What are the Limitations of Life Cycle Assessment?

LCA provides valuable environmental insights but has inherent limitations. Understanding these limitations helps interpret results appropriately and avoid overreliance on methodology that can’t answer every question.

Scope Boundaries#

System boundaries exclude some impacts by definition. You must draw a line somewhere, but this means potentially significant processes might sit outside assessment.

Capital equipment and infrastructure typically get excluded. Building a factory creates substantial impacts, but these get amortised across millions of products. The per-unit contribution becomes negligible, but the absolute impacts are real.

Services and indirect effects often fall outside boundaries. The environmental impact of office work supporting manufacturing, research and development, or marketing activities typically aren’t included. These matter at organisational level even if not at product level.

Geographical boundaries limit completeness. A European LCA uses European background data. Global supply chains actually source from diverse regions with different environmental profiles. Generic data masks this variation.

Data Quality and Uncertainty#

LCA results carry substantial uncertainty. Multiple data sources with varying quality combine through complex calculations, propagating uncertainty throughout.

Secondary database data represents averages. Your specific supplier might perform much better or worse than the average. Generic data for “European steel production” doesn’t distinguish between efficient modern plants and older facilities.

Measurement uncertainty affects primary data. Utility meters have accuracy limits. Production yields vary. Transport distances are estimated. Small measurement errors compound through calculations.

Temporal mismatches create uncertainty. Current electricity grid data represents today’s mix, but products use energy over years or decades as grids change. Historical data might not represent current operations.

Data gaps force assumptions. Missing information gets estimated, substituted with proxy data, or excluded. Each gap introduces potential error.

Methodological Choices#

Different practitioners applying LCA to the same product can reach different conclusions based on legitimate methodological choices.

Allocation rules substantially affect results. Mass allocation, energy allocation, and economic allocation produce different outcomes. Each approach has valid justification, but they create different product-level impacts from multifunctional processes.

Impact assessment method selection changes results. Different methods use different characterisation factors, particularly for toxicity impacts. Method choice can alter which option performs better.

System expansion versus allocation creates divergent results. Consequential approaches model system changes differently than attributional approaches. Both are valid for different questions, but they don’t produce comparable results.

Cut-off thresholds determine what gets excluded. A 1% cut-off might exclude different processes than a 5% cut-off. Multiple small exclusions accumulate. The threshold choice is somewhat arbitrary even within ISO guidance.

Incomplete Impact Coverage#

LCA doesn’t measure all environmental concerns. Standard impact categories cover many issues but miss others.

Biodiversity loss lacks robust assessment methods. Habitat destruction, species extinction risk, and ecosystem function degradation are acknowledged as important but poorly characterised in LCA.

Noise pollution typically falls outside LCA scope. Aircraft noise, traffic noise, and industrial noise create real environmental and health impacts but don’t appear in standard assessments.

Microplastic pollution is emerging as significant but isn’t captured in current impact methods. Standard toxicity assessments don’t fully account for persistence and accumulation of microplastics.

Radiation impacts from nuclear power appear in some methods but not others. The controversy around nuclear makes this a contested methodological area.

Social impacts sit entirely outside environmental LCA. Worker safety, fair wages, community impacts, and human rights issues matter but aren’t environmental impacts as LCA defines them.

Temporal Limitations#

LCA provides a snapshot assessment. Dynamic changes over time are difficult to model.

Technology improvements happen during product lifetimes. A vehicle manufactured in 2025 uses electricity from 2025 grids during 2040. Prospective LCA attempts to address this but introduces additional uncertainty.

Market changes affect consequential models. Predicting how industries respond to demand changes requires economic modelling beyond typical LCA expertise. These predictions often prove wrong.

Long-term impacts like groundwater contamination or climate change occur over decades to centuries. LCA collapses these into single impact scores, losing temporal information about when impacts occur.

No Absolute Scale#

LCA provides relative comparison, not absolute environmental performance assessment. Knowing Product A has lower impacts than Product B doesn’t reveal whether either is actually “sustainable” or within planetary boundaries.

