States facing a costly no-agreement outcome

The costs associated with the no-agreement outcome affect a state’s interest in reaching a substantively strong agreement, and, thus, its interest in using an issue-addition strategy. These costs are shaped by the substantive nature of the issues being negotiated. If the agenda deals with an issue area where the policies of other states are consistent with a state’s interests, that state faces few costs if a cooperative agreement is not reached. However, if other states’ current policies regarding the issues on the agenda go against a state’s interests, the inability to reach a cooperative agreement on those issues will be costly for that state. This type of state has an incentive to reach an agreement that moves other states’ policies regarding the issues on the agenda as far as possible away from the ones they currently have in place. Issue addition can help it do so by fostering an exchange of concessions that allows it to win on the issues currently on the agenda. However, a state can only effectively use issue addition when the promises underpinning this strategy can be credible.

The distribution of states’ interests across the issues on the bargaining agenda affects the credibility with which the promises underpinning an issue-addition strategy can be made. As demonstrated by the issue-linkage literature, if the issues on the current agenda are differently valued, an agreement is likely to be reached on those issues. Regardless of the costliness of a state’s no-agreement outcome, it is unlikely to have to face those costs. Using issue addition and offering concessions to the other side in order to secure a beneficial agreement on the original issues is unnecessary and costly. A state is, therefore, unlikely to use an issue-addition strategy when negotiating over an agenda characterized by issues that are differently valued.

In contrast, if the issues on the agenda are not differently valued, an agreement is unlikely to be reached. This forces states to face the costly no-agreement outcome. A state can use issue addition to avoid that outcome by offering concessions on an outside issue in order to foster an exchange of concessions between that issue and the original ones, thus bringing about an agreement. The greater the costs associated with that no-agreement outcome, the more likely it is that the benefit of using an issue-addition strategy – even minus the cost of doing so – will yield a higher payoff than remaining with the current bargaining agenda and having to face the cost of not reaching an agreement. When a state finds itself in this type of strategic situation, an offer of concessions on an outside issue can be credible, and thus effective in achieving its interest in reaching a beneficial agreement on the original issues.

Consider the 1982 negotiations within the European Economic Community (EEC) over the renewal of EEC sanctions against Argentina.Footnote ²⁴ Most EEC states argued that the cost of renewing these sanctions outweighed the benefits. However, renewing them was very important to the United Kingdom, as they provided leverage in its dispute with Argentina over the Falkland Islands. Absent a renewal, the EEC sanctions on Argentina would expire, and the leverage the United Kingdom gained from them in the dispute would disappear. Consequently, the United Kingdom faced a costly no-agreement outcome associated with this bargaining agenda.

At the same time, this one issue – the question of renewing sanctions on Argentina – was the central focus of states on both sides of the negotiation. This is a key feature of a bargaining agenda characterized by issues that are not differently valued. Consistent with this type of agenda, bargaining impasse resulted in these negotiations, forcing the United Kingdom to face its costly no-agreement outcome.

The United Kingdom, therefore, had an incentive to use issue addition to put an issue of importance to the other EEC states on the bargaining agenda, so that the new set of issues being discussed would be differently valued. A beneficial agreement on the original sanctions issue could then be reached. This is precisely what it did. It brought the negotiation of the EEC budget into the discussion of the sanctions, offering to accept a larger agricultural budget than it wanted in exchange for concessions by the other EEC states to renew the sanctions on Argentina. The issues now being jointly considered were differently valued by the EEC states. The sanctioning issue was most salient to the United Kingdom, while the budget issue was most salient to other EEC states.Footnote ²⁵ Once these issues were linked together, an exchange of concessions (and an agreement) was possible.

Overall, a state that has a more costly no-agreement outcome and is bargaining over issues that are less differently valued is more likely to derive benefits from using an issue-addition strategy that outweighs the costs of doing so. It is thus more likely to adopt this type of bargaining strategy, all else constant. This argument leads to a first testable hypothesis:

States facing a cost-imposing agreement

In contrast, states facing the potential to have to make costly policy adjustments if an agreement is reached on a given set of issues have little incentive to use issue addition. Such a strategy not only keeps the current set of (costly) issues on the table, but it also imposes an additional cost on the adding state to do so. The greater the policy-adjustment costs that would be imposed by an agreement, the less likely it is that a state would want to use issue addition to help bring about that agreement and/or to further strengthen it.

