Temporary vs permanent fiscal policy in a small open inflation targeting economy

Just a short note as a backgrounder for the current Canadian debate about fiscal policy.

This is totally unoriginal boring textbook stuff (at least I hope it is, anyway). Prerequisite: intermediate macro (or special permission to skip to the results if you promise not to ask me daft questions about where they came from).

The simple second year textbook Mundell-Fleming ISLMBP model under flexible exchange rates makes two assumptions that we need to modify:

1. The simple textbook version assumes the money supply is fixed. In other words, it assumes the Bank of Canada targets the money supply. But the Bank of Canada targets the inflation rate, and it adjusts the money supply (adjusts the nominal interest rate and allows the stock of money to adjust, if you prefer) to keep its internal forecast of the inflation rate at 2%. The easiest way to build this into the textbook picture is to assume the Bank of Canada shifts the LM curve to try to keep the ISLMBP equilibrium point at (what the Bank of Canada thinks is) potential output Y*.

2. The simple (or simplest) textbook version (normally) assumes the exchange rate is not expected to change. The expected rate of appreciation/depreciation is assumed to be zero, so the BP curve is horizontal at the exogenous world interest rate. But if the change in fiscal policy (e.g. an increase in Government expenditure) is temporary, the exchange rate will appreciate when G increases, and will depreciate again in future when G falls back to normal. And if the foreign exchange market is not "dumber than a sack of hammers" (in Stephen Gordon's unforgettable phrase) it might figure this out. If so, the expected exchange rate depreciation following the initial (surprise) appreciation will cause the BP curve to shift up. Canadian interest rates will rise relative to world interest rates, because the Looney is expected to be depreciating back to its original level. But if the change in fiscal policy is expected to be "permanent" (or at least be wound down slowly enough that this effect can be ignored) the BP curve will not shift up.

Permanent increase in G.

The increase in G causes the IS curve to shift right, which causes an incipient (lovely word) increase in the Canadian interest rate which causes an incipient excess demand for Looneys on the foreign exchange market, which causes an exchange rate appreciation which causes Net eXports to fall which causes the IS curve to shift left again to where it started. The Bank of Canada does nothing because it doesn't need to do anything.

TLDR: A "permanent" increase in G (or one that is expected to be unwound very slowly over time) causes an appreciation of the exchange rate and a fall in Net eXports equal to the rise in G, with no change in interest rates, output, inflation, whatever.

Temporary increase in G.

Same as before, the IS curve shifts right and the exchange rate appreciates. But this time the BP curve shifts up, because the exchange rate is expected to depreciate as the increase in G is unwound. If the Bank of Canada held the money supply constant (like in the textbook version) the Aggregate Demand curve would shift right, output would increase, and inflation would rise above the Bank's forecast target. So the Bank of Canada will decrease the money supply to shift the LM curve to the left to prevent this happening. This is called "monetary offset"; the Bank of Canada is in charge of the AD curve, and does what it needs to to shift the AD curve to where the Bank of Canada thinks (rightly or wrongly) it ought to be to keep future inflation at the 2% target.

TLDR: An increase in G that is expected to be temporary (unwound relatively quickly) will cause both an exchange rate appreciation and an increase in Canadian interest rates. The fall in Net eXports plus the fall in Investment and Consumption will equal the rise in Government expenditure. No change in output or inflation.

Bottom line: if (for some reason) you want temporarily higher Canadian (real and nominal) interest rates then you should ask for an increase in Government spending or Tax cut, but it should be very clearly announced that the higher deficit will be wound down quickly to 2x4 the foreign exchange market into figuring out that expected depreciation is a rational expectation. If the forex market doesn't get the message, or doesn't believe the message, it won't work.

[Of course, if you are a microeconomist, who doesn't understand all this macro cr*p, but reckons that low real interest rates raise the Net Present Value of some government investment projects, and makes some of them them now NPV > 0 in a standard micro Cost Benefit Analysis, then you go right ahead.]

P.S. I ignored Ricardian Equivalence, but if you did want to assume it: a permanent increase in G won't even change the exchange rate (unless the government has different preferences for domestic vs imported goods than the private sector); a temporary increase in G has roughly the same effects as what I said above regardless of whether Ricardian Equivalence is true (though smaller effects the longer it is expected to last), but temporary cuts in T don't matter at all.

