Read on Twitter
One of the things the Federal Reserve realized over the last two years is that data can get in the way of information. We might be at that point with parties and polls. /1
As part of an 18-month policy review, the Fed put on "Fed Listens" events at all of the regional Fed banks: bring in community members, have a conversation. What could have been a pro-forma marketing exercise ended up affecting policy /2 https://www.ft.com/content/d2652066-32af-11ea-9703-eea0cae3f0de
The Fed learned that there was way more "slack" in the economy than it had thought -- way more potential to pull up wages and bring people into the workforce. They also discovered that most Americans really, really, really don't care about inflation /3
Now: the Federal Reserve has a deep bench of A-Team research staff. There is no organization better able to pull in and analyze data. Even so: they learned things that weren't in the data by just, simply, talking to people /4
The blind spot for macroeconomists (sorry guys) is qualitative data: talking to people in a structured way, posing open questions, listening to what people care about. I wonder whether polls and consultants don't do that to political parties, too /5
What if you built a political party around the Fed Listens model: be in every county. Have regular conversations. Learn what people care about, and what motivates them. Build the whole party structure itself around listening and drawing conclusions from what you hear /6
It's not that polls are unnecessary. It's that they're insufficient. They're clearly failing to capture something. I think the false precision of a poll gets in the way of learning things -- the same way that the false precision of an economic data median can hide things /end
