Which emissions belong to us? The case for contributory value-chain emissions accounting

Abstract

States and other climate actors now commonly set ‘net zero’ targets – pledging that, by a certain date, they will put no more carbon into the atmosphere than they take out. However, there is controversy over what exactly should count as attaining such targets. The method of emissions accounting that states currently use – territorial emissions accounting – is often criticized as problematic, but a fully satisfactory explanation of the problem is needed. We argue that the key both to understanding the problem and to solving it is that there is a specific kind of cooperative contribution – a mereological contribution – that a national net zero commitment needs to make to global climate action in order to meet the standards that global citizens should hold states to. On this basis, we propose an alternative method – Contributory Value-Chain Accounting – showing how it can be qualified to reflect trade relationships between net-zero-aligned countries, and defending it against the objections that it involves double counting, is infeasible and too informationally demanding, and that it would send the wrong incentive signals.

Introduction

States, cities, firms, universities and many other climate actors have increasingly been adopting net zero emissions targets – declarations of intent to ensure that by a certain date they will make no net addition to the global atmospheric concentration of CO₂ and other greenhouse gases (GHGs). ¹ In doing so, they commit themselves to attaining a state of balance between GHG emissions and removals. However, there is no agreed standard for exactly what counts as achieving this balance: different climate actors count different activities as GHG-emitting, different activities as GHG-removing, and their net zero commitments vary accordingly. This is troubling, since ‘net zero’ commitments that rely on an unduly narrow understanding of which emissions belong in a net zero ledger or an overly expansive understanding of what is emissions-removing invite suspicion that they create only an illusion of climate action. This raises a pressing ethical question: just what standards should govern legitimate emissions accounting for the purpose of assessing progress towards net zero?

This article narrows that question in two respects: it focuses on the case of national states and on the emissions (not removals) side of the net zero ledger. We ask:

What criterion of emissions allocation should we use in assessing whether a state has attained net zero?

The ‘we’ in this question refers to global citizens, holding national governments to account for their climate action. Whether states themselves should use the same criterion is a separate question, which we return to later. States typically use a standard method – territorial emissions accounting – which would work fine, in a world in which every state was securely committed to attaining net zero as measured by that method. But since we do not live in that world, that method is problematic. Our aim is to explain exactly what the problem is, ethically speaking, and what emissions accounting system is needed to avoid it.

After setting out our framing assumptions and examining the three most prominent proposals for emissions accounting in the climate economics and policy literature, we will argue that the key both to understanding what makes those proposals problematic and to formulating a proposal that solves the problem comes from seeing that there is a specific kind of cooperative contribution – we call it a mereological contribution – that a national net zero commitment needs to make to global climate action for it to meet the standards that we (global citizens) should hold states to.

We then draw on this to make the case for our positive proposal: Contributory Value-Chain Accounting (CVCA). Our proposal has affinities with a prominent type of proposal for corporate emissions accounting – that best practice should include not only Scope 1 (direct) and Scope 2 (power-consumption) emissions, but also (at least some) Scope 3 emissions that lie further upstream or downstream in the economic ‘value chain’ in which a firm participates. ² Proposals for value-chain emissions accounting for states have also been advanced by some economists. ³ But while value-chain emissions accounting may seem intuitively attractive, the climate change literature lacks both a systematic articulation of the philosophical case in its favour and a clear account of the criteria determining which value-chain emissions should be included. We aim to fill these gaps, explaining why a form of value-chain emissions accounting properly applies to national states, showing how it can be qualified to reflect trade relationships between net-zero-aligned countries, and defending it against some objections to which it might seem vulnerable, including misgivings about double-counting, infeasibility, informational demandingness, and perverse incentives.

Context and framing assumptions

We begin by drawing attention to some significant features of the context in which our question arises, and setting out the framing assumptions we make in answering it.

First, anthropogenic climate change is itself an ethical issue, in large measure because of the burdens it imposes on climate-vulnerable people, now and in the future: the harms caused by the increased frequency and severity of extreme weather events; the prevalence of serious diseases and malnutrition; the risks imposed on those who are not actually harmed; and the other threats to livelihoods and ways of life that are posed by sea-level rise, desertification and other environmental changes. ⁴

Second, climate change mitigation requires collective action. ⁵ Currently, the vehicle for coordinating global climate action is the United Nations Framework Convention on Climate Change (UNFCCC), and the goal towards which global cooperation is directed is to reach net zero GHG emissions globally by mid-century. ⁶ Attaining global net zero (GNZ) requires reaching and sustaining a state in which there is no net anthropogenic addition of long-lived GHGs, such as CO₂, to the atmosphere, with any CO₂ emissions thereafter being matched by equivalent CO₂ removals. ⁷ Most of the world's states now participate in the UNFCCC's annual Conference of Parties meetings, through which they communicate their commitment to ‘nationally determined contributions to the global response to climate change’ (NDCs). ⁸ Either through NDCs or other policy declarations, many countries have now committed themselves to national net zero (NNZ) targets as their contribution towards reaching GNZ. ⁹

The recognition of these two points has encouraged a common way of framing the issue of emissions accounting: the Responsibility Framing. ¹⁰ This sees the proper task of emissions accounting as identifying the emissions for which a party is responsible, in a way that reflects an equitable distribution of the costs of climate action. However, this framing is problematic, for two main reasons. First, it does not allow us to make the following sensible claim. A just distribution of responsibilities in contributing to the common goal of GNZ by 2050 is not one in which every country attains NNZ by 2050: instead, it requires some countries to go further (maintaining net negative emissions) and does not require others to go that far. ¹¹ Saying this can only make sense if measuring NNZ and assessing climate responsibilities are two different things.

The other reason is that not all of the emissions for which a state can be held responsible are properly recorded on its own emissions ledger. To illustrate this, consider:

Outbidder. Wealthy state Postora outbids its competitors in the market for the scarce minerals needed to make the transition to a low-emissions economy. As a result, Postora makes the transition, while other less wealthy states remain dependent on high-emissions forms of production, sending global emissions up.

If Postora's actions impede progress towards GNZ, that could be something for which it is responsible, both causally and morally. But that does not mean that the competitors’ emissions should be recorded on Postora's ledger. We take this to be a point of consensus in the emissions-accounting literature: there is disagreement about just which value-chain emissions should be recorded on a country's ledger, but not about whether the emissions of competitors, which are not part of value chains in which a country's economic actors participate, should be recorded on its ledger.

For those two reasons, we should reject the Responsibility Framing: the emissions to be recorded on a country's emissions ledger cannot be identified with the emissions for which it is responsible. To avoid the first problem, we should make a distinction. Within any collective enterprise in which one of the parties declares that it will make contribution C towards the attainment of a common goal G, two questions can be asked:

The Fair Share Question. Is C that party's fair share of what we all need to do to reach G?

The Attainment Question. What is required for that party to make contribution C?

