Sleepwalking With the Bomb (36 page)

In the 1950s, the Strategic Air Command introduced the Special Weapons Emergency Separation System (SWESS). If an aircrew were disabled during an attack while over enemy territory, SWESS would automatically release bombs on board, once the plane fell below a specified altitude. This became known as the “dead-man’s switch.”

For 30 years the Strategic Air Command kept an airborne command post aloft, originally named National Emergency Airborne Command Post, or NEACP (pronounced “kneecap”), and ultimately called Looking Glass. This ever-flying patrol enabled the U.S. to retaliate, even after a surprise attack. When the Cold War ended, so did the 24-hour airborne command post patrols.

M
UCH HAS BEEN MADE OF THE “ACTION-REACTION” INTERPLAY OF
the superpower arms race and of imaginary missile gaps allegedly invented to spur the U.S. arms buildup. The truth is more complex: what many observers (and even some senior-level policy makers in various administrations) thought was conspiracy was in fact, as is usual in human affairs, a case of blunder.

In 1976 Albert Wohlstetter proved that American intelligence estimates consistently underestimated Soviet deployments, with even high-end estimates often below the actual Soviet numbers.
Far from improving with experience, these estimating errors grew worse with time
. Convenient assumptions guided intelligence policy, rather than logical inferences from incoming evidence.

Intelligence underestimates beginning in the mid-1960s are best understood against the earlier overestimates of the late 1950s and early 1960s. Then, the focus was on a “missile gap”—the idea that the United States, with some 200 ICBMs, had fewer than the Soviets. Because these early overestimations of the Soviet arsenal have been a huge marker in nuclear policy debates for half a century, they merit closer examination.

In
The Wizards of Armageddon
, Fred Kaplan discusses “the gap that never was.” In 1960, President Eisenhower’s last year, intelligence analysts projected that the Russians would field 50 to 200 ICBMs in the early 1960s. But U-2 reconnaissance flights—flying up to 70,000 feet above land—revealed no ICBMs. On August 10, 1960, the air force launched the first Discoverer satellite (circling in highly elliptical polar orbits, these satellites swoop low—a few hundred miles
⁵⁶
—over the target area). Only then did the U.S. acquire the ability to cover all of Soviet Russia’s vast territory, which spans 12 time zones. The new satellite found only four ICBMs, sited at Plesetsk in northeast Russia. In February of 1961, Kennedy’s secretary of defense, Robert McNamara, concluded that there was and had been no ICBM missile gap (or rather, the gap went the other way—America had more ICBMs than Russia). The issue that the new president had flogged so successfully in his campaign was mooted. In June, the CIA issued an intelligence estimate for 1961 that said the Soviets might have up to 50 to 100 ICBMs—and potentially up to 200 by next year (the high end of their earlier 1960 estimate, which they only repudiated that September).

In his landmark book,
One Minute to Midnight: Kennedy, Khrushchev and Castro on the Brink of Nuclear War
, author Michael Dobbs writes that in 1962 the Pentagon estimated the Soviets had 86 to 110 ICBMs (versus our own 240), but that the actual Soviet total was 42. Surely a contributing factor was Khrushchev’s public bluffing as to how the Soviets were growing ICBMs like sausages, while privately telling his son Sergei that the USSR had little of either product.

However, Paul Nitze explained that there was a second gap: Russia led in medium- and intermediate-range ballistic missiles. These were the main spearhead of Russian missile deployments in Cuba. The threat posed by these missiles guided U.S. policy during the 1962 crisis. Nitze also cited a Soviet budget expert’s assessment that the beginning of a 25-year Soviet strategic force buildup began at least a year before the Cuban Missile Crisis. The early intelligence overestimates of Soviet ICBM deployments surely were a large factor in later intelligence underestimates, via the classic pendulum swing that often follows major organizational failures.

Nitze points out that the theory of National Intelligence Estimates (NIEs) is that they look to non-U.S. capabilities only, and do not attempt “net assessments”—those based upon comparing forces. Yet forward net assessments became what most NIEs did, due to bureaucratic biases in favor of trying to look ahead. It was a task rarely done well, due to biases built into assumptions.

