U.S. travel and tourism satellite accounts for 1992.

Okubo, Sumiye ; Planting, Mark A.

This article presents prototype travel and tourism satellite accounts (TTSA'S), which are a second set of extensions developed in 1998 to the U.S. input-output (I-O) accounts.(1) Satellite accounts are rearrangements of information from the national economic accounts and other sources for the purpose of analyzing specific economic activities more completely than is possible within the structure of the basic accounts (see the box "Satellite Accounts").(2) The TTSA'S integrate information on the flows of commodities that are related to travel and tourism activities but that are not identified in the standard presentation of the I-O accounts.(3)

There is strong motivation to develop analytical measures of travel and tourism activities in the United States. Travelers are important consumers of U.S. production, and the industries that cater to travelers use a substantial share of output from other industries, add substantial economic value to other industries' outputs, and employ large numbers of people. Furthermore, both travel in the United States by foreigners and travel abroad by U.S. residents have grown dramatically in recent years.

The TTSA'S define travel and tourism as the economic activity generated inside the United States by "visitors" of all types--for business and pleasure, by residents and nonresidents alike--and outside the United States by U.S. residents.(4) The TTSA'S extend the I-O accounts in that they attempt to measure an economic activity (travel and tourism) undertaken by only a subset of purchasers (visitors) and involving only a subset of purchases (tourism demand). This task first requires the identification of the commodities that are purchased by visitors and the corresponding industries that produce these commodities. This task is further complicated because tourism is inherently defined in relative geographic terms-like distance from home--and because many of the activities that are undertaken by visitors--such as dining out in restaurants--are also undertaken by nonvisitors, that is, people who are close to their homes. Therefore, deriving the output and value added of tourism industries is less straightforward than for a conventional industry producing a conventional commodity, such as iron and steel.

The basic building blocks of I-O accounts are commodities, most of which are not readily distinguishable by type of consumer. Therefore, in developing the TTSA'S, the share of each commodity purchased by visitors had to be estimated. The information available to allocate commodities between visitors and nonvisitors is generally based on relatively small sample surveys and indirect methods. In the prototype TTSA'S, three different methodologies were used, and estimates are presented as a range, rather than as a single estimate.

The following are highlights from the new TTSA'S for 1992 (table 1):

* Value added in travel and tourism represented 1.9-2.2 percent of U.S. gross domestic product (GDP). The industry with the highest value added was the hotels and lodging industry.

* Expenditures for travel and tourism accounted for 4.6-5.3 percent of U.S. GDP. The largest category of expenditures was expenditures for passenger air travel services.

* Employment in travel and tourism activities accounted for 3.2-3.7 percent of total employment in the United States. The average compensation per tourism employee was $21,400 per year, but compensation varied widely by industry.

Table 1.--Key Indicators of Tourism Activity: Range of
Estimates 1992

 De- Value Employ- Percent
 mand added ment Share of GDP Share
 (billions (billions (thou- of
 of of De- Value employ-
 dollars) dollars) sands) mand added ment

Method 1. 284.2 120.5 3,749 4.6 1.9 3.2
Method 2. 294.9 124.5 3,933 4.7 2.0 3.3
Method 3. 332.8 135.7 4,353 5.3 2.2 3.7

NOTE.--See the section "Methodological Overview" for a discussion of the three methods.

The first section of this article describes the development of the TTSA'S. The second section provides a conceptual overview of the TTSA'S, including their relationship to the I-O accounts. The third section describes the major components of the TTSA'S. The fourth section provides an overview of the methodology used to estimate the TTSA'S. The fifth section summarizes the TTSA estimates for 1992. The final section outlines future work and extensions.

Background

The activities of travel and tourism are covered in the national economic accounts, but the system underlying the classification of output in the I-O accounts--the Standard Industrial Classification (SIC) system--does not facilitate separately identifying tourism, as the SIC was primarily designed to present industry statistics without regard to the purpose of the purchase of output. A measure of tourism activities would be understated if it included only the output of industries that are typically associated with tourism activities--hotels and air, water, and rail transportation--because it would exclude expenditures on other types of commodities, such as eating and drinking places, that represent a relatively important share of tourism expenditures but that cannot be separately identified. On the other hand, that measure would be grossly overstated if it included all the expenditures on eating and drinking because it would also include expenditures by local residents.

Various measures of tourism have been developed, including the number of travelers, the number of trips made by U.S. residents in the United States and abroad, and the level of expenditures of U.S. resident and nonresident visitors on passenger fares, lodgings, and other goods and services.(5) However, these measures do not provide a consistent way to compare travel and tourism with other economic activities.

A White House Conference on Travel and Tourism, held in October 1995, highlighted the difficulty in linking these measures to other production and consumption activities in the economy. The Conference delegates recommended that the U.S. Department of Commerce, in partnership with the travel and tourism industries, develop travel and tourism satellite accounts to provide measures that are consistent with the U.S. national economic accounts. This recommendation was strongly supported by the Tourism Policy Council, headed by the Secretary of Commerce. As a result, in 1997, the Tourism Industries Office of the International Trade Administration, U.S. Department of Commerce, entered into an agreement with the Bureau of Economic Analysis to develop the U.S. TTSA'S.

Efforts to develop travel and tourism satellite accounts

The definitions, framework, and estimating methods used for the U.S. TTSA'S follow, as closely as practicable, the guidelines for similar accounts that were developed by the World Tourism Organization (WTO) and the Organisation for Economic Co-operation and Development (OECD). Over the past decade and a half, the WTO and the oEcD have prepared a series of reports that define tourism, provide recommendations on the collection and organization of tourism statistics, and establish guidelines for estimating travel and tourism satellite accounts.(6) The definitions and concepts developed by the WTO have been followed, with minor modifications, by the OECD, by the Statistical Office of the European Communities (EUROSTAT), and by countries collecting tourism statistics and estimating tourism satellite accounts. The WTO and the OECD are working cooperatively to develop manuals for estimating tourism satellite accounts. With these TTSA'S, the United States joins Canada and Norway as countries that have designed and implemented travel and tourism accounts in accord with the WTO and OECD initiatives.(7)

The OECD and WTO frameworks include sets of tables that measure similar aspects of travel and tourism. The measures common to both frameworks include the production and value added of tourism industries, purchases of tourism commodities and other commodities by domestic and international visitors, employment in tourism industries, and investment and capital stock in tourism industries.(8)

Conceptual Overview of Travel and Tourism Satellite Accounts

The purpose of the travel and tourism satellite accounts is to provide a framework for analyzing tourism expenditures in a systematic and consistent way that links tourism demand expenditures to the industries that produce tourism goods and services.

TTSA definitions

The TTSA'S for the United States rely primarily on the WTO and OECD definitions of visitors, tourism, and tourism expenditures (table 2). The TTSA'S define visitors as persons taking a trip or traveling outside of their "usual environment," and tourism as the activities of visitors while traveling. The accounts require the following definitions: The visitor and the usual environment.

Table 2.--Alternative Frameworks For Measuring Travel And Tourism
Activities
 BEA

Statistical unit Visitor
Concept of visitor Person traveling outside of
 usual environment
 for less than 12 months.
Concept of usual environment Place of usual
 activities--residence, work,
 leisure,
 Minimum distance determined
 by available data
 sources--between 50 and 100
 miles from residence.

Criteria distinguishing
 tourism from non- Direct contact between visitor
 tourism expenditures, and supplier
 of tourism commodities.
Tourism demand Expenditures by visitors
Tourism commodities/tourism
 industries Determined by what U.S.
 visitors do

Infrastructure
 investments--private and Future extension of TTSA'S
 public.

 OECD

Statistical unit Visitor
Concept of visitor Same as BEA

Concept of usual environment Place of usual
 activities--residence, work,
 leisure.
 Tourism determined by
 minimum distance
 from usual environment.
 Minimum distance defined
 by country

Criteria distinguishing
 tourism from non- Same as BEA
 tourism expenditures,
Tourism demand Same as BEA
Tourism commodities/tourism
 industries Determined by share of
 commodity purchased by
 visitors or produced
 primarily as an attraction
 for visitors.
Infrastructure
 investments--private and Private purchases of fixed
 public. assets, for example, capital
 investment in hotel structures.
 Public purchases include
 airports, long-distance
 bus stations.
 List still under discussion

 WTO

Statistical unit Visitor
Concept of visitor Same as BEA

Concept of usual environment Same as OECD

Criteria distinguishing
 tourism from non- Same as BEA
 tourism expenditures,
Tourism demand Same as BEA
Tourism commodities/tourism
 industries Same as OECD

Infrastructure
 investments--private and Private purchases of fixed
 public. assets are same as OECD

BEA Bureau of Economic Analysis OECD Organisation for Economic Co-operation and Development WTO World Tourism Organization

Visitor. A visitor is a person who either travels outside of his or her "usual environment" for a period of less than 1 year or who stays overnight in a hotel or motel. The visitor may travel for personal pleasure or on industry or government business. Visitors do not include travelers whose main purpose is to be compensated within the place visited--such as migrant workers, persons traveling to new assignments, and diplomatic and military personnel traveling between their duty stations and home countries. The 1-year time period is consistent with the internationally accepted limit that is usually used to define a "resident."

Usual environment. The usual environment is defined as the place of normal (or everyday) activities--such as residence, leisure, study, and work--and the criterion is distance. For the U.S. TTSA'S, the usual environment is defined as the area within 50-100 miles of home, depending on available data sources.(9)

Tourism demand, commodities, and industries

In the TTSA'S, tourism activities are measured by tourism demand, which is defined as the travel-related expenditures made by all visitors, before, during, and immediately after each trip taken. Tourism demand consists of business travel and travel by government employees inside and outside the United States, U.S. resident household travel inside and outside the United States, and travel in the United States by nonresidents (international visitors).(10)

Tourism commodities are the commodities that are typically purchased by visitors directly from producers. The identification of tourism commodities partly depends on the locale and the activities of visitors, but several commodities, such as hotels and transportation services, are obvious. Classification of tourism commodities in the TTSA'S is based on a list of predominant activities of visitors that was developed from WTO and OECD recommendations and from five different sets of surveys of U.S. visitors.(11) The commodities so classified are grouped into the following broad categories: Tourism commodities, such as hotels and lodging, eating and drinking places, other types of leisure activities, and modes of transportation; and non-tourism commodities, such as gasoline and oil (see table 3).