Our approach to emissions accounting treats it as an answer to the Attainment Question, not the Fair Share Question. The latter is, of course, of critical importance. And we do not think that every country's fair share of GNZ is NNZ. But the Fair Share Question is not our focus in this article. Instead, we seek to know: when a state does make an NNZ declaration, how should we (global citizens) assess whether NNZ really has been attained or whether a state has made progress towards attaining it? How should NNZ be measured, as a contribution to GNZ?

In approaching these questions, we make some further framing assumptions. States are morally accountable for their climate actions, to three different kinds of agent. ¹² They are accountable to the climate-vulnerable for reducing their imposition of burdens on them. They are accountable to each other, for joining the global effort to reach GNZ, and for making their promised contribution to that collective effort. And they are accountable to their citizens (whether climate-vulnerable or not) for acting on their behalf. They are accountable to these three groups for climate actions of various kinds. One such action is sustaining net positive emissions, but there are various others: depriving other states of the resources they need to reduce their emissions, failing to share green technology, obstructing international climate agreements and so on. In Outbidder, Postora is accountable for causally contributing to making global emissions go up, but not for having net positive emissions itself.

When we say that states are ‘accountable’ to these groups for their climate actions, what we mean is that these groups have moral claims that states do or do not act in certain ways. We leave the concept of a ‘moral claim’ unanalysed, working with the standard idea that when someone has a claim that one do A or not do B, they are wronged if one fails to do A or does B, absent a valid justification. So ‘accountability’, as we are understanding it, is not all-things-considered culpability for having done something: it is a matter of there being a moral case (but perhaps an outweighed moral case) for or against doing something. Speaking this way, when a just distribution of climate responsibilities permits developing state X to sustain net positive emissions, X is accountable for the emissions burden it is imposing, but it has an adequate justification for doing so.

Thus, we approach our question as follows. When a state acts in a way that carries net positive emissions, that is one way of acting for which a state may be morally accountable (and for which it may or may not have a good justification). Our question is: what is it for a state's actions to ‘carry net positive emissions’, and therefore to be subject to that moral claim?

We add one further framing assumption: although many states have made NNZ commitments, none can be securely relied on to meet those commitments. For that, a state would have to have a clear and credibly detailed implementation plan for reaching net zero, with interim targets, and a transparent reporting process providing evidence of meeting those targets: it is far from clear that any existing state currently meets those conditions. In the “Net zero alignment” section, we will explore the implications of our proposal in a world in which states could be securely relied upon to meet their NNZ commitments.

Many other ethical questions can be asked about NNZ commitments. Which states morally ought to make them, all things considered, and why? Which states ought to go further, which can justifiably go less far, and what ought to be done to cover the resulting shortfall? What can legitimately be claimed as an emissions removal? Given that declaring the target of reaching NNZ at some future date (e.g. 2050) leaves open a range of different pathways, in which the bulk of emissions reduction happens earlier or later, what pathway to NNZ ought a given state to adopt? ¹³ And when we reach GNZ, the atmosphere will contain excess emissions that will have to be removed somehow: How should that task be allocated? These are all questions of great importance, but they fall outside the scope of this article.

Proposals for emissions attribution

Which emissions should we attribute to a state in assessing its progress towards net zero? An answer to this question, has to allow for the complexity of the relationships through which goods and services are produced, traded, supplied and used – with emissions being generated at all these points. Consider this simple example:

Mobile Phones. In response to consumer demand in state A, a firm in state B produces mobile phones for export to A. The production facility in B is powered through the burning of coal purchased from C. Some of the components are sourced from D, and the finished product is transported through E for delivery to A. ¹⁴

Several agents are involved in this economic activity – they all contribute causally, in one way or another, by initiating or helping to sustain processes that release emissions. So which emissions should we assign to the national inventories of A, B, C, D and E, or distribute among them in some measure?

Territorial emissions accounting

According to the accounting method assumed in the current practice of states, all and only those emissions that are released in a state's territorial jurisdiction are attributed to it – excluding so-called bunker fuels, which are used in aviation and the operation of some maritime vessels. ¹⁵ Applying this to Mobile Phones, we would attribute the emissions from the operations of B's factory to B, the emissions from C's coal mines to C, the emissions released in producing the components in D to D, the emissions released during transport across E to E, and to A would be attributed any emissions resulting from the use and disposal of the phones themselves within its territory. This practice is often referred to as ‘production-based’ emissions accounting, but we will use the more accurate label ‘territorial emissions accounting’. ¹⁶ We’ll say that a country that reaches net zero as measured by this accounting practice attains territorial net zero (TNZ).

As several critics point out, territorial emissions accounting has a significant drawback. ¹⁷ A state can attain TNZ, or make progress towards TNZ, without contributing towards the attainment of GNZ. This can happen through what economists commonly call ‘carbon leakage’, where domestic emissions reductions resulting from policies in one country lead to a transfer of production to other countries, often with less stringent environmental policies. ¹⁸ This is illustrated by:

Greened Production. State X restructures its national economy by shutting its consumer goods factories and meeting domestic demand for those goods through imports from state Y. Y's consumer goods factories are more emissions-intensive than X's were.

Here, territorial emissions accounting says that X's emissions have declined. But this way of attaining NNZ should not be counted as a contribution towards GNZ. In making this economic restructure, X does nothing to reduce the emissions produced in meeting its consumer demand: it merely displaces them. Indeed, if Y's production processes are more emissions-intensive than X's were, this could take us further away from GNZ rather than closer to it. ¹⁹

This objection raises some questions that warrant closer examination. Exactly what is meant by saying that when restructuring a state's economy ‘merely displaces’ emissions it is not a contribution to GNZ, and exactly what features must a method of measuring NNZ possess to avoid this objection? And how could it make sense, in Greened Production, to count the emissions released in Y on X's ledger, if the extra emissions in Outbidder do not belong on Postora's? We will return to discuss those questions in detail in §III, but let us first consider two other prominent types of proposal that are motivated by the ‘carbon leakage’ challenge for territorial emissions accounting.

Consumption-based accounting

Consumption-based accounting attributes all and only the emissions that are ‘embedded’ in goods and services to the state in which they are consumed. ²⁰ As applied to Mobile Phones, the attribution of emissions is now much simpler: all the emissions released at all points in this process are attributed to A, where the goods are consumed.

Consumption-based accounting is not vulnerable to the objection we made to territorial emissions accounting. ²¹ It does not allow the emissions released in producing a consumer good to be removed from a state's ledger merely by displacing them offshore, upstream in the value chain. In Greened Production, X's shift from producing goods for domestic consumption to importing the same goods from Y, where they are produced more emissions-intensively, would be registered by a consumption-based scheme as a net increase in X's emissions.

However, consumption-based accounting is vulnerable to an analogue of the objection to territorial emissions accounting. It too allows a state to qualify as making progress to NNZ by merely displacing the same economic activities across international borders, without contributing to GNZ. This time, the problematic form of displacement involves the relocation of consumption rather than production. ²² Consider:

Greened Consumption. X has been producing its consumer goods in two ways: through a first set of low-emissions production processes and a second set of high-emissions processes. It restructures its national economy, greatly increasing its high-emissions goods production, exporting those goods for consumption in other countries, and using the income to finance its own consumption of low-embedded-emissions goods.

