MLR December 2005 article

Program Report

Program Report

IPP introduces

additional Locality
of Origin import
price indexes

the statistical stability of calculated indexes. This article discusses trends
found in these newly-published Locality of Origin import price indexes.

Background
Helen McCulley
and
Melissa Schwartz
The International Price Program (IPP ) of
the Bureau of Labor Statistics, as the primary source of data on price changes in
the foreign sector, publishes monthly indexes of import and export prices for
U.S. merchandise.1 While such indexes
convey price information across product categories of goods traded between
the United States and the rest of the
world, there is evidence that price trends
further vary by the geographic source of
the product being traded.2 U.S. Locality of Origin (LOO) import price indexes
were first published by IPP in 1992 for
the following groupings, geographic regions, and countries: industrialized and
other countries, Canada, European
Union, Japan, and the Asian Newly Industrialized Countries (NICs); in 1997
the Latin America locality was added to
publication.3 Since 1992, other countries and regions such as China and
Mexico have emerged as important trading partners with the United States.
Thus, in January 2005, price index series for these two countries were added
to the set of published LOO price indexes
along with six other localities: France,
Germany, the United Kingdom, the Pacific Rim, the Association of South East
Asian Nations ( ASEAN), and the Asia Near
East.4 (See box.) The new localities were
determined according to customer interest, having a sufficient number and variety of usable item prices to reflect the
actual dollar value and type of trade, and
Helen McCulley is a mathematical statistician,
and Melissa Schwartz a former economist, in the
International Price Program, Bureau of Labor
Statistics.
E-mail:
McCulley.Helen@bls.gov
Melissa.Schwartz@gmail.com

36

Monthly Labor Review

The motivation for producing price indexes by geographic region of origin is
twofold. First, the types of products being traded differ across localities; therefore, price indexes across localities
should exhibit different trends that could
not otherwise be observed from the
world goods price indexes. For example, the proportion of trade in manufactured goods is relatively higher for
industrialized countries than for developing countries.5 As such, petroleum
and other raw materials prices have a
lesser impact than manufactured goods
prices in the industrialized LOO price index than in the other LOO price index.
(See chart 1.)
Second, the U.S. dollar’s fluctuation
against foreign currencies has an impact
on internationally traded products. The
magnitude of the influence of currency
fluctuation on price levels (often referred to as the pass-through rate) depends on a variety of factors. 1) Historically, raw materials prices have been
more independent of exchange rate fluctuations than finished goods prices. 2)
The magnitude and duration of exchange rate movements also impact the
pass-through rate—larger and more permanent fluctuations are more likely to
pass through to prices. 3) A particular
industry’s pricing conventions, such as
longer durations between negotiations
among buyers and sellers, tend to result
in less responsiveness to exchange rate
fluctuations. 4) The impact of exchange
rate movements on transaction prices
between trade partners may vary depending on whether said trade is intrafirm or not. 5) Finally, the degree of
competitive pressures in an industry can
determine whether a seller absorbs exchange rate fluctuations or passes them
on to selling prices.

December 2005

The set of LOO price indexes selected
for publication was determined according to the current levels of pricing data
collected monthly by the IPP as part of
the voluntary survey sample of importing and exporting U.S. companies. To
guarantee accuracy and stability of a
price index by locality of origin, an index must contain consistent and abundant price information. The methods for
selecting potentially publishable LOO
indexes based on accuracy and stability
are outlined in the “Other decision criteria” section of this article. Data on
imported products from the Consumption Entry Documents collected by the
U.S. Customs Bureau serve as an information source on the value and type of
trade with foreign countries. These data
serve as weights across product categories within an LOO index and are updated
annually to reflect frequent shifts in
trade. The preferred price basis for imports is f.o.b. (free on board), which is
the price at the foreign port of exportation before insurance, freight, or duty
are added. The product universe for
constructing price indexes is defined as
all merchandise that is consistently
traded, excluding works of art, military
items, and used items.
The LOO indexes are constructed using a modified Laspeyres index formula
and the North American Industrial Classification System (NAICS) for the aggregation structure. An updated classification system reflecting new and emerging
industries, NAICS has been implemented
or is in the process of being implemented across many Federal statistical
agencies to provide a consistent conceptual framework. The NAICS further allows the LOO indexes to be published—
publishability standards permitting—at
the disaggregated “manufacturing” and
“nonmanufacturing” categories.