Comments

[John Handley]

Nick,

Let's change the assumptions just a little bit here and have output start out below potential. In this case, assuming the central bank refuses to loosen monetary policy [I'll try to justify this in a little bit, bear with me], fiscal policy can effectively return the economy to equilibrium. Is Canadian GDP currently below potential? Well, you would know better than I do, but, considering the trickle of news about Canada that reaches California, this seems to be the case.

Regarding the central bank's 'refusal' to loosen monetary policy, I offer the following explanation:

Suppose there is an inflation targeting central bank that has committed to grow the money supply at a rate consistent with its inflation target to the infinite future. Let's say that the divine coincidence doesn't hold perfectly in this particular situation, so there is a recession even though inflation is on target. The central bank now decides to engage in a temporary monetary stimulus -- i.e., a monetary stimulus that won't compromise it's long run inflation target. This means that the current money supply must expend, but that future money growth is expected to be lower so that the inflation target is kept [The reason money growth has to decline to keep inflation on target is that prices are sticky, so a permanent increase in the money supply met with constant money supply growth in the future will mean higher than normal inflation as prices fully adjust to there new equilibrium level]. If interest rates are already low, as they are in Canada, this temporary increase in the money supply might mean that the central bank hits the zero lower bound at which point the usual liquidity trap stuff happens; bonds and money are perfect substitutes and further monetary expansion is hoarded because agents refuse to economize on their cash balances (you're probably itching to try to refute this, but at least hear me out for the sake of argument).

I digress,

If monetary policy is in some way constrained, say by a commitment to an inflation target combined with already very low interest rates, so that it can't effectively keep output at potential or counteract the demand-side effects of fiscal stimulus, then fiscal policy is an adequate stabilization policy, at least in textbook models like ISLMBP.

I think to simply assume that monetary policy is effective at all times makes your analysis of fiscal policy in this case too simplistic. If the overnight rate in Canada were significantly above 0.5%, then the basic monetary offset analysis would be completely valid, but currently I think you at least should make the case that monetary policy remains unconstrained in Canada.

[Nick Rowe]

John: shorter version: yes, if you assume the Bank of Canada is unable or unwilling to loosen monetary policy to depreciate the exchange rate, you have effectively the fixed exchange rate version of ISLMBP, where fiscal policy works.

[Makroint]

I would slightly question the first result because thinking further ahead, is a permanently higher exchange rate even possible? That would be a country getting richer simply by permanently increasing government spending. I guess in case a deficit could occur forever this could be right, but assuming that the current account over the very long run has to be in equilibrium, the exchange rate might be unchanged. I would suspect that in this case one would switch to the temporary expansion case, as though G is expanded forever, other domestic demand has to be reduced in the future. Flawed thinking?

[Nick Rowe]

Makroint: I think that's correct thinking. Net exports are lower, so we are borrowing from abroad. We can't keep doing that indefinitely (unless it is being used to finance investments that pay for themselves). But you are now getting into long run theory, and the distinction between GDP and GNP, and this little model is really a short run model, that isn't designed to handle questions at that sort of time-horizon. That's why I said the "permanent" increase in G was really one that was unwound "slowly" over time. How slow is "slow", is a question I ducked. (And it depends on the term of the interest rate).

[John Handley]

Nick,

Isn't the question of whether or not monetary policy is impotent much more important to the current situation, though? Everyone agrees that monetary offset can theoretically contradict the demand side effects of stimulus, but not everyone agrees that this is always the case. Is this the case in Canada? If not, why not (rhetorical questions, but feel free to answer anyway if you want to)

[Nick Rowe]

John: the BoC is well above its self-declared Effective Lower Bound. And that's even before we start talking about "forward guidance", "QE" etc. Plus, the idea that the exchange rate is fixed, and independent of BoC announcements, does not look very plausible right now.

[John Handley]

Nick,

"the BoC is well above its self-declared Effective Lower Bound."