This would count as progress towards consumption-based net zero (CBNZ); but, again, it should not be counted as a contribution towards GNZ. It is directly putting more GHGs into the atmosphere than previously, through production for trade with other countries. As before, this economic restructure could even drive global emissions up. ²³

Input-based accounting

Territorial emissions accounting locates the site of emissions attribution where emissions are released in producing or using a product; consumption-based accounting locates it at the downstream end of the product's value chain, with the consumer. The other main approach instead locates the site of attribution at one or more locations upstream in the value chain. ²⁴ We’ll classify the different variants of this idea together as forms of ‘input-based accounting’. The simplest is extraction-based accounting, according to which the emissions entered on a state's national inventory are those released through the use of the fossil fuels extracted from the territory under its jurisdiction. ²⁵ As applied to Mobile Phones, this would assign to the coal-exporting country C all of the emissions released when its coal is burned in B; and the other value-chain emissions – those released when making the components in D, assembling the phones in B, transporting them through E, and using them in A – would need to be attributed by tracing them back to the states from whose territory the fossil fuels were extracted to supply those various processes.

Extraction-based accounting would prevent a state from claiming progress towards NNZ merely through the displacement of production or consumption, and thus would avoid the objections we made above to territorial and consumption-based accounting. No matter where the coal extracted by C is burned, or the products it is used to manufacture are consumed, the emissions are still allocated to C.

However, while extraction-based accounting covers the emissions from some important sources, it misses others (e.g. agriculture and industrial processes such as cement production). ²⁶ To rectify this, environmental economists have made a more general proposal: income-based accounting, which takes all of the primary factors of production (land, labour, capital and entrepreneurship) that are upstream from the emissions-releasing use of a product, and allocates the emissions in proportion to the income generated by those inputs. ²⁷ This method can therefore be applied to all anthropogenic emissions, not just those produced from fossil fuels, and to all the economic inputs that enable those emissions, extending beyond extraction to encompass transport, finance and other services. ²⁸

However, neither of these proposals avoids a version of the displacement problem. While extraction-based accounting is not sensitive to the displacement of production or consumption, it is sensitive to the displacement of extraction. This can be seen in:

Greened Extraction. X has been extracting coal to supply its coal-fired power stations. It restructures its economy, shutting down its own coal mines but not its power stations, which it now supplies with coal imported from other countries. It simultaneously expands its emissions-intensive production facilities.

X has now attained extraction-based net zero – yet, intuitively, this should not be counted as a contribution towards GNZ. It does nothing to lessen the additions X makes to GHG concentrations through burning fossil fuels. Moreover, income-based accounting is sensitive to the displacement of production in the same way as territorial emissions accounting. In Greened Production, when X switches to consuming emission-intensive imports from Y, it makes progress towards income-based net zero without contributing to GNZ. So, again, neither of these proposals supplies a method of emissions accounting for which attaining NNZ according to that method is sufficient to ensure a genuine contribution to GNZ.

What is the problem with displacement?

We have raised essentially the same objection against territorial, consumption-based, and input-based accounting methods: each allows a state to claim NNZ, when all it does is displace emissions in a way that does not contribute to GNZ. But let us now examine this objection more closely. Exactly what is problematic about the forms of emissions displacement that these methods allow?

The answer might seem obvious: these methods all allow national emissions to count as going down through doing things that cause global emissions to go up. However, on closer examination, this answer should be rejected. It mishandles Outbidder, where Postora's contribution to causing additional emissions does not put those emissions on its own ledger. And if that claim about Outbidder seems questionable, then consider:

Independence. Solitaria has previously used low-emissions processes to produce consumer goods for export, while consuming emissions-intensive imports. It now restructures its economy, producing its own consumer goods using emissions-free processes and withdrawing from the export market. New emissions-intensive producers enter the market vacated by Solitaria, with the result that global emissions go up.

Here, too, there is carbon leakage: an increase in global emissions is a causal consequence of Solitaria's actions. But surely that should not stop us from saying that Solitaria's own national emissions have gone down: the extra emissions should be attributed to the new market entrants, not to Solitaria. ²⁹

So: how can the problem with emissions displacement be explained, without mishandling cases like Outbidder and Independence? Our answer is this: the problem with emissions displacement is not that it fails to make a causal contribution to GNZ; it is that it fails to make a mereological contribution to GNZ.

To explain this, we need to start by clarifying the idea of a mereological contribution to a common goal, and then apply it to national emissions accounting.

Mereological contribution

When the members of a group contribute towards a common goal, a mereological contribution (we stipulate) is a contribution that constitutes a part of the attainment of that goal. This is not the only form that contributions to collective action can take. For example, suppose a community bands together to protect their town from a wildfire, with some people maintaining the fire trucks, some cooking meals, and others going out to fight the fires. Here, the contribution that the firefighters make towards the attainment of the collective goal is mereological, but the contribution of the cooks and mechanics is not: the cooking of meals and maintaining of trucks enables, but does not constitute a part of, putting out the fires.

Common goals and mereological contributions to them can belong to significantly different kinds. Sometimes, the common goal is to produce a good; sometimes, it is to solve a problem. And when the common goal is to solve a problem, we can distinguish between initiative cases, where the problem is not caused by us and solving it requires new actions (as in fighting a wildfire); and inhibition cases, where the problem is produced by our existing actions, and solving it requires that we modify our actions to stop producing it (as in fighting pollution or environmental degradation). In both of these kinds of case, an action can fail to make a mereological contribution to the common goal when it merely displaces the problem we are trying to solve.

To see this, we can consider a simple initiative case:

Snow. After a snowfall, the neighbours along the street agree to clear the footpath for pedestrians. As your contribution, you say you will clear the path in front of your own house.

Here, a Fair Share Question can be asked; and perhaps (if you are infirm, and a lot of snow has drifted onto your path) doing this is more than your fair share, perhaps less. But let us focus instead on the Attainment Question: what should count as making the contribution you said you would make? Suppose you proceed as follows:

Dumping. Your neighbour, Eira, clears all the snow from the path in front of her house. You then shovel the snow from your path onto Eira's, explaining: ‘When I said “I’ll clear the snow from my path”, I made no commitment about where I would clear it to’.

Obviously, that is unsatisfactory. Is the problem just that Eira hasn’t consented to your dumping? No – consider:

Deal. You pay your other neighbour Lumi for her permission to shovel your snow onto her path. However, Lumi has not joined the neighbourhood agreement, and you know she will leave your snow there without clearing it.

In both Dumping and Deal, you have done something that meets the description ‘clearing your path’; but the standard to which the other snow-shovellers can hold you accountable is determined not by your thoughts about the commitment you have made, but by what makes a mereological contribution to the common goal. Here, your mereological contribution can be calculated by quantifying the total volume of snow that needs to be removed from the whole footpath, and asking what reduction you have made to that total. Merely displacing some snow from one part of the footpath to another makes no such reduction: you have left just as much snow to be removed by others as if you had done no shovelling at all. This way of ‘clearing your path’ is not a mereological contribution to the common goal.