This lack of incentive to use issue addition exists regardless of whether the issues on the agenda foster or hinder agreement. If the issues on the agenda are structured in a way that hinders agreement, a state that faces having to make costly policy adjustments if an agreement is reached has little incentive to pay the cost of using issue addition in order to create an issue- linkage structure that will bring about that (cost-imposing) agreement. If the issues on the agenda are structured in a way that fosters agreement, using issue addition and offering concessions on a new issue would only strengthen the policy-adjustment requirements that state would be required to make. A state that would have to make significant policy adjustments if an agreement was reached on the original issues, alone, therefore has little incentive to use an issue-addition strategy. Instead, when an agreement is likely to be reached, it has an incentive to try to reduce the policy changes it would be required to make – either by watering down the agreement or preventing the agreement from being reached, altogether.

Consider the UN Security Council negotiations in the 2000s that were designed to respond to the humanitarian crisis in Darfur by placing sanctions on Sudan. In these negotiations, China faced significant policy-adjustment costs from agreement. It had strong trade links with Sudan – both economic (involving oil trade, in particular) and military (involving arms trade). If an agreement was reached, China would have to pay the cost of cutting off the flow of oil it was importing from Sudan, as well as the cost of cutting off arms exports to it. Given the costly policy-change requirements associated with agreement on this sanctions issue, China had no interest in further increasing the costs of such an agreement by offering side-payments to help bring it about. Indeed, rather than add an issue to try to secure the sanctions agreement, China (informally) vetoed the potential agreement several times.Footnote ²⁶

Overall, the greater are the policy-adjustment costs a state expects will stem from agreement on a given set of issues, the less likely it is to have an incentive to use an issue-addition strategy.Footnote ²⁷ This argument leads to a second testable hypothesis:

States facing a cost-imposing agreement

The costliness of the policy adjustments an agreement would require a state to make affect its interest in reaching a substantively weaker agreement, and thus its interest in using an issue-subtraction bargaining strategy. These costs are shaped by the substantive nature of the issues being negotiated. An agenda that deals with a policy area in which a state is not significantly involved, or in which it is already a policy ‘entrepreneur,’ will impose few policy-adjustment costs on that state if a cooperative agreement is reached. However, an agenda designed to move policy forward in an issue area in which a state is a ‘laggard’ can impose potentially significant policy-adjustment costs on that state if an agreement is reached. This type of state has an incentive to water down the agreement and lessen those potential policy-adjustment costs – i.e., to use issue subtraction. However, it can only effectively do so when the threats underpinning this strategy can be credible.

The distribution of states’ interests across the issues on the agenda affects the credibility of issue subtraction strategies. If the issues on the agenda are not differently valued, it is unlikely that an agreement will be reached.Footnote ³⁴ Because an agreement is not likely to be reached, any policy change that would have been required by that agreement is not likely to be imposed. Regardless of the costliness of the policy changes that potential agreement would require, a state would have little incentive to pay the costs of issue subtraction to water down the ‘already-unlikely’ agreement.

However, if the issues on the agenda are differently valued, an agreement is likely to be reached,Footnote ³⁵ and states would have to pay the costs associated with the agreement’s required policy changes. The greater the costs associated with those required policy changes, the more likely it is that the cost of using issue subtraction and remaining at the status quo would be offset by the benefit of not having to pay the policy-adjustment costs associated with reaching an agreement. The veto threat underpinning an issue-subtraction strategy can be credible, because the state would have an incentive to carry it out. When issues are more differently valued, a state facing the possibility of having to make more costly policy adjustments if an agreement is reached is, therefore, more likely to use issue subtraction.

