These would be the only people with movement patterns like that. They wouldn't blend at all: they'd stick out like a sore thumb on any type of large scale stati

These would be the only people with movement patterns like that. They wouldn't blend at all: they'd stick out like a sore thumb on any type of large scale statistical analysis as anomalies. "being really secret" doesn't work because the attempt to do so means you're doing a whole lot of stuff that no ordinary person does. That's a statistical anomaly - that's easy to find. And again: all this information is public. It would be widely and well known that this is what is done, so what to look for would be well advertised. The locations of these buildings would then be well known. And since you've now added multiple intervening steps, nailing down who's going where is also easier - i.e. if someone drives into a known parking structure at time X, then you can draw a circle around that which is the mean transport time to get to any interesting federal buildings in the area before the work day starts. Map this across a few months and you'll pin down exactly which one they work at (I mean, you wouldn't bother if you were a foreign government though - a couple days worth of casual surveillance work would also grab every number plate and face you can see). This is quite aside from the why of doing this: it would be a huge pain in the ass for the employees, who are not paid nearly enough to bother with it and become a hiring problem, for...what gain?