Sovereign insolvency and illiquidity

If a country has a debt/GDP ratio of 100%, and is paying 9% interest, and nominal GDP is not rising, then it's got a solvency problem. It needs to run a budget surplus of 9% of GDP just to stop the debt/GDP ratio rising further. And that is very hard to do.

But why would a country ever be paying 9% interest and have 0% nominal GDP growth? By David Beckworth's simple, crude, but nevertheless useful measure of the tightness of monetary policy -- the gap between nominal interest rates and expected nominal GDP growth rate -- a gap of 9 percentage points is very tight monetary policy. No country that had control over its own monetary policy would set monetary policy that tight (OK, unless its population were falling exogenously at around 9% per year, or if it were a purely temporary measure to reduce entrenched inflation).

Countries like the UK, US, Japan, have large debts and deficits. But they control their own monetary policy, and none of them have monetary policy anywhere near that tight. If they did set monetary policy that tight, they too would have a solvency crisis.

Ireland has a solvency crisis, but only because it has a liquidity crisis. Ireland does not control its own monetary policy. Money is the most liquid of all assets.

If monetary policy were less tight, so the gap between nominal interest rates and nominal GDP growth were 1%, a country with a 100% debt/GDP ratio would only need a budget surplus of 1% of GDP to prevent the debt/GDP ratio from rising. That's doable. That country is solvent.

Comments

[Ian Lee]

With thanks to David Rosenberg's commentary today for the quote of the day by Ken Rogoff:

“It is preposterous for European politicians to claim that the whole crisis is being caused by irrational markets. The debt levels across much of Europe, both private and public, are off the charts.”

Nick, can you illuminate your argument with actual current numbers from Europe. I thought that the ECB rate was very low along with a low rate of inflation and low economic growth rates.

It was my understanding - influenced strongly by Rogoff - that debt levels across the EU are very high and in some instances, e.g. Ireland, Greece, Portugal, are essentially insolvent (defined as "unable to pay one's debt obligations as they become due").

[Jim Rootham]

The only reason that Ireland is insolvent is because of the foolish promise to bail out the banks. It remains to be seen whether the upcoming IMF riots there can push that back.

[Nick Rowe]

Jim: the promise to bail out the banks is what pushed the debt/GDP ratio up for Ireland. But if monetary policy were looser, and so interest rates were lower, and nominal GDP were rising (not falling), those banks would be a lot more solvent than they are, and the sovereign debt/GDP position would be a lot lower. And even if it were still as high as it is,...

Ian: but even if we ignore the above, and take IIRC 130% as Ireland's debt/GDP ratio, the UK has had debt/GDP ratios of over 200% a several times in the past couple of centuries. And that's when the government and taxation was a smaller share of the economy too. If monetary policy were looser in Ireland, so the gap between interest rates and nominal GDP growth were say 1%, a 130% debt/GDP ratio means you need a surplus of 1.3% of GDP. No big deal. Canada has done that.

Ireland is only insolvent because tight monetary policy is making it insolvent.

[RSJ]

Nick,

That the central bank sets an overnight interest rate of 1% does not mean that government bonds yield 1%. You are assuming the expectations hypothesis here, but when attempting to describe what happens in the real world -- c.f. the issue I raised in "What we research and what we believe". I understand that if in the models, it's too hard to think of money market rates as being separate from bond rates, which are themselves separate from equity rates, but you have to abandon that fudge when describing the real world. And no amount of backtracking that "It's only a model, it's only *useful*" is credible because of economists then turn around and make statements such as these when describing the real world.

First in the EU, nations are paying a risk premium that nations such as the UK or US are not. Which is why UK rates, for example, are quite low even though their public finances are worse than that of Ireland. Nations such as the US and UK have the option of money financing, which means that there is no default risk. Ireland does not have this option, and therefore there is a default risk, so it pays higher rates, which makes the default risk worse, etc. Such a cycle doesn't happen in the U.S., U.K, or in Japan, because these are currency issuers whereas Ireland is not. Despite the warnings that a bond crisis in Japan is "just around the corner", they continue to enjoy low (and falling) rates as the debt ratios go up, to the rising anger of the of the hard money crowd.

It's not because Ireland doesn't have control over the OIR, but because Ireland doesn't have the option of money financing that makes the difference.

You can imagine a situation in which a nation has control over the OIR -- for example the government can tax banks with excess reserves -- seizing those reserves -- and lend the reserves out to other banks, keeping the total level of currency constant, or growing according to whatever k% rule is externally imposed on the government.

