Mostly for non-macroeconomists.
I first learned macroeconomics in the very early 1970's in the UK. I learned that the macroeconomy was not automatically self-equilibrating, and that the government should use fiscal policy to target "full employment" (aka "potential output"). The government should loosen fiscal policy when the economy was below potential and tighten fiscal policy when the economy was above potential. We didn't pay much attention to monetary policy. And we didn't pay much attention to inflation either, at least in the models, though it was getting harder and harder to ignore in reality. I think we hoped for what we would nowadays call "Divine Coincidence": that if we were successful at targeting "full employment" then inflation would take care of itself. And if it didn't, then maybe we needed some additional policy lever, like wage and price controls.
Let's call that "Old Keynesian" fiscal policy. Many people will still think of "Keynesian" fiscal policy as something like that.
But times change.
I used to be a New Keynesian, until the recent recession. (I still am a bit of a New Keynesian, just like I still am a bit of an Old Keynesian, if the moon is full, and I hear the ancestral spirits calling me.) So it's very easy for me to pretend to be a New Keynesian.
New Keynesians believe that the macroeconomy is not automatically self-equilibrating, and that the central bank should use monetary policy (interpreted as setting a rate of interest) to target inflation. (Though some New Keynesians are recently moving away from inflation targeting.) The central bank should raise interest rates if it forecasts that inflation will otherwise rise above target, and lower interest rates if it forecasts that inflation will otherwise fall below target. Simple New Keynesian models have Divine Coincidence: if the central bank is successful at targeting inflation, then employment and output will take care of themselves.
New Keynesians (with an exception I will come to later) see fiscal policy as unneeded for macroeconomic stabilisation, and therefore useless for macroeconomic stabilisation. If the government loosens fiscal policy, and if this would increase Aggregate Demand and inflation if the central bank did nothing in response, then the central bank will do something in response.
Canada is a New Keynesian country. The New Keynesian perspective is baked into our institutions. The Bank of Canada is held accountable for keeping inflation at the 2% target. If the Bank of Canada sees a change in fiscal policy that would change AD away from where it wants it to be, and thereby change its forecast inflation away from 2%, then it will respond to fully offset that change in fiscal policy. It does not matter whether the Bank of Canada's forecast of inflation is correct or not. The only thing that matters is that the Bank of Canada thinks its forecast of inflation is correct.
We saw a clear example of monetary offset in the mid 1990's. Fiscal policy tightened a lot, the Bank of Canada loosened in response, and there was no recession and inflation stayed roughly on target. And Canada is not the US. Canada is a small open economy. The exchange rate matters as much if not a lot more than the interest rate. We dare not say the Bank of Canada "depreciates" the exchange rate to loosen monetary policy, but we dare say that the Bank of Canada does not prevent the exchange rate depreciating when monetary policy needs to loosen. (Saying it that second way doesn't antagonise the cousins so much.)
There is one exception to the New Keynesian view that fiscal policy is unneeded and therefore useless for aggregate demand stabilisation. That exception is the Zero Lower Bound on nominal interest rates, where the central bank may be either unable or unwilling to use "unconventional" monetary policies to keep forecast inflation on target and offset fiscal policy changes.
Let me assume that is the case (for the sake of argument).
According to (simple) New Keynesian models, a (perceived) permanent increase in government spending will not help the economy escape the Zero Lower Bound. The fiscal multiplier in that case is zero. But a (perceived) transitory increase in government spending will help the economy escape the ZLB. The government increases spending above normal at the ZLB, and reduces it back to normal immediately after the ZLB is past. The fiscal multiplier in that case is one. And bringing forward government spending from the future into the present will help even more. The government increases spending above normal at the ZLB, and reduces it below normal (at least temporarily) immediately after the ZLB is past. The "preponed" government spending multiplier is two. Read my crappily written old blog post if you want to maybe understand why. [Update: plus, unlike the transitory increase in government spending, preponing government spending leaves the debt/GDP ratio unchanged when it's all over, so tax rates don't need to rise to service a permanently higher debt/GDP.]
When I look back at Canadian fiscal policy over the last few years I see something that looks very close to an application of the preponed government spending multiplier idea. It was very New Keynesian fiscal policy.
If you want to tell me that the Bank of Canada raised interest rates a bit too much a bit too soon to keep inflation at the 2% target, I would agree. But it's irrelevant to whether fiscal policy tightened too much too soon.
If you want to tell me that Divine Coincidence fails, and we should stop targeting 2% inflation, and target something like NGDP instead, I would totally agree. But it's irrelevant to whether fiscal policy tightened too much too soon.
The actual existing rules of the game are that the Bank of Canada targets its internal forecast of 2% inflation. It is held accountable for doing so, and is thereby held accountable for fully offsetting the effects on aggregate demand, output, employment, and hence inflation, of any changes to fiscal policy. The Bank of Canada only gets a break at the ZLB, where fiscal policy might be asked to share accountability. And the most effective way for fiscal policy to help, according to New Keynesian theory, is to increase government spending above normal while at the ZLB, and reduce government spending below normal immediately after. Prepone government investment, in other (Indian English) words.
You wanna change the rules of the game? Great! Let's have the Bank of Canada target NGDP instead. It won't work perfectly, but it should work better than the two failed extremes, where we crossed our fingers hoping that Divine Coincidence was true. Plus, it should help keep us off the ZLB, so fiscal policy can stick to its very important micro knitting, and not get messed up by helping the Bank of Canada do its job. [Update: so the government builds bridges when drivers need help driving places, and not when the central bank needs help controlling aggregate demand.]
"Keynesian" fiscal policy in Canada today is not what I learned in Kansas school in England in 1972. Times change.