I ENJOY reading the Atlanta Fed's macroblog. Most of the time what draws me is the analysis—of labour markets, macro conditions generally, that sort of thing. Lately, however, it has become useful as a source of insight into the mindset of the inflation-averse central banker, thanks to a some recent writing by the Atlanta Fed's executive vice president and research director, David Altig. Two weeks ago, Mark Thoma wrote a piece questioning whether the Fed's approach to its 2% inflation target is actually symmetric (such that downside misses generate as aggressive a response as upside misses). In recent years it has certainly seemed that while the Fed is interested in avoiding deflation, it is much more comfortable with an inflation rate just below 2% than one at or above 2%. Mr Altig took to his blog to defend the central bank, using this image:

The chart shows the results of an Atlanta Fed business survey question: "Projecting ahead, to the best of your ability, please assign a percent likelihood to the following changes to unit costs per year over the next five to 10 years." Mr Altig notes, "The obvious pattern in these survey responses is their asymmetry to the upside." It's very interesting to me that that is what Mr Altig sees in this chart. I'll tell you what I see. Despite the fact that the Fed chairman is famous for his research on the dangers of deflation and despite the fact that the Fed has repeatedly intervened to fend off deflation, most respondents expect unit costs to rise at 3% or less per year, on average, over the next five to ten years, with a meaningful probability of a rate of increase of 1% or below per year. I don't take a great deal of comfort from this image. Especially since 0% inflation seems to be much more costly than 4% inflation and since the Fed seems to much less comfortable raising the inflation rate than lowering it. A chart that suggests to me a Fed that's dangerously complacent about deflation risks looks to a central banker like evidence of upside asymmetry. (And of course, market expectations of inflation reinforce the notion that inflation is likely to be at or below historical levels for the foreseeable future.)

Then yesterday, Mr Altig addressed an argument made by economist Simon Wren-Lewis:

In another recent blog item (also with a pointer from Mark Thoma), Simon Wren-Lewis offers the opinion that acknowledging uncertainty about size of the output gap actually argues in favor of being "less cautious" about taking an aggressive policy course. The basic idea is familiar. It is a simple matter to raise rates should the Fed overestimate the magnitude of the output gap. But with the short-term policy rates already at zero, it is not so easy to go in the opposite direction should we underestimate the gap.

Mr Altig says he has no argument with this. But he appears to have an argument with this:

On the opposite side of the ledger, we know little about the conditions that would cause the Fed to lose credibility with respect to its commitment to its inflation goals, and very little about the triggers that would cause inflation expectations to become unanchored. Thus, I think it not difficult to construct a plausible argument about the risks of being wrong about the output gap that is exact opposite of the Wren-Lewis conclusion.

This sort of language is quite similar to that used by Ben Bernanke in defending his choice (or the Fed's choice) not to aim for a short period of above-target inflation. Unfortunately, Mr Altig puts no new meat on the bones of what strikes me as a very dubious position. Mr Wren-Lewis is right; we have an overwhelming body of evidence from recent decades and from across the rich world indicating that central banks can quite easily establish expectations of low and stable (with a downward bias) inflation. We have, to my knowledge, no recent examples of a major, rich-world central bank that wanted to keep expectations anchored at low levels and failed to do so. Central bankers like to say that low inflation expectations were "hard won". That seems like an iffy claim to me. Yes, the 1981-2 recession was a deep one, but the ensuing recovery was extremely rapid; if you asked them, I imagine most Americans would clearly trade this business cycle for that one if given the choice. And the Fed would almost certainly not have to work as hard now to bring inflation down given 1) the previous example of the 1980s disinflation and 2) the fact that inflation in the two decades prior to that was extraordinarily high while inflation over the past two decades has been extraordinarily low.

All of recent history, in other words, suggests that Mr Wren-Lewis is exactly right: it's much easier for central banks to go in one direction than in the other. Now one could, as Mr Altig says, come up with a "plausible argument" in which things don't work like that. Given the very large and ongoing costs of labour-market weakness, I would certainly expect America's central bankers to do better than that. I would like to see some very clear evidence that a year or two of 4% inflation poses more of a threat than at least a year or two more of unemployment well above the natural rate. What we're getting instead is little more than hand-waving.

Mr Altig is giving us a glimpse inside the mind of those making monetary policy. Unfortunately, the view is pretty disappointing.