Chronicling 100 Years of the U.S. Economy

June 2021
Volume 101, Number 6

Territorial Economic Accounts for American Samoa, the Commonwealth of the Northern Mariana Islands, Guam, and the U.S. Virgin Islands

New Estimates of GDP for 2019, New Estimates of GDP by Industry and Compensation by Industry for 2018, Revised Estimates for 2002–2018


In collaboration with the U.S. Department of the Interior Office of Insular Affairs (OIA) and with staff from the territorial governments, the Bureau of Economic Analysis (BEA) produces annual economic accounts for American Samoa, the Commonwealth of the Northern Mariana Islands (CNMI), Guam, and the U.S. Virgin Islands (USVI). The OIA provides funding for the work and facilitates interactions between BEA and the territorial governments.1

The purpose of this ongoing project is to provide data users with comprehensive, objective measures of economic activity for these four U.S. territories. For the United States, BEA produces gross domestic product (GDP) and other related economic measures as part of its National Income and Product Accounts (NIPAs); however, these estimates cover only the 50 states and the District of Columbia. Transactions with the territories are classified as transactions with the “rest of the world.”

Consistent measures of economic activity are critical for understanding the territorial economies and how they have developed over time. Without such measures, it is difficult for businesses and governments to make informed economic and financial planning decisions and for policymakers to assess the impacts of their decisions on growth. A primary challenge associated with producing these statistics is the lack of coverage of these four territories by most of the major surveys used by BEA to produce estimates of U.S. GDP and related economic measures.2 As described in the methodology appendix to this article, the work thus relies heavily on the provision of data from the territorial governments.

Between December 2020 and May 2021, BEA released new estimates of GDP for 2019 and revised estimates for 2002–2018.3 Real GDP growth rates for each territory are shown in chart 1.

These estimates reflect the results of a comprehensive, or benchmark, update to the territorial economic accounts. Comprehensive updates provide an opportunity to incorporate major improvements in how BEA measures the economies of the territories. This benchmark update incorporated improvements to source data and methods, including information from the following:

  • The Census Bureau's 2017 Economic Census of Island Areas, which provides the most thorough and detailed information on the structure of the territorial economies,
  • Newly available or revised source data provided by the territorial governments, and
  • Results from BEA's latest comprehensive and annual revisions to the U.S. national accounts.

Although the levels of GDP have been revised, the year-to-year patterns of inflation-adjusted growth are similar to the previously published estimates for all four territories (tables A.1.6, B.1.6, C.1.6, and D.1.6).

Highlights of the latest GDP estimates for each territory are described below.

American Samoa

Real GDP decreased 1.4 percent in 2019 (table A.1.3).

  • The decline in the American Samoa economy reflected decreases in government spending, exports, and private fixed investment (table A.1.4).4
  • Government spending decreased, primarily reflecting a decline in territorial government spending following the 2018 completion of a multimillion-dollar telecommunications construction project.
  • Exports of canned tuna and related products decreased in 2019, reflecting a decline in output of the tuna canning industry.
  • Private fixed investment declined sharply. This reflected decreases in construction activity and purchases of equipment following the completion in the prior year of reconstruction activities associated with Tropical Cyclone Gita.

Commonwealth of the Northern Mariana Islands (CNMI)

Real GDP decreased 11.2 percent in 2019 (table B.1.3).

  • The decline in the CNMI economy primarily reflected declines in gaming industry revenues and private fixed investment.
  • Exports of services was the largest contributor to the decline in real GDP in 2019, reflecting decreases in visitor spending, particularly on casino gambling (table B.1.4). According to publicly available financial statements, revenues from casino gambling dropped over 80 percent.
  • Private fixed investment decreased in 2019 but remained at historically high levels, supported by continued development of the casino resort on Saipan and reconstruction activities associated with Typhoon Yutu.


Real GDP increased 2.0 percent in 2019 (table C.1.3).