Recent trends
The United States is the world’s largest market for other exporting countries. In 2004, it imported more than

been one of world’s
fastest-growing; its
gross domestic
product (GDP ) has
Industrialized countries
grown from around
Other countries
$600 billion in
Canada
1978 to well more
European Union
than $5 trillion in
France
2003. Direct forGermany
eign investment inUnited Kingdom
creased to nearly
Latin America
$50 billion, up from
Mexico
less than $300 milPacific Rim
lion in 1978, while
China
the volume of imJapan
ports and exports
Asia Newly Industrialized Countries (Asian NICs)
both increased by
Association of Southeast Asian Nations (ASEAN)
well more than
Asia Near East
1,000 percent over
the same period. 7
NOTE: Localities in bold were added to publication in 2005.
All other LOO indexes were first published in 1992, except Latin
More recently, the
America which was first published in 1997
growth of China’s
imports, particularly for raw mate$1.3 trillion worth of merchandise—
rials, has been attributed as a factor in
accounting for more than 17 percent
rising world spot prices for raw materiof the total value of world-wide imals and energy.8 Their fuel and raw ma6
ported goods. Of the total 2004 U.S
terials imports have increased more than
imports, half came from the four top
250 percent over the last decade. Chitrading partners: Canada (17 pernese demand for finished goods has also
cent), China (13 percent), Mexico (11
increased substantially over the same
percent), and Japan (9 percent). (See
period: manufacturing imports have inchart 2.) When the Locality of Origin
creased more than 200 percent.
indexes were first introduced in 1992,
China has also become a major proCanada and Japan were overwhelmducer of manufactured products. Its exingly the top suppliers of merchandise ports of manufactured products have into the United States. (See chart 3.)
creased well more than 400 percent in
In the time since, the volume of trade
the past decade. The United States is
with the United States for both
China’s largest market, and approxiMexico and China substantially in- mately 40 percent of China’s exports
creased; imports from China have
were purchased by the United States in
grown a staggering 665 percent since
2003, consisting mostly of the follow1992, and imports from Mexico have
ing manufactured items: office and
grown 343 percent over the same pe- household machinery, telecommunicariod. What has changed?
tions and electronic equipment, furniture, textiles, clothing, and footwear.
Asia-Pacific region. In 1978, the year
Imports from the Pacific Rim region
Deng Xiaoping announced China’s
as a whole totaled nearly $500 billion in
Open Door Policy, market-oriented eco- 2004, more than doubling the import
nomic reform began in China, including
dollar value since 1992. Adding to the
the opening up of markets to world
set of price indexes for imports from
trade. Since then, China’s economy has
Asian-Pacific economies, the InternaU.S. locality of origin (LOO) import price indexes

tional Price Program is introducing indexes for the Pacific Rim region and the
Association of South East Asian Nations
(ASEAN ). Comprised of 14 Eastern
Hemisphere nations—including China,
Japan, and Australia—the Pacific Rim
is the most aggregated regional price
index for that part of the world. The
ASEAN was established in 1967 with the
mission of providing not only economic
integration, but also cooperation in social areas such as health, labor, poverty,
women’s and children’s issues, education, and disaster management. Its
population extends to nearly 500 million
people with a collective GDP of nearly
$686 billion, and exports to the United
States totaling nearly $82 billion in
2003.
Chart 4 displays the China, Pacific
Rim, and ASEAN Locality of Origin import price series for 2004. Because such
small percentages of imports from these
regions are of nonmanufactured goods,
the price index for all imports excluding petroleum is included in the chart for
comparison purposes. The data show
that the price index for imports from
China is stable throughout 2004; during
this time the U.S. dollar equaled roughly
8.28 yuan, which has been the exchange
rate since October 1998. Prices for imports from the ASEAN have drifted
slightly downward over the year, evidence of the falling price trend for world
computers and electronics, an industry
area that comprises approximately 57
percent of imports from the ASEAN region. Prices for imports from the Pacific Rim region as a whole, like those
from China, were relatively flat during
2004.
Mexico. Mexico has also become an
increasingly important trade partner
with the United States. Since the enactment of the North American Free
Trade Agreement (NAFTA ) in 1994,
Mexico’s trade with the United States
and Canada has tripled.9 Despite the
economic crisis that began in late
1994 and resulted in a large current

