Two-level games on the trans-boundary river Indus: obstacles to cooperation

This synthesis paper explores the reasons hindering water cooperation between India and Pakistan on the Indus River Basin. It argues that both domestic and international-level elements narrow the size of the ‘win-sets’ which make water cooperation between the two states highly challenging. Not only state actors but also the domestic actors in both India and Pakistan have repeatedly played ‘water games’. Further, due to long-standing geopolitical and territorial conflicts between India and Pakistan, the strategies pursued so far by these states including ‘securitization’, ‘issue-linkage’ and ‘alliance strategies’ as leverage mechanisms, have also contributed to the lack of cooperation in their water realm.


Graphical Abstract
Introduction When India and Pakistan gained independence in 1947, partition ruined the interdependent hydraulic system (Ranjan, 2016) as it was not based on water resource but on religious identity (Wirsing, 2008). As former World Bank Senior Water Advisor Briscoe stated, the borderline between the two countries was determined on the basis of religion in 1947, without giving heed to hydrology. As a result, the headwaters of irrigation canals in Punjab, Pakistan's most fertile region, were in Indian-held Kashmir (Briscoe, interviewed by Lipschutz, 2010). However, this dividing line gave around 90% of the irrigated land of the Indus Basin to Pakistan (Condon et al., 2014).
Since then, a prolonged chain of water disputes ensued between the two states. An explosive situation appeared to be forestalled when, after several rounds of negotiations, they concluded the Indus Water Treaty (IWT) with the assistance of the World Bank in 1960 (Nazakat Ali et al., 2015). However, their water conflicts have flared up again in recent decades. As the lower riparian, Pakistan has been the natural objector in this unresolved dispute.
The present water disputes India and Pakistan are mostly in relation to projects on dam construction, storage and water diversion. Indian projects on Pakistan's Indus tributaries, including the Kishanganga Dam (330 MW), the Baglihar Dam (850 MW), Ratle Dam (810 MW) and the Tulbal project (Wullar Dam), which turned a natural lake into an artificial storage reservoir, are the main schemes over which Islamabad has unresolved disputes with Delhi (Riffat & Iftikhar, 2015). The main complaints of Pakistan's Indus River system (IRSA) authority's experts are that India has carried out massive water storage works not only on the three rivers under its absolute control but on all six rivers belonging to the Indus basin (Zaman & Abubakr, 2016). According to Pakistan, these water infrastructures have violated the Indus Agreement and Pakistan's water rights on the two rivers, Chenab and Jhelum, due to the defective and incomplete design of water projects by India (Kiani, 2018), while these projects create severe water shortages or flood extremes in Pakistan. Abid Shirali, Pakistan's Minister of Water and concerned with their security requirement and therefore, they follow their own national interest (Tema, 2014).
Realists consider international politics as a continuous battle for power (Lee 2018). In the classical Realist view, unlike military security and diplomatic relations, economic and environmental problems are not considered high-politics issues. Apart from military force, however, they also endeavour to increase their capabilities in other domains including demographic resources and natural resources such as water resources (Warner, 2016, pp. 743-744). In this case, they may shed tears to preserve their national interest, particularly when water can enhance their power. Indeed, states may use water resources as emergent and sometimes intentional, political tools to control a space and people and strengthen state-building process to increase their power (Molle et al., 2009). Realist leaders attempt to decrease interdependence, as it increases vulnerability (Warner, 2016). According to Lowi (1993), the chances for cooperation on transboundary waters are low, as upstream countries have little to gain from cooperation.
Among other possible issues with Realism, the emergence of non-state actors and nonconventional conflict between international agents challenges its explanatory power (Orsi et al., 2018). For Realism, domestic politics is unnecessary to account for important foreign policy decisions or, at least, international political consequences.
With the decline of the Cold War during the 1970s, however, the importance of international and nonstate actors and multinational corporations increased, leading to a revival of the idealist view of IR, which became known as liberal (neo)institutionalism or pluralism (Tema, 2014). Keohane & Nye (1972), for instance, highlight the significance of non-state actors, such as multinational corporations and their transnational relations, the role of revolutionist groups and the economic relations in the world interactions. As a consequence, states have less control over events. Consequently, pluralist thinkers believe in the reduction of conflicts and the achievement of peace through trade, communication and international institutions.
Liberal (neo-)institutionalism draws attention to the countries' reasons (or cost-benefit calculations) for engaging in cooperative processes and establishing regimes (Leb, 2009). Robert Keohane (1984), a leading neo-institutionalist theorist, seeks to show how cooperation will take place even despite the features that neo-realists single out: the anarchic international system; states as main actors, acting on the basis of rationality and their national interest; capability being important to states. According to this school, states are atomistic actors seeking to increase their absolute achievements, so if cooperation is more profitable, they will cooperate (Keohane, 1984).

