Paul Krugman poses a very difficult question

Here’s Paul Krugman:

A question (to which I don’t have the full answer): why are the interest rates on Italian and Japanese debt so different? As of right now, 10-year Japanese bonds are yielding 1.09%; 10-year Italian bonds 5.76%.

I ask this because in a number of ways the two countries look similar. Both have high debt levels, although Japan’s is higher. Both have awful demography. In other respects, the numbers if anything favor Italy, which has a much smaller current deficit as a percentage of GDP.

So what’s going on? I normally argue that members of the euro zone that have excessive costs “” which certainly includes Italy “” face a straightjacket in the sense that they will be forced to go through a period of grinding deflation to restore competitiveness. But while Japan has its own currency, it’s suffering from its own deflation all the same.

What is true is that the Bank of Japan is keeping rates at zero, while the European Central Bank seems determined to raise rates. Is that enough to explain the difference? Or is it something about the absence of a proper lender-of-last-resort function?

I don’t have a completely satisfying explanation, but I might have one or two insights that would help a bit.  And I can’t resist a challenge.

Let’s start from the fact that Italian bonds are euro bonds, just like German government bonds (albeit with a higher default risk.)  Let’s also assume that both Japanese and German government bonds have virtually no explicit default risk (although obviously there is some risk of inflation that reduces the real value of the principal and interest.)

As you know, I like to argue that long run nominal interest rates are closely linked to long run NGDP growth.  This is because long term real interest rates are usually pretty close to long run real GDP growth.  (Less true over the business cycle.)  In Japan, NGDP is almost unchanged since 1993, whereas in Europe, NGDP tends to grow at closer to 3% or 4% a year.  Thus it would not be at all surprising if the risk-free rate were substantially higher in Europe.  Indeed, the current yield on 10 year German bonds (2.7%) is only 1.6% above the current yield on Japanese 10 year bonds.  So there’s no big mystery there.

Furthermore, the fact that the gap between German and Japanese long term rates is currently smaller than the gap between trend rates of NGDP growth, suggests that Japanese bonds may be viewed as “riskier” in a inflation sense, not a default sense.  Thus because of the big debt and deficits that Krugman points to, it’s possible that investors are more afraid of the Japanese government resorting to inflation, then the German government (actually the ECB) resorting to inflation.  So I don’t find any great mystery in the yield spread between German and Japanese government bonds.

The second part of the puzzle is explaining the gap between the yield on German and Italian euro bonds.  And in one sense that’s easy to explain, it must be default risk.  What else could it be?  And given how rapidly things have deteriorated on the periphery of Europe, I can’t really argue that a 3% premium due to default risk is unwarranted (although I’d actually prefer the Italian bonds right now.)

And yet I can’t help feeling this explanation was a bit too glib.  For one thing, I’ve never been able to reconcile my view of nominal interest rates being linked to NGDP growth with the interest parity condition, combined with the Balassa-Samuelson effect.  Those two theories suggest interest rates should be higher in slow growing economies, to compensate for the B-S effect leading to currency depreciation in slow growing economies.  Second, I’m not sure my argument really explains the size of the gap between Italian and Japanese yields.  After all, doesn’t it imply that if Japanese conditions are as bad as those in Italy (as Krugman suggests) then Japanese risk premia should be just as high (3%.)  But that means a theoretical (inflation and default) risk-free Japanese bond should offer a negative nominal yield, if somehow the Japanese government could convince investors that there was no inflation risk.  I’m not sure what it means for a theoretical non-default risk, non-inflation risk bond to have a negative yield, but I’m guessing the answer is to be found in It’s Baaack.

