Riccardo M Masolo and Francesca Monti.
Newspapers and other media outlets regularly speculate about what the Bank of England might do in response to current economic conditions. Curiously, however, most of the models we use to carry out our economic and policy analysis completely disregard this type of uncertainty. Many of them consider how people would behave when uncertain about the state of the economy, yet everyone is assumed to know for sure the variables that the central bank will respond to, how aggressively and why. To try and fill this gap between the models we typically use and the reality we actually face, in our paper we explore the effects of Knightian uncertainty about the behaviour of the policymaker in an otherwise standard macro model.
We think Knightian uncertainty best captures the situation we want to model, because it represents an environment in which agents are not just uncertain about the macroeconomic outcomes, but are also uncertain about the odds of such outcomes. The simplest example is one in which a draw is made from an urn containing red and black balls. If we were told we’d win one pound when a red ball is drawn and we knew half of spheres were actually red, we would clearly be uncertain about our future income, but we would at least know that we stand a 50 percent chance of winning. Knightian uncertainty, on the other hand, refers to the case in which we do not know the share of red balls in the urn, and therefore do not know the odds.
Moving to our macroeconomic application, models usually posit that the central bank will follow a simple Taylor-type rule, which assumes that policymakers respond to deviations of inflation from its target and output from its potential level. Crucially, standard macro models also assume that all agents in the models perfectly know that this is indeed the rule that will determine policy rates at present and in the future.
If that were really the case, central bank communication would be redundant; in particular there would be no need to explain how policymakers intend to respond to certain economic conditions. Forward guidance, as implemented in summer 2013, would have no reason to exist in that everyone, by definition, would know how the central bank intends to respond to unemployment, inflation expectations and the like.
In such an environment, agents would be uncertain about future policy rates because of the inherent uncertainty surrounding inflation (black or red ball draw) but could accurately assess the probability of various outcomes given the model of the economy (they know the shares of black and red balls in the urn). Naturally, that assessment would be complicated if agents did not know for certain how the Central Bank would respond to inflation.
In our work we study the situation in which economic agents have in mind a set of possible models for the behaviour of the central bank. Because agents cannot form probability assessments about these alternatives, they will adopt a decision rule that is robust to such model uncertainty. Namely, they will base decisions on the assumption that the worst case will materialise (which is sometimes referred to as max-min, because the decision maker chooses actions that maximise their welfare under the assumption that the worst case scenario has materialised) because they do not like this type of uncertainty. In so doing, we apply an insight explored by Ilut and Schneider (2014), among others, to the monetary policy rule.
Some effects of Knightian uncertainty
An immediate effect of Knightian uncertainty is that inflation will persistently deviate from its designated target, while the output gap will not close, even in the absence of shocks buffeting the economy. This is because, while the central bank is implementing the optimal policy, the private sector makes its decisions about consumption and saving as if, instead, they expected a different, inefficient, policy to be followed, according to the max-min logic.
Interestingly, it turns out that, in this world, the responsiveness of the policy rate to inflation serves as a hedge against the worst consequences of uncertainty. In our simple setup, the inefficiency determined by Knightian uncertainty enters the economy via a suboptimal level of inflation. A stronger response to inflation deviations is a way of reducing the scope for this pernicious effect. But there are clear limits to the extent to which the policy rate can respond to inflation deviating from target, most notably the effective lower bound on interest rates. So reducing the underlying Knightian uncertainty by communication is a way of achieving better macroeconomic outcomes.
Overcoming the inefficiency
The negative effects of Knightian uncertainty arise because the private sector does not behave according to the (correct) assumption that the central bank will follow the optimal monetary-policy response function – which, in turn, makes the response function suboptimal. While reality is obviously more complicated than our simple model economy, we see communication as a way of increasing the private sector’s confidence in policymakers. Importantly, we are not concerned here with communication about the state of the economy; rather, we are considering communication of what could be considered contingency plans, i.e. how a central bank will respond if a set of conditions should materialize. The forward guidance issued by the Fed in December 2012 and Bank of England in August 2013 – which related the policy decisions to thresholds for unemployment, subject to not breaching the price stability objective – could in part be viewed as examples of this communication strategy.
Admittedly, this is easier said than done, because in the model the optimal response is obvious and fully known by the central bankers. In reality, policymakers face model uncertainty, as well as data uncertainty (e.g. due to data revisions) and generally uncertainty about the state of the economy; so they are themselves uncertain about the best response and might not be in a position to be more explicit about their decision process. On the whole, however, we think our analysis makes a compelling case for clear policy communications to ensure that the largest number of economic agents is confident that they know what policymakers intend to do should certain events materialise.
Riccardo M Masolo works in the Bank’s Conjunctural Assessments and Projections Division and Francesca Monti works in the Bank’s Monetary Assessment and Strategy Division.
If you want to get in touch, please email us at email@example.com
Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.