November 17, 2008
Triple "Ut-Oh": September Trade Release and the End of the Consumer of Last Resort
Brad Setser says "Ut-Oh", beating me to the punch on the September trade release, which showed US exports plunging. It's a post that Paul Krugman rightly expresses some angst upon reading. And now I'm going to add two more reasons to worry (not that I think Setser and Krugman aren't aware of these points).
First, after reflecting upon the collapse of exports noted by Setser, think "disaggregate".
Figure 1 depicts exports of capital goods; they declined 10%, exceeding the 8% reported for all goods exports (all calcuations in log terms). That's 10% for September alone. Since the standard deviation of monthly log changes is 3.1% (2004M02-08M9), well, that's pretty significant... Why focus on capital goods exports (more so than say ag exports), given their volatility? Because they represent foreign demand for goods that can be used to produce things; as demand for capital goods goes down, so too should one's inferences about future growth prospects abroad. And that growth abroad (and the associated US exports) has been what's been keeping the US economy out of recession. Well, as I noted earlier, exports alone have not kept the US economy out of recession in past episodes , and this time doesn't seem like an exception.
Figure 1: Real exports of capital goods (blue), and real imports of consumer goods (red), seasonally adjusted Ch.2000$. Tan shaded area denotes period after 2007M12. Source: BEA/Census, September trade release.
Figure 1 also depicts imports of consumer goods. That series declined 8.1% (compared 2.7% for all non-oil imports). What this seems to suggest is that the consumer has "hit a wall" (as if we needed any additional confirmation of that).
In figure 2 I present a more long run view of these two series, using the NIPA series, from the October 30th release (so predating the release of the September trade release -- but this won't affect the picture save for the 2008Q3 observation for consumer goods imports, which will be revised downward). The key point here is the the import series has flattened out in the period before the consumption downturn.
Figure 2: Log real exports of capital goods (blue), and log real imports of consumer goods (red), SAAR Ch.2000$. NBER defined recession dates shaded gray. Tan shaded area denotes period after 2007Q4. Source: BEA, NIPA advance release of 30 October, NBER, and author's calculations.
That consumption downturn is illustrated in Figure 3.
Figure 3: Log real consumption (blue, left scale), log real goods imports ex.-oil (green), and log real consumer goods (red), SAAR Ch.2000$. NBER defined recession dates shaded gray. Tan shaded area denotes period after 2007Q4. Source: BEA, NIPA advance release of 30 October, NBER, and author's calculations.
Lastly, perhaps an obvious point, but it bears restating: as US consumption declines, this will induce a reduction in imports, which will decrease foreign exports and hence foreign income -- which then reduces US exports, etc. What's the implication of this? Let's use some concrete numbers in a simple two-country Keynesian model (math here). If the marginal propensity to consume out of total income is 0.7, and the marginal propensity to import is 0.3, then the multiplier for an autonomous drop in consumption is 1.67, when there are no repercussion effects. But if the drop in US imports decreases foreign income, then (assuming symmetry for the sake of simplicity), the multiplier is 2.22. The multiplier worked on the way up; it will work on the way down.
Now add that idea to the estimate I presented in this post on the consumption prospects -- the implied 2% decline of 164 bn Ch.2000$ (if treated as largely due to wealth effects) will imply a larger decrease in output than implied by the closed economy multiplier.
So, to sum up:
- The sharp and significant decline in exports is worrisome, since exports have been maintaining US GDP growth thus far.
- The sharp and significant decline in capital goods exports is worrisome, because it presages a decline in future output, insofar as expenditures on capital goods are a forward looking indicator of investment prospects.
- The sharp and significant decline in consumer goods imports further substantiates the view that consumption is sharply dropping. Even if the drop is reversed next month, the sustained flattening out of consumption imports (already it looks longer than in the previous two recessions) represents a big departure from the past.
- If the US has truly stopped being the consumer of last resort, then to the extent the impending consumption decline in the US is autonomous (largely unrelated to income), we should expect the repercussions to be widely felt amongst our trading partners.
