Bond vigilante attack = bursting the bond bubble = target NGDP

The title says it all. There's not much to add. "Bond vigilante attack" is just another name for "bursting the bond bubble". And loosening monetary policy would do that.

Just over two years ago I wrote a post "The bond "bubble", and why we should be worried about it". I don't see any need to change anything in that old post, except to add that it has taken two years longer than I hoped it would when I wrote that last line: "We need to burst the bond bubble. Bursting the bond bubble will help the economy recover more quickly."

Scott Sumner says we should "Never reason from a higher risk premium", and that we should "(f)irst ask why the risk premium rose". Fair enough.

Scott says "...it’s almost impossible to imagine a shock that would be expected to promote both a higher risk premium and a faster recovery.". I disagree. I find it very easy to imagine a shock that would promote both a higher risk premium on US government bonds and a faster recovery: the shock would be for the Fed to do something like what Scott wants it to do -- target NGDP. With currently tight monetary policy, US government bonds (and money) look safe relative to other assets. With a looser monetary policy, other assets would look less risky relative to US government bonds (and money).

Call it "decreasing the fear safety discount [thanks Phil Koop] on bond yields", if you like.

Reframe/remodel.

 

When Eno still had hair. Now my daughter is studying his theories at university.

Comments

[Frances Woolley]

Nick: From a recent interview, where Brian Eno talks about economics Brian Eno: There's an issue we're both interested in – this middle ground between control and chaos.

[Gregor Bush]

"I find it very easy to imagine a shock that would promote both a higher risk premium on US government bonds and a faster recovery."

That's dead right. The term premium on the US 10 year bond (the difference in the 10-year expected return from going long a 10-year bond and going short the 3-month T-bill) is in deeply negative territory. When the economy is close to full employment the term premium tends to be in the +40-80 bps range. But investors are willing to hold 10-year bonds with a negative expected return because the correlation between stocks and bonds is much more negative than normal right now - Treasuries have become much more powerful hedging tools. And I think this negative correlation is very much related to the uncertainty about the future level of NGDP.

If a shock caused the people to expect a higher future level of NGDP, and for the uncertainty around that expectation of future NGDP to be reduced, two things would happen:

1) the expected future path of short-term rates would rise (higher expected NGDP)
2) the term premium would rise because it would make much less sense to hold a 10-year bond at a negative expected return (for any given path of short rates) if investors were much more certain about future NGDP

And of course both of these factors occurring at once would lead to a sharp selloff of Treasuries – but also rise in stock and commodity prices followed by stronger output and employment growth.

[Ritwik]

*Holding the yield on non-govt bond assets constant*

Huge ceteris paribus disclaimer missing from the post, don't you think?

[Ian Lippert]

"With currently tight monetary policy, US government bonds (and money) look safe relative to other assets. With a looser monetary policy, other assets would look less risky relative to US government bonds (and money)."

But does this necessarily make other assets less risky? If we make US bonds more risky than other risky assets those other assets may remain risky even though US bonds have become more riskier.

"Call it "decreasing the fear discount on bond yields", if you like."

I'd call it increasing the fear discount in all assets :)

[Frank Restly]

"Bond vigilante attack is just another name for bursting the bond bubble".

If the monetary authority really wanted a "bond vigilante" attack, they could become that vigilante. If the federal government (Treasury Department) wanted a "bond vigilante" attack, they could become that vigilante.

Before talking about "bond market vigilantees" you might first want to grab a clue about who makes such attacks possible - hint it was not just Paul Volcker. There was also a guy named Blumenthal.

[Phil Koop]

Nice language you write in, but I prefer English, in which "risk premium" already has an established meaning: the extra amount of return required from a risky asset over a riskless one with the same expected return (real-world measure.) So you and Scott are in fact saying the same thing.

If you want to take a risky asset as your numeraire, you need to talk about "shrinking safety discount" or something similar.

[Nick Rowe]

Phil: "If you want to take a risky asset as your numeraire, you need to talk about "shrinking safety discount" or something similar."

That's what I was trying to say. I have edited the post. Thanks.

Frances: good find! I have come across some of Eno's writings on politics and economics a little. Not sure if it's my cup of tea (but then not all his music is, either).

Gregor: "But investors are willing to hold 10-year bonds with a negative expected return because the correlation between stocks and bonds is much more negative than normal right now - Treasuries have become much more powerful hedging tools. And I think this negative correlation is very much related to the uncertainty about the future level of NGDP."

Very important point, and one that I had missed. I was thinking about means and variances, and your point is about covariances.

Ritwik: I thought I was deliberately *not* holding the return on other assets constant.

Ian: I think I worded that bit about "fear discount" wrong. Phil's wording is better, I think.

Frank: sorry, you lost me.

[Frank Restly]

Nick,

"With a looser monetary policy, other assets would look less risky relative to US government bonds (and money)."

"Frank: sorry, you lost me."

