The recent policy debate over whether its time for interest
rates to start to rise after being at the lowest levels since the Great
Depression for nearly five years shows just how much of a policy box
governments are in when it comes to fiscal and monetary policy. Never mind the debate over whether there is a trade-off between monetary stimulus and financial stability. Raising rates to reduce the perverse
incentives on long-run economic behavior and activity comes with the risk of destabilizing
public finances by raising debt service costs.
On the one hand, the global economy has been stagnating for
five years and it can be argued still requires monetary stimulus in the form of
low interest rates and fiscal stimulus in the form of deficits and expansionary
fiscal policy. On the other hand,
low interest rates are having perverse economic effects in the form of
encouraging excessive risk taking and potential asset bubbles as well as
reducing the returns to savers.
Moreover, much like the case for home buyers, low interest rates have also been a factor in reducing the
burden to governments of acquiring debt.
Here is the policy box. Low interest rates and the size of public sector debts
and deficits are not mutually exclusive policy choices. Any move that raises interest rates
will substantially raise debt-servicing costs and actually worsen many
countries fiscal positions at least in the short run. Let’s take the example of Canada. For the 2011-12 fiscal year, total expenditures for the
federal government were 271.423 of which 31 billion dollars was debt service
costs with a net debt of 650.1 billion dollars (see Federal Fiscal Reference Tables). The debt service share of total
federal government expenditures is 11.4 percent and the “effective interest
rate” on the net debt was 4.8%. If the effective interest rate is simply one
percentage point higher – at 5.8% – then debt service costs would have been
37.7 billion dollars or 14 percent of total federal expenditure. Just one percentage point in the
effective interest rate on the entire net debt increases spending by almost
seven billion dollars without any stimulus increase on government programs or spending
on goods and services. It is
simply a transfer to bondholders.
Take the United States federal budget as a second
example. In 2012, total spending
by the US federal government was 3.537 trillion dollars of which 220 billion
was net interest for a debt service share of 6.2 percent of federal expenditure
(see US federal budget documents). Net financial debt in 2012 was 10.282
trillion dollars resulting in an “effective interest rate” of only 2.1
percent. An increase in the
effective interest rate by 1 percent to 3.1 percent raises debt service costs
to 319 billion dollars thereby raising federal spending by 99 billion dollars
and bringing the debt service share of federal spending to 9 percent.
By historic standards, these are small
increases in the effective interest rate on net debt. Larger increases would
have even greater impacts on the bottom line. A return to the rates of even the mid to late 1990s would be
nothing short of catastrophic in terms of the havoc they could wreak on public
sector budgets. Now of course,
interest rate increases are phased in gradually as debt rolls over but my point
remains the same. Despite the
perverse long-term impact of low interest rates on economic incentives when it
comes to private saving and investment behaviour as well as government debt
accumulation, the short-run impact of raising interests on government budgets
will also be harsh and substantial.
Borrowing from Nick Rowe’s recent pole analogy – policy makers are like
tightrope walkers with a balancing pole that has interest rates on the one side
and government budgets on the other. If interest rates are raised, the other side will not necessarily move up
to counterbalance without deliberate fiscal policy action – it can move down and toss you over. At the same time, raising interest rates and new austerity on the fiscal side to counteract the effect of high interest rates on government budgets – well, that is the kind of nightmare that must keep central bankers and finance
ministers on the edge of their seats. The C.D. Howe piece by Paul Masson on why interest rates should go up outlines and deals with two major objections to raising them – yet the effect on public sector budgets is not one of them.
Worthwhile Canadian Initiative 
If interest rates rise -yes that increases the cost of debt service. Interest rates for many governments like Canada where markets are unlikely to question– in the short run– the solvency of the sovereign will only rise as the pace of economic activity quickens and resources are more fully employed. This of course will mean a revenue windfall for government–increasing tax revenues-and reduce spending on various social insurance programs. Additionally measures take to stimulate growth will also be wound down. You therefore can’t make an apriori assumption that governments face an interest rate risk of any great magnitude without explaining why rates will go up but the economy remain below it’s trend growth rate. For those who manage government debt no doubt a key question right now is whether to take advantage of these rates and try to tilt the balance of outstanding debt toward longer term bonds.
Livio, I don’t get your reasoning here.
Raising rates to reduce the perverse incentives on long-run economic behavior and activity comes with the risk of destabilizing public finances by raising debt service costs.
We’re in a Liquidity Trap.
On the one hand, the global economy has been stagnating for five years and it can be argued still requires monetary stimulus in the form of low interest rates and fiscal stimulus in the form of deficits and expansionary fiscal policy.
It’s been stagnating because we’re in a Liquidity Trap. There is an excess of Savings over Investment, which makes the clearing rate of interest negative. Due to ZLB, this cannot be realized easily, leading to the Trap condition. We haven’t faced this particular condition in Canada since the 1930’s, but it’s back, which is why this Recession is the Lesser Depression and it has every appearance of being the Great Depression’s kid brother.
