Monetary policy is not interest rate policy - Japanese version

Suppose I announced I would be buying an asset, both now and in future. And suppose people believed my announcement. It would be paradoxical if my announcement caused the price of that asset to fall. It would be even more paradoxical if I had said I was buying the asset because I was trying to make its price go up. That's what's been happening in Japan. Bond prices have been falling (and so bond yields have been rising), as a result (probably) of Abenomics.

There are a few cases where this paradoxical result might happen:

Maybe I have a reputation for being a really bad investor, who always loses money. So when other people see me buying houses, they all decide to sell their houses.

Maybe the sheer knowledge that I own an asset makes that asset less valuable. Because I'm so very unfashionable that if I buy a Burberry swimsuit everybody else stops buying Burberry.

Or maybe I'm buying taxi medallions, and paying for them with newly-produced bicycles. With a flood of bicycles on the market, fewer people want to take a taxi, so the medallions are worth less.

That last example is the closest to the current Japanese case. It's not just what you buy, but what you buy it with. An increase in the stock of bicycles in circulation on the streets, both now and in future, will have an effect on the demand for taxi medallions. An increase in the stock of money in circulation, both now and in future, will have an effect on the demand for bonds. Because it can raise the expected future price level, and/or expected future real income, which would reduce the demand for bonds.

If you think of monetary policy as interest rate policy, it's hard to escape the paradox. It's like if your mechanic hooks up your car's power steering hoses the wrong way around. So if you try to turn the steering wheel clockwise, to turn the car right, you are surprised to feel the steering wheel fight back, actually turn counterclockwise, and the car turn left. Once you figured out how it worked, and got used to it, you might be able to steer the car. But is the English language up to the job of clearly describing what you are doing when you steer the car? "Which way are you trying to turn the steering wheel??". My guess is that the Japanese language would have the same problem.

If you are aware that printing money and buying bonds and "trying" to force bond yields down will in fact result in economic recovery, and expectations of recovery will cause a rise in investment, a fall in saving, and less demand for the liquidity and safety of government bonds, and these will cause a rise in bond yields, and if you intend that economic recovery to happen, and you are willing to accept that rise in bond yields as a side-effect of that recovery, are you really truly trying to force bond yields down? Well, yes and no. But the rise in bond yields does not mean your policy has failed. On the contrary, it means your policy is succeeding.

Is the Bank of Japan trying to push down bond yields? Well, yes and no. Yes, it is fighting a battle to push down bond yields, but that battle is part of a wider war for economic recovery. And if it wins that war for economic recovery, it will lose that battle to push down bond yields. So it wants to lose that battle to push down bond yields.

This is what Ambrose Evans-Pritchard says that Richard Koo says:

"As I reported last night, Mr Koo thinks the Abenomics plan of monetary reflation is madness. “Once inflation concerns start to emerge the BoJ will be unable to restrain a rise in yields no matter how many bonds it buys.” This could lead a “loss of faith in the Japanese government” and the “beginning of the end” for Japan’s economy."

Yep. When the Japanese economy recovers enough from deflation and recession, the BoJ will be unable to restrain a rise in bond yields. Because the harder it "tries" to restrain them, the more the equilibrium yield will rise. But when the recovery looks strong enough, the BoJ will stop "trying" to restrain bond yields, and start "trying" to increase them, at which point those bond yields will stop rising.

Ambrose also says that Richard Koo is "... Japan’s most famous economist and an arch-Keynesian." There's a lot of weight placed on the "arch" bit in "arch-Keynesian. Paul Krugman is a Keynesian, but I don't think Paul would have a problem recognising that Abenomics, if successful, would drive up bond yields. I think he would cheer the rise in bond yields, and say it means that the BoJ has escaped the liquidity trap, so that monetary policy once again has "traction". Is Richard Koo really a Keynesian? Or is he just a finance guy, who doesn't really get macro/money? (I don't know the answer to that question.)

This is not the "beginning of the end" of the Japanese economy. But it might be, and I hope it is, the beginning of the end of Japan's long recession.

And if this is indeed the beginning of the end of Japan's long recession, it will also be the beginning of the end of Richard Koo's thesis that monetary policy is powerless in a balance sheet recession, and that only fiscal policy can offset private sector deleveraging. And we can only regret that Japan did not do this many years earlier, instead of wasting all those years and letting Japan's government debt/GDP ratio climb. Because that high debt/GDP ratio is the only reason why someone might want Japan's economic recovery but not want the higher interest rates that will accompany that recovery. Which is no reason to try to stop the recovery. Though it is one additional reason to regret not having done something like Abenomics a lot earlier.

Comments

[Anon ymous]

I think a much simpler explanation for the paradox is bond term structure.

The BOJ is buying short term bonds, enabling deficit spending by the government.

The BOJ is not buying long bonds - those thus reflect the market's expectation of future Japanese growth.

So Abenomics has convinced the market that Japan will grow in the future, escaping deflation and ZIRP.

And in fact Japan has already started growing: 3.5% GDP growth.

So Abenomics works - there's no paradox: combined fiscal and monetary policy can fight deflation and can create growth out of thin air, where the free market was unable to escape deflation for 20 years. If it keeps up it's an incredible success story.

[Andrew F]

The sad thing is that you were talking about this on your blog years ago. If this works, then the last twenty years--a whole generation--suffered deflation needlessly. And if it works, perhaps it will be instructive to the West to not ease up on QE too soon...

[Andrew F]

An aside, Nick, but I've noticed much more chatter about currency wars with Abenomics. Is 'competitive devaluation' really zero sum? If there isn't enough inflation to go around, is it not a good thing for all parties to be engaged in competitive devaluation? If the money supply doubles for Yen and USD and their exchange rate remains unchanged, there is still (desirable at this point) inflationary pressure?

I guess I just don't understand the griping about currency wars--do you understand it?

[K]

"It would be paradoxical if my announcement caused the price of that asset to fall."

Exactly. Bonds, in fact, rallied huge on the day of the announcement. Anything that happened on any other *non-news* day could be due to any one of monetary, fiscal, or general economic conditions. I say the recent bond puke-out is due to better than expected economic news and increased expectations of stimulative fiscal policy. Also, targeted purchases of equity ETFs by the BoJ.

"An increase in the stock of money in circulation, both now and in future, will have an effect on the demand for bonds. Because it can raise the expected future price level, and/or expected future real income, which would reduce the demand for bonds."

Except, inflation expectations didn't move on the announcement. Not one iota. How do you account for a money channel that doesn't raise inflation expectations?

"If you think of monetary policy as interest rate policy, it's hard to escape the paradox."

There's no paradox. The long end of the curve tends to move in the opposite direction of unexpected rate changes in the short end of the curve. It's a necessary consequence of Wicksellian inflation dynamics. If you don't distinguish between different parts of the curve, it does in fact become very difficult to understand interest rate policy.

"Or is he just a finance guy, who doesn't really get macro/money?"

Like Woodford? Also doesn't believe in the QE channel. 

"it will also be the beginning of the end of Richard Koo's thesis that monetary policy is powerless in a balance sheet recession, and that only fiscal policy can offset private sector deleveraging."

Only if you can demonstrate that there was no change in economic fundamentals and the outlook for fiscal policy. That's going to be tough.

