Skip to content

Data and the work of economic knowing

"Data-dependence" is, these days, a particular contingency of monetary policy-making. For example, FOMC Chair Jerome Powell said in his remarks following the December 2018 FOMC meeting, at which the target for short-term interest rates was increased by 25 basis points, that

[w]e always emphasize that our policy decisions are not on a preset course and will change if the incoming data materially change the outlook. And, given recent developments, the statement notes that we "will continue to monitor global economic and financial developments and assess their implications for the economic outlook." [@powell18, 2]

Powell is at pains to differentiate monetary policy decisions, discussed and ratified in ten FOMC meetings per year, are not "on a preset course"—but then what kind of course are they on? Each meeting is an occasion at which data are incoming, at which new knowledge is entered into the decision-making process, an outlook is formed, and the course of policy is changed. The outlook entails a likely future path for policy, but the statement emphasizes that the path is contingent. Policy will change with the outlook, and the outlook will change with developments [@neilson15].

What developments? Powell:

This illustrates the nature of data dependence that we always emphasize. In 2018, the economy was somewhat more robust than expected, and this led to a slightly faster pace of policy normalization than had been projected. When the economy has, instead, turned out weaker than expected, the Committee has slowed or paused the pace of rate increases as we did in 2016. And when the economy has performed about as expected, the Committee has generally moved in line with the median projection as we did in 2017. [@powell18, 3]

One cannot help but notice the constant interplay between expectation and contingency. The inherently forward-looking nature of monetary policy, whose effects seem to play out over an extended period of time, pushes against the insistence that decisions are made anew, that the present will not be held hostage to the past. This makes plain the constitutive status of data in monetary policy's ontological: data, when they "come in," are the arrival of the world in the knowledge-space of the FOMC, and thus they become the basis for reaction, deviation from the previous plan. Arrival is the force of data in the policy-making process, understood as the indirect impingement of material conditions upon the field of vision of policy makers.

The arrival of data is a discrete event, occurring at a particular moment in time, after which what is thought to be known about the world is changed, updated. In Powell's logic, the anticipated eventual arrival of data is what will make the future different from the present, and which will cause future FOMC decisions to differ from current FOMC decisions. More data will arrive at future moments in time, that is, and the fact that that data is not knowable in advance is what creates the contingency of "data-dependence." Data arrival, we could observe, is the expression of that more general quality of time emphasized by Keynes ("We simply do not know" [@keynes37].) and [@shackle88]. (I explore this in chapter 5 of my book: @neilson19.) Things will happen, the Fed says, and when they happen, they will generate data, and when the data get to us, we will react to them.

This would seem to imagine a clean line from events to data to policy-making, but the shutdown of the US government shows that the generation of data is also contingent: "Federal Reserve Bank of Atlanta President Raphael Bostic said Monday that the lack of some data will likely hinder the central bank in its decision making" [@dmitrieva19]. The shutdown brought a halt to the process of data production. It is an interruption of knowing, not of happening: things in the world continued to happen, but the FOMC, like other market participants, was not able to know about them. It is a contingency not imagined by Powell—he was attuned to the future arrival of unknown data on its known schedule, not to data's unanticipated failure to arrive at all.

What then is the function of expectations in the FOMC's debates? Data-dependence, the contingency of future policy, is the assertion that today's decisions will not bind future decisions, which will (after all) be made in light of new (and better) data that will by then have arrived. Nonetheless there is a view today of what that data will say and what future policy will therefore be. Yet these expectations are formed along some lines and not others: the failure of data to arrive, in particular, is a wrench in the works, apparently paralyzing decision-making.