October 28, 2011

Defense Whiz to Pentagon: Your Predictions Are Destined to Fail

The Arab Spring. The Fall of the Soviet Union. India’s nukes. The U.S. government has a perfectly awful track record of predicting future events. And there’s a good reason why, says the chairman of an influential think tank: it’s friggin’ impossible.

Dr. Richard Danzig, the former Navy Secretary and current chair of the Center For a New American Security (CNAS), has published a comprehensive report on what he convincingly argues is a pretty huge problem for the Pentagon. After all, basing billion-dollar, life-or-death decisions on forecasts practically guaranteed to be wrong isn’t exactly an ideal management strategy.

“It’s good to consider the range of the possible; certainly, it’s better than not considering possible outcomes,” Danzig tells Danger Room. “But we’re never going to foresee everything, so we have to be planning for that failure as well.”

The Pentagon has been on a different track, however. Instead of planning for failure, they’ve been stocking up on polish for their crystal balls. Not only has the military spent $125 million in the last three years alone on computer software to predict political unrest, they’re also funding a ton of initiatives, from Internet mining to network science, to upgrade their forecasting. So far, the improvements have been modest, at best.

And because the Defense Department takes decades to develop weapons, it’s forced to make 20-year guesses about the kinds of enemies it will face. And those guesses are almost always wrong, wrong, wrong. Even making shorter-term predictions (Egypt’s uprising, say) depends on so many variables that it’s “wildly improbable” officials will get it right.

“I accept that the inclination to predict is deeply embedded in U.S. institutions and in human nature,” Danzig writes. “[But] long-term national security planning…will inevitably be conducted in conditions that planners describe as ‘deep’ or ‘high’ uncertainty, and in these conditions, foresight will repeatedly fail.”

The longer the forecast, the less likely it is to be. So Danzig wants the Pentagon to buy gear for the nearest possible future — and make sure that equipment can change with the times.

That means speeding up decisions about what to buy, how to design it and where to deploy it; troops can’t afford to wait years and years for armored vehicles, like they did in Iraq. Make ‘em and deploy ‘em,  and by keeping troops safer you’ll manage to — surprise! — improve future outcomes in the meantime.

Oh, and stop being so damn sophisticated. The military doesn’t just plod on development, it also hemorrhages money — high-tech planes, boats and weapons are designed to last thirty years, and it’s a policy that makes for expensive manufacturing and equipment that works in one war and winds up irrelevant in the next. Those Humvees designed for the Cold War? Troops could still drive them in Iraq, but they had to solder armament onto them to protect against prevalent improvised bomb attacks. A bare-bones vehicle, with add-ons like detachable armament and a lifespan of 10 years, would have been a cheaper, more strategic investment.

Of course, Danzig doesn’t want officials to give up on trying for a better future. He just wants them to have the right premise before they make decisions to get us there.

“It’s like systems built into cars to prevent crashes — I’m all for those,” he says. “But if you’re a good car-maker, you need to assume there will be crashes.”

Bottom line? Unless top brass get their hands on a crystal ball or two, even their best forecast can only illuminate so much. “Policymakers will always drive in the dark,” Danzig writes. “However, they must stop pretending that they can see the road.”