Perhaps the disaster feared by investors in Italian bonds is much scarier than the disaster feared by investors in Japanese bonds.  Consider that when a crisis forces a devaluation, the devaluation is almost always extremely large (relative to nominal interest rates.)  We are talking about 25% to 50% in a single day.  In contrast, if things in Japan get worse, it may be that several years down the road the BOJ is forced to produce modestly higher inflation rates.–with much of the future inflation coming after the 10 year bond matures.  In Italy, the crisis could come very fast, and quickly reduce the (euro) value of 10 year bonds.  Italy is also paying a price for the contagion effect, which seems linked partly to whether countries are in the euro, and partly to the actual state of their budget finances.

Finally, I’d like to give a slightly different take on this comment by Krugman:

What is true is that the Bank of Japan is keeping rates at zero, while the European Central Bank seems determined to raise rates. Is that enough to explain the difference?

I think this misleads the average reader into thinking Japan might have lower rates because they have easier money.  As Friedman pointed out, the ultra-low rates in Japan were a sign that money had been tight.  The earlier episode of tight money drove inflation below zero, and led to sluggish growth.  Under these conditions market interest rates fall to very low levels.  I agree with Krugman that the ECB is now making a serious error in raising rates prematurely.  However, even if the ECB followed the dream policy of Krugman and I, the short term rate might well rise sharply a few years in the future, precisely because of a vigorous recovery (and that’s even if they followed the Krugman/Woodford/Eggertsson policy recommendation for a liquidity trap–hold rates near zero a bit longer than the Taylor rule would suggest.)  So the yield curve would be expected to price in higher future short term rates in Europe, even if the ECB wasn’t erring toward tight money today.  Indeed I think their policy error is one reason long term bond yields in Germany have fallen in recent months.

To summarize, I agree with Krugman that there is a puzzle here, as the gap seems too large.  But Krugman’s post seems to imply it’s hard to explain why there is any gap at all—the Japanese case looks just as bad as the Italian case.  But that’s not true.  At least half the gap can be explained by the different trend rates of NGDP growth in the euro zone and the yen zone.   After all, we’re talking about euro bonds and yen bonds.   I suspect the other half has to do with the type and timing of crises anticipated in each country.  Japan’s endgame may be more gradual, and occur much later.  But even I am not completely satisfied with that answer.

PS.  Obviously most economists would prefer to do my NGDP gap analysis with the inflation gap.



32 Responses to “Paul Krugman poses a very difficult question”

  1. Gravatar of Cliff Cliff
    17. July 2011 at 09:24

    I think people are talking about devaluing the Euro to save Europe.

  2. Gravatar of onliberty onliberty
    17. July 2011 at 09:33

    Yep I agree with Cliff. Italy has no easy way out because they are locked in to a fixed exchange rate. I think the market knows that fixed exchange rates are prone to collapse.

    The yen not only floats, but it’s often a semi-reserve currency.

  3. Gravatar of onliberty onliberty
    17. July 2011 at 09:44

    It’s one thing if Greece is experiencing trouble because of the euro. They could leave the eurozone and devalue relatively easily (not that that would prevent a default).

    The story is totally different with Italy. Italy is a core member of the eurozone and if Italy left, it would cause massive uncertainty (it already is).

    Those concerns simply don’t exist with Japan.

  4. Gravatar of JW JW
    17. July 2011 at 09:55

    It’s funny, I read Krugman’s post and I immediately had two thoughts: 1) I wonder what Scott Sumner thinks about this? and 2) I bet he has a partial explanation for this puzzle. Thanks for your thoughts. I agree with you that a significant part of the answer relates to the immediacy of the danger faced by investors in Italian bonds.

  5. Gravatar of flow5 flow5
    17. July 2011 at 10:03

    “I like to argue that long run nominal interest rates are closely linked to long run NGDP growth”

    The function is slightly more complicated. You have to take into account TIME. And TIME is a function of the deflator, not real-output.

  6. Gravatar of JPIrving JPIrving
    17. July 2011 at 10:11

    I am satisfied with your the NGDP growth rate and greater urgency in Italy explanations. Forget devalued Euros, what’s the probability an Italian bond wont be mostly paid back in lira?

