Over What Fraction of our Share of Emissions do we have Agency?

Let’s start by defining what one’s ‘share’ of emissions are, beginning with the Global Average Citizen (GAC). A GAC is the in-every-way-average human on the planet, at least as far as climate is concerned. We can determine the emissions profile of a GAC by simply dividing total emissions by total population.^*Yes, of course this is absurd

Background (Emissions Reminder)

We’ll use the IPCC AR5 breakdown of emissions:

Reference: IPCC AR5 WG3

Of the ~49 billion tons of CO₂e emissions in 2010, approximately 18% were associated with the operation of Buildings, ~14% with Transportation, ~32% with Industry, ~25% with Agriculture, Forestry, and Other Land Use (AFOLU), and ~11% with Other uses of Energy. These emissions are broken into Direct Emissions from the sector (i.e. when cement is produced, CO₂ is directly emitted as process emissions in the process) and Indirect Emissions stemming from the use of Electricity or Heat that is provided, typically, from the energy sector.

In 2010, the population of the planet was ~6.96 billion people.

Reference: Our World in Data

Therefore, the GAC’s emissions were approximately 7 tons of CO₂e^*The ‘e’ stands for equivalent emissions (meaning that other GHGs are described in terms of their equivalent impact on a CO₂ basis). In this post we will use CO₂e to mean ‘equivalent emissions impact, as measured by radiative forcing integrated over 100 years.’ in 2010, spread out across sectors as shown in the figure below:

This is great data for helping inform the guilt that we feel about our role in climate change, but it doesn’t tell us much about 1) what we can do or 2) what fraction of these emissions our choices as individuals can really affect.

To get at these questions, a different (and bottoms-up) slice of emissions is reflected in the Sankey diagram below:

Reference: Bajzelj et al. 2013

The breakdown by “Final Service” and “Sector” can help us understand what fraction of emissions are associated with the various activities that we, as lowly consumers, participate in throughout the course our lives. Note that energy use has now been integrated into sectors (as opposed to the IPCC data shared above). Moving forward we integrate the use of Electricity and Heat (i.e. indirect emissions) by sector.

A first-order naive attempt to classify the fraction of emissions over which we might have agency takes the sector classification of the above Sankey diagram and determines whether a consumer has any agency over that source of emissions. The chart below shows the outcome of assuming 100% agency over Passenger Transport, Residential Building Energy Use, and Food (Agriculture) and assumes 50% agency over Waste (the municipal solid waste fraction of the global waste footprint). The fractions of emissions over which the GAC is presumed to have agency are highlighted blue. This amounts to a total of ~47% agency, meaning that a GAC might hope to have agency over up to 47% of their associated emissions.

Following the same logic, but instead using the Final Service classification, these assumptions lead to a ~53% agency for our GAC.

We can perform the same analysis for an AAC (American Average Consumer) and a CAC (Chinese Average Consumer). As shown below, these two individuals have a nominal ~55% and ~41% agency over their emissions, respectively.

On naive inspection, it appears possible that our GAC does indeed have some agency over a significant fraction of their footprint. Of course, this analysis was hand-waving at best. In the following sections we’ll dive a bit deeper into a bottom-up assessment of the GAC’s agency.

Assumptions

So far all that has been provided are data. We now enter territory where judgement is applied, so reader beware.

Let’s use a framework of emissions scope organized as follows:

• Level 1 – Emissions directly controlled by consumer without requiring 3rd party behavior change
• Level 2 – Emissions over which the consumer has direct influence, subject to 3rd party consent or support
• Level 3 – Emissions over which the consumer has no direct influence

We’ll walk through each of the major sectors to try and assign Level 1, 2, and 3 emissions agency to each.

Agriculture (Food)

Level 1

A consumer can (assuming access to a reliable and varied supply chain – a poor assumption for much of the world to this day, unfortunately) control their diet directly. The choice to be a vegetarian, for example, is available to much of the planet.

Approximately 12-15 Gt of CO₂e is associated with the production of food. For the purposes of this discussion we will use the more conservative 12.7 Gt in 2010 (or 26% of global emissions).

Reference: Our World in Data

Thus, for the GAC of 2010, approximately 1.8 tons of CO₂e per year of emissions come from the consumption of food. The average person is supplied with 2870 kcal per day. This implies an average diet emissions intensity of 1.74 kg CO₂e per 1000 kcal of food.

It is theoretically possible to consume a net-zero GHG diet. Indeed, it’s possible to consume a net-negative GHG diet. However, for the sake of this analysis we will assume that the maximum range of agency is 100% (i.e. the minimum carbon footprint for food consumption is 0 ton per person per year) for a reduction potential of 1.8 tons of CO₂e per year. A more typical low-GHG diet that might be presumed a normative case for an individual choosing to exert agency in their diet is ~0.9 kg CO₂e per 1000 kcal of food for a reduction potential of 48%, or 0.87 tons of CO₂e per year. The average American vegan, for example, has a footprint of approximately 60% that of the average American, per the image below. The reason why the average American vegan diet contributes substantially higher total GHGs than the GAC vegan diet is that we supply far more than the global average 2870 kcal per person per day in the United States.

Source: Shrink That Footprint

Similarly, the average UK vegan has an annual footprint of approximately 1 ton CO₂e compared to the average UK citizen’s footprint of approximately 2 tons CO₂e, for an approximately 50% reduction.

Level 2

Per the above description of Level 1 emissions for Food, theoretically a consumer could exert up to (and perhaps exceeding) 100% agency. However, the average consumer would have to go to considerable lengths to source carbon-neutral food satisfying dietary preference and nutritional requirements. Support from at least one 3rd party (e.g. a grocery store, food logistics provider, farmers) is likely required. We will assign the residual 52% of agency (~0.93 tons of CO₂e per year) over Food to Level 2.

Level 3

We have already accounted for 100% of food-associated emissions for the GAC. However, it is very likely that a GAC has a certain share of Food related emissions that are not at all subject to their agency merely by their existence in society. An example might be the deforestation associated with certain practices of the community in which they reside, or the food footprint of animals at their local zoo. It is thus likely overly optimistic to presume that 100% agency is possible for Food, but for the sake of this analysis we will leave the numbers as calculated.^*This sort of discrepancy is all-too-frequent when using bottom-up analysis to fractionate a top-down quantification. We do what we can.

Transportation

Of the GAC’s ~1.4 tons of CO₂e per year of transportation-associated emissions, roughly 0.8 tons are personal transport and roughly 0.6 tons are freight-related. We will address each in turn.

Level 1

It is conceivable that an individual could exert complete agency over their personal transport related emissions, for a maximum potential reduction (aggressive) of 100%, or 0.8 tons CO₂e per year. However, this would require orienting a substantial fraction of one’s life around this goal. Proximity to work, willingness to remotely engage with far-flung friends, avoidance of air travel, and minimizing the distance between oneself and basic needs (i.e. groceries) is certainly achievable, and yet it is obvious upon inspection that within the current cultural context, for most individuals these choices aren’t truly within the realm of agency. How many of us are sufficiently privileged to choose a job that does not require commuting? How many of us are sufficiently privileged to co-locate with our families? How many of the substantially privileged are willing to avoid travel to see loved ones?

Instead of the impossible task of determining what is ‘reasonable’ to assume one might achieve in reduction via lifestyle choices, as a heuristic we will instead attempt to consider the variation in miles traveled in countries of similar levels of quality of life. As shown below, in 2005 the average distance traveled (by land) in the United States was ~26,000 km per person, compared to just ~10,000 km per person in Japan.

Source: Shrink That Footprint

Unfortunately, annual passenger miles per capita appears to be highly correlated with GDP per capita:

Source: EIA

What is more challenging to ascertain is to what degree this relationship expresses a causal connection between GDP per capita and traveled passenger miles (i.e. the preference for travel in light of increased income) or the inverse causality (i.e. the productivity associated with additional transportation). The average household passenger transport miles by purpose in the United States is shown below (courtesy of ORNL). As above, shaded in blue is the fraction which I judge to be subject to agency.

In total, of the average ~36,000 passenger-miles per household traveled in 2017, ~18,000 (or 50%) are subject to discretion.