In that case, any bank that is short of reserves could borrow at the OIR in order to replenish those reserves.

The government would have complete control over the OIR, but because it would not have the option of money financing, it might be charged a risk premium and you could get an Ireland-type situation. So control over OIR != control over your own interest rate.

And as I argued previously, the long term risk-free nominal rates will, over long time periods, be equal to the NGDP growth rate, and as government finances itself with maturities that are less than long term, the interest rate on government debt will, over long time periods, be less than the GDP growth rate, assuming government debt is risk-free. Which is why, after 200 years of ridiculous spending on wars, boondoggles, and everything in between, governments continue to have very reasonable sovereign debt to GDP ratios, even though they almost never run primary surpluses, but pretty much run deficits every year. Government debt is never "repaid", but is just rolled over, and Governments that are currency issuers are never insolvent, it's the private sector, which does not enjoy risk-free rates or infinite liquidity, that becomes insolvent.

The specific issue with Ireland is that their sovereign debt to GDP ratio is less than Germany's, but they guaranteed their private sector bank bondholders, so they have enormous off-balance sheet liabilities.

Basically they tried to extend their low funding costs to the 10xGDP of external debt held by the private sector, and realized that their own funding costs went up, rather than having the private sector's funding costs go down. They were operating under your hypothesis, that their funding costs would be low because the ECB rates are accomodative, and this bit them. Only by recovering the option of money financing their debts would they be able to pull a Japan, and try to bail out the private sector while keeping sovereign rates low.

[RSJ]

FYI,

Here are some sovereign debt to GDP ratios (for 2009):

https://www.cia.gov/library/publications/the-world-factbook/rankorder/2186rank.html

Japan 192%
Italy 116%
Ireland 65%
Singapore 110%
Greece 113%
UK 68%
Egypt 81%
US 54%
Germany 73%
Netherlands 61%
Spain 53%
Greece 113%
Portugal 77%
Canada 83%
Israel 78%
France 78%

[Kevin Donoghue]

The only reason that Ireland is insolvent is because of the foolish promise to bail out the banks.

True enough, but the implication is that the banks were insolvent when the promise was given. So Rogoff is right about aggregate (public+private) debt levels. It's just too bad that the debts of property developers were transformed into government debt.

[RSJ]

"It's just too bad that the debts of property developers were transformed into government debt."

Wise words, and this reflects the anger occurring everywhere. Governments are engaging in "austerity" -- a transfer of wealth from households to bank creditors -- in order to bail out their financial sector, when they should be spending more to compensate for the shock of financial sector defaults. The transfer that is needed is in the other direction, from creditors to debtors.

When a business that produces goods can no longer sell those goods at a profit, it goes bankrupt, but when a financial intermediary is no longer profitable, then for some reason the government decides to make it profitable, transferring real resources from debtors to creditors. How can this not lead to a permanent excess savings demand on the part of both creditors and debtors? Debtors need to save in order to repay debt that cannot be repaid, and Creditors become accustomed to earning unrealistic returns.

[babar / q]

Thanks Nick, it's far from clear to me there is any real issue with govt debt at US levels. There's plenty of fear about it and that fear is somehow easy to use to get people riled up. Fear has the interesting characteristic that it gets people to stop thinking. For that reason it's difficult to get through with arguments of the form "Don't be afraid; there's no problem here."

> Ireland is only insolvent because tight monetary policy is making it insolvent.

If monetary policy were looser, presumably fewer of the debts that Ireland guaranteed would be bad, and the economy would be greater in nominal terms, so either Ireland would have an easier time repaying or defaults would be fewer and less harsh on credit nations.

[OGT]

This is presuming the country can borrow in its own currency, correct?

[Nick Rowe]

OGT: Correct.

babar/q: "If monetary policy were looser, presumably fewer of the debts that Ireland guaranteed would be bad, and the economy would be greater in nominal terms,.."

Yep. I'm ignoring that in my post, just looking at the simple relationship between debt/gdp, the gap between interest rates and nominal GDP grwoth, and the surplus needed.

The only reason to be worried about the US debt is US politics. Can the US, politically, close its deficit?

RSJ: Sure, the average interest rate on government debt is usually a percent or so above the overnight rate. And financing government spending by printing money is not at root different from open market operations, and the profits the central bank earns from the interest from the bonds it holds. And the MMTers do forget that there *is* a real limit on how much revenue that can be collected from seigniorage. But let's not go there on this post. I'm making a simple point here, about the arithmetic of debt and deficits.

