« The Federal Deficit: A Long Term View | Main | Ontario Budget Commentary »


Feed You can follow this conversation by subscribing to the comment feed for this post.

The economists in Canada must have a lot of extra time on their hands.

Nowhere mentioned in this article is that the goal of central banks and the fiscal authorities should be to maximize real production and employment. Yes, they must also answer to financial interests when they scream too loudly.

And since when can a country/central bank with its own currency "go bankrupt"?

Ed Rector: No need to be rude.

Point #2 seems spurious. Investors don't use timelines long enough to make it important, and other risks (legislative changes, revolutions ect) probably dominate it in long term effect.


"What sort of long-term target would best allow short-term flexibility without compromising long-term commitment: a deficit target; or a debt target?"

It's a false choice. Both a deficit and debt target can be achieved by adding equity financing.

You list the advantages of a debt (or debt / GDP target) as being:
1. Thinking about targeting the level of debt/NGDP ratio forces you to adjust the deficit for inflation and real GDP growth (=NGDP growth).
2. Given a long enough time horizon, a deficit target means you will almost certainly end up either bankrupt, or communist (where the debt/NGDP ratio is so negative that the government owns everything).
3. The actual debt/NGDP ratio, relative to the target debt/NGDP ratio, keeps track of your history of declaring that circumstances are special.

Now let me list the advantages of a deficit target
1. Independence from monetary policy. The redemption value of debt is influenced by monetary policy decisions (both past and present). Thinking about targeting the level of debt/NGDP ratio also forces you to adjust the deficit for monetary policy stance.

And so hit the deficit target using equity finance. Hit the debt target using debt finance.

Ed Rector is not just rude, he is ignorant.

Monetary and fiscal policy were used in the past to target "full employment". Then in the 1970's they learned it didn't work.

A country with its own currency, and that borrows in its own currency, may (if its debt/NGDP ratio gets too large) face a choice between hyperinflation or default. Some have chosen default. Hyperinflation is just default by another name.

Hyperinflation also allows all of those who have debts denominated in the national currency to default as well, robbing their creditors as well. It is almost like they are required to default, though I suppose that nothing prevents such a debtor from paying extra to compensate their creditors.

As a practical matter, a debt/tax revenue ratio might be better than a debt/gdp ratio. Your approach treats all national income as if it belongs to the government.

Is a debt target really the equivalent of a price-level target?

A price level target has units of dollars; an inflation target has units of dollars per year.

A debt-by-GDP target has units of years; a deficit-by-GDP target is unitless.

I agree that a target based on the stock of debt is probably a more stable policy tool, but I think the better way to define that is to normalize by another stock, rather than by a flow (GDP). Perhaps divide by national net worth? That gives a public-sector leverage ratio.

I don't agree Monetary Policy was targeting full employment in the 70's. They spent most of the time trying to stomp inflation as shown by policy rate goals(which were never met). Monetary Policy trying to create "full employment"(whatever that means) is a modern post early 80's phenom.

bacon: For a long enough timeline, any rule will be broken or changed. But what we are doing is comparing two worlds: one in which the perspective is "the future debt/GDP ratio should stay where it is right now" to "the debt/GDP ratio is above/below its long run target, and we should aim to reduce/increase it".

Frank: "equity" in this case would mean something like NGDP-linked bonds. Yes they have advantages. But that still leaves open the question how many bonds we issue.

Bill: "As a practical matter, a debt/tax revenue ratio might be better than a debt/gdp ratio."

Hmmm. Interesting idea. Good point. Especially in the context of fiscal federalism, in a country like Canada (or the US), where the relative size of federal vs provincial governments may change over time. (Just like the ability of any government to raise revenue, as a percentage of NGDP, may change over time, because some things are easier to tax than others, and the tax technology may change.) I think you may be right.

Majro: "Is a debt target really the equivalent of a price-level target?"

Well, not really. But a debt target is to a deficit target what a price level target is to an inflation target. In both cases the second is the time-derivative of the first. They are parallel lines connecting 4 points.

Bert: which country are you talking about? In the 1980's the Bank of Canada said it was trying to bring down inflation (though it didn't say by how much). From 1992 it was targeting inflation.

