Asking the right question about fiscal policy

If you can't get a good answer to a question, it might be because you are asking the wrong question. Good answers need good questions.

Before 1975 (roughly), we used to ask questions like: "What is the effect of increasing the money supply today, when unemployment is bigger than normal?" After 1975, economists learned to stop asking questions like that. We learned instead to ask questions like "What is the effect of following a monetary policy rule in which the money supply increases whenever unemployment is bigger than normal and decreases whenever unemployment is smaller than normal?".

We stopped talking about monetary policy actions. We switched to talking about monetary policy rules. Would a countercyclical monetary policy rule be better than a rule in which the money supply was constant?

Now many economists switched to talking about monetary policy in terms of interest rates rather than the money supply. But we kept on talking about monetary policy rules, not actions. "Should monetary policy follow a Taylor Rule?" "Should monetary policy follow a rule of keeping forecast inflation at a fixed 2% target?"

Most of us stopped talking about fiscal policy altogether, in terms of using it for macroeconomic stabilisation. But over the last three years we have started talking about fiscal policy again.

But we aren't talking about fiscal policy the right way. We are asking the wrong question. We are talking about fiscal policy actions. We should be talking about fiscal policy rules. We aren't applying to fiscal policy what we learned about how to talk about monetary policy.

We keep asking questions like: "What is the effect of increasing government spending today, when unemployment is higher than normal?" Instead we should be asking "What is the effect of following a rule in which government spending increases whenever unemployment is higher than normal and decreases whenever unemployment is lower than normal?".

There are three reasons why we should talk about monetary and fiscal rules, not actions:

1. Because the effects of the current action depend on whether or not it was expected.

2. Because expectations of future actions matter too.

3. Because it forces you to think symmetrically, so you are forced to think about monetary or fiscal policy being tightened under some conditions as well as loosened under other conditions. You don't get fooled into some sort of ratchet-thinking where they can only be loosened, never tightened.

It was Lucas who taught us we had to think in terms of monetary policy rules, not actions. And when Keynesians accepted that Lucas was right on that point (even if they disagreed with him on some other points) the Old Keynesians became New Keynesians.

But, because we had forgotten about fiscal policy altogether, we forgot that what Lucas taught us about monetary policy applies equally to fiscal policy too.

So we are still talking about fiscal policy the way Old Keynesians did. So we get the slightly bizarre spectacle of New Keynesians asking an Old Keynesian question in a New Keynesian model.  And the New Keynesian has to torture his model to force it to answer the question, because the model knows that the question doesn't make any sense, and really doesn't want to answer it.

This is the Old Keynesian question that the New Keynesians are forcing their models to try to answer: "What is the effect of an increase in government spending in an economy at the Zero Lower Bound?"

This is the question a New Keynesian should be asking: "What is the effect of a fiscal policy rule in which government spending is higher than normal whenever the economy is at the ZLB, and lower than normal whenever the economy is not at the ZLB?" (And comparing that rule to an alternative policy rule in which government spending is constant at the same average level as in the first policy rule.)

Here's why it matters to get the question right:

1. The model won't explode into hyperinflation if you try to ask it "what would happen if we increased government spending for a very long time and the central bank kept the real interest rate constant so that consumption stays constant?" and you won't tie yourself in knots trying to explain that this might not happen because of a peculiar feature of the Calvo Phillips Curve that you don't really believe in yourself.

2. You won't get confused between the effects of countercyclical government spending and the effects of higher average levels of government spending.

3. You will be able to see that the expectation that government spending would increase at the ZLB [Update: or maybe even the threat that government spending would increase if the economy hit the ZLB] might help prevent the economy hitting the ZLB in the first place.

4. You will be able to see that the expectation of lower government spending and higher consumption when the economy leaves the ZLB might increase consumption during the ZLB, by consumption smoothing.

5. You will be forced to think about commitment to fiscal policy rules and monetary policy rules symmetrically, and ask yourself whether one would be any more or less credible than the other.

6. The intuition would be much clearer

7. It will be good for your soul to do the right thing.

I really wish I had thought of this point before writing my other posts on fiscal policy. But the neat thing about the blogosphere is that it lets you change your perspective very quickly.

