According to one senior air force official, “In the early 1990s, the air force was thinking of one-way missions. . . . [I]t was unlikely that the pilot deployed on a nuclear attack mission would have made it back.”
A senior Indian defense official privy to this effort disclosed to the author that, until [Kargil in 1999], the air force had no idea (1) what types of weapons were available; (2) in how many numbers; and (3) what it was expected to do with the weapons.
He makes a conceptual distinction (p79-80) between a nuclear device (“an apparatus that presents proof of scientific principle that a nuclear explosion will occur”) and actual weaponization (“building compact reliable rugged weapons and mating them with delivery vehicles”), and then asks why the “process of weaponization in South Africa, India, and Pakistan took eight, fifteen and ten years, a nearly twenty-eight-fold increase on average” when compared to the P5 states. India conducted its first nuclear test in 1974, but only developed a weapon much later.
His answer is secrecy (p81-2):
Indian political leaders feared pressures for nuclear rollback from the United States. These pressures pushed the weaponization process underground, deep into the bowels of the state. To safeguard secrecy, policy planning was weakly institutionalized. Sensitive nuclear weapons–related information was tightly compartmentalized and hived off within an informal social network consisting of a small number of scientists and civilian bureaucrats. Secrecy concerns prevented decision-makers and policy planners from decomposing problem sets and parceling them out simultaneously for resolution to multiple bureaucratic actors, including the military …
In the absence of holistic planning stretching back to the 1980s, many technical problems, particularly those related to the integration of weapons with combat aircraft, were only partially anticipated. In other instances, policy planners remained unaware of the technical challenges until they demanded resolution. All of these factors became roadblocks on the path to weaponization. Secrecy concerns similarly prevented policy planners from institutionalizing the soft organizational and training routines between the scientific and military agencies necessary to move weapons from the stockpile to the target, in effect attenuating the state’s capacity to make good on its insinuated threat to punish a nuclear aggressor via a retaliatory response.
Kampani’s article is based on a large number of interviews with high-level Indian civilian and military officials associated with India’s nuclear weapons programme. As such, he reveals a number of new details that are of great interest whether or not one agrees with his theoretical explanation.
On India’s ability to deliver nuclear weapons (pp94, 97-8, 101):
The DRDO first conducted trials in the early 1980s to test the Jaguar combat aircraft, which India had purchased from Britain in the late 1970s, as a potential delivery vehicle … Having found the Jaguar unsuitable because of the low ground clearance between the aircraft and the nuclear weapon container, DRDO next identified the Mirage 2000 as its choice for a delivery system …
The trouble, recalls another senior air force officer who served at the time, was that “the boffins developed it [the device] independently without reference to the delivery platform. There was a problem with carriage because the weapon was too long.” This was cause for concern especially during the “rotation maneuver during the takeoff stage. A skilled Mirage pilot could have pulled it off . . . but not just any pilot,” a senior air force officer with an intimate view of the program told the author. The “size of the weapon itself, its length and weight upset the aerodynamics and center of gravity of the aircraft.”
Other aspects that needed resolution were the aircraft’s electronic interface and sighting systems to enable the arming and release of the weapon. The electronic interface could not be reconfigured without what one air force officer described as access to the “manufacturer’s database” and computer source codes. The aircraft also required extensive rewiring for electrical connectivity to enable the bomb’s functions. The Mirages that India had acquired from France in the mid–1980s were not nuclear certified. There were thus concerns that a post-detonation electromagnetic pulse could interfere with the aircraft’s computer-controlled fly-by-wire, communications, and other electronic systems.
According to one senior air force official, “In the early 1990s, the air force was thinking of one-way missions. . . . [I]t was unlikely that the pilot deployed on a nuclear attack mission would have made it back.” … Until 1994, DRDO conducted experimental modifications on just one Mirage 2000 with a single test pilot. There was no backup .. [but] an Indian air force study conducted in the early 2000s highlighted the logistical challenges of planning nuclear missions against Pakistan. It showed that a single mission alone could tie up as many as sixty aircraft to assist the penetrating nuclear aircraft.
