If you allow for the 30% contingency that was applied on top, Zubrin's figures look very close to ESA (Z= $71.5Bn, ESA = $74). NASA comes out high bid, due to several cost loading elements. Many of these involve the setting-up of large administrative organisations at system level, project level, and programme level. ESA just get on with the job.

OTOH, Beagle 2 shows what can happen if you let people get on with their job without adequate oversight....

Interesting that ESA reckons it can train from scratch a mission control team for the same cost that NASA can retrain their experienced shuttle team . FOr subsequent ops, ESA budget for mission control is far lower. Is this realistic? Well, only if there's no problem! :-)