Comments

[Bill Woolsey]

Great post.

Do you really have in mind "saving up" good government projects just in case
the economy hits the lower bound?

Forget government consumption, and think about government investment (bridges.)

You build less than optimal numbers during normal times, and then bunch some
projects when the economy is at the zero nominal bound.

[Nick Rowe]

You are very quick Bill! Thanks!

"Do you really have in mind "saving up" good government projects just in case the economy hits the lower bound?"

That's half of it. The other half of it is bringing some good government projects forward in time and doing them earlier than you would have done when the economy hits the ZLB. We should postpone government investment in good times and prepone government investment in bad times. It should be roughly symmetric.

I got the preponed bit in this old post, but missed the postponed bit because I wasn't thinking in terms of rules:

http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/03/the-preponed-government-spending-multiplier-may-exceed-one.html

[Kevin Donoghue]

"So we are still talking about fiscal policy the way Old Keynesians did."

We? You really are inviting the Tonto riposte. I'm pretty sure that a review of the literature on fiscal rules would throw up a lot of references. But I heartily agree that that's the way to go.

I'm not sure that the NK model is the best place to start. The monetary OLG model, which Benassy is fond of, looks much easier to handle at least for didactic purposes.

[Nick Rowe]

Kevin: yep. I think some of the Bank of Canada guys have been asking the right question (probably others too). But when "We" here includes Michael Woodford, not just those of us in the econoblogosphere recently, that's hardly nobody.

"The monetary OLG model, which Benassy is fond of, looks much easier to handle at least for didactic purposes."

Can you give me a quick sketch of what you mean by that sort of model?

[Kevin Donoghue]

Here's his homepage. Scroll down to numbers 49 and 51:

http://www.pse.ens.fr/benassy/artihist.html

[Nick Rowe]

Kevin: thanks for those Benassy links. He was one of my favourites in the distant past, but I had lost touch with what he's been up to.

From a quick skim: Basically, he is using Samuelson 58 and adding shocks and sticky prices/wages. And asking the fiscal question the right way.

[Lord]

Very good. Even more important than what is done at ZLB is what is done away from it as the economy is there most of the time. Fiscal policy in terms of debt reduction during normal times can help stabilize the economy as much as monetary policy can.

[Michael Carroll]

This post clears a lot up from the last few weeks. I really hope all the usual suspects take the time to respond and critique.

[Lars Christensen]

Nick, I actually think that the order of your posts where correct.

I still think that there really is no such thing as fiscal policy - in the sense that I fundamentally can't see how fiscal policy directly will be able to influence nominal spending. You make that wonderfully clear in your first post.

And finally I am not sure that I understand you right - should fiscal policy be used at ZLB? We all after all know that ZLB is not really a problem - it is never a problem to print money and fiscal policy don't have to be utilized to do that. But ok I am pretty much a traditionalist.

[Lars Christensen]

...was correct

[Steve Waldman]

This post is a gem.

[tjfxh]

See Martin Wolf today at FT. Apparently he has understood Wynne Godley who explained quite clearly the sectoral balance approach to fiscal policy (that MMT later adopted adopted).

http://blogs.ft.com/martin-wolf-exchange/2011/12/05/understanding-sectoral-balances-for-the-uk/#axzz1k6lRuMkF

[Nichol]

The reason that people have been arguing for 'independence' of central banks is that they can function quite well with a policy based on relatively simple rules. If more complicated judgement were necessary, and choices between 'winners' and 'losers' .. then that would be not good for democracy.

The conclusion is that 'rules-based policy' can be decided for the long term (democratically), to be implemented by a rather 'independent' agency like a central bank. But sometimes real choices must to be made: war or peace, big infrastructure investment, big changes in the way we do things like using energy, how we protect nature, etc. Those choices can not be made on a rules-based policy, though it may be possible to have some rules-based agency hold the purse strings that somewhat modify the speed at which those policies can get implemented. But: there will always be giant emergencies, that will provide good reasons for rules to be broken. Or situations not foreseen in the design of the rules.