What you, and everyone else, is calling a "risk premium", I say is tight money. Why isn't there a similar risk premium on UK debt? There's only a risk because monetary policy is tight. There are two equilibria. And the job of monetary policy, as lender of last resort, is to get the country to the "good" equilibrium.

[Gregory Sokoloff]

Nick, could you comment a little on what precisely justifies labeling a state insolvent? Recently, the Eurogroup announced a new policy whereby member states will be divided into the solvent but needy versus the insolvent and needy, and receive different medicine for each malady (see Simon Johnson's article). How does a state get the former diagnosis? Apparently, only by the unanimous vote of the Eurogroup Ministers!

So how would a right thinking macroeconomist like yourself judge the solvency of country? You must need a model with a number of variables some of which can't be very precise. For example, exactly how much sacrifice will the population put up with to raise the primary surplus, and so forth.

[RSJ]

"There's only a risk because monetary policy is tight. There are two equilibria. And the job of monetary policy, as lender of last resort, is to get the country to the "good" equilibrium."

OK, but I'm trying to argue for precision, because precision is important. If by "tight money", you mean, "the inability to issue your own money", then I agree. But the latter is a power, not a policy. But if by "tight money", you mean overnight interest rates are too high, then I don't see how 1% OIR is too high. The ECB does not have a tight money policy.

Being able to issue your own money is crucial. It's like deposit insurance. Because there is deposit insurance, then there are no bank runs, and you don't actually need to bail out depositors. But I wouldn't call that a "loose money" policy. In the same way, being able to issue your own money means that Government debt is risk-free. And so you don't need to actually issue your own money in large amounts, since selling debt is affordable. It's not because of the policy, but the currency issuing power.

In terms of the limits of seignorage, I don't think you are accurately describing the MMT position. The position is that there isn't a whole lot of difference between issuing zero maturity currency and longer maturity debt. Because both are liquid risk-free assets; they will have roughly the same effects on inflation.

If government buys too many goods, then prices go up, independently of how those purchases were financed. Any differences, in terms of inflation, will be because of interest rate effects, in the sense that bond sales are currently used to maintain the OIR at target. But there are other mechanisms to maintain the OIR that do not require bond sales. Canada, for instance, doesn't rely on OMO.

There is no a priori reason why government liabilities should be 7 Trillion in bonds and 1 Trillion in currency, as opposed to 7 Trillion in currency and 1 Trillion in bonds. In both cases, households deposits are about 8 Trillion. To see that, you need a financial sector in your model.

Either way you cut it, the amount of "money" in the economy is going to 8 Trillion. And you can see this correlation across the board, just by comparing, say, MZM to CB liabilities + Government Debt held by the public. It's basically the same time series. Not exactly the same, as money has an endogenous component, but it will be roughly the same.

So in that case, you have to ask why the government bothers selling debt, and why it can't manage OIR without selling debt. The argument is a lot more subtle than just talking about seignorage.

[Scott Sumner]

Nick, I don't think you can say there is tight money merely because rates are high on government debt. I happen to think money in the eurozone is tight, but for entirely different reasons. The high rates in Ireland reflect default risk. In order to claim they reflect tight money you'd have to show that borrowers in the same currency with lower default risk also had to pay high interest rates. But that's not true, German rates are very low.

It would be like arguing that a 9% interest rate on GM bonds shows money is tight in America. No, it shows GM has a high default risk, as the US Treasury is only paying 2%.

Think of Ireland as a like a company in greater Germania.

Again, money is very tight in Ireland, but not because of high rates on government debt. It's tight because nominal GDP growth in the eurozone is below target.

[Nick Rowe]

Gregory: "Nick, could you comment a little on what precisely justifies labeling a state insolvent?"

There is no precise way of telling. Not even vaguely anywhere near precise. Even if you grant all my assumptions, that the interest rate would stay permanently at 9%, and nominal GDP growth at 0%, and debt/GDP at 100%, then it's a judgment call of whether it would be possible to run a budget surplus of 9% of GDP. My guess is that it would be very hard, politically.

So when the Euro ministers get together and decide, it will be a political decision.

But presumably the bond market is judging that Ireland has a significant risk of default.

RSJ: If Ireland had its own money, it could (if it were really stupid) have monetary policy so tight that it had 9% interest rates on its government bonds and falling nominal GDP. But it wouldn't be so stupid. If it ever found itself in that position, it would loosen monetary policy.