Failing to make a mereological contribution is not the only way of being uncooperative towards your neighbours in Snow. Another way is illustrated by:

Shovels. You clear all the snow from the path in front of your house into your front yard, then buy up all of the snow shovels from the local store, offering to sell them to your neighbours for ten times the price you paid for them.

Here, you have made a mereological contribution to the common goal: in clearing your path, you do reduce the amount of snow that remains to be cleared. But you have made it more costly for others to make their mereological contributions. We don’t claim that this kind of unhelpful action is any better, overall, than Dumping or Deal; but it is different. If the neighbourhood is interested in accounting for how much snow-clearing each neighbour has contributed towards the common goal of clearing the footpath, the answer is none in Dumping and Deal, but not in Shovels.

Now let's turn to inhibition cases – those where a problem is being produced by our own actions, and we need to modify those actions to solve the problem. One common kind of inhibition case concerns the performance of privately beneficial actions that carry external costs – as in cases of pollution. To solve a problem of that kind, we need to modify our actions in one of three ways: not performing the externality-producing actions, preventing the actions from carrying the externalities, or cleaning up after ourselves. We can illustrate this with:

Smoke. We heat our houses with coal fires; the smoke creates air pollution that causes respiratory disease. As your contribution to solving the problem, you say you will stop burning coal to heat your house.

Inhibition cases like this require a different way of measuring mereological contributions. Since the problem is created by our actions, we cannot assess the total magnitude of the problem independently of everyone's conduct and then measure mereological contributions as reductions to that total (as we did in Snow). Instead, we have to measure each agent's mereological contribution to the problem, and then quantify their mereological contribution to the solution as the opposite (the ‘additive inverse’) of that. If you emit 1 tonne of pollutants, your mereological contribution to solving the pollution problem is −1 tonne. ³⁰

Here, too, it is obvious that there are forms of displacement that make no mereological contribution to solving the problem – for example:

Short Pipe. You pay your neighbour to heat your house by burning coal in his fireplace, directing the smoke up his chimney, and supplying heat through a pipe to your house.

Now you have managed to heat your house while being able to say that you are ‘not burning coal’; but here, too, your action involves a form of displacement that fails to make a mereological contribution to the common goal. It does nothing to eliminate the external costs that are carried by your benefit-procuring action (the heating of your house). You have simply displaced the location at which the external cost is produced.

Moreover, the same point can apply when you are not yourself cooperating directly with the cost-producer. Consider next:

Longer Pipe. You pay your neighbour Kemuri to heat your house. Kemuri pays his other neighbour Reykur to burn coal in Reykur's fireplace, directing the smoke up Reykur's chimney, and supplying heat through a pipe that runs from Reykur's house through Kemuri's to yours.

Here, you are cooperating with Kemuri, not Reykur. But obviously, the move from Short Pipe to Longer Pipe makes no contribution to solving the problem we confront in Smoke. You are part of a cooperative network that merely separates the location of the benefit you receive from the location at which the externality is directly produced. This would remain true however far away the coal that heats your house is burned. The problem we are seeking to address can be constituted either by a collection of independent agents each of whom secures a private benefit by acting in a way that directly imposes an external cost, or by agents who secure the same benefits by participating in a cooperative network that supplies the benefit while locating the external cost elsewhere. Switching from the first way of securing the benefit to the second leaves the problem unchanged, so it makes no mereological contribution to solving that problem.

Notice next that the problem of external costs attaches to production activities, as well as consumption activities. To illustrate, we can add a further backstory to Short Pipe. Kemuri used to work as a blacksmith, burning one sack of coal per day to make horseshoes, which he sold to make an income. His production activity carried an external cost – the contribution he made to air pollution by burning his daily sack of coal. You now offer to pay him the same income if he switches from making horseshoes to heating your house: he accepts, and Short Pipe is the result. Here, Kemuri has switched his production activity (from producing horseshoes to producing heating); he gains the same benefit from that production, with the same external cost. So he, too, makes no mereological contribution to solving the air pollution problem by changing his production activity.

Thus, in inhibition cases, mereological contributions to an externality problem can be made by both consumers and producers, and those contributions are not lessened (hence, mereological contributions to solving the problem are not increased) by displacement across a cooperative network. However, again, mereological and causal contributions to solving our common problem must be distinguished. We saw the need for this distinction in initiative cases, as illustrated by Shovels. The same holds for inhibition cases. Consider:

Clean Monopoly. You monopolize the supply of smokeless coal, using it to heat your own house while raising the price at which it is available to everyone else. As a result, other households switch to using dirtier coal.

Here, too, your action unhelpfully adds to the problem we are aiming to solve. But your contribution is causal, not mereological. The problem is constituted by the burning of dirty coal to heat other houses (whether it is burned in those houses or elsewhere). It is the others whose contribution to the problem is mereological; yours (which may be no less bad) is causal. In ceasing to heat your own house in that way, you make a mereological contribution to solving the problem – while also doing something (monopolizing the smokeless coal supply) that makes a causal contribution to worsening it.

The application to emissions displacement

We can now apply these lessons to the case of emissions accounting.

When, as a global community, we set ourselves the common goal of attaining GNZ as a solution to the problem of rising emissions, we face an inhibition case, not an initiative case. GNZ should be thought of not as merely a state-of-affairs-at-a-time (there being no more emissions than removals on 1 January 2050, say), but instead as an ongoing pattern of global economic activity in which we no longer generate net positive emissions. The common goal is to solve a problem that we are creating through actions that secure private benefits while carrying external costs: the costs of global warming. Accordingly, when a national state makes an NNZ pledge as its ‘nationally determined contribution’ to GNZ, the standard to which it is properly held accountable – by the climate-vulnerable, other states, and its own citizens – in meeting that commitment is fixed not by its own interpretation of that commitment, but by what counts as a genuine mereological contribution to that goal.

Since the emissions problem shares this structure with other inhibition cases, a mereological contribution to the common goal (GNZ) cannot be attained by displacement, any more than it can in Smoke. When the problem we are trying to solve is a pattern of privately beneficial action that carries external costs, the problem is not altered by displacing the location at which the external cost is produced, within a cooperative network. That is what is happening in the displacement cases described in the ‘Proposals for emissions attribution’ section: Greened Production, Greened Consumption, and Greened Extraction. These, too, are cases in which the external cost associated with a beneficial action (of either consumption or production) is relocated within a cooperative network, through international trade relationships. In those three cases, there is a direct cooperative relationship between a state X and its trade partners, displacing the location of emissions release in one direction or the other across that relationship. The same point applies to forms of displacement in which that location is further removed along a value-chain; as the move from Short Pipe to Longer Pipe, shows, displacing the location at which the external cost is produced to another location in a cooperative network leaves the problem unchanged. So it makes no mereological contribution to solving the problem.

