My last three years were spent turning my 20% project into a product, and my job now is spent turning another 20% project into a product. There was never any management pressure from any of my managers to not work on 20% projects; my performance reviews were consistent with a productive Googler.
Calling 20% time 120% time is fair. Realistically it's hard to do your day job productively and also build a new project from scratch. You have to be willing to put in hours outside of your normal job to be successful.
What 20% time really means is that you- as a Google eng- have access to, and can use, Google's compute infrastructure to experiment and build new systems. The infrastructure, and the associated software tools, can be leveraged in 20% time to make an eng far more productive than they normally would be. Certainly I, and many other Googlers, are simply super-motivated and willing to use our free time to work on projects that use our infrstructure because we're intrinsically interested in using these things to make new products.