The Difference Between Reducing Carbon and Shifting It: Why Some Supplier Switches Make Your PCF Worse

A manufacturer decides to switch to a lower-cost supplier based in a different region. The new supplier quotes a lower carbon intensity for the same material. On paper it looks like a clear PCF win. The switch is made. Six months later, when the product carbon footprint is recalculated, the number is higher than before the change.

The culprit is not the material. It is the transport leg — which tripled in distance, moved from sea to air freight for part of the route, and introduced a Scope 3 transport contribution that the original supplier comparison had never modelled. The carbon that appeared to leave the product through the material switch came back through the logistics tail — and brought more with it.

This is carbon shifting. It is one of the most common, least discussed errors in product-level decarbonisation. And it happens because manufacturers evaluate supplier carbon performance in isolation rather than as part of the full product system.

Carbon shifting, in the context of a product carbon footprint, occurs when an action that appears to reduce emissions in one part of the product system actually relocates those emissions to another part — leaving the total unchanged or making it worse.

It is distinct from carbon leakage, which is a policy concept describing how emissions reduction in one country can trigger increased production in another country with weaker regulation. Carbon shifting in a supply chain context is a product-level accounting problem. When a manufacturer changes a supplier, a material, a process, or a logistics route, emissions do not disappear. They are recalculated across a different set of inputs, geographies, and emission factors. If those secondary effects are not modelled before the decision is made, the outcome can contradict the intention completely.

The core issue is system boundary thinking. A product carbon footprint calculation covers the full system — raw material extraction, upstream processing, inbound transport, manufacturing, outbound distribution, and for cradle-to-grave analyses, use phase and end of life. A supplier switch that improves one node in that system while worsening others does not reduce the product's total footprint. It shifts it.

This is the pattern described in the opening scenario and it is extremely common. A supplier in a geographically distant region produces a material with a lower emission factor per kilogram — perhaps because their facility runs on a cleaner energy grid, or because their production process is more efficient. The material-level carbon comparison looks favourable.

But the mode and distance of transport changes significantly. Air freight emits approximately 500 to 1,050 grams of CO₂ per metric tonne per kilometre of transportation. Sea freight emits 10 to 40 grams per metric tonne per kilometre — making air freight roughly 20 to 50 times more carbon-intensive per tonne-km than sea freight. For high-value, low-weight, or time-sensitive components, air freight is often used regardless of sustainability intentions. A supplier switch that moves sourcing from a regional supplier reached by road to an overseas supplier requiring air freight for part of the journey can easily negate, or exceed, any material-level carbon saving.

Even a full switch to sea freight over a substantially longer route adds transport emissions that must be modelled accurately, not assumed to be negligible.

Manufacturing location matters significantly in PCF calculation. The same production process in a country with a coal-heavy electricity grid generates substantially more Scope 2 and embedded manufacturing emissions than in a country powered by renewables or nuclear.

The GHG intensity of electricity production differs significantly between countries and regions. Within Europe alone, [CORRECTED: Poland, Cyprus, and Estonia recorded the highest electricity carbon intensities in 2023, with Czechia and Germany also ranking above the EU average], driven by continued reliance on solid fossil fuels and limited renewables penetration. Meanwhile, countries like Norway and Iceland have among the lowest grid intensities globally — Norway dominated by hydropower, and Iceland by a combination of geothermal and hydropower.

A supplier in a lower-wage manufacturing region may offer a cheaper, or nominally cleaner, unit product while their factory's electricity grid intensity means that the actual embedded manufacturing emissions are significantly higher than those of a supplier in a high-renewable-grid country. Without applying location-specific electricity emission factors to the manufacturing stage in the PCF calculation, this difference is invisible until the recalculation reveals it.

Suppliers increasingly market recycled content as a carbon reduction feature. Secondary aluminium, for example, requires substantially less energy to produce than primary aluminium because the energy-intensive electrolytic reduction of bauxite ore is bypassed. Recycled aluminium uses roughly 5% of the energy required to produce primary aluminium from bauxite. This is a real and significant emissions difference.

However, how recycled content is treated in a PCF calculation depends on the allocation method applied — specifically how the emissions credit for recycling is distributed between the product that supplied the end-of-life material and the product that consumes it. Under different allocation approaches, the same recycled-content material can carry very different emission factors in a PCF model. A supplier claiming low carbon for recycled content may be using an allocation methodology that is incompatible with the buyer's PCF methodology. When the buyer recalculates the PCF using their own methodology correctly applied, the benefit is smaller than advertised — or disappears entirely.

A supplier may have invested in energy efficiency at their own facility, reducing their direct and energy-related emissions measurably. This is a genuine improvement. But if those efficiency gains were achieved by outsourcing energy-intensive upstream processing steps to a third party — a sub-supplier whose own emissions are not accounted for in the supplier's PCF claim — the buyer's Scope 3 calculation must capture those upstream emissions regardless.

The GHG Protocol's Phase 1 Progress Update, published March 2026, proposes a 95% coverage floor for Scope 3 reporting, with exclusions requiring documented justification. The direction is toward comprehensive upstream accounting, not selective reporting. A supplier PCF that covers only the supplier's own operations without tracing upstream inputs can produce a misleading number that causes the buyer to make a sourcing decision based on incomplete information.

Most supplier carbon comparisons are conducted using the supplier's own disclosed carbon data — a carbon declaration, a product-level EPD, or a response to a supplier questionnaire. These documents report the supplier's emissions as they have defined and measured them. They do not automatically reflect the full-system PCF impact of using that supplier in a specific product, in a specific supply chain configuration, with a specific logistics route and manufacturing location.

The buyer's PCF model must integrate the supplier's data into the full product system and recalculate across all stages. This means applying the actual transport emission factors for the specific route and mode used to bring materials from that supplier to the manufacturing site. It means applying the correct grid intensity for the manufacturing location. It means ensuring that the allocation methodology used by the supplier for recycled content or by-product credits is compatible with the methodology used in the buyer's PCF.

When this integration step is skipped — when a supplier switch is evaluated only on the supplier's own carbon disclosure without modelling the full-system PCF impact — carbon shifting is not a risk. It is close to a certainty in any case where the supplier change also alters geography, transport mode, or upstream processing structure.

Before making a sourcing change on carbon grounds, a full-system PCF delta analysis should answer three questions.

A supplier switch that passes all three tests produces a genuine system-level PCF reduction. A supplier switch that skips this analysis and relies only on a supplier's self-disclosed carbon intensity is a carbon shift waiting to be discovered in the next recalculation.

The difference between genuinely reducing a product's carbon footprint and shifting it to a less visible part of the system comes down to one question: has the total system been modelled, or just the part being changed?

Manufacturing location matters significantly. The same production process in Norway versus Poland versus China carries different energy-related emissions depending on the electricity grid carbon intensity. Mode of transport and distance both matter, as sourcing a material locally versus from a distant region will produce very different transport emissions.

These are not edge cases or rare complications. They are standard variables in any product carbon footprint calculation. When they are treated as variables — modelled explicitly before a sourcing decision is finalised — carbon shifting is detectable and avoidable. When they are treated as assumptions — rounded away or left out of the supplier comparison — the PCF can worsen from an action designed to improve it.

The manufacturers who are building credible, defensible product carbon footprints in 2026 are the ones who evaluate supplier changes as full-system questions, not component-level comparisons. The carbon is always going somewhere. The only question is whether the model accounts for where.