Posted by Menzie Chinn at November 17, 2008 09:48 PMdigg this | reddit
the one (tiny) good bit of news is that the October capital goods exports data was influenced by the boeing strike (only 5 deliveries v 12 in sept and way more in august ...)
that said, the boeing strike likely pulled forward bad news that was already on the way; there is a host of other indicators that suggest a significant slowdown in exports and in all probability a slump (meaning an outright fall)
Posted by: bsetser at November 17, 2008 10:23 PM
How much of the drop in the value of exports is a price effect? The dollar strenghtened significantly against most currencies (the DXY index was up 10% QoQ in Q3!) surely affecting foreign demand. Volume data should provide more meanignful clues.
keep up the the excellent posting!
Posted by: Andrea at November 18, 2008 03:28 AM
Paul Krugman wrote:
...the reason the economy didnít fall off a cliff immediately when the housing bubble burst was that, for a while, export growth took up the slack.
The sharp and significant decline in exports is worrisome, since exports have been maintaining US GDP growth thus far.
If the US has truly stopped being the consumer of last resort...
I am always amused when illusions are taken as cause. Krugman's comment is the most amusing because the US economy has been off the cliff for a long time. The truth is that the inclusion of imports in the statistics has actually simply masked the fact that we have been in recession for months.
Now in one sense that was a good thing because congress did not panic and send the stock market crashing until a few weeks ago. Had they been dealing with the recession earlier they might have caused the crash earlier. But then that gave the crash more time to build up potential energy so that when the congress did push the market over the edge it fell much faster and harder.
All this may be fun to watch but it is much ado about nothing. If your expenses are greater than your income you go get a job at a convenience store not pay your bills on a credit card - that is unless you are a government.
Posted by: Anonymous at November 18, 2008 06:13 AM
Andrea: The series plotted are "real" (Chained 2000$), so are volume numbers. I do agree the appreciation of the dollar should have an effect on trade volumes, but I think there will be some lags in these effects. Hence, the September figures are probably driven by activity (and perhaps trade credit/credit crunch issues).
Anonymous: I don't understand your point. The issue has been for the past decade or more, for households expenses have been greater than income, and we did pay our bills on credit card (same for the US government, except for 2000, in the last year of the Clinton Administration).
Point of information: You are claiming that if Congress had done nothing, then the stock market would be where it was back in say August?
Posted by: Menzie Chinn at November 18, 2008 09:37 AM
The conribution of the trade sector to GDP is exports less imports. Why not focus on this difference, rather than on exports alone? For example, if exports stay the same while imports fall, trade will contribute to GDP growth. If exports and imports fall in tandem, the effect of trade is nil.
As for the composition of trade changes, how about disaggregating between currency manipulators and all others, as well as between oil exporters and all others?
Posted by: don at November 18, 2008 10:01 AM
don: I'm not focusing on the growth decomposition arising from net exports because I've made this general point numerous times already, most recently in regard to the 08Q2 release.
In any case, here I'm trying to infer something about the future, rather than doing a mechanical decomposition.
Which countries are "currency manipulators"? Are all currency peggers manipulators. Are only countries that keep their exchange rate undervalued "manipulators"? And what constitutes "undervalued"?
Posted by: Menzie Chinn at November 18, 2008 10:32 AM
Point of information: You are claiming that if Congress had done nothing, then the stock market would be where it was back in say August?
Pretty much. Right now the whole world economy is frozen as everybody watches the FED, Treasury, and congress to see what they are going to do next. Markets can live with bad news or good news. What kills markets is uncertainty and that is exactly what we have now. Statistics don't mean a thing when you don't know whether the government is going to own your business tomorrow.
Posted by: Anonymous at November 18, 2008 01:17 PM
As exports shrink, imports also shrink, according to this discussion.
Therefore, won't oil fall even further in price?
Won't China also see a hard landing, economically?
Posted by: GK at November 18, 2008 02:18 PM
I define currency manipulators to be countries that engage in noticable official purchases of foreign currencies. Although there is the problem of implicit guarantees (such as offered by Japan, which allows the yen carry trade to be so strident).