    I am dumbfound at the ECB’s tight policy. They run the ship like it is a normal currency area, not an increasingly fragile monetary alliance where compromise is needed. Norway, Sweden and Switzerland are looking pretty smart these days. Central Europe too.

  7. Gravatar of Mark A. Sadowski Mark A. Sadowski
    17. July 2011 at 10:11

    Krugman raises the question merely because he observes that despite Japan’s monetary independence it seems to be following a relatively tight policy. I think Krugman’s mistake is that he is looking at the current figures for inflation and is failing to take into account overall growth trends.

    From 1997-2007 Japan’s nominal GDP did not grow at all (modest deflation coupled with modest real growth). In contrast Italy’s nominal GDP grew at an average annual rate of 4.0%. The IMF is forecasting that from 2007-2016 that Japan’s nominal GDP will grow at an average annual rate of 0.1%, a tiny increase in the rate of growth. In contrast Italy’s nominal GDP is forecast to grow at a 2.2% annual rate, a substantial decrease in the rate of growth.

    The point is that Japan’s monetary stance is not forecast to be much different between 1997-2007 and 2007-2016. Italy’s, however, is forecast to become much tighter.

    So it’s not just the fact of Italy’s lack of monetary independence at work here. It is also the fact that its monetary policy is expected to be much tighter than in the past. And this dramatic tightening of monetary policy makes holding Italy’s debt much riskier.

  8. Gravatar of Tomasz Wegrzanowski Tomasz Wegrzanowski
    17. July 2011 at 11:01

    What are full yield curves of both? If the gap doesn’t disappear at any time it rules out a lot of explanations.

  9. Gravatar of Indy Indy
    17. July 2011 at 11:51

    One part of the puzzle may be to look at currency exchange rates. The Yen is at an all-time high of 79 to the dollar (having steadily appreciated 56% in only four years) and now within striking distance of the all-time high vs. the Euro (having appreciated 50% in only three years, and mostly in the second half of 2008).

    In fact, let’s look at a table of major currencies and our two favorite commodities using the minimum on June 22, 2007 vs the Dollar as our index of 100. The table today would look like this:

    UK: 81.7
    US: 100
    EURO: 105.7
    CHINA: 118
    Oil: 142 (WTI), 163 (Brent)
    Japan: 156.0
    Gold: 245 (if Pounds were the index, it would be 300)

    Now, I’ve remarked before that conventional Econ-101 thinking tells us that at least in terms of purchasing power in globally-traded commodities (excluding local wages and real-estate), US inflation should have been running at about 12% higher than Japan’s, though we know that the core formulas for both over this period have been very close and low.

    Obviously, it takes time for the currency effects to ripple and propagate through the systems, but four years is actually a long time to my mind. For some reason the effects have a latency that has yet to fully manifest. Maybe the interest rate gap has something to do with that.

  10. Gravatar of Benjamin Cole Benjamin Cole
    17. July 2011 at 12:01

    Well, as a layman I say this:

    1. The yen has been appreciating forever. A foreigner who buys a Japanese bond can expect interest plus an exchange-rate return.

    2. The Japanese owe money to themselves only, and even have foreign reserves earning interest. In a sense, default isn’t possible in Japan, as all the interest paid just goes back into the pockets of Japanese citizens who pay Japanese taxes.

    That said, japan is suffocating itself with its tight-money policies. While their economy is not shrinking, it is very sluggish, and a whole generation of Japanese have seen nothing but declines in the values of equities and real estate.

    China and Korea have been booming, and print lots of money.

  11. Gravatar of onliberty onliberty
    17. July 2011 at 12:26

    “what’s the probability an Italian bond wont be mostly paid back in lira?”

    It could be huge. The liabilities will be in euros and will have to be paid back in lira. This basically means that their debts will be dollarized (in this case euro-ized). History has proven that when that is the case default frequently occurs.

