First, I take it as a good sign that Summers understands that the Phillips curve is not all it's cracked up to be:
...the idea that below normal unemployment will necessarily lead to accelerating inflation as suggested by the so called Phillips curve is very uncertain. Contrary to such predictions, inflation did not decelerate by much even a few years ago when unemployment was in the range of 10 per cent. Nor was there much evidence of accelerating inflation in the 1990s when the unemployment rate fell below 4 per cent.But Summers is trying to make the case that, for the Fed, continuing with ZIRP (zero interest rate policy) would be a good thing. He says:
...if inflation were to accelerate a bit this would be a good thing. It is now running and is expected to run below the Fed target. Prices are about 4 per cent below where they would have been if 2 per cent inflation had been maintained since 2007. So there is a case for some inflation above 2 per cent to catch up to the Fed’s price level target path. There may also be a case for inflation a little bit above 2 per cent for the next few years to allow real interest rates low enough to promote recovery when the next recession comes.So, here he seems confused. If the Phillips curve doesn't explain what's going on, how do we get more inflation with continued ZIRP except through a Phillips curve mechanism? Further, Summers seems worried about the "next recession." Presumably if the Fed still has ZIRP at that point, it's powerless (except perhaps with unconventional tools) to do anything about it.
Next, we enter the realm of the bad analogy:
...a plane that accelerates too rapidly as it takes off may cause passengers discomfort while a plane that accelerates too slowly may crash at the end of the runway. Historical experience is that inflation accelerates only slowly so the costs of an overshoot on inflation are small and reversible with standard tightening policies. In contrast, aborting recovery and risking a further slowing of inflation is potentially catastrophic — as Japan’s experience demonstrates. So in a world where economic forecasts are highly uncertain, prudence in avoiding the largest risks counsels in favour of Fed restraint in raising rates.His assumption, again, is that continued ZIRP will make the inflation rate go up. But "as Japan's experience demonstrates," 20 years of ZIRP just serves to produce low inflation.