I'm still thinking about Bill's point.

Suppose you had a (say) 100% debt/tax revenue target. And that tax revenue is initially 50% of NGDP. Now suppose temporary special circumstances increase the debt by 10%. After the special circumstances are over, you could get back to target immediately by increasing taxes by 10% too. The debt/NGDP ratio would rise to 55%, but would slowly come back down again.

That seems to make sense as roughly reasonable policy.

My favorite fiscal policy is to tax the rich and the money for government programs like infrastructure investment and programs to help poor and middle class

This help reduce deficit without increasing debt

And usevthe money for government programs that is


Your 2nd last para contradicts the rest of your article. As you rightly say in that para, “There are circumstances where it is good policy to run a large deficit, and increase the debt/NGDP ratio. There are (other) circumstances where it is good policy to run a large surplus, and reduce the debt/NGDP ratio.”

Quite right. In short, ANY specific target for the deficit or debt (in terms of a percentage of GDP) is nonsense. That is, the deficit (or surplus) should be what the needs of the day dictate. E.g. given excess unemployment, the deficit should be increased.

And that’s MMT in a nutshell.

djb: we are already taxing the rich to give transfers to the poor and pay for government expenditures. You are just repeating the boring conventional wisdom. But that says nothing about what the debt or deficit should be.

Ralph: we have monetary policy to control AD. We have fiscal policy for intertemporal allocation.


Monetary policy in the form of interest rate adjustments are DISTORTIONARY: they influence just one form of spending, loan based spending. Might as well do helicopter drops just on houses inhabited by blondes and red-heads while those with brown and black hair wait for a trickle down effect. Unless interest rate adjustments work much more quickly and predictably than fiscal, which they don't, I don't see the case for monetary.

As for fiscal policy and intertemporal allocation, I’ve never known politicians or others discuss the “optimum amount of intertemporal allocation” when discussing how much government should borrow. The only reference you normally get in IA from politicians is a load of nonsense about government debt being a “burden” on future generations.

Also, increasing public debt does not burden future generations UNLESS the population at the same time saves more for pensions. To illustrate, if I during my working life abstain from consumption and buy government debt, but I DON’T sell the debt in my retirement and my children DON’T save up and buy the debt from me, then the next generation will be no better or worse off: they inherit an asset (government bonds) plus a liability in the form of a requirement that they pay back the holder of those bonds when the bonds mature.

Ralph: Using fiscal policy to prevent interest rates changing is distortionary. Should we use fiscal policy to buy stawberries in summer to keep the price of strawberries constant between summer and winter?

But this discussion is going nowhere -- we have been here before.


"Equity" in this case would mean something like NGDP-linked bonds. Yes they have advantages. But that still leaves open the question how many bonds we issue.

As many as needed to hit the deficit target...

And really because government does not have a claim on All of NGDP, I don't think NGDP-linked bonds are a valid comparison. It would be like Boeing trying to sell bonds linked to total aircraft sales for the airline industry. Boeing sells equity claims only on it's own revenue stream.

I was thinking of equity more along the lines of direct claims on tax revenue where either
1. Total payouts (dividends) from available tax revenue are set by Congress as a direct expenditure
2. Return on investment on government equity is achieved through the tax payment mechanism

"There are (other) circumstances where it is good policy to run a large surplus, and reduce the debt/NGDP ratio."

Can you amplify when this would be the case for a government that has its own currency and borrows in that currency? I have read that, at least in US history, every period of aggressive deficit reduction has led immediately to a recession. Likewise the situation of Greece suggests that it is quite strenuous on an economy to run surpluses.

I appreciate your efforts to educate general readers such as myself.

William: if you look at the graph on Livio's recent post you will see that Canada went from a large deficit to surplus in the mid-1990's, but there was no recession, because the Bank of Canada responded to the change in fiscal policy to prevent a recession. (It won't work in Greece, because Greece does not have its own monetary policy).

The old example was that governments should run a surplus in times of peace, and a deficit in times of war. Demographics might be another example: run a surplus when the baby boomers are all in the labour force, so you can run a deficit when they are all retired and need extra medical care. Norway has a large negative government debt, saving up its oil and gas revenues for when they run out in future. Etc.