  12. Gravatar of Cliff Cliff
    17. July 2011 at 12:28

    I also think that the current account surplus of Japan plays into this. With Japan having a huge current account surplus there exists huge demand for the Yen and therefore any monetary action by its central bank would tend to have less effect on the currency value as there are more buyers of Yen then sellers.

    On the other hand, Italy (and the EU as a whole) has a current account deficit. Therefore, there are more sellers of the Euro then buyers. Thus, a monetary expansion into this could have a drastic effect on the value of currency as there might not be buyers (at least not without a devalue).

  13. Gravatar of Richard W Richard W
    17. July 2011 at 13:06

    A few reasons why Italy and Japan with awful demographics could be different.

    Japan’s income surplus on the current account from overseas investments are recycled into JGBs depressing yields.

    The EZ monetary union may have reduced the home country bias. Whereas, in Japan it still exists. Moreover, an investor holding yen can only invest those yen in Japan. They could exchange the yen with somebody for some other currency and now there is a new holder of the yen asset. However, in the EZ pulling euro denominated investment from Italy or Greece can be reinvested anywhere else in the EZ. Therefore, German bund yields would decline and the bund spread vis-a-vis the periphery should widen.

    A general dearth of investment opportunities in Japan. An Italian can still find good investments in other parts of the EZ so is less reluctant to recycle savings into government bonds.

    Concern over the solvency of some Italian banks would cause government bond yields to rise because the banks are holding so much of the outstanding stock.

    Political instability in Italy is almost a permanent feature. The Prime Minister public argument with the finance minister spooked the markets a few weeks ago.

    Rollover risk. I think they need to refinance something in the order of 40% of the total stock in the next few years.

    The market has realised after a period of convergence that there is effectively no risk-free rate in the EZ. It is a fixed exchange rate regime rather than a monetary union and convergence is reversing to a divergence reflecting something like the true risks. Nobody truly knows what the true risks are so the market continually overshoots until there is clarity. The overshooting is a self-fulfilling prophecy.

  14. Gravatar of MTD MTD
    17. July 2011 at 14:22

    Scott — I like to think of the difference between peripheral debt markets and Japan’s debt market (and the US for that matter) as the difference between being funded in your own currency versus not being funded in your own currency. The PIIGS look like Mexico in 1994 and Argentina in 2001. Doesn’t this account for the bond rate differential? Japan’s way out is likely a depreciating yen with gradually rising inflation (and NGDP). Italy will get a big burst of NGDP after they default/exit the euro, not before. So there’s more explicit default risk in Italy versus Japan.

  15. Gravatar of Scott Sumner Scott Sumner
    17. July 2011 at 14:39

    Cliff, They should have done that years ago.

    OnLiberty, I agree. But it’s still slightly odd that their doesn’t seem to be inflation risk in Japan–their deficit is huge.

    JW, Thanks.

    flow5, I thought I did take time into account.

    JPIrving, I agree about the ECB. I think we can safely say they’d be paid back in lira–the question is what would those lira be worth?

    Mark, That’s a very good point. Indeed I think there is much more that could be said about the question. What do you think of my breaking it into two parts; the German rate, presumably reflecting eurozone-wide economic conditions, and the Italian risk premium?

    What your comment implies (I think) is that although the specifically Italian inflation rate doesn’t matter via the Fisher effect, a low inflation rate makes default more likely, and actually raises interest rates–a reverse Fisher effect!! Is that worth a post?

    Tomasz, Good question. All I know is the Japanese short rate is near zero.

    Indy, I know I’m supposed to believe in the EMH, but doesn’t the J-bond look like the worst investment in the world? You are buying a currency when it has just appreciated far more than PPP would imply, and then you get a measly 1% interest.

    No sooner do I type that, when Benjamin presents the other side.

    onliberty, No, I think the debts would be converted to a devalued lira–that’s what Argentina did with its dollar debts.

    Cliff, Yes they have a CA surplus, but I’m surprised people aren’t mentioning their fiscal debt–it’s almost Greek-like. Krugman does have a good point.