Impact categories have units (kg CO₂-eq, mol H⁺-eq) but these don’t translate to ecological thresholds. 100 kg CO₂-eq might be “good” or “terrible” depending on context that LCA doesn’t provide.

Normalisation attempts to address this by comparing to per-capita or regional total impacts. However, these reference values are themselves uncertain and normalisation involves value judgements.

Local Context Ignored#

LCA uses average impact factors. Local conditions substantially affect actual impacts.

Water use in water-abundant regions differs from water-scarce regions. Standard LCA might use the same impact factor. Water scarcity indicators attempt corrections but remain imperfect.

Toxicity exposure depends on population density, ecosystem sensitivity, and environmental fate. Emissions in remote areas affect fewer people than urban emissions. Standard characterisation doesn’t fully capture this.

Ecosystem vulnerability varies. Acidification matters more in already acidified regions. Eutrophication sensitivity differs between water bodies. LCA impact factors are generalised.

Rebound Effects#

LCA assesses direct product impacts but doesn’t model system-level rebound effects.

Efficiency improvements might reduce per-unit impacts but increased consumption could negate savings. More efficient vehicles might encourage more driving. LCA doesn’t model these behavioural responses.

Substitution effects extend beyond product comparison. A more efficient product that costs less might enable consumption elsewhere. The saved money gets spent on other impacts.

Subjectivity in Interpretation#

Interpretation involves judgement. Determining which impacts are “significant”, how to weigh trade-offs between categories, and what recommendations follow from results all require subjective decisions.

Two practitioners might emphasise different hotspots from the same data. One focuses on climate change, another on water scarcity. Both interpretations are defensible but lead to different recommendations.

Weighting impact categories involves value choices that science can’t resolve. Is climate change more important than biodiversity? How much water scarcity equals what amount of acidification? These are societal values, not technical questions.

Limited Forward-Looking Capability#

LCA describes systems as they are or were. It doesn’t inherently predict futures or model transformative changes.

Emerging technologies lack data. Laboratory processes don’t represent future industrial operations. Scaling relationships introduce additional uncertainty.

Disruptive changes resist modelling. If a new technology eliminates entire process steps or creates new value chains, historical data becomes irrelevant.

Resource Requirements#

Conducting rigorous LCA requires expertise, time, and money. Not all organisations can afford comprehensive assessment.

This creates equity issues. Large companies conduct LCA and demonstrate environmental performance. Small companies lack resources for equivalent assessment. Their actual performance might be similar or better, but they can’t prove it.

Resource constraints might force simplified approaches that miss important factors. Screening assessments identify major issues but might overlate minor hotspots or trade-offs.

Potential for Misuse#

LCA can be manipulated through methodological choices. Selective boundaries, cherry-picked impact categories, or favourable allocation rules can make products appear better than they are.

Transparency helps address this. ISO 14044 requires documentation and critical review for comparative assertions. However, non-comparative internal studies face less scrutiny.

Oversimplification in communication creates misunderstanding. Complex multi-dimensional results get reduced to single scores or simple statements. Nuance gets lost.

What LCA Does Well#

Despite limitations, LCA excels at several tasks:

• Comprehensive environmental burden quantification across life cycles
• Hotspot identification showing where to focus improvement efforts
• Comparison of alternatives under consistent methodology
• Avoiding burden shifting between life cycle stages
• Providing data for environmental product declarations
• Supporting environmentally informed decision-making

Understanding limitations doesn’t invalidate LCA. It requires appropriate interpretation. Use LCA for what it does well. Supplement with other tools for questions it doesn’t answer.

Addressing Limitations#

LCA practice continues developing:

• Improved impact assessment methods addressing current gaps
• Better data quality as measurement improves
• Standardised approaches reducing methodological variation
• Integration with other assessment methods
• More transparent documentation enabling critical review

Perfect assessment isn’t possible. The goal is useful insights despite imperfect methodology. Acknowledging limitations builds credibility rather than undermining it.