Mining European data
from Follow the Money

Mining European data

More on:


I spent some time over the weekend delving into the Eurozone balance of payments data published by the ECB in its monthly bulletin, looking for insights into the global flow of funds.

But to jazz this up, I want to draw on this data to push back a bit - or perhaps to nuance a bit - three arguments that appeared in the US last week:

  1. Greg Ip's argument that foreign central bank financing for the US has dried up in 2005, wit h no impact on interest rates: "foreign central bank purchases of US bonds have financed a significantly smaller share of the US trade deficit this year than they did last year, with no discernible impact on interest rates."
  2. Caroline Baum's argument that US manufacturing is in inexorable decline because of rising productivity in the manufacturing sector and the trade deficit has little to do with the health of the manufacturing sector.
  3. Kristin Forbes' argument (reported by Ip) that "Europe" not only needs to do its part to address global imbalances. 

Three things are needed to rebalance the global economy: a smaller US budget deficit, flexible Asian exchange rates, and stronger European growth.  "On [the first ]two of the three planks we've seen progress, which increases the chances of a smooth reduction in global imbalances."

Roubini also likes a variant of the Forbes argument, though he would argue the recent improvement in the US fiscal deficit is cyclical and almost certainly transitory.  I agree that all parts of the global economy need to do their part, and certainly think stronger European growth would help.  But I do think Europe should get a bit more credit for letting the euro appreciate in 2003 and 2004 than it often gets -- it let the Euro move by a bit more than 2.1%.  Europe is not the only region that is not doing all that it needs to do for orderly global rebalancing.

A region's contribution to global rebalancing is a function of both currency movements and domestic demand growth - Europe provided one without the other.  But euro appreciation alone did have an impact in 2004, and is still having an impact in 2005.

Leaving aside my (small) attempt to enliven the post by structuring it as commentary on these three points, make no mistake - this is a very wonky, data intense kind of post.  In the future it may make more sense for me to write up commentary on this kind of data as a short paper, not as a blog.

First, Ip.  His Friday Wall Street Journal article suggests - echoing many others - that foreign central bank support for the US fixed income market is falling.  That is clearly the story that the US data tells.  It is a story that the Fed, among others, likes.   In q1, net central bank inflows into the US were only $25 billion - only about 1/8th of the almost $200 b quarterly current account deficit.   In 2003 and 2004, central bank inflows accounted for over ½ the current account deficit.  A plot of central bank demand for Treasuries over time tells a similar story - with a fall off starting in q4 of 04 that continues in 2004 (not coincidentally, that falloff coincides the with the end of Japanese purchases)

My problem?  Simple.  Global reserve accumulation has not fallen off anywhere near as sharply as the fall off in recorded central bank inflows to the US.  It is running at a roughly $125b a quarter pace, with about ½ that increase coming from China.   If central banks are only putting 20% of that increase into the US, what are they doing with the rest of it?   If China adds close to $300 b to its reserves this year, and puts only a third of that increase into dollars, that alone would generate $100 b of net inflows.  Yet if the q1 quarterly data is taken at face value, the US is on track to receive only $100 b in net inflows from all the central banks of the world.

The eurozone data might help clarify the mystery - and help us track down what foreign central banks are doing with these funds if they are not investing in the US dollar.  Unfortunately, I could not find a breakdown between "central bank" purchases of European securities and private purchases.    However, comparing the European data to the US data is still interesting.




$ billion unless otherwise noted





2005 ytd

Central bank reserve increase, valuation adjusted









(estimate largely explained here, remember that the 11 economies are not the world)

o/w $ reserves




At least 400 ($395 b of inflows recorded in the US data), probably more like 500

At least 50

(25 in q1)

o/w non $ reserves










Total non-resident purchases of US long-term debt securities






o/w private




@ 301 (through may)

o/w official





o/w official purchases of Treasuries











Total non-resident purchases of euro long-term debt (in euro billion)




102 (through April)

Central bank purchases of euro debt

less than $74 b/ euro 65-80b (depending on the timing of the purchases)

less than $61 b/ euro 50 b



Central bank reserve data: BIS, and Setser estimates. 2003 data adjusted to reflect Chinese reserves transferred to two state banks.  2004 and 2005 minimum dollar reserve increase come from the US BEA.   Foreign purchases of US securities from the TIC data set.   Foreign purchases of euro securities from the ECB, specifically from table 5 in section 7.1 of the statistical annex to the ECB's monthly bulletin.