    Richard, Those are all good points. I was thinking of mentioning the home country bias, as it would help me dig out from under the fact that my NGDP argument violates the Balassa-Samuelson effect. In general, faster growing countries should see expected currency appreciation, and hence lower interest rates, but Japan is slow growing. The home bias could create demand for their bonds. I suppose that’s why I mentioned above that they seemed like a horrible investment to me–I have no home bias from Japan.

    And the exact opposite is true of China, where I have invested heavily.

  16. Gravatar of Scott Sumner Scott Sumner
    17. July 2011 at 14:42

    MTD, That may be it. But I think Krugman’s point is that the advantage of having your currency is that you can inflate out of huge debts. And Japan has a huge debt. But in that case why doesn’t Japan have an inflation premium? The timing issue may be it, which I also mentioned.

  17. Gravatar of Richard W Richard W
    17. July 2011 at 15:20

    HSBC did some interesting research on what hypothetically would happen with a EUR-core and EUR-periphery currency. They suggest that the euro for the core is currently 28% undervalued vis-a-vis the USD. Therefore, there is a huge overvaluation in the periphery. The pressure is moving to the government bond market because the adjustment can’t come through the exchange rate.

  18. Gravatar of Steve Steve
    17. July 2011 at 17:54

    Scott wrote:
    “Italy is also paying a price for the contagion effect”

    If you believe in EMH, wouldn’t the contagion effect not exist? Instead, it would be a rational belief about the consequences of a binary policy choice.

    Also, I think the mystery is why are German Bund yields so *HIGH*.

    Here’s the doomsday scenario:
    Greece defaults and drops out of the Euro. This hurts capital in Eurozone banks, while also causing the Euro to rise because it is more heavily weighted toward Germany.

    Other countries are now less competitive and at greater risk of having to bail out their banks. This pushes Portugal, and possibly Ireland or Spain into default. More Euro appreciation, and more banking sector and competitiveness pressures on the remaining countries such as Italy.

    Ultimately, only Germany remains, everyone else has defaulted, and the German Euro trades at $2. All the other countries are de facto operating currency boards pegged to the German Euro.

    Are people underestimating the huge appreciation potential of the German Euro?

    Also, I think Benjamin Cole was right about Japan.

  19. Gravatar of Steve Steve
    17. July 2011 at 17:58

    Oh, and Scott, one of the great things about your blog is that you’ve attracted a very knowledgeable readership.

    Paul Krugman is weighed down by the less savvy crowd who read the NY Times.

  20. Gravatar of Rien Huizer Rien Huizer
    17. July 2011 at 20:05


    Trading volumes in JGBs are at least 10K larger, so the price signal is probably much clearer. There is no demand for CDS on Japanese gvt laibilities and the authorities still wield strong, but mainly informal control over the market.

    I doubt the EMH applies to both the markets in Japanese and Italian gvt and of course, Italian gvt does not have a lot of control over ECB policy (although what about mr Draghi’s past indiscretions (parking tickets etc) still on file in Rome?

    I wonder what mr Krugman was expecting to achieve: it is not a very productive line of enquiry.

  21. Gravatar of Mark A. Sadowski Mark A. Sadowski
    18. July 2011 at 02:19

    Scott you wrote:
    “What do you think of my breaking it into two parts; the German rate, presumably reflecting eurozone-wide economic conditions, and the Italian risk premium?”

    The problem is that in my opinion the risk premium itself the result of the eurozone tightening having impacting the core and the peripheral regions asymmetrically. For example from 1997-2007 Germany’s NGDP grew at a 2.4% average annual rate. The IMF forecasts it will grwo at a 2.3% annual rate from 2007-2016, hardly a difference. meanwhile the eurozone as a whole (eurozone-17) will see its average annual NGDP growth rate slow down from 4.3% to 2.0%. Similarly Portugal, Spain and Greece will see their average annual NGDP gowth rates slow down from 5.3%, 7.7% and 7.9% respectively to 1.6%, 2.3% and 1.7% respectively over those same time periods.

