October 02, 2003

Thinking About Puzzling Anomalies in the Flow of Macroeconomic Data

A precis of a discussion at dinner last Friday night. Things that readers think are smart should be attributed to Paul Krugman or to Janet Yellen. Things that people think are dumb should be attributed to me.

We currently have two large, puzzling anomalies in the macroeconomic dataflow. First, productivity growth is ludicrously, ridiculously, unbelievably rapid. Second, the high current level of the U.S. trade deficit fits very uneasily with the relatively high value of the dollar and the lack of large interest rate differentials in favor of the U.S. relative to other countries. Could one or both of these anomalies be the result of data errors? Could--as Morgan Stanley economist Stephen Roach has hinted--much of the recent data oddities be the result of our failure to adequately track the ongoing process of international trade and globalization?

Suppose that our system of national accounts is missing a lot of service-sector outsourcing. Suppose that $100 billion a year of call-center and other service-sector services provided by foreign companies to American companies is simply missed. And suppose that this missing component has all grown up in the past two and a half years. Then our current real GDP is overstated by one percent--measured imports are too low. And the rate of productivity growth is overstated by 0.4% per year.

There are two observations to make. First, even a productivity growth overstatement of 0.4% per year is less than one-fifth of the productivity growth anomaly over the past two years. Missing outsourcing can only be a small part of what is going on with measured productivity growth. Second, if $100 billion of service-sector exports each year are missing, this means that the trade deficit is not $500 billion a year but $600 billion a year. And the mystery of the high value of the dollar in the absence of large interest rate differentials becomes even more puzzling.

The same argument works in reverse if you try to account for the high value of the dollar by saying that U.S. service-sector exports are understated. Suppose $200 billion a year of service-sector exports are unrecorded. Then our trade deficit is only $300 billion--and perhaps the value of the dollar is reasonable.

But a $300 billion trade deficit is still an enormous one, and there still remains a substantial dollar-valuation puzzle. And if we are missing $200 billion of service-sector exports, we are understating real GDP by $200 billion. If this understatement had grown up over the past four years (say), then productivity growth is understated by 0.5% per year, and the productivity growth puzzle is correspondingly greater.

Either way, the possible data errors in tracking international trade make only a small step toward resolving one anomaly, and increase the magnitude of the other.

So we could see no way of making our current macroeconomic situation less bizarre and puzzling by assuming that there were substantial and rapidly-growing errors in tracking service-sector international trade.

Posted by DeLong at October 2, 2003 01:47 PM | TrackBack


I agree with your conclusion that data errors in the CA are probably not the solution to the two anomalies that you present.

Let me offer alternatives. First, one could argue that there is no dollar/CA balance anomaly, if one looks at the causation running the other way. The US has a huge CA deficit because of a savings-spending imbalance. The strong dollar is simply the *mechanism* by which that S/I (really G/T) imbalance gets translated into a CA deficit. According to this logic, the dollar will not weaken until the CA balance is ready to move back toward zero. Put yet another way -- the dollar is so strong exactly because the US is running a huge CA deficit right now. Thus there is no anomaly.

Regarding productivity: I know relatively little about productivity, but Metlzer proposed one alternative explanation in his WSJ op-ed. He thinks that the employment data is faulty -- more people are working than the establishment survey suggests. I have no idea how likely this is, personally. One final explanation might be that we're getting the typical bump in productivity for this stage of the business cycle, plus a secular increase in productivity that has been ongoing for the past 2 or 3 years due to IT. Yeah, it's a tired story, but if you add a bit of IT productivity to business cycle gains, you might be able to explain current rates of productivity growth.

Posted by: Kash on October 2, 2003 02:16 PM

If you want to explain the "ludicrous" growth in productivity, look at the bottom number, not the top. It isn't clear to me from reading the technical notes on calculation that ANY of the following are correctly accounted for:

1) Hours worked by contract workers, who are neither employees (and so aren't counted as hours worked in the CES) nor firms (and so aren't SURVEYED in the CES). In the surveys, are these people describing themselves as "employed by others" or "self-employed"? I'm sure the employers aren't calling them employees. (Look at the simultaneous divergence of the two big measures of labor force participation - the household and payroll surveys. When two strange things are happening at once, they may well be linked.)

2) How is vacation/sick time accounted? If someone earns 2 weeks vacation every year, but can never take it, does it really exist?

3) Are hours worked correctly calculated? My understanding is that most hours-worked measures are based on hourly workers. Is it reasonable to use these hours worked as a proxy for salaried workers hours? Are BLS's "surveys" used to calculate typical hours for salaried employees in non-production tasks still valid? If salaried employees hours are changing(increasing) over time, this alone could create problems.

