Fewer Fiscal Multipliers and more Clarity from the Bank of Canada

At the risk of being thought "cavalier", I don't like what Bank of Canada governor Stephen Poloz is reported as saying by Kevin Carmichael.

Inflation targeting is supposed to be inflation forecast targeting. The Bank of Canada is supposed to do what is needed to ensure that its own internal forecast of future inflation (at a roughly 2-year horizon), conditional on the Bank's information, remains at the 2% target.

Inflation targeting is supposed to be symmetric. The Bank is supposed to respond to new information about any shock to Aggregate Demand to prevent that shock affecting its forecast of inflation, whether that shock is positive or negative.

The Bank of Canada is not supposed to follow any simple instrument rule, like a Taylor Rule, if that simple rule leaves out any information that might be relevant for the Bank's internal inflation forecast. For example, the only information in the Taylor Rule is the output gap (the difference between current output and potential output) and the inflation gap (the difference between current inflation and target inflation). Using a Taylor Rule implies ignoring any information the Bank of Canada might have about shocks to the natural rate of interest. [I say this notwithstanding the internal inconsistency, like the Cretan Liar Paradox, of the Bank's own Totem model assuming that its own future actions will follow a slightly-modified Taylor Rule.]

The whole point of inflation forecast targeting is that the Bank of Canada does not "accommodate" any shock to Aggregate Demand that would change its internal forecast of future output and hence future inflation. And that includes fiscal policy shocks. The Bank of Canada is supposed to offset any such shock, 100%. It makes no sense for the Bank of Canada to provide any estimate of fiscal policy "multipliers" other than zero. Even calculating a hypothetical "multiplier", for how output would respond if the Bank did "nothing", makes no sense, because the definition of "doing nothing" is arbitrary. Even if we agree that "doing nothing" means "doing nothing with the nominal rate of interest", that leaves open the question of how long the Bank holds the nominal rate of interest constant. And "forever" is not a meaningful answer to that question, because standard New Keynesian models (and the Bank is New Keynesian) tell us that monetary systems will either explode or implode if the central bank holds the nominal interest rate constant forever. [The Bank says it does accommodate the "first-round effect" of short-run supply-side or price shocks, like changes in indirect taxes, but that's not what we are arguing about here.]

When we talk about the Bank of Canada offsetting rather than accommodating changes in fiscal policy, it is important to understand that we are talking about changing the nominal interest rate relative to what it would have been otherwise without the fiscal policy change, and not relative to what the nominal rate was in the past. It's a statement about counterfactual conditionals, not about changes over time. For example, if the Bank gets new information that leads it to expect inflation will fall below target unless it cuts the overnight rate of interest, and at the same time gets new information about fiscal policy, the Bank might decide to leave the rate of interest where it is. The Bank offsets the change in fiscal policy by cancelling a planned cut to the overnight rate.

If the Bank of Canada does what it is supposed to do, and what it says it does, then a temporary increase in the fiscal deficit will cause a temporary rise in the nominal and real interest rate (and nominal and real exchange rate), relative to what would have happened otherwise. [See my old post.] It will have zero effect on the Bank's expectation of the output gap and zero effect on the Bank's expectation of inflation. The Bank's expectations will almost certainly be wrong, of course, because it lacks a true crystal ball model of the economy. But those errors could go in either direction.

I want to interpret Senior Deputy Governor Carolyn Wilkins in Kevin Carmichael's interview as saying:

"Yes I agree with all that, and we welcomed the change in fiscal policy because it meant we could keep forecast inflation on target without having to cut interest rates, which we would otherwise have done. This is a good thing because needing to cut interest rates would have increased the risks to financial stability." That would be an internally consistent thing for the Bank of Canada to say.

If she had added: "Plus, even though we are currently above the Effective Lower Bound on nominal interest rates (which is probably below 0%) we are worried that the margin of safety is getting a bit small, and are pleased that fiscal policy is making that margin of safety a bit bigger than it otherwise would be" that would also be an internally consistent thing for the Bank of Canada to say.

I don't get this bit though:

"Fiscal policy at low interest rates is also just more effective."

You can get that from a model where the Bank of Canada holds the money stock constant, and the demand for money gets increasingly interest-elastic at lower interest rates so the LM curve gets flatter. But the Bank of Canada does not hold the money stock constant. If the Bank of Canada held the nominal interest rate constant the LM curve would be horizontal. If the Bank of Canada adjusts the interest rate and/or money stock to keep output at potential, and inflation on target, then it tries to make the LM curve vertical at potential output.

Back to marking exams.

Comments

[Scott Sumner]

Excellent. As long as Canadian rates are not at infinity, the BOC still has room to raise them (and hence offset fiscal expansion). And 1% is a long way from infinity.

[Nick Rowe]

Thanks Scott!

[Majromax]

Being charitable, we might also think that the Bank of Canada is admitting either a nonlinear policy response or a "latch" policy response.

Suppose, for example, the Bank of Canada foresees inflation returning to its target over four years with a given set of CB actions. That would be consistent with a medium-term inflation target. Now, if the fiscal authority conducts deficit spending such that the CB thinks inflation will now hit its target in two years, that's still consistent with the same set of CB actions (over the first two years, anyway), yet it is also stimulative.