In addition to the lifestyle choices that one might make to influence the magnitude of travel, we can also consider agency over the emissions intensity of a given mile of travel. Today, both globally and domestically, passenger vehicle transport (i.e. cars) still dominates personal transportation emissions (~70% of the GAC’s personal transport emissions and ~80% of the AAC’s). The average emissions intensity of the average vehicle is a function of fuel type (e.g. gas, diesel, electricity), vehicle size (e.g. sedan vs. SUV), and vehicle efficiency. In addition, the emissions intensity depends on factors such as the emissions intensity of manufacturing (for which, in the present analysis, we will consider in the category of ‘industry’) and local factors such as the grid emissions intensity.

Source: Our World in Data

What is the level of agency that an average consumer might have over the average emissions intensity of their primary means of vehicular passenger transport?

To identify the range of agency, we can consider the bounding cases. If the average AAC (for which there is higher-fidelity data to perform this analysis) were to swap their road transportation to the average electric vehicle (assuming average grid emissions intensity in the US, fuel-cycle emissions alone (i.e. not lifecycle emissions, herein counted under industry), and average vehicle occupancy) the emissions intensity of a given passenger-km would be reduced by approximately 65%. See US Road Transportation for the analysis.

A 65% reduction of 80% of an AAC’s personal transportation emissions footprint translates to a 52% reduction in overall personal transportation emissions. But could one go even further? Perhaps by swapping out the use of an electric bike for that morning commute, or using public transportation (especially rail), one could further reduce one’s personal transportation footprint. It is difficult to assess this on a rigorous basis without detailed demographic data that are challenging to capture.

We will therefore assign a potential 50% reduction associated with discretionary miles traveled (total) and a 52% reduction in emissions intensity per mile (for road transportation, which is responsible for 83% of the GAC’s personal transportation emissions). This totals to 77% Level 1 emissions agency for personal transportation.

It is assumed that the consumer as no direct agency over associated freight emissions.

Level 2

Where a person lives and works and their distance from family and friends has substantial influence over their personal transport emissions. For example, the emission intensity of the grid (on a generation basis) in Idaho is ~0.086 kg CO₂e per kWh while in Wyoming it is ~0.926 kg CO₂e per kWh which means that charging your electric vehicle in Wyoming does a hell of a lot less to reduce your footprint than charging it in Idaho. Public transportation availability (and emission intensity) varies wildly by location. Urban dwellers tend to drive fewer miles during their daily commute and are more able to swap out their mode of transportation for biking or public transit. See the chart below that shows the negative correlation between population density and transportation-associated emissions.

Source: Climate Change and Cities

A plurality of factors contribute to the available choices to a consumer regarding the emissions intensity of their transportation footprint. Clearly, without extreme measures it is unlikely that a GAC could reduce their transportation-related emissions to zero. However, and especially with the agency to choose where to live, substantial reductions are possible. We will consider the residual 23% of personal transportation emissions to be of Level 2 agency.^*For an existence proof, it would be possible to purchase residential solar generation to power one’s electric vehicle and for one to refuse to fly.

Level 3

Per the above discussion, freight transportation emissions have been considered to be without direct influence of the GAC. Consequently, 100% of freight will fall into Level 3.

Industry

Dissecting the vastly complex landscape of Industry emissions with a lens for agency is not for the faint at heart. At the risk of oversimplifying, the consumer is assumed to have no direct influence over their industrial footprint. Perhaps the greatest single choice, should a consumer have it, surrounding their industrial emissions is where to live. Steel emissions and building codes (and the associated required volumes of cement) vary by geography. Grid and fossil fuel LCA emission factors vary too. While it is true that a consumer could choose to buy goods based on the emissions intensity of the manufacturing processes associated with them, this level of data is not typically accessible to the GAC, nor is the agency to change behavior without substantial accommodation from society. Consequently, the vast majority of Industry emissions are at best Level 2, and more probably Level 3. For the sake of simplicity (and at the risk of indicating to the GAC that there is no hope for them to control their industry footprint), we will assign 100% of Industry to Level 3. In a future piece a more actionable set of recommendations will be provided that can have influence over even Industry emissions.