Consider the Strategic Arms Limitation Treaty (SALT) II negotiations that took place between the United States and the Soviet Union in the 1970s. These negotiations were designed to reduce the American and Soviet manufacturing of nuclear weapons. The United States placed most importance on limiting the nuclear capabilities of the Soviet Union, while the Soviet Union placed most importance on limiting the nuclear capabilities of the United States. The main issues on the table were, indeed, differently valued by the two states, allowing a trade-off to be possible between them. At the same time, each side faced significant policy-adjustment costs if such an agreement was reached. They would each be forced to limit their own nuclear arsenals, which were seen as important for protecting their strategic interests. Both sides responded to this bargaining situation with an issue-subtraction strategy. The United States removed the requirement for the reduction of what it considered to be essential defensive weapons (such as the Trident and cruise missiles). The Soviet Union removed the requirement for the reduction of most of its ‘heavy ICBM’ launchers. The result was a less robust agreement than it could have been if it required a limitation of all strategic weapons. However, because of their interest in limiting the requirement to reduce weapons they considered central to their own strategic position in the system, both used issue subtraction in order to water down the agreement and protect those interests.

Overall, a state that faces having to make more costly policy adjustments if an agreement is reached and is bargaining over issues that are more differently valued is more likely to derive benefits from using an issue-subtraction strategy that outweigh the costs of doing so. It is, therefore, more likely to adopt this type of bargaining strategy, all else constant. This argument leads to a third testable hypothesis.

States facing a costly no-agreement outcome

In contrast, states facing a costly no-agreement outcome have little incentive to use issue subtraction. For these states, remaining at the status quo comes with costs. They would prefer to have the agreement move the policy outcome as far from that status quo as possible. However, a state risks losing the entire agreement and having to remain at the (costly) status quo if it wields a veto threat – a tactic that often characterizes an issue-subtraction strategy. Even if this is not the case, subtracting an issue ‘weakens’ the substance of the potential agreement – i.e., it leads to an outcome that does not move policy as far from the (costly) status quo as an agreement that included all of the issues would have moved it. Even further disincentivizing the use of issue subtraction, states pay costs when they try to remove issues from the agenda. Using issue subtraction would not only leave the policy outcome closer to the costly status quo, but a cost would have to be paid to do so. A state that faces a costly no-agreement outcome, therefore, has little incentive to use an issue subtraction strategy.

This lack of incentive exists regardless of whether the issues on the agenda foster or hinder agreement. If the issues on the agenda are structured in a way that hinders agreement, a state facing a costly no-agreement outcome would have an incentive to bring about an agreement in some way. However, it would have an incentive to do so in a way that keeps the agreement on the original issues as strong as possible, rather than in a way that waters it down. It would thus have an incentive to use issue addition, not issue subtraction. If the issues on the agenda are structured in a way that fosters agreement, a state facing a costly no-agreement outcome would have an incentive to leave the agenda as it is. Doing so would allow an agreement to be reached and the costly no-agreement outcome to be avoided. In both cases, this type of state has little incentive to use issue subtraction.

Consider the negotiations in the Copenhagen climate-change conference in 2009. Many issues were on the table, revolving around questions related to the mitigation of greenhouse gas emissions and providing aid to developing states to help them adapt to the adverse effects of climate change. If an agreement was not reached, the status quo would remain – an outcome characterized by the fact that many major greenhouse gas emitters were not required to reduce their emissions, as well as the fact that all emissions reduction requirements currently in place would lapse in 2012 with the expiration of the Kyoto Protocol.

Small island states like Tuvalu faced significant costs if the status quo continued unabated – costs related, in particular, to the rise in sea levels associated with climate change. Absent a reduction in greenhouse gas emissions, this problem would only continue to worsen. Indeed, these states argued that allowing things to continue without taking significant action posed a threat to their very survival.

So we would expect these states to have little incentive to water down the agreement through the use of issue subtraction. Consistent with this expectation, these small island states did not try to limit the scope of the agreement by removing issues from the table. Instead, they were some of the most active states in pushing for a comprehensive and strong agreement that would make important policy advances on all of the issues on the agenda. They even pushed hard for the inclusion of issues that were, at best, tangentially related to them – for example, support for states dealing with drought and desertification.Footnote ³⁶ Rather than try to water down the agreement, the small island states were some of the biggest supporters of having a strong agreement that covered all issues related to climate change.