    And you wrote:
    “What your comment implies (I think) is that although the specifically Italian inflation rate doesn’t matter via the Fisher effect, a low inflation rate makes default more likely, and actually raises interest rates-a reverse Fisher effect!! Is that worth a post?”

    I’m merely thinking that a great deal of that debt was acquired under old NGDP growth trends that are simply no longer true. Tight money creates fiscal crises by reducing the growth rate of NGDP and consequently leading to dramatic reductions in revenue growth. What was sustainable before is no longer. But I am interested in your thoughts concerning a reverse Fisher effect.

  22. Gravatar of Steve Steve
    18. July 2011 at 05:10

    Mark A. Sadowski wrote:

    “For example from 1997-2007 Germany’s NGDP grew at a 2.4% average annual rate. The IMF forecasts it will grwo at a 2.3% annual rate from 2007-2016, hardly a difference. meanwhile the eurozone as a whole (eurozone-17) will see its average annual NGDP growth rate slow down from 4.3% to 2.0%. Similarly Portugal, Spain and Greece will see their average annual NGDP gowth rates slow down from 5.3%, 7.7% and 7.9% respectively to 1.6%, 2.3% and 1.7% respectively over those same time periods.”

    Actually, I think the problem is far, far worse. Portugal, Spain, and Greece NGDP will be forced to return to the German trend line, not merely slowed to a lower trend line. That mean 10 years of negative NGDP. The monetary union enforces NGDP level targeting.

  23. Gravatar of Mark C Mark C
    18. July 2011 at 07:12


    I am happy that you raise this, this is a topic of great interest which me and my colleagues have had a few debates on this.

    First, I would say the prevalent belief in the financial market is that JGBs (Japanese Government Bonds) are being owned by Japanese people, like what Cole said, the Japanese owe money to themselves only.

    But one observation, 10yr JGB yield has gone up more than 80 bps from 1.17% in July 2005 to a high of around 2% in early 2006 and managed to stay above 1.40% through 2007, that also happened to be a period where BoJ hiked rate by 50 bps and the Nikkei Index reached a post yr 2000 high of 18,000 (compared that with the current level of 10,000), conincidence? I think probably not.

    I tend to believe that longer term yields are determined by longer term NGDP growth expectation. As long as the government owns the currency, there shouldn’t be any risk premia in it. (After all risk premia is relative, Italy has a 3% risk premia against German bunds which is considered as risk free, but should JGB has a risk premia, then what asset should the risk premia be measured against? What asset in JPY could be considered safer than JGB?) Like MTD said, Japan always have the option of inflating itself out of debt and default risk is minimal.

    Of course you will ask “But in that case why doesn’t Japan have an inflation premium? ” and I agree with you that it’s a timing issue. If we think about it, Japan has been suffering from deflation and very low inflation for decades, any market participant wouldn’t price in much inflation premium until they see any sign of BoJ or Japanese government changing their stance.

    Another interesting point is when the credit rating agencies cut Japan’s credit rating by one notch earlier this year, the JGB market barely moved a bit. I would bet things would be very different if they cut Italy’s ratings by one notch, and that is quite likely to happen within the next few years.

  24. Gravatar of Mark C Mark C
    18. July 2011 at 07:39

    Scott you said:

    “faster growing countries should see expected currency appreciation, and hence lower interest rates”

    If that’s what the B-S effect predicted, it certainly doesn’t apply to the current situation, truth is, emerging economies from Indonesia to Mexico not only have appreciating currencies but also a much higher interest rates.

  25. Gravatar of Scott Sumner Scott Sumner
    18. July 2011 at 07:58

    Richard, That’s interesting, although I have trouble imagining a 1.83 euro. I’m not sure the SF is the ideal comparison, as Switzerland is a small and special place.

