The introduction of nuclear weapons into any rivalry, let alone one as entrenched as that between Pakistan and India, is bound to change both the stakes and the rules of engagement. Paul Kapur and Sumit Ganguly proposed theories to explain whether the possibility of nuclear warfare has made the situation more or less stable. Yet neither scholar satisfactorily explains the muted Indian military response to the Pakistan-sponsored 2008 terrorist attack on Mumbai. The fact that India and Pakistan did not wind up fighting directly in the attack’s wake challenges the extent to which the presence of nuclear weapons is destabilizing. While there was no escalation, pessimistic theories still cannot explain the degree and lack of Indian mobilization.
Understanding this theoretical outlier is crucial for policy towards the subcontinent. This paper draws on Ganguly and Kapur’s theories about nukes in the Indo-Pak dyad and accounts of the Mumbai attacks to illustrate this dilemma. In the end, I argue that while Kapur and Ganguly both describe aspects of nuclear weapons’ impact on the rivalry, the Indian reaction to Mumbai can best be explained by combining the structural effects of nuclear weapons with the proximate impact of diplomacy by the United States and India’s civilian leadership’s lack of will for war.
To advance this argument I will first recount Ganguly and Kapur’s theories and offer a more satisfactory rejoinder. Then, I will describe the 2008 Mumbai crisis and apply my revised theory to explain the lack of escalation using proximate and structural arguments while considering its limitations. Finally I will conclude with the policy implications of my findings.
Much Ado about Not Much
Sumit Ganguly and Paul Kapur have engaged in a protracted theoretical debate about the impact of proliferation on the subcontinent, but alone both of their theories fail in key ways. While Ganguly essentially argues that proliferation has prevented escalation of conflicts and Kapur finds that nuclear weapons have destabilized the situation by promoting low-level conflict in the first instance, in reality both may have happened at the same time. While possession of nuclear weapons has allowed Pakistan to become more provocative at low levels of violence, sustained high level violence has become less likely because of the risk of nuclear escalation.
Ganguly argues that proliferation has made the subcontinent safer since the risk of crisis escalation has decreased following the introduction of nuclear weapons. Pakistan and India have engaged in full-scale war on three separate occasions—1947, 1965 and 1971—without ever resolving the core causes of conflict like the Kashmir dispute. Yet escalation has been capped since the countries began acquiring nuclear capabilities. India and Pakistan first tested nuclear weapons in 1998, the culmination of two parallel, and ostensibly covert, processes. Pakistan’s program was started by Zulfiqar Ali Bhutto, who sought to develop an “Islamic bomb” to ensure his country’s security vis-à-vis India. Both countries were aware of the other’s developments, though it was not necessarily clear what capability each possessed at a given time. Regardless, Pakistan followed India’s 1998 test with one of its own. Although each side successfully initiated atomic explosions, it was not obvious whether and how these weapons could be delivered or deployed. This constituted the formal introduction of nuclear weapons into the Indo-Pak dyad, after which Ganguly claims nuclear deterrence took hold of crises between the two.
Ganguly’s theory may have explained the situation well before that point. For instance, concerns about the nuclear capabilities of Pakistan and India capped the 1987 Brasstacks crisis. Pakistan precipitated the crisis by intervening to assist insurgents causing unrest in the Indian Punjab, leading India to launch large-scale military exercises called Brasstacks. Pakistan began exercises as well, and miscalculation that could have caused war was avoided by the United States and Soviet Union stepping in to resolve the dispute and end the conflict.
Though Ganguly does not directly assimilate this crisis into his theory of nuclear deterrence preventing escalation, he does mention that “A number of scholars and analysts have claimed that Pakistan delivered a veiled nuclear threat toward the end of the Brasstacks crisis.” India was most likely aware of Pakistan’s nuclear program, though unsure about the extent of Pakistani nuclear development. Ganguly’s thesis may ring truer earlier than he argues if the risk of nuclear retaliation contributed to India choosing to resolve Brasstacks peacefully. If the U.S. and USSR intervened with added urgency because of their own knowledge of Pakistani and Indian nuclear capabilities, then it would also bolster Ganguly’s argument.