Debt is an instrument, like rates. It's crazy to target an instrument, you set the deficit/debt to be whatever is needed for targeting employment and inflation. Some people have a problem with interest rates being too low for a period of time -- lower than their numerology tells them it "should" be, or debt being too high, higher than some magical threshhold. But if the debt was really too high (or if interest rates were too low), you would see accelerating inflation. You don't target rates or debt, you target inflation and let the instruments float.

Btw, historically, the U.S. has run primary deficits across the averages of both war and peace time. I'm not aware of any nation that has historically run primary surpluses across long periods of time -- that would suggest that interest rates are higher than growth rates, which at this point we have pretty convincing data is not the case.

The interesting thing about this wave of numerology that has taken hold is that it is really ahistorical in nature. Even when on a gold standard, England ran up huge debts and had no target for its debt level. Like other nations, it was concerned with foreign gold outflows, not the level of domestic debt, which was allowed to float even as England ran primary deficits (on average) throughout both gold-standard and fiat money eras.

Also, you don't really have control over the debt/GDP ratio, so it is silly to target it. That would be like claiming control over (nominal) interest rate*inflation. If you think this quantity is too high, so you cut rates, you end up with an even higher quantity. You can't control the product of an instrument and a target. But there is no relevance to the Debt level alone (without comparison to GDP), so there is nothing to target here.


"Using fiscal policy to prevent interest rates changing is distortionary." I didn't suggest that. My point was that using interest rate changes ALONE to regulate demand is distortionary.

If there is an increase in AD, and that increase is dispersed over a WIDE RANGE of types of economic activity (which it should be), the RESULT is a change in interest rates, I've no objections.

rsj: tax rates and government spending are the instruments. The only *ultimate* target is human well-being, somehow defined.

Ralph: ultimately, the optimal response to any shock that affects AD depends on the nature of the shock. Compare two shocks: an increase in demand for money; an increase in the desire to save. Only the second shock requires a fall in r. And I fall in r will (generally) make it optimal to increase both government and private investment. But we have a division of labour, and need to assign a target to the controllers of each instrument.

I think fiscal policy targeting debt levels is a great idea. Where do I sign up?

The notion of having a "rule" for fiscal policy analogous to that used for monetary policy is very appealing. It would be interesting to see a debt target adopted for several reasons not the least of which is that it would require a (spirited) discussion on how governments ought to have zero debt (the underlying belief being that government finances should be run just like a household budget). If all we did was get beyond that debate our country would be well-served.

Three thoughts come to mind:
1. What would the target be? You say 40% which sounds reasonable but I am not sure why it sounds reasonable. Some empirical evidence to support the rate would be required and I am not sure there is much evidence (even 90% didn't seem to stifle economic growth).

2. If there were a recession then the % would shoot up simply because NGDP has fallen. This would raise cries of "cut the deficit" "run a surplus" to reduce the ratio back to target. I still recall people freaking out about health care suddenly taking up 50% of Ontario's GDP when all that had changed was GDP had taken a nose-dive. People were advocating huge cuts to health care when in fact all we needed to do was wait for GDP to recover.

3. Would the debt target be across all levels of government Federal, Provincial and Municipal? I have recently adopted the opinion that in Canada, where monetary policy is determined independently from all levels of government (including Federal), fiscal policy is best implemented at "lower" levels e.g. the Provincial level where there is a higher likelihood that the economy faces similar negative shocks.

Too bad governments don't have true balance sheets since debt as a percentage of equity/assets is one of the things people look at when determining if a company is over-leveraged.

How do we count debt? A province or city sign a long term contract with a private operator. Is that a debt? Do we substract the "value" of service rendered while we don't do it if the service is provided directly? Or do we coun,t the value of the borrowing/equity needed by the private contractor?
Debt might be a metaphysical concept. Can we use it as a policy objective or an intrument?
Musing from an humble IO guy,

Kathleen: "1. What would the target [debt/NGDP ratio] be?"

That's the hardest question, and one I have carefully ducked. AFAIK, there is little good empirical or theoretical work that gives any sort of precision in answering that question. Sometimes though, it matters less what the numerical target is than that we have a target, and what sort of target it is. (It doesn't matter what side of the road we drive on, but we must pick one.) We always end up acting as if there were *some* sort of target, and a debt/GDP target seems better than a deficit target.

