I just read an article by Marcus Nunes on his Historinhas blog, entitled "Bean counting at the BoE." In this post, Mr. Nunes discusses misunderstandings that the Deputy Governor of the Bank of England has about NGDPLT, where a central bank would manipulate the money supply so as to achieve a targeted path of nominal national output (NGDP).
I want to get past the question of whether NGDPLT should be tried now an think, instead, of whether it could be a policy for all seasons, so let us consider what the results would be of following NGDPLT under different scenarios.
First, consider the possibility that the growth rate of potential real GDP declines. Under this situation, the rate of inflation would rise somewhat. Over time this could be mitigated (if it is actually necessary) by lowering the growth rate of the NGDP target.
Second, think of the converse: the real growth rate of potential GDP rises. The natural outcome is that inflation declines, which sounds good. From experience, however, we know that low inflation will be reflected in the target interest rate, so if a large demand shortfall occurred, interest rate cuts would not be able to deliver as much help, although there are other potential channels for monetary expansion to help stabilize GDP along its nominal path (and they are less conventional mechanisms: think Operation Twist, the current MBS-based round of QE, or currency market manipulation, to name a few).
Those are the two main long-run scenarios, but consider the argument that NGDPLT would cause inflation expectations to become "unanchored." This idea does not make sense to me. If NGDP achieved a steady growth path of nominal national output, then the deviations in inflation would merely be complementary to any deviations in the growth of real GDP. The long-run inflation rate would simply be the difference between the target growth rate and the growth rate of real potential GDP. This does not sound like anything that should unanchor expectations. If anything, the outcome should be roughly as good as targeting inflation, except that negative or positive demand shocks would be counteracted more effectively.
Even if NGDP were not on target, a central bank with any credibility will "lean" in the direction of policy that will return NGDP to its target growth path, so a similar long-run inflation rate should develop.
The fact is that I am very sympathetic to the idea of NGDPLT. I am still not 100% sure it is always the right policy to follow, but I doubt it would result in inflation expectations becoming unanchored (as I have heard some suggest from time to time).