[rsj]

"If you are aware that printing money and buying bonds and "trying" to force bond yields down will in fact result in economic recovery"

I would like a single example of this occurring, because most people are not, in fact, aware of this "fact". There has been no economic recovery as a result of either Japanese or U.S. QE, even though both nations have beat this drum repeatedly to no effect.

[Petar]

Wasnt Koo the guy who advised Japan government to spend all that money to boost income claiming that Japan is in a liquidity trap. That worked out well. Now he uses the deleveraging recession theory or something like that, to prove how government stuffed a credit hole the BoJ couldnt

[Mark A. Sadowski]

"Is Richard Koo really a Keynesian? Or is he just a finance guy, who doesn't really get macro/money? (I don't know the answer to that question.)"

I finally went to the trouble of reading one of Koo's papers recently just to see what all of the fuss was about and I'm almost sorry that I did:

Koo's paper is here:

http://snbchf.com/wp-content/uploads/2013/04/Koo-Ineffectiveness-Monetary-Expansion.pdf

The passage I take the most issue with is on pages 11-12:

"The post-1990 Japan managed to maintain its money supply (Exhibit 7) and GDP (Exhibit 4) from shrinking because the government was borrowing the deleveraged and newly saved funds from the private sector (Exhibit 9) Unfortunately there was a period in economics profession, from late 1980s to early 2000s, where many noted academics tried to re-write the history by arguing that it was monetary and not fiscal policy that allowed the US economy to recover from the Great Depression. They made this argument based on the fact that the US money supply increased significantly from 1933 to 1936. However, none of these academics bothered to look at what was on the asset side of banks’ balance sheets.

The asset side of banks’ balance sheet clearly indicates that it was lending to the government that grew during this period (Exhibit 10). The lending to the private sector did not grow at all during this period because the sector was still repairing its balance sheets. And the government was borrowing because the Roosevelt Administration needed to finance its New Deal fiscal stimulus. In other words, it was Roosevelt’s fiscal stimulus that increased both the GDP and money supply after 1933."

There's a reason why none of those academics attached much importance to the asset side of banks' balance sheets.

Koo acts like there's a one-to-one correspondence between the banking sector's assets and public and private sector liabilities, but that's not at all true. For example, as Koo shows in Exhibit 9, banks held about 500 trillion yen and 250 trillion yen in private sector and public sector debt respectively in December 2007. But at the end of 2007 the total amount of non-financial private sector and public sector debt was about 850 trillion yen and 900 trillion yen respectively:

http://4.bp.blogspot.com/_up3_ViopRks/SrjWz9o6zgI/AAAAAAAACnU/TZNfdoqSucA/s1600-h/JapanDebt.png

Similarly in Exhibit 10 Koo shows that banks held about $16 billion in private and public sector debt each in June 1936. But in June 1936 the total amount of non-financial private sector debt and public sector debt was about $116 billion and $54 billion respectively (page 989:

http://www2.census.gov/prod2/statcomp/documents/CT1970p2-11.pdf

(Incidentally, why did Koo put the cutoff in 1936 instead of 1937? The recession didn't start until 1937 after all. It's probably because the asset side of the 1937 bank balance sheets are a lot less supportive of his already flimsy argument.)

In both cases banks held only a fraction of each kind of outstanding debt. There is simply no clear relationship between debt and money supply. (Monetary policy is not credit policy.)

More importantly, the period 1997-2007 includes the first Japanese QE (2001-2006) and the corresponding Koizumi Boom (2003-2007), which was also a period of pronounced fiscal tightening in Japan. Similarly, although the Federal Emergency Relief Work and public works projects funded by the New Deal were helpful in creating the swift recovery experienced from the Great Depression during 1933-36, as several academic studies have shown (e.g. Romer, Eichengreen, etc.) the recovery was primarily due to FDR's ending of the gold standard and the subsequent devaluing of the US dollar.

[Peter]

"Because that high debt/GDP ratio is the only reason why someone might want Japan's economic recovery but not want the higher interest rates that will accompany that recovery."

Higher interest rates due to higher nominal growth doesn't make the debt any harder to maintain. If NGDP grows 1%-point faster and interest rates go up 1%, it makes no difference in the cost of maintaining debt/NGDP.

The same is true for mortgages and debt/nominal wages. I blogged about this recently. But it's in Swedish, so I wont ask you to read it. :)

[Lorenzo from Oz]

My take on Koo (having read some of his papers and articles) is that he has not read any of the literature on the gold standard and the Great Depression. Or if he has, he has not understood it. Or utterly forgotten it. Or something.

And no, he does not get macro/money. In particular, he does not get monetary policy and income expectations.

[Nick Rowe]

Anon ymous: yes, but I think that bringing in the term structure is rather like ducking the paradox. It's a cop-out. I would rather face the paradox head-on. For example, we can certainly imagine a world in which all Japanese bonds are perpetuities, so the BoJ can only buy perpetuities. And if the BoJ buys perpetuities, and raises expected NGDP growth, the price of perpetuities falls.

Andrew: yep, but I suspect that economists like Richard Koo have been much more influential in setting Japanese policy. And one thing I didn't mention in my post is that Richard Koo has, I think, been arguing that the BoJ should fight to keep its reputation as an inflation-fighter. So it's not just that he has ignored monetary policy as a tool to help Japan escape the recession, he has been advocating a counter-productive monetary policy. As long as people expected the BoJ would act to prevent NGDP rising, Japan stayed stuck.

"Currency wars" doesn't have to be a zero sum game. If one country alone reflates, the gains to that country will be bigger than the losses to other countries. So if all reflate, they all gain.

K: I acknowledge that the timing of events doesn't fit any clear pattern. Why did the bond market first move in one direction, then move in the other? Was the bond market just slow to realise that Abe was serious? Or was it slow to figure out the general equilibrium/macro effects? And if the bond market is expecting higher NGDP growth, how much of that is higher RGDP growth and how much is higher inflation? The BoJ communications policy is less than ideal: partly because it talked about inflation rather than NGDP; but mostly because it talked about pushing down interest rates.

I don't think Woodford would have any problem understanding that loosening monetary policy via a higher inflation or NGDP target would cause bond prices to fall, and that that would be a sign of successful loosening.

Yep, empirical macro is hard. Because we can never assume other things are equal. That's why I stuck in the word "probably" right at the beginning. But Richard Koo is also saying the rise in bond yields is due to Abeonomics.

rsj: the only thing a central bank controls is its own balance sheet. "QE" is just a silly name for what central banks always do. They buy and sell IOUs. If you say "QE" doesn't work, you are saying that monetary policy never works. Canada mid-1990's is one counterexample, where the BoC successfully kept inflation on target despite a massive fiscal tightening.

Petar: I think that's right.

Mark: I too haven't been really following Richard Koo. But now I regret not having followed him. Not because I think he is right. But because I think he has maybe been very wrong and very influential. But I'm not 100% sure on that, simply because I haven't been following him.

Peter: you are right. It all depends on which rises more: nominal interest rates, or NGDP growth. But theory says, that for an economy in equilibrium, the higher the debt/GDP ratio the higher the rate of interest will be, for a given NGDP growth. So I'm a bit concerned that in Japan interest rates may rise more than NGDP growth as the economy returns to equilibrium. But again, all that means is that we should be even more upset that they didn't do Abenomics earlier, because the longer they leave it the higher the debt/GDP ratio will be.