Waste

Waste is broken into two main buckets: wastewater and municipal solid waste (MSW, otherwise known as trash), representing approximately 0.1 tons of CO₂e per GAC each.

Level 1

There are two obvious vectors by which a consumer might influence their MSW waste footprint: generation and disposal. Decisions around consumption influence the generation of waste. The vast majority of the GHG footprint of MSW comes from the generation of methane during anaerobic digestion of organics in landfills and/or dumps (with a small fraction of the residual resulting from incineration or burning of plastics). Of this, the majority of the organics comprise food waste (with yard trimmings, paper/paperboard, and textiles making up the balance). Consequently, much of the Waste footprint is indeed a consequence of agency exerted over food choices. Consumers could choose to generate less organic waste by not overbuying perishable food and by managing their yards with practices less generative of organic waste. Beyond organics, a consumer’s buying habits can influence the mitigation of lifecycle emissions when combined with recycling by choosing goods (and packaging) that are either 1) low-emissions to begin with or 2) easily (and efficiently) recyclable given the community’s recycling capabilities.

Once generated, organic waste is only significantly detrimental when exposed to the anoxic conditions of landfills and dumps. Composting, disposal of food waste down the garbage disposal, and diligent source-separation by the consumer in communities with municipal composting facilities or paper recycling capability can dramatically reduce a consumer’s MSW emissions footprint. Additionally, choosing to recycle, when possible, can offset some of the emissions associated with the lifecycle emissions of goods (herein counted under Industry).

Of the biodegradable (i.e. that which can be anaerobically digested to methane) fraction of MSW, approximately 56% is associated with food and yard trimmings.

Because it is quite possible for a consumer to ensure that minimal organics end in the landfill/dump, the Level 1 emissions agency over MSW is approximately 56%.

Level 2

Per the above discussion, choices that consumers make regarding recycling can offset emissions from the supply chain of Industry. However, these choices are only possible when a community supports recycling practices. Unfortunately, recycling is not as common as one might expect given the publicity… and it is even less effective.

However, where available, the ability to recycle or compost can eliminate the majority of the remaining biodegradable waste stream as well as reduce the emissions footprint of the GAC’s waste dramatically. Consequently, we will ascribe the residual 44% of MSW waste emissions to Level 2.

Level 3

Wastewater treatment is typically a municipal issue. Consequently, the consumer has no direct agency over the emissions profile of their wastewater footprint (aside from choice of where to live). 100% will be ascribed to Level 3.

Buildings

This fraction of emissions is a strange catchall for, primarily, energy use by society. The vast majority of energy is consumed in the context of buildings, and based on reporting frameworks used by major emissions data aggregators, it provides a convenient bucketing for capturing energy use associated with the GAC. Buildings can be broken down into Residential and Non-Residential.

Residential energy use is primarily comprised of electricity and home heating fuel. Home heating fuel varies by geography and property, and can include everything from biomass (e.g. wood burning stoves) to natural gas or electricity. Residential uses of electricity are abundant and diverse, and include illumination, appliances, electronic devices, and air conditioning.

For both electricity and heating fuel applications in the home, we can break down the agency of the GAC into the categories of demand for amount of energy and the emissions intensity of that energy.

Decisions that control the demand for energy include efficiency of appliances, degree of home insulation, use of demand-management optimization tools (e.g. a smart thermostat), geography, etc.

The breakdown of residential building emissions is shown in the figure below.

Note: while cooling only comprises 3% of the GAC’s residential emissions, for wealthy societies cooling can account for a much larger fraction. As the world continues to develop, estimates are that global use of air conditioning could comprise up to 13% of electricity demand (and still yet more emissions from HFCs). Further, even in 2010, 12% of non-residential buildings emissions came from cooling. Cooling is a ticking time bomb for emissions.

The majority of us have some degree of agency over our residential energy use. It is challenging to identify the maximum amount of agency we might have over our residential footprint. In theory we could all choose to live in darkness in caves without any energy demand… but to avoid absurd speculation we will instead evaluate a series of potential interventions that one could execute.

Level 1

We will consider three interventions: changes to the building envelope (i.e. how insulated and optimized for energy management is the home itself?), changes to major appliances, and the use of demand management (i.e. a smart thermostat).