What jumped out at me from this data?  A few things:

The surge in central bank demand for US debt securities led to a noticeable pick up in overall foreign demand for US debt.  That is pretty clear in the 2003 and 2004 data.

Presumably, a surge in demand from central banks for euro-denominated debt would have a similar effect on the overall eurozone data.  We know central banks only added about $74 billion to their non-dollar reserve portfolio in 2002, and another $61.5 billion in 2003 (data from the BIS).  Obviously not all non-dollar reserves go into euros, and not all euro reserves are invested in euro-denominated bonds, but some of those funds presumably were used to buy euro-denominated securities. 

Unfortunately, if we don't have a breakdown between dollar and non-dollar reserve accumulation for 2004.   But if central banks added more euros to their portfolio in 2004 AND private demand stayed constant, there should have been a pickup in demand for eurozone debt securities.   Indeed, there was a small bump up appeared in the European data, as foreign purchases of eurozone debt securities increased from euro 170 billion to euro 210 billion - a euro 40 billion pickup.  That is consistent with some small pick up in central bank demand, but not an enormous increase in central bank demand.   

Inflows into Euroland bonds seem unlikely to explain the $X b gap between total reserve accumulation (net of a valuation gains) recorded central bank inflows into the US.  Take a very-high end estimate of foreign central banks purchases of euroland debt securities, and say they bought euro 100 billion of euro-denominated debt in 2004 ($135 b using end 2004 exchange rates).   That would still have left about $500 b to invest in other things, and a $ gap between overall reserve increase and inflows into the US.   And remember, 100 euro billion would be a very high estimate of central bank demand for euro-denominated securities. 

There also is not a strong correlation between the quarterly pace of non-Japanese reserve accumulation in 2004 and inflows into the eurozone debt markets.  Remember, in q4 2004, emerging market central banks, including the PBoC, were adding to their reserves like mad.  They presumably are more willing than the BoJ - which was out of the market by then -- to hold non-dollar reserves.  But total purchases of European debt securities, public and private - were only about euro 32 billion in q4 2004 - the lowest quarterly total of the year.   

Similarly, it is hard to believe that the inflows to the eurozone debt in q1 - which were in line with the quarterly average for 2004 - explain the fall off in recorded central bank inflows to the US.  ½ of euro 44 billion is euro 22 billion, or $27-28 billion.   That is no where near enough to explain the "missing" $100 b in central bank reserves.

Obviously, I don't know the breakdown between central bank purchases and private purchases of eurozone securities.  Private sales or smaller private purchases could mask bigger central bank purchases.  And certainly central bank purchases could account for less than the ½ of total purchases - that is a made up number that I used merely to set some parameters.

One last point: Demand for eurozone debt stepped way up in April - to euro 58 billion. But demand for European equities also fell off, big time, in April.   Indeed, shifts in equity flows far more than shifts in debt flows tend to correlate with euro moves.  The big inflows into the eurozone in q4 went into equities, the pace of inflows fell a bit in q1, and foreign investors (say Americans) were net sellers in April.  The preliminary May data suggests a fall off in external demand for Eurozone debt, but a pickup in demand for eurozone equities.

My overall point:  central banks have not stopped adding to their reserves, so if they really are buying less dollar debt at the margin, what are they doing with their money instead?  The $100b in global reserves in q1 that did not show up in the US data ought to be leaving a few traces somewhere.  The US data implies a huge surge in demand for non-dollar assets.  $100b annualized is $400b, or about 40% of the expected 2005 global current account surplus.   It ain't small change.  Until it can be accounted for, I would be somewhat cautious in interpreting the US data.   We know central banks are increasingly holding a more diverse portfolio of US debt, and it is at least possible that the US recording system is increasingly failing to pick up all central bank inflows into the US.

Second, is Baum correct to argue that manufacturing employment in exorable decline because of rising manufacturing productivity, not because of trade?

It [a yuan revaluation] won't help the U.S. manufacturing industry. The number of U.S. manufacturing jobs peaked at 19.6 million in June 1979, when factory workers comprised 21.7 percent of the workforce. Today the U.S. boasts 14.3 million factory jobs, 10.7 percent of all employed workers.

The villain, if you choose to look at it that way, is productivity. Manufacturing productivity, or output per hour worked, has risen at an average 4.5 percent rate over the last decade.