Table 3.--Classification of Commodities in the Travel and Tourism
Accounts

Description of commodity Content

Tourism commodities:

Hotels and lodging places Includes lodging receipts from
 hotels, motels, guestrooms, and
 rooming and boarding houses
 serving the general
 public; other receipts of
 hotels and motels, sporting and
 recreational camps, and
 recreational vehicle parks and
 camp sites
 Excludes meals served by hotels
 or motels

Eating and drinking places Includes food and beverage
 receipts and tips
 Excludes catering services and
 school lunch sales by State and
 local governments

Passenger rail Includes receipts from rail
 passengers for travel and
 dining and tips

Passenger bus and other
 local transportation Includes receipts from passengers
 for intercity, charter, and
 local bus services and subway
 and limousine services

Taxicabs Includes taxi fares and tips

Domestic passenger air fares Includes receipts from domestic
 air passengers for airfares,
 meals and beverages, movies,
 and other receipts

International air fares Includes receipts from
 international air passengers

Passenger water Includes receipts from passengers
 for water transportation

Auto and truck rental Includes receipts for rental of
 automobiles and trucks

Other vehicle rental Includes receipts for rental of
 recreational vehicles and utility
 trailers

Arrangement of passenger
 transportation Includes commissions for the
 arrangement of passenger
 transportation and net receipts
 for tours

Recreation and entertainment Includes miscellaneous
 entertainment receipts such
 as amusement parks, fairs,
 museums, gambling, and other
 recreation and amusements

Participant sports Includes participant sports
 such as golf and tennis

Movie, theater, ballet, and
 musical events Includes receipts for admissions
 to movies and theater and
 music programs

Sports events Includes admissions to sports
 events

Petroleum retail margins Includes retail margins on
 petroleum sales

Other retail margins Includes retail margin on all
 other goods

Travel by U.S. residents
 abroad Includes travel expenditures
 by U.S. residents abroad

Nontourism commodities:(1)

Gasoline and oil Includes sales of gasoline,
 diesel fuel, lubricating
 oils, and grease

PCE nondurable commodities Includes sales of all other
 nondurable commodities

Selected services Includes receipts for selected
 services that may be used by
 tourists on, during, or after
 a trip, such as parking,
 tolls, and automotive repair
 services

Wholesale trade margins and
 transportation costs Includes wholesale margins and
 transportation costs on all goods

All other commodities Includes all other commodities
 not considered above

(1.) Nontourism commodities are commodities not classified as tourism commodities. PCE Personal consumption expenditures

Several commodities are included in order to account for shopping and other purchases and to provide estimates in purchasers' prices (that is, to compute values in producers' prices from the production table). These commodities are petroleum products retail margins, other retail margins, gasoline and oil, and wholesale trade margins and transportation costs. Petroleum retail margins and other retail margins are classified as tourism commodities because the retailers have direct contact with visitors. Other commodities, while supplying tourism demand, are classified as nontourism commodities because the producers do not have direct contact with visitors.

The TTSA'S exclude a number of commodities that could be viewed as tourism commodities--consumer durables, imputed rents from vacation homes, skiing, health spas, financial services, and retail food (off-premise food consumption); many of these commodities are not separately identified in the I-O accounts. However, two of these commodities--consumer durables and imputed rents from vacation homes and related lodging accommodations, such as time-shares--are important tourism commodities, and their exclusion results in an understatement of travel and tourism activities. Including them would require additional analysis to develop estimates of tourism's share and, in the case of consumer durables, to determine which items to include.

Tourism industries are identified by analyzing the relationships shown in the I-O accounts between tourism commodities and the producing industries. Industries that include tourism commodities as a primary product are classified as tourism industries. These industries generally sell a significant portion of their output to visitors, where "significant" indicates that the industries' revenues and profits would be substantially affected if tourism ceased to exist. Examples include airline transportation, hotels, and local public transportation.

With modifications, the TTSA'S follow the tables developed by the OECD. Of the 13 tables suggested by the OECD, the TTSA'S include 4 that are considered high priority--the tables for production, demand, value added, and employment. A fifth table, on estimates of gross domestic investment in tourism industries, is also considered high priority but is left for future work. The WTO and OECD definitions, classification of tourism commodities and industries, and tables have been modified to account for differences in U.S. national accounting practices, tourism activities in the United States, and the availability of data on tourism.

Relationship to the I-O accounts

The I-O accounts formed the basis of the preparation of the TTSA'S in three ways. First, the I-O accounts provided detailed measures of output by commodity and industry that were used to identify commodities purchased by visitors. Second, the I-O accounts provided the detailed estimates of industry and final use expenditures required to identify tourism expenditures by type of visitor. Third, the I-O accounts provided the analytical framework that links these expenditures to industry output and to national aggregates, such as GDP.

The TTSA'S are adapted from the I-O accounts by rearranging selected outputs and inputs in the I-O accounts to fit the classification of the TTSA'S. The TTSA'S generally maintain the conventions of the I-O accounts, but they differ in the following ways:

* I-O industries and commodities are regrouped to follow the classification system for the TTSA'S.

* Personal consumption expenditures (PCE) for tourism commodities is disaggregated into resident and nonresident purchases--a distinction that is not made in the I-O accounts.

* Resident household and nonresident tourism expenditures from nonprofit institutions and from government--primarily admissions to national parks, museums, and other services sold to visitors--are included as admissions paid by visitors. In the I-O accounts, the outputs of nonprofit institutions and of government are measured as current expenditures.(12) Estimates of employment for tourism industries are included; the I-O accounts do not include employment by industry. Employee and business travel expenditures are included in tourism demand. The I-O accounts include these expenditures as intermediate inputs and not as final expenditures.

In the TTSA'S, there is generally a one-to-one relationship between industries and commodities (table 4). Exceptions are the "passenger air transportation" and "automotive rental" industries, both of which produce more than one tourism commodity. In addition, the commodity "travel by U.S. residents abroad" has no industry counterpart, because no U.S. production is associated with overseas expenditures by U.S. residents.