In our earlier illustrations of initiative and inhibition cases, we saw that an action's mereological contribution towards a common goal is not to be identified with its causal contribution: Shovels and Clean Monopoly showed this. The same point can be applied now to Independence and Outbidder. When Solitaria becomes economically independent, the causal consequence is that other emissions-intensive producers enter the market it has vacated; but it is the new market entrants whose privately beneficial and externality-carrying actions constitute the problem that Solitaria is helping to cause. And when Postora outbids its competitors for scarce minerals, then, like the monopolists in Shovels and Clean Monopoly, it is acting unhelpfully; but its unhelpfulness is a matter of impeding the mereological contributions of other agents, not of failing to make a mereological contribution itself.

We wanted an explanation of what is wrong with allowing a state to count as attaining NNZ by merely displacing emissions within a value chain, without carrying the false-looking implication that Independence and Outbidder do not attain NNZ. We now have that explanation. When a state makes a commitment to attaining NNZ, it is ostensibly offering this as a mereological contribution to GNZ. Given the problem to which GNZ is a solution, forms of value-chain emissions-displacement cannot be mereological contributions to solving that problem. So, if an emissions accounting system is going to be adequate to measure mereological contributions to the common goal of GNZ, it will need to include on a state's emissions ledger the offshore emissions that lie upstream or downstream in its value chains.

Our discussion has identified three kinds of economic activity through which a state can mereologically contribute to the problem that GNZ aims to solve – and which therefore need to be included on a ledger that measures its mereological contribution to the solution. The first and most obvious is through territorial emissions. The second is through demand for imported goods with embedded emissions, released upstream in the value-chain. And the third is through the production or supply of goods exported to locations where their use will release emissions, downstream in the value-chain. Therefore, a method of emissions accounting on which NNZ really does amount to an adequate mereological contribution to GNZ must include the emissions to which a state is related in all three of these ways.

Contributory Value-Chain Accounting

Let us sum up the argument so far. When a state makes an NNZ commitment, it is morally accountable – to the climate-vulnerable and to the other participants in the global effort to reach GNZ – for fulfilling its commitment in a way that mereologically contributes to that global goal. But a mereological contribution to GNZ cannot be made by merely displacing the location at which emissions are released to a different location within a value-chain. So when we ask, ‘What method of emissions accounting needs to be used in order for the attainment of NNZ, as assessed by that method, to meet this standard?’, territorial, consumption-based, and input-based accounting methods all fall short.

To avoid this problem, we need an emissions accounting method that includes all three types – territorial emissions, emissions embedded in imported goods and emissions released in the use of exported goods – on a country's ledger. Here is an initial statement of our proposal for doing that.

Contributory Value-Chain Accounting: Unqualified Version

Emissions are attributed to state X's national inventory when either:

Those emissions are produced on X's own territory, or

X pays producers or suppliers in state Y to produce or supply goods or services for import into X, and as part of producing or supplying those goods Y conducts activities through which those emissions are attributed to Y, or

X gains an income from the production or supply, or inputs to the production or supply, of exported products the use of which will produce those emissions.

In a nutshell, CVCA requires a state's emissions ledger to include the emissions attributed to it by territorial emissions accounting and consumption-based accounting and income-based accounting.

We call our proposal Contributory Value-Chain Accounting because it is the form of emissions accounting that is appropriate to assessing mereological contributions to a collective effort to attain GNZ. What is immediately obvious is that CVCA attributes the same emissions to more than one party. If C sells its coal to B, where it is burned to produce mobile phones for export to A, then the emissions are attributed to all three countries. This is an essential feature of the proposal, not a bug (a feature it shares with those corporate emissions accounting proposals that include Scope 3 emissions). If we were looking for an emissions accounting system that allowed us to calculate global emissions by summing national emissions, CVCA would obviously be unsuitable. ³¹ But the purpose of CVCA is different; it gives us an accounting method capable of measuring progress towards the kind of NNZ that could indeed be what it purports to be: an adequate mereological contribution to GNZ. Double-counting was already implicit in the earlier treatment of Short Pipe: there, both you (the heating consumer) and Kemuri (the heating producer) receive a benefit from the transaction, and the same externality attaches to the benefits received by both parties. CVCA correctly reflects the fact that, given the externalities that need to be eliminated to solve our global problem, none of the participants in the value-chains containing those emissions can claim to have reached NNZ, as a mereological contribution to GNZ, while those externalities have not been internalized. As the global community progresses along a pathway to GNZ, the quantity of multiply attributed value-chain emissions will reduce.

Notice that, when applied to a country's upstream emissions – those embedded in the goods it imports – CVCA works recursively. Consider the application to Mobile Phones, where B manufactures a product for export to A, sourcing components from D. By condition 1, the emissions released in producing the components are attributed to D. By condition 2, D's production emissions are also attributed to B, since those emissions come from activities conducted as part of producing and supplying manufacturers in B with the necessary components. And then by a further application of condition 2, the same emissions are also attributed to A, since A is paying B to produce the phones, and as part of producing them B conducts activities (namely, sourcing the components from D) through which those emissions are attributed to B. Thus, CVCA attributes as many embedded emissions to a country through its imports as are captured by repeated applications of condition 2. This accurately reflects the general point that the displacement of an externality to distant locations within a cooperative network does not improve an agent's mereological contribution to solving an externality problem.

CVCA applies to mediated value-chain relationships in the downstream direction too. In this direction, no recursive reapplication of CVCA is required: mediated relationships are covered directly by condition 3. If, in Mobile Phones, C first sells its coal to some intermediate state that then sells it on to be burnt in B, it remains true that C gains an income from the production of a product the use of which in B produces emissions.

We might wonder whether CVCA really has solved the displacement problem we found for the other emissions accounting methods. Just as production-, consumption- and extraction-based accounting allow NNZ to be reached through the displacement of production, consumption, and extraction, won’t net zero as measured by CVCA (CVCNZ) be something a state can reach through the displacement of emissions from one value chain to another?

Yes, it will; but now we no longer face the same problem. To produce value-chain displacement, we need a case with this structure:

Greened Value Chain. Y has attained CVCNZ, relying on the import of scarce minerals (with no embedded emissions). X commits itself to CVCNZ, and in order to attain this, it outbids Y to secure those minerals. The result is that X attains CVCNZ, while Y reverts to fossil-fuel-reliant forms of production.

But now we have a version of Outbidder, and the points we made earlier about that case apply. It is true that X's restructure could causally contribute to raising global emissions, that this causal contribution could be unhelpful to the global goal of GNZ, and that X may be morally accountable for that. But X would be morally accountable for its causal contribution to our not attaining GNZ, not for any failure to make a mereological contribution to our attaining it. We do not claim that the former matters less than the latter; but our point is that they are different, and that the proper role of emissions accounting is to measure mereological, not causal, contributions. In Greened Value Chain, X does make a mereological contribution to GNZ; in the other displacement cases, it doesn’t.

