The $12 billion figure was for developing the Airbus A-380. Designing and certifying new plane is tough, but it's not exactly rocket science. By way of comparison, I heard that GM spent over $3 billion developing the Saturn car.

To Think you can develop major new space systems like those needed for sending humans to Mars for $20 billion is to ignore the history of aerospace development. Creating a Mars lander alone will almost certainly be a bigger technical challenge than developing the Lunar Module was in the 1960s. That program took many years and was billions over budget. The Cassini space probe cost somewhere around $3 billion, and while it's a remarkable piece of machinery, it pales in complexity and cost to what it'd take for a manned mission to Mars. Things aren't cheap. Aerospace things most definitely aren't cheap. Cheap implies cutting corners. Cutting corners leads to failure and death.

Now, Zubrin may try to arrive at his $20 billion price tag by conviently shifting major parts of the development cost to other programs. For example, he may not be counting the cost of creating a heavy lift derivative of the Shuttle. However, if the only reason you're developing that launcher is for his Mars program, that would be dishonest accounting. It'd be like not including the cost of developing and building the Saturn V in the Apollo program. I've seen a LOT of that in the government sector.

There's also "low balling", where you give an absurdly low figure to get a program started and hope to make up for the overruns with more appropriations. After all, you've already spent so much money, it'd be a waste to not see it through. For an example, see the "Big Dig" in Boston. That's a figging hole in the ground that's cost over $12 billion. $20 billion to Mars? Give me a break. If you ever worked in this sector, you'd see just how absurd that number is.