[K]

Nick,

Your main argument here is that "It's not just what you buy, but what you buy it with." This is where both Richard Koo and Michael Woodford would disagree with you. Woodford is excruciatingly careful in distinguishing different types of "monetary policy": pure QE, Operation twist, forward rate guidance, and classical spot rate policy. He is very clear that the "what you buy it with" channel (i.e. pure QE) has no effect at the ZLB:

"A policy of creating additional reserves in order to purchase long-term Treasury securities can equivalently be analyzed as the composition of two policies: one under which new reserves are created by purchasing short-term Treasuries (“pure QE”), and another under which the central bank sells short-term Treasuries to buy long-term Treasury securities, with no change in the size of its balance sheet (“Operation Twist”). To the extent that a reserve-financed purchase of long-term Treasuries has effects at the zero lower bound, on the price of Treasuries or anything else, these effects should be identical to the effects of “Operation Twist” alone."

"I don't think Woodford would have any problem understanding that loosening monetary policy via a higher inflation or NGDP target would cause bond prices to fall, and that that would be a sign of successful loosening."

I think he would assume that it would have occurred via changes in expectations of central bank actions with respect to some effective policy instrument. Perhaps implied forward rate guidance (i.e. future short rate policy) or expected effective targeted asset purchases. A change in the target will not have any effect unless there exist some set of changes in policy instrument actions consistent with achieving said target. And changes to the monetary base at the ZLB is not such a policy action.

Koo is not a finance crank who differs in some radical way from standard macro theory. Not believing in pure QE doesn't set Koo apart from standard modern monetary theory. As far as I can tell, where he differs from Woodford is that Koo doesn't have any faith in the forward guidance channel, whereas Woodford clearly has *some* faith. I don't think that's a huge difference.

You didn't address how monetary expansion can effect the macroeconomy without any intermediating impact on inflation expectations.

If you are skeptical about Woodford's skepticism about the "what they buy it with" channel, you should read section 2.1 of his Jackson Hole Paper.

[K]

Obviously, "standard modern monetary theory" not = MMT above.

[K]

Also, by "expected effective asset purchases" I meant stocks, not long bonds.

[Nick Rowe]

K: whether Woodford and Koo agree on the ineffectiveness of "QE" at the ZLB is one thing. Whether Woodford and Kee agree on the relation between looser monetary policy and nominal bond yields is quite a different thing. My post here is about that second thing.

"You didn't address how monetary expansion can effect the macroeconomy without any intermediating impact on inflation expectations."

It can affect the macroeconomy *either* via an impact on expected future price level *or* via an impact on expected future real GDP. We could lump those two together into expected future NGDP, if we wanted to simplify. The division of higher expected future NGDP into higher expected P and higher expected RGDP depends on subjective expectations of the slope of the Short Run Phillips curve. I don't know what Japanese think about the slope of the SRPC. Moreover, as Ambrose says above, Koo is talking about "inflation concerns", so Koo seems to think that expected inflation has increased.

"Obviously, "standard modern monetary theory" not = MMT above." ;-)

Lorenzo: so both you and Mike are confirming my suspicions.

[K]

Nick,

I find it really tough to have a clear conversation when you use terms like "looser monetary policy" without being excruciatingly clear about what you mean. You definitely said above that it was all about *what* they buy it with. So I addressed that. Now I suspect that you are merely talking about changing the target, irrespective of any changes in the outlook for use of monetary *instruments*. I just can't tell what you mean. Its so easy to read and agree or disagree with Woddford since he says exactly what he means. Can't we be more clear?

[K]

Assuming that by "looser monetary policy" you just mean more base: do you mean

1) more base now
2) more expected base while at the ZLB
3) more expected base post ZLB
4) more expected base at ZLB for longer thus delaying exit from ZLB?

Also, most central banks pay IOR and so can determine the path of rates independently of the quantity of reserves. Is QE when money and tbills always earn the same rate "easing?"

[Nick Rowe]

K: asking a monetary economist for a clear definition of what he means by "looser monetary policy" is like asking a philosopher what he means by "truth", "knows", or "good", or asking a biologist what precisely he means by "species". You aren't going to get a simple answer. What you get is a discussion of different meanings and the pros and cons of each.

If I defined "looser monetary policy" as "the central bank lowering (a) nominal interest rate" that would be very clear, but also very stupid. That is the whole point of this post (or one of the points).

I would prefer something like "raising the target path of NGDP". But even that's not precise, because two target paths might cross, and so it's unclear which one is "looser". And an inflation targeter could reasonably object, and say it's raising the inflation target. And someone who prefered the gold standard would say it's raising the price of gold. Etc.

BTW, what do you think Woodford's definition of "looser monetary policy" is? (I don't think Woodford would define it as "lower nominal interest rates".)

[Nick Rowe]

And I don't think "more base" is a good definition either. Because if monetary policy succeeds in raising the time-path of NGDP, so Japan recovers, I think the demand for monetary base would fall, so the equilibrium time-path for monetary base would be lower, at least in some periods.

[Diego Espinosa]

The back up in JGB's may be a reflection of fear over a JGB selling cascade by banks and pension funds. This fear is a financial friction. It is like Godzilla standing between Chuck Norris's threat of higher NGDP and the banks' desire to make more loans. Godzilla makes banks and pension funds risk averse; he forces them to worry about their capital being stepped on and squashed by bond losses. As they cower, they rush to cut duration risk. Not in the direction of risk assets, but towards s.t. bills and FX instead. In Chuck's world, that FX demand just increases NGDP. In the world of financial friction, a disorderly Yen devaluation would lead to yet more JGB selling.

"Chuck just needs to kick Godzilla's ass", you might argue. So the BOJ pulls forward its already-massive QE to swamp bond-loss fears. Except markets see that the BOJ will not keep up that pace for long. They continue to sell. The BOJ faces a choice: promise to be irresponsible (buy at that pace no matter the inflation rate), or be prudent and back off Abenomics.

The difference between the Fed's and BOJ's QE is that the Fed harnesses financial frictions. Agency effects encourage bond QE front-running by banks and hedge funds: they earn out-sized bonuses on the capital gains. In Japan, the banks and pension funds are already "all in" the trade. Frontrunning would be dangerous to capital levels. So instead of frontrunning, they flee.

[K]

"What you get is a discussion of different meanings and the pros and cons of each."

Agreed. I wasn't really asking about the "looser" part. It was the "monetary policy" part I was questioning. I.e. are you talking about spot rates, quantity, forward quantity, forward rates, targeted purchases etc? If you just say "monetary policy is looser" I don't know which policy you are talking about. This is the part Woodford is so incredibly clear about in the Jackson Hole paper and it's really critical to having a productive conversation. So some monetary theorists *are* really specific about what they mean by "monetary policy" and what exactly they think each kind of policy can or cannot do.

"what do you think Woodford's definition of "looser monetary policy" is?"

While I think it's fairly easy to come with a good definition of "more optimal" I think it's tough to come up with a coherent definition of "looser." Any suboptimal policy that involves an asymptotically stable inflation rate will necessarily trade off excess consumption in some periods against too little consumption in others. I.e. any asymptotically stable policy cannot be characterized as unequivocally "loose" or "tight." It will always be loose in some forward periods and tight in others. I guess there may be policies that raise the asymptotic inflation rate which might *possibly* be qualified as a overall "loose", though they will always have periods of both excessive and inadequate output. Only runaway economies can truly be characterized as loose or tight.