You cannot look at the overnight rate of interest alone as a measure of whether monetary policy is tight or loose. You cannot look at *any* nominal interest rate alone. David Beckworth's suggestion, that we look at the gap between nominal interest rates and and nominal GDP growth, is a much better measure. Scott Sumner, and I, have spilt a lot of pixels on why nominal interest rates are a very poor measure of the stance of monetary policy.

No, sorry, but money and debt are not the same thing. But I'm not going into that here.

The MMT position

[Nick Rowe]

Scott: "Nick, I don't think you can say there is tight money merely because rates are high on government debt. I happen to think money in the eurozone is tight, but for entirely different reasons. The high rates in Ireland reflect default risk."

But that default risk is not independent of monetary policy. If Ireland had it's own central bank, to act as lender of last resort to the government, it would immediately do an OMO, nominal interest rates on bonds would come down, growth rate of NGDP would go up, and the default risk would fall.

You cannot separate default risk from tight monetary policy.

Think of it as like a Diamond Dybvig bank run. There's a run on Ireland, and no run on Germany.

[RSJ]

"If Ireland had its own money, it could (if it were really stupid) have monetary policy so tight that it had 9% interest rates on its government bonds and falling nominal GDP"

That's an interesting thought experiment. I claim the following would happen:

The CB of Ireland does OMO so that interbank rates go to 9%. Long term rates would shoot up, but borrowers could not credibly commit to paying 9%, so they would not be able to borrow. Thus there would be an excess of lenders over borrowers, and long term rates would start dropping like a rock, together with NGDP growth.

You would get an inverted yield curve, in which long term rates are below short term rates. Something similar happened when Volcker hiked rates. The Long Term rates were below the interbank rates.

Now you could argue that long term rates are below the short term rates because the CB cannot credible commit to keeping short term rates so high. At that point, it's observationally equivalent, since if the CB can only credibly commit to keeping ST rates at a level consistent with NGDP growth, then CB can't really set nominal short term rates except except in some zone of plausibility.

There is some freedom in that zone, but not complete freedom.

Alternately, you can argue that borrowers are constrained with what rates they can commit to paying, even in nominal terms. For short term borrowers, there is a somewhat inelastic need to borrow, but even there, at some point they will stop borrowing. Banks must have reserves, but not if they go out of business. Then they don't need reserves.

"No, sorry, but money and debt are not the same thing. But I'm not going into that here."

Clearly deposits are not debt. :)

But seriously, is it asking so much for you to define what you mean, instead of using strategically vague terms such as "money"? Not to mention changing the meaning of said term based on the phase of the moon? It would really clarify the discussion, and force some intellectually consistency.

Regardless of what you believe, you have to be able to specify what you mean by "money" -- currency, reserves, deposits, checkable deposits, MZM -- what? Why all the hand waving, if not to hide some confusion about what is going on?

But if you are unable to specify which asset you are talking about, no matter how hard you try, then statements such as "This asset is not debt" are meaningless. In that case, I would suggest that you don't really have a theory. You have a hunch, dressed up as theory. So first tell me precisely which asset you like to call money, and then we can decide whether said asset qualifies as debt, and in what way.

[OGT]

Nick: "But that default risk is not independent of monetary policy. If Ireland had it's own central bank, to act as lender of last resort to the government, it would immediately do an OMO, nominal interest rates on bonds would come down, growth rate of NGDP would go up, and the default risk would fall."

That's one of the reasons I asked about the presumption of borrowing in its own currency. Ireland gave up the power to do that, and can't take it back now. Their debt is in Deutsche-Francs.

If a US state or Canadian Provence had a business cycle particularly out of step with the broader nation and there was no fiscal transfer union and limited labor mobility they could certainly become insolvent (though labor mobility may be a mixed blessing to an indebted state).

Also, there have been repeated soveriegn defaults in the "fiat" era internationally. Pretty much all of these have involved countries that through the repeat game of borrowing and printing lost the ability to borrow in the currency they controled.

[Nick Rowe]

OGT: Yes, having your own currency doesn't guarantee you can't get a sovereign default, if the country is fiscally irresponsible enough on an ongoing basis, or if the debt/GDP ratio is just way too high. But that doesn't sound like Ireland.

[vimothy]

Ireland had a debt/GDP ratio of about 25% in Q4 07, IIRC.