Table 4.--TTSA Industries and Commodities

Industry Commodity

Hotels and lodging places Hotels and lodging places

Eating and drinking places Eating and drinking places

Railroads and related services Passenger rail

Local and suburban transit and
 interurban highway passenger
 transportation, except taxicabs Passenger bus and other local
 transportation

Taxicabs Taxicabs

Air transportation Domestic passenger air fares
 International air fares

Water transportation Passenger water

Automotive rental and leasing,
 without drivers Auto and truck rental
 Other vehicle rental

Arrangement of passenger
 transportation Arrangement of passenger
 transportation

Miscellaneous amusement and
 recreation services (except
 membership sports and
 recreation clubs); racing,
 including track operation;
 marinas; libraries, museums,
 art galleries, and botanical
 and zoological gardens. Recreation and entertainment

Membership sports and
 recreation clubs Participant sports (golf,
 tennis, etc.)

Motion picture theaters; dance
 studios, schools, and halls;
 theatrical producers (except
 motion pictures), bands, Movie, theater, ballet, and
orchestras, and entertainers. musical events

Professional sports clubs and
 promoters Sports events

Gasoline service stations Petroleum retail margins

Retail, excluding eating and
 drinking places and gasoline
 service stations Other retail margins

Industries producing nondurable
 PCE goods PCE nondurable commodities

Automobile parking, automotive
 repair shops and services, and
 toll highways Parking, automotive repair,
 and highway tolls

All other industries Wholesale trade margins and
 transportation costs
 Gasoline and oil

(1) Travel by U.S. residents abroad

(1.) Travel by U.S. residents abroad has no industry counterpart. U.S. residents traveling abroad purchase commodities that are produced abroad, and the TTSA'S include only domestically produced commodities.

PCE Personal consumption expenditures

Components of TTSA'S

The TTSA'S expand the four "priority" OECD tables to five tables to show clearly the major components of tourism. The production of tourism commodities is shown in table 5; the supply and consumption of tourism commodities, in table 6; tourism demand, in table 7; tourism gross domestic product (CDP), in table 8; and tourism employment, in table 9.

[TABULAR DATAS 5 to 9 NOT REPRODUCIBLE IN ASCII]

The production account of tourism industries

The production table (table 5) is similar to the I-O make table, but with three modifications. First, the rows and columns are reversed; TTSA commodities are shown across the rows, and TTSA industries down the columns. Second, detail is shown only for TTSA commodities and industries; the production for all other commodities and industries is aggregated. Third, the intermediate inputs and the value-added components--compensation of employees, indirect business taxes, and other value added--are shown as rows at the bottom of the table.

Similar to the I-O make table, the TTSA production table shows the value of each commodity produced by each industry in producers' prices. Each cell on the main diagonal (that is, from the top left to bottom right) shows the value for the commodity produced by the industry that has been designated its primary producer. For example, the "hotels and lodging places" industry (column) produces mostly the commodity "hotels and lodging places" (row). "Eating and drinking places" are the main producers of meals and beverages (the "eating and drinking places" commodity). The other cells in each column show the value of production by that industry of commodities for which the industry is a secondary producer. For example, the industry "participant sports" is a secondary producer of "hotels and lodging." The sum of the row entries is the total output for the commodity. Because tourism industries are the primary producers of tourism commodities, and only a few other industries are secondary producers, most of the cells outside of the main diagonal are zero. For example, the "railroad" industry produces only transportation services sold to visitors.

Supply and consumption of tourism commodities

The table on supply and consumption of tourism commodities (table 6) rearranges and modifies the I-O use table to show the supply of tourism commodities and all other commodities, as well as the intermediate and final demand for these same commodities. Supply is defined as the total amount of the commodity available to be purchased by business, households, and government and to be exported. It is the sum of domestic production, imports, government sales, inventory changes, wholesale and retail margins, and transportation costs. On the consumption side, it shows the intermediate (business) and final purchases (personal consumption expenditures, investment, exports, and government expenditures excluding sales) of these commodities in purchasers' prices. The categories of consumption in the TTSA'S are the same as in the I-O use table.

Tourism demand by type of commodity and type of visitor

The TTSA demand table (table 7), a subset of the I-O use table, rearranges information from the supply and consumption table (table 6) and separates tourism demand from nontourism demand. It shows the major components of tourism demand in purchasers' prices, including the following: Consumption by business as intermediate use and consumption by government, resident households, and nonresidents as final uses. Household consumption is equivalent to PCE that has been adjusted to exclude the expenses of nonprofit institutions that are not covered by admissions fees and to exclude travel by nonresidents in the United States. Nonresident tourism includes PCE nonresident travel expenditures plus exports of international air and water transportation fares.

Table 7 also shows the proportions of tourism commodities compared with nontourism commodities. These proportions provide the basis for estimating tourism value added and tourism employment. They are also used to derive the "tourism-commodity ratio" for each tourism commodity--that is, the proportion of the supply of the commodity that is purchased by visitors. For example, 100 percent of supply of "hotels and other lodging" was used by visitors, but only 18 percent of the supply of "eating and drinking places" was purchased by visitors.

Tourism GDP of tourism industries and other industries

The tourism GDP table (table 8), which is derived from the supply and consumption table (table 6) and the demand table (table 7), shows the relative importance of tourism industries and other industries in producing tourism output and tourism value added. For each tourism industry and for "all other industries" it shows intermediate consumption, value added, and the "tourism-industry ratio," that is, the share of an industry's output that is purchased by visitors.

The tourism-industry ratio for each industry is derived by applying the tourism-commodity ratio to each of the tourism commodities produced by that industry. For example, the hotel industry's ratio is derived by summing the output of hotel services times 1.0 (the tourism-commodity ratio for hotel services), the output of eating and drinking places times 0.18, the output of recreation and entertainment (gambling) times 0.90, and the output of other retail times 0.03 and then dividing that sum by hotel industry output. The tourism-industry ratio for each industry is applied to the industry output, intermediate consumption, and value added. The ratio applied to the industry value added is the estimated tourism-industry contribution to GDP.

