What we don't measure, we can't manage. What we don't measure accurately, we manage badly.  (And that may be to the huge advantage of special interests.)

Because we lack some of the data we should have, we end up substituting data which points in the general direction of what we need to know.

IRS data, for example, is often a rather poor proxy for the data we should be collecting.  While the data itself may be quite accurate, it has severe limitations for many kinds of analysis.  And to the extent that certain kinds of income simply don't get reported, the remaining data makes a poor tool for analysis.

Mason Gaffney:  Sounding the Revenue Potential of Land: Fifteen Lost Elements