by Calculated Risk on 5/24/2005 12:27:00 AM

The April trade balance will be reported on June 10th. I'm trying to develop a simple model to help predict the trade deficit. Last Thursday, I posted an oil import model (I need to add exports and Seasonal Adjustments). This post will address China. I'll post my methodology and hopefully others will offer suggestions and improvements.

My general approach is to divide the deficit into two components: petroleum energy related products and everything else. This is a mixed model, by "goods" for petroleum, by country or region for everything else.

CHINA

My approach to China is to assume trade follows the normal seasonal pattern and recent growth trends. The seasonal number will be adjusted according to container traffic at the Port of Long Beach. If anything unusual has happened (change in exchange rate, labor strike, new tariffs, etc.) I will try to factor that into the estimate. For April, the only "unusual" event was the soft patch in the US economy. This probably did not impact imports, so I will assume nothing unusual for April.

US IMPORTS

First, here are the monthly imports from China for the last 6 years.



Click on graph for larger image.

Imports from China have increased every year, except in 2001 when imports were relatively flat (during the US recession). A couple of features stand out: There is a consistent seasonal pattern to imports; low early in the year, building towards the Holiday season and then dropping off at the end of the year.

The low for the year is usually in February, but it occasionally occurs in March like this year.

The second graph shows the last three years plotted against the best fit trend line. This clearly shows the seasonal pattern to imports.

Comment on Seasonal Adjustment: The seasonal adjustment is intended to remove the seasonal fluctuations from the data. This type of data is usually adjusted with a multiplicative approach: A = C x S x I where:
A = Observed Series (Not Seasonally Adjusted or NSA)
C = Trend Cycle
S = Seasonal Component
I = Irregular Component (weather, strikes, etc.)

Looking at the above graph, if the actual line is typical below the trend line for a given month, the observed numbers are adjusted upwards and if the line is usually above the trend, the oberved number is adjusted down. For April, the typical pattern shows about 4% below the trend line, so the observed number (NSA) will be adjusted upwards (SA). Errors occur if the trend changes. Also, the steeper the trend line, the more error prone the adjustment.


After reviewing the data, it appears that imports from China track inbound container shipments at the port of Long Beach. Of course this is comparing dollars to volume, so the mix of products has to remain relatively stable. The Port of Los Angeles also tracks imports relatively well, but there was a labor shortage last year and LB fit the data better.

This brings me to the first prediction for April: The trend for NSA imports for China is $17.8 Billion. Inbound container traffic was up 29% at both LB and LA ports (more than the usual March increase), so I'm going to adjust NSA upwards to $18.5 Billion.

US EXPORTS

The next step is to estimate the US exports to China. I'm going to use the same approach as for imports.


Exports to China do not show any seasonal pattern. There has been a steady increase with a slight jump in exports in 2003.

Since there is no seasonal pattern, the initial estimate is based on the trend line and the Port of Long Beach data. LB reported a 3% increase in loaded outbound containers, so the estimate will be the 3% higher than March or $3.4 Billion for April.


The final graph shows the relationship between containers and exports from LB. The bad news is the correlation is not as strong as for imports from China.

The good news is exports to China are relatively small compared to imports from China, so any error will have a minimal impact on the overall estimate.

The final step is to convert the NSA numbers to SA. Since there is little or no seasonal trend to US exports, $3.4 Billion is also the SA export number. For imports, the $18.5 Billion number is adjusted up by to $19.3 Billion.

The following table presents the actual for February and March (with estimate of the SA numbers) and the estimates for April.

CHINA TRADE BALANCE: Table numbers in Billions $
NOT SEASONALLY ADJUSTED

MONTH	NSA Balance	NSA Exports	NSA Imports
February	-$13.9	$3.08	$16.95
March	-$12.9	$3.3	$16.21
April	-$15.1(est)	$3.4(est)	$18.5(est)

SEASONALLY ADJUSTED (all estimates)

MONTH	SA Balance	SA Exports	SA Imports
February	-$18.1	$3.08	$21.19
March	-$15.1	$3.3	$18.42
April	-$15.9(est)	$3.4(est)	$19.3(est)

Note that February (usually a weak month for imports) was relatively strong and the SA number was probably over $21 Billion for imports from China, contributing to the record reported SA trade deficit.