Similar dynamics were at play in the 1990 Compound Crisis. Pakistani governmental intervention in a native-born Kashmiri insurgency transformed a local insurrection into a full-fledged revolt. Indian policymakers became worried after aggressive Pakistani war exercises and rhetoric were added to the mix. Concerned by the escalating rhetoric and risks of instability, the U.S. sent Robert Gates, then the deputy national security adviser, to the region. Gates was able to induce cooperation from the parties and deescalate the situation. Crucially, Ganguly comments that the existence of nuclear weapons in a volatile region brought in outside intervention—not for the last time—and helped de-escalate the crisis. He claims that nuclear deterrence alone has prevented escalation, yet it is clear that the side effect of increased international attention can also play a role in preventing war. Yet Ganguly concludes that nuclear deterrence has continued to work even when the diplomacy it spurs has failed.
After each side tested nuclear weapons in 1998, a conflict between Pakistani and Indian troops touched off in Kargil. Pakistan occupied empty Indian posts, eventually crossing the Line of Control—the de-facto border in Kashmir—with regular and paramilitary soldiers. India was caught off guard and had an initially lackluster response, but eventually escalated with air and artillery assets and regained some initiative. Ganguly claims that even as Pakistan threatened nuclear use—a threat relayed implicitly through overt possession as much as by clumsy rhetoric—India focused on avoiding escalation by winning the conflict conventionally, which it did. Despite attempted intervention by the United States, the conflict continued to simmer and Pakistan refused to accept responsibility for the incursions. Yet India refrained from escalating inside of the Kashmir theater—carefully avoiding crossing the LOC with its planes—and chose not to escalate horizontally across the international border as it had in 1965. Ganguly says only nuclear deterrence, not failed international diplomacy, explains India’s decision not to escalate. India had the means to escalate and an operation would have been passable domestically, yet it chose not to act. Restraint flowed directly from the risk of nuclear escalation.
According to Ganguly, similar dynamics were at play in the subsequent Twin Peaks crisis. Against a tense backdrop, Jaish-e-Mohammed attacked the Kashmir state legislature and months later attacked parliament in New Delhi. Even though India mobilized troops, had the means to strike Pakistan across the international border, and Pakistan refused to bend to Indian pressure to hold state-sponsored groups accountable for the attack, it did not start a conventional conflict. U.S. desire to retain Pakistan as a partner in the war on terror also limited U.S. pressure. Six months later, Pakistani proxy Lakshar-e-Taiba (LeT) launched an attack against an Indian military base, yet India avoided escalating. India continued mobilizing its military, but chose not to launch an attack. Even though the international community had not punished Pakistan and India had the assets and international and domestic legitimacy to strike, it refrained. This leads Ganguly to conclude that Kargil and Twin Peaks demonstrate that “nuclear deterrence is robust in South Asia” as both “were contained at levels considerably short of full-scale war.” He reasons that proliferation was unequivocally good for stability on the subcontinent, since the end of the Cold War made increased belligerence by Pakistan inevitable, and only the risk of nuclear war has prevented Indian escalation during crisis and confrontation.
While Ganguly claims nuclear weapons have made the subcontinent safer, Kapur thinks they make the dyad more dangerous by increasing Pakistani aggressiveness towards India. Given that Pakistan is a revisionist state and conventionally weaker than India, Kapur observes that the risk of nuclear escalation opens up possibilities for Pakistan to pursue its objectives by limiting Indian options during a crisis. This is what Kapur calls the instability-instability paradox—it is harder for India to punish Pakistani low-level violence since escalation runs the high and deadly risk of nuclear force. Pakistan takes advantage of India’s bind by attempting to act on its claims in Kashmir, which sets off Indian retaliation and risks escalatory conflict. Therefore, Kapur claims that nuclear weapons have increased the risk of conflict on the subcontinent.
There is also a progressive component to Kapur’s argument: as India and Pakistan have proliferated, the tendency for Pakistan to initiate insurgent or terrorist violence under the nuclear umbrella has increased. Using Correlates of War Data, Kapur finds that the number of months of interstate disputes between India and Pakistan has increased as the Pakistani and Indian nuclear programs have moved from opaque to overt. The claim is causal—it has become easier for Pakistan to start conflict because the price it has to pay for doing so has decreased over time. While states often start limited wars for limited aims, the causative relationship between nuclear weapons and conflict solidify Kapur’s argument.
Kapur demonstrates his claim by referring to many of the same conflicts as Ganguly. Kargil and Twin Peaks are both Pakistani provocations enabled by a perception of nuclear weapons tying India’s hands. He notes that Pakistan became even bolder—explicitly involving Pakistani regulars in its operations against India, for the first time since 1971, during these crises because it felt protected by its now-overt nuclear deterrent. While Kapur concedes that nukes may have prevented escalation during these conflicts, he claims that this was luck that will not necessarily hold in future conflicts. Moreover, he points to circumstances on the ground like surprising Indian conventional success in Kashmir during Kargil—not nuclear weapons—as precluding the need for further conventional escalation. Kapur concludes that nuclear weapons have left the subcontinent worse off by progressively increasing the amount of conflict.
The flaws in Kapur and Ganguly’s logics are exposed when they attempt to refute each other and argue that nuclear weapons are unequivocally good or bad for the Indo-Pak dyad. In fact, nuclear weapons have had multiple effects at different levels of stability and there is no real contradiction between the scholars’ claims. There are fundamentally different dynamics at play at the lowest and highest levels of violence and the same change—proliferation—can have different effects on the likelihoods of insurgent or terrorist violence and all-out war without nuclear weapons being good or bad for the risk of conflict in the absolute. Proliferation has progressively increased the risk of low-level conflict, so Kapur is right. At the same time, the presence of nuclear weapons has constrained India’s options in responding to provocations and decreased the risk of total war, so Ganguly is too.
Beyond asserting that their views contradict, Ganguly and Kapur each advance an unconvincing argument to refute the other. Ganguly claims that Kapur conflates the cause of Pakistani belligerence, which he declares to be the end of the Cold War. In the face of the causative or at least correlative data offered by Kapur, this claim breaks down, especially since the trend line is visible back into the early 1980s during earlier stages of proliferation, not just after the Cold War. Kapur takes his stab at Ganguly by claiming that escalation has been avoided by circumstantial factors like conventional Indian victory as much as fears of nuclear escalation. While the risks of miscalculation and accidents mean that nuclear weapons can never guarantee escalation will not occur, Ganguly’s examples of Indian restraint strategically and tactically demonstrate that the risk of nuclear escalation has clearly impacted Indian decisions during crises in a way that has decreased the chances of full-scale conflict. Pakistan has also showed restraint by not giving militants advanced anti-aircraft or missile capabilities.
Usually the truth between two polarized views lies somewhere in the middle, but it is not clear that these theories are polarized. They are probably complementary as they describe the varied effects of proliferation on disputes between Pakistan and India. The theories do describe the lenses through which each country views nuclear weapons in the conflict—Pakistan sees an opportunity to advance its goals through irregular violence without repercussions while India looks for ways to leverage its conventional superiority without triggering escalation. As the conflict has gone on, each side has and will continue to devise strategies—tactical nuclear weapons for Pakistan and Cold Start with lessened but significant reliance on international pressure for India—that allow them to use their asymmetric assets to advance their own goals.
Pakistan will continue to try to use the threat of nuclear exchange to limit Indian options for retaliating against terrorist and insurgent attacks, while India will do all it can to sketch clear redlines that preclude Pakistani aggression, whose enforcement would be credible, automatic, and devastating without nuclear use. Meanwhile, improvements in nuclear signaling, clarification of redlines, and better rhetoric will militate against escalation. This process still has dangers and it is doubtful that Pakistan has or will soon be persuaded to give up militant aggressiveness shielded by nuclear weapons as an instrument of foreign policy.