Kampani concludes that “the author’s interviews with several senior retired Indian air force officers at the highest levels suggest that India achieved an air-deliverable capability sometime in 1995” (p86), a much later date than is usually assumed in the literature. He also notes (p97, footnote 73):
There is evidence to suggest that Indian defense agencies were able to modify warheads for ballistic missile delivery by 1996–97. It is uncertain if the systems met operational standards of reliability, however. According to a former commander in chief, Strategic Forces Command, who spoke to the author on the condition of anonymity, even as late as 2003–04, combat aircraft were the most flexible and reliable nuclear delivery systems India possessed.
On civil-military coordination (pp102-3):
[P]rior to 1999, the air force did not know who possessed the codes for arming nuclear weapons and how those codes were to be deployed during a mission. Indian weapons at this time did not incorporate permissive action links that would enable arming the weapons at will. The assumption in the air force was that the task of arming the weapon would fall on the pilot at a designated time during oight. The air force and the scientific agencies, however, did not conduct practice drills to test the communication and weapon arming protocols during a potential nuclear mission …
As the Indian government secretly prepared for an all-out war with Pakistan [during Kargil, in 1999], the spotlight turned to the nuclear aspect and the lack of operational planning with the Indian military. A senior Indian defense official privy to this effort disclosed to the author that, until then, the air force had no idea (1) what types of weapons were available; (2) in how many numbers; and (3) what it was expected to do with the weapons. All the air force had was delivery capability in the form of a few modiaed Mirage 2000s. At that point, only the air chief, the vice air chief, and two other individuals at Air Headquarters had knowledge of the program.
On dealing with nuclear attack in the 1990s (p100):
In the event of the prime minister’s incapacitation, power would devolve upon the Cabinet Committee on Security, but the likelihood of that event happening was thought low. A Pakistani nuclear attack, the officials believed, would be limited and symbolic and leave the functioning of the federal government relatively undisturbed.
But in the worst-case “bolt-out-of-the blue” scenario in which Delhi did go up in a mushroom cloud, power would devolve upon a hierarchy of state governors and principals in the state civil service who would assume responsibilities of the federal government, while the military would function under a reconstituted civilian authority. India, the leaders of the nuclear network believed, was a “big country. It would survive!” But how, they could not tell.
Similarly, a spare oral brief was made to new holders of the prime minister’s office. If, however, they were deemed disinterested [!!!], and at least three incumbents in the 1990s were, their principal secretaries were briefed instead. Beyond prime ministers and their principal secretaries, no information was shared with ministers on the Cabinet Committee on Security or with federal governors and provincial civil service chiefs who might be called to assume responsibilities.
Kampani rejects the view that India’s slow weaponization was to do with India’s civilian-dominated civil-military relations (pp108-9), as others have argued:
If civil-military institutional tensions were the cause, however, one would see greater aggregation of information among civilians, but the regime of information scarcity operated with nearly equal severity on both the civilian and military sides of the nuclear equation. As a senior Indian defense official at the heart of the nuclear network put it: “Yes, the military was kept out of the information loop. There were no serious reasons to bring the military into the loop because of the danger of secrecy being compromised. The chiefs of staff are trustworthy. But who can vouch for the trustworthiness of their staff, their drivers? The latter could be spies and the weak link in the chain. The military’s complaints have more to with a sense of privilege and pride. Why should they be told? The cabinet ministers weren’t told, the defense minister, their political boss was not told. So why should the armed services chiefs be told?” … More significant, India’s civilian leaders have shown little hesitation in institutionalizing the military’s role in nuclear planning post-1998, once India stepped out of the nuclear closet. This change has occurred without any fundamental rewrite in the DNA of India’s civil-military relations.
The implication of all this, of course, is that as India’s nuclear weapons programme institutionally matures (see my post last week and Shyam Saran last year) and internal secrecy diminishes, the assumed civil-military or cultural constraints will not be as constraining as is assumed.
 Gaurav Kampani, “New Delhi’s Long Nuclear Journey.” International Security 38, no. 4 (April 1, 2014): 79–114