If the effective interest rate is simply one percentage point higher – at 5.8% – then debt service costs would have been 37.7 billion dollars or 14 percent of total federal expenditure. Just one percentage point in the effective interest rate on the entire net debt increases spending by almost seven billion dollars without any stimulus increase on government programs or spending on goods and services. It is simply a transfer to bondholders.
Wow, there is a huge unstated assumption there that income is constant. That’s untenable, in my opinion. Right now we are in a classic Keynesian Liquidity Trap, with inflation low and falling. Nowhere is there any sign of runaway inflation. Any interest rate increase at this point would be from increased economic activity, which leads to higher tax receipts and higher ability to carry debt. What’s the problem, that’s exactly what we want!
I’m sorry, without consideration of the income side of the equation I can’t see that there’s a problem. ISTM that your argument falls on that.
I have a cynical view of this debate – some people want higher rates because they produce an immediate profit boost for them. The arguments in favor are just insincere blather.
As an example, Schwab has published multiple editorials in the WSJ calling for higher rates. One of Schwab’s major profit streams is interest on customer cash, so near-zero rates directly reduce Schwab’s profits. Of course they can’t say “gimme more profit!” in their editorials – they need to come up with something more persuasive.
Livio,
“encouraging excessive risk taking and potential asset bubbles as well as reducing the returns to savers.”
Reducing returns to savers is designed to get them to take investment risk. The whole point is to encourage risk-taking! If risk-taking was excessive there wouldn’t be over 7% unemployment.
Where’s the evidence of a bubble or any sort of trade off whatsoever?
“A return to the rates of even the mid to late 1990s would be nothing short of catastrophic in terms of the havoc they could wreak on public sector budgets.”
Except, of course, that NGDP growth was correspondingly higher back then. In fact, if you look back over the past 50 years of North American economic history, apart from one or two years in the early eighties, there is not a single period where the policy rate exceeded NGDP growth. There is just no theoretical or empirical justification for postulating interest service being a significant contributor to the debt/GDP ratio.
lol, if rates returned to late 90’s levels, there would be no public debt to service. You simply aren’t getting that.
Matter fact fact, those nominal rates were in fact nominal.
I don’t understand the need for such incendiary language — “time bomb”. The whole article relies on a lot of non-sensical claims that serve no purpose other than move the scare story forward:
1) The CB is lying when it says it will keep rates low even once the recovery starts in order to try to generate excess future inflation
2) That government revenues are independent of changes to NGDP, so we can play a game of pretend in which an increase in interest payments necessarily must cause an increase in the revenue share of interest payments.
3) That the CB will, inexplicably, raise interest rates above the NGDP growth rate in an environment of depressed demand
4) That household savings demands will suddenly and rapidly decrease
I have to think that given the irrelevance of huge deficits run by nations with their own currency, the deficit-scolds are forced to invent their own reality in which ticking time bombs are just around the corner.
Deficit-Scolds, next it’ll be Very Serious People. Paul Krugman’s articles are having a worldwide effect in a way he didn’t intend, I see. 😉
Ha! You are right!
Vladimir is right: this is one huge non-problem.
Assuming a “consolidated and coordinated behaviour” of the US Fed + federal gov I can see this being a non-issue (so others argue). What about Canada, with very large provincial debts? Is it in a more precarious position?
Livio,
“Here is the policy box. Low interest rates and the size of public sector debts and deficits are not mutually exclusive policy choices.”
They can be IF:
1. The fiscal authority switches from coupon bonds to accrual (non-coupon) bonds
2. The fiscal authority does not limit the duration of the securities that it sells.
The limiting factor on a the size of the public debt is the annual interest payments that must come from tax revenue.
For coupon securities, the amount of annual interest paid is:
Interest Expense = Total Debt * Annual Interest Rate
For accrual securities, the amount of annual interest paid is:
Interest Expense = ( Total Debt / Average Duration ) * ( 1 + Annual Interest Rate ) ^ Average Duration
For accrual securities the Annual Interest Rate is some Term Premium times the Short Term Interest Rate
Interest Expense = ( Total Debt / Average Duration ) * ( 1 + Term Premium * Short Term Interest Rate ) ^ Average Duration
With accrual securities, extending duration can lower the annual amount of interest paid.
It seems to me that if the effective interest rate is currently 4.8%, while 1 years bonds are yielding 1.05% (30 year 2.53%), then that rate is currently falling as debt rolls over. Further more, even if the Bank of Canada raises its overnight rate by the 1% you use in your analysis (to 2%) those rates would continue to fall.
This may be considered a good article in regular business newspaper promoting the “good thing” of not being scared of low interest rate using flawed language that readers are used to. But it is still wrong. Doubly so that it was posted on a blog where reasons why it is wrong are spelled every other day.