This does not necessarily mean that they could return output to potential without having to go below the their effective lower bound. With the overnight rate at 0.5%, there is at least a case to made that this is the case, especially since the BoC is committed to 2% inflation in the long run and won't engage in any large permanent stimulus for this reason.

Regarding 'QE,' the explanation I gave for monetary policy impotence above explicitly explains why QE should be useless, especially if the BoC only interferes in the Canadian government bond market. Forward guidance is also explicitly dealt with in my explanation above; forward guidance requires for the central bank in question to allow inflation to be temporarily above target, which won't happen under strict inflation targeting. At least when it comes to concrete steppes, it appears to me that the BoC's hands are basically tied. But, fine, suggest that agents will irrationally believe in promises that can't or won't actually be kept (e.g., excessive additional monetary stimulus at the effective lower bound, forward guidance combined with a period of inflation in excess of the inflation target, effective large non-permanent monetary expansion).

[Nick Rowe]

John: when the BoC announced "no change" this week, the exchange rate jumped nearly 1% on the news. Forward guidance on interest rates means the BoC can operate on the whole term structure of interest rates, the long end not just the short end, where the long end may be well above the ELB. And even if this does mean a trade-off between future inflation rising above target to prevent present inflation falling below target, why not exploit it? And until the BoC runs out of assets to buy, we cannot say that QE is tapped out. Plus, empirically it wasn't tapped out in the US, and Canada is in better shape.

But yes, this would all be easier with a level-path target.

[Scott Sumner]

John, Nick's right the BOC can always depreciate the Canadian dollar. The zero bound is not an issue in Canada.

Nick, You said:

"Of course, if you are a microeconomist, who doesn't understand all this macro cr*p, but reckons that low real interest rates raise the Net Present Value of some government investment projects, and makes some of them them now NPV > 0 in a standard micro Cost Benefit Analysis, then you go right ahead."

I get a lot of Keynesians making this argument in my comment section. But it proves too much. If it were true for government projects, it would also be true for private investment projects. In that case you'd expect investment to be higher when rates are low. But we observe the opposite, investment tends to be higher when interest rates are high.

One possibility is that corporate executives are dumb, and don't understand that the NPV of their potential investment projects is higher at low rates. A more likely explanation is that the low rates are CAUSED by a shift left in the investment schedule, for both corporate and government investment.

In other words, never reason from a price change.

[John Handley]

Nick and Scott,

Both of you like to refer to the effects of central bank announcements on exchange rates, but you miss the point. In the US, the lack of inflation given the size of QE is evidence in and of itself of a liquidity trap. There are only three explanations that I've heard and/or can think of for extremely high money demand: 1) Liquidity traps exist and the fact that the FFR was about equal to IOR made excess reserves completely useless. This explains why the velocity of the monetary base stopped looking like pretty much a linear function of the FFR; velocity plummeted further every time we have QE even though interest rates went nowhere, so obviously excess reserves don't do much of anything. 2) The money demand function is exponential, so the marginal effect of monetary expansion on NGDP becomes extremely small at high levels of the monetary base (this is basically the MIUF prediction; the velocity has to rise to infinity if FRR - IOR = 0). 3) It's all IOR (I've only heard this from Market Monetarists, specifically Scott); 0.25% IOR was responsible for soaring base demand.

Naturally, I'm most sympathetic to the first explanation, as it is the only one that I think readily explains declining velocity at a basically constant FFR and I'm least sympathetic to the last view because it treats IOR as a distortion in its own right as opposed to what is actually important; the spread between the short run risk free interest rate and IOR.

If you ignore markets for a minute and focus only on relevant aggregate time series (e.g., inflation, real GDP, employment), then QE certainly looks like an abject failure, unless your counterfactual is deflation and constant unemployment, I guess.

Besides, the existence of monetary offset is not itself an argument against fiscal stimulus. You said it earlier, Nick: "if (for some reason) you want temporarily higher Canadian (real and nominal) interest rates then you should ask for an increase in Government spending or Tax cut." If looser fiscal policy means that the BoC will be able to raise interest rates without causing a recession, then why not have looser fiscal policy?