Overall, states that face costs if an agreement moving policy beyond the status quo on the original issues is not reached, therefore, have little incentive to water down the agreement by using issue subtraction. The greater are the costs a state faces if status quo policy remains, the less incentive it has to use of this type of strategy. This argument leads to a fourth testable hypothesis:

Dependent variables: issue addition and subtraction

To test Hypotheses 1 and 2, the first analysis focuses on states’ use of issue addition as a bargaining strategy. The dependent variable is dichotomous – coded 1 if a state adopted an issue-addition bargaining strategy in a particular negotiation, and 0 if it did not.Footnote ⁴¹ A state’s bargaining strategy is coded as being characterized by issue addition under two conditions. First, a state is coded as having used issue addition if it suggested that an issue not already on the bargaining agenda should be brought into the discussions.Footnote ⁴² A state’s bargaining strategy is also coded as being characterized by issue addition if it suggested creating a working group (‘Contact Group’) to discuss a new issue. By doing so, a state is essentially pushing for that issue to be added to the Contact Group level of negotiations.Footnote ⁴³

To test Hypotheses 3 and 4, the second analysis focuses on states’ use of issue subtraction. The dependent variable is dichotomous – coded 1 if a state adopted an issue-subtraction strategy in a particular negotiation, and 0 if it did not. A state’s bargaining strategy is coded as being characterized by issue subtraction under two conditions. First, a state is coded as having used an issue-subtraction strategy if it explicitly suggested that an issue currently on the bargaining agenda be removed from the discussions.Footnote ⁴⁴ A state’s bargaining strategy is also coded as being characterized by issue subtraction if it suggested moving discussions of an issue from the current negotiation to another level of negotiations within that COP – i.e., removing it from the bargaining agenda in the current negotiation.Footnote ⁴⁵

Costs from agreement

As discussed above, the substantive nature of the bargaining agenda influences the costs associated with reaching an agreement. In the climate-change negotiations analyzed here, one of the main goals was to reduce global greenhouse emissions by placing specific emission reduction requirements on states. The emissions reduction actions states were expected to adopt in the resulting agreement were to be ‘nationally appropriate,’ based on a principle of ‘common but differentiated responsibilities,’ and applied to both developed and developing countries. States responsible for a greater degree of emissions were, therefore, expected to take on greater commitments to reduce emissions.

An indicator of the costliness of the policy adjustment requirements likely to stem from an agreement is therefore the level of greenhouse gases currently being emitted by a state. I include a measure each state’s level of carbon dioxide (CO₂) emissions (kt) in the models to capture the costs that reaching an agreement would likely impose on a particular state.Footnote ⁴⁶ Larger values of this variable indicate that a state would likely have to implement more significant policy adjustments to comply with any climate-change agreement that was reached.

Hypothesis 2 predicts that states facing more costly policy changes if an agreement is reached should be less likely to use issue addition. As a stand-alone factor, this variable should therefore exert a negative and statistically significant effect on the likelihood a state will use an issue-addition bargaining strategy.

No-agreement costs

Another factor influencing a state’s incentive to use an issue-linkage strategy is the costliness of its no-agreement outcome. In the context of climate-change negotiations, the costliness of a state’s no-agreement outcome is characterized by the adverse effects of unregulated greenhouse gas emissions that if faces, and its ability to deal with those effects. Therefore, I code the costliness of each state’s no-agreement outcome by drawing on two factors. I use the ‘Environmental Vulnerability Index’ (EVI) indicators to construct a measure of each state’s vulnerability to climate change.Footnote ⁴⁷ I then weight this measure by a state’s gross domestic product (GDP) per capita, which captures its wealth, and thus its own ability to deal with any environmental problems it faces. The resulting measure captures a state’s environmental vulnerability to climate change, as well as its ability to deal with those climate-change problems unilaterally, making it less dependent on an agreement for help.Footnote ⁴⁸ States with the most costly no-agreement outcome face significant environmental impacts from climate change and have few resources to deal with those problems. States with the least costly no-agreement outcome face few adverse effects from climate change and have significant resources to deal with any problems they do face.Footnote ⁴⁹

Hypothesis 4 predicts that states facing a more costly no-agreement outcome should be less likely to use issue subtraction. As a stand-alone factor, this variable should, therefore, exert a negative and statistically significant effect on the likelihood a state will adopt an issue- subtraction bargaining strategy.