If anything one should learn from market monetarist blogs that interest rates are unreliable. One cannot look at solely at nominal interest rate and claim that they are low or high. Interest rates should be always seen in context of the monetary policy stance, which can be observed from thing like inflation and nominal income growth. Are inflation and NGDP growth “low” (lagging behind trend)? Then it is a very bad idea to have CB increasing interest rates. Not because it has this or that impact on fiscal policy. But because it will make already tight monetary policy tighter – which will mean that sometime in the future CB will have to ease it (by lowering interest rates) or accept higher unemployment and depression. And because people do not like unemployment and depressions, it is more likely that we will see the former – look how much good did ECB interest rate hike do in 2011. The economic condition worsened so much that ECB did not only revert the interest rate hike but cut it a little bit more. And looking at NGDP and inflation in Eurozone it is by far not enough. ECB baby steps toward easier monetary policy can be seen as passive tightening if one looks at relevant making macroeconomic indicators like inflation and NGDP. This is why MM sometimes say that low interest rates may be a result of past tight money.
I find it amusing that economists are faulted when they fail to anticipate future economic situations and told they are presenting non-problems when they lay out what future economic issues may arise. I still think that increases in the interest rate on government debt has the potential to become a more important public finance problem for governments given the mass of debt that has been accumulated to date as a result of fiscal stimulus as well as the low cost of acquiring more debt due to low rates. I can recall the early 1980s and the early 1990s when debt service costs soared due to rising interest rates and deficits. The economic and policy conditions of that period that drove the high rates were of course very different then and not being replicated at present. But given the level of debt that has been incurred, if rates rise, debt service costs will rise. In Canada, the combined level of federal and provincial net debt has been estimated at about 1.2 trillion dollars – with GDP at about 1.7 trillion. Federal debt service costs in 2011-12 were 31 billion dollars. Compare that to what the federal government spent on Old Age Security (38 billion), Employment Insurance (17 billion dollars) or the Health and Social Transfer (38 billion). In Ontario, the amount of debt service (10.6 billion dollars) exceeds what it spends on colleges and universities (7.6 billion) or transportation (2.8 billion). If debt service is starting to rival government expenditure categories on programs at “low”interest rates, then we should be concerned about interest rates rising and their impact on government finances.
There are two stories one hears a lot nowadays. In the first story governments spend too much on wasteful projects crowding out private spending and investment. In this story easing monetary policy will achieve nothing – besides higher inflation. Many proponents of this theory actually shout loudly that money is already too easy despite low inflation and low NGDP growth. They explaining it either by pretending that this question was never asked, by pretending that low interest rates means easy money or by bad data, or by supporting volume of flawed evidence etc. But this is basically a story of supply-side recession.
Then there is another story where economy experienced a large nominal shock to incomes and is it is now stuck in a bad equilibrium where some sticky prices (especially wages) will not clear leaving us with unemployment and unused production capacity. MM claim that the nominal shock was brought upon us by mismanagement of central banks back in 2007 and continuation of bad monetary policy. The best solution for this is to jump-start demand prefferably by “unorthodox” monetary policies – such as creating the money printing path that is consistent with economy with sufficient demand (yeah how “unorthodox” for central bank to print money, right)?
Then some people – like Scott Sumner – say that we may have both. We may have experienced shifts in both – aggregate demand and aggregate supply. But the first one is way easier and quicker to solve by printing money. Then we may think about the supply side problem.
So where comes your story into this? It comes nowhere. Because you focused on the wrong problem. Because “central bank rising interest rate” is insufficient information. Because there are two ways how CB can “increase interest rates”. The first one is that it will tighten the monetary policy which means deflation and unemployment. But sure, crazy enough central bank can certainly do this, especially if noone important seems to understand that it was CB that caused the trouble (like ECB in 2011). Or there is a way where Central Bank may achieve interest rates increase by printing money – enough to bring equilibrium to the labor market. And then at the point where the money velocity picks up, when inflation will threaten to push NGDP growth way too much, then CB can with clear conscience increase interest rate – without even thinking about fiscal policy.
The moral of the story is that it is just wrong to shackle monetary policy with fiscal considerations. It is never good for a nation to have a bad aggregate demand policy. There is no situation where bad aggregate demand policy is a price to pay for some particular thing. Not for supposed fiscal problems. Or for soothing egoes of hedge fund managers who made bets against CB. Or to “protect European savers”. Or to prevent runaway lending by banks. Or whatever. The price is just too high and many times you end up with the exact opposite anyways.
I’m quite open to Aggregate Supply problems, I acknowledge they exist. But they don’t exist right now, not on the order of our Aggregate Demand problems.
Respectfully, Livio, your post appears to embody a “Anchoring Effect” in which economic problems are compared to the 1970’s/80’s in a Stagflation framework. It’s a common human bias to make such analyses, but the problem in economics is that the present crisis is not the 1970’s, it’s the 1930’s. There are far more similarities with the Great Depression than there are with Stagflation, which is why Paul Krugman and his traditional Keynesianism has done so well as an theoretical framework.
Starting in 1981 (when we were as far from a Keynesian paradigm as we were ever going to get) we managed to blow ourselves back to 1935 in the space of 30 years.
