You are driving a car with rear-wheel steering and no reverse gear. You are driving alongside a wall. If you drive too close to the wall you are trapped, because you would need to steer your rear wheels into the wall in order to get your front wheels to move away from the wall. If your car had front-wheel steering you couldn't get trapped like that.
Using an interest rate instrument to control monetary policy is like having rear-wheel steering to control your car. If you get too close to the Zero Lower Bound wall, you are trapped. Raising the inflation target to reduce the chances of hitting the ZLB is like choosing a route further away from the wall to reduce your chances of getting trapped. It would be much more sensible to choose a different car, with front-wheel steering.
Irving Fisher's Compensated Dollar Plan is like a car with front-wheel steering. Under the Gold Standard, central banks held the price of gold fixed. Irving Fisher said that central banks should adjust the price of gold to try to keep the general price level fixed. If the price level started to fall, the central bank should raise the price of gold to create an offsetting force raising the price level. If the price level started to rise, the central bank should lower the price of gold to create an offsetting force lowering the price level. The front wheels (the price of gold) move in the same direction you want the whole car (the general price level) to move. Irving Fisher's Plan was never adopted, but Roosevelt raised the price of gold to end the deflation in the 1930's. So did other central banks.
Choosing Irving Fisher's method for steering the car (the monetary policy instrument) does not mean choosing Irving Fisher's route (the constant price level monetary policy target). But choosing a car with front-wheel steering does let us choose to drive closer to the wall (have a lower inflation target) if we wanted to do that, because it eliminates the risk of getting trapped. If we did find ourselves alongside the wall, we would simply need to choose above-target inflation temporarily.
The price of gold seems like a very anachronistic monetary policy instrument. But it's not the only form of front-wheel steering. One alternative method of front-wheel steering would be for central banks to use a stock price index as the monetary policy instrument. Instead of varying the price at which it will buy and sell gold, the central bank could vary the price at which it will buy and sell an index of stocks. Instead of keeping bars of gold in the basement and on the asset side of its balance sheet, central banks would keep market baskets of stocks. Stocks are better than bonds for central banks. Because stocks are front-wheel steering, and bonds are rear-wheel steering. If a central bank wants to raise the general price level, it would raise the price of stocks, but lower the price of bonds.
If all cars had rear-wheel steering, and then you saw a lot of cars trapped against walls, I think that car designers would be working hard designing new models with front-wheel steering. Any automotive engineers who protested "But cars have always had rear-wheel steering; that's how cars work!" would be ignored.
I have thought about front-wheel steering a bit, in some older posts I can't find now. Roger Farmer thinks about it. It should be at or near the top of the money/macro research agenda.
Sure it's just a metaphor. And models aren't? (Irving Fisher and Bill Phillips built mechanical/hydraulic models.)
[I stole the metaphor from Brad DeLong, but am using it in a way he may or may not agree with.]