Monetary and Fiscal Federalism, Debt, Canada, and the Eurozone.

A government that undertakes a commitment to target 2% CPI inflation does not, strictly speaking, "borrow in its own currency". Its bonds are an indirect promise to pay, via transversality transitivity (damn!), a specified quantity of CPI baskets of goods and services. In much the same way that bonds under the gold standard were an indirect promise to pay a specified quantity of gold. See my old post.

It is true nevertheless that a government whose bonds are promises to pay money that it itself prints, even if it targets 2% CPI inflation, cannot suffer a liquidity crisis. If there were a run on those bonds, where bondholders wanted to sell and hold money instead, the central bank could and would (to keep inflation on target) simply print as much extra money as people wanted to hold and use it to buy back the bonds they did not want to hold.

It cannot suffer a liquidity crisis, but it can nevertheless suffer a solvency crisis. The Magic Money Tree is real, but not very big. If we assume a currency/NGDP ratio of 5%, and Nominal GDP growing on average at 4% (2% Real GDP growth plus 2% inflation), the Magic Money Tree gives the government that owns the printing press profits worth 0.2% (=5%x4%) of GDP per year. Which isn't very big, and is already built into the government's budget anyway, as the profits from the central bank. In a solvency crisis the government would face a choice between defaulting on its bonds and defaulting on its commitment to not let inflation rise above 2%. Much like a government under the gold standard would face a choice between defaulting on its bonds and defaulting on its commitment to making its money worth a specified quantity of gold. But a solvency crisis like that could not coincide with a deficiency of Aggregate Demand where inflation threatens to fall below the 2% target. Instead it could coincide with an excess of Aggregate Demand, where inflation threatens to rise above the 2% target.

The above describes the Canadian Federal government.

Canadian Provincial governments are different. They do not "borrow in their own currency"; they can't pay their debts by printing money. They can suffer a liquidity crisis and/or a solvency crisis.

Canada is not like other countries. Canadian provinces play a big role fiscally, in taxing, spending, and borrowing. [Can anybody find that lovely graph I saw a few months ago showing Canada at the extreme end of the fiscal federalism measure?] Update: Thanks to Jim Sentance for pointing me to Kevin Milligan's post.

If you drew a line, with a unitary state at one end, and the Eurozone at the other end, Canada is halfway along that line. If you abolish the Federal government, leaving only the Bank of Canada, you get the Eurozone. Eurozone national governments are like Canadian provincial governments.

So what happened in the Eurozone could happen here, at the provincial level, even though it wouldn't be exactly the same. The biggest difference is that someone is in charge who can ultimately control both monetary and fiscal policy at the Federal level and decide who gets bailed out under what conditions.

I am not aware of any problems in the near and foreseeable future. But then I don't trust my foresight. And it may be a good idea to think through how we should respond to problems before they happen, and how we might reduce the probability that they do happen. Under what conditions should the Bank of Canada do a Draghi whatever it takes buying provincial government bonds?

Comments

[Brian Romanchuk]

Umm, it did happen here - Alberta in the Depression.

[dlr]

Under what conditions should the Bank of Canada do a Draghi whatever it takes buying provincial government bonds?

That partly depends on what you think mattered about "Whatever it takes." Imagine a slightly modified 2012 Draghi speech, where he was equally forceful and bottomless, except that the immediate mechanism of his preserve-the-Euro operation was to tirelessly print Euros and swap them for foreign equities and real estate (oh and lower the main refi rate since it was still .75) instead of bonds from struggling EZ members. One possibility is that this works just as well and solves the liquidity crisis by virtue of an expected higher price path for the Euro alleviating the market's expectation of spiraling nominal frictions in distressed countries. If this would have been true, is a monetary policy that avoids providing liquidity to provinces preferable? I would say yes, on the basis that the first-best solution is always the one with the least heterogeneous intervention. Similarly and maybe less controversially, all the arguments about Bear and Lehman and AIG should begin with whether the Fed was already doing it's absolute best to achieve its first-best non-bailout price path target (narrator: It wasn't).

If provinces faced a liquidity crisis, the first question is whether the BOC has already applied "whatever it takes" to its primary nominal target in the most homogeneous manner possible. Admittedly people may not agree on the answer. It's possible to have seemingly on-track nominal expectations in the EZ or Canada that still result in expected nominal frictions in weaker countries or provinces if dispersion is high enough and mobility is low enough. But let's assume that the first-best policy is optimal and there is nonetheless a liquidity crisis in countries/provinces that is not caused by (and not ultimately generating) expected nominal frictions. Instead we're in the always impossible situation of determining whether markets are just running themselves into a bad equilibrium (liquidity becomes solvency eventually) or are accurately diagnosing provincial insolvency. If it's the latter, you still might decide to intervene, just not the BOC.

But since there will almost always be some *possibility* it's the former (hey, maybe Lehman could have ended up solvent), there will always be an argument for the BOC involvement as a provincial Bagehot/FDIC to the provinces. While most people think that providing liquidity to "solvent" (haha we never really know) institutions is one of the core purposes of a central bank, I doubt that's the optimal institutional structure. I think a better setup would be to allow the BOC to do only what it deemed necessary to achieve its nominal target and have a separate entity make decisions on whether to try to jolt a particular bank, country or province out of a supposed run with targeted liquidity or guarantees.

[Nick Rowe]

Brian: yes. And it could happen again. And to what extent does it make a difference that there's now a 2% CPI standard vs the gold standard?

dlr: I think those are exactly the right questions to ask.

[Jim Sentance]

http://blogs.ubc.ca/kevinmilligan/2017/11/17/canadas-radical-fiscal-federation/

[Nick Rowe]

Thanks Jim! That's the one I was trying to remember. Post updated.