I guess the conclusion is that the division of 'long-term-rules-based' vs 'democratic-choice' isn't exactly along the lines of monetary vs fiscal policy? Is that a problem? And should there be a rule to break the rule, in case you find the rule was not universally applicable?

[Scott Sumner]

Great post, but I'd take this even further. The optimal fiscal policy rules needs to incorporate the expected monetary policy reaction. If monetary policy is optimal (NGDP futures targeting), we may not need any real fiscal "policy" at all (other than the usual cost-benefit calculations.)

But here's what I find most interesting. When you frame the question this way, an argument for a monetary policy rule is really an argument that fiscal policymakers can implement a rule more effectively than monetary policymakers. Thus Krugman is essentially arguing that he trusts America's Congress to have a better understanding of optimal intertemporal policy rules than his colleague Ben Bernanke. Or maybe he thinks Congress is less corrupt than Bernanke. Meanwhile, in the real world most legislatures adopt the following policy rule: "If we have money, we spend it; when we don't, we don't." There may be a few optimal policies in places like Singapore, Sweden, and Chile, but they are few and far between.

[PrometheeFeu]

I think the problem with that approach is that it is too far from the real world in fiscal matters. We all have an understanding of what monetary policy rules look like. Almost everyone believes the Fed followed a Taylor rule for a while. Everyone understands the ECB targets 2% inflation. But what do fiscal rules look like. All I can see if a bunch of legislators who tax, borrow, spend and cut more or less willy-nilly. Sure, sometimes, they agree to put together some sort of "counter-cyclical" program in place, but it's a one-off with no guarantee it will actually keep being counter-cyclical when recovery happens, or maybe be cut-off too soon to be really counter-cyclical, or whatever. I'm not claiming central bankers are immune from political forces or that legislators are immune from the effects of the business cycle, but I'm at a loss as to what a real-world fiscal rule looks like.

[Ralph Musgrave]

Bill Woolsey, Infrastructure projects are a poor way of dealing with recessions. First, most of those projects take years to plan and get going, by which time the recession may be over. Second, to the extent that they employ those with construction skills, that’s OK, but going much beyond that involves having those without the relevant skills trying to build bridges: not a bright idea.

There is a case for boosting ALL TYPES of government spending in a recession: this is that the effect of increased government spending on employment is more CERTAIN that tax cuts. You never quite know what households will do with the extra cash: spend it or save it.

[paine]

nick

actually has a practical post not a chin scratcher bravo

the rule v action distinction is fine
in model world

we now have the autostablizers

the system ought to move to a meade -lerner full auto stablizer
algorithmic complex

for get G spending flux

use the tax/borrow and transfer systems
set them all to counter changes in household demand trade balance
and corporate net spending

design and implement a sort of automaton macronautic
fiscal budget great helmsman

yes the accumulated deficits to gdp ratio
would need a price level rate of change policy mechanism too

and trade deficits suggest a forex mechanism agreeed on by our major trading partners

but you start where you can eh ??

[Michael Carroll]

Scott, you are ignoring the obvious possibility that fiscal policy might simply be faster at eliminating unemployment.

[OGT]

I thought the Old Keynesians were for "hard Keynesianism" that balanced the budget over the business cycle. Chile has something like this put in by 'the Chicago boys,' that keys off of copper prices I believe.

Even from a cost benefit perspective preponing investment makes sense, it's just cheaper in terms of opportunity cost. Doing something like that isn't out of the realm of possibilities with something like an infrastructure bank, which like the Fed or automatic stabilizers means lessening the Congress's short term discretion

[OGT]

Scott- No, I think, it is an argument that your monetary reaction function is fundamentally flawed from a political science perspective. Whenever you talk about monetary policy in the US during the Depression you write about FDR, not the Fed Chairman. Nixon took us off the Gold Standard, not Burns. The Congress created the dual mandate, the Congress decides how often the Fed Chair testifies before them, whether there are camera's and how comfortable his chair is.

Notice the pattern?

As someone once wrote, the Supreme Court can read the election returns. So can the Fed, if the elected branches were united in pursuit of faster NGDP growth and it made it clear with fiscal policy and rhetoric there's no plausible case that the Fed would be able to stand in the way for long.

