Eggertsson and Woodford’s 2003 paper was written when Japan’s policy rate had sat near zero for years. It starts from a simple fact: nominal interest rates cannot fall much below zero.
A central bank conducts policy by setting a short-term nominal rate. In the US this is the federal funds rate, the overnight rate at which banks lend reserves to each other. Lower rates make borrowing cheaper and saving less attractive, so spending rises.
The floor comes from cash. Cash pays a nominal return of zero, so if a bond paid −2%, investors would hold cash instead. Hence $i \ge 0$. Storing and insuring large amounts of cash is costly, so in practice rates can go slightly negative (the ECB went to −0.5%), and the floor is now often called the effective lower bound.
What matters for spending is the real rate, the nominal rate less expected inflation:
$$r = i - E\pi.$$Usually the central bank can move $i$ far enough to put $r$ where the economy needs it. The bound bites when a large shock, such as households and firms suddenly wanting to save much more, calls for a deeply negative real rate. With expected inflation near 2% and $i$ stuck at zero, $r$ cannot fall below −2%. Spending stays low, prices fall, expected inflation falls with them, and the real rate rises, the opposite of what is needed.
Eggertsson and Woodford’s answer is that one lever remains: expectations. Spending today depends on the whole expected path of real rates, not only today’s. A credible promise to keep $i$ at zero after the shock passes, and to tolerate some inflation then, lowers expected real rates now. The difficulty is credibility. Once the economy recovers, the bank will be tempted to break the promise, and the public knows it.
At zero, a central bank’s most useful instrument would be its reputation.
sources
- Eggertsson, G. and M. Woodford (2003), “The Zero Bound on Interest Rates and Optimal Monetary Policy,” Brookings Papers on Economic Activity.