Such beggar-thy-neighbor policies are much more effective than import tarrifs at 'exporting unemployment.' I am just curious about their importance, although I realize assigning the U.S. share of the adverse effects can be a problem.
Posted by: don at November 18, 2008 03:16 PM
If you make a chart comparing exports versus currency exchange rates, it will be quite apparent that the recent rise in the dollar compared with most European and Asian currencies explains much of the drop in sales of goods to foreign countries.
Also, our economic woes affected much of the world, so foreign consumers are buying less (just as we are buying less).
Posted by: Dr. T at November 18, 2008 04:26 PM
Good post...better than the comments (but then they are log-graph deprived...a distinct disadvantage) and although it was nice to see bsetser's comment...it was tiny...picky even...and might some of Boeing's orders be cancelled/postponed in any case, owing to the unprecendented diveandall?
But the jewel for me was the unique expression Ut-oh. which betters even Setser's prize winning "off" (as in "anony, your illusory remarks about causes were not amusing, but off.")
There is some tradition in the literature for the expression, "Uh-oh".
But Setser's is new.
The turbo-charged (ok, possibly slapped-in-the-gut...or other sensitive spot of your choice) "Ut-oh" capturing not only the staccato, but the animatto, the urgency...so missing in the Economic literature.
Posted by: calmo at November 18, 2008 09:12 PM
GK: Yes, cet. par. Maybe, depending upon your definition of "hard landing". Is 5-6% y/y growth a hard-landing in your definition? If so, seems plausible.
don: By your definition, and depending on one's view of the word ("noticeable"), most of the world's countries could be currency manipulators (just check the Reinhart-Rogoff categorization of how many countries are "free floaters" vs. total). In addition, please see this post.
Dr. T: It's true that the dollar has risen 11.4% (in log terms) July to October. However, a two month lag in the impact of exchange rates to exports is much shorter than typically identified in export-trade equations . I think it much more plausible to think of a common shock hitting export demand and the dollar (including revisions to expectations of future income). In other words, the elasticity of export demand with respect to foreign income is higher than with respect to the exchange rate at high frequencies.
calmo: I too wondered what the etymology of "ut-oh" was, since I was only familiar with "uh-oh". Perhaps somebody can add in whether this is a regional variant, or one that indicates greater urgency.
Posted by: Menzie Chinn at November 19, 2008 07:37 AM
Thanks for your responses. In the paper you cite, you talk about reversion to mean of the CA. But what if currency purchases keep the mean CA surplus high (and capital restrictions or other things keep private capital flows from being perfectly accommodating)?
I agree it might be hard to identify 'currency mercantilists,' but I suspect such policies impose bigger trade distortions than import restrictions and that efforts to curb them are on the way.
Did the Teletubbies use 'ut-oh'?
Posted by: don at November 19, 2008 10:43 AM
Menzie says, "My interpretation of this figure is that the consumption slowdown   signals the end of the role of the US as consumer of last resort."
This is what I have been trying to bring to the attention of your more optimistic co-blogger. All signs suggest that the US economy must transition from being a consumer economy to a savings/investment economy. Some other economy, such as India and/or China has to take over the role of temporarily spending beyond its means. Fortunately, they have lots of productive things to spend money on, things that we might well call investments.
The US has a LOT to do to get its house in order. Fortunately, the main steps that need to be taken to rationalize the economy are very obvious: cut health care costs and defense spending; re-build manufacturing. Less fortunately, there are a lot of trogs who will make doing these things difficult.
Posted by: Charles at November 19, 2008 03:51 PM
Don, "ut-oh" is a ScoobyDoo phrase.
Posted by: Charles at November 19, 2008 08:32 PM
don: I see your point, but using mean values over long time spans (as is done in this paper) means that we mitigate this problem; and note that until 2002, the Chinese CA was less than 2 ppts of GDP.
Since tariffs and quotas are more persistent than exchange rate misalignments (measured as deviations from relative PPP), I'm not sure how one can make the conclusion that you do.
Posted by: Menzie Chinn at November 21, 2008 01:59 PM