So I'm hoping to sharpen the discussion with respect to the meaning of "monetary policy," not with respect to the meaning of "loose." In particular, I'd like to find out *exactly* how you think the "what you buy it *with*" channel "can raise the expected future price level, and/or expected future real income". In what period, and conditional on what states of the economy, will it change the incentives/behaviour of economic agents and how?

[Mark A. Sadowski]

Nick Rowe:
"yep, but I suspect that economists like Richard Koo have been much more influential in setting Japanese policy. And one thing I didn't mention in my post is that Richard Koo has, I think, been arguing that the BoJ should fight to keep its reputation as an inflation-fighter. So it's not just that he has ignored monetary policy as a tool to help Japan escape the recession, he has been advocating a counter-productive monetary policy. As long as people expected the BoJ would act to prevent NGDP rising, Japan stayed stuck.I was glad to see Nick Rowe bring up Richard Koo."

I was startled by the following passage that I found from a report Koo wrote in December:

"But nightmare scenario awaits when private loan demand recovers. The problem is what happens when private loan demand recovers. Loan books could grow more than tenfold in the US and five fold in Japan and Europe if bank reserves remain at current levels, triggering inflation rates of 500% to over 1,000%.

To avoid this outcome, central banks will have to mop up excessive reserves by raising the statutory reserve ratio, raising the interest rate paid on reserves, and selling government bonds. All of these measures will serve to lift interest rates, sending bond yields sharply higher and triggering a possible crash in the bond markets.

A sharp increase in government bond yields could lead to fiscal collapse in countries with a large national debt. For Japan, where the national debt amounts to 240% of GDP, the results would be catastrophic.

Expanding quantitative easing because it appears to be doing no harm is grievous error. Mr. Abe and his advisors may believe that all they have to do once their anti-deflationary policies succeed and JGB yields start to rise is have the BOJ buy more bonds. However, bank reserves under quantitative easing have risen to a level capable of fueling a 500% inflation rate, in which case the BOJ would have to sell, not buy, JGBs.

Nomura | JPN"

http://moslereconomics.com/2012/12/13/koo-on-reserves-time-bomd-500-inflation/

Yes, you read that correctly. Koo is predicting that the current Japanese large scale asset purchases have the potential to create 500% to 1,000% rates of inflation, a crash in the bond market and a complete fiscal collapse.

I think it's high time that people realize that the Emperor Koo has no clothes.

[Mark A. Sadowski]

I'm somewhat puzzled how "I was glad to see Nick Rowe bring up Richard Koo." got appended to my quote of Nick Rowe (evidently sloppy cut and paste) but it's true in any case.

[Nick Rowe]

Mark: that bit from Koo is deeply weird. Sure, recovery will cause interest rates to rise. And *if* interest rates rise by more than NGDP growth (they might) that will make it harder for Japan to service the national debt. But "catastrophic" seems a little bit on the strong side. Furthermore, the *reason* Japan's debt/GDP ratio is so high is precisely because it has been doing what Koo wanted all these years, rather than doing something like Abenomics a lot earlier. Furthermore^2, is Koo saying that it is now far too late for Japan ever to escape its ZLB prison, and must stay there forever, for fear of fiscal catastrophe? God help us!

[Nick Rowe]

I don't normally question people's motives. Because it isn't very productive. But in this case... does he work for people who are invested up to the hilt in long Japanese bonds, and who can't cover their positions, for some reason?

[DorganG]

"This is not the "beginning of the end" of the Japanese economy. But it might be, and I hope it is, the beginning of the end of Japan's long recession.

And if this is indeed the beginning of the end of Japan's long recession, it will also be the beginning of the end of Richard Koo's thesis that monetary policy is powerless in a balance sheet recession, and that only fiscal policy can offset private sector deleveraging. And we can only regret that Japan did not do this many years earlier, instead of wasting all those years and letting Japan's government debt/GDP ratio climb. Because that high debt/GDP ratio is the only reason why someone might want Japan's economic recovery but not want the higher interest rates that will accompany that recovery. Which is no reason to try to stop the recovery. Though it is one additional reason to regret not having done something like Abenomics a lot earlier."

The strengthening of the Japanese economy in 2009/2010 despite the global recession shows that the Japanese balance sheet recession was already finished in 2010. One main reason was that Japanese investors repatriated their foreign investments. This has not changed in 2013, many Japanese investors sold their foreign bond holdings. The devaluation of the yen has further fueled potential inflation pressures in an economy that was not concerned by euro crisis or the long-term US weakness. Apart from demographic issues, recent deflation in Japan was caused by global slowing, the hedonic effect of a technology nation and by things like lower student or scholarship fees.
The balance sheet recession had already finished before and this is the problem of Abeconomics: Addressing a problem that does not exist anymore.

[Charlie]

K,

Inflation expectations have risen quite a bit in Japan. Both on news about Abe's election chances and on the big QE announcements.

http://badoutcomes.blogspot.com/2013/04/inflation-expectation-in-japan.html?m=1

Here is various looks at inflation indexed bonds. Expectations have risen since. I haven't updated the data, but Lars Christensen has links on his blog.

[Philo]

“. . . expectations of recovery will cause a rise in investment [and] a fall in saving . . . .” But Scott Sumner assures us that investment = saving. Did you, perhaps, mean, “a fall in *the hoarding of base money*” (rather than "a fall in saving")?

[K]

Charlie,

Right, which is exactly my point. Japanese inflation expectations do move in line with economic news. They moved up on the election and they have followed stocks up in the last few months. The one announcement they did *not* follow was the huge QE announcement on April 4. What's being missed is that "Abenomics" is also a huge change in fiscal policy. Monetarists are wanting to attribute all the progress to monetary policy but it's really principally about fiscal policy. If monetary policy was relevant inflation expectations would have moved on the one announcement that exclusively represented a huge change in the expectation of BoJ policy (see the huge reduction in bond yields on the announcement). They did not.

[Charlie]

K,

The 1 year break even rate rose 12.5% on April 4th from .2 to .225.

My interpretation is that QE works through the expectations channel. The market was already convinced the BOJ was pretty credible. QE signaled something about the timing the BOJ was aiming for. That's why the 5 year and 4 year moved much less.

What is your interpretation of the bond market movement? I think the day 1 effect was mostly a liquidity effect.

[Charlie]

K,

When Kuroda was appointed an announced a two year two percent inflation target on March 19th, the two year break even went up 270%. What fiscal policy do you attribute that to?

[rsj]

I think it's funny to

a) talk about movements of values very close to zero in percentage terms
b) attempt to interpret brownian motion in terms of what political announcement was made on the day the motion occurred if the motion goes in the right direction.
c) attempt to interpret brownian motion in terms of already pricing in the announcements if it goes in the wrong direction.

[rsj]

I think a much better gauge would be to see whether sustained inflation is generated on the ground.

Japan has in the past seen isolated quarters of small positive inflation, but could not sustain it.