Two-level game theory
In this context, two-level game theory may be an appropriate synthesis, because its win-set concept endeavours to show under what conditions the possibility of cooperation in international negotiation (concluding an agreement) would enhance and when the states fail to reach cooperation or an agreement (the possibility of conflict) (Bowles & McGinn, 2008;Menga, 2016).
Given its focus on the leaders of a state as the major actors in international negotiations, two-level game theory is considered to be state-centered. However, Putnam (1988) believes 'the state-centric literature is an uncertain foundation for theorizing about how domestic and international politics interact'. Unlike Realism, this approach recognizes the importance of domestic players and institutional constraints. Thus, while showing similarities with a liberal vision, it emphasizes that managers have a certain degree of autonomy in their decision-making (da Conceição-Heldt & Mello, 2017). Indeed, twolevel game theory overemphasizes neither cooperation nor conflict; rather this theory is a kind of political model of international conflict resolution between countries derived from game theory. Putnam (1988) argues that diplomatic interactions should be considered as negotiations interconnected with domestic politics. He argues that government leaders simultaneously engage in 'twolevel games' in multilateral or bilateral negotiations: Level I (diplomatic) and Level II (domestic). At the diplomatic table, bargaining between negotiators could lead to an early agreement. But after Druckman (1977), Putnam observed that a negotiator tries to create a package that is both acceptable to other players at the diplomatic table and to actors within the country.
The executive branch of states as 'major negotiators' in international negotiations will ultimately have to gain the approval of domestic actors. However, the approval process may be formal or informal. One of the main motivations of the negotiator is to enhance his/her position at Level II by adding his/her political resources or minimizing potential losses. It is reasonable to assume for the case of international double bargaining, that if a choice is to be made, the senior negotiator generally gives priority to the domestic community, in particular, because his position and responsibility often depend on his position inside the country. In general, the success of the talks depends on the size of the 'win-sets' (Putnam, 1988).
A win-set is defined as a set of all possible agreements at Level I that would gain the approval of the majority in the domestic level. The larger the win-set inside the country, the greater the chances of the success of the negotiations; but, in turn, it will weaken the bargaining power of the country in the negotiation process. On the contrary, the smaller the win-set, the greater the risk of involuntary defection and the more likely that negotiations would fail for the lack of 'approval' by each negotiating party. Only if the win-sets of all sides overlap, they can reach an agreement. Putnam claims that the outcome of the win-set shows a zero-sum game between X and Y represented in Figure 1. XM and YM show the maximum results for X and Y, while X1 and Y1 are the possible approved results. At this point, any agreement between the X1 and Y1 can be approved by both sides. The implications between Y1 and Y2 will no longer be possible if the size of the win-sets of Y is reduced to Y2, and therefore the range of possible agreements would be truncated in Y's favour. However, if Y continues to decrease its win-sets to Y3, the win-sets are no longer overlapping and negotiators suddenly face a deadlock (Putnam, 1988).

Conditions affecting win-set size
For Putnam (1988), the three sets of factors are particularly important for influencing win-sets: (1) Priorities and coalitions at Level II, (2) Institutions at Level II and (3) Negotiator strategies at Level I.
The distribution of power, preferences, and possible coalitions among Level II players has an effect on the win-set's size. Issue linkage tactics which can cause two levels to become entangled affect the size of the win-sets. For instance, this occurs when the majority of the second level actors are opposed to a specific policy (for example, oil prices), but some are willing to vote in favour of this deal, in exchange  (Putnam, 1988). for creating more jobs in the export industry, the negotiator can do an international deal which offers more jobs in the field, so it can change the initial result at the domestic level and the negotiations will come to an end (Putnam, 1988).
Each party can also use issue-linkage tactics as a strategy in bargaining in regional and international contexts. Just & Netanyahu (2000) show that the complexity of international negotiations can be better modelled by linking independent games. They claim that issue linkage strategy has a tendency to be preferred in the particular conditions, such as Israeli-Palestinian water negotiations circumstances.
Domestic political procedures can also affect the size of the win-set. For example, the strong order in the ruling party boosts the win-set by expanding the range of agreements that the Level I negotiator calls for support. The more autonomy the central decision makers have from Level II, the bigger their winsets and the more the likelihood of reaching the international agreement too. Thus, Putnam argued that 'as a state is stronger in terms of autonomy from domestic pressures, the weaker its relative bargaining position internationally.' A senior negotiator whose political position is high may easily obtain approval of his domestic constituents that it is useful in the expansion of the win-set which as a result will strengthen the Level I agreement.