Certainly rising productivity is part of the story.   But I would argue that the fact that a rising share of US demand for manufactured goods is being satisfied with imports, and the fact that the US is paying for those imports by exporting debt rather than goods or services has added to the decline in manufacturing employment.  Look at the calculations that Ronald McKinnon has done.  If the US had to pay for its manufactured goods imports with manufactured goods exports, manufacturing employment would be higher than it is now, even with strong productivity growth in manufacturing.

That doesn't imply that a rising trade deficit has reduced total employment in the US - the fact that the US is exporting debt helps interest-sensitive sectors, and increases employment in those sectors.   It does imply that the composition of US output is somewhat different than it would be if US trade was closer to balance.  I suspect that there are proportionally fewer manufacturing jobs and more housing jobs than otherwise would be the case.  

What does the eurozone data tell us?  That high-wage countries do not necessarily run a deficit in "manufactured" goods trade.  Like the US, European countries compete with low wage labor -- but the eurozone still ran a euro 229 billion surplus in trade in manufactured goods with the rest of the world, more than offsetting a euro 155-56 billion deficit in commodities and raw materials.  Put differently, the US pays for its imported oil  by exporting debt; the euro-area pays for its oil imports by exporting manufactured goods.

Finally, is Europe the weak link in the global rebalancing story, as Kristin Forbes  argues (full disclosure: I briefly worked for Kristin when she was at the Treasury). 

No doubt stronger growth in Europe would help. But I also think the US should give Europe some credit for what it did do - namely, let its currency appreciate from 2002 to 2004. 

The strong euro led to strong growth in European imports even though European growth lagged.  I have argued previously that US exports to slow-growing Europe are doing quite well - far better this year than US exports to fast-growing East Asia -- and the strong Euro presumably has something to do with that.  The same story can be told using European data.   From 2002 to 2004, Eurozone import volumes increased by more (9.6%) than Eurozone export volumes (9.2%).   We know that Chinese exports to Europe are way, way up on the back of the strong euro (weak RMB), at least in dollar terms.   That increase also shows up in the 12.2% increase in the volume of Eurozone imports of consumer goods from 2002 to 2004 (all numbers are y/y changes in annual averages).

European exports also grew strongly during this period - driven by strong demand growth in the US, strong demand for European goods in eastern Europe and the oil exporting countries, and some increase in European market share relative to American market share in "oil exporters" from 2002 to 2004.  But European imports also grew - and grew far more rapidly that one would expect from simple domestic demand growth.  That could have helped to make a dent in global imbalances if US import demand had not been absolutely exploding at the same time.

What the US really needs for the softest of all soft landings is for all other regions of the world economy to experience strong domestic demand led growth even as their currencies are appreciating.  The euro rose a lot from 2002 to 2004, but internal demand in the eurozone was weak.  Asia grew like mad from 2002 to 2004, but its growth was export-led and it resisted currency appreciation.  The US needs Euro style currency appreciation combined with Asian-paced growth, but with the growth coming from domestic demand, not growing exports to the US.

All my quibbling aside at the end of the day Forbes is right, if for no other reason than the difficulty sustaining a strong euro in the absence of strong eurozone growth. 

The interesting question is what - policy wise - should Europe do to generate the demand growth needed to support global adjustment. 

In my view, it won't come just from structural labor market reforms.  Such reforms may be necessary to reduce unemployment in Europe, but in the short-run, it also increases job market uncertainty.  And that may get in the way of the consumption growth needed for strong demand growth.  Look at Germany: by inflating at a slower pace than the rest of the eurozone and keeping wage growth down it has improved its competitive position.  But during that process, German domestic demand has been in the doldrums.   European monetary and even fiscal policy stimulus may also have a role to play during the global adjustment process.

While I do think Europe should get some credit for letting the euro appreciate, and not joining East Asia in fighting dollar weakness, I also do think Europe should attribute its current woes exclusively to the strong euro.   Let me end by quoting Ted Truman:

Although euro area policy makers periodically speak out against the brutal 50% appreciation of the euro against the dollar [Editor's note: and the RMB], the simple fact is that the impact on their economy so far has been limited.  Over the past three years, euro exports to the US as well as the world have continued to expand at a double digit rate.  ... Europe as a whole has yet to feel substantial pain from US external adjustment because there has not been much pain.

Cutting into the US external deficit won't be fun either for the US or for any region that relies on US domestic demand to offset weak demand at home.   The brutal part of the adjustment has not even started

p.s. I have not used the Eurozone BOP data before.  I hope I am interpreting all the pluses and minuses correctly, and I apologize if I missed a key chart that would help clarify a complicated story.

More on: