Over 20 years after the Cold War, America’s nuclear arsenal remains bloated. True, it now deploys only about 1,600 strategic nuclear weapons — down from 12,000 in 1990 — and the Obama administration has proposed to cut the number to as few as 1,000 if Russia agrees.

Still, while America’s nuclear arsenal is useful to deter attacks on the United States and its allies, it is much larger than necessary.

So what is holding the United States back? Recent improvements in missile technology and international politics allow deeper reductions. The real problem is Washington’s outdated nuclear strategy, and the internal Pentagon politics that drives it. That strategy is built on maintaining a triad of long-range delivery systems — bomber aircraft, intercontinental ballistic missiles and submarine-launched ballistic missiles — developed early in the Cold War.

The triad grew from compromises between the United States Air Force and Navy, not from a coherent strategy to protect American interests.

In the 1950s, the Air Force had a near monopoly on nuclear delivery, until thermonuclear warheads made ballistic missiles viable. The Navy, hustling to regain relevance and budget share, managed to deploy them on Polaris submarines. The Air Force, meanwhile, developed various land-based missiles.

During the Kennedy and Johnson administrations, interservice nuclear competition abated. The Navy and Air Force learned to cooperate to grow the whole budgetary pie. Pentagon and congressional support for each leg of the triad meant that no administration had the stomach to eliminate one.

During the Cold War it was more reasonable, within the logic of nuclear competition with the Soviet Union, to justify such large arsenals. Different deployment capabilities made it harder for the Soviets to knock out America’s ability to retaliate.

But while the Cold War is long over, the triad remains.

The main justification continues to be “survivability” — maintaining mutually assured destruction with enemies. That keeps enemies from using the threat of a first strike as blackmail to prevent the United States from defending allies.

But as is often the case in politics, the public justification differs from reality. In fact, America’s nuclear weapons themselves are made to sidestep the MAD trap: warhead design and ever-improving accuracy optimize the ability to destroy enemy nuclear forces before they launch, not retaliate afterward. Contrary to much official rhetoric, Washington’s nuclear war plans have always focused on a pre-emptive strike against enemy weapons.

One reason for that posture is the doubt that the United States would invite its own destruction to protect foreigners. As a result, deterrence theorists tell us, protecting allies from nuclear rivals requires the ability to escape enemy retaliation by destroying all of their nuclear weapons first.

In the early days of the triad, that mission gave each delivery system a discrete role. ICBM’s had the accuracy and reliability to target most missile silos. Bombers would deliver warheads powerful and accurate enough to destroy especially difficult targets. And submarines, because their missiles were relatively inaccurate but essentially invulnerable, would be held in reserve to threaten Soviet cities in a second strike, thus encouraging Soviet leaders to sue for peace.

Pre-emption is pricey, though. It takes several nuclear weapons to ensure that you can destroy one enemy weapon. Worse, pre-emption encourages an arms race. Fear of a first strike encourages enemies to build more weapons for defense, requiring more weapons to pre-empt them, and so on. That helps explain why U.S. military budgets have long been insufficient to achieve clean first strikes against all rivals.

In short, America’s nuclear policy is a contradictory muddle: an underfunded first-strike force justified by second-strike rhetoric. These days, a submarine-based monad makes more sense. For one thing, survivability is easier to achieve. America’s current adversaries are unable to track its submarines, let alone target them. Moreover, leaps in accuracy have reduced the size of the force needed. Submarine-launched missiles are actually more accurate than the land-based kind and, with conventional weapons, can now threaten any enemy arsenal.

And potential targets for American nuclear weapons are growing scarcer. New nuclear powers like North Korea struggle to deploy even a handful of delivery vehicles. Targeting China’s few long-range missiles demands intelligence to find them, not sheer numbers of warheads to hit them. Russia can no longer afford nuclear parity, especially given its plans to modernize its nonnuclear forces.

Whatever aggressive plans these rivals have are deterred by America’s allies and conventional military superiority, making nuclear weapons overkill in most cases.

Most important, deterrence is easier to achieve than nuclear weapons enthusiasts typically admit. Even the Soviet Union, we now know, was eager to avoid a major conventional war, let alone a nuclear escalation. Today’s rivals are even more easily contained by American and allied conventional strength.

The idea of nuclear weapons cannot be abolished. And because nuclear weapons contribute to deterrence, they remain a wise investment for the United States, but one that need not cost so much.

Moving from a triad to a submarine-only monad wouldn’t be easy, but the political situation is changing. Budget-conscious service chiefs may see nuclear weapons as an attractive target, especially given their irrelevance in recent wars.

Pentagon competition helped create the triad; restored competition could help kill it.