Approval procedures at the domestic level clearly affect the size of the win-sets. For example, if a two-thirds vote is needed for approval, the win-set is definitely smaller than a simple majority. This is effective when two sides on negotiation reach a primary agreement.
Some governments look to spread each other's win-sets; their leaders may provide foreign aid to a friendly but unstable government which has an effect on the size of the win-set. They seek to restrict internal restrictions that might otherwise prevent their governments from cooperating (Putnam, 1988).
The common win-sets of the two parties have been explored by systematically examining the proposals put forward by both sides in the negotiation sessions (see Introduction), as well as the responses of the two sides to these suggestions expressed in primary and secondary sources, such as news sites, reports from international experts engaged in the negotiations (e.g. the report by Lafitte, a neutral expert for the Baglihar project appointed in accordance with the provisions of IWT's Article IX/Annexure F) and other secondary data (articles and documents).
Therefore, in line with Putnam (1988), we have interpreted win-sets for the Indus case as follows: 1. Resolution of water disputes based on the Indus water treaty. This treaty is a common agenda between the two parties at Level I for resolving water disputes so that the negotiations between the two states have been held based on the Indus Treaty's articles, concluded in 1960. 2. Continuation of the dam construction in accordance with the Indus water treaty. 3. Minor modifications to the technical specifications of projects, such as changes in dam design, changes in storage and changes in the power generation capacity.
We shall now apply the theory to our case study.

Pakistan's elements at Level II
Given the historically hostile and competitive relations between Pakistan and India, water has been politicized from the beginning of partition. Indeed, water is heavily politicized not only in India and Pakistan but in the south Asian region (Price et al., 2014). According to Putnam (1988), politicization is one factor affecting the size of the win-sets. Politicization often activates groups that are less concerned with the cost of disagreements, as a result, it decreases the size of the win-sets. However, as water has found a strong bond with food security and livelihood of a large number of the agriculturaldependent people in this region, it is related to national security in the region so that its crisis can be a source of instability for the countries. Therefore, along with tension over access to water and increasing water scarcity, securitization discourse has been also heavily enhanced in the region (Price et al., 2014). An issue may be framed within a nonpoliticized, politicized, securitized and violized range (Neumann, 1998). Politicized issues are part of public policy and need state decisions and allocation of resources or community management; while securitized issues are presented as an existential threat, requiring emergency actions and justifying actions outside the political process (Buzan et al., 1998).
Thus, securitization, over and above politicization, may also affect the size of the win-sets. Some scholars have argued that securitization can be a positive process in that in special circumstances, it has provided an institutional solution to environmental problems and enhance cooperation between the parties like India and Nepal in the case of the Tanakpur Barrage project (Mirumachi, 2013). However, the argument is not fully applicable to the whole of South Asia, particularly in Indo-Pakistan relations; while India-Nepal water relations have thawed, there has been acute hostility between India and Pakistan from the beginning of the separation of the two countries in 1947 and unresolved geopolitical disputes between them are serious barriers to positive cooperation.
In Pakistan, both state and non-state actors have securitized the shared water with India on the Indus River. Diversion of water and flood policy are two domains of threat perception among Pakistani actors. Water diversion has created a threat perception in Pakistan connected to the 1948 event that India stopped the flows of water in the canals on the eastern rivers of Ravi and Sutlej and it was only after the agreement of May 1948 that New Delhi allowed the water to enter into Pakistan again (Hill, 2006). Further, the intentional release of water has been often on the minds of Pakistanis as another threat (Malik & Tanveer, 2014) promoting securitization. Indeed, Pakistan is highly flood-prone. For example, during the 2010 monsoon season, flooding cost the country billions of dollars and impaired its overall economy (Looney, 2012).
Water has been securitized using anti-Indian discourse. Indeed, India's supposed antagonistic intentions against Pakistan constitute the most obvious variable legitimizing securitization in the Pakistani water discourse triggering Pakistan's various domestic players including the military, the extremist groups, politicians and also civilian groups and institutions.