Monthly Labor Review

December 2005

37

Program Report

Chart 1.

Comparison of world import price indexes with industrialized and other countries
LOO indexes
Annual percent change

Annual percent change
15

15

10

All imports

10

Industrialized countries
Other countries

5

5

0

0

-5

-5

All imports less petroleum
-10

-10

-15

-15
1993

Chart 2.

94

95

96

97

98

99

2000

01

03

04

Proportion of U.S. imports value held by top trading partners, 1992 and 2004, in percent
1992

2004

(Total = 100 percent)

(Total = 100 percent)

Japan
18
Mexico
7
China
5
Germany
5

Mexico
11
Canada
19

Canada
17

Germany
5
Other
42

United
Kingdom
4
United Kingdom
3
NOTE: Data are in millions of U.S. dollars, not rebased because these are percentages of total.

Monthly Labor Review

Japan
9

China
13

Other
42

38

02

December 2005

Chart 3.

U.S. imports from top trading partners, by U.S. dollar value of trade

U.S. dollars
(in millions)

U.S. dollars
(in millions)

300,000

300,000
Canada

Japan

Mexico

China

250,000

250,000

200,000

200,000

150,000

150,000

100,000

100,000

50,000

50,000

0

0
1993

Chart 4.

94

95

96

97

98

99

2000

01

02

03

04

Comparison of Association of Southeast Asian Nations (ASEAN), China and Pacific Rim
locality of origin indexes with the nonpetroleum import index, 2004

Index
(Dec. 2003 = 100)

Index
(Dec. 2003 = 100)

105

105

104

104

Nonpetroleum import index
103

103

102

102

Pacific Rim
101

101

100

100

China

99
98

99
98

ASEAN

97

97

96

96

95

95
Jan.

Feb.

Mar.

Apr.

May

June

July

Aug.

Sept.

Oct.

Monthly Labor Review

Nov.

Dec.

December 2005

39

Program Report

account deficit necessitating the
Mexican government to float the
peso—which subsequently lost half of
its value against the U.S. dollar—
Mexico’s export sector rebounded
quickly. Mexico’s GDP grew an average of 12 percent per year from 1996
to 2000; and in 1995, Mexico began
running trade surpluses with the
United States. Such surpluses have
grown to more than $40 billion, and
in 2001 Mexico overtook Japan as the
United States’ second largest trading
partner (after Canada) before more
recently being surpassed only by
China. Approximately 85 percent of
Mexico’s exports go to the United
States, comprising nearly one-quarter
of Mexico’s GDP.10 Mexico is not only
the fourth largest supplier of petroleum to the United States, but also a
significant supplier of manufactured
goods such as motor vehicle parts and
electronic equipment. Furthermore,
intra-company trade plays a key role
Chart 5.