Thankfully, the Department of Energy has done a more thorough job assessing the energy savings potential of residential buildings than we could hope to accomplish here, published in their Quadrennial Technology Review.

Source: DOE Quadrennial Technology Review

Combining improvements in building envelope and appliance efficiency upgrades, the following reductions in energy demand are enabled by the ‘best available technology’ meaning what is accessible to a consumer to purchase if desired.

Cooling: 61%

Source: DOE Quadrennial Technology Review

Heating: 100%

Source: DOE Quadrennial Technology Review

Illumination: 93%

Source: DOE Quadrennial Technology Review

Appliances: 30%

Refrigerators, washing machines, dryers and similar equipment have large potential for reduction. Cooking devices are much more challenging to improve (especially for gas-based systems) with today’s technology. However, electrified cooking appliances do have potential for reduction. The 30% number will be used across appliances for our analysis, but this is no doubt a rough assumption.

Hot Water: 50%

Many of the same improvements that affect the heating and cooling loads can influence the hot water energy requirement. Additionally, the hot water heater itself and demand for hot water can be managed more directly.

If we assume a 1:1 relationship between energy saving and emissions reduction, the above-described interventions amount to ~64% emissions savings.

Demand management (i.e. modulating demand vs. time to optimize for energy efficiency and cost) in the form of smart thermostats can reduce the demand for energy in the home by ~15%.. When compounded with the efficiency savings calculated above, this results in a total Level 1 agency over residential buildings emissions of ~69%.

Level 2

There are many additional hypothetical interventions that require some degree of 3rd party support. One could acquire home solar and a solar- or geothermal-heating system. Once could even choose to move to a high-density urban environment with efficient construction and low-emissions intensity grid. While it may not be practical for all parties, it is now possible as a consumer to effectively reduce the emissions intensity of home energy use to close to zero should agency be exerted. Consequently the remaining 31% of residential buildings emissions will be assigned to Level 2.

Level 3

While we could exert agency over where we live (and, consequently, the grid emissions intensity and typical heating fuel / heating & cooling requirements of a building) and thus impact commercial building use associated with our lives to some degree, this is considered a very low-grade level of agency and in the modern global economy most products are not produced locally enough, nor is the information about emissions intensity available (similarly to the Industry footprint), for this to capture a large fraction of the commercial buildings energy footprint. Thus, the entirety of commercial building energy use will be considered Level 3.^*There are additional Level 3 interventions that affect the emissions profile of the buildings sector, such as grid emissions factor

Summary

The following chart summarizes the Level 1 (light blue) Level 2 (dark blue) and Level 3 (gray) agency that the GAC has over their emissions footprint.

In total this amounts to 29% Level 1, 19% Level 2, and 52% Level 3 agency for our GAC. Level 1 + 2 agency is about 48%, which is approximately what we calculated with our top-down assumptions at the beginning of this post.

This is substantial agency. The interpretation is that nearly 50% of the emissions of our species could be reduced through choices made by individuals that do not require systemic shifts. Of course, these choices may require lifestyle changes, expenditures, and time that many cannot (or are not willing to) sustain. But perhaps we should find some reason for optimism in this analysis – we have more control than might be intuitive. For at the core of anthropogenic climate change are anthropic choices – it is the summation of our individual actions that results in our species footprint.

The charts below show the GAC’s Level 1 and Level 1 + Level 2 agency in order of magnitude by sector.

Of course, depending on your specific geography, the interventions with the greatest impact may vary (e.g. in the United States the Transportation footprint over which the consumer has agency is significantly higher than that of the GAC).

In future posts the specific actions that individuals can take to reduce their footprints will be discussed, as will the means by which we can all exert some agency over our Level 3 emissions (hint… vote). However, from this analysis we can begin to understand where to concentrate our efforts for reducing our emissions footprints.

We will also discuss other loci of agency – in this post we considered the individual consumer as the agent. However, if we consider corporations as agents the above analysis will show a different fraction of emissions associated with Level 1, 2, and 3 agency. Similarly, other agents in society (such as politicians) have other ‘slices’ of the agency pie. To be continued…

References