Although Buzan et al. (1998) doubtless had state leadership in mind when they developed securitization theory, securitization does not need to be spoken only from an authoritative position to resonate. Security threats may be successfully invoked from a 'subaltern' position, for example, Litfin (1999) claims that non-state actors can securitize an issue successfully; Barthwal-Datta (2009) claims that 'there are well-informed and influential non-state actors raising issues as threats to security and therefore should be acknowledged as legitimate securitizing actors' (Barthwal-Datta, 2009). In Pakistan's case, civilian and technical institutions have engaged in the securitization process using anti-Indian discourse, too. In 2010, there was a mass demonstration against the Indian water activities launched by a farmers' collective, the Pakistan Muttahida Kissan Mahaz, with the participation of a large number of farmers, students, businessmen and civil society members from different places such as Vehari, Rawalpindi, Lahore and Multan (Dawn, September, 2009). They issued a joint statement saying that as India is 'stealing' the water from the western rivers through dam construction, 80% of the farmers had been adversely affected. Further, the Engineers Study Forum, a water expert panel, also released a report claiming 15-20% of water is 'stolen' by India costing Pakistan's agricultural sector $12 billion. The Punjab Water Council, a collective of farmers in Punjab has stressed the necessity of 'talking' with New Delhi about water to mitigate Pakistani worries about Indian hydroelectric infrastructures which could dry up Pakistan's rivers (Bisht, 2011). Thus, like other countries in South Asia, an agriculture-based economy as well as lacking the capability to manage water disaster risk impelled Pakistani actors to perceive water activities from upstream states as a threat to their security.
Therefore, in a region with territorial and (geo-)political rivalries, a securitization strategy may shrink the win-sets, reducing the likelihood of reaching an agreement; securitization creates conditions in which one county's water projects (India) are considered as threats to the security of another country (Pakistan).
In reducing the size of the win-sets, securitization in turn decreases the possibility of internal consent to an agreement. Likewise, Pakistani actors' demands opposing India's water projects on the Indus, mostly limited to halting India's activities on water projects reduced the size of Pakistan's win-sets. Indeed, securitization of Indian water activities seems to have reduced the negotiating flexibility on both sides of the diplomatic table. Securitization made it especially difficult for Pakistani negotiators to negotiate with India, given the strong metaphors and narratives levelled at India. For instance, any compromise on the part of Pakistan with India will not easily be accepted by the extremists who have a strong influence on their society and strongly used the 'war narrative' against the Indian water policy on Indus which can arouse people's feelings against India. Coupled with linguistic structures used to securitize water, extremists have been trying to create or strengthen the anti-Indian sentiments among the people in the flood-hit areas while distributing food and medicine there (Malik & Tanveer, 2014). This policy can enhance their influence over people and intense anti-Indian feeling among the general public. In the environmental realm, non-state actors (for example, civilian institutions and groups in Pakistan and extremists) can impact on the negotiations in a variety of ways, such as lobbying government officials, mobilizing people in protests, publishing reports to the media, challenging existing discourse and using legal courts to persuade subordinate actors to follow environmental laws (Warner & Zawahri, 2012). In this case, anti-Indian discourse has assisted the actors in mobilizing people for protestation. As there is no consensus on moderate thinking about relations with India in Pakistan, this is the win-set for the army and its related groups.
Securitization helps opponents to negotiation with India in disallowing proponents of negotiations with India negotiate with New Delhi on its terms. Pursuing this discursive strategy, opponents also create hurdles for India's dam construction (Pakistan's win-set) through controlling their officials who are in favour of negotiating with the Indian side.
Islamist groups including Jamaat al-Dawa, a political party, and the militant anti-Indian Hizbul Mujahideen organization are major actors in Pakistan politics. Due to the proliferation of religious parties into politics (Shahzad & Johnson, 2018) as well as political polarization (Shams, 2018), and coalitions of other parties with Islamist parties ( Jamal, 2018), the extremist Islamic groups have influenced the political arena and published radical narrative in the country, and captured the imagination of the common people (Weinbaum, 2011). Through the exploitation of the media 2 , the increase of its social services, and its capability to support affected people during natural disasters, the JuD has increased its influence even in the most distant part of the country (Roul, 2015).
According to two-level game theory, the strong order in the ruling party boosts the win-set by expanding the support of negotiators from domestic actors. In Pakistan besides the jihadist groups, another main player in politics is the Army. In this country, the polity is divided into political, military and religious groups. Political parties are weak in Pakistan; they are defined not by their policies but by particular individuals, groups and clans which are of great influence (Business Monitor International, 2014).
In Pakistan, unelected institutions like the military have taken control of policy processes. The Military has been active in foreign policy even when the executive power has officially been held by civilians. This has been repeated several times since the independence of Pakistan (Tudor, 2014). These circumstances have led to political instability which decreases the government's ability to adopt a coherent and active policy. There is little chance that a long-lived consensus can be forged due to severe competition between civil and military officials there. Elected civilian government has not been able to dominate security policy and take control of policy-making. The Armed Forces in Pakistan have directly taken control of the country for several decades, enhancing its influence not only over the state but over the political area and policy-making process, too (Rumi, 2015). Analysts claim the most significant structural obstacle to a strategic change in Pakistan's policy towards India is the continuing role of its influential army which dominated public discourse, too (Ahmad & Ebert, 2013). With the military defining India as the threatening 'other' (Ahmed, 2011), a long history of armed conflict and overt military clashes in 1947, 1965, 1971 and 1999, it has become natural for Pakistan to be concerned about India's actions as a long-standing enemy ( Jawed, 2019).