Tourism employment and compensation of employees

The tourism employment and compensation of employees table (table 9) has no counterpart in the I-O accounts, but it includes I-O compensation values. Total employment is equal to the number of full- and part-time employees; compensation of employees consists of wages and salaries and supplements to wages and salaries, such as employer contributions to pension plans, social security, and fringe benefits. The table shows total employment, tourism employment, tourism compensation, and average compensation in each tourism industry. Tourism employment is determined by the total employment in each industry and the tourism-industry ratio (from table 8). Similarly, tourism compensation is derived by multiplying compensation of employees in each industry by the tourism-industry ratio for that industry. Average compensation in each tourism industry is tourism compensation divided by tourism employment.

Methodological Overview

The TTSA estimates were based on the 1992 benchmark I-O accounts, but not all were taken directly from them. The estimates that were taken directly from the I-O accounts include the output of tourism and nontourism commodities, the output of the industries that produce these commodities, the value added of these industries, compensation of employees in these industries, and the consumption of the tourism and nontourism commodities (tables 5 and 6). The tables show the supply of output of tourism commodities that are produced and are available for purchase, and they identify the economic units that consume TTSA commodities: Business, persons, and government.

The estimates that were not taken directly from the I-O accounts include tourism demand and tourism aDP (tables 7 and 8). For a number of tourism commodities, tourism expenditures were separated from nontourism expenditures on the basis of assumptions about allocations between the two types of expenditures. There are several sources on which to base these allocations, and each provides significantly different allocations. Because BEA was not able to reconcile these differences, three alternative sets of TTSA estimates were prepared.

Tourism expenditures in the I-O accounts

In the I-O accounts, the expenditures of visitors are included as purchases by business and final users. Expenditures for business travel are included in intermediate purchases, and travel by government employees are included in government final expenditures. Expenditures by nonresidents for international air and water fares are included in exports.

All other travel expenditures by visitors (resident and nonresident) are included in personal consumption expenditures (PCE). PCE for selected commodities consists of all purchases by resident and nonresident visitors in the United States, and it excludes purchases abroad by U.S. residents. In contrast, total PCE consists of all expenditures by U.S. residents, including expenditures made during foreign travel, and it excludes expenditures made in the United States by nonresidents. The differences between PCE for specific commodities and total PCE are in two PCE categories: "Foreign travel by U.S. residents" and "Expenditures in the United States by nonresidents."(13)

Alternative methods of estimating tourism demand

The tourism expenditures in the TTSA'S were derived from the I-O estimates of consumption. The procedures used to develop the estimates of the visitors' share of consumption depended on the source of demand: Consumer demand, business and government demand, and international demand.

Consumer demand.--The tourism commodities purchased by consumers were separated into two types. "Pure-tourism" commodities are commodities for which all or most of the expenditures are by visitors, such as hotels and lodging places. "Mixed-use" commodities are commodities for which the expenditures are by both visitors and nonvisitors, such as restaurant meals.

For pure-tourism commodities, commodity-flow estimates taken directly from the relevant PCE categories in the 1992 I-O accounts were used for each of these commodities.(14) To provide separate figures for expenditures of resident households and nonresidents, estimates of the expenditures by nonresidents were subtracted from the total (see the section "Other procedures" for a description of these estimates).

For mixed-use commodities, the Bureau of Labor Statistics (BLS) Consumer Expenditure Survey (CEX) is the only national source of data that is available to allocate the shares of consumer spending between visitors and nonvisitors.(15) The CEX is a quarterly survey of 5,000 U.S. households that collects data on consumer expenditures, including expenditures on tourism, or out-of-town trips, for selected categories of commodities.(16)

Surveying households to collect expenditure data is difficult, and several evaluations of the CEX have highlighted some of these problems.(17) The problems include small sample size; the length, complexity, and degree of detail of the questionnaire; high nonresponse rates; and high rates of recall error. As a result, certain types of expenditures tend to be underreported or misreported. Large purchases, such as automobiles, and regular and recurring expenditures, such as rent and utilities, appear to be well reported. However, smaller and/or infrequent expenditures (such as food, apparel, public transportation, alcohol, tobacco, and gambling) appear to be underreported.

Unfortunately, there are few studies that have evaluated the relative accuracy of the individual components of the PCE and the CEX or the completeness of reporting in the CEX.(18) The PCE estimates are largely based on Census Bureau data on sales of goods and services and BEA input-output estimates of the allocation of purchases of goods and services between consumers, business, and government. The PCE estimates may also be subject to errors; for example, some of the PCE estimates may be overstated because they include purchases by businesses not correctly identified in the I-O estimates. For a comparable list of commodities, the CEX estimates tend to be lower than PCE estimates for less frequently consumed or small-value items, but similar for big-ticket and frequently purchased items. Overall, the CEX estimates are about 30 percent lower than the PCE estimates for 1992.

Because the CEX is a general-purpose household survey and was not designed as a travel survey, it is difficult to assess how well tourism expenditures are reported. Some of the limitations of the CEX for use as a travel survey are the following: The questions on out-of-town trips are at the end of the survey; a single respondent, rather than the individual traveler, is asked about all expenditures for the household; the survey is a large questionnaire that covers all household expenditures rather than a smaller one that focuses on specific types of expenditures such as travel; and the sample size for households reporting travel is likely to be far smaller than the samples for other types of consumer spending. Surveys specific to travel have tended to yield higher estimates of travel expenditures than has the CEX.