Let's see if Japan can see 4 consecutive quarters in which inflation is 1% or higher, no? This will happen sometime in our lifetimes, at which point we can argue about how much of the credit should be attributed to fiscal policy versus monetary policy versus some other structural adjustment.

[K]

Charlie,

The Kuroda appointment was also the end of Shirakawa. Removing Shirakawa significantly reduced the risk of premature rate hikes, which is a big deal, and therefore the raised inflation expectations.

"The 1 year break even rate rose 12.5% on April 4th from .2 to .225."

Like rsj says, 2 basis points??? I was watching the break evens moving intraday and the announcement was literally undetectable. Volatility was the same as on any other day.

"QE signaled something about the timing the BOJ was aiming for."

OK, but timing of what? First rate hike? If so, why not just tell us when and contingent on what, they are going to start hiking? Why do we have to go through this elaborate ritual QE dance?

Apart from that, I totally agree with rsj. Lets see some actual signs of sustained *current* economic vitality before starting the victory dance. That applies to all stimulus measures, fiscal or supposed monetary.

[Charlie]

rsj,

1. Would it be better to say the contract was at 20 basis points and increased 2.25 basis points, a 12% increase? You are getting confused by scaling, this is what percentage changes are for.

2. I think it's funny to say we should look at market movements to distinguish monetary from fiscal policy, and then when someone points out that your argument is factually incorrect to switch to saying we shouldn't try to interpret market movements.

3. There is nothing Brownian about the 270% rise in the two year over three days. You can model it as a random process, but you'll need jumps or stochastic vol or something.

[Charlie]

"The Kuroda appointment was also the end of Shirakawa. Removing Shirakawa significantly reduced the risk of premature rate hikes, which is a big deal, and therefore the raised inflation expectations."

In what world is this news about fiscal policy? This is the basic definition of monetary policy. Changing the time path of future inflation expectations.

"OK, but timing of what? First rate hike? If so, why not just tell us when and contingent on what, they are going to start hiking? Why do we have to go through this elaborate ritual QE dance?"

I think the BOJ could accomplish all of their goals through a 2% level target, but that is a quite radical view, especially among central bankers. Look at the reaction Woodford got at the last IMF meetings. He was way out of step with the other panelists and then the questions reinforced that he was way out of step with the audience.

"Apart from that, I totally agree with rsj. Lets see some actual signs of sustained *current* economic vitality before starting the victory dance. That applies to all stimulus measures, fiscal or supposed monetary."

Some of us believed strongly that 1) monetary policy can raise inflation expectations 2) a rise in inflation expectations will increase real growth. Many economists think both views are wrong. I think it's perfectly reasonable to narrate our model in real time and explain the different predictions and interpretation it makes. The idea is to think through your model ahead of time, hear competing views, and let the data settle the dispute (or at least alter the posteriors).

rsj and K,

If you both think we should sit and wait, why are you commenting. I signed on to point out an error I saw repeated in K's comments. I really have done much, but point out errors. I don't recall ringing the victory bell.

[Charlie]

rsj,

"Let's see if Japan can see 4 consecutive quarters in which inflation is 1% or higher, no? This will happen sometime in our lifetimes, at which point we can argue about how much of the credit should be attributed to fiscal policy versus monetary policy versus some other structural adjustment."

1) The only way we will be able to distinguish between the two is to look for news events that represent regime changes (policy shifts) and analyze their effects. We'll argue about what the right news events are, but it's exactly the right method. This is why K was pointing to a monetary policy news event to begin with.

2) Is it your view that the inflation expectations are wrong and Japan can't have positive inflation?

[K]

Charlie,

"In what world is this news about fiscal policy? This is the basic definition of monetary policy."

Absolutely. This was an important monetary policy event. My only real point here is that there is no evidence whatsoever for the effectiveness of QE. I'm all for rates policy including forward guidance.

The point about the victory bell wasn't really principally for you. There are a lot of people trying to claim a victory for QE in Japan. I don't think they have a leg to stand on, and I'm commenting because I think it's important to protest against nonsense before it becomes established "truth".

"Some of us believed strongly that 1) monetary policy can raise inflation expectations 2) a rise in inflation expectations will increase real growth."

You are using the term "monetary policy" without specific reference to what policy you are talking about. There are many different types of policy that are commonly called "monetary" (see my comments to Nick above). This post is about the impact of monetary *expansion*, ie what Woodford calls "pureQE". Like I said, I'm all in favour of rates policy.

[Scott Sumner]

Nick, Very good post, but one quibble. At the end you seem to suggest that future events will help us better understand whether Koo was right in claiming that the BOJ was incapable of stimulating the economy. That's not really true. The claim also applied to the value of the yen. And that's been clearly demonstrated--the BOJ has driven the yen sharply lower. Thus there is really nothing more for us to learn. There is no liquidity trap in Japan, and there never was. Koo was wrong, and Krugman was half wrong (the 1/2 of the time when he suggested they were in an expectations trap.)

Instead we'll learn whether Japan's problems were supply-side or demand-side.

[Max]

This is a little off topic, but one fairly clear effect of Abenomics is increased volatility of the Japan stock and bond market and exchange rate, presumably reflecting greater economic uncertainty.

Was a lack of uncertainty holding back the Japan economy?

More seriously - although volatility is usually related to bad stuff happening, this is not always the case. People who attribute macro problems to "uncertainty" are confusing symptom with cause.

[Nick Rowe]

Thanks Scott. Fair point. If Koo's original position was that monetary policy could do nothing at the ZLB, it seems he was wrong, and has already shifted his position. He now seems to be saying that monetary policy is to blame for the increase in expected inflation and higher bond yields. But I think it will take an increase in both actual inflation and RGDP to put the final nail in Koonomics. Because only then will people say "Why did we wait this long to do Abenomics?", and start asking who it was that was holding Japan back.

Max: yep. "certainty" in Japan would mean "more of the (bad) same". Any new policy that might work and might be implemented must increase uncertainty.

[rsj]

Charlie,

The only way we will be able to distinguish between the two is to look for news events that represent regime changes

No. The reason why QE is impotent is because you are conducting asset swaps with the financial sector while leaving household deposits unchanged. The fact of the matter is that there has never been a "hot-potato" effect with regards to household deposits or currency, as they quantity of these is demand determined.

The way you distinguish between models is not by reading news headlines, but by assessing the model's accuracy in making out of sample predictions.

As there are many models often making similar predictions, you are left to build a case for the model. Those who are all-in to the old monetarist school, like Nick, are not going to change their minds. They will always argue, when current policies don't work, that "more is needed".

That was the same argument for price controls.

It is the younger generation who needs to decide which model is more convincing.

For the quantity type views, which rely on debunked fixed ratios between reserves and deposits, the mainstream has already rejected Nick's view. As have the central banks, when they abandoned attempts at setting targets for monetary aggregates, as the higher monetary aggregates are demand determined and not under the central banks control.

This is why the mainstream does not share Nick's view, they share Woodford's view. It has nothing to do with reading headlines. The next generation will have some other view.

The current fashionable view is that even though QE does nothing to increase the quantity of deposits, nevertheless it serves as a type of signal that the CB is "serious" about keeping rates lower than necessary during the recovery.