It is generally accepted in Pakistan that it is also in the corporate interest of the Pakistani military to take an uncompromising stance towards India (Tellis, 2017). Asymmetric power relations between the domestic players (the non-military ruling party and other players), with the army perching at the top of the power hierarchy in Pakistan, make the win-sets smaller. Indeed, in the water realm, as the water issue in Pakistan is linked to the issue of India's partition, and the Kashmir issue 3 (Bisht, 2011), the military considers both water and Kashmir issues very important; for example, General Ashfaq Kayani stated that as long as the Kashmir dispute and water issues are unresolved, India will remain the focus of Pakistani military (Gupta, 2010). So Pakistan's armed forces and extremists do not accept any negotiation on Indian terms with water because they consider water negotiation as an issue related to Kashmir. The collapse of negotiations on the Ratle project, after attacks by Pakistani groups, has been one example. Talks between Pakistan and India on this project ended in tensions after terrorist groups in Pakistan staged a series of terrorist attacks in India (Economic Times, 2017).
Therefore, there is no consensus on the issue of India between the non-military state, the army and extremist groups in Pakistan which impacts on the size of the win-sets. Lack of domestic consensus is a main barrier for Pakistan to reach an agreement on water with India.
Indeed, the more autonomy the chief decision makers from domestic players enjoy, the bigger their win-sets and the more the likelihood of reaching an international agreement (Putnam, 1988). In Pakistan, not all Pakistani leaders have fuelled the flames. Nawaz Sharif, former Prime Minister of Pakistan, believed that if Pakistan were able to make an agreement with India it could save significant amounts of money on defence and in turn, spend them on social and economic areas (Bray, 1992).
Nawaz Sharif sought to develop Pakistan relations with India to improve the languishing Pakistani economy (Business Monitor International, 2014). For example, on his visit to Turkey in 2017, he claimed that Pakistan wanted friendly relations with India. He said that his party did not pursue an 'India-bashing' policy even during its election campaign. Also, he stated that Pakistan supported stability in Afghanistan in its own interest (Dawn, 2017). Afghanistan's stability is also important for India's security since India has economic, political and social interests in post-Taliban Afghanistan (Wang, 2017). After his dismissal, there has been an extensive perception in Pakistan and abroad that Nawaz Sharif was ousted from the prime minister position for crossing the military's red line on Afghanistan, India and United States (Khattak, 2018). For example, Husain Haqqani, Pakistan's ex-ambassador to the United States claims that what caused the elected Prime Minister Nawaz Sharif's failure to complete his third term was not just corruption, but basically the military, as the Pakistani military clearly wanted an elected government that followed its policies on India, Afghanistan, jihadi extremism, China and the US (Haqqani, 2018). From the two-level game perspective, in international negotiations, domestic support is necessary for the Executive as 'major negotiators'. Obtaining the support of the domestic players can be easy for a senior official with a high political position that it is useful in the expansion of the win-set which as a result will strengthen the possibility of an agreement; and any leader who fails to meet the demands of the players on the domestic table runs the risk of being driven out of power (Putnam, 1988). But Pakistan's non-military state is not so powerful vis-a-vis the army on the issue of India in getting the approval of other actors easily. Therefore, this also reduces the size of Pakistan's win-sets which decreases the possibility of the internal approval of the agreement where support among the domestic players is lacking.

India's elements at Level II
Indian Hawks who have frequently accused their governments of not adopting an aggressive approach against Pakistani attacks are important actors in Indian politics (Kumar, 2016). PM Narendra Modi himself criticized the government, under Manmohan Singh, for its soft stance on the water issue against Pakistan in spite of Pakistani attacks, particularly in Kashmir (Bano, 2015).
In line with Putnam (1988), it stands to reason to assume that the senior negotiator of India generally gives priority to the domestic community, in particular, because his position and responsibility are often depended on his position inside the country. Therefore, the public criticism has made Indian government utilize a rigid approach towards Pakistan on water negotiations. Indeed, the government fears it would face adverse criticism from public and the opposition parties (the domestic players) and would lose its power position inside the country. So it is perceived by the government that any efforts to conduct negotiations with Pakistan would adversely impact on the government influence at the domestic level and weaken it. Thus, they selected a tough approach against Pakistan, which prevented the Indian Government from changing the original position (Kaushik, 2017). Therefore, public criticism impacted on the size of the win-sets and made them narrower.