__________________

56.
Anything less than a thousand miles above Earth is “low” for a satellite, whereas a hundredth of that is high for a plane. Spy planes like the U-2 (and its faster-flying successor, the SR-71 Blackbird, which could cruise at 85,000 feet, or 16 miles above earth) are among the highest-flying manned non-rocket aircraft. The SR-71 was retired in the 1990s.

T
HE FUNDAMENTAL CONCEPTS OF NUCLEAR ARMS CONTROL WERE
developed in the West long before small powers of questionable stability came into possession of nuclear weapons. Calibrated to the threat from a hostile superpower, decisions taken 40 years ago created a mindset that persisted past the demise of that superpower 20 years ago.

Failure of the U.S. to deploy an effective missile defense against a small-power attack is a product of superpower arms control. Binding arms-control constraints began with SALT I in 1972. The Anti-Ballistic Missile (ABM) Treaty severely limited missile-defense design and deployment in the United States. Defensive system design since then has aimed not for the best products that technology and innovation can produce. Rather, system design has been governed by the maximum technological result deemed permissible under strategic arms-control principles as they were understood forty years ago. The result has been systems of perilously stunted capability, making a successful strike by a small power achievable.

Just how this came to pass teaches a crucial lesson in arms-control efforts: how limitations that to many appeared reasonable in one strategic context—the Cold War face-off against a massively armed superpower—proved obsolete and even dangerous decades later, when emerging powers in possession of or seeking small arsenals of far less sophistication menace the free world. Missile defense against 1,000 ICBMs might never work within the limits of existing technologies; defense against 10 or 20 ICBMs might work.

Missile defense became inextricably intertwined with MIRV—multiple independently targeted vehicles (warheads). Put simply, the more warheads could be directed at targets, the harder it would be for defensive systems to intercept them. During the late 1950s and early 1960s, missile defense systems increasingly faced offensive systems whose growing size and hence payload capacity enabled carrying initially lightweight decoys and then, as warhead sizes drastically shrunk, multiple warheads. As attacking warheads increased, the burden on missile defense increased commensurately. As decoys confused sensors, the task of shooting down warheads became far more daunting.

These large offensive systems with multiple warheads were first deployed in 1964 on the U.S. Navy’s Polaris A-3 submarine-launched ballistic missile. Soon after, it became possible to design a missile that dispensed a series of independently targeted warheads. The navy was first to deploy these MIRV systems in 1971. MIRV developments in America and Russia went along roughly in parallel—most American officials were convinced that American restraint on MIRV would not be reciprocated by the Soviets.

Missile defense capabilities deployed to date cannot intercept ICBMs, which travel at four miles per second (nearly equal to the five-mile per second orbital velocity of satellites), twice the speed of intermediate-range missiles and about four times the speed of a short-range missile (like the Scud). Such superfast warheads cannot be tracked and intercepted by existing defensive systems.

But the Russians feared that America would be able to surmount missile defense limitations. When U.S. Secretary of Defense McNamara lectured Soviet Premier Alexei Kosygin on the dangers of missile defense at the 1967 Glassboro (New Jersey) Summit, Kosygin countered him with Occam’s Razor (a rule of preference for the simplest explanation): “When I have trouble sleeping, it’s because of your offensive missiles, not your defensive missiles.” McNamara was focused on defensive missiles because he accepted MAD. There is no credible evidence that the Soviets accepted MAD, except, perhaps, as Mainly America’s Destruction. The Soviet Union’s extensive civil defense program indicated a desire to save its population, which is utterly inconsistent with MAD. Even if the shelters would have proven useless, the government’s intention in building them was to protect the very people MAD was supposed to hold at risk. (Nor did America fully accept MAD, as noted in the text.)

As arms talks progressed in the Nixon administration, domestic opposition to ABM—an acronym of Cold War origin that denotes anti-ballistic missiles, still used by many—began to build. Such systems were far more widely known than MIRV, and thus became the primary focus of arms-control attention.