Expanding the sample and making other improvements to the CEX may help address some of these issues. In the meantime, for purposes of this study, the following three methods were used to provide a range of estimates for mixed-use TTSA commodities.

* Under "Method 1" the CEX estimates for tourism expenditures were assumed to be fully and accurately reported. Thus, tourism expenditures were estimated under this method as the CEX estimates minus estimated overseas expenditures.

* Under "Method 2," the CEX estimates for tourism expenditures were assumed to be as accurate as the estimates of nontourism expenditures for the same commodities; thus, PCE was used as the control total for total spending for each commodity. Tourism expenditures were estimated under this method as the ratio of CEX expenditures on tourism activities (less estimated overseas expenditures) to total CEX expenditures (less overseas expenditures), multiplied by PCE less nonresident expenditures. Under this method, the tourism estimates for mixed-use commodities were about 16 percent above those under method 1.

* Under "Method 3," the CEX estimates for tourism activities were assumed to be more understated than the estimates of other consumer expenditures; thus, the travel estimates were first adjusted using trade-source data and information on the reporting ratio of CEX to PCE for comparable expenditure categories. The adjustment factor for the CEX was 1.5.(19) Tourism expenditures were estimated under this method as the ratio of adjusted CEX expenditures on out-of-town trips (less estimated overseas expenditures) divided by the total CEX expenditures (less overseas expenditures), multiplied by PCE less nonresident expenditures.

For eating and drinking places, the method 3 estimates were further adjusted using estimates from two trade sources--the Travel Industry Association (TIA) and D.K. Shifflet and Associates.(20) This adjustment consisted of using a weighted average of the TIA estimate, the Shifflet estimate, and the method 3 estimate for resident households.

The estimate of PCE on nondurable commodities, other than gasoline and oil, by resident households was based on an average of the ratios from the D.K. Shifflet and Associates survey and the In-Flight Survey for visitor shopping to the sum of expenditures for hotels, meals, and recreation.(21) This average ratio was applied to the three sets of estimates of the sum of resident household purchases that were based on the three alternative methods of the following TTSA commodities: Hotels and lodging places; eating and drinking places; arrangement of passenger transportation (tours); recreation and entertainment; participant sports; movie, theater, ballet, and musical events; and sports events (table 7).

For all three methods, the estimate of resident household purchases of auto repair was based on the ratio of CEX resident household purchases of gasoline on out-of-town trips to PCE gasoline purchases, multiplied by PCE purchases of auto repair.

Business and government demand.--The estimates of the share of business and government expenditures for tourism were based on a variety of sources and different methods, depending on the commodity.

For pure-tourism commodities, the estimates were taken directly from the I-O tables, following the method used for consumer demand.

For mixed-use commodities, the business and government tourism expenditures for various commodities were estimated differently from those for consumers. Expenditures on "eating and drinking places" were estimated by applying the ratio of meals and beverage expenditures to hotel expenditures from the American Express Survey of Business Travel Management to I-O purchases of this commodity.(22) Expenditures for "local transportation," "taxicabs," and "gasoline and oil" were estimated using the ratio of resident household tourism purchases to total PCE by category. Because the resident household tourism purchases were calculated using the three alternative methods (see the discussion on page 19), the business and government expenditures also show three different estimates. Business expenditures on "participant sports, movie, theater, ballet, and musical events" and on "sports events" were estimated from underlying data in the I-O accounts on expenditures for business travel and entertainment by industries.

International demand.--The estimates of total travel expenditures by U.S. residents abroad and by foreign residents in the United States were largely based on BEA balance of payments data. Data from the In-Flight Survey were used to separate total travel expenditures in the United States into five categories: Hotels and lodging places; eating and drinking places; transportation within the United States; recreation and entertainment; and shopping. Expenditures for transportation, recreation, and entertainment in the United States were disaggregated into TTSA commodities using weights calculated from resident household tourism expenditures. A single value for each category was estimated under all three methods.

Estimating tourism employment and compensation of employees

The TTSA estimates of tourism employment were developed from BLS estimates of average monthly employment by industry at the four-digit sic level and from BEA estimates at the two-digit SIC level. Employment was estimated at the four-digit SIC level by applying employment weights from the BLS estimates to the BEA estimates.(23) Employment by SIC industry was assigned to the TTSA industries. Tourism employment and compensation of tourism employees were each estimated by multiplying industry employment and industry compensation, respectively, by the tourism-industry ratio for each of the three methods. Estimates of compensation of employees were from the I-O accounts.

Estimates of Travel and Tourism for 1992

The 1992 estimates of tourism expenditures help to gauge the size and importance of travel and tourism in the United States. Because three different methods were used in estimating tourism expenditures (see the section "Alternative methods for estimating tourism demand"), these estimates are presented as ranges.

In terms of output, travel and tourism represented 2.1-2.4 percent of total U.S. production or output in 1992 ($230.8--$259.5 billion as a share of total industry output) and 1.9-2.2 percent of U.S. GDP (tourism value added of $120.5-$135.7 billion as a share of GDP). These shares are similar to those derived in Canada's tourism satellite accounts, which showed a 2.3-percent share of Canadian GDP; Norway's tourism satellite accounts showed a 4.3-percent share.(24)

The "hotels and lodging" industry had the highest value added among the tourism industries, at $41.6-$42.7 billion, or 31-35 percent of tourism GDP (table 8). The second highest industry was "passenger air" (domestic and international), at $30.5 billion, or 22-25 percent of tourism a GDP. Third was "eating and drinking places;" at $17.7-$22.7 billion, or 15-17 percent of tourism GDP. Shopping ("retail excluding eating and drinking places, and gasoline service stations") was fourth, at $9.0-$11.3 billion, or 7.4-8.3 percent of tourism GDP. In terms of employment (table 9), "hotels and lodging" was the largest tourism industry employer, with 1.3-1.4 million employees, and "eating and drinking places" was the second largest, with 1.1-1.4 million employees. "Air transportation" employment was only 0.5 million, less than half that of hotels.

