I cannot, of course, speculate. I wasn't there, and even if I had been there, I wouldn't be allowed to speculate about why the decision was made, so I am in no position.
I can make two comments, though, on your question.
First, it certainly wasn't to save money, because my understanding is that this exercise ends up costing $30 million more and results in a great deal less usable data. It certainly was not to save money.
The other little point I want to make is that, as you mentioned, we might find out years from now whether the data are biased. I'm afraid we might not find out even then, because bias is inherently not knowable. Some bias we will know, because there will be contradictory information from independent sources, but much of the bias will be hidden. That's just in the nature of bias.