I suppose we all have our biases and mine have been shaped by coming of age during the recessions of the early 1980s and early 1990s. True, our current situation is not the 1970s but it is also not really the 1930s either. The present situation is quite unique in that it was marked by Depression-like economic forces but many of the worst effects were partly mitigated by successful and quick deployment of monetary policy and fiscal stimulus – unlike the 1930s. As well, interest rates were already low prior to 2008-09 as a result of dealing with the tech crash. Also, international trade has not dried up like the 1930s. Unemployment in North America during the 1930s was over 20%. At present, there are some European countries that are marked by Depression-like conditions (eg. Spain) but others that are not – such as Germany. Canada itself has escaped relatively unscathed. However, one of the side effects of expansionary monetary policy and low interest rates has been low debt servicing costs which have made it easier to acquire more debt. Its not wrong to be concerned that if interest rates rise, debt service costs will squeeze out other government spending. It is also not wrong to assume that monetary policy can influence government fiscal situations and vice versa. Monetary and fiscal forces and policy do not operate in watertight compartments.
Canada itself has escaped relatively unscathed.
I do not agree. The voluminous lines of young graduates trying to get a job (and the recurring articles this fact generates in newspapers) says that we were, and still are, very much scathed.
True, our current situation is not the 1970s but it is also not really the 1930s either.
Blunting the knife’s edge does not diminish the fact that it is still a knife and still caused a wound.
Monetary and fiscal forces and policy do not operate in watertight compartments.
And yet I am mystified that you have not considered the increase in tax receipts that higher inflation,, high employment and higher NGDP would entail. This is exactly the result that Keynesian policies are supposed to deliver via the Fiscal Multiplier. Why are we concerned about our own success?
I graduated high school during a recession and from graduate school into another. Many in my cohort did not find decent work right away and their lifetime earnjngs have suffered so I can sympathize with current young graduates. It is worse now I suppose given that those recessions had relatively sharp recoveries whereas the current downturn has been protracted. Still, unemployment peaked at much higher rates during those recessions.
Right, the early 1980’s recessions were “planned” as Dr. Torben Drewes, a prof at Trent memorably said in a talk he gave that I attended. The Bank of Canada and the Federal Reserve deliberately raised interest rates to eye-watering levels to suppress inflation. It was a deliberate lowering of demand. Once inflation was suppressed, the recovery commenced, as one would expect if demand was artificially, mechanically and deliberately lowered.
But today we have an “unplanned” recession, demand has fallen short of its own accord; it has to be stoked. That was the world Keynes talked about. As Paul Krugman so effectively said, the map is not the territory; the problems, causes and effects of the 1980’s recessions don’t apply today to a much different world. It’s apples and oranges. So we need a different map and coincidentally the map Keynes drew is the map we need.
Livio,
The problem here is that over the last 100 years of interest rate history, you are not going to find a single example of the overnight rate being consistently higher than the NGDP growth rate. Here, I define “consistently” to be for a period of 4 years or more. The call money rate averages significantly below the NGDP growth rate. When it rises above this rate, it is because 1) there is a financial panic in a gold standard environment, or 2) in a gold standard or pegged regime, nations are trying to engineer internal devaluation.
So you really need to make a case as to why you think Canada, which does not have a currency peg, would try to engineer internal devaluation for a prolonged period of time.
It just doesn’t make any sense, you have no theoretical model to back up your fear, and the data doesn’t support you that interest burdens are something to be afraid of.
Sure, you can find lots of handwringing and political fears of interest burdens but there is no real solvency or accounting fear. None.
Nor was there any real excess interest burden in the early 80s-90s (when rates were falling, not rising). The only burden is political and self-imposed.
This blog is scare mongering pure and simple — and is pretty damn irresponsible.
I think I’ve laid out the evidence for any concerns I have. I am neither irresponsible nor a scare monger. If and when interest rates rise, they will raise debt service costs – ceteris paribus.
If and when interest rates rise, they will raise debt service costs – ceteris paribus.
That’s a Macro 1000 error, as Nick would say. Or Macro 1A03, in my alma mater’s course terminology. Which I have repeatedly pointed out, as have others.
Governments tax a proportion of NGDP, so the interest burden does not go up when interest rates go up, provided that interest rates stay in line with NGDP. That is the key issue at stake here, and we have data tracking both the call money rate and NGDP growth rate for many nations. Interest rates rise above NGDP growth rates when nations are trying to create deflation so that they internally devalue. For example in the gold standard era when nations needed to defend a gold peg even at the expense of domestic output.
I think it highly unlikely that we will see a situation like that again. It is not a “ticking time bomb”. There is always the possibility that the future may hold fantastically bad policy decisions, but those decisions are not created by current (good) policy decisions, nor made inevitable by the present deficit stance.