in U.S.-Mexico trade: about 64 percent of U.S. imports from Mexico and
about 35 percent of U.S. exports to
Mexico represent related party trade.11
Chart 5 displays the path of the import price index for goods from Mexico
in 2004 along with the all imports price
index. Import prices from Mexico
trended upward over the year in a tendency similar to overall import prices,
though the former showed a more pronounced increase and subsequent decrease resulting from the sharp petroleum price movements in the fall.
European countries. The International
Price Program has published a Locality
of Origin price index for the European
Union since 1992. In 2005, price indexes for imports from France, Germany, and the United Kingdom were
individually added to the set of published indexes to provide a more comprehensive picture of the behavior of
import prices from that region. To-

gether, the three countries make up more
than half of the total dollar volume of
imports from the European Union—and
each is a significant contributor to the
region’s production of motor vehicles
and chemicals, the industries accounting for the largest share of U.S. imports
from the European region.
Chart 6 plots the import price indexes from Germany, the United
Kingdom, and France along with a
U.S. dollar-euro exchange rate index
for comparison. 12 The U.K. import
price index diverges from the two
other countries’ indexes and is attributable to the effect of world petroleum
prices on the United Kingdom’s refined petroleum industry. A noteworthy phenomenon in recent years is the
U.S. dollar’s weakening against major foreign currencies, particularly the
euro. Between its June 2001 peak and
January 2005, the dollar has lost more
than 30 percent of its value against the
euro.13 However, chart 6 reveals that

Comparison of the import price index with the Mexico locality of origin (LOO) index, 2004

Index
(Dec. 2003 = 100)

Index
(Dec. 2003 = 100)

110

110

108

108

106

106

All import
104

104

102

102

Mexico LOO

100

100

98

98
Jan.

40

Feb.

Mar.

Monthly Labor Review

Apr.

May

December 2005

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

Chart 6.

Comparison of France, Germany, and United Kingdom locality of origin indexes with U.S.
dollar to 1 euro exchange rate index, 2004

Index
(Dec. 2003 = 100)

Index
(Dec. 2003 = 100)

110

110

108

108

United Kingdom
106

106

France

104

104

102

102

100

100

Germany

98

98

96

96

U.S. dollar to 1 euro exchange rate

94

94

92

92

90

90
Jan.

Feb.

Mar.

Apr.

any impact of the exchange rate
movement appears inconclusive in
both the Germany and France series,
which were flat over the year—suggesting that exchange rate fluctuations were not passed through to import prices from major European
trade partners to any notable degree.
Other Asia. The Asia Near East price
index, which represents more than
$40 billion in import merchandise
value in 2003, is expected to be dominated by the behavior of petroleum
prices, which account for nearly 60
percent of its exports to the United
States.14 Indeed, it can be seen in
chart 7 that the series tracked closely
with the world price index for petroleum in 2004. The remaining composition of imports from this region includes apparel, chemicals, and
diamonds.

May

June

July

Aug.

Other decision criteria
The feasibility research for determining
acceptable additions to the set of published Locality of Origin indexes incorporated the evaluation of several criteria:
annual dollar values of trade, goodnessof-fit measures, and variance analysis.
First, the country’s or region’s trade dollar value with the United States generates
customer interest—presumably, higher
trade flows garner more public interest,
particularly with individual countries such
as China and Mexico.
In addition to customer interest, two
goodness-of-fit statistics were used to
compare the distribution of price quotes
across disaggregated index strata to the
distribution of trade dollar values across
those index strata. Goodness-of-fit is
especially important in determining the
robustness of Locality of Origin indexes
because the International Price Program

Sept.

Oct.

Nov.

Dec.

samples from the universe of import and
export transactions according to trade
dollar values across product categories,
rather than the trade dollar values across
localities.15 The goodness-of-fit measures thus provide a picture of how well
the sampling process represents the distribution of trade by locality.
The first goodness-of-fit statistic
bases the distribution of price quotes on
the total number of prices requested,
while the second goodness-of-fit statistic bases the distribution of quotes on
the total number of usable prices. The
general form of the goodness-of-fit statistic is

GOF

=

 γ
φ
∑ γ i  Γ i − Φ i
i∈ S
 S
S
∑ γi

Monthly Labor Review






2

i∈ S

(1)

December 2005

41

Program Report

Chart 7.

