Response to Simon Wren Lewis: the output gap and inflation, necessity for a fiscal stimulus, and all that

Simon was very kind to dignify my blog post on why there should be no more fiscal stimulus with a response.   I don't have much to add:  Simon inhabits (is in fact a far more senior participant in) the same empirical/New Keynesian macro world that I do, and probably knows what I think better than I know myself.  However:

I accept that the low level of output, relative to its level pre-crisis trend, and relative to either a straight or bendy line drawn through a long time series of output, is neither normal nor desirable.  I also accept that high unemployment is neither normal nor desirable.  My position is that seen through the lense of almost all coherent accounts of the macroeconomy that we have currently, high and constant inflation is another way of inferring that the undesirably abnormally low output and high unemployment is not something that monetary or fiscal policy can do anything to alter permanently.   Once again, looking at the issue through these same models, both instruments could be used to boost output temporarily, at the cost of a temporary surge in inflation.  But no permanent improvement in the real economy could be bought without ever increasing inflation, which, ultimately would impoverish all of us, and hit those least indexed (protected from inflation) first, which usually means the poorest.

I don't mean to imply that just because conventional monetary or fiscal policy stimulus shouldn't be used to try to lower unemployment and increase output, there is no role for public policy to improve things.

I think it entirely plausible that there are other reasons beside deficient demand for high unemployment and low output.  The macro-models of financial frictions that we have explain how starving the productive sector of access to funds will reduce the demand for factors (eg labour), and inhibit the reallocation of funds across different sectors in the economy.  The recent literature on 'misallocation' is set in the stark world of real business cycle models, which I think are question-begging, but they demonstrate that there can be very chunky changes in the aggregate economy's capacity to produce coming from changes in how well funds are allocated across sectors, (eg in the 5-10% range).  I don't claim that this is the whole story, but it convinces me that this is a plausible account of high unemployment coexisting with high inflation.  And if this is part of the story, then the role for public policy to improve output and lower unemployment lies in unlocking the flow of finance to firms who can't get enough of it.  This could in fact involve more government borrowing (eg if it entailed funding some new body to lend to SME's, or recapitalising further existing Banks).  (A little confusing, that, since up to this point I have been arguing against further government borrowing.)

I don't accept the argument that the usefulness of the inflation rate as a window into the output gap is any diferent at high or low inflation.  This is not a statement you can make in the context of empirical models only.  At least if you define the output gap to be the difference between actual output and the imaginary level that monetary or fiscal policy should seek to bring about.  It's interesting to observe that the correlation between a change in inflation, and the difference between actual output and alternative trend lines, is itself dependent on the level of inflation.  However, in the absence of a sound theory to explain them, in particular to connect those gaps between actual and trend to quantities that monetary or fiscal policy could or should do anything about, I don't think they are suggestive of whether there should or should not be more fiscal stimulus.   The variability of this correlation also does not invalidate many of the kind of models I'm talking from:  in fact, my hunch is it only invalidates the simplest New Keynesian model with one shock, in which there is no structural change.  That's not to say that data-based, what we could call atheoretical, empirical research isn't valuable.  (Most of what I have done falls into this category!).  On the contrary, it's essential.  However, it's hard to draw policy conclusions from it alone; or, at least, if one does, you run the risk of falling foul of the Lucas Critique.  [of finding that the correlation you exploit doesn't survive the attempt to exploit it].

The use of 'sound' and 'coherent' is a shorthand really for  models that the successors to Kydland, Prescott and Lucas would approve of, at least in the sense that the models display an attempt to grapple with the issue of what people and firms are trying to do in the model and how they do it.  (All three would disapprove of the sticky price sticky wage world of modern monetary policy models, and the many other reasons those models give for why business cylces are not efficient, but they would at least recognise the recipe used to build them).   For examples of the kind of models I do think are capable of speaking to fiscal policy, well, see Simon's webpage, and his previous research on fiscal policy with, eg, Campbell Leith!

The comments I make about the undesirability of more fiscal stimulus do not hinge on assuming that there is linearity in the world.  For example, constant inflation would indicate a zero output gap [once again, not necessarily desirable or high output] in the original nonlinear sticky price model.

All of my comments hinge on believing that the mainstream macro models are what we should be using to answer the big macro questions of our day. I don't believe that these models are correct or should be taken literally.  The crisis is warning enough that many of them are missing lots of things whose ignorance hampered macro and financial policy.  The appropriate way to answer these questions is not to put your faith in any one model either, but, if you were to do it mechanically, lay out all the models you might believe in, put probability weights on them, and analogously probability-weight the corresponding policy implications.  Its certain that all of the models you will have at your disposal are wrong.  And so the precise truth about what fiscal policy should be won't ever be known. But it's not at all obvious to me that the as yet unknown right model, articulating all the exotic nonlinearities we often sweep away, would tell you that demand is deficient and more fiscal stimulus is needed.  Even if such a model did, one would have to contemplate the equally plausible exotic possibility that we are on the knife edge of causing an unstoppable inflationary spiral, as observers conclude the consensus for low inflation is evapourating.   I don't personally believe this is a material risk, but equally, looking at the evidence, I don't think it's any more outlandish than discarding the diagnosis that no more stimulus is needed encoded in the high and stable inflation rate.

Finally, if its not apparent already, I accept that reasonable people can disagree on this, and Simon's challenge to me (and to the macro literature which I am speaking from) is very fair, and, given the evaporation of the 'great moderation/stability' recently, far from answered in this post.