4) Is the exclusion of government workers' hours valid? It seems to me that a larger government share of the work force could be a drag on productivity, almost by definition.

5) How accurate/honest are firms in providing this data?

I'm sure if I had time, I could find other objections.

Posted by: rvman on October 2, 2003 02:32 PM

The new BLS statistics ftp://ftp.bls.gov/pub/news.release/cewbd.txt have a lot of information about the current employment situation. In Q4 of 1999, There were over 9 million jobs created. Job loss was 8.1 million. In Q3 of 2001, job creation drops 13 % below Q4 99 to 7.9 million. Job loss peaked at 9.0 million

By Q4 2002, job loss had dropped to 7.8 million, but job creation had dropped to 7.7 million, the lowest since Q2 1995.

Therefore our current jobless recovery is lack of creation of new jobs by either business expansion or new openings. New openings have dropped to 1.6 million, the lowest since 1994, but the overall statistics are dominated by expansion and contraction of existing businesses.

The productivity gains are for real. Companies are making more products and they have ruthlessly cut excess employees. Companies are removing duplication that accompanied automation at the early stages. Plus, building web sites and e commerce requires a lot more employees than maintaining sites once they are built.

The trade deficit is being driven by a shift in parts manufacturing from the US to other countries. Steel tariffs have made it more profitable to import parts manufactured elsewhere with access to cheap steel. The ongoing process has been accelerated. Part of the trade deficit occurs within multinational corporations that buy parts from their foreign assets. If the money gets recycled why should this lead to a devaluation of the dollar? It does lead to more wealth inequality between the investor class and the working class.

Posted by: bakho on October 2, 2003 03:26 PM


Some questions.

1) Is not the trade deficit tied to the large budget deficit as a significant domestic dissavings, requiring the import of foreign financial capital, especially since the household savings rate is quite low? The trade deficit started to trend sharply higher during the last
Clinton years, when the economy was booming and the stock market inflating, drawing in large flows of foreign capital, but, obviously, these conditions no longer obtain.

2) If a large portion of foreign trade is in fact
infra-corporate transfers on the part of multinational corporations, is not this case different from the classic case of comparative advantage for trade between domestically based industries, in that what the corporations are actually seeking is lower wage and tax rates rather than optimalization of sectoral productivity distributions?

3) If there is, for the near to middle term, at least, a rise in the trend productivity growth rate, together with stagnant employment rates, precisely how are these productivity gains to be distributed economy-wide to raise the overall standard of living? Would not such productivity gains to some degree be subject to capture as quasi-monoploy rents within corporate organizations and in profit distributions, especially since such productivity gains increase both the market power and strategic organizational flexibility of large corporations, while the labor "market" is already weak and under-organized?

4) Over the long term, would not high productivity
gains and the redistribution of employment they imply require the development of new economic sectors? And if so, how can we be sure that they are forthcoming, since future technical development is inherently unpredictable? I have long suspected that the expanding health care sector, since its demand is undelimitable, has been one place such gains are parked.

Posted by: john c. halasz on October 2, 2003 04:33 PM

What about the "black economy", I understand in Italy it is a very important part of their economy. With more people out of work, there must be more work being done for cold cash.

Posted by: big al on October 2, 2003 04:47 PM

Is high producitvity not sufficient to explain an over valued dollar? If foreign investors are willing to buy $ assets in the belief that there is a US/rest of the world productivity differential resluting in a higher gain on US assets, then we could have a strong US$ and a high CA deficit at the same time.

It would be interesting to know if the the composition of unconstrained investors investments in the US (defined as non central bank investors) has shifted appreciably towards equity and corporate/retail debt assets rather than sovereign debt.

Seems to be a self re-inforcing cycle - higher apparent/anticipated productivity keeps demand for $ assets high - high demand for $ assets, keps the $ strong.


Posted by: Peeyoosh on October 2, 2003 06:12 PM

Is high producitvity not sufficient to explain an over valued dollar? If foreign investors are willing to buy $ assets in the belief that there is a US/rest of the world productivity differential resluting in a higher gain on US assets, then we could have a strong US$ and a high CA deficit at the same time.

It would be interesting to know if the the composition of unconstrained investors investments in the US (defined as non central bank investors) has shifted appreciably towards equity and corporate/retail debt assets rather than sovereign debt.

Seems to be a self re-inforcing cycle - higher apparent/anticipated productivity keeps demand for $ assets high - high demand for $ assets, keps the $ strong.