[JKH]

It seems to me you’re framing liquidity in nominal terms and solvency in real terms.

Is your distinction between liquidity and solvency consistent with your distinction between nominal and real?

And are both of those consistent with the distinction between currency issuers and currency users?

[Nick Rowe]

JKH: As long as the Bank of Canada keeps inflation at the 2% target, nominal and real variables are tied together (with a 2% drift apart of course).

I'm thinking of a "liquidity crisis" as something that makes bonds less liquid (harder to buy and sell) than before, or people having a stronger preference for more liquid assets than before, so they are less willing to hold bonds until maturity, and want to hold more money (the most liquid asset) instead. No problem if it's federal govt bonds, because the Bank of Canada just does an open market operation. And there's no risk to the 2% inflation target, because the BoC is just increasing the money supply in line with increased money demand.

And a solvency crisis as when the bondholders doubt the government will have enough tax revenue (plus seigniorage from the Bank of Canada) to service the debt in future. So they want to sell government bonds, but don't want to hold extra money, they want some other interest-earning asset instead. The BoC would need to increase inflation above 2% to reduce the real value of the debt enough (raise nominal tax revenues enough) that the government would be able to service the debt.

dlr: On re-reading, I really like your comment.

[JP Koning]

IIRC, Alberta refused loans from the Bank of Canada during the Great Depression--it was run by Socreds, after all, who hated the banking establishment--but Saskatchewan was more than happy to accept them. To this day it remains the only example of a province getting direct funding from the Bank of Canada.

Just thinking out loud here, but if a province approaches bankruptcy, I suppose the Federal government would have to step in. Or the Bank of Canada. Either way, would it make a big difference what institution takes the lead? If it was a small province, the Federal government could easily handle the hit to its finances. And so could the BoC, since it could lend to the government for a while and easily sterilize the operation, so that it wouldn't fail to hit its 2% target. If it was a big province like Ontario, and the Federal government stepped in, the BoC might have to monetize much of the new Federal debt, putting its finances in jeopardy and causing it to fail to hit its target. But if the BoC lent directly to Ontario, the same result would occur.

I like dlr's comment too.

Apparently Roy Romanow was close to getting a BoC bailout in 1993:

https://cs.uwaterloo.ca/~alopez-o/politics/GandMarticle.html

[Nick Rowe]

JP: do you know if Saskatchewan had to accept any conditions from the BoC before getting a loan?

I used to still see SoCred pamphlets in my mail when I first moved to Ottawa 1981. Don't think I see them much on the blogosphere nowadays.

"Either way, would it make a big difference what institution takes the lead?"

In terms of economics, I don't see any difference (as long as the loan is smaller than the BoC's existing balance sheet, so the BoC could, as you say, simply sell federal bonds and buy provincial bonds). Politics could make it a bit tricky though, if there was any perceived risk the loan might not be paid back. Think Germany vs Greece, and Target2, only with provinces playing the role of Germany and Greece. I think it would have to be a political decision at the federal level to take the perceived risk "for the good of the country".

[JKH]

I asked the question because - for one thing - the following two points are not consistent:

“cannot suffer a liquidity crisis. If there were a run on those bonds, where bondholders wanted to sell and hold money instead, the central bank could and would (to keep inflation on target) simply print as much extra money as people wanted to hold and use it to buy back the bonds they did not want to hold”

“In a solvency crisis the government would face a choice between defaulting on its bonds…”

There's no need to consider default in light of the liquidity solution.

[Nick Rowe]

JKH: An increase in the supply of BoC money (to buy bonds that would otherwise default), unless it is in response to an equal increase in the demand for BoC money (which it wouldn't be, if it were a solvency not liquidity crisis), would cause inflation to rise above the 2% target.

Or I'm not understanding you.

[Majromax]

> Under what conditions should the Bank of Canada do a Draghi whatever it takes buying provincial government bonds?

Doing a Draghi is an admission that the relevant jurisdiction is not an optimal currency area, and that heterogeneous aggregate demand means that different constituent areas can have meaningfully different (demand-driven) output gaps / inflation rates.

In Canada, the institutional answer to this question is "practically never." We've already tacitly agreed to use transfer payments / equalization to level long-term aggregate demand between provinces, but over the short-term economic fluctuations are more likely due to supply-side factors rather than demand-side factors.

Even something as severe as Ontario defaulting on its debt would be a supply-side shock; in and of itself it would be a policy-related decision that impacted previously-believed-secure property rights. There might be resulting AD-affecting contagion from a consequent financial crisis, but there's no reason to believe that contagion would remain localized: firms in Vancouver and Montreal would probably have as much difficulty securing new private credit as firms in Toronto.

[JKH]

In your first paragraph, you define default as a failure to deliver the full money value of a committed real basket of CPI goods and services – i.e. failure to deliver full real value in the face of unexpected inflation.

In your third paragraph you seem to be defining a nominal dollar default. That make no sense for a currency issuing government, because the central bank can always buy maturing bonds. And this is supported by what you say about liquidity.

And you also say:

“In a solvency crisis the government would face a choice between defaulting on its bonds and defaulting on its commitment to not let inflation rise above 2%.”

If inflation rises above 2 per cent, you have a real default by definition. So defaulting on the real value of bonds and default on the inflation commitment are the same thing. So there is no “choice”.

And nominal default is not an issue.

I still can’t make any sense of it.

[Henry Rech]

"But a solvency crisis.............. could coincide with an excess of Aggregate Demand, where inflation threatens to rise above the 2% target."

How would that come about?

[Jussi]

"But a solvency crisis.............. could coincide with an excess of Aggregate Demand, where inflation threatens to rise above the 2% target."

These claims are IMO always hard to assess because the only cases this can happen is that the tax base or ability is demolished - both far fetched in the case of Canada.