Possible supplementations

CVCA as set out above is the core of a proposal for emissions accounting that can be supplemented in various ways. Since the merits of the alternative possible supplementations can be assessed independently of the case we have made for the main proposal, it is clearest to set it out in the simple form above. However, part of its attraction is that it can be readily supplemented to deal with issues that we identified earlier. We mention three of these briefly now, and a fourth is discussed in the next section.

First, CVCA provides a ready mechanism for implementing the widely acknowledged principle that – given states’ different levels of wealth, different emissions histories, and different levels of capacity – the burdens it is reasonable to require them to bear in contributing to the attainment of GNZ are different; they are bearers of ‘common but differentiated responsibilities’. ³² This could be done within CVCA by agreeing a specified net positive emissions ceiling for those states with a sound justification for it, and then requiring their trade partners to commit to CVCNZ. ³³ The partners making that commitment would in effect be undertaking to absorb the cost of the extra net emission reductions required in order to reconcile the attainment of GNZ with permitting developing countries to maintain a level of net positive emissions.

Second, it also provides a mechanism for dealing with the foreseeable scenario in which some states refuse, without good justification, to commit to NNZ. This could be handled, again, through other states’ implementing CVCNZ. Progress towards meeting those national commitments would still amount to progress towards GNZ, and the extra costs incurred in dealing with the recalcitrant countries’ emissions could be passed on to them.

Third, if GNZ is to be reached, bunker fuels need somehow to be accounted for. There are various proposals for doing this, and most could be combined with CVCA in some way. ³⁴ The most straightforward would be to treat international aviation and shipping as the equivalent of another state in the value chain. To see how this would work, suppose that A imports goods from B by transporting them across international territory using F's freight service. F is selling a product – a freight service – the use of which in international territory produces emissions, which, by the application of condition 3, are attributed to F. They are also derivatively attributed to A by the application of condition 2: A is paying F to supply the imported goods. ³⁵ In addition, they are attributed to B through the application of condition 3, since B gains an income from the production or supply of inputs to the product (i.e. the goods that are transported by the freight service), the use of which in international territory produces emissions.

Net zero alignment

In §I, we made the simplifying assumption that none of the states that have made NNZ commitments can be securely relied on to meet them. What difference does it make if that assumption is false – or becomes false in the future?

To appreciate how this matters, we can return to the Short Pipe example, and add some extra details. Suppose that, first, we have set ourselves the common goal of eliminating air pollution from coal fires by the end of the year; second (the Promise Condition), your neighbour Kemuri promises to transition to burning smokeless coal by the end of the year; and, third (the Reliance Condition), it is reasonable to rely on Kemuri to keep his promise. Under these conditions, you can now say that you are making a mereological contribution to the common goal when you burn no coal yourself and source your heating from Kemuri. And you can say this before Kemuri has completed the transition. Kemuri is in effect saying to you, ‘I will take responsibility for eliminating the external costs associated with my coal-burning by the end of the year’. Given this, there is nothing further that you need to do as your mereological contribution to the common goal, beyond not burning any dirty coal yourself.

In saying this, however, two points need to be emphasized. First, this requires that the Reliance Condition, and not only the Promise Condition, is met. It is true that by promising to remove the externality from the process that generates your heating, Kemuri acquires an obligation to do that. But that is not sufficient to transfer all accountability for securing that outcome to him from you. You are accountable to two groups – those who are vulnerable to air pollution, and other householders who are committed to the same common goal – for eliminating the externalities associated with your heating. In order to be able to justify your conduct to them, it must be true not just that someone has said he will eliminate them for you, but that it is reasonable to rely on him to actually do so.

The second point is that in a case like this, we must carefully distinguish the destination from the pathway. Under the conditions just described, I can say that I have made my mereological contribution to the common goal of eliminating our air pollution by the end of the year, and this can be true when in January the Short Pipe arrangement is set up, even if Kemuri's transition to smokeless coal will only be completed in December. However, this does not remove my accountability for the smoke released during the intervening year in heating my house. If there is a less polluting pathway I could have taken, I am accountable for not doing that. I am accountable for the external costs carried by my actions (via the cooperation with Kemuri) before the common goal is reached; and if (once it is reached) residual air pollution needs to be removed, some of which is attributable to me, I may be responsible for removing it.

Returning now to GHG emissions, the corresponding points apply. When state X makes a commitment to reaching NNZ by date D (thereby meeting the Promise Condition) and it is reasonable to rely on X to do so (thereby meeting the Reliance Condition) let us say that X is ‘NZ-aligned’. Of course, this raises the question of what exactly it takes for a state to meet the Reliance Condition. We won’t attempt a full answer to that here. As we said in the ‘Context and framing assumptions’ section, we think this would include at least that X has a clear and credibly detailed implementation plan for reaching net zero, with interim targets, and a transparent reporting process providing evidence of meeting those targets; and that no state currently meets those minimum conditions. However, suppose that either there are now NZ-aligned states or that, in future, states become NZ-aligned. What then?

Then we should treat interactions with those states in the same way that we treat interactions with Kemuri in the case just described. To illustrate this, consider:

Early Adopter. Every state in the global community is securely committed to TNZ by 2050. However, states vary in their capacities for transition to a decarbonized economy and hence are on different trajectories towards this target. One country, Viridia, is the first to reach TNZ. It exports electric vehicles to other countries that have not yet fully transitioned their electricity grid to renewable sources of generation.

Here, Viridia can say not only that it has reached TNZ, but that it has completed its mereological contribution to the global goal of GNZ and so has genuinely attained NNZ. Since its value-chain trade partners are NZ-aligned (with a 2050 target date), Viridia has done everything that is needed from it for us all to reach GNZ by 2050: its mereological contribution to that common goal has already been completed.

The same applies to any exports that Viridia makes to NZ-aligned states during their period of energy transition – including fossil fuels. Provided the other parties along its value chains really are themselves on a pathway to TNZ, Viridia is not impeding the attainment of GNZ by the target date. If those parties are phasing out their use of fossil fuels on a trajectory to that destination, Viridia's attaining TNZ itself is sufficient for it to have completed its own mereological contribution to GNZ.

The same point applies not only to the downstream value-chain emissions released in using the products a country exports, but also to the upstream emissions embedded in its imports. If an NZ-aligned state is currently releasing emissions as part of producing or supplying goods to Viridia, those emissions need not be attributed to Viridia in assessing Viridia's own progress towards NNZ. The reason is the same: Viridia has done everything that is needed from it for us all to reach GNZ, and is just waiting for its trade partners to modify their actions in the ways they have reliably promised to do.