Mumbai: Failure to Launch?
For all the ink Ganguly and Kapur spilled on their theories, neither deals with the 2008 Mumbai crisis well: Pakistani-sponsored terrorists attacked India in a spectacular manner, Pakistan refused to take responsibility and punish the perpetrators, and despite high tensions and military maneuvering no retaliatory attack took place. While Ganguly would correctly predict escalation would be capped, he cannot explain why India did not even mobilize in response like during Twin Peaks. Kapur anticipates the initial terrorist attack, but if instability is an attack without retaliation, his warnings are less ominous and absent a massive number of attacks, the total months of conflict will start to trend down. Clearly Kapur’s theory also cannot forecast the size of India’s response. The rejoinder outlined above would not necessarily have predicted the Indian response to Mumbai either, but it does explain it better than the other two. A sketch of the Mumbai crisis and investigation into the causes of Indian behavior will show why.
Beginning on November 26, 2008, a group of militants with Pakistani origins and support attacked two luxury hotels, the central train station, and a Jewish center, among other targets in Mumbai. As Polly Nayak and Michael Krepon document, the shock value of the attacks and their coverage on traditional and social media make them an Indian analogue to America’s 9/11. Security forces took 3-4 days to end sieges of the targets, leaving 160 or so dead. Indian police responded poorly giving attackers added time to kill more individuals within their targeted sites. Indian leaders did not instantly know the attacks’ gravity and Pakistani origins, but after intelligence intercepted communications between militants and Pakistan-based commanders and interrogated a captured suspect it was clear that LeT was behind the strikes. While Indian and American officials could not be sure Pakistan’s intelligence services—which train, fund, and often organize the group—had knowledge of the attacks beforehand or directed them, the Indian public began pressuring the government to retaliate against Pakistan. The Indian government had to weigh the option of attacking Pakistan while fearing that further militant attacks could be in the offing. A public outraged over the attacks and the burden of preventing more created a high-pressure scenario for Indian leaders.
U.S. officials in Mumbai, New Delhi, Islamabad, and Washington, D.C. began to focus on preventing the conflict from escalating soon after the attacks. Security officials became worried about Pakistan and India repeating Kargil or escalating further. A lack of intelligence in the first few days left attribution to Pakistani proxies to be a matter of reasonable guesswork for the U.S. and India. The benefits of this plausible deniability for Pakistan were somewhat limited as the U.S. and others still had to engage in rigorous high-level diplomacy—including calls by President Bush and President-elect Obama to their counterparts—to calm leaders and the public on both sides. Western leaders had to talk Pakistan out of preempting feared Indian military action. Despite Prime Minister Singh’s disinterest in starting a war, public and military pressure plus upcoming elections meant U.S. leaders needed to focus on preventing Indian escalation. This risk was acute given Indian Cold Start doctrine designed to enable a quick and irreversible military response to Pakistani militancy. It would have been difficult to stop the process if it had started.
To help calm the situation, the U.S. began cooperation with India on addressing this attack and improving its counterterrorism policies in general, bolstering relations between the two. This included an investigation into the perpetrators of the attack, the results of which soon pointed towards Pakistani groups and eventually government officials. While investigations made clear recently elected Pakistani civilian leaders probably did not know about or authorize the attacks, it was less clear in the case of the more powerful military and intelligence. U.S. dialogue with the military was more productive than with civilians and made Pakistan less likely to initiate military conflict.
However, Pakistani cooperation on arresting or punishing LeT members was extremely limited. The U.S. and Indian expectation that Pakistan would hold responsible a long-time proxy seems odd, but the level of cooperation was still lower than they had hoped. This remained true even as new evidence made clear the attack was perpetrated by Pakistanis from Pakistan with government ties. Though civilian leadership may have been willing to punish LeT members, the military and intelligence determined crisis behavior. Those groups began denying that Pakistan played any role in the attacks—walking back previous positions—and made no effort to harm their proxies.