  • The increase in the Guam economy reflected increases in exports, private fixed investment, federal government spending, and consumer spending (table C.1.4).
  • Exports of services was the largest contributor to the increase in real GDP in 2019. This reflected a record high in visitor arrivals for the year, according to statistics published by the Guam government. Average spending by Korean and Japanese tourists, who make up much of Guam's tourist market, also increased.
  • Spending on construction and equipment for defense and private projects increased. Business spending on construction increased in 2019, following project delays in 2017 and 2018.
  • Consumer spending also increased, largely reflecting growth in retail trade activity and health care services.

U.S. Virgin Islands (USVI)

Real GDP increased 2.2 percent in 2019 (table D.1.3).

  • The growth in the USVI economy reflected increases in private fixed investment, exports, and consumer spending (table D.1.4).
  • Private fixed investment doubled from the previous year. A refinery and oil storage facility on St. Croix invested in major capital improvements during 2019. Additionally, disaster-related insurance payouts and federal assistance supported the reconstruction of homes and businesses that were destroyed or heavily damaged in 2017 by Hurricanes Irma and Maria.
  • Exports of services increased in 2019. This reflected an increase in visitor arrivals, as reported by the USVI government.
  • Consumer spending increased, reflecting growth in household purchases of goods, including motor vehicles and food and services, such as health care.

New estimates of GDP by industry and compensation by industry for 2018 were also included in the 2019 GDP news releases for each territory. These estimates, along with GDP for 2019, are presented in detail in the accompanying tables.5

The assistance and information provided by each of the territorial governments and the support from OIA will continue to be critical to the success of this project. See “Acknowledgments” for a list of key contributors.

In the near term, BEA is collaborating with the territorial governments and OIA to identify statistical improvement opportunities. Advances that support the development of high quality, more timely economic indicators for territorial decisionmakers will also improve the data used in the GDP estimates.

In the long term, BEA's goal is to integrate all of the territories into the full set of U.S. NIPAs. Although the inclusion of the U.S. territories in more federal economic surveys remains a key challenge, a recent achievement by the Bureau was the development and publication of prototype GDP statistics for Puerto Rico; see “Spotlight: Puerto Rico GDP.” With this step, BEA is estimating GDP for all five territories using modern national accounting methods, enabling direct comparisons with other BEA statistics.

The methodologies used to estimate GDP for American Samoa, the Commonwealth of the Northern Mariana Islands, Guam, and the U.S. Virgin Islands are consistent with the methods used to estimate GDP for the United States (excluding the territories). Information from the U.S. Census Bureau Economic Census of Island Areas (ECIA) is used to establish levels of spending for specific components for the benchmark years 2002, 2007, 2012, and 2017. For other years, annual estimates are developed using a combination of data sets collected by BEA from federal agencies, the territorial governments, and private sector sources. These data sets are listed in appendix tables A–D.

The U.S. territories are not included in many federal economic data sources used in national (or state-level) GDP. BEA relies on territorial government data in conjunction with the limited federal data that are available for its annual estimates of territory GDP. As noted in a May 2021 Survey of Current Business article, there are a number of implications to BEA's reliance on territorial government data.6 Notably, in addition to having distinct key industries and economic structures, each territory has a different set of survey and administrative data. Thus, while rooted in the same core expenditure-based GDP estimation framework that is used for BEA's national estimates, the specific methods and information used for each territory are unique to that territory's economic structure and available data.

Estimates of Current-Dollar GDP

Personal consumption expenditures

Personal consumption expenditures (PCE), or “consumer spending,” measures the goods and services purchased by “persons”—that is, by households and by nonprofit institutions serving households (NPISHs)—who are resident in each territory.

ECIA data on the consumer shares of sales by industry are used to benchmark the estimates of household purchases of most goods and selected services. However, for multiple categories of services, the ECIA does not fully reflect all consumer expenditures measured in GDP and so other data sources are used to benchmark these estimates. These other data sources are described below.