As I do not believe that QE constrains future monetary policy in any way (monetary policy is just interest rate policy), I do not subscribe to this view and therefore believe that QE will continue to fail to generate inflation in Japan. Japan will be stuck in a liquidity trap until the underlying problems are worked out, regardless of what the current mainstream believes, or what the previous generation of monetarists such as Nick believe.

I could well be wrong, but this is not a matter to settle by reading news headlines, but by analysing out of band predictions and the overall plausibility of any given model.

[Charlie]

"Japan will be stuck in a liquidity trap until the underlying problems are worked out, regardless of what the current mainstream believes, or what the previous generation of monetarists such as Nick believe."

Most mainstream models are constrained by some sort of rational expectations. Thus, there is some sort of market efficiency imposed. If we don't have market efficiency, you are right, the changes in inflation expectations evident in the yen, stock and inflation indexed bond markets are meaningless, just the whims and fancies of the market. I agree that in such a world, we do have to wait to see out of sample things like GDP and prices. For many models actual market movements on news are out of sample, since most macro models make testable predictions about how financial markets will react to policy changes, but I can see how with your views that you would not see it that way.

Of course, if you are right, you can make a lot of money over the next two to five years. I hope you are not.

[rsj]

Charlie, the financial markets are forward looking, so they can be both rational and yet exhibit brownian motion. It is *insane* to look at bond market performance over small time periods and try to conclude something about policy from this.

I remember -- do you remember? -- when there was a rise in US yields that brought out all the "bond market vigilante" talk. Of course that was just a blip and a few months later rates went back down. Now, in retrospect, the original bond market vigilante appearance was also random noise. There has never been an appearance of bond market vigilantes on U.S. soil, it can all be explained as random noise. The financial media spins a lot of narratives about this.

The above does *not* mean that I can make a lot of money by attributing to brownian motion what others attribute to policy reactions. I can only say that, ex post, longer term yields in Japan will remain low because the short rate will remain at zero.

Whether "low" means 10 b.p for a 1 year bond or 20 b.p. for a one year bond is noise.

On other blogs, commentators are achieving orgasm because 10 year yields in Japan are at 0.9%. 0.9%! How on earth is borrowing money for a decade for 0.9% inconsistent with a long run deflationary scenario? Don't you think that longer term yields might be a bit higher than 0.9% if the market expected inflation to be north of 1%?

One can argue that even 2% is a low yield for a 10 year bonds, and is completely consistent with an expectation of 0% short rates for many years.

As to TIPS break even rates, you have to deal with liquidity premia, which is another way of saying that the TIPS breakevens tend to overestimate inflation by about half a percent, although this varies, with this residual being defined as a "TIPS premium". More information here: http://www.frbsf.org/publications/economics/letter/2011/el2011-19.html

Univ. of Michigan inflation surveys also tend to overestimate inflation and tend to assume that inflation will be 3% or so.

[rsj]

OOps, was late about JGBs. Current (as of 5/24) JGB 10 year yields are 0.82%. This time last year, they were 0.87%. I see no evidence that markets are expecting sustained periods of positive inflation over the next decade, or that anything meaningful has happened in outlook from today versus one year ago.

[Charlie]

"As to TIPS break even rates, you have to deal with liquidity premia, which is another way of saying that the TIPS breakevens tend to overestimate inflation by about half a percent, although this varies, with this residual being defined as a "TIPS premium"."

You have it backwards. The liquidity premium causes TIPS to understate inflation.

Here's why:

Treasury = r* + IP
TIPS = r* +LP

Treasury - TIPS = TIPS-based inflation = IP - LP, thus you need to add the liquidity premium to the break even to get true expected inflation. Actually, the link you provide has it correct, maybe you should read it. Perhaps, you have confused it with the inflation risk premium. Usually that biases the break even up, though it's smaller than the Liquidity premium. In Japan currently I wonder it the inflation risk premium is even negative.

"the financial markets are forward looking, so they can be both rational and yet exhibit brownian motion."

Of course, this is true. I'm not sure how you misunderstood. I was arguing that the 2 to 5 year breakevens show positive inflation expectations (and of course they are biased down). Thus, your statement that inflation will be zero for the considerable future is not consistent with the market. Either you are wrong or the market is inefficient.

"I remember -- do you remember? -- when there was a rise in US yields that brought out all the "bond market vigilante" talk."

Now you write a lot about how some people said stupid things about market movements, so no one should ever say things about market movements. Actually, during all that bond vigilante talk, CDS spreads on Treasury debt were down (and stocks were up). The evidence was much more consistent with a positive outlook for the US economy.

[Charlie]

"Don't you think that longer term yields might be a bit higher than 0.9% if the market expected inflation to be north of 1%?"

Almost all of the rise in inflation expectations has been a fall in real rates. Very little (about .2 - .3) has come from a rise in nominal rates. That is quite consistent with Japan being stuck at the ZLB and the market clearing real interest rate being negative.

You seem to be looking at one price in isolation (nominal bonds), which is always the wrong thing to do. Any model can explain one price movement. The way to test models with financial markets is to look at comovements. How do you explain the rise in Japanese stocks and the fall in the yen? For instance, stocks can go up if future dividends are expected to be higher, future inflation is expected to be higher or the risk premium is lower. Even if the movements were completely unpredictable, they have to be the result of shocks to one of those three things (in an efficient market).

Lastly, I hope you do also realize that the underlying measures are stochastic as well. Even if actual (came from God) Japanese expected inflation over the next two years is 2% today, it could still end up being zero in two years. The underlying measure of inflation is, itself, stochastic. Japan could be hit by random shocks that cause prices to be lower than the mean. It seems quite arbitrary to worry so much about the random noise in markets and not at all at the random noise in data.

[K]

Charlie,

There is no liquid repo market for Japan real return bonds, so the price is basically set by consensus of the longs. I.e. I think there is good reason to suspect that they trade at a premium to an efficient market equilibrium because there is no way for me to express my opinion that they are overpriced.

As to the inflation risk premium, I tend to agree with you that there are good reasons to suspect that it might be negative. But I'm wondering what is your reason for believing that?

[rsj]

I believe my comment was swallowed

[rsj]

Charlie,

I would recommend the paper by the newyorkfed about measuring inflation expectations in Japan. I can't post the link (spam filter), but google for it. Once someone unswallows my comment, you will see it.

With regards to nominal rate-based analysis always being a bad idea, and the market clearing rate being negative, I don't think this is relevant. If the 10 year rate is 0.82%, then your expected average of short rates over the next 10 years is zero. That means that the market does not believe that Japan will be able to escape the liquidity trap. I.e. if Japan could really control inflation with the current policies, it would set a level of inflation sufficiently high so that regardless of how negative the real rate was, short rates would be above zero. The market would expect this and price long rates higher.

Therefore there is no need to look at (unreliable) inflation swap spreads, we can look at long bonds directly to determine whether the market believes that Japan can generate inflation with current policies.

[paule]

K - smartest guy in the room. By a country mile.

[Charlie]

"But I'm wondering what is your reason for believing that?"

The states of the world with higher inflation in Japan are likely better than the states of the world with low inflation in Japan. Thus, investors will want to hedge low inflation states relative to high inflation states, so the inflation risk premium is negative. In normal times, in well functioning economies, the opposite argument holds.

"then your expected average of short rates over the next 10 years is zero."