Posted by: Peeyoosh on October 2, 2003 06:13 PM

The deficit issue is a wash. Clinton ran surplusses by collecting over 20% of GDP as revenue. Bush is collecting what, 16.3 %?? IN a ten trillion $ economy, 4 % is $400 billion or not far off the deficit. This means the taxes that would have been paid to the government are available to purchase government bonds issued to service the debt. This is just a direct transfer of treasury funds to Bush campaign contributors and paying them interest on the money. No wonder he is able to raise so much campaign cash. The people making money off this scam would lose big time if Bush is fired. A couple of thousand in insurance to pay for negative TV ads to trash the Democratic challenger are worth a $400 billion pie distributed to brown nosers.

Posted by: bakho on October 2, 2003 09:53 PM

Halasz- Re your number 4 point.

Do you really think this adiminstration gives a rats ass about poor people or black people or anyone who does not cut a check for $2000???? If you think that , you are wrong. This administration is about nationalizing Texas. In Texas, Bush cut deals for the people that funded his campaign. This administration is all about cutting deals for the people that will fund 2004. This is all about overwhelming whoever survives the Dem primary with negative ads linking them to axe murderers, father rapers and mother beaters. The model is South Carolina. If your own negatives are high, make the opponent have higher negatives. John McCain has a BLACK BABY!!! Dukakis would stand by and appluad if Willie Horton raped his wife!!! This is what money buys. The Rove strategy is to use money to overcome whatever negatives his boy has. It just might work.

Posted by: bakho on October 2, 2003 10:04 PM

HALASZ- regarding points 2 and 3

#2 Of course they are seeking lower wages. That IS the comparative advantage.

#3. This is not an egalitarian administration. This is about winners and losers. This is about JR Ewing screwing the people he does not like. The intention is to have concentration of wealth, not redistribution of wealth. This administration has radical goals and does not play by the "rules" as some of you like to think exist. They plan to concentrate the wealth and use the concentration of wealth to BUY the next election and future elections.

Why was popular election of senators so controverisal???

Prior to popular election, it was possible for very large corporations (mostly railroads) to buy state legislature majorities and appoint their very own toadies to the US Senate. THis is just a return to big money interests getting their way to grease the skids to perpetuate their own self interest at the expense of everyone else. If you don't stand up and fight it, you deserve to be screwed. The alternative is to join the Pioneer club and pony up the money for the negative TV ads that so many of us detest.

In the long run, this type of policy will drag down our country. Many will lose to benefit the lucky duckies.

Posted by: bakho on October 2, 2003 10:22 PM

My questions were well-meant and more general or "theoretical" than polemical. They were occasioned by your tightly argued post about Bush policies leading to the increased capacity to export capital. I do not know the answers to this questions. Suspicions are not answers.

Perhaps you are right about these deficits being a wash, at least, the 35%(?) of them caused by the tax cuts. But it has seemed to me that the Bushies have assumed by their policies a continuing long-term ability of the U.S.A. to import capital, even if, as well, they have facilitated the possibility of investing abroad. But the rest of the world (RoW) is not in great shape and there was a post at the Bonoboland website today saying, in effect, that the U.S.A. economy is on its own, that there are no other significant sources of demand growth in the world today. Furthermore, as Anne, the ever well-wishing left Keysian, is always at pains to remind everyone, there are large opportunity costs to the policies already enacted. Pointing out that there is the revenue potential to cope with such deficits does not deal with question of how to recoup already lost revenues.

My fourth question was not aimed at what the Bushies would do about the longterm prospect. Obviously, I do not cherish any hopes about there intentions and motives. But recently, rumours, planted by anonymous sources, suggest that the Bush administration has begun to implode. If so, the question would then face their successors. Is it not about time that people began to address the reconstruction of the U.S.A. after regime change has occurred?

Posted by: john c. halasz on October 2, 2003 11:14 PM

"But a $300 billion trade deficit is still an enormous one, and there still remains a substantial dollar-valuation puzzle." - But can you really refer to a puzzle when just looking at one piece, the trade balance? Isn't the good relative productivity and the even better relative productivity gain reasons enough for a strong dollar?

Posted by: Mats on October 2, 2003 11:59 PM

Brad, Paul Krugman, Kenneth Rogoff and a lot of other people who are much smarter than I am continue to make Econ. 101 level mistakes about the current account deficit and the dollar.

These authors--and the conventional wisdom generally--miss two, closely related points. First, foreign investors' risk exposure to dollar assets has essentially nothing to do with the current account deficit. Second, the ability of U.S. households, firms, and the government to service their liabilities have essentially nothing to do with whether those liabilities are owed to foreign or domestic investors.