Comparison of the Asia Near East locality of origin index and the import petroleum index
monthly percent changes, 2004

1-month
percent change

1-month
percent change

15

15

Petroleum imports
10

10

Asia Near East
5

5

0

0

-5

-5

-10

-10

-15

-15
Jan.

where

Feb.

Mar.

Apr.

γ i is the sampled dollar value for

stratum i ;

ΓS is the sampled dollar

value of trade in stratum i’s parent stratum S;

φ i is the number of prices in stra-

tum i;

Φ S is the number of prices in

stratum

i ’s

parent

stratum

S;

and GOF

∈ [0,1] . As the value of
16

GOF gets closer to zero, the closer the
distribution of prices is to the distribution of trade dollar value, indicating that
the number of prices requested or collected is appropriately distributed to
match trade patterns within a particular
price index.
Acceptable upper-bounds for the two
goodness-of-fit statistics were found by
applying the statistics to price indexes
for import products from the world published under the Harmonized products
classification system. 17 The upper-

42

Monthly Labor Review

May

June

July

Aug.

bound was set at .05 for both goodnessof-fit statistics because approximately
95 percent of the two-digit-level Harmonized strata produced values of less than
.05 for both versions of the goodnessof-fit statistic.18 Therefore, LOO price
indexes considered for publication
should fall below the same upper-bound
as the Harmonized strata; that is, a potentially publishable LOO price index
should have a goodness-of-fit result
equal to or less than .05.
Estimating the variances of price indexes is desirable as a measure of accuracy and stability. Variance estimates
were obtained through a bootstrapping
method to estimate the variability of the
annual change of price index values.19
The set of prices (known henceforth as
item sets) for each potentially publishable country or region was independently re-sampled with replacement to
obtain equivalent stratum-level item set

December 2005

Sept.

Oct.

Nov.

Dec.

sizes. This re-sampling was performed
50 times to create 50 item set realizations for each locality. For each replicate item set, bootstrap item weights
were calculated by multiplying the original item weights by the number of times
each item was randomly selected. 2 0
These re-sampled item sets, along with
the adjusted item weights, were then
used to create 50 realizations of each
locality’s price index series. Letting PI
denote the price index value, l the locality, s the stratum of interest, i the replicate number, and t the time period (representing a monthly observation), the
replicate annual changes in the price index values for each stratum within a locality were calculated as

?θ

l, s, i, t

=

PI
PI
(2)

l, s, i, t

l, s, i, t − 12

And variances were calculated in the
usual way as

Varl, s,t

2
1 50
=
∑(?θl,s,i,t − ?θl,s,t )
50 − 1 i =1
(3)

where

?θ

l, s, t

=

1
50

50

∑

?θ

l, s, i, t

i=1

(4)
Baselines for acceptable variance levels
were established by calculating variances for the LOO indexes already published and for the Harmonized classification system import index. In general,
variances for LOO indexes exceeded the
variances for Harmonized import price
indexes—an expected result because the
number of prices in the world Harmonized indexes are greater than the number of prices in the LOO indexes. However, most fell within the Harmonized
variances and the existing Locality of
Origin indexes’ variances. Locality of
Origin countries and regions were then
ranked and selected according to the
number of periods that variances fell
below the lowest variances for existing
LOO indexes; the number of periods that
variances fell between the lowest and
highest variances for existing LOO indexes; and the number of periods that
variances fell above the highest variances for existing LOO indexes.

Conclusion
The addition of newly-published Locality of Origin import price indexes to
data offered by the International Price
Program enhances the set of price indexes available to measure different aspects of inflation in merchandise markets. In 2004, price indexes for imports
from Mexico, the United Kingdom, and
the Asia Near East have trended with

world petroleum prices, while the indexes for China, the Pacific Rim,
France, Germany, and the Association
of Southeast Asian Nations have been
comparatively flat. The eight new Locality of Origin import price indexes are
publicly available dating back to December 2003. It was not feasible to calculate indexes prior to December 2003
because the classification structure at
the most disaggregated level changes so
frequently that the market basket beyond a 1-year history cannot be reconstructed. Trade shifts are especially
critical for LOO indexes; U.S. importers
regularly change suppliers, which may
not reside in the same locality as previous suppliers.