The short rate can be 0 and inflation can be 2%, because the real short rate is negative. How can they be in a liquidity trap, if they can raise expected inflation.

"if Japan could really control inflation with the current policies, it would set a level of inflation sufficiently high so that regardless of how negative the real rate was"

How can you live in the world of the last 5 years and not believe suboptimal monetary policy exists or at least that it could exist in theory? Moving to a 2% target was radical as is. The Yen has depreciated 25% since November. What if the BOJ traded 150 yen for every dollar, still think they couldn't create inflation? Why don't they just buy up all of Japan's outstanding debt? Then the BOJ could forgive it, and Japan would never have to pay it back. They could buy other assets and just fund their government off the interest payments. No need for taxes. If Japan was really in a liquidity trap, wouldn't they jump on this free lunch?

[rsj]

Charlie,

Lots of confusions between monetary and fiscal policy in your last post. Hint, "forgiving" debt, or purchasing risky debt is not monetary policy, and the BoJ cannot do it anymore than it can do helicopter drops. The Fed can't do this, either.

But these types of category errors are to be expected from those who believe that monetary policy is the key to managing aggregate demand (it never was) in a ZLB context. They immediately switch to advocating fiscal policy (e.g. giving money away or debt forgiveness), but call it monetary policy.

Absolutely with sufficient fiscal policy, Japan escapes the liquidity trap, and lack of sufficient fiscal policy is the main reason why Japan has remained in this trap. Debt forgiveness, wealth transfers, more social insurance, giving people money -- yes, all of that is needed, and none has anything whatsoever to do with monetary policy or what any central bank is allowed to do.

[Nick Rowe]

rsj: since the government of Japan owns the Bank of Japan, and (sooner or later) gets all the BoJ's seigniorage profits, the link between monetary policy and fiscal policy is impossible to escape. If the BoJ bought all of the government debt, and held it permanently, it would be exactly as if the BoJ then "forgave" that debt. Because all the BoJ is saying is that it makes no sense for the government to hand over billions of yen to the BoJ every year for interest on debt that the BoJ then hands straight back to the government.

[rsj]

Nick, you need to be a bit more precise when you say "link". The issue is one of authority and responsibility. Yes, the Japanese treasury owns the BoJ and collects all bank profits and pays for all bank losses. Therefore central bank actions have fiscal consequences for the Japanese government -- there is a link. Because of this link, governments keep their central banks on a short leash and do not authorize them to bear a lot of risk. Any asset purchased by a central bank is effectively guaranteed by the treasury, so the set of assets that Treasuries allow central banks to buy is (primarily) limited to assets already guaranteed by the treasury. The link explains why central banks are kept on a short leash so that fiscal policy remains the purview of elected officials rather than bank appointees -- because the elected officials prefer to keep this power to themselves and they are the ones calling the shots vis-a-vis what the central bank can do.

A notable exception would be the EU, where the ECB is trying to call the shots, telling legislatures what fiscal measures they can and cannot take. We see from this how grotesque such an inversion of control is, which should explain why other nations do not have such arrangements.

[rsj]

But if the larger point is that fiscal measures are the key to re-igniting healthy levels of inflation, and out of ideological purity, you will only agree to undertake such measures if you call them "monetary" -- e.g. redistribution of wealth, income guarantees, etc -- then what the hell, call the measures monetary policy if you want, as long as the measures are undertaken. But no, forcing banks to hold a few more reserves isnt going to do anything for inflation, because goods and services are still too expensive in Japan for the middle class to increase consumption spending on them. Japan is a very expensive country, with median wage shares of GDP falling more rapidly in Japan than in any other OECD nation. As long as these wage repression policies remain in place, you are not going to ignite inflation regardless of how many excess reserves the BoJ is forcing on the financial sector.

A good comparison nation here would be S. Korea, which is also an export oriented nation, is also resource constrained, and yet unlike Japan has a rich history of union actions, strikes, and rapidly rising wage shares. Korea, paradoxically, is not stuck in a ZLB environment and faces positive levels of inflation. The difference between Korea and Japan is really striking.

[Nick Rowe]

rsj: the "link" I'm talking about here is simply accounting. The central bank's monetary policy affects the central bank's profits, and those profits are (sooner or later) given to the government, and the government (sooner or later) spends those profits (by increasing G and/or cutting T), which is fiscal policy.

[rsj]

Nick, yes, but I'm making the point that because CB P & L shows up on Treasury's balance sheet, none of the policies mentioned by Charlie or, in other cases, yourself (such as helicopter drops or private equity purchases), can be undertaken by Central Banks because of their fiscal consequences.

But too often, I get the sense that monetarists believe everything to do with "money" is controlled by the central bank, while everyone else is exchanging CPI baskets.

That isn't the case in a monetary exchange economy.

IMO a more correct dividing line is that nominal income flows are affected more by fiscal policy whereas interest rates are affected more by monetary policy. Yes, there are links between these, but to the degree that there is overlap, it is the fiscal authority that needs to authorize/allow the monetary authority, and the monetary authority limits the overlap as much as possible in deference to the fiscal authority. ECB is the reverse.

[Nick Rowe]

rsj: then since *all* CB actions will affect their profits and losses, which will show up on Treasury's balance sheet, you are saying that *no* CB actions can ever be undertaken.

[rsj]

No, I didn't say that. I said it was about authority and responsibility. The government has given the CB responsibility for interest rate management and this effects seignorage income, but CB can do it. But paying interest on reserves, for example, required an additional act of congress because this would affect seignorage income as well. It remains the case that CBs cannot take losses, they have to buy primarily guaranteed debt. That is a big difference between some of the conversations here, where CBs can do whatever they want as long as it somehow involved money (e.g. charge negative interest rates, purchase shares of private businesses, etc.)

[Charlie]

Nick is of course right.

How about this then: the BOJ buys all of the outstading Japanese debt, then all of the outstanding US debt, then all of the outstanding Euro debt. Each year it will receive a tremendous amount of interest payments, which it will remit to the Japanese gov't. Everything still works and no laws are broken!

Also, you made quite a few statements that are just factually wrong, for instance, " so the set of assets that Treasuries allow central banks to buy is (primarily) limited to assets alreadyguaranteed by the treasury. "

The BOJ holds commercial paper, stock and real estate etfs, corporate bonds as well as gold and FX. They can certainly buy more as well. All of this info is publicly available.

http://www.boj.or.jp/en/statistics/boj/other/acmai/release/2013/ac130531.htm/

[Noah Smith]

1. How come WCI always gets more comments than my blog? WAAAAAHHH.

2. In a standard New Keynesian model, the interest rate would not rise immediately as a result of expansionary monetary policy.

[Nick Rowe]

Noah:
1. Last time I checked, your blog posts seemed to be getting lots of comments!

2. I'm trying to work out whether you are right on that, in all cases:

Suppose the central bank announced an increase in the inflation target. That would certainly increase future nominal interest rates in a standard NK model. Even if the 1-period nominal interest rate fell, the current multi-period nominal interest rate would rise (if there are enough periods). We are talking 10 years bonds here.