Kashmiris are another domestic actor impacted (even if indirectly) on possible negotiations between the two countries. There have been formal complaints about the Indus Agreement in the Jammu and Kashmir state. To stem discontent, India's government in 2000 signed an MoU with the local government of the state to build hydroelectric power projects: Uri II and Sewa II (Kumar Sinha, 2008). In 2002, the Jammu and Kashmir state legislative assembly passed a resolution demanding the termination of the IWT treaty. Since then this treaty has been regularly debated in the state's assembly. Opposing Kashmiris have been repeatedly objecting to the 'hijacking' of their waters and asking the Indian central government to review the treaty and pay compensation to the state (Parvaiz, 2016). They believe the agreement has imposed limitations on the use of the river waters and also has adversely affected the state's agricultural production and its power generation, which led to the damage of its industrial sector and its economic overall growth and development (Nazakat & Nengroo, 2012).
Owing to Kashmiri protests in official and unofficial circles, the instability in this disputed region, the issues of national sovereignty, territorial integrity and the stability of the country are highly significant. As a result, India's strategic vulnerability in Kashmir is not only challenged by Pakistan but also by the local people and local authorities of Kashmir province. India has always faced challenges in Jammu and Kashmir state where it is of strategic significance for India. Thus, the Indian negotiators in the negotiation process need to pay attention to the demands of these actors in the region.
Therefore, activists who generally disapprove of the Indus treaty make narrow the size of the win-sets because they do not accept any negotiation easily but want to disrupt the treaty. In this case, the size of the win-sets may be reduced.
India needs to meet the needs of Kashmiris to reduce the instability in the region. On the other hand, although the suspension of IWT in acute crisis situations may be used as a strategy, it is not viable under relatively normal circumstances, as it would lead to India's condemnation on the international stage and face strong Pakistani backlash that would further instability and insecurity in the region. Thus, India also needs to negotiate with Pakistan to reduce tensions and avoid conflict, while also trying to keep domestic actors satisfied in Kashmir's volatile, sensitive and strategic region 4 . In fact, here both the lack of attention to domestic actors and the breach of the treaty are linked to the security and stability of the country. Therefore, in such situations, a country can partially overcome these pressures by fudging and buying time as Pakistanis officials characterized Indian negotiating strategies as: '(1) it was one of delay, of foot-dragging, of 'tiring you out'; (2) of 'creating facts'proceeding with construction plans, even when aware that the plans might well violate the treaty, so that Pakistan, confronted eventually with fait accompli, would have no choice but to cut its losses and accept an unfavorable compromise settlement; and (3) insisting on a bilateral framework of talks, without intending ever to settle on any but India's terms' (Wirsing & Jasparro, 2006).

India and Pakistan's negotiator strategies at Level I
On September 18, 2016, after an attack on the Uri area in Kashmir under Indian control and killing and injuring Indian soldiers, Delhi accused Islamabad of being behind that event. The attack caused an outrage in India, Indian politicians and pundits blamed the Pakistani government and urged the Modi government to take serious action (Rowlett, 2016). Shortly afterwards, due to the fury of Indian people, India decided to stop the Indus Water Commission negotiations until the 'Pakistan-backed terrorism' in India would end. So, India immediately suspended the India-Pakistan Indus Commission session. Subsequently, the Indian Prime Minister Narendra Modi publicly stated that 'blood and water cannot flow together' (Bagchi & Mohan, 2016). Indeed, the idea of the cancellation of the agreement was publicly announced in 2002 after the attack on the Indian Parliament by Pakistani radical Islamist groups. G. Parthasarathy, a former Indian high-ranking commissioner to Pakistan, suggested that New Delhi should call off the Indus Agreement, saying that 'We should make it clear to Pakistan that if it can bleed us in Jammu and Kashmir, we have the capability to starve them' (Kraska, 2003). Although, at that time, India kept to the treaty, this idea gained widespread support among Indian extremists and hard liners (Economy, 2008-9).
India's decision to suspend the Indus Agreement has led to increased tensions between the two countries, heightening the security discourse. The special assistant to the Pakistani prime minister stated that Pakistan would not accept any changes or modifications of the IWT (Qureshi, 2017). So, the threat to war was intensified in Pakistani discourse; headlines in the news channels of television and the newspapers highlighted the discourse that Prime Minister Modi has planned to 'steal' Pakistan's water and dry up the country due to intense tensions in Uri. Also, Pakistan declared it 'will not accept Indian aggression in any form, and any Indian step for disrupting water flow as (an) upper riparian state will pertain to considerable risk of war and hostilities' (Ahmad, 2016).