India continued to use harsh rhetoric without taking military action. The greatest risk of escalation came between December 15, 2008 and the end of the month. Publicly available evidence of Pakistani intelligence’s involvement in the attacks added to the pressure faced by Indian leaders. India began military exercises and it was unclear if they were part of Cold Start. Pakistan and India continued brinksmanship that could have easily have led to incidental or intentional escalation. Both sides trumpeted high alert levels and Pakistan threatened to retaliate against even limited strikes. India seemed intent on pressuring Pakistan to arrest members of LeT, but in spite of the military posturing and pressure this could not be accomplished. Indian officials may have postured to please their public but could have been prevented from backing down by those who warned against coming off as weak. Despite pressure from within and outside of the government from the public and former officials, Singh avoided escalation and had his methods praised as “mature” by U.S. officials. The situation de-escalated unceremoniously as India reduced its demands and Pakistan made efforts at taking short term, minimal responsibility. India and Pakistan never fully mobilized, and the crisis gradually faded away.
There are a few explanations for Indian restraint during the conflict. The first is that U.S. high-level dialogue and the investigation into the attack created time, space, and calm for Indian leaders. Engagement may have reduced Indians’ “sense of crisis” and allowed officials who were already “genuinely conflicted about how to respond”—balancing avoiding nuclear exchange, undermining a weak Pakistani civilian government, and seeming weak themselves—to realize how limited their options were and decide not to escalate. Despite efforts to craft a Cold Start doctrine for these exact scenarios, India realized it had no good strategy for dealing with the attacks without incurring unnecessary risk. By creating time, U.S.-led dialogue helped the pacifist impulses win out. Singh’s “maturity” is also an important variable here. A more trigger-happy or uncertain leader may have felt the need to respond kinetically in some manner.
The second explanation entails India’s lessons from past crises like Twin Peaks and Kargil. Nayak and Krepon argue that “New Delhi learned during the Twin Peaks crisis that its nuclear arsenal did not deter terrorist attacks and that mobilization of its conventional forces did not compel corrective actions by Pakistani authorities.” Rather than mobilize and then back down—India was convinced the U.S. would not let war break out—Indian leaders avoided starting the cycle. This is not entirely separate from the first reason, as it illustrates why Indian leaders felt as though they had no good options. Past crises had taught them that they had no tools with which to compel Pakistan and achieve their goals.
Applying a similar argument to Nayak’s and the rejoinder above, Vipin Narang argues that Pakistani threats to use nuclear weapons inhibited Indian responses more than U.S. diplomacy. For Narang, Mumbai demonstrates that Pakistan’s nuclear possession deters India from any large scale military response to militant attacks. The risk of escalation—aided by unclear Pakistani redlines for nuclear use—limits India’s options. Further, he asserts that India could shirk U.S. pressure if it so desired, like it did in 1998, and therefore did not escalate only because of Pakistani nuclear possession. Along the lines of the rejoinder, Narang concludes that, “[T]he adoption of an asymmetric escalation posture—not simply acquisition of nuclear weapons—seems to be the critical shift that erected a shield behind which such [revisionist territorial] preferences have been more assertively pursued.” Narang claims that since overt proliferation in 1998 Pakistan has felt bolder in using militants and India less certain about how to respond. Narang’s observations support the rejoinder outlined above, though using the three crises from 1998 to track a trend seems questionable, especially because of Kapur’s correlative data. India also responded kinetically in Kargil in spite of Pakistani nukes, so it might be better to argue that India is most deterred from responding in cases in which it cannot do so proportionately.
Conclusions and Implications
Any attempt to explain this crisis with one theory—diplomacy, Indian leadership, or nuclear deterrence—is intellectually irresponsible and greatly understates the tremendous instability of this deterrent pairing and the significant danger it still poses. The rejoinder of Kapur and Ganguly outlined above does offer the best jumping off point for understanding the dynamics that were at play in Mumbai. Pakistan felt comfortable launching the attack because it had nuclear weapons and could deter Indian conventional retaliation. India felt uncomfortable retaliating because of unclear redlines and the risk of nuclear escalation. International diplomacy, led by the U.S., bought India the time and cover to avoid a military response. Along with Singh’s disinterest in starting a war, these factors helped avoid escalation.