Notes
1
The Bureau of Labor Statistics also produces import and export price indexes for a
set of services industries. The services sector
is not included in locality of origin price indexes, and so is excluded from the discussion
here.
2
See “New international price series published by Nation and region,” by Michelle Albert
Vachris, Monthly Labor Review, June 1992, pp.
16–22.
3
The Locality of Origin import price indexes
measure prices at the point of exportation to the
United States. Prior to January 2003, the “Industrialized Countries” and “Other Countries”
categories were termed, respectively, “Developed Countries” and “Developing Countries.”
“Industrialized Countries” includes Western Europe, Canada, Japan, Australia, New
Zealand,and South Africa, and “Other Countries” includes all other countries not comprising “Industrialized Countries.” The Asia NIC s
category includes Hong Kong, Singapore, South
Korea, and Taiwan.
4
The Pacific Rim countries are Australia,
Brunei, China, Hong Kong, Indonesia, Japan,
Macao, Malaysia, New Zealand, Papua New
Guinea, Philippines, Singapore, South Korea, and
Taiwan.
5
In 2003, the dollar value of nonmanufactured items comprised just more than 8 percent of industrialized imports compared with
nearly 17 percent for other countries.
6
Trade data collected from the Foreign Trade
Division of the U.S. Census Bureau.
7
World Bank 2004 World Development Indicators CD-ROM.

8
See “A hungry dragon,” The Economist,
September 30, 2004.
9
See CIA -The World Factbook, on the Internet
at http://www.cia.gov/cia/publications/
factbook/geos/mx.html.
10
Trade data collected from the Foreign Trade
Division of the U.S. Census Bureau and from the
Organisation for Economic Co-operation and
Development.
11
Related-party trade information collected
from the U.S. Bureau of the Census, on the
Internet at http://www.census.gov/foreigntrade/Press-Release/2003pr/aip/rp03-exh1.txt.
12
The exchange rate is defined as the
monthly average of the U.S dollar to 1 euro, and
the exchange rate index is set according to December 2003=100, and then using the monthly
percent changes to create subsequent index values. Note that the United Kingdom employs the
British pound as its currency rather than the
euro.
13
As of January 2005. In June 2001, $1=1.172
euros; in January 2005, $1=.762 euros.
14
The countries included in the Asia Near
East region include the following: Bahrain, Iran,
Iraq, Israel, Jordan, Kuwait, Lebanon, Oman,
Qatar, Saudi Arabia, Syrian Arab Republic,
United Arab Emirates, and Yemen. This definition is based on that used by the U.S. Census
Bureau. Other definitions include countries in
northeastern Africa and/or all countries along the
southern and eastern parts of the Mediterranean
Sea. The area is also frequently referred to as
the “Middle East.”
15
For additional details of the International
Price Program’s sample design, see Chapter 15
of the BLS Handbook of Methods, on the Internet
at h t t p : / / s t a t s . b l s . g o v / o p u b / h o m /
homch15_a.htm.
16
Disaggregated strata are termed “child
strata” when considered relative to “parent”
strata, which are the next broadest level in an established classification structure.
17
The Harmonized system is used for product classification during the sampling process
in the International Price Program and so is thus
assumed to offer the most appropriate baseline
measure.
18
Harmonized import and export price indexes are published at the following levels: Section, Chapter (2-digit), and 4-digit levels. The
HTUSA (import) codes are maintained by the U.S.
International Trade Commission and the Schedule B (export) codes are maintained by the U.S.
Census Bureau.
19
Annual percent changes are less noisy than
monthly percent changes.
20
The original item weights are those used
in the calculation of the import and export price
indexes and are based on probability sampling
techniques. The weights are a function of the
product category’s and company’s importance in
trade.

Monthly Labor Review

December 2005

43