Monetary policy is dead and won’t be coming back unless labor share increases. So in the meantime, monetary policy has to “do like the Romans, when in Rome”. Monetary has to adjust to the conditions in which it is operating and the conditions have shifted after being stable for decades. If monetary policy continues to operate under past conditions that don’t exist now, it is making an error. The new conditions warrant a change in calculation. The central bank interest rate in the US should have been rising slowly over the past 2 years. It is now way behind and will not soon catch up.
Here is the explanation of the model behind the thinking…
http://effectivedemand.typepad.com/ed/2013/05/monetary-policy-of-effective-demand-the-basics.html
RSJ,
“The problem here is that over the last 100 years of interest rate history, you are not going to find a single example of the overnight rate being consistently higher than the NGDP growth rate. Here, I define “consistently” to be for a period of 4 years or more.”
The cost of debt service is a function of the nominal interest rate AND the total amount of debt outstanding. Governments typically do not borrow at the overnight rate, but instead across a spectrum of maturities. Look instead at growth rate of debt + average interest rate and compare it to the nominal GDP growth rate.
A few points:
1) The Paul Masson piece does mention the effects on public sector budgets. But he portrays it as an advantage (or rather as a disadvantage of low interests). In his view, low interest rates are an incentive for fiscal irresponsibility. Increasing short term rates, would force the provinces to make hard choices (which he seems to favor).
2) Obviously, this would mean that the economy would feel both fiscal and monetary contraction if rates increase (c.f. Ireland, or perhaps Spain). It could get very ugly.
3) If we suppose that rates (both short and long-term) increase only once nominal GDP (and thus nominal government receipts) growth has reached a fast pace, then this extra revenue will help cover the higher interest expenditures. (Say government sees revenue growth at 5% instead of 3%, this would cover almost a full percentage point increase in the “effective interest rate” prof. Di Matteo gave in example. By the way, this means that half our present debt should be rolled over at 6.8%. That is quite an interest hike!)
5) In that same scenario where NGDP growth is strong, then we would expect some government expenditures related to social services (e.g. unemployment insurance) and other recession-related spending to be quite lower. By how much? I do not know. This lower mandatory spending combined with higher revenue growth could help keep public sector deficits under control.
6) If NGDP growth is strong, we would expect private sector investments to be strong (which is of course a cause of long-term interest increases). Significant government spending in these circumstances would be expected to crowd out investment and thus be quite counter-productive. In this environment, budget cuts could be economically well justified.
The cost of debt service is a function of the nominal interest rate AND the total amount of debt outstanding.
Yes, but government revenue grows with NGDP, so independent of the size of the debt, if the weighted average interest rate is less than the growth rate of GDP, then the Debt/GDP ratio will stabilize without the need to run any primary surpluses. You can run primary deficits every year and have the debt/GDP ratio be bounded because the numerator is growing at a smaller rate than the denominator.
Significant government spending in these circumstances would be expected to crowd out investment and thus be quite counter-productive.
That is a red herring. The reason why we have such large deficits now is because of countercyclical spending which will decrease when aggregate demand is robust. The issue being debated is whether the “overhang” of the present level of countercyclical spending\ will prove to be a burden in future years, and there is simply no evidence for this argument.
A crude but useful measure of debt-service costs would be DSC = (i-n)(NDEBT/NGDP)
where i is average nominal interest rate on govt debt, n is growth rate of nominal GDP, and NDEBT/NGDP is the nominal debt/GDP ratio.
This measures what fraction of GDP is needed to service the debt to keep the debt/GDP ratio constant over time.
It does not matter for this question if n > i (so DSC is negative). You get exactly the same effect if i rises relative to n. A reduction in a benefit is the same as an increase in a cost. Both are bad things.
Currently, for Canada, i on new debt is abnormally low relative to n (because world interest rates are abnormally low, and the Canadian economy is (or at least has been) recovering, so n is not that low). But it is plausible to predict that i will rise relative to n in Canada, at least on new debt. (That’s an important caveat, since a lot of Canadian debt is old debt, issued when interest rates were higher than they are now, which explains that 4.8% number, and so there will be a temporary offsetting effect coming from the vintage effect as old debt gets rolled over at lower rates.)
If i-n is going to increase in future, and it probably will, then Livio is right that we need to take that into account. But I would be more worried about Japan than Canada. First because Japan has a higher debt/GDP ratio. Second, because Japan has been in recession with low interest rates for much longer, so there will be less of a “vintage effect” going the other way in Japan. Lets just hope that Abenomics works, and works quickly.
Nick,
If the NGDP growth rate is greater than the weighted average interest service, then we completely ignore Debt/GDP. We never raise taxes to “pay down” debt, and we never cut any spending that we would have made without debt. We do whatever we would have done with zero debt, just rolling the debt over.
Therefore there is no cost. It does not “require” any portion of GDP to service the debt, since the debt services itself via refinancing operations.
And this is the world that we — by and large — live in now and have lived in the past.
I say “by and large” because there can be periods of manias in which we do choose to suffer through austerity not because we need to, but because we want this number to be lower more quickly. The pleasure of seeing this ratio fall, or the fear of seeing this number go up, causes a lot of people to believe that austerity is necessary. But these are just scare tactics, and often attempts to reduce this number cause the ratio to go up, as GDP declines.