In terms of demand, the TTSA'S show that tourism purchases in the United States were $284.2--$332.8 billion, or 4.6-5.3 percent of GDP. The TTSA'S also show expenditures by type of visitor. In 1992, resident households' expenditures in the United States accounted for 30-47 percent of domestic tourism expenditures (total tourism demand less travel expenditures by U.S. residents abroad); business sector expenditures, for 27-30 percent; and government expenditures, for 5-6 percent. Expenditures by nonresidents were $71.5 billion, or 21-25 percent of tourism expenditures in the United States. Expenditures by U.S. residents overseas accounted for 11-12 percent of total tourism purchases (table 7).

By category, the largest expenditures were in passenger air travel, followed by hotels and lodging, meals and beverages, and shopping (table 7). Expenditures for passenger air travel services (domestic air, at $48.5 billion, plus international air, at $32.2 billion) were 22-25 percent of total tourism expenditures. Expenditures for hotels and lodging places were $56.6 billion, or 15-17 percent, and those for meals and beverages were $45.4-$58.5 billion, or 14-15 percent. Shopping expenditures (personal consumption expenditures on nondurable, nontourism commodities) were $35.4-$47.4 billion, or 11-13 percent.

Tourism employment in 1992 was 3.8-4.4 million in 1992, or 3.2-3.7 percent of total employment in the United States; this is similar to the 3.9-percent share of employment estimated in Canada's tourism satellite accounts. Tourism's share of employment is much higher than its share of value added to GDP (1.9-2.2 percent), indicating that tourism industries are more labor intensive than the economy as a whole.

Compensation of tourism employees was 2.2-2.5 percent of total compensation of employees. The average compensation per tourism employee was $21,393 per year, but it ranged from a high of $80,783 in professional sports clubs and promoters to a low of $11,917 in eating and drinking places. The average compensation per tourism employee is lower than the average compensation per employee for the economy as a whole ($30,891).

Future Work and Extensions

This prototype satellite account represents a first step in producing satellite accounts for travel and tourism for the United States. Depending on additional funding, the next steps may include the following: Developing point estimates to replace ranges of travel and tourism expenditures; updating the TTSA'S annually; improving the quality of the estimates by collecting additional data; adding estimates of investment in tourism industries; expanding tourism commodities to include consumer durables, imputed rents for vacation homes, and the provision of public facilities used by visitors; and adding estimates by U.S. region (or by State).

Point estimates, updates, and additional data requirements

To develop point estimates, the accuracy and reliability of source data by tourism commodity must be examined to determine the adjustments required for a number of commodities, especially for "eating and drinking places," "shopping," "recreational" activities, and business expenditures. Over time, the proposed expansion of the CEX should provide improved data on tourism expenditures.

Updates of the TTSA'S could be made annually using NIPA estimates of final uses and gross product by industry estimates. This updating would require assuming that the input-output relationships in producing tourism commodities remain the same, that the proportions of tourism demand for many tourism commodities do not change, and that the mix of output by the tourism industries does not change.

Improved estimates of tourism demand are needed in a number of categories of tourism commodities. Specifically, better estimation approaches or the collection of additional data would improve measures of "eating and drinking places" "shopping on out-of-town trips, "recreation and entertainment," "participant sports," "movie theater, ballet, and musical events," "sports events," and business expenditures made on trips.

Investment in tourism industries

One important estimate that could be added is private and public investment in fixed capital by tourism industries. The OECD- and WTO-recommended criterion for the inclusion of this investment in the satellite accounts is that the main use of the output of the industry should be by tourism or by visitors. For private investment, investment in fixed capital by hotels, airlines, and restaurants should be included. For public investment, investment in airports and long-distance bus stations should be included. Both public and private investment in railroads, highways, and seaports should also be included if a method could be developed to separate use for freight transportation services from use for passenger travel.

Consumer durables

Consumer durables--such as recreation vehicles, automobiles, and sports equipment (for example, skis)--are currently not included in the TTSA'S. Whether and how consumer durables might be incorporated requires additional analysis. First, which consumer durables to include needs to be determined. Should only small, low-value items such as cameras be included? Should large, single-purpose consumer durables such as recreational vehicles be included, or should multipurpose consumer durables such as personal automobiles be included? Second, how these durables would be included needs to be determined. Should they be treated as a final consumption item, as they are now in the I-O accounts and the NIPA'S, or should they be treated as investment, as is owner-occupied housing and government purchases of plant and equipment?(25)

Rental of vacation homes

Rents from vacation homes used for tourism purposes conceptually should be included in the TTSA'S. The inclusion of imputed rents for owner-occupied homes is a standard treatment in the NIPA'S, but applying this approach to vacation homes is difficult. One difficulty is separating vacation homes from primary residence homes. A second difficulty is the lack of data on the length of time the vacation home is used by the owner and the time it is rented to others. If the home is rented to others, information on rental costs is needed; these costs are not readily available, and indirect estimates are difficult to make.

Public facilities used by visitors

Public facilities, such as parks and museums, are attractions to visitors as well as nonvisitors. However, the costs of providing these public facilities are not included in the TTSA'S. Estimating these expenditures would be a lengthy and time-consuming project.

Regional analysis

One extension of the TTSA'S would be to estimate travel and tourism expenditures by region (or by State) in the United States. It would focus on tourism CDP and employment and employee compensation by State.