[vimothy]

I mean national debt there, of course.

[vimothy]

Kevin O'Rourke sums up the whole sorry affair in this rather brutal "letter from Dublin": http://www.eurointelligence.com/index.php?id=581&tx_ttnews[tt_news]=2973&tx_ttnews[backPid]=901&cHash=484db55c3a

[RSJ]

"Also, there have been repeated soveriegn defaults in the "fiat" era internationally. Pretty much all of these have involved countries that through the repeat game of borrowing and printing lost the ability to borrow in the currency they controled."

There hasn't been a single sovereign debt default by a government whose debts could be extinguished with floating non-covertible currency that it could issue.

There have been sovereign defaults by nations that either promise some form of convertibility (in the form of a peg or currency board arrangement) and/or borrow in a foreign currency. I.e. there are non-fiat nations in the "fiat era" and these are the ones that default.

[Sergei]

Is my reading of your post correct that you define insolvency as increasing debt/GDP ratio?

[Nick Rowe]

Sergei: not quite, but close. If the debt/GDP ratio increases indefinitely, without limit, then eventually you pass an (undefined) point of no return, where you can't possibly ever pay off the debt.

[Sergei]

You can not make claims on "undefined" points. 9% rates and 0% GDP growth will not stay forever at these levels in order to test your claim about "undefined" point of insolvency.

It is like string theory. A nice but not testable construct which allows many physicists of other "beliefs" laugh at them.

[Nick Rowe]

Sergei: Oh yes I can! Just watch me!

If nominal debt exceeds the present value of nominal GDP, you can't pay it off. Assuming constant growth, that puts an upper bound on the debt/GDP ratio of 1/(r-g). The undefined actual bound must be somewhere below that upper limit.

[Sergei]

Nick, Oh no you can not! Just watch me! :)

You nationalize the financial sector (pretty much where we are in some places anyway) and do left pocket-right pocket accounting. Why do you think your fundamentalistic financial economy is anyhow more plausible or better than my state controlled one?

You do not like a nationalized financial sector? Then what about Basel 4 regulations making even stronger (absolute) focus on liquidity? Banks anyway repo this stuff with central banks so no big problem for them. So just a bit of higher cost of business. They still make way too much money if you ask me :)

Or at a point of time X you introduce a change to the tax code which brings back the 90% tax rate (on interest rate gains). You will have to really-really struggle with your GDP in order not to make your ratio as close to infinity as possible.

They might be even more perfectly legal "solutions". If you legally allow central bank to play with interest rates why should you deny the ultimate player in any country (congress) the right to play with other solutions?

The point is that a truly sovereign country has all such solutions available. So debt/GDP ratio is never a problem. It does not mean that a truly sovereign country can not decide to default but then they can decide to default at any point in (0, infinity) without consulting with your formula.

[Nick Rowe]

Sergei: Oh Yes I ...:-)

They will all emigrate. You'ld need a Berlin Wall!

[Scott Sumner]

Nick, Yes, I think you can separate default risk from tight money, as with a junk bond. Again, I agree money is tight in Ireland, but not because rates are 9%. Suppose eurozone money had not been tight, say NGDP was rising at a steady 4% rate year after year, and yet Ireland ran a reckless fiscal policy. They might be paying 14% interest, but that would merely reflect default risk, not tight money. Even in that case however, currency depreciation would solve their debt problem. But merely pointing to the fact that currency depreciation will solve the debt problem of a country facing a high default risk, doesn't in and of itself prove tight money. I still think you are mixing up two issues, which often tend to coincide, but not always.

[David Pearson]

There are two components of a sovereign risk premium. The first is compensation for default risk. The second is compensation for accepting the uncertainty of future inflation. In Brazil during its inflationary period, there was little default risk in local currency sovereign debt, yet the term premium made issuing long term debt prohibitively expensive for the government. Thus, inflation prone economies tend to only have short term debt markets in local currency, and they rely on foreign currency issuance for longer term debt.

So you are basically saying that developed country governments can raise NGDP at will, and are far from being charged a high premium for inflation uncertainty. The problem with this assertion is that the view of future inflation uncertainty does not change in a linear fashion as the central bank raises NGDP. This is not something you can determine running historical regressions. At what combination of expected QE and large, long term deficit projections does inflation fear set in to markets? If the Fed thinks it knows this function, it is sadly mistaken.

[Determinant]

Nick loses the argument through the Post-War Economics Interpretation of Godwin's Law.