Here is a graph of the primary budget surplus/deficit since 1947 as a percentage of NGDP (the time we started “repaying” WW2).
http://research.stlouisfed.org/fred2/graph/?g=iBV
You will note that pretty much every year, we ran a primary budget deficit, and yet the debt/GDP ratio today is about the same as it was in 1947. We ran an average primary deficit of -1.92% of GDP over that time period.
rsj: see where I said in my comment: “It does not matter for this question if n > i (so DSC is negative). You get exactly the same effect if i rises relative to n. A reduction in a benefit is the same as an increase in a cost. Both are bad things.”
(And who is “we” in your last two lines? On this blog, “we” means “Canada” 😉 )
“we” means “Canada”
Heh. Point taken. Well, point me to a good website disseminating free Canadian data, and I will run the numbers for you, too. Btw, I think Livio made an error in the post because I strongly suspect he is using ‘interest payments’ when he should be using ‘net interest payments’. But, I can’t prove this as I don’t know where you Canadians store your data.
A reduction in a benefit is the same as an increase in a cost.
No, in that case we should run deficits up until the point where i = n. As long as i < n, then we are leaving a free lunch on the table because we are effectively borrowing money that never needs to be paid back.
OK, OECD to the rescue!
http://www.oecd.org/eco/outlook/economicoutlookannextables.htm
For Canada, we have general government net interest payments of 0.4% of GDP in 2012, or about 7.3 Billion (assuming a 2012 GDP of 1.8 Trillion) — not 31 Billion. You cannot include principle repayment when determining interest expense, as this confuses stocks with flows and in any case will vary with the maturity structure of the debt.
“I am neither irresponsible nor a scare monger.”
When you put ‘time bomb’ in the title of a post, I’d have to disagree.
Over the past 62 years US nominal GDP increased by a factor of 58. Over the same period, if the US government had merely let the debt compound at the t-bill rate and never paid off any interest, the nominal amount of debt would have increased by a factor of 17. I.e. debt/GDP would have decreased by more than a factor of 3. This chart should also help clarify the picture:
Nick,
“A reduction in a benefit is the same as an increase in a cost”
I’m with rsj, here (what else is new). If the economy is dynamically inefficient (and there are good theoretical reasons as well as empirical evidence that it is – see above graph), then there’s a free lunch to be had in increasing unfunded public expenditure. This is the market’s way of telling us that we can increase the intertemporal consumption optimum via increased investment in public goods.
And while we are setting the facts straight, the effective interest rate for canadian government net general debt is 1.26%, not 4.8% (that’s another conflation of including principle repayment as an interest expense). Moreover, an increase of 1% in the effective interest rate would cause the interest expense to go up to 13.1 B from 7.3 B, nowhere near 14% of general expenditures, and of course the cited estimates for the U.S. suffer from similar errors.
You just cannot include principle repayment, which is a financing operation on stocks, as a part or fraction of general expenditures, which is a flow.
If you do, you will massively overstate expenditures as well as the debt service share of expenditures.
As a simple example, suppose all debt was financed with a 1 year bond, so principle repayment included the entire debt outstanding each year. Now you’ve ballooned expenditures and suddenly “debt service” is by far the largest share of expenditures. Good for scare-mongering, bad for estimating debt burdens.
Nick: “If i-n is going to increase in future, and it probably will, then Livio is right that we need to take that into account”
What do you mean by “need to take it into account”? Should we take it into account when deciding about premature tightening right now that would lead us to higher unemployment and reduction of utilization of production capacity anyways? Or should we worry about increasing ‘i-n’ when deciding about monetary easing right now in order to end recession, increase employment and closing output gap?
Is there actually some literature when aggregate demand management is analyzed from welfare point of view that takes into account all these particular things that should be taken into account? Should we for instance take into account profit/loss of central bank on OMO. Or the level of bank lending tied to particular aggregate demand policy? Or impact of monetary policy on savers? Or utility loss caused to goldbugs that would preffer to pay with gold coins?
I personally think that you are just trying to salvage an untenable position. If there is a lesson from monetary history that I am aware of it is that aggregate demand policy should not be subordinate to these particular interests, because the costs are high and there are always better ways to achieve particular goals anyways. But maybe I am wrong, then please can you point me to a macro-literature that seriously analyzes when fiscal considerations should be taken into account when forming monetary policy, preferably something that has at least remote relation to the condition that are similar to those in today Canada?
RSJ,
“Yes, but government revenue grows with NGDP, so independent of the size of the debt, if the weighted average interest rate is less than the growth rate of GDP, then the Debt/GDP ratio will stabilize without the need to run any primary surpluses. You can run primary deficits every year and have the debt/GDP ratio be bounded because the numerator is growing at a smaller rate than the denominator.”