Suppose we modified the standard simple NK model, so that there was a lag between the loosening of monetary policy and the increase in real output. (It takes time for firms to implement new investment projects.) The increase in expected future real income, relative to current real income, would increase the current equilibrium real interest rate. You can think of this as the loosening of monetary policy causing the IS curve to shift right, as well as causing a movement down along the IS curve. Though I prefer to think of it as an IS curve that slopes up, so that the real effects of a loosening of monetary policy causes an increase in the real interest rate at which desired S = desired I.

[Noah Smith]

I'm pretty sure that rates are lower with the monetary expansion than without...I think that just follows from everything being concave...

[Noah Smith]

Also, see this: http://noahpinionblog.blogspot.com/2013/06/the-zero-upper-bound.html

[Nick Rowe]

Noah: It's pretty easy to get *nominal* rates being higher with the monetary expansion. Just assume, quite reasonably, that monetary expansion implies higher expected inflation. And if monetary expansion leads to higher expected growth of real income, you could get higher real rates too.

I always get concave and convex muddled. Plus it's late here, and my brain is tired. But I think I see where you are coming from there. But here's another way of thinking about it: assume people expect the economy to be in recession next period. Now the BoJ announces this period that it will loosen monetary policy next period, and people think this will cause the economy to recover next period, and so they now expect both a higher price level and higher real income next period. here's my old post on upward-sloping IS curves

I had a quick skim of your post. Will give it a more careful read, and try to respond.

[PierGiorgio Gawronski]

JP is an open economy. Long term interest rates are determined through the uncovered interest parity condition (corrected for the liquidity trap). So suppose that something (fear that political pressure form abroad - or monetary intervention/easing abroad) is temporarily blocking half way the yen depreciation. Then nominal long term interest rates will rise above what you suggest (the level apt to accomodate an improved outlook for growth), and most likely real interest rates will rise too, and too much.
JP would be better off by announcing that the short term $/Yen exrate target is 120 (so as to reach it overnight) and that from there on they expect a very very slow appreciation. This announcement would cause a Yen collapse (to 120), but long term interest rates will stay low 'forever'.
Regards.

[Nick Rowe]

PierGiorgio: OK. If monetary loosening causes an expectation of ongoing real exchange rate depreciation, that would tend to force up Japan's real iterest rates. But my sense is that the real exchange rate has already depreciated, and from now on there will be expected appreciating. So the effect would go the other way. Trying to remember how overshooting models apply in this case....

[Nick Rowe]

But more generally, yes, the BoJ/Abenomics communications policy leaves much to be desired, in telling people they were trying to push interest rates down. It would have been better, IMO, if they had instead said they were trying to increase future NGDP.

[Ashok Rao]

Nick, you have a – well – knack for great analogies. The bicycle really should help the confused understand Abenomics. I have a response to DeLong in which you might be interested (I think we're striking the same point from a different angle). I discuss this in full here – http://bit.ly/165liic – but this is a concise comment I left for DeLong:

"1. No reason to believe that the wedge between public-private yields wouldn't happen where both fall (in real), just private more than public. But even if you are correct, lower private cost of capital implies higher wages and capital income, and definitely higher profits. This makes debt more sustainable through tax revenues. The effect of each may not cancel out, but it can't be perceived so simply.

2. Money non-neutrality in the medium-run doesn't require necessarily Keynesian assumptions, just the relaxation on an exogenously-set time preference. Endogenizing this adds a whole lot of problems, but there are descriptive and agent models that support non-neutrality where time preference is a positive function on real financial wealth. The models which maintain neutrality with an endogenized time preference make really weird assumptions. Becker and Mulligan (1997) show ways in which this can be endogenized.

3. If we're talking about "unsustainable debts" – and we are – more "risk tolerance" can also imply an increased demand for Japanese public debt, since the premise itself is "risky", it seems like a self defeating question. Or maybe my point is circular here?

4. Japan should move to a 4% rather than 2% interest rate. Mundell-Tobin talks within the context of a one-time increase in anticipated inflation expectations. If Japan moves to a 2% and then 4% it looses credibility. But if BoJ goes hard at 4% today, it increases the duration of the short-run (probably) and thereby reduces the probability of tail risks in your post.

All said and done, I really like the Mundell-Tobin "descriptive" way of looking at things (though talk about Cam 2000 optimizing model to the same effect). Inflation decreases financial wealth, other things equal, which decreases money demand, increases velocity and hence encourages capital creation and a fall in real interest rates.

I worry however that increases in velocity will have funny, nonlinear effects under an inflation-targeting system with current expectations. Interested in all of your thoughts.."

A followup on that last point is the adoption of a healthy NGDP target. And another worry that emerges is the international considerations of Mundell-Tobin esque analysis. Though, for an economy as large as Japan's, I don't know if it would matter.

[Nick Rowe]

Ashok: I think I'm following you.

Sidrauski's model of long run superneutrality is my "benchmark" case. But sure, money can't be precisely super-neutral, if currency doesn't pay interest, and so higher inflation will cause lower real currency balances, and this can in principle affect lots of other real variables, in various ways. The Mundell-Tobin effect is one of those ways.

But my gut says that the Mundell-Tobin effect isn't big enough to matter much, unless we are talking very large changes in inflation for countries where people use a lot of currency. Multiply currency/NGDP times the interest elasticity of the demand for currency times the change in inflation. That's three smallish numbers multiplied together. The capital/GDP ratio will be an order of magnitude bigger. The potential for debt/GDP to crowd out capital (which is essentially the same mechanism as Mundell-Tobin) would be bigger. Not to mention the non-neutralities from nominal rigidities.

[Ashok Rao]

Nick, I think we principally agree, then. However, to the extent a small Mundell-Tobin effect is working, wouldn't the risk that debt/GDP to crowd out capital significantly fall? Right now, at or near the zero bound that is obviously not the case. Inflation should just erode that, so the Mundell-Tobin would pick up over time, and the risk of crowding out should fall. If our initial premise that r < g is correct. The extension here is that it *stays* correct until debt becomes fundamentally manageable.

In some sense, the question here is protecting against tail risks, and the way to do that would be to make the short run, perhaps non-neutral, period as long as possible. The best way to do this is to maximize change in inflationary expectations: target NGDP or 4% inflation (I mistakenly said 4% rates).

But let's say we're in DeLong's medium run where expected real yields are increasing. By definition we're no longer at the zero lower bound, which means an unexpected monetary base expansion can also trigger Mundell-Tobin like effects (without the contingent "credibly commit to irresponsibility" clause we have today) and provide the economy breathing room, though it seems we both doubt this is very likely.

All said, however, the number of feedback loops in this mechanism *does* challenge my confidence..

[PierGiorgioGawronski]

You may be right about the future, but I doubt there were many people willing to buy JP bonds in the last 2 months, unless they were offered a discount, while the currency was consistently plunging and the authorities were cheering at the plunge asking for more. Anyway I agree with your broader point & Outlook. Good post! (By the way, could you send me an email address where I could send you an original idea, and ask for your opinion? It's only 795 words.)

[Nick Rowe]

PierGiorgio: Nick dot Rowe at Carleton dot ca. No promises I will read it and respond though.

[PierGiorgioGawronski]

This email address does not work right now. But I see it's the same you use at your university so I'm puzzled.

[Nick Rowe]

Hmm. I just tried it, and it worked for me. Did you remember the e in Carleton?

[PierGiorgio Gawronski]

Yes I did. I sent you now the mail from another server, hope it works. Lmk.
G