Point one. Foreign investors don't hold an asset called "net claims" on the U.S. Rather, they hold various gross claims: bonds, equities, FDI an the like. These gross claims are the source of foreign investors' risk exposure to changes in the value of the dollar and the price of dollar assets. (If the dollar falls, for example, foreign investors suffer a home currency capital loss on their dollar assets; liabilities to the U.S., denominated largely in the home currency, aren't much affected.) The right metric for analysing foreign investors' risk exposure to dollar assets is the share of dollar assets in total financial wealth. This share--gross exposure--is, ceteris paribus, boosted by a U.S. current account deficit, but ceteris often isn't paribus. Changes in the value of the dollar and dollar asset prices are also important source of changes in gross exposure. Indeed, from end-2000 to end-2002, foreign claims on the U.S. rose by $97 billion, against the two-year current account deficit of $875 billion.

Foreign investors' exposure to dollar assets, rightly measured, is in fact quite small. Roughly 12% of the total foreign foreign portolios of marketable equities consisits of holdings of U.S. companies--against the 47% U.S. share of global market capitalization. This share is no higher than in 1982. Roughly 11% of the total foreign fixed income portfolio is in dollars--against the 44% U.S. share in global market capitalization, and the 59% U.S. share in the AAA universe. This 11% share is up from 7% at end-1989; but all the increase comes from foreign central bank reserves accumulation, not increased private exposure.

In order to claim that the U.S. current account deficit is generating foreign "overexposure" to dollar assets, you need to say what share of dollar assets in foreign portfolios is "too high."

Point two. Essentially all U.S. foreign liabilities are denominated in dollars. Whether a U.S. household, the U.S. government or a U.S. corporation can make good on its obligations thus doesn't depend on whether they are owed to domestic or foreign investors. As it happens, U.S. household debt relative disposable income isn't particularly high by G-7 standards; neither are corporate or government debt relative to GDP. Granted, debt isn't a complete measure of balance sheet health. But the general point is that there's no reason at all to think that large U.S. current account deficits translate into difficulty in servicing U.S. external obligations.

Brad, and other, might justifiably complain that the U.S. savings rate is too low. But that's not the stuff of a dollar meltdown.

The two points above are closely related. In popular dollar meltdown stories, worries about the current account deficit itself is often the trigger for a flight from dollar assets.

Two disclaimers, and two other quick points.

Disclaimer one. I don't know what will happen to the dollar, tomorrow or at any other horizon.

Disclaimer two. Yes, I know the national income and balance of payments identities. With today's current account deficit, U.S. net foreign liabilties will EVENTUALLY reach something like 90% of GDP, and something like 5% or 6% of GDP would need to be shipped abroad as service income every year. But eventually is a very long time, and U.S. per capita income wil probably double by the time we get there. As long as U.S. foreign liabilities are denominated in dollars, there's no implied payment difficulty.

Quick point one. Global financial integration means that global investors' exposure to foreign assets is rising, both to current account surplus and current account deficit countries. This simply underscore the fact that focusing on the current account defict is the wrong way to think about the issue.

Second, Obtsfeld/Rogoff-type stories call for a large dollar drop due to "sudden stop" in capital flows to the U.S. But remember, the U.S. is big. The $500 billion has to go someplace else. Maybe non-U.S. investment will grow by 20% for two years running, even as the U.S. tanks. But somehow I doubt it.

Posted by: matt on October 3, 2003 07:14 AM

Great post matt.

This administration truly believes that the economy will eventually expand and outgrow the deficits they are running. So they do not worry about it. This administration KNOWS 2 things:

1) Giving tax cuts to campaign contributors and having those tax cuts subject to reversal if Bush loses is designed to keep the big campaign contributions coming in to the GOP. From an investor view, what is $2000 campaign gift compared to $100,000 tax cut. Take the Hillary Clinton $8 million book deal. Dropping the top tax rate by 1% is $80,000 tax saving already. A successful $2000 check would be a 40 fold return on investment. Since elections only come every 4 years, that is only $500 per year. Someone making $1 million would save $10,000 per year or a 20X return on investment every year for four years. Where else does one find those returns on investment?

2) The Administration KNOWS that deficit speding (if high enough) eventually will stimulate the economy they BELIEVE it will reverse all of the revenue loss. Therefore they have no incentive to reverse course on their policy. Reversing course on the tax cuts would alienate the base and lose the election. It is better to take chances that however bad the economy is, the campaign cash guarantee election success.

Posted by: bakho on October 3, 2003 07:41 AM

From 2001 to 2003, National income increases from 10.1 trillion to 10.8 trillion. Job loss is 3.3 million jobs out of 129.9. So a 6.7% increase in national income is accomplished by 2.5% fewer workers. That makes the high productivity numbers seem reasonable.

Posted by: bakho on October 3, 2003 08:05 AM
Post a comment