TR = Effective Tax Rate (Taxes as percentage of NGDP)
NGDP = Nominal GDP
INT = Weighted average interest rate across all maturities of government debt
D = Total Debt Outstanding
dNDGP/dt = Growth rate of NGDP
dD/dt = Change in growth rate of D
The only condition that the government must satisfy:
TR * NDGP > INT * D
After that the government is borrowing from one person to make the interest payments to another (Ponzi finance)
TR * NGDP = k * INT * D : k is a number larger than one
TR = k * INT * D / NGDP
Assume a constant tax rate and weighted average interest rate:
dNGDP/dt * TR = k * dD/dt * INT + dk/dt * D * INT
For dk/dt = 0 ( constant debt service as a percentage of tax revenue )
dNGDP/dt * TR = k * dD/dt * INT
TR / k = dD/dt * INT / ( dNGDP/dt )
In the U. S. that has not been the case because a lot of the increase in the federal debt has been a debt swap from financial sector to public sector. Federal debt growth rate has been much larger than nominal GDP growth rate because of debt swap. If the nominal GDP growth rate is larger than the product of the debt growth rate and the weighted average interest rate, your percentage of tax revenue dedicated to debt service – the fraction 1/k – will get smaller.
“You can run primary deficits every year and have the debt/GDP ratio be bounded because the numerator is growing at a smaller rate than the denominator.”
The debt/GDP ratio is not a binding concern of governments. That is the point I was trying to make. The only binding concern is the amount of tax revenue available to make the interest payments.
I have used total public debt charges (which means I have included principle repayment) because those are obligations that a government must meet and therefore represent a burden on current resources. Governments do include interest on the principle when it comes to paying the bills. Frank is right in stating that ultimately “the only binding concern is the amount of tax revenue available to make interest payments” but it still means if the payments get large enough, spending is moved to debt service and away from other government programs. An example from Canadian fiscal history – in 1971/72, federal program spending as a share of total federal spending was approximately where is is now with debt service at 11 percent of total spending). By 1981/82, the program expenditure share was 82 percent and debt service share was at 18 percent. By 1990/91,the program expenditure share was 71 percent and debt service was taking up 29 percent of the federal budget. The result was a major restraint program in the 1990s that reduced federal transfers to health and education and led to reductions in services. Fortunately, the economic boom after the mid 1990s to 2007 raised tax revenues and reduced the burden of debt and deficits. Rising GDP reduced debt to GDP ratios. Moreover, declining interest rates yielded a fiscal dividend that allowed governments to reduce their debt service payments and put money into both programs and tax reduction. We are now in 2013. Unlike the period after WWII which saw high debt to GDP followed by a robust economic boom and low interest rates or the mid 1990s where high debt to GDP was again followed by a boom period and falling interest rates there does not seem to be a boom on the horizon after five years of low growth, low rates and substantial debt accumulation. If interest rates start to rise, debt service costs will rise and without strong GDP growth those payments will impose additional fiscal burden on government. The large mass of accumulated debt is tinder awaiting the match of higher interest rate. Higher nominal GDP growth is the fire extinguisher. If this is scare mongering, so be it.
Livio,
I have used total public debt charges (which means I have included principle repayment) because those are obligations that a government must meet and therefore represent a burden on current resources.
No, rolling over of principle is not a burden on current resources (a flow). It is a refinancing operation.
You cannot subtract principle payment from current income anymore than you can add borrowing to income.
Moreover you cannot say count principle repayment as “transfer to bondholders” — only the interest income is a transfer to bondholder.
Also, you cannot divide total debt by the principle repayment to get an “effective interest rate” — your interest rate is 3 times too high!
Finally. you cannot assume that a 1% increase in rates will cause both the interest and principle payments to rise — that is double counting!
Yes, governments typically use cash-flow based accounting, but that is a classification that you need to understand and not confuse with more standard accounting terms used by the private sector.
RSJ,
IF you’re going to cite the OECD net debt interest figures for Canada – which are facially ridiculous as measure of Canada’s government interest obligations – to refute Livio’s point, you might want to read the accompanying notes a bit more closely. From the accompanying OECD notes:
“Net property income is used as a proxy for net interest payments, which are not separately available. For Canada this includes significant levels of royalties from the exploitation on natural resources, which may bias corresponding measures of primary balances compared with other countries.”
I don’t know about you, but that to my mind that renders the OECD numbers (at least for Canada – given the not insignificant role that royalty revenues play in provincial finances) more or less useless in this discussion.
And if you want to check Livio’s numbers, you might try looking at our public accounts (at least for the federal government). The $31 billion public debt charge comes from that. http://www.tpsgc-pwgsc.gc.ca/recgen/cpc-pac/index-eng.html
RSJ: The $31 billion does NOT include capital repayment, it’s a pure interest calculation. In addition to the $31 billion in interest charges the federal governm also had to finance $41 billion in maturing debt in 2011-12.
Also, using the treasury bill rate to measure government debt cost is seriously misleading, seeing as the bulk of the government’s marketable securities is made up of longer term debt which is significantly more expensive than the T-bills.