However, in accordance with the settlement procedures of the IWT which could involve the Court of Arbitration (Salman, 2008), Pakistan brought the dispute before the court. However, the World Bank brought the lawsuit to a halt to protect the IWT (World Bank, 2016), Thus, there was no effective result.
Indeed, the Modi government's suspension of Indus commissioners' meeting is indicative of India seeking to use the Indus Agreement as a bargaining tool ( Jacob, 2016;Kaushik, 2017). In this case, the negotiations tactic of 'issue-linkage' has been utilized as a form of bargaining power with distributive ends in the discursive interaction between the two governments. Coupling the revision of the Indus Agreement issue with the security-related issue of terrorismsubstantially supported the relatively stronger side's effort (India) to challenge the Pakistan support for Islamic terrorists at least temporarily.
The containment of terrorism is very important to New Delhi as around 200 areas in India are impacted by insurgents or terrorist groups. The Indian intelligence siege by Pan-Islamic fundamentalists has been a permanent concern for India (Singh, 2014). After September 11th, 2001, violent attacks on the Parliament of India and the hotels siege in Bombay (Mumbai) by Islamist groups indicate the vulnerability of India to terrorism (Osiur, 2015).
According to two-level game theory, issue-linkage policy can impact on the negotiation procedures. Indeed, linkages can be attempts to break a deadlock or otherwise improve one's bargaining situation on a specific issue by linking it to another, unrelated issue (Bow, 2009) as Gupta et al. (1995) explain: for fear of retaliation in other areas, linkages 'encourage compliance' and can help 'package deals and compensation'. However, linkage politics may also be aimed at complicating things; linkages can create deadlocks through 'negative linkage', meaning a bargain would be complicated due to 'unnecessarily adding an unwelcome condition' (Warner, 2016).
However, while Pakistan's concern for maintaining the Indus Agreement seems to provide an opportunity for India to bargain for making terrorists retreat from its borders to create stability in the area, control Kashmir and Kashmiris which lead to the continuation of its dam building in Kashmir (India's win-set), it failed to reach a conclusion; firstly, as 'war narratives' used by hardliners make Pakistan government face challenges to negotiate for terrorism with India. According to Cohen, the Indo-Pakistan conflict can be considered a 'paired minority' conflict that stems from the perceptions of important groups on both sides that feel threatened by each other. In 'paired minority conflicts', conflicts would gain their energy from the endless distrust caused by historical disputes; so the weaker party thinks it is vulnerable against a more powerful party and exposed to attack from it, and therefore tries not to make concessions while the stronger side believes it must not compromise as it has the ability to bend the other side to its will. So it is difficult for either side to compromise even on trivial issues, because this may be perceived by others as weakness, and invite further demands. Therefore, the responsible diplomatic team resists making concessions because compromising may be understood at home as a sign of weakness, or worse, as a sign of cooperation and collaboration with the hostile. The result is that the negotiating team on either side loses its capacity to solve problems it has with the enemy (Cohen, 2003). Thus, for Pakistanis negotiation on terrorism is not easy.
Secondly, from the Pakistani point of view, negotiating with India on terrorism issues through official channels would spill over to other significant issues of Kashmir and as a result weaken Pakistan's ability in the control of domestic issues (Kaushik, 2017).
Further, generally, Pakistan has always adopted a benevolent approach to extremist groups that have committed violent acts against Afghanistan and India (Milam & Nelson, 2013); the United States froze financial support to Pakistan citing the country's failure to take serious action against terrorist and extremist groups (Khattak, 2018).
Thus, in a region where countries have territorial and geopolitical disputes, linking water issues to nonwater ones which are of significance to both sides including territorial issues or other important issues, may decrease the size of the win-set and make it difficult for two sides to reach an agreement. This reasoning contrasts with some analysts' argumentation. Bennett et al. (1998) for example argued that players negotiating on separate issues can gain by linking the issue under negotiation (here: water) with an issue in another game (non-water) as the issue-linkage tactic increases the bargaining chip by allowing countries to condition the results of the water negotiations with the results of the past non-water games.
Indeed, India has changed its negotiation range by adopting an 'issue-linkage' strategy and linked the issue of Indus treaty to Pakistan's terrorism and reducing its win-sets range and Pakistan by rejecting the proposal as an unwanted condition (a negative linkage) reduced the size of its win-sets too, which made it difficult to continue collaborating on water-related issues linked to terrorism in this condition. So the negotiations face deadlock because the parties' win-sets do not overlap. The acceptance of the treaty is one of the basic conditions of the negotiations, and the negotiations between the two countries are based on the terms of the treaty so disruption of the treaty makes the win-sets of two sides fail to overlap.
