The market massacre following Fed chairman Ben Bernanke's June 19 press conference shows the problems of a central bank trying to have its cake and eat it too – trying to convince Mr Market that policy will be both stable and data-driven. The chairman can't have it both ways.
Back in October 2012, the Fed pledged to keep policy highly accommodative into mid-2015. This presumed that the Fed could accurately forecast low inflation and economic slack three years out, since the pledge only made sense if it could anchor market expectations that it would not need to tighten before then.
Then in December the Fed shifted from a date-based pledge to a data-based one, promising to keep rates near zero until the unemployment rate fell below 6.5%. With unemployment at an elevated 7.8%, however, the Fed still appeared to be well away from pivoting towards a less accommodating stance. Mr Market remained calm.
But the pledge strategy was always, in my view, a train-wreck waiting to happen.
The Fed is fundamentally a data-obsessed institution that overestimates its predictive abilities. Even in normal times, the Fed is simply not very good at forecasting output, unemployment, or inflation. When the Fed studied its own staff's forecasting performance over the period 1986 to 2006, it found that it was worse than that of market-watchers outside the Fed. The Board of Governors' and Reserve Banks' first three-year forecasts in October 2007 were, in particular, wildly off the mark: actual 2010 gross domestic product, unemployment, and inflation were all outside the entire range of the 17 separate forecasts.
Yet on May 22 Bernanke spooked Mr Market by revealing that the Fed was looking to glean policy-relevant information from incoming economic data, announcing that it would consider "a recalibration of the pace of its [asset] purchases… in light of incoming information". He did not provide any guidance on the incoming information that could precipitate a recalibration, but he did suggest that there would be a "considerable" period of time between the end of the asset purchase programme and the end of the period of highly accommodative monetary policy, indicating that the Fed would want to cease the asset purchase programme well before unemployment reached 6.5%. And it would likely want to decrease the size of the purchase programme well before that.
The Fed's attempt to fine-tune the pace of asset purchases was bound to give Mr Market a bad case of the shakes, as "incoming information" has been extremely volatile throughout this economic recovery. As the above figure shows, using the six-month average of employment gains to project the unemployment rate going forward suggests vastly different metrics of how close the Fed is to achieving its 6.5% unemployment-rate objective.
If the average pace of job gains in the six months leading up to and including the March unemployment report had been extrapolated forward, the Fed would have expected to reach its 6.5% unemployment target in August 2015.
Yet with just one additional month of employment data, an extrapolation of the six-month average gain in employment through April shows the unemployment rate falling below the Fed's 6.5% threshold in August 2014, a full year earlier.
And the pattern over the past two months is not an anomaly. A projection based on the six months of employment data available in November would have the Fed reaching its employment objective as soon as May 2014; but a projection based on data in January of this year, just two months after the November unemployment report, doesn't have the Fed reaching its objective until September 2015.
On June 19 Bernanke went into calibration overdrive. The Fed, he said, "may adjust the flow rate of purchases month to month to appropriately calibrate the amount of accommodation we're providing given the outlook for the labour market".
The Fed would "keep the level of accommodation consistent with the outlook." It would use "models and other indicators of the state of the labour market to try to make a good estimate of how much we need to change the rate of flow".
But any data-based indicators of the state of employment, as the other projections in the figure show, are going to be highly volatile. It's no wonder Mr Market went into convulsions. Asset purchases are not a precision tool, so the idea of continuously calibrating them to volatile economic data is a particularly bad one, as the market reaction has borne out. When the Fed announced that it would start buying an additional $40bn a month in mortgage-backed securities last September, in order to push down mortgage rates, the 30-year fixed rate was 3.57%. Today it is near 4.5%.
Bernanke himself back in 2004 highlighted the role of asset purchases as a powerful signalling tool. "Quantitative easing," he wrote, "may complement the expectations management approach by providing a visible signal to the public about the central bank's intended future policies… this policy provides a way of underscoring the central bank's commitment to keeping the policy rate at zero for an extended period."
But this works in both directions: just as the initial announcement of future asset purchases signalled the Fed's commitment to the zero interest-rate policy, so now the announcement of "calibration" underscores the absence of commitment.
Calibration is a strategy badly in need of recalibration.
Benn Steil is director of international economics at the Council on Foreign Relations and author of The Battle of Bretton Woods, which led off this weekend's FT "books of the year so far" list. Dinah Walker is an analyst at the Council.
This article appears in full on CFR.org by permission of its original publisher. It was originally available here (Subscription required).