However, this is only plausible if, as before, we observe the distinction between destination and pathway. If the other parties in Viridia's value chains are NZ-aligned, and Viridia has reached TNZ, that suffices for it to have completed its mereological contribution to the destination of GNZ. But it does not remove its accountability for the actions it takes on the pathway to that destination. As we pointed out in the ‘Context and framing assumptions’ section, our question ‘What should count as attaining NNZ?’ is different from the question ‘What should be done about the emissions released on the way to GNZ?’. When we reach GNZ, we will then have the further task of removing the excess GHG that was added to the atmosphere during the pathway to that destination. In discharging that task, we will face the latter question. And in relation to that question, Viridia is plausibly accountable for remedial action in relation to the emissions released in burning the fuels it exported. Viridia need not include those emissions on a ledger that assesses whether it has attained NNZ; but it remains accountable for how it attains NNZ. ³⁶

There is one further point to notice. When a party meets the Promise and Reliance Conditions, the scope of the promise is important. Suppose that Kemuri's promise is that he will transition to burning smokeless coal, but the scenario is Longer Pipe, where he sources your heating from his dirty-coal-burning neighbour Reykur, who has made no such promise. Then the externality associated with the benefit I receive does not fall within the scope of Kemuri's promise. Since no one else has taken responsibility for eliminating that externality, I cannot now say that I have done all I need to do in order to make my mereological contribution to the common goal.

This point also extends to emissions accounting. Currently, states employ territorial emissions accounting, not CVCA. So, when a state makes an NZ commitment, it is TNZ that figures in its target. Given this, when A invokes B's secure NNZ commitment to justify excluding from A's own inventory the value-chain emissions embedded in the goods it imports from B, the justification only covers B's territorial emissions. If C and D, situated further upstream in the same value chain, have made no such commitment to reduce their territorial emissions, then those emissions, which are also embedded in the products A imports from B, must still be included in A's inventory. If, however, B were securely committed to CVCNZ, A would not then need to add to its own inventory any emissions from the value chains in which B participates, since those would be covered by B's commitment.

To accommodate these points, our statement of CVCA needs to be qualified. If it is to have fully general application – application not only to our current situation, in which no state is NZ-aligned, but possible future situations in which some states are NZ-aligned – then it must be formulated in a way that excludes from a state's emissions ledger those value-chain emissions that are covered by the NZ commitments of NZ-aligned states. We can do so as follows:

Contributory Value-Chain Accounting: Qualified Statement

Emissions are attributed to state X's national inventory when:

Those emissions are produced on X's own territory, or State Y either is not NZ-aligned, or is TNZ-aligned and the emissions are not produced on Y's own territory, and either

X pays producers or suppliers in Y to produce or supply goods or services for import into X, and as part of producing or supplying those goods Y conducts activities through which those emissions are attributed to Y, or

X gains an income from the production or supply, or inputs to the production or supply, of products the use of which in Y will produce those emissions.

This qualifies the earlier statement of CVCA by adding the italicized clause. It has the effect of excluding from C's ledger any value-chain emissions that are covered by the NNZ commitments of other parties in its value chains.

Objections and clarifications

Before concluding, we briefly describe and respond to some objections we have commonly encountered in discussing CVCA. It will help to frame our treatment of them to recall our opening question:

What criterion of emissions allocation should we – global citizens – use in assessing whether a state has attained net zero?

Our claim is that CVCA is the standard that should be used for this purpose. This is compatible with thinking that, given the current political context, states should continue to practise territorial emissions accounting; and we do indeed think that. The emissions accounting method that states should use at a given time is that which has the best prospect of supporting equitable progress towards GNZ; and there is a danger that the more demanding accounting standard set by CVCA would make it demotivating to adopt it.

Nonetheless, we think that the idea that states ought to adopt CVCA at some future time should be taken seriously. There are several possible futures in which CVCA would become redundant: these include a future in which every country becomes TNZ-aligned (in which case TNZ would be an adequate mereological contribution to GNZ), one in which a global carbon price is adopted, and one in which the global goal of GNZ is abandoned (perhaps in favour of a geoengineering solution to climate change). But as long as the global community retains the goal of GNZ while not every country is NZ-aligned, we will need some way of ensuring that pathways to NNZ really are pathways to GNZ and not simply exercises in emissions displacement. This in turn requires that either (a) states eventually adopt CVCA themselves or (b) some other mechanism is devised for disincentivizing the kind of displacement activity that allows NNZ to diverge from CVCNZ. And although in scenario (b) states would not be using CVCA themselves, it would still be needed to measure whether the mechanism was working or not. Some states already practise a form of dual emissions accounting, recording their territorial emissions along with a record of some of the value-chain emissions covered by CVCA. ³⁷ So although we are not advocating the immediate adoption of CVCA by states, we think it should not be ruled out as a desirable future enhancement of national emissions accounting practices.

Against this background, we now describe and respond to some of the main objections that CVCA tends to encounter.

Incentives

A common objection is that CVCA, as a system of national emissions accounting, would create perverse incentives with respect to trade between NZ-aligned and non-NZ-aligned states. If Viridia sells its electric vehicles to non-NZ-aligned state Ignavia, CVCA requires Viridia to include Ignavia's territorial emissions on its own ledger. However, suppose Ignavia's alternative to importing Viridia's electric cars are buying combustion engine cars from another non-NZ-aligned country. Should we not then encourage the trade between Ignavia and Viridia? Wouldn’t the adoption of CVCA by Viridia, in disincentivizing its trade with Ignavia, impede progress towards GNZ?

To this objection there are two replies. First, it at most supports option (b) over option (a) above. It is at best an argument against states’ adopting CVCA themselves, not an argument against CVCA as the correct answer to our opening question. But, secondly, while it is a sensible concern to raise about (a), we don’t see it as decisive. Whether this objection to (a) succeeds is circumstance-dependent, and the circumstances in which it does not succeed are not especially far-fetched. Since this issue is peripheral to our main line of argument, we will not pursue the point in detail here. But, in brief, what will matter is whether there is enough demand for the products of a genuinely NZ-aligned economy, and enough recognition that a genuinely NZ-aligned economy must be CVCNZ-aligned, to allow a state to sustain an economy of this kind by pricing its products at a level that makes this possible, and passing on any associated costs to its trade partners. These are the economic dynamics that allow some firms to make NZ commitments that employ a measure of value-chain emissions accounting ³⁸ ; the same dynamics could, in suitable circumstances, make it possible to operate in the same way at the national level, passing on the associated costs.

Informational burdens

A different kind of worry about the practical implementation of CVCA concerns the informational burdens it imposes. Wouldn’t CVCA require information about value-chain emissions on a scale, and at a level of detail, that are not practically attainable? And isn’t this worry amplified further when we insist that our argument defends CVCA as an accounting method for use by global citizens rather than national states (for now)? How can we, as global citizens, make the calculations CVCA would require us to make?