Differently-valued issues

I argue that the degree to which the issues on the bargaining agenda are (or are not) differently valued exerts important effects on states’ likelihood of adopting an issue-linkage strategy. A variable is included in the models to capture the degree to which the issues on the bargaining agenda were differently valued in these various climate-change negotiations.

The coding rule stems from a logic derived in the formal issue-linkage literature, which demonstrates that linking issues together that are more differently valued creates a larger zone of agreement,Footnote ⁵⁰ while the linkage of issues that are less differently valued reduces (or eliminates) the zone of agreement, all else constant.Footnote ⁵¹ I have built on this logic elsewhere, proposing empirical coding rules that capture the average size of the zone of agreement created by linking together a given set of issues. These rules were used to code the bargaining agenda in various negotiations among the states of the EU.Footnote ⁵² Here, I used this same coding procedure to measure the degree to which the issues on the table were differently valued in the climate-change negotiations I am analyzing.Footnote ⁵³ Larger values indicate that a larger zone of agreement was created by the linkage of the issues on the agenda, and thus that the issues on the agenda were more differently valued, all else constant.

Interaction effects

The predictions of Hypotheses 1 and 3 involve the interaction of these variables. Therefore, I interact the Differently-valued issues variable with the No-agreement costs and CO ₂ emissions variables in order to test these hypotheses.

Larger values of the first interaction term, Differently-valued issues × No-agreement costs, indicate a state with a more costly no-agreement outcome bargaining in a negotiation characterized by issues that are valued more differently. Because Hypothesis 1 predicts that a state with a more costly no-agreement outcome will be more likely to adopt an issue-addition strategy in negotiations characterized by issues that are valued less differently, this interaction term should exert a negative and statistically significant effect on the likelihood that a state will adopt an issue-addition bargaining strategy.

Larger values of the second interaction term, Differently-valued issues × CO ₂ emissions, indicate a state with greater levels of CO₂ emissions engaged in a negotiation characterized by issues that are valued more differently. Hypothesis 3 predicts that a state facing greater costs associated with the policy-adjustment requirements a climate-change agreement would impose (i.e., with greater CO₂ emissions) will be more likely to adopt an issue-subtraction strategy in negotiations characterized by issues that are valued more differently. This interaction term should, therefore, exert a positive and statistically significant effect on the likelihood a state will adopt an issue-subtraction bargaining strategy.

Control variables

Variables are also included in the models to control for variation in potential mitigating factors. First, a variable is included to control for a state’s material power, which could impact its ability to use either type of issue-linkage strategy effectively. States with greater economic resources are more likely to be able to use issue addition and offer side payments to other states. They are also more likely to be able to sustain the reputational costs of issue subtraction, given that they are potentially more important cooperative partners in the international system than are weaker states. A measure of each state’s GDPFootnote ⁵⁴ is included in the models to control for its potential effect on states’ ability to adopt issue-linkage strategies.Footnote ⁵⁵

Three additional variables are included to account for differences in states’ likelihood of adding or subtracting issues in the different bargaining forums of each COP. First, because states are likely to use issue linkage to add or remove Contact Groups from the conference, negotiations within these Contact Groups, themselves, might differ from other COP bargaining interactions, where creating and/or removing an entire Contact Group might be a more likely strategy. A variable is included in the models to control for this potentially mitigating effect.Footnote ⁵⁶ Second, parties to the Kyoto Protocol that take place during each COP are specifically designed to deal with Kyoto Protocol-related issues and the reduction commitments of Annex I states. However, given that other issues such as emission reduction requirements for developing countries are included in other parts of the conference, it is possible they may be linked to these Kyoto Protocol negotiations as well.Footnote ⁵⁷ This might lead to the use of issue linkage strategies -- both by states that want to make such linkages, as well as those that want to prevent them. A variable is included to control for these potential effects. Finally, negotiations in the high-level plenary session fall at the end of each COP, after all other negotiations have taken place. A state’s likelihood of using issue addition or issue subtraction is likely to differ from its likelihood of using these strategies in earlier bargaining interactions in the conference. A variable is included in the models to control for this potentially mitigating effect.Footnote ⁵⁸