3. QE3: Data-Driven, Not Date-Driven

Unconventional: A Policymaker's Reflections on Crisis to Recovery

Unconventional A Policymaker's Reflections on Crisis to Recovery

"Jim Bullard was an early advocate of state-contingent, or data-driven, quantitative easing in 2009, when date-driven QE policy was the approach of choice. Eventually, by 2012, the entire Committee came around to the view that QE should be data-driven, not date-driven. QE3 was designed on the concept of state-contingent, data-driven policy."

—Christopher Waller, Executive Vice President and Director of Research

3. QE3: Data-Driven, Not Date-Driven

James Bullard shared some reflections on his first 10 years as Bank president during recent conversations with staff. The following are excerpts from those discussions.

In 2012, about three years post-recession, the U.S. economy wasn’t growing as fast as people would have liked, and the pace of improvement in the labor market slowed. Furthermore, inflation wasn’t as high as people expected it to be. As mentioned earlier, headline inflation was above the Fed's 2 percent target, and core inflation was right around the target in early 2012. During the first half of that year, however, inflation began declining and went below target (although not as far below as in 2010).

Consequently, many policymakers felt like they wanted to do more to help stimulate the economy. The FOMC voted in September 2012 to begin a third quantitative easing program, known as QE3, and stated that the program would continue until the labor market outlook improved substantially.

I was not very supportive of QE3 at that time because, in my view, the data didn’t support such a major decision. For instance, while job growth wasn’t as robust as people would have liked, I thought that the slower job growth perhaps had become the norm, since we were several years past the financial crisis.1 Furthermore, the U.S. economy wasn’t in recession, nor did a recession look imminent. Those are among the reasons that I opposed beginning a new QE program at that particular time.

However, I did support QE3's open-ended aspect, which is a form of state-contingent or data-dependent policymaking and which stood in contrast with the fixed end dates associated with QE1 and QE2. As early as 2009, I had advocated for balance sheet policy to be state-contingent and adjusted depending on economic conditions, much like interest rate policy had been prior to the financial crisis. For instance, I argued that the FOMC should say a QE program would continue until the desired results for the economy were achieved, instead of saying it would end on a particular date that did not depend on goals being met.

This was not a very popular idea at first. But the FOMC eventually came around with QE3. In addition, the recent QE programs of the Bank of Japan and the European Central Bank (ECB) took the open-ended, state-contingent form.

At the December 2013 meeting, the FOMC decided to begin reducing the pace of asset purchases the following month. In October 2014, the FOMC determined that substantial improvement in the labor market outlook had occurred and ended the QE3 program.


  1. I also thought that some other labor-market trends, such as the decline in the labor force participation rate, were largely due to demographic changes rather than cyclical factors. For more, see my speech from Feb. 19, 2014, The Rise and Fall of Labor Force Participation in the U.S. [ back to text ]

Additional Resources


State-Contingent Policy

One of the consistent themes underlying James Bullard’s thinking has been the importance of state-contingent, or data-dependent, monetary policy, even when the FOMC uses unconventional policy such as QE and forward guidance.1

"State-contingent policy means that you should react to economic events and not do things according to the calendar ... I've tried to be an advocate for this at the FOMC."

—James Bullard, President and CEO

"State-contingent policy means that you should react to economic events and not do things according to the calendar," Bullard said. "I do think it’s a problem in monetary policymaking that there's somehow an overwhelming urge to say that you're going to do certain things at certain times, regardless of what's going on in the economy. But everything we know about economics and economic policy says that, 'No, the policy should be calibrated to what's actually happening in the economy,' which means reacting to what's actually going on.

"And so, I've tried to be an advocate for this at the FOMC. I think that we've had only mixed success, and sometimes I think we've slipped back more into calendar-style policy instead of state-contingent policy."


  1. For examples, see Bullard, James. Three Lessons for Monetary Policy from the Panic of 2008, a presentation delivered in Philadelphia, Dec. 4, 2009; and Bullard, James. President’s Message: What Does Data Dependence Mean? The Federal Reserve Bank of St. Louis The Regional Economist, January 2016. [ back to text ]

BACK: 2. Fear of a Deflationary Trap

NEXT: 4. Preferred Approach to Normalization