The only truly independent CB is the ECB. There is no Chuck Norris, only Jean Claude Van Damme.

[Scott Sumner]

OGT, If Congress wants monetary policy X, why would Congress produce a fiscal policy calling for Y? That makes no sense. If Congress determines monetary policy, then there is no need for fiscal policy.

Michael Carroll, In my view monetary policy is much faster, fiscal policy has to go through Congress. Both policies affect AD from the point at which they are anticipated, BTW. Another reason why it's important to think in terms of policy rules.

[Kevin Donoghue]

There's no basis for the belief that fiscal policy rules have to be slow-acting. For example, let the law say that if inflation falls below a certain threshold, the Treasury will send a voucher for $1,000 to every taxpayer, redeemable at any retail outlet if spent within 60 days. Similarly, automatic tax increases can be triggered by well-defined measures of overheating.

[Kevin Donoghue]

And Scott, just to keep you happy, the vouchers could instead be dispatched whenever NGDP is below some desired growth-path. You may not even need that furures market.

[Nick Rowe]

Thanks for the good comments everyone!

A few general observations:

Asking the right question is the first step, but the right answer to that question will depend on: the model; what the monetary policy rule is.

You could talk about what the best fiscal policy rule is for any given monetary rule; or what is the best monetary policy rule for any given fiscal policy rule; or you could look for the best pair of two rules at the same time.

And if you believe that a good monetary policy rule, if credible, can do all the stabilisation needed, then the remaining arguments for countercyclical fiscal policy are: standard micro cost-benefit analysis arguments (it makes sense to invest more/less when real interest rates are low/high); timing (fiscal can respond quicker than monetary); if, for some reason, the optimal monetary policy rule is not credible (or likely to be followed in practice), and the optimal fiscal policy rule is credible (and likely to be followed in practice).

Put it another way: if we start out with some randomly chosen monetary policy rule (e.g. "hold the real interest rate constant") I'm pretty sure that in any vaguely sensible model the best fiscal policy rule will not be "hold government spending constant".

There's also the micro question of how costly it is to postpone/prepone government projects. Obviously, some projects could be postponed or preponed at very little cost (the optimal timing is a U-shaped curve, with a rather flat bottom of the U, rather like when exactly I change my car's engine oil). But other projects will have more of a V-shaped curve, like filling up with gas.

[Nick Rowe]

Come to think of it, do I remember a recent paper by Mankiw, where he asked the right question, and also solved for the best pair of monetary and fiscal rules at the same time?

[OGT]

Scott- Normally congress doesn't produce monetary policy. But, there is no question of the institutional hierarchy, and the Fed knows this hierarchy and acts accordingly. If the elected branches want faster nominal growth and make that known, the Fed will accommodate. Fiscal policy is a commitment device in that sense.

Sometimes you use the analogy that you are driving the car and fiscal policy is like your daughter trying to grab the wheel. But that is a bad analogy. The Fed is more like a football playing high school senior son driving the car with the father in the passenger seat, technically he can easily resist his middle aged father's feeble attempt to physically grab the wheel, but he also knows he is dependent on his father.

The point being that one shouldn't take your 'reaction function' too seriously, the Fed is not independent enough to completely resist serious attempts by Congress to boost AD through fiscal policy.

The ECB, on the other hand, is willing and able to completely dictate to national governments in Europe.

[Greg Ransom]

What is a "good government project" depends on interest rates, changing relative prices, and changing regulatory and labor environments, and changing special interest rent seeking activities.

The WWPPS / BPA nuclear reactor project -- larger in size than the Apollo Moon project -- in the 1980s was a "good government project", until it wasn't, and it wasn't the moment interedt rates went up and the price of electricity went up as the full cost of nuclear power began to enter the price structure for electricity in the Pacific Northwest.

tens of billins of 1980s dollars in construction were left rusting across the state of Washington.

I've left out the story of the interest groups behind the projects, the "distributive justice/share the wealth" mandates that made a mess of construction, the highest in the nation union wages, etc.

But in the magic land of nirvana fallacy economic models this stuff is always imagined away.