[Nick Rowe]

Scott: I've been (silently) thinking over your point about government vs private investment. I think I get your point, but I'm not sure it's right. And not just because the government and private sector invest in different sorts of things, so their NPV's won't be perfectly correlated.

This, I *think* is the key difference:

If recessions are caused by an excess demand for money (which you and I think they are, though you would emphasise MOA and I would emphasise MOE) then it matters a lot whether the (future) benefits of an investment project will be sold on the market in a monetary exchange. Consider for example home production of consumer durables, which is home production of investment goods. You "sell" those benefits to yourself, with no money changing hands, so sticky prices in terms of MOA, or excess demand for MOE, don't matter. So home production of investment goods for home use should be acyclical or even countercyclical. Like investment in human capital.

What I conjecture is that many government investment projects are more like home production than business investment in machines.

That is not very clear. Here is my related post. Monetary Business Cycles and Countercyclical human capital investment.

[John Handley]

Nick,

"And even if this does mean a trade-off between future inflation rising above target to prevent present inflation falling below target, why not exploit it?"

I'm not arguing that they shouldn't, I'm arguing that they will not. Just look at the US, for example. The Fed has refused to allow current inflation to go above target, which, so long as that's what everyone expected (which would be the case with rational expectations), would mean that forward guidance failed in the US. Maybe the BoC will be better, but, at least with strict inflation targeting, I doubt it.

[Donald Pretari]

Hey Nick, I just got my proof of Canadian citizenship, so I'm now a very proud citizen of both Canada and the U.S., and can comment on this blog now as a Canadian. It still freaks me out that we first met in September of 08, when you helped me understand helicopter money, among other things. Cheers. Don

[Nick Rowe]

Don: Congratulations!

[Donald Pretari]

Thanks Nick

[Scott Sumner]

John, I most certainly do not claim that the high demand for base money is all due to IOR. Base demand soars at zero rates, even without QE.

You are looking at the causal effects of QE, whereas it makes more sense to view QE as the effect of a tight money policy that drives rates to zero. If you do a more expansionary monetary policy, such as currency depreciation, then you do not need as much QE. QE is a defensive mechanism, monetary policy needs to be viewed in terms of the policy goals of the central bank, and in terms of whether it will do whatever it takes to reach those goals.

Nick, I don't quite buy that argument. Most government investment is not like home production, it's a close complement to private investment. So in a deep slump the optimal amount of new houses, offices building, shopping malls, etc., falls sharply. And with it the optimal amount of new roads, schools and sewer systems also drops sharply, because this infrastructure is mostly built to serve (geographically) growing communities. So I continue to believe it is reasoning from a price change. Take an extreme case, an economy with lots of SOEs. Does it make sense that the optimal amount of new steel mills built by SOEs would rise in a recession, but the optimal amount of new steel mills built by a private sector competitor would fall?

[Nick Rowe]

Scott: My argument wasn't very clear. And it does depend on what caused the rate of interest to fall. But Canada is a small open economy, that takes the world interest rate as given, (just like a Hayekian individual who does not need to know why the price of copper (or whatever it was) increased). Whatever it was, it is unlikely to have been perfectly correlated with any shock internal to Canada. Plus, the microeconomist doing my hypothetical NPV calculation would be doing fundamental analysis, starting from scratch. He would not just *assume* that the other terms in his NPV calculation stayed the same when r fell; he would check.

[Nathanael]

In your first result, you're assuming fully competitive markets in exports and imports. Which may be the case for Canada, which sells a lot of fairly common commodities and doesn't seem to import anything much which it can't produce.

But which is not true in general; for a lot of one-source or two-source goods, the export market is quite fixed and the customers aren't going to go somewhere else or find substitutes. I suspect the import market can also be uncompetitive for the same reasons.

[Nick Rowe]

Nathanael: No. I'm (implicitly) assuming the opposite. As Canadian produced goods become closer and substitutes to rest of world goods (as the elasticity of net exports with respect to the real exchange rate gets larger), we need a smaller and smaller appreciation to reduce NX by an amount equal to the rise in G. In the limit, in a one-good world, with no transport costs, an increase in G causes an equal fall in NX with no appreciation, because the law of one price holds (Purchasing Power Parity).