The truth behind each of these statements is more complex. Pakistan has felt more comfortable using militants with the improvement of its nuclear weapons program, but in the last ten years Mumbai has been the only major attack. Periodizing Pakistani proliferation in a certain way could still make it seem like the period since 1998 has been more violent than those that came before it. However, if the argument is true that Pakistan feels more comfortable using jihadists than it has before, it is unclear why they would exhibit anything that resembles restraint. Brasstacks and Kashmir were closer together than Twin Peaks and Mumbai, yet Pakistan should now be more confident in its ability to use jihadists.
That paradox shines light on the issue in claiming that escalation has been deterred. In fact, Pakistan does not think escalation is impossible and India has not taken it off the table. While parts of the Pakistani military and intelligence organizations feel like they can initiate asymmetric attacks with fewer consequences than before, the frenzied manner in which Pakistan blurs redlines and postures to deter any Indian retaliation—in the wake of Mumbai and as a broader strategy—demonstrates that Pakistan has “probable hope” and not “certainty” that India will not respond.
Escalation is also not off the table for India. It has to be ruled out during every crisis. Even if nine times out of ten Indian leadership concludes Pakistani proliferation prevents retaliation, saying escalation has been deterred is not a certain prediction. There was a military response to Kargil and there have been mobilizations since. Certainty of any future Indian response requires an overly optimistic reading of past crises and ignores the dynamic risks and chances for miscalculation each crisis presents. This is why Kapur and Ganguly struggle to account for variance in Indian action in crises. Indian responses have become more muted over time, which may be the effect of nuclear deterrence. Yet that should increase Pakistani militants’ use, and it has not increased in the last decade. There is a chance the risks of the situation have deterred Pakistan as well, though the data to verify this claim is beyond the scope of this paper.
International diplomacy is not a silver bullet either. India clearly benefits from pressure on Pakistan that allows it to claim the moral high ground and avoid a military response. The time and space intervention produces also helps avoid conflict. Still, concluding from Mumbai that the U.S. has the effective means by which even similar Indo-Pak crises could be prevented from escalating is foolish. There are interests both countries would fight to defend and at best diplomacy could talk one side down from defending a gray area. U.S. interests and effort have been fairly constant since Brasstacks, and while U.S. methods may have changed, they are not enough to explain the variations in outcomes. Additionally, were someone other than Singh Prime Minister of India, the outcome of the crisis could be entirely different.
In the case of Mumbai, uniquely restrained leadership and international diplomacy accompanied by structural factors combined to prevent escalation. Neither proliferation optimists nor pessimists can reliably predict any Indian response, let alone one this restrained in the face of spectacular provocation. While a rejoinded argument that recognizes both how nukes make Pakistani jihadist use more likely and how nukes deter Indian retaliation comes closer to explaining Mumbai, causes like diplomacy and restraint by Indian leaders are also needed. Even the most synthetic explanation cannot satisfactorily deal with Mumbai. The extreme lack of Indian retaliation is not something that can be reliably predicted even by a combined theory. There are simply too many circumstantial factors that vary from crisis to crisis. Different targets, death counts, or leaders in any country involved could produce different results.
The real lesson from Mumbai is that proximate solutions cannot be relied upon to prevent Indo-Pak conflict. Reactive diplomacy, in spite of nuclear deterrence, is no guarantee against escalation in the future. Pressure could mount that no Indian government could shirk. Either side could miscalculate and strike first, even if they knew they lacked “good options.” Going forward, policymakers on and outside the subcontinent must do more to deal with the origins of this scenario for nuclear escalation. Stabilizing deterrence by improving nuclear safeguards and communications between the countries could help. Such measures could reduce the risk of nuclear escalation, but unless Pakistan can somehow be prevented from using jihadist proxies against India there will continue to be an unacceptable risk of a crisis getting out of hand.