    Steve, No, the contagion effect is 100% consistent with the EMH (plus imperfect information). There are many unobserved aspects of a country’s economic position. These may include accounting (as in Greece) but also likely political reaction to imbalances. When something happens in one country, it reveals a new set of probabilities about what is likely to happen in other, similar countries.

    Steve, That might happen, but there are plenty of other possibilities. First of all, the euro core would probably include much more than Germany. Second, if there was a major crisis the ECB might adopt a more expansionary policy to prevent a depression–perhaps even depreciating the core euro. I really don’t know what would happen, but there are many possibilities.

    Thanks for the support Steve, I’m very aware of how lucky I am.

    Rien, The Japanese government may have important control over the market, but would that really explain those low rates? Is the marginal buyer controlled by the government?

    mark, It seems to me that you aren’t really criticizing my attempt to split the problem into two, but rather you are suggested I overestimated the forward-looking eurozone NGDP growth rates (and you’re probably right.) But in that case is seems like both Germany and Japan have about the same perceived inflation risk, and the extra 1.6% is completely explained by the higher eurozone NGDP growth expectations. Which still leaves the puzzle of why Italian debt is viewed as risky, but not Japanese debt. It’s 3.0% mystery, not a 4.6% mystery, as Krugman suggested.

    I certainly agree that the slowdown in NGDP growth triggered the crisis.

    Steve, You said;

    “The monetary union enforces NGDP level targeting.”

    Yes, but not for each country. Your comment assumes all the Spanish catchup growth in RGDP was illusory–which isn’t necessary so.

    Mark C., I agree there is no default risk premium in Japan, and said so in the post. But why is there no inflation risk premium. Isn’t it just as hard for Japan to pay back its debt as Italy? You can say that Japan controls their own currency–yes but the only way they could put that to use (reducing their debt burden) is by inflating.

  26. Gravatar of Scott Sumner Scott Sumner
    18. July 2011 at 08:00

    Mark, C, But it’s the expected currency appreciation rate that matters. And I think that forward markets show some of these currencies expected to depreciate–at least where there are no capital controls. That’s simple arbitrage.

  27. Gravatar of David Stinson David Stinson
    18. July 2011 at 08:20

    Scott said:

    “MTD, That may be it. But I think Krugman’s point is that the advantage of having your currency is that you can inflate out of huge debts. And Japan has a huge debt. But in that case why doesn’t Japan have an inflation premium? The timing issue may be it, which I also mentioned.”

    I wonder if the better way to think of it is that when one does not have control of one’s monetary policy (or one’s own currency), there is a greater chance of monetary disequilibrium. This has two effects on sovereign debt rates – one on expected inflation (i.e., Fisher effect) and one on default risk (via the risk of excess money demand).

    Two questions related to your post:

    1) Why is/was the carry trade sustainable, if not some sort of continued monetary disequilibrium? This, I think, is another way of asking why there is a continuing gap between Japan and other countries’ rates.

    2) I can see why real rates are related to expectations of real GDP growth. Given global capital markets, however, I don’t understand why real rates (excluding default risk and inflation expectations) are related to expectations of real growth in any one country (rather than expectations of global real growth).

  28. Gravatar of Ales Ziegler Ales Ziegler
    18. July 2011 at 13:42

    I try a very layman explanation – spreads reflect inflation risk and default risk. Thus, first part of the puzzle is inflation gap between eurozone and Japan, which Scott Sumner (btw. I am not a native English speaker, I really have no idea whether or not I should use third person here) describes in terms of NGDP.
    One might asks why such a gap exists, but answer on that would require (or maybe more precisely, is nothing else than) explanation why Japan is so deflationary country. I do not know that, I suppose that possible answer could be found at some previous post here.

