Maybe we are measuring happiness incorrectly. Instead of most people being happy, what if we made a small number of people REALLY happy? Maybe the 0.1% are so happy that it balances out the sadness of everyone else?
As sarcastic as this is, you still end up with the same just less intensity. If statistics and estimation are involved, this proposed happiness index will leave some behind and end up with everybody the same.
This is actually a real problem in population ethics.
Why is total utility the measure though? Why not median
If you used the median then only actions which affected the utility of someone in the top 50% and bottom 50% would have any moral impact. e.g. two miserable people torturing each other or helping each other would be neither good nor bad, since the median would not change.
Ah thanks for the context
You have to remember that the happiness will trickle down from that 0.1% to the rest of the population.