Maybe people are paid to do little or are counted as employed when they aren't really, which improves employment, but they are still paid, and they activity or lack thereof is then reflected in productivity numbers. Employment is counted relative to total population, which is a pretty well known number.

Unemployment is counted relative to "active population", an altogether fuzzier concept (and the easiest way to cheat on unemployment statistics is to move the active unemployed to the inactive - it's just a labelling trick and it costs nothing, as opposed to payign them to pretend to be working)

In the long run, we're all dead. John Maynard Keynes