Since the Federal Reserve raised its short-term interest-rate target in December 2015, much has changed for the worse: Job gains and overall output have disappointed, projections of future growth have declined, depressed inflation expectations have cast doubt on the Fed's credibility, and threats such as Britain's possible exit from the European Union have added to global uncertainty.

This all suggests that the central bank should be contemplating much more radical actions than it did at Wednesday's policy-making meeting, where officials decided to leave interest rates unchanged.

Despite all the disappointing data, there's no sign that Fed officials are even considering an interest rate decrease. For example, in the economic projections released along with today's policy statement, not a single member of the Federal Open Market Committee indicated that it would be appropriate for the central bank's target rate to be lower at the end of the year than it is today.

The Fed is thus signaling that it is highly unwilling ever to cut interest rates, a message that is problematic for a couple reasons. First, by communicating that it would take a big shock to get it to think about lowering rates, the central bank is essentially telling businesses and investors that they will get no protection from smaller downward shocks. This leaves them feeling vulnerable, creating a drag on economic activity.

Second, the lack of flexibility makes raising rates more difficult. Because the Fed doesn't want to ease, it must view each interest-rate increase as semi-permanent, meaning that it must be exceedingly cautious about making any move. As a result, its progress toward normalization — that is, toward getting interest rates closer to where they have been historically — is slower than it otherwise could be.

The Fed likes to emphasize that it focuses on the incoming data in deciding where to set interest rates. This is a great core message that cannot be repeated too often. The Fed is charged with achieving economic outcomes — stable prices and full employment. Hence, economic conditions should determine what it does to generate those outcomes — and the required policy can change over time, perhaps by a lot. 

The Fed’s view of data dependence, however, seems oddly asymmetric. Bad data can keep interest rates on hold, but they must be overwhelmingly bad to turn around the three-year trajectory toward tighter policy. This upward bias is visible even in the t-shirts that John Williams, president of the San Francisco Fed, famously had printed up last year:

Maybe what the Fed needs is a new design, with arrows that can go down as well as up.