Liftoff: A Comparison of Two Normalization Cycles

Many Federal Open Market Committee (FOMC) participants have said that the policy rate (i.e., the target for the federal funds rate) should come off the zero lower bound in 2015, with the exact timing dependent on how key macroeconomic indicators evolve. Given that this initial increase would mark the start of a normalization cycle, now is a good time to review the previous two major normalization cycles to see what we can learn from them.1

The first normalization cycle for comparison began in 1994. The policy rate since September 1992 had been at 3 percent, which at the time was considered exceptionally low relative to the federal funds rate during the 1970s and 1980s. U.S. macroeconomic data indicated a strong economy toward the end of 1993. For instance, real gross domestic product (GDP) growth accelerated in the fourth quarter, job growth was slightly stronger on average and inflation was threatening to move higher. In what was largely a surprise to financial markets, the FOMC began a normalization cycle in February 1994 and continued raising rates throughout that year.2 In contrast to the second normalization cycle I will highlight, the FOMC raised the policy rate by 25 basis points sometimes, by 50 basis points other times and by 75 basis points on one occasion. Also, the policy rate was left unchanged at a few meetings. The pace was adjusted in reaction to the incoming macroeconomic data and in this sense was data-dependent, or state-contingent. The normalization cycle ended in February 1995, with a policy rate of 6 percent.

Financial markets generally viewed this adjustment to higher interest rates as disorderly. In fact, the bond market had one of its worst years in 1994. The 10-year Treasury yield, for instance, rose roughly 2 percentage points that year. Despite being disorderly, the 1994 normalization turned out to be a success for the U.S. economy. The policy rate was returned to a more normal level, and the economy boomed in the second half of the 1990s—one of the best periods for economic growth in the postwar era.

The second normalization cycle for comparison took place in 2004-06. The policy rate had been 1 percent since June 2003. Leading up to the June 2004 FOMC meeting, real GDP growth remained solid, gains in nonfarm payroll employment had increased in recent months and inflation had risen. The FOMC raised the policy rate to 1.25 percent in June 2004 and continued with a mechanical pace of increase of 25 basis points at each of the next 16 meetings. Thus, there was almost no state contingency with this normalization cycle. In terms of communication, the FOMC was more transparent regarding its expectations for future increases in the policy rate than it had been previously. This cycle ended in June 2006, bringing the policy rate to 5.25 percent.

Financial markets viewed this form of normalization as much more orderly than the 1994 case and, therefore, a success. However, this normalization cycle may have been counterproductive. The housing bubble inflated even more during this two-year period as financial markets found ways to create investments in housing based on cheap financing—investments that ultimately proved disastrous. Although policymakers were cognizant that house prices were rising and that mortgage finance was increasing, the general view was that the air could be let out of the bubble slowly and without dramatic macroeconomic consequences. In actuality, the opposite occurred. The housing bubble burst, starting in 2006, right about the time the normalization cycle ended. House prices fell about 30 percent, and the U.S. experienced a severe recession.

What are the lessons from these two episodes? Although the 1994 normalization cycle was considered disorderly (i.e., uneven amounts that were somewhat unpredictable), it seemed to set up the U.S. economy for success in the second half of the 1990s. On the other hand, the 2004-06 normalization cycle was considered orderly (i.e., perfectly even amounts that were generally anticipated) but, in retrospect, turned out to be suboptimal because it allowed for the continuation of speculation in housing markets and in mortgage finance. For the upcoming normalization cycle, some combination of the two—the data dependency from the 1994 case and the transparency from the 2004-06 case—would probably provide the optimal method of returning the policy rate to normal.