Bob,
I was using table 31, “General government net debt interest payments”. As you know, this page has 62 tables, with many footnotes. The footnote you cite is not for this table and has nothing to do with the calculation of net interest payments. You need to make a better case than this
The reason why you cannot site a “public debt charge” as interest expense is that it includes principle repayment and so is manifestly not an interest expense. Because it is not an interest expense, it cannot be used to calculate the effective interest rate, or the amount transferred to bondholders, which is the point of Livio’s post. It has meaning — just not for this discussion.
Seriously, it makes a big difference if the government’s interest burden is 1.26% or 4.8%, no? For example, the former is less than inflation, therefore in one case the government gets to borrow at negative rates, and in the other case the government does not. When deciding whether or not to borrow, this makes a difference no? The amount “transferred to bondholders” in one case is positive and another negative in real terms. That makes a difference.
Bob,
From Statscan:
“H29. Public debt charges are gross interest and carrying charges on public debt.”
http://www.statcan.gc.ca/pub/11-516-x/sectionh/4057752-eng.htm
Bob,
Even if you understand the difference between accounting interest and principal, accrual accounting numbers are completely irrelevant to understanding the economics of the debt burden. If a government locks in all its borrowing in 30 year bonds at 1%, and inflation and the 30 year rate then rise to 10%+, then the debt, as structured, will be easy to carry. If on the other hand, they lock in 30yr debt at 10% and rates then drop to 3% (see the Great Moderation), then carry costs will appear high. These are just examples of good or bad trading and the effect of carrying liabilities at book, rather than market value. In both cases, what is really happening is a huge change in the market value of debt, which is then carried forward at the spot short rate.
If governments keep the average duration of their liabilities relatively short (e.g. less than 5 years), then the carrying costs over the period of a business cycle will tend to average around the short rate plus a bit of term premium. Looking at accounting numbers which reflect some random weighted average of coupon rates incurred over the past 30 years is completely pointless.
If you want to understand the expected economic carry costs of additional debt, you need to understand the relationship between nominal growth and the equilibrium short rate. The empirical evidence (see my graph above) is very strong here and gives no support at all to the existence of any burden whatsoever. The future reality may be different (hey, maybe this time is different) and it’s worth thinking about. But you are not going to establish anything by comparing accounting cash flows that were baked in over the last 30 years with current economic size and growth numbers.
RSJ: “I was using table 31, “General government net debt interest payments”. As you know, this page has 62 tables, with many footnotes. The footnote you cite is not for this table and has nothing to do with the calculation of net interest payments. You need to make a better case than this.”
Naughty, naughty, RSJ. You really should have read the “Notes to the Economic Outlook Annex Tables”, specifically, the notes to Table 31, and more specifically, the country specific notes under that section from which that excerpt is drawn (http://www.oecd.org/eco/outlook/notestotheeconomicoutlookannextables.htm). Is it too much to expect you to have looked at the notes to the table you were referring to?
RSJ: “H29. Public debt charges are gross interest and carrying charges on public debt.”
Sorry, in what universe does “gross interest and carrying charges” include principal repayment? I don’t know what you think carrying charges mean, but in the rest of the world, they are a cost of financing, not principal repayment. In any event, the suggestion that “public debt charges” includes principal repayment, apart from being absurd, is refuted by the Public Accounts numbers for 2011-12, in which “public debt charges” were $31 billion, and $41 billion worth of debt matured (since on your interpretation the former must always be larger than the latter – at least for any non-zero government interest rate).
Bob,
Oops and now I am wrong — a good definition was on Fred (ha!)
http://research.stlouisfed.org/fred2/series/GGGDTACAA188N
“Gross debt consists of all liabilities that require payment or payments of interest and/or principal by the debtor to the creditor at a date or dates in the future. This includes debt liabilities in the form of Special Drawing Rights (SDRs), currency and deposits, debt securities, loans, insurance, pensions and standardized guarantee schemes, and other accounts payable. Thus, all liabilities in the Government Finance Statistics Manual 2001 (GFSM 2001) system are debt, except for equity and investment fund shares and financial derivatives and employee stock options. ”
By this definition, gross debt was 85.641% of GDP, so that debt charges of $31 B is not going to be an effective interest rate of 4.8%, but rather 1.7%. So you are looking at other things — e.g. pension liabilities, etc., when measuring gross debt. Net debt = gross debt – assets, and net interest = effective interest on net debt.
For good reason, we generally look at net debt rather than gross debt when calculating debt/GDP ratios, but then it is an error to take the gross interest payments, divide by net debt, and get an effective interest rate.
In terms of measuring what is transferred to bondholders, net interest is the more appropriate measure as well.
The measure that Livio was using before is called (on the page you referenced) the “interest ratio”, which is debt charges/revenues. It is not the effective interest rate.
I would point out that whether you are looking at gross interest/gross debt, or net interest/net debt, you are still getting interest rates less than the NGDP growth rate.