    Second part of the puzzle is default risk, which, I think, exist for every country, even if “government owns the currency” – what is that mean anyway? In an uncertain world, there is always a small risk for every government that it would not be here after ten years (imagine coup, or some similar catastrophy).
    And default risk is significantly higher for Italy than for Japan because of huge difference in socio-politico-cultural-whatever environment in both countries. I admit that it is not an economic approach – but sovereign default is political problem as much as economical (best regards to the US). To start from the top, Italy is the country whose citizens put Silvio Berlusconi, entrepreneur in the area of showbusiness, in charge of their State. In charge of Japanese State is formally an Emperor from oldest dynasty in the world. That alone explain a lot about differences between those two countries.
    Italian society, especially at south, is very “mediteranean” and shares many characteristics with Greece society, namely broad distrust and animosity to the government and extraordinary ability to neutralize legal restrictions through informal means. Japan society is, or at least it is common western layman’s understanding, deeply inequal and hierarchical, with its career ladders in huge corporations and bureaucracy, and legendary emphasis on maintaining personal reputation. To put it most simply possible, we have one society which is notoriuosly undisciplined, and second which is notoriusly disciplined – which one is more likely to default on its national debt, ceteris paribus?

    I think that Japanese bondholders rationaly believe in willingness and ability of their government to implement drastic austerity measures in order to avoid defaut (and face loss associated with it) far more than Italian bondholders.

    P.S. I think that without euro gap between Japanese and Italian spreads would be measured in light years by now, because of expected inflation gap.

  29. Gravatar of MarkS MarkS
    18. July 2011 at 15:53

    Scott, you’re as clueless as Krugman. Japan can print their own currency. So there’s no such thing as a solvency risk there. It is impossible for them to not be able to pay off a debt that is denominated entirely in their own currency. The only risk then that bond traders see is the inflation risk. Obviously, there is no inflation in Japan so rates have remained low. The case is the same in the USA.

    Germany has a very real solvency risk as they can’t print Euro, but their financial position is far superior to that of the other Euro nations. Therefore, they’ve remained the safehaven in the EMU. There’s nothing that says that will last forever though. As long as they remain in a monetary system in which they are always revenue constrained they will always have a solvency risk.

    Setting you straight,


  30. Gravatar of Mark C Mark C
    19. July 2011 at 07:43

    Scott, you said

    “And I think that forward markets show some of these currencies expected to depreciate-at least where there are no capital controls.”

    Forward FX rates, unlike forward interest rates seldomly contain any expectation of future exchange rates. The fx fwd swap points that’s being added (trading at premium) or deducted (trading at discount) from the spot fx rate depends almost entirely on the interest rates differential. Take for example the currency pair AUD/USD which is currently traded at around 1.0650. If at anytime, the AUD money market rate is trading higher than the USD money market rate (which is what we are seeing now) then the forward rate will always be a discount, so if you buy 3mth fwd AUD/USD, the fx rate that you will get will be lower than 1.0650, for sure! And it doesn’t really mean market thought AUD will depreciate against the USD 3 mths from now.

    If at anytime in the future we have a higher USD money market rate than AUD, then the currency pair will be trading at a premium.

    Furthermore, any person who’s taking a bet on the future movements of a currency pair through fx forwards will affect the spot market in the same direction, so the fact that a currency is appreciating today more or less tells you what people’s expectations of that currency in the future.

  31. Gravatar of Scott Sumner Scott Sumner
    19. July 2011 at 08:39

    David Stinson, I agree with your second point, and indeed raised that problem in my post. (When I mentioned interest parity.) Some argue for market segmentation, especially in Japan.

    On the first point, I’ve always assumed that people borrowed in Japan and invested in Australia because Japan had a high saving rate and a dearth of investment opportunities.

    Ales, I have made similar remarks about Berlusconi, and southern Italian society in general (along with Greek society.) Societies with less “civic virtue” have bigger economic problems.

    MarkS, Thanks for setting Paul and I straight (rolls eyes.)

    Mark C, I understand all that, but where do we get the expected change in forex rates?

  32. Gravatar of Fred F Fred F
    4. August 2011 at 07:29

    Hey Ziegler, read Mein Kampf lately? Sheesh, unbelievable racist stereotyping.

Leave a Reply