2. Yep, that's the flexibility vs commitment question. Cyclical adjustment seems to give us some sort of help there, but cyclical adjustment turns out to be a very fuzzy concept itself.

"3. Would the debt target be across all levels of government Federal, Provincial and Municipal?"

I think so. But as Bill says above, a debt/tax revenue ratio seems to make more sense than a debt/NGDP ratio, especially in a federal/provincial/municipal context.

Jacques: unfunded pension liabilities (or health care liabilities for an aging population) is probably the biggy there. In principle we should count them the same way as debt that is explicitly on the books.

"2. Yep, that's the flexibility vs commitment question. Cyclical adjustment seems to give us some sort of help there, but cyclical adjustment turns out to be a very fuzzy concept itself."

So take: Country B has a lazy government, which likes the idea of "fire-and-forget" fiscal policy. It passes a law before the recession, which ensures that government spending automatically rises and taxes automatically fall whenever there's a recession. Then it goes back to sleep

and change it to:

Country B has a lazy government, which likes the idea of "fire-and-forget" fiscal policy. It passes a law before the recession, which ensures that government spending automatically rises and taxes automatically fall whenever there's a recession AND the ratio of Debt/NGDP stays within a set range of 40% to 45%. Then it goes back to sleep.

If the recession begins with the country at 40% then there may be room for debt financed deficits (since the ratio could shoot up simply because of the recession). If the recession begins with the country already at 45% then there is no room for debt financed deficits.

One piece of context for my thinking: I am operating at the Provincial level where there is no monetary authority who will offset fiscal policy. At the Provincial level there is only one policy lever. The monetary authority acts at the National level where there are, unfortunately, liely to be heterogenous shocks to the economy.

Kathleen: OK. I see where you are coming from. If a province had a *low* *long-run* debt/provincial NGDP target, this would (in principle) give them more room to use countercyclical fiscal policy without risking ending up like Greece (which is like a Eurozone province). Because when the bad times are over, it would be required to bring the debt/NGDP ratio back down again to target, whereas a deficit/NGDP target would let the debt/NGDP ratio follow a random walk, staying permanently higher (until the next shock hits).

Well, you turned my argument upside down but I like your interpretation much better. Having targets I assume is meant to constrain/enable monetary and government authorities in their actions. I was thinking a debt/NGDP target would remove fiscal policy as an option all together since the ratio could exceed the target just by virtue of the fact that a recession had occurred (i.e. no new debt issued and no deficit spending yet reduced NGDP meant the ratio exceeds target).

Taking your interpretation here is my ideal world:
1. Monetary policy targets NGDP.
2. Federal level targets a small ratio of debt to target NGDP (Debt/TNGDP). Federal debt matches national infrastructure of which we no longer have all that much.
3. Provincial level targets a larger ratio of Debt/TNGDP. Counter-cyclical fiscal policy operates primarily at the provincial level.

Thanks for helping me think about this.

Nick: even capitalized pension plan rest on assumption over future returns, bets on future GDP growth, bets on future interest rate, future risk premium and so on. Not only debt but most assets are metaphysical. It's hard to predict the future, espacially if it hasn't happened yet.

Jacques Rene: "Not only debt but most assets are metaphysical. It's hard to predict the future, espacially if it hasn't happened yet."

Yep. I'm currently on a couple of University finance committees. Looking at a snapshot of a current balance sheet is.......inadequate. Nearly all the stuff that really matters doesn't appear. For example, our biggest asset is (probably) our reputation, which isn't on the balance sheet at all. It's nearly all about making guesstimates about the future.

@Jacques Rene: "Debt might be a metaphysical concept."

And so is money - isn't it just irredeemable zero coupon debt? But yes it is medium of exchange which makes it special.

Government's liability is liability against the whole and asset for the holder. In two person economy it can be thought as a pure private debt relationship - though it's only "a half the nominal" in the sense half of the asset will be "paid" by the holder (to the holder, netting zero). 3-person economy blurs it further but still we can deduct who owns to whom and how much in the end. N-person works the same.