For services sold by governments, including hospital services, utility services, and higher education services, the data sources used to establish levels of spending for benchmark years are the same as the annual data sources listed in appendix tables A–D.

Data sources covering consumer purchases of housing services, financial services, and insurance services are currently limited. For these estimates, spending in most years is interpolated or extrapolated using broad nominal indicators of spending such as resident population multiplied by a relevant consumer price index. Data sources that are used to establish the levels of these services for select years are described below.

For housing services, information on the number of owner- and tenant-occupied housing units and average rental rates reported in the decennial Census of Population and Housing is used to estimate values in census years. When available, intercensal household expenditure surveys conducted by the territorial governments that report number of housing units and household spending on rents are also used.

For financial services, a key data source used to estimate spending on banking services was the Federal Deposit Insurance Corporation call report data. However, commercial bank mergers and acquisitions that cut across territories and states have created challenges in using these data to estimate more recent years of spending, as the activity of establishments in a single territory are no longer separately identified in most cases.

For insurance services, private-sector data sources on premiums collected by type of insurance were used in the calculation of consumer spending for some territories. However, these data sources are no longer available. Territorial government offices of the insurance commissioner may have more recent data that can be used to update and improve the estimates of insurance services; more research needs to be conducted in this area.

Note that the goods and services estimates described in this section can reflect on-island spending by both residents and nonresidents. To ensure that aggregate PCE reflects spending by residents of each territory (including their spending while traveling abroad), foreign travel expenditures by residents must be added and expenditures by nonresidents must be removed.7 Net foreign travel equals foreign travel expenditures by residents less expenditures in the territory by nonresidents and is shown as a separate series in the PCE tables for each territory. For American Samoa, Guam, and the CNMI, resident spending abroad is benchmarked using information from household surveys conducted by the territorial governments. For the USVI, resident spending abroad is estimated for all years using USVI resident spending data from Puerto Rico's balance-of-payment accounts and a scaling factor to capture spending by USVI residents in locations other than Puerto Rico. Data sources used to benchmark nonresident spending in each territory include visitor exit surveys conducted by the territorial governments and ECIA data on shares of sales to visitors by industry.

Private fixed investment

Private fixed investment (PFI) measures spending by private businesses, nonprofit institutions, and households on structures and equipment that are used in the production of goods and services.

ECIA data on capital expenditures by private-sector establishments are used to benchmark most of PFI. Construction of new homes is not covered in the ECIA capital expenditures data except when it is intended to be used as rental property for the real estate industry. Investment in new homes is benchmarked separately using ECIA product line data for the construction sector or territorial government data for residential building permits.

Change in private inventories

Change in private inventories (CIPI) is a measure of the value in current prices of the change in the physical volume of the inventories—additions less withdrawals—that businesses maintain to support their production and distribution activities.

Because of limited data, CIPI, or “inventory investment,” is estimated only for commodities that are especially significant to the territorial economy, can be stored for a relatively long time, and account for a substantial share of imports of goods. For American Samoa, this includes cannery supplies such as tin plates that are used to manufacture cans for tuna. For the USVI, this includes crude oil and other petroleum products that are stored at the refinery and oil storage facility in St. Croix.

There is currently no annual data available on the physical stocks of these products. Therefore, inventory investment is estimated based on imports, exports, and domestic consumption of these products. The key data sources used are listed in appendix tables A and D.

Net exports of goods and services

Net exports represents the difference between a territory's exports and its imports. Exports measures the portion of total production of goods and services in a territory that is provided to the rest of the world (including other countries, the 50 states and the District of Columbia, and other U.S. territories); imports measures the portion of total expenditures in a territory that is accounted for by goods and services provided by the rest of the world.

For exports of goods to the rest of the United States (that is, the 50 states, the District of Columbia, and other U.S. territories), all years reflect values for shipments published in the Census Bureau's U.S. Trade with Puerto Rico and U.S. Possessions (FT895) reports. For exports of goods to other countries, the Census Bureau's U.S. International Trade in Goods and Services (FT900) report covers shipments from the USVI and Puerto Rico only. For the other three territories, data on exports of goods to other countries are produced by the territorial governments.

For imports of goods from the rest of the United States and from other countries, the FT895 and FT900 cover shipments to the USVI and Puerto Rico only. For the other three territories, data on imports of goods are produced by the territorial governments. The data vary greatly in coverage, detail, and timeliness.8

The estimates of exports of services primarily reflect spending by visitors, which is estimated using the same data sources that are described above in the net foreign travel component of PCE. For American Samoa, exports of services also includes an estimate of services provided to foreign ships by the government port authority. Information on exports of other services is limited.

The estimates of imports of services include freight and insurance charges on imported goods, business and personal insurance services, construction-related services (for example, architectural, design, and other support services), and travel and transportation services. For the CNMI, imports of services also includes other business services that are assumed to be purchased primarily by the casino industry. These imports are derived by comparing estimates of intermediate expenses of CNMI businesses, including casinos, with information on revenues of establishments located in the CNMI; intermediate expenses that are not paid to CNMI businesses must be imported.9

Like PCE insurance services, data sources covering imports of insurance services are limited. Private-sector data sources are used when available; other years are extrapolated using broad nominal indicators such as employment.

Imports of construction-related services is the value of construction spending that reflects support services produced elsewhere. This equals the difference between the value of all construction spending (that is, the sum of private-sector investment in structures and government investment in structures) and the value of construction performed by domestic construction establishments.10 Except for American Samoa, the value of construction performed by domestic construction establishments is benchmarked using information on construction-sector revenues from the ECIA. For American Samoa, imports of construction services cannot be benchmarked in this manner due to limitations in the detailed construction-sector data from the ECIA.11 Instead, imports of construction services are estimated for years in which they are assumed to be substantial, using government financial statements and news reports.

Imports of travel and transportation services are estimated using the same data sources that are described above in the net foreign travel component of PCE. For American Samoa, imports of services also includes territorial government imports of transportation services.

The key annual data sources used to estimate exports and imports are listed in appendix tables A–D.

Government consumption expenditures and gross investment

Government consumption expenditures and gross investment—or “government spending”—measures final expenditures accounted for by the territorial government (including the primary government and its component units) and the U.S. federal government.

Government consumption expenditures consists of spending by government agencies, except government enterprises, to provide goods and services to the public.12 Consumption expenditures includes compensation of government employees, consumption of fixed capital, and purchases of intermediate goods and services. Sales to other sectors, which is the value of goods and services sold to the household and business sectors, is netted out to prevent double-counting of the expenditures across the components of GDP.

Gross investment consists of spending by all government agencies, including government enterprises, for structures and equipment that are used in producing the goods and services provided by the government.

The key data sources used to estimate federal government spending and territorial government spending are listed in appendix tables A–D.

Estimates of Real GDP

Inflation-adjusted, or “real”, estimates of GDP and its components are derived within a chain-type Fisher index framework.13 For most of the detailed components of GDP, inflation-adjusted estimates are calculated by deflating each component using an appropriate price index. Because of the limited availability of price indexes that are specific to the territories, U.S. price indexes are used for deflation in many cases. The territorial governments each produce their own consumer price indexes, which BEA uses within the estimates of consumer spending, but there are no producer price indexes for any territory, nor are there price indexes that cover trade in goods and services between each of the territories and the rest of the United States.

Personal consumption expenditures

Consumer price indexes (CPI) produced by the territorial governments are used to deflate many detailed components of consumer spending. In cases where the composition of consumer spending categories does not align well with the territory-specific CPIs, other price data are used, such as U.S. PCE price indexes from BEA or U.S. import and export prices from the U.S. Bureau of Labor Statistics (BLS). Exceptions include consumer spending on electricity for all territories and gasoline for American Samoa and Guam. Real consumer spending on electricity is estimated using quantity information on residential consumption by territory from the U.S. Energy Information Administration. Real consumer spending on gasoline for American Samoa and Guam is estimated using quantity information on gasoline consumption or sales provided by the territorial governments.

Private fixed investment

U.S. PFI price indexes from BEA and Producer Price Indexes from the U.S. BLS are used to deflate most of the detailed components. Currently, there are no territory-specific price indexes covering capital goods purchased by businesses in the island areas.

Change in private inventories

For American Samoa, cannery supplies held in inventory are adjusted for inflation using U.S. import, export, and producer prices from the U.S. BLS. For the USVI, inventories of crude oil and petroleum products are adjusted for inflation using U.S. import and export prices from the U.S. BLS.

Net exports of goods and services

For exports and imports of goods, a variety of price indexes are used, including territory CPIs and U.S. import, export, and producer prices from the U.S. BLS. Note that the U.S. BLS International Price Program treats the territories as part of the United States; thus, import and export price indexes in theory reflect prices for goods traded between the territories and other countries. However, the prices paid and received by the territories cannot be separately identified from the prices paid and received by the rest of the United States. For exports and imports of services, territory CPIs and price indexes from BEA estimates of U.S. trade in services and U.S. consumer spending are used.

Government consumption expenditures and gross investment

For compensation, information on average wages and salaries derived from administrative or survey data is used to deflate compensation. For other components of government spending, government price indexes from BEA's U.S. National Income and Product Accounts are used.

Additional information on source data

The annual data sources listed in appendix tables A–D include a variety of economic measures, such as sales or receipts, expenses, wages and salaries, and tax collections.

For many components, the indicators are derived from source data that are “value data”: they encompass both the quantity data and the price data required to prepare the current-dollar estimates. For indicators that are not derived from value data, a “quantity-times-price” method is used. For example, the indicator for Guam consumer spending on gasoline is calculated as gasoline sales in gallons times the Guam consumer price index for motor fuel.

All fiscal year data sources are adjusted to a calendar year basis for use in estimation.


  1. OIA is the federal agency that manages the federal government's relations with the governments of American Samoa, the CNMI, Guam, and the USVI. It works with these territories to encourage economic development, transparency of government, financial stability, and accountability.
  2. These surveys include the Annual Wholesale Trade Survey; the Annual Retail Trade Survey; the Annual Capital Expenditures Survey; the Value of Construction Put in Place Survey; the Service Annual Survey; the Annual Survey of Manufactures; the Manufactures' Shipments, Inventories, and Orders Survey; the Annual Survey of State and Local Government Finances.
  3. Individual news releases for each of the territories are available on BEA's website.
  4. In this article, “consumer spending” refers to “personal consumption expenditures” and “government spending” refers to “government consumption expenditures and gross investment.”
  5. The industry detail shown for GDP by industry and compensation by industry varies depending on the territory.
  6. Sabrina L. Montes and Aya Hamano, “Estimating GDP for the U.S. Territories,” Survey of Current Business 101 (May 2021).
  7. “Foreign travel expenditures by residents” includes spending in the rest of the world, that is, other countries, the 50 states and the District of Columbia, and other U.S. territories.
  8. Guam government data on imports of goods are especially limited. Data are not collected for all periods and do not represent the universe of imports of goods.
  9. Intermediate expenses are purchases of goods and services that are used in the production process of other goods and services and are not sold in final-demand markets.
  10. The estimation of private-sector and government gross investment in structures is described in the PFI and government spending sections of this appendix, respectively.
  11. For Guam, imports of construction services for defense projects are benchmarked separately using information on U.S. Department of Defense contracts reported by the Guam Bureau of Labor Statistics.
  12. Government enterprises are government agencies that cover a substantial portion of their operating costs by selling goods and services to the public and that maintain their own separate accounts.
  13. For additional information, see J. Steven Landefeld, Brent R. Moulton, and Cindy M. Vojtech, “Chained-Dollar Indexes: Issues, Tips on Their Use, and Upcoming Changes,” Survey of Current Business 83 (November 2003):8–17.