Forward guidance: talking about it, changing it, and disagreements with the yield curve
Recently the Fed was criticised for talking about tapering and then, when it saw how much the yield curve and asset prices reacted, backing off. The UK's Monetary Policy Committee has likewise been criticised as somehow messing up forward guidance because UK expectations of interest rates price in a rise in rates sooner than is consistent with the MPCs own forecast. Seen through the lens of modern macro, these central bank behaviours are perfectly proper, and don't indicate incompetence, contrary to what you might read in the newspapers.
How so? If the world were populated by agents who had what economists called 'rational expectations', who knew how the economy worked, and this knowledge was shared by central banks, we could not have seen either phenomenon. The Fed would have known what would happen when it talked about tapering, and there would have been no need to back off. In fact, there would have been no need for talking about it, as, unless there had been a change of plan, all there would be to do would be to execute it according to how the data dictated as it unfolded. And there could be no divergence between the UK's MPC forecast of its instrument, and forecasts of markets. At the most basic level, all that these recent events reveal is that the world is not like this. But that should not be news to anyone, except a few nutty rational expectations literalists.
A plausible conception of the world is that the central bank and markets have different views of how the recovery will pan out, because they have different views of how strong the effects of QE and interest rate stimuli are, and how they come about, different views of what caused the contraction in the first place. And markets may understandably be unsure of the central bank's precise intentions. (Neither central bank has tried to spell them out very precisely). The point of forward guidance is that you are trying to lower long rates because you have run out of room to cut short rates. And you do this by manipulating forecasts of future short rates. Central banks in this world of differing and imperfect world views ought to be forming views about how private forecasts of what it will do are formed, how they will react to events as they unfold, including how they react to information about its own intentions.
Seen this way, the Fed's actions are entirely reasonable. The Fed started out with a conception of private sector forecasts of the short rates (and associated wider asset price movements) that was less responsive to what it was contemplating saying about tapering than proved to be the case. When the taper talk caused all the fuss, the Fed revised its conception of how private forecasts were made. Tighter rates and lower asset prices naturally led the Fed to back off its initial taper talk somewhat because tapering was always going to be data-dependent, and included in that data would be data on asset prices. The Fed have not opted to describe what they did in these terms exactly, but a story like this is consistent with what they did. And if it is the right explanation, then what the Fed did is perfectly reasonable.
One can detect elements of this kind of thinking in the Bank of England's behaviour too. Faced with the yield curve that predicted rates were going to rise before it was forecasting them to, the MPC determined to re-emphasise the benefits of forward guidance, to try to assuage concerns that they were really committed to it. And to clarify elements of the guidance, such as the status of the unemployment rate 'staging post', which was not to be considered a 'trigger' for rate rises. The MPC was clearly attempting to diagnose the 'incorrect' yield curve, and provide information to private forecasters that would pull their forecast into line with the MPCs.
But there is something missing from the MPC's execution. Although there's evidence of the MPC forming a conception of how the yield curve is arrived at, there's no sense communicated of how the interest rate plan responds to that, if at all. To see why such a response might be necessary, imagine a world in which, once Carney announced forward guidance, market forecasts of central banks coincided perfectly with the MPCs. There would be nothing more to do in this world except PR and responding to data as it came in. (Which is pretty much what the MPC have done so far). However, we are not in that world. In the real world, the yield curve has turned out tighter than the MPC hope. Some of the reasons why that might be the case would actually warrant looser policy than in the imaginary world the MPC were hoping for. For example, if markets thought there was a fair chance that the inflation and financial stability knockouts, notoriously vague as they are, could be manipulated to provide a way for the MPC to renege on its forward guidance, and forecast tighter rates as a consequence, MPC would plan to keep rates lower for longer (or undertake more QE). (To be fair, there are also arguments weighing in favour of tighter policy too, but just for simplicity's sake let's omit them). What is missing from the MPCs execution so far is the explanation of how the differing conception of the world that gives rises to the discordant yield curve affects the MPCs own plans. Such a response is consistent with tying forward guidance to the unemployment rate, but is so far kept private.
So, I think the widespread media commentary on the two central banks is wrong to ridicule forward guidance. There is nothing wrong with changing your mind about a taper per se based on what you find out about people's reaction to it. And the fact that you can't ensure the private sector shares your forecast doesn't mean forward guidance has not worked - their forecasts of rates may be lower than in the absence of forward guidance. But executing forward guidance in a world of evolving an imperfect private sector interest rate forecasts involves forming a model of those forecasts, and how they react to what you do and say, and devising the guidance accordingly. Central banks cannot simply say what they plan to do and pretend that they will stick with it regardless of how others see the world unfolding, even if the central bank is, in a sense, right about how the world is going to unfold.
You could imagine a policymaker saying: Hang on, I'm not going down this road, because down this road leads to infinite regress. This might well be true. As well as forming a conception of how private sector forecasts evolve, central banks might, under some such conceptions, have to form a model of how those forecast evolutions depend on how the central bank responds to the private sector's forecasts, and so on, ad infinitum. However, if it were true, it would not necessarily do simply to brush it aside. And you can't just cut through it by pretending you won't react, or hoping that the world will somehow become more like a rational expectations world if you set the problem to one side.
This infinite regress would not be there for all possible models of private sector forecasts. For example, the central bank might believe that the private sector will set rates equal to their long term average plus something times the output gap, plus something times the inflation rate. On this view, the private sector aren't trying to pick the central bank's brains. They are using a mechanical tool to forecast rates. So there is no need for a view of what the central bank would do about a yield curve disagreement to enter their forecast. And hence no need for the MPC to worry about infinite regress! Unfortunately, in a world like this, there would also be no real room for forward guidance either, as no-one is listening. All that is going on is the mechanical feeding through of data.
As you read this, you might start to feel as though this is all just academic fun and games. However, it's not. In some sense, forward guidance, as conceived in the original academic work, and from which both the Fed and UK varieties draw inspiration (no matter what the UK's MPC says), embodies a contradiction. The original work was expressed in models of rational expectations. In those models there is no need for any announcements. There is only the matter of figuring out how to bind the hands of the policymaker (to stop her tightening rates when the recovery gathers pace, and the old worries about recession and deflation are long forgotten). The private sector know all there is to know, so they don't need Beige Books or fancy Inflation Report forecasts. The fact that all these things exist illustrates the obvious, that we are not in a world of rational expectations. However, since we are not in that world, the appropriate forward guidance will differ from the one devised in the academic studies (and presumably not just by the trivial fact that the real one has to be announced and described and reported on). The (somewhat unrealistic) example above with the private sector that simply forecasts rates based on a Taylor rule is a case in point, a case, in which, to repeat, forward guidance would be futile.