The ABM/MIRV case was a classic example of strategic systems whose development was so closely linked that the “action-reaction” cliché often used by arms controllers—that each side’s programs were primarily driven by similar moves by the other side—held an initial measure of validity. That theory, however, suggested that American restraint would have been reciprocated. By the mid-1970s it became clear that far from emulating American decisions, the Soviets were continuing their massive military buildup despite considerable American restraint, including freezing offensive forces at 1967 numbers. This should not have come as a surprise, in that American and British restraint during the 1920s and 1930s pursuant to the interwar naval treaties did not dissuade Nazi Germany and militarist Japan from rushing pell-mell to build far beyond the limits they had nominally agreed to accept. Nor have the post-1967 proliferators—India, Pakistan, South Africa, North Korea, and nuclear-club wannabe Iran—followed U.S. nuclear restraint.

The 1972 ABM Treaty did not halt development of MIRV. The price of gaining broad support for the first arms-control treaty between the U.S. and USSR included deployment of several modern strategic systems, including those incorporating MIRV. The rationale driving deployment was that as Soviet missiles became more accurate a smaller number of missiles would survive a surprise attack, and these would need enough warheads to be able to fully retaliate and thus preserve deterrence.

While MIRV development continued, ABM development was brought to a virtual standstill. The ABM Treaty permitted each side to deploy 100 missiles to defend a chosen land-based missile-silo basing site and another 100 to defend the national capital city. America deployed its Safeguard ABM in 1974 at the missile base in Grand Forks, North Dakota. Safeguard consisted of a two-layer defense: the Spartan missile designed to intercept ballistic missiles above the atmosphere, and the Sprint missile designed to intercept at low altitude missiles that Spartan missed. The system was never deployed around Washington, D.C., due to very understandable popular resistance to deploying five-megaton warheads close to heavily populated areas. The system at Grand Forks was dismantled in 1976. A 1974 protocol (add-on) to the ABM Treaty limited Russia to 100 ballistic missile interceptors.
⁵⁷

The systems ultimately deployed by the United States were “dumbed down”—deliberately made less capable of intercepting incoming warheads—in order to conform to arms-control agreements as interpreted by arms controllers. Specifically, radar capabilities and access to satellite tracking data were restricted. Thus when a Scud missile (a short-range, primitive Soviet ballistic missile system sold to several Mideast countries) destroyed a barracks and killed American servicemen in Dhahran, Saudi Arabia, near the end of the Gulf War, the dumbed-down Patriot-3 system failed. It is reasonable to believe, though not definitively provable as it is the road not taken, that unfettered development of missile defense technology would have produced a system able to destroy the Scuds launched during the Gulf War (most landed in Israel).
Thus arms agreements already have plausibly prevented deployment of lifesaving defensive systems.

Much of the opposition to missile defense was based upon the sheer infeasibility of defeating a massive missile salvo of the kind the Soviet Union could have launched, using the kinds of systems deployable within arms-control constraints. The uncertainties were similar to those faced by prospective attackers using a large fleet of missiles. Put simply, systems were tested in small numbers, with many tests solo. There is no way for technologists to gauge from such tests how the same systems will perform when used on a large scale. Test trajectories and war trajectories differ, with aim “bias” introduced by asymmetries in the Earth’s magnetic fields. System performance, in a nutshell, may not scale in uniform, linear fashion. Thus offensive system behavior in situations other than those specifically tested cannot confidently be predicted by attacker or defender.

Systems currently deployed intercept missiles either in their final (terminal) phase of flight or in midcourse. Terminal-phase intercept involves separating heavier warheads from lighter decoys in the closing seconds, made possible when warheads encounter friction in the atmosphere, which then separates the two based upon weight and density differentials. But with time so short, taking out a large salvo—even if defense radars were not destroyed, a highly shaky assumption—is a complex task. Midcourse intercept targets ballistic missiles coasting in space on unalterable trajectories (like artillery shells), but where the zero gravity of space makes separating warheads and decoys extremely difficult.