Maybe this is just obvious and doesn't add anything as it assumes away effects of fiscal/monetary policy. I cannot connect the dots but something along these lines often bothers me when government debt is discussed. Even if burden is not mentioned.

E.g. given the economic situation government can A) run deficit B) run balanced budget. Both worlds can be deducted to private only debt relationships. But why and how these worlds differ? If they differ that's due to policy choice maybe of a democracy. Does it mean in theory that the deficit "intervention" is a better world for the most if compared to pure capitalism? But then can we as a democracy say ex-ante that we should have a target for the deficit/debt level?

Statistical relationships can influence people's thinking about what to do. Good to have these as aids to discussions about govt policy.

I very much want to re-surface what majromax said above - he noted that the relationship should be related to Net Wealth (a stock-ratio).

My suggestion would be to build a rule or guideline as an aid for further discussion that looks at both 1) flow and 2) Net Wealth. After all, public debt is first a sign of an insufficient revenue system, but there are more judgments to make here, than just meeting the sufficiency-test - affecting economics and society now and in the future. Let us not just focus on flow over a 12-month period (GDP) and miss the point that the society may indeed be very wealthy.

What I do not want people to conclude or adopt as their rationale some thinking that actually hides its purpose of protecting New Wealth from consideration. Simple rules applied so that a wealthy society does not do anything because the current flows (GDP) are problematic, so there is nothing done counter-cyclical in nature to help the bulk of society (where individuals have little Net Worth above their need for basic savings, savings that are of interest to society and of interest to AD and flow itself; e.g., emergency money, need to improve housing situation, need to replace household durables, more education/training, retirement, others). By definition, if you allow it to be part of the guideline, the society certainly may have the resources available to justify action (financing technique to follow then).

So aids for discussion, great, just make the relationship-guideline based on both Net Wealth and GDP. Just don't make it a "rule" by which to govern, especially if it simply protects Net Wealth by leaving it out of the discussion.

Kathleen: for a national gunment with its own central bank, somme deby objective, however defined, may be appropriate. For provincial ones, who deal in "real things" including paying for human capital, debt objectives make no sense as they become a capital objective which may need to be raised or lowered according to real needs, not those of a macro-stabilisation policy.

The problem with aiming at “target” debt level is that it ignores a few aspects of reality:

1) Politics I – In an economic downturn, if the debt level was previously already “on target”, what is the likelihood political leaders would cut spending, or raise taxes to meet the new, lower, nominal target? If a government is below the target in good times, they must be initiating some new program of expenditure to meet the target. Whatever it may be, those new programs will tend to be rather sticky, if not become permanent.

2) Politics II – Even if we ignore the incentives in the political system that tilts toward more spending, this assumes that successive governments of significantly different economic philosophy would have the discipline to stick to a “target” debt level. History seems to show that the debt level can reach extremes, perhaps with political leaders conveniently redefining the “target” as they go.

3) Quality of Spending – The assumption seems to be that the “benefits” from the spending is flat across all levels. In reality, if there is any benefit, it ought to be diminishing the higher the existing government expenditure is as a percentage of NGDP.

4) Interest on the Debt – Interest rates will fluctuate and will cause a higher swing in spending objectives to meet the target debt level, the higher that target debt level to NGDP is. When interest rates rise, will the “investment” from the additional spending return a “benefit” greater than the additional interest that will have to be paid? - back to quality of spending.

5) Taxation Drag – this all ignores the additional taxation level to support the debt. Theoretically, this is limited to the interest coverage, as the underlying assumption must be that the debt never gets paid off. Still, it gets back to quality of spending – does the “benefit” outweigh taking that additional level of funds out of the private marketplace?

Given these problems, it is a wonder if we even have a political system capable of determining and living up to an “appropriate” debt target. The above suggests that, perhaps, we are better off with a target as close to zero as possible – what is a “healthy range” of debt level anyway?

What might be more useful would be to define what that “healthy range” is, given the above. But, even then, politics being what it is, it would only encourage a target at the upper end of the range, and much debate about that even being “too low”.

So, it is nice to discuss interest / inflation rate targets and deficit / debt targets as stand alone optimizations, but it is a long way from sensible policy discussion where the rubber meets the road on issues like this.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad