Paul Krugman looks back on the past twenty years of macroeconomic policy and finds that his 1998 paper was more prescient than he or anyone could have imagined. Back then many observers assumed that central bankers--particularly those at the Bank of Japan--need only increase the monetary base to increase the price level. It was that simple.

Ken Rogoff, for example, said the following in commenting on Krugman's 1998 article:
No one should seriously believe that the BOJ would face any significant technical problems in inflating if it puts it mind to the matter, liquidity trap or no. For example, one can feel quite confident that if the BOJ were to issue a 25 percent increase in the current supply and use it to buy back 4 percent of government nominal debt, inflationary expectations would rise.
Krugman disagreed in his 1998 paper. He showed, using a New Keynesian model, that it was more complicated than many imagined. It depended on whether the monetary injections were expected to be temporary or permanent. Here is how he summarizes his 1998 article (my emphasis):
[T]he proposition that money issuance must raise the price level was false. Or if you like, it was missing a word: permanent money issuance would raise the price level. But a monetary expansion the private sector expected to be temporary, to be wound down after the crisis had passed, would do nothing at all: the extra monetary base would just sit there. Furthermore, it was reasonable for the private sector to assume that even large increases in the monetary base in a liquidity-trap economy would be temporary, to be wound down after the crisis had passed, would do nothing at all: the extra monetary base would just sit there
He goes to note that the public should, in fact, expect large expansions of the monetary base to be temporary. Otherwise, it would imply an implausibly large jump in the price level that would not be politically tolerated. For example, if the several-hundred percent increase in the U.S. monetary base expansion under the Fed's QE were expected to be permanent then the price level would have proportionally jumped several hundred percent as well. 

Krugman notes that the actual performance of Japan's first QE program of 2001-2006 and the Fed's QE programs bore out his predictions. These large-scale asset purchase programs ultimately proved to be temporary monetary programs. 

This is an important point and one that I stress in my own work. Just to be clear, permanent means an exogenous increase in the monetary base that (1) is beyond that required to accommodate normal money demand growth and (2) is not expected to be reversed. 

To illustrate this point, I want to repeat what I showed in an earlier post. There I used the Fed's median forecast of its assets through 2025 from its 2016 SOMA Annual Report to create projections of the Fed's balance sheet. These projections show the trend growth path of currency and a series I call the 'permanent monetary base' extrapolated to 2025. The latter series is the monetary base minus excess reserves. These two measures, which reflect the liability side of the Fed's balance sheet, are plotted along side the forecasted path of the asset side of the Fed's balance sheet.

Note that the Fed's median forecast of its assets eventually converges with the trend growth of currency which historically has made up most of the monetary base. Consequently, the permanent measure of the monetary base roughly tracks currency's trend path.

The figure implies the Fed's balance sheet forecast confirms the temporary nature of the monetary expansion  under the QE programs. That is, the Fed expects most of the permanent growth in the monetary base in 2025 to have come from the normal currency demand growth. This endogenous money growth would have happened in the absence of QE. There is no sign of an exogenous permanent increase in the monetary base. 

Krugman's bigger point is that to have robustly raised nominal demand over the past decade required a permanent increase in the monetary base. This did not happen at the Fed or the ECB. Krugman thinks it sort of happened in Japan under Abenomics. Maybe so, but I am not completely convinced. In my view, the shackles of inflation targeting made such permanent increases very hard to do at most advanced-economy central banks over the past decade. 

So what would a permanent monetary base expansion look like in practice? In my paper, I argue one could look to the experience of Israel over the past decade. An even better example comes from the U.S. economy coming out of the Great Depression. It is fairly easy to see the permanent jump in the monetary base during the early 1940s:

That is what Krugman would call being credibly irresponsible. It took a war to accomplish this permanent jump in the monetary base. The same can be done more efficiently and in a rule-like manner with a NGDP level target. Credibly going to a NGDPLT, however, would require a major regime change to U.S. monetary policy. And that brings us back to Krugman's article:
Given the way experience has undermined much of the original case for a 2 percent inflation target, and given the severity of the economic crisis, you might therefore have expected some  revision – a rise in the inflation target, or a shift to some other kind of targeting – price level or nominal GDP targeting. But that hasn’t happened... This is quite remarkable. If the worst economic crisis since the 1930s, one that cumulatively cost advanced nations something on the order of 20 percent of GDP in foregone output, wasn’t enough to provoke a monetary regime change, it’s hard to imagine what will.
His pessimism is understandable. However, I am more hopeful as noted in my last post. We are making progress. We have changed the conversation and that is the first step forward.