So...an economy can be thought to be something like an engine. Runs fine most of the time. Runs badly sometimes.
There are several stories as to why the economy runs badly, but most of us (non-technical economists) are really as clueless as we are (non-automotive software engineers) about how the software in our car regulates fuel consumption for maximum efficiency.
My thought for the best metaphor is that the money supply is the oil that keeps the engine running. No oil, no worka. Too little oil, grinding sounds, and burnouts. Enough oil...engine runs smooth. Metaphor fail when there's too much money.
The alternative (interesting) perspective is that the economy is NOT like an engine. It's like a Walmart (that offers Haircuts, financial services, tune-ups, food, and clothes). This line says that our problem is that we had a lot of folks doing haircuts, and what we need now is more auto mechanics, and it takes quite a while for folks to (a) learn that there's a glut of hairstylists, and (b) learn the skills to be a mechanic.
AFAICT...the evidence is pretty strongly on the side of the oil-metaphor. Textbook macro argues that "quantitative easing" or "NGDP Targeting" or "putting more money in the system any way we can" would solve this problem. Indeed, Sumner argues that during every recession we've ever seen, the macro folks say, "this one is unique," while after every recession, the macro folks say..."Ya know, it seems to have been a money supply problem again". If that's true (and I'm taking Sumner's word for it), then NOTt trying NGDP targetting is idiotic.
Maybe we're wrong...but when you have a consensus about all prior recessions, and a new one comes along...ignoring the consensus seems dumb.
I am not a macro-economist, and I'm basically an anarchist who thinks government causes most problems (including this recession). So...add salt liberally.