On the other hand, in line with its long-standing geopolitical economic partnership, Islamabad and Beijing have cooperated with each other in the transboundary water area to confront against India. China and Pakistan concluded a Memorandum of Understanding for the construction of five dams along the Indus River which has worried India (Chandran, 2017). Further, China's hydroelectric projects on the main branch of Brahmaputra have been one of India's serious water security concerns due to their geopolitical-economic rivalries and territorial conflict ( Jiang et al., 2017). Thus, the week after India had suspended the Indus Commission due to the outbreak of Islamist attack on Uri, China initially shut down some water flows in the Brahmaputra tributary to start building a dam (Hussain, 2016). China went public with the construction of a huge dam on the Xiabuqu River near the city of Xigaze (Shigatse), a very strategic location near Bhutan and Sikkim; this project would cause dehydration in northeastern India and water tensions for India (Sehga, 2017), it would be a security threat to India due to its close proximity to Sikkim too. Sikkim was incorporated as the 22nd state of India in 1975 (Arora, 2006). China's investment in the construction of a dam near Xigaze (Figure 2) put pressure on India while it could release Pakistan from the crisis.
In fact, China has repeatedly stated its support to Pakistan, for example, the Chinese Consul General in Lahore, Yu Boren, said: 'In case of any [foreign] aggression our country will extend its full support to Pakistan' (Mahmood, 2016;Kaushik, 2017). As Kaushik (2017) argued the time of announcing the dam construction along the river Brahmaputra indicates that China has been pointing out that the India-Pakistan relationship is intertwined with a complicated geopolitical relation. Therefore, as India is a downstream country in the Brahmaputra, China as an upstream state can follow the Indian way on the Brahmaputra; development of China's projects along the river Brahmaputra may have catastrophic effects on northeastern of India. Perhaps the greatest threat to the confrontation of China against India can be perceived in the former foreign minister of Pakistan, Sartaj Aziz's statement: 'If India tries to interrupt water flow into Pakistan, it will not only violate the Indus Waters Treaty, but also set a regional state practice under which international law can be serve [sic] as a precedent. It will provide China, for example, a justification to consider of suspension of waters of the Brahmaputra River' (Dawn, 2016). As noted, Putnam (1988) claimed governments may look to spread each other's win-sets providing foreign aid to a friendly government. Here China's intervention into the dispute between India and Pakistan at that time made it difficult and costly for India to suspend the implementation of the Indus Agreement. Indeed, China and Pakistan strategic cooperation (alliance) in the water realm used to balance India has been one factor which impacted on the size of the win-sets. The use of strategic alliances, whether with the opponent or with other actors to improve bargaining power, belongs to leverage mechanisms in transboundary water interaction (Zeitoun et al., 2017).
Given Pakistan's economic weakness and the growing economic and military developments in India, Pakistan seeks to use its partnership (alliance) with China to control India and increase its own influence in the region and Kashmir to create hurdles for India's development process including water projects in the region and its influence over Kashmir (Pakistan's win-set). In the alliance strategy between China and Pakistan, China used issue-linkage policy against India. Indeed, the implicit threat of retaliation (issue-linkage strategy) posed by upstream China, the presumed hydro-hegemon of the region and Pakistan's strategic partner, to downstream India nullified the Indian issue-linkage strategy and has led to a reduction of the pressure on Pakistan and an increase in Pakistan's power, so that there is no need for Pakistan to negotiate with India based on her terms and conditions. Thus, there is no overlap between their win-sets. So China acted as a third country which has not directly attended the water negotiations of the two sides but engaged indirectly, reducing the size of the win-set particularly for Pakistan.

Conclusion
In the Indus Basin, both domestic and (geo-)political issues at the regional level have affected water interaction between India and Pakistan. Political conflicts with India since independence in 1947, coupled with domestic challenges have been the main drivers contributing to the securitization of water in Pakistan. Securitization, issue-linkage and alliance building (between China and Pakistan) acted as barriers to their water cooperation.
The state is not the only actor in decision-making in Pakistan's foreign policy; other actors including the army, Islamist parties and radical groups are engaged in its political arena. In the case of India, Indian politicians and pundits and the Jammu and Kashmir state government, politicians and water experts also have played the water game, impacting on the Indian water policies on Indus River, too. These domestic problems provide an important explanation why negotiations between the two sides have failed to reach a productive conclusion. The recent development in the status of Kashmir, now formally incorporated into India, is bound to complicate the water negotiations between the two countries even more (Dar, 2020).

Data availability statement
All relevant data are included in the paper or its Supplementary Information.