"I'm not going to listen to you. I know about you. You are a powerful advocate for your province and you have your reasons for wanting money in your province. Maybe the man in Samawah is less good as an advocate. So I am just going to have to rely on my numbers. Some numbers are better than no numbers."
I told him that was nonsense.
"How do you propose I allocate money then?"
I suggested he gather us all at a table, ask tough questions, let us justify our requests, listen to us argue, get to know our personalities, correct for those who seemed more forceful, and come to a decision. I said, "You can't get around problems with numbers. It doesn't matter how many people died; it matters how they died and why and who killed them, and for that you need political officers. There is no exact relationship between your indexes and policy. You may want to put more money into a richer place if it is about to collapse into civil war; you may-"
"What you suggest is very messy, imprecise and imperfect," he interrupted. "There will be biases. That is why we use metrics."