We think the force of this worry is blunted by three observations. First, CVCA can draw on (and conjoin) three developed proposals for emissions accounting that use available data: territorial, consumption-based and income-based accounting. Second, it is true that implementing any such accounting system requires filling data gaps through the use of averages, but that is fully defensible. ³⁹ It properly reflects the way in which our epistemic position with respect to climate change is limited: given this, we are accountable for reducing the risk imposed by our emissions, where the risks an agent is morally accountable for reducing (the risks that are in question when we make moral judgements of negligence and recklessness) depend on the epistemic probabilities it is rational for agents to associate with their actions, given the evidence. ⁴⁰ The use of averages accurately reflects this: it supplies the basis for making the epistemic probability assessments that are relevant to determining the risk imposed by our emissions. And, third, when we say that CVCA is the correct answer to our opening questions, we are not thereby recommending that private individuals make their own calculations of states’ value-chain emissions. This is a task for economists – conjoining three already developed proposals, and using publicly available data – in serving the interests we all have in holding states morally to account for their climate actions.

Double-counting

CVCA leads to the double-counting – indeed, multiple counting –of emissions. We claimed that, given its purpose, this is a feature not a bug. However, the worry provoked by this feature might persist. In particular, it might be asked: if the purpose of CVCA is to measure mereological contributions to GNZ – that is, parts of the attainment of that common goal – how can it make sense to attribute the same emissions to two different parts of the same whole?

This can make sense because of the structure of moral accountability. When you suffer the imposition of a burden, more than one party can be accountable for the imposition; and when the burden has parts, more than one party can be accountable for the imposition of the same part. This is especially obvious when burden-imposers cooperate. When your house is ransacked by looters, and Remy carries off your Rembrandt while Renee and Renata together carry off your Renoir, then Remy is accountable for depriving you of one part of your art collection, while Renee and Renata are each accountable for (together) depriving you of another part of it. This general point applies to cases of mereological contribution to a common goal. In Deal, where you pay Lumi to take your snow, knowing she will not clear it, you are both accountable for the uncleared snow. In Short Pipe, you and Kemuri are both accountable for the smoke he releases in heating your house. And in the cases where CVCA makes value-chain emissions attributions – cases where one state cooperates by trading with another, non-NZ-aligned state – there are (at least) two parties cooperating in a mutually beneficial activity that produces an external cost in the form of an emissions burden, and those who are burdened have claims against both parties for the imposition of the burden. If either removes it, the claim against the other is eliminated. But when neither does so, a claim is directed against each of them.

Consequently, if an accounting system is going to cover the mereological contributions to GNZ that states are morally accountable for making, it must double-count.

Vagueness

A different criticism concerns the vagueness of several of the concepts that CVCA deploys. For example, according to condition 2, Y's production emissions are added to X's inventory when Y's activity is ‘part of’ producing or supplying something for X. One might wonder exactly what this includes. Suppose that when A's mobile phones are produced in B's factories, the workers get to the factories using high-emitting private transport. Are the workers’ travel emissions ‘part of’ B's supplying mobile phones to A? The same sort of question can be raised for several of the concepts used in formulating CVCA: ‘production’, ‘supply’, ‘inputs’, ‘gaining an income’ and so on. None of these has completely sharp boundaries: all admit of borderline cases.

This means that to implement CVCA in a way that generates a consistent set of emissions accounts, further decisions are needed to operationalize those vague concepts, making their boundaries calculably precise. That seems right to us – but it is not a good objection to CVCA. CVCA seeks to identify the morally significant relations that make a state accountable for having net positive emissions. So it is phrased in terms of those morally significant relations. Condition 2, for example, claims that when X pays Y to perform an activity, the relation of being part of that activity is itself morally significant in determining which of Y's emissions should be attributed to X. It is true that this relation admits of borderline cases. But many important moral relations are vague and admit of borderline cases. And while we do not here supply a further analysis of the mereological relation of one activity's being part of another, any such analysis could be incorporated into our account (with rival analyses producing rival variants of CVCA).

So if the example of the workers’ travel was a borderline case, that would not be a problem for CVCA; but in fact, we do not think it is a borderline case. The workers’ travel to work may be a necessary condition for producing mobile phones, but it is not part of the activity of producing mobile phones. (Of course, this doesn’t mean that their transport emissions should go uncounted. They should be included in B's inventory as production emissions, but not in A's inventory as supply-chain emissions.)

Over-inclusiveness

A final worry concerns how far upstream in the value chain we must go in tracking the emissions embedded in an imported product. Discussions of Scope 3 emissions accounting commonly talk of the ‘cradle-to-grave’ emissions profile of a product, with the cradle for a product like a mobile phone being the mine from which its materials were extracted. ⁴¹ But, in CVCA, there is no cradle. To do its mining, the mining company needs to buy machinery, the manufacture of the machinery in turn has inputs further upstream, and so on indefinitely. This might seem to produce an absurd implication: that, given the interconnectedness of the global economy, all anthropogenic GHG emissions will eventually end up on state A's ledger, and on every other state's ledger too.

But that worry is misplaced. When B supplies goods to A, only the proportion of B's overall emissions that derive from activities it conducts as part of supplying A are attributed to A. It is true that the emissions released in manufacturing mining machinery lie upstream as distant inputs in the value chain of the mobile phones A imports. However, the further upstream we travel, the smaller the proportion of those emissions that will be attributable to A. In practice, it is reasonable for a calculation of the upstream emissions embedded in a product to stop at an intuitive ‘cradle’ point, such as the mine from which its materials were extracted, if the emissions to be attributed from processes further upstream are too small to justify the work involved in accounting for them.

The same point applies to the attribution of downstream emissions by condition 3. This, too, is formulated in a way that can extend several steps along a value chain. When, in mobile phones, B burns coal sourced from C, the resulting emissions are attributed to B by condition 1, and to C as the coal supplier by condition 3; but if another state G has financed the coal-mining operations in C, emissions are also attributed to G in proportion to the income it derives from its inputs to the process. Worries about over-attribution have the same answer: attribution should be proportional to the income derived from the input to the process, using the techniques of income-based emissions accounting.

Conclusion

Many previous discussions of emissions accounting have noticed the problem that can be posed by the ‘leakage’ or displacement of emissions. Some of those discussions propose forms of value-chain emissions accounting as the remedy; but they do so while adopting a Responsibility Framing. ⁴² Within that framing – which sets emissions accounting itself the aim of guaranteeing an equitable distribution of the costs of climate action – the double-counting of emissions clearly is a problem. However, as we explained in the ‘Proposals for emissions attribution’ section, that framing should be rejected. The question ‘How should responsibility for climate action be equitably distributed?’ is very important, but it is a mistake to try to answer it through a method of emissions accounting. The question of what a country's net emissions are is an ethical question, but it is not that ethical question. It is a prior question – one that has to be answered first, before we can specify fair shares of the common task of reaching GNZ. Our key claim has been that when a state makes an NNZ commitment, it is morally accountable – to the climate-vulnerable, to the other participants in the global effort to reach GNZ, and to its own citizens – for fulfilling its commitment in a way that mereologically contributes to that global goal, by eliminating its mereological contribution to the problem that that goal aims to solve. What we need a method of emissions accounting to do – whether we are assessing a commitment to NNZ or any other net emissions target – is to measure that mereological contribution; and the right way to do so is by using Contributory Value-Chain Accounting.