In case it's not clear from my comments, let me try to make it clear at the outset that most of what I say will disagree largely with not everything that Jim and Mario have said but with elements of what Jim and Mario have said and with only a little bit of what Scott has said.
I think it's important at the outset when thinking about monetary policy to recognize the modest impact that central banks can actually have on the economy over a sustained period. In fact, I think the disagreements that I have with Jim and Mario and possibly with Craig will come down to that point: what kind of effect can the central bank have on the economy over short periods of time, and what kind of effect can it have and on what variables over long periods of time?
I believe that Canada's inflation targeting framework has been a great success. The system began in 1991, as you know. The inflation target has been at 2% since 1995, and since that time inflation has been lower and more stable than it was in the previous 30 years. I think that stability of inflation is good for the stability of the economy. I think it is good for planning. A very important part of inflation targeting is the extent to which it helps to anchor the expectations of inflation--the expectations of households, of firms, of participants in financial markets--in response to shocks of various kinds. And of course our economy is subject to shocks of various kinds.
As you all know, the bank and the government recently announced their new agreement for five more years, which will leave the status quo largely unchanged, and, all things considered, I think that was actually a good decision. So let me be here today as a defender of the status quo, although there were some alternatives the bank and the government considered, which could have been chosen.
Targeting nominal GDP, as Scott said, is one alternative. As Scott explained, a perceived advantage of targeting nominal GDP is that it gives the bank--or any central bank, but the Bank of Canada in this case--more flexibility in being able to take its eye off of inflation and be concerned a little bit more about real GDP growth in the short run.
In my view, the problem is that makes the bank officially indifferent, if you like, to various combinations of real GDP growth and inflation. The sum of those two is nominal GDP growth. So if it is going to put more emphasis on real GDP growth in the short run, by definition it must put less emphasis on inflation. The result of that, I think, is almost certainly that inflation will be a little bit more volatile over the business cycle.
The danger with that is that it may dislodge those well-entrenched, well-anchored expectations. I am concerned about that possibility, given that it takes a long time for monetary policy to develop the credibility necessary to actually establish and anchor those expectations.
Scott also said that you could have the same long-run trend of inflation in a world of nominal GDP targeting, and that's true under some conditions. Canada, more so than the United States, is faced with an aging population. They have aging there as well, but their demographics are a little bit better than ours, and we will have a declining rate of real GDP growth over the next 15 years, due simply to the aging of the population and the withdrawal of retired Canadians from the workforce. Other things being equal, that reality will reduce the rate of GDP growth.
If we were targeting nominal GDP from now to the future, in a world where real GDP growth on trend would decline, there would be a trend to increase in inflation. I see no reason why we should be sanguine about an increase in the costs that come from that inflation.
Another option is targeting full employment. Mario talked about how in the past the bank did not work hard enough to combat the unemployment rate, and Jim talked about how the bank might pay more attention to variables other than just inflation. So, at least on the surface, targeting full employment is an attractive option.
The problem is the concept of full employment is very slippery. You can look all you want at the data published any place by Statistics Canada and you will not see a measure of full employment. Why? Because it is not something that is directly measurable. It is a concept that is very useful and that economists use to organize their thinking about the operations of the economy, but it is not something observable in the data. You can only use our models to back out an estimate of full employment, but the models of course are contentious.
Mario's preferred model will almost certainly be different from Jim's, mine, and Craig's. That means if we're going to disagree about the models, we will disagree about the estimate of full employment that comes from those models. Then if you have a target for the central bank around which there is disagreement on the measurement of the variable, what basis is that for actually targeting monetary policy?