Earlier this week Boston Fed chief Eric Rosengreen in an interview on CNBC said that the Federal Reserve could introduce a forth round of quantitative easing – QE4 – since the beginning of the crisis in 2008 if the outlook for the US economy worsens.

I have quite mixed emotions about Rosengreen’s comments. I would of course welcome an increase in money base growth – what the Fed and others like to call quantitative easing – if it is necessary to ensure nominal stability in the US economy.

However, the way Rosengreen and the Fed in general is framing the use of quantitative easing in my view is highly problematic.

First of all when the fed is talking about quantitative easing it is speaking of it as something “unconventional”. However, there is nothing unconventional about using money base control to conduct monetary policy. What is unconventional is actually to use the language of interest rate targeting as the primary monetary policy “instrument”.

Second, the Fed continues to conduct monetary policy in a quasi-discretionary fashion – acting as a fire fighter putting out financial and economic fires it helped start itself.

The solution: Use the money base as an instrument to hit a 4% NGDP level target

I have praised the Fed for having moved closer to a rule based monetary policy in recent years, but the recent escalating distress in the US financial markets and particularly the marked drop in US inflation expectations show that the present monetary policy framework is far from optimal. I realize that the root of the recent distress likely is European and Chinese rather than American, but the fact that US inflation expectations also have dropped shows that the present monetary policy framework in the US is not functioning well-enough.

I, however, think that the Fed could improve the policy framework dramatically with a few adjustments to its present policy.

First of all the Fed needs to completely stop thinking about and communicating about its monetary stance in terms of setting an interest rate target. Instead the Fed should only communicate in terms of money base control.

The most straightforward way to do that is that at each FOMC meeting a monthly growth rate for the money base is announced. The announced monthly growth rate can be increased or decreased at every FOMC-meeting if needed to hit the Fed’s ultimate policy target.

Using “the” interest rate as a policy “instrument” is not necessarily a major problem when the “natural interest rate” for example is 4 or 5%, but if the natural interest rate is for example 1 or 2% and there is major slack in the economy and quasi-deflationary expectations then you again and again will run into a problem that the Fed hits the Zero Lower Bound everything even a small shock hits the economy. That creates an unnecessary degree of uncertainly about the outlook for monetary policy and a natural deflationary bias to monetary policy.

I frankly speaking have a hard time understanding why central bankers are so obsessed about communicating about monetary policy in terms of interest rate targeting rather than money base control, but I can only think it is because their favourite Keynesian models – both ‘old’ and ‘new’ – are “moneyless”.

I have earlier argued that the Fed since the summer of 2009 effectively has target 4% nominal GDP growth (level targeting). One can obviously argue that that has been too tight a monetary policy stance, but we have now seen considerable real adjustments in the US economy so even if the US economy likely could benefit from higher NGDP growth for a couple of year I would pragmatically suggest that the time has come to let bygones-be-bygones and make a 4% NGDP level target an official Fed target.

Alternatively the Fed could once every year announce its NGDP target for the coming five years based on an estimate for potential real GDP growth and the Fed’s 2% inflation target. So if the Fed thinks potential real GDP growth in the US in the coming five years is 2% then it would target 4% NGDP growth. If it thinks potential RGDP growth is 1.5% then it would target 3.5% growth.

However, it is important that the Fed targets a path level rather than the growth rate. Therefore, if the Fed undershoots the targeted level one year it would have to bring the NGDP level back to the targeted level as fast as possible.

Finally it is important to realize that the Fed should not be targeting the present level of NGDP, but rather the future level of NGDP. Therefore, when the FOMC sets the monthly growth rate of the money base it needs to know whether NGDP is ‘on track’ or not. Therefore a forecast for future NGDP is needed.

The way I – pragmatically – would suggest the FOMC handled this is that the FOMC should publish three forecasts based on three different methods for NGDP two years ahead.

The first forecast should be a forecast prepared by the Fed’s own economists.

The second forecast should be a survey of professional forecasts.

And finally the third forecast should be a ‘market forecast’. Scott Sumner has of course suggested creating a NGDP future, which the Fed could target or use as a forecasting tool. This I believe would be the proper ‘market forecast’. However, I also believe that a ‘synthetic’ NGDP future can relatively easily be created with a bit of econometric work and the input from market inflation expectations, the US stock market, a dollar index and commodity prices. In fact it is odd no Fed district has not already undertaken this task.

An idealised policy process

To sum up how could the Fed change the policy process to dramatically improve nominal stability and reduce monetary policy discretion?

It would be a two-step procedure at each FOMC meeting.

First, the FOMC would look at the three different forecasts for the NGDP level two-years ahead. These forecasts would then be compared to the targeted level of NGDP in two year.

The FOMC statement after the policy decision the three forecast should be presented and it should be made clear whether they are above or below the NGDP target level. This would greatly increase policy-making discipline. The FOMC members would be more or less forced to follow the “policy recommendation” implied by the forecasts for the NGDP level.

Second, the FOMC would decide on the monthly money base growth rate and it is clear that it follows logically that if the NGDP forecasts are below (above) the NGDP level target then the money base growth rate would have to be increased (decreased).

It think the advantages of this policy process would be enormous compared to the present quasi-discretionary and eclectic process and it would greatly move the Fed towards a truly rule-based monetary policy.

Furthermore, the process would be easily understood by the markets and by commentators alike and it would in no way be in conflict with the Fed’s official dual mandate as I strongly believe that such a set-up would both ensure price stability – defined as 2% inflation over the cycle – and “maximum employment”.

And finally back to the headline – “Time for the Fed to introduce a forward McCallum rule”. What I essentially have suggested above is that the Fed should introduce a forward-looking version of the McCallum rule. Bennett McCallum obviously originally formulated his rule in backward-looking terms (and in growth terms rather than in level terms), but I am sure that Bennett will forgive me for trying to formulate his rule in forward-looking terms.

PS if the ECB followed the exactly same rule as I have suggested above then the euro crisis would come to an end more or less immediately.