to make our spend-based estimates granular, timely and specific to the country where these transactions are occurring.
So, that’s data localisation. To learn more about how calculating CO2 estimates works, check out our white paper on the power of carbon emissions data.