I think this is consistent with global CB actions so far. Between low inflation rates, low inflation targets, and the ZLB (at least in terms of policy), central banks have great difficulty in setting large negative (effective) real interest rates. The best we can hope for is -1% to -2%, which should result in a return to target but may not do so quickly. It's hardly the -10% to +10% range we could see in the early 80s.

Finally, there may also be "concrete steppes" concerns. The two most recent BoC rate cuts, for example, were met with smaller reductions in commercial banks' prime lending rates to consumers, which limits the influence of those rate cuts on general private-sector demand. The BoC could continue the beatings (that is, cut rates further) until morale improves, but that would be unconventional policy.

[Tel]

That interview reads like so many party apparatchik buzzwords strung together to the point of being meaningless. I can't believe anyone takes that stuff seriously.

The issue is about control, in particular who can control the flow of money and how to redistribute to the politically connected interest groups. Governments have more options in this regard... drawing an example from Australia we had the "National Broadband Network" which was a project clearly not commercially viable, but government dumped taxpayer money into it, no bank would ever have offered money for that. The Australian government also put through a bunch of anti-competitive laws in the hope of removing some of the market competitors but regardless of that the project failing anyway and the government is having difficulty fully enforcing such anti-competitive laws in practice.

The problem for government is that these heavy handed "infrastructure" projects are also very obvious, and hard to hold at arm's length, so everyone knows exactly who the winners and losers are. Central banks can be a bit more subtle about the whole thing, which also ties their hands to some extent.

Cantillon effects.

[Nick Rowe]

Majro: you might, just, be onto something there. The Bank has always interpreted the "flexible" part of "flexible inflation targeting" to mean "we don't always try to get back to 2% instantly". (For good reason, since doing so might cause some very severe relative price changes and real distortions, given some prices are stickier than others). Originally it defined the "medium term" as "6 to 8 quarters" (meaning 12-month inflation ending 6 to 8 quarters ahead). More recently it has adopted a more "flexible" definition of "flexible", meaning "the length of the horizon depends on how big the missed shock was".

But it is not obvious to me why fiscal vs monetary would affect the optimal horizon. Plus, 12 month core inflation was 1.9% in February and 2.1% in March. (The Bank targets core at a medium horizon and headline at a longer horizon).

Tel: try being interviewed when you are #2 in a large organisation! You have to be very careful what you say. Especially when lots of people could make big money if they could interpret what you are saying better than everyone else.

Whether we like it or not, central banks exist. They have control. The question is: how best to use that control.

[Tel]

Nick, I'm well aware of how Fedspeak operates, that doesn't mean I have to respect such behaviour.

Whether we like it or not, central banks exist. They have control. The question is: how best to use that control.

Fatalism, do you look both ways before crossing the road?

The answer is already well known, the best thing central banks can do is stabilize the currency. They even give lip service to this goal themselves, they just consider "stability" to mean steady inflation plus QE whenever they feel like it. Clearly this is not stability under the normal sense of the word.

The problem is Keynesian stimulus, and central bankers are as guilty as anyone of propagating the myth that this can work. Dual mandate implies no mandate at all.

[jonathan]

It's difficult to interpret her answer because she sort of vaguely waves in the direction of several ideas:

"There are concerns about the effects of persistently low interest rates and the quantitative easing that other countries have done, in terms of increasing risk-taking by financial market players and individuals. If fiscal policy can do some of the heavy lifting, that’s a positive thing. Fiscal policy at low interest rates is also just more effective. In a world where growth is going to be structurally slower because of demographic changes, monetary policy can’t fix that. If we want sustainable growth, we need to boost productivity, not only in Canada [but in] the global economy. That’s the only place growth can come from."

The first part (about quantitative easing) seems to be talking about some possible problems of low interest rates for financial stability. This is sensible, as you indicate in your post.

The "fiscal policy at low interest rates is also just more effective" bit is a little ambiguous. But immediately after this she argues that low interest rates indicate higher returns to public investment relative to private investment. So that's one possible interpretation of the sentence -- it's not that fiscal policy is "more effective" because there's a big multiplier, but because public investment carries a high return relative to private investment, and so will boost growth.

[Nick Rowe]

jonathan: I missed that bit about relative returns to public vs private investment. In principle, just as lower interest rates are supposed to make private investment more profitable, they should do the same to public investment. (Though that's holding a lot of other things equal, of course.)

[Majromax]

> In principle, just as lower interest rates are supposed to make private investment more profitable, they should do the same to public investment.

The statement makes more sense if we interpret it as referring to real interest rates. If market interest rates are low, then that suggests that the return on the marginal private-sector investment is also low due to either overinvestment or poor a supply-side environment.

However, the public cannot contribute to the marginal public-sector investment, since the level of public investment is determined by the government and not markets. If government has been following an a-priori prescribed level of public investment, then it's quite possible that the rate of return on the marginal public investment is greater than the rate of return on the marginal private investment.

This is the opposite of the crowding-out effect, where positive real-return private investments are foregone because the public sector is overpaying for lower-return public spending.

I wonder what would happen if the monetary authority committed to inflation targeting and the fiscal authority committed to a fixed target of debt service (say, 2% GDP). With this pair of instruments, might the Wicksellian rate be controlled?