You are lost because:

1. You don't look at the supply of government bonds handled by the US Treasury department. The Treasury department could increase the risk of buying government bonds in a number of ways:
a. Sell callable bonds - increases prepayment risk
b. Sell nonmarketable bonds - increases liquidity risk
c. Sell long term bonds - increases inflation and interest rate risk
d. Sell equity that has a higher potential return on investment than bonds

2. You don't acknowledge how the federal reserve affects the profitability of lending to the federal government. By setting short term interest rates below long term interest rates, leveraged buyers can earn a risk free rate of return by borrowing short and lending long. If you want to increase the risk of doing that, then the federal reserve can just push short term interest rates above long term interest rates. This would have the effect of making the US financial system less stable, but if we are going to abandon government bonds, why not pitch the financial system out with them.

[Frank Restly]

Frances,

"There's an issue we're both interested in – this middle ground between control and chaos. Some economists say you can only have a control model or a chaos model, that you're either a socialist or it's all about the free market. Whereas you say: Let's find a place in between."

"The problem is more with the way people think and not the content of it. Human beings are very prone to this black-and-white dichotomous thinking, so if you're a socialist country you allow no market and squash any dissent, if you're a capitalist country you're supposed to – although in fact, many countries don't – you're supposed to put profit and economic growth before any human values."

A socialist system is a system where a governing body controls the means of producing goods. A private enterprise system is a system where production is controlled by the citizenry.
A free market system is a system where all parties resolve disputes without resorting to a legal authority. A closed market system is a system where parties are required to settle disputes using a legal authority.
A capitalist system is a system of liability claims against future revenue (equities and bonds). A fiat system is a system where liability claims against future revenue do not exist.
A democratic system is a system where the governing body and/or legal authority are elected by the citizenry. An imperialist system is a system where the governing body and/or legal authority are given power by edict / lineage / force of will.

I added a fourth to introduce the political element that is missing from the other three. Any myriad of combinations are conceivable. A private enterprise system can exist without capitalism. Likewise a capitalist system can exist under a socialist regime. A free market system can exist with or without capitalism and with or without a private enterprise system.

[Ritwik]

Nick

You're modelling the IBV attack as an increase in the risk premium of govvie bonds and decrease in relative risk premium of non-govvie assets. So, overall, non-govvie asset yields remain same-ish.

The way I think about a bond vigilante *attack*, it will most probably lead to a steepening of the yield curve. When long govvies are sold off, in all probability, risk assets will see a sell-off as well.

[David Beckworth]

Nick,

I was worried at first that we disagreed, but then Phil Koop helped me see we actually agree. We just seemed to define risk premium differently. See lower figures on this post and this post. They show the risk premium is too high and needs to come down. This is the same as saying the low yields on government debt have to go up. And if that happens, it is equivalent to saying excess demand for safe assets will end.

[Scott Sumner]

Nick, Yes, I confused "risk premium" with "default risk." I suppose my (lame) excuse is that lots of people out in the real world who worry about the national debt were also thinking in terms of default risk. But you are right.

I love those early Roxy Music albums.

[Peter N]

But how hard do you have to work to shrink the safety discount? The other central banks have problems:

The Bank of England now holds 27% of UK debt and is buying more. Mervyn King said in a recent speech "As for the MPC, you can be sure we shall be looking for as much guidance as we can find, divine or otherwise. What better inspiration than the memory of those children on Rhossili beach singing Cwm Rhondda."

The BOJ is faced with a crashing balance of trade. But it can't afford to raise interest rates too much because of the cost of servicing the national debt, which debt is exploding.

The ECB is faced with recapitalizing the European banks, but France, Spain and Italy represent 40% of the EMU and would be bailing out themselves by, in effect, buying their own debt. Germany has a national debt of around %80 GDP. It isn't keen on doubling it to bail out other peoples' banks.

This is an explanation from a Kyle Basz newsletter:

"The ECB is going to buy bonds of bankrupt banks, just so the banks can buy more bonds from bankrupt governments [whose central banks are the basis of the ECB]. Meanwhile, just to prop this up, the ESM will borrow money from bankrupt governments to buy the very bonds of those bankrupt governments."

Not an unfair description.

This graphic is from http://redstar.hr/blog/248/european-debt-crisi-by-info-graphics/

"13 Banks biggest banks.
Each bank has its own lane.
Total Convoy Length: 26.7km (16.59 miles) if you put all 13 lanes together.
The people of Greece, Ireland, Italy, Portugal & Spain owe €2,91 trillion Euros to these 13 banks.
These are only the 13 biggest banks, there are many more that lent money, that together add up to a much bigger debt numbers.
3 pallets of cash would cover the cost of all the trucks shown below, if you bought them new. You’d have enough left over for a lifetime of joy. Each truck has 20 such pallets."

These trucks are filled with pallets of 100 Euro notes. Each truck holds 2 billion Euros.

The Fed and China are the only really healthy central banks, so it would take powerful mojo to overcome the safety discount. Maybe too powerful?