[paine]

"If Congress determines monetary policy, then there is no need for fiscal policy."

no need for fiscal policy ??
how 80's of u

scotty is like a bourbon

learn nothing forget nothing

[paine]

"There's also the micro question of how costly it is to postpone/prepone government projects."
exactly why you don't alter discretionary spending you use the tax /borrow trad off and the transfer system

[Nick Rowe]

Pain: "exactly why you don't alter discretionary spending you use the tax /borrow trad[e] off and the transfer system"

Maybe. But it may also be costly to postpone/prepone taxes and transfers.

Changing VAT (GST) rates temporarily (Stephen Gordon's favourite countercyclical fiscal policy) is the most plausible in my view. Because it has both an income effect and a substitution effect. (And the substitution effect would work even if Ricardian equivalence were correct).

[paine]

nr:
vat is the answer ...exactly !!!!!
in System with a vat
also health premiums
and retirement payments

here in the states no vat
gotta be payroll taxes employee side
and state sales taxes

as to "costly "

we can look this debt /tax impact pathway
straight in the eye 24/7 365 year after year
and get it right

not as today here in the states
with a sudden mad dash to find
enough
uncle spending opportunities
to counter action such

[paine]

the transfer system is faster cleaner less targeted more neutral and popularly prefered

[Jacques René Giguère]

A $ 1000 voucher is no good in time of panic. People will sustitute it for the cash they still intended to spend.
The real effect of fiscal policy is that something is sure to happen. A bridge or a pyramid will be there. And employment is what calms people. It bring certainty. That's a potentially huge effect that you don't get in any other way , except a magnificently designed and implemented monetary policy, which is rarely the case.
And anyway, who will build the bridges?
Stephen is right that VAT variations are the best theoretical way. Except, that nowhere are VAT rates so high ( say 100%) that variations huge enough to be effective can be made. Especially in times of fear.

Independance? red herring. An independant Argentinian central bank will be led incompetently following absurd rules. A dependant Swedish bank will be managed competently according to sound rules. Whatever the legal framework, Mark Carney will lead the BoC competently and so would his successors for a foreseeable future.
The problem comes when you get a UFO: a competent organisation taken over by a madman. Whatever the rules ,salus populis suprema lex esto, (let the people salvation be your supreme law). Maybe Draghi,as an italian, remembered that old piece of Roman political wisdom that Trichet forgot( along with Scaüble and Merkel)

Maybe fiscal policy should be available only to country on a select list of functionnal countries ( Scandinavia, Canada , Australia, NZ, Netherlands) the usual suspects for good governance. Not Germany or especially the U.S.

[Russil Wvong]

Scott Sumner: "When you frame the question this way, an argument for a [fiscal] policy rule is really an argument that fiscal policymakers can implement a rule more effectively than monetary policymakers. Thus Krugman is essentially arguing that he trusts America's Congress to have a better understanding of optimal intertemporal policy rules than his colleague Ben Bernanke. Or maybe he thinks Congress is less corrupt than Bernanke."

Or perhaps Krugman and Bernanke believe that in the current situation, monetary policy alone is not sufficient to increase demand. Bernanke advocated greater fiscal stimulus in his August 2011 speech at Jackson Hole. The Globe and Mail: "With global financial markets zeroed in on his annual August speech at Jackson Hole, Wyo., Mr. Bernanke made the case Friday for more fiscal stimulus, something the White House and other supporters of short-term government spending haven’t been willing to push with force."

Scott: "Meanwhile, in the real world most legislatures adopt the following policy rule: 'If we have money, we spend it; when we don't, we don't.'"

I think there's two common rules:

1. Balance the budget over the course of the business cycle.

Typically there's some automatic counter-cyclical stabilizers in place: in a recession, tax revenues go down and unemployment payments go up. When the economy recovers, so do tax revenues, and the government can pay down the debt. (The Canadian government did this, starting in 1995.)

According to Wikipedia, Switzerland has implemented a law setting limits on government spending, adjusted to the business cycle.

2. Always balance the budget.

As I understand it, this rule is mandatory for state governments in the US, which leads to pro-cyclical policies (cutbacks during a recession, making the problem worse).

The Tea Party is pushing for a constitutional amendment requiring balanced budgets at the federal level as well.