(1). The first set of extensions covered transportation activities; see Bingsong Fang, Xiaoli Han, Ann M. Lawson, and Sherlene K.S. Lum, "U.S. Transportation Satellite Accounts for 1992," Survey of Current Business 78 (April 1998): 16-27. These accounts were developed jointly with the Bureau of Transportation Statistics, U.S. Department of Transportation.

The TTSA'S were developed by the Bureau of Economic Analysis with the support of the Tourism Industries Office of the International Trade Administration, U.S. Department of Commerce.

(2.) For descriptions of the other satellite accounts that have been developed by BEA, see "Integrated Economic and Environmental Satellite Accounts" and "Accounting for Mineral Resources: Issues and BEA'S Initial Estimates," Survey 74 (April 1994): 33-72; and "A Satellite Account for Research and Development," Survey 74 (November 1994): 37-71.

(3.) For a description of the I-O accounts, see Ann M. Lawson, "Benchmark Input-Output Accounts for the U.S. Economy, 1992: Make, Use, and Supplementary Tables," Survey 77 (November 1997): 36-82; and Ann M. Lawson, "Benchmark Input-Output Accounts for the U.S. Economy, 1992: Requirements Tables," Survey 77 (December 1997): 22-47.

(4.) The term "visitor" is used because it is more descriptive of the travel activities included in the TTS'S than the term "tourist" which connotes a person who travels for leisure only.

(5.) U.S. Travel Center, Travel Industry Association, National Travel Survey (Washington, DC: Travel Industries Association of America, 1992); U.S. Department of Commerce, International Trade Administration, Tourism Industries Office, Summary and Analysis of International Visitors to the United States (Washington, DC: U.S. Government Printing Office, 1992); U.S. Department of Transportation, Bureau of Transportation Statistics, American Travel Survey (Washington, DC: U.S. Government Printing Office, 1995).

(6.) See World Tourism Organization, "Tourism Satellite Account (TSA)," 3rd Draft, Rev. 1 (Madrid, January 1998) for the most recent description of WTO definitions and guidelines, and Tourism Committee, Organisation for Economic Co-operation and Development, Manual on Tourism Economic Accounts, OECD/GD(91)82 (Paris, 1991).

(7.) See Jocelyn Lapierre and Duane Hayes, "The Tourism Satellite Account," National Income and Expenditure Accounts, Quarterly Estimates, Second Quarter 1994 (Ottawa: Statistics Canada); and Trude Nygaard, "Satellite Accounts for Tourism in Norway" (Division of National Accounts, Statistics Norway, August 1996). Other OECD countries that are currently developing their own accounts include Australia, Spain, France, New Zealand, Switzerland, and Poland.

More recently, the European Union issued a directive to its member states to collect harmonized European Community statistical information on tourism supply and demand; EUROSTAT began collecting data in 1997.

(8.) The OECD recommends a set of 13 tables that show the following: The production account of tourism industries; the demand for tourism commodities; the value added of tourism industries; employment in tourism industries; visitor characteristics; investment by tourism industries; capital stock; and 6 other tables measuring these variables in real terms. The WTO recommends a set of 12 tables that consist of the following: Domestic tourism consumption by commodity and origin (domestic and foreign); outbound tourism consumption; inbound tourism consumption; internal and national tourism consumption by forms--summary; tourism consumption by payer (by sector, or households, business, government); tourism consumption by function (domestic tourism, outbound tourism, and inbound tourism); importance of tourism consumption within supply; supply of tourism commodities according to activities; production accounts of tourism industries; tourism-related net acquisition of nonfinancial assets; employment in tourism industries; and imports and exports of goods and services generated by tourism.

(9.) The distance criterion differs by survey: The Consumer Expenditures Survey (Bureau of Labor Statistics) uses 75 miles from home; the American Travel Survey (Bureau of Transportation Statistics), 100 miles from home; private surveys by the Travel Industry Association, 50 or 100 miles from home; and surveys by D.K. Shifflet and Associates, 50 miles from home.

(10.) Resident household travel refers to tourism of residents within the country, and travel by nonresidents refers to tourism of nonresidents within the country (inbound international tourism).

(11.) The WTO and OECD recommendations were modified to agree with the U.S. national income and product accounts and expanded where additional data on tourism demand were available. The five sets of surveys reviewed were the Consumer Expenditures Survey, prepared by the Bureau of Labor Statistics; the In-Flight Survey, prepared by the International Trade Administration, U.S. Department of Commerce; the American Travel Survey, prepared by the Bureau of Transportation Statistics, U.S. Department of Transportation; the National Travel Survey of the Travel Industry Association; and surveys by D.K. Shifflet and Associates.

(12.) In the I-O accounts, government expenditures are the net of expenditures less government sales.

(13) The expenditures in these two PCE categories are also included in net exports. Expenditures in the United States by nonresidents consist of both travel and other (not travel-related) expenditures. For the purposes of the TTSA'S, only travel expenditures are included as part of tourism demand. Excluded are medical expenditures by nonresidents, nonresident student expenditures, expenditures by nonresidents in the U.S. working for foreign governments and international organizations, expenditures by Mexican, West Indian, and Puerto Rican workers in the United States, and expenditures by foreign ocean and air crews in the United States.

(14.) Household tourism expenditures from nonprofit institutions include only the portion of expenses of nonprofit institutions covered by admissions.

(15.) The acronym "CEX" is used for this survey because the acronym "CES" is usually used for the BLS Current Employment Survey.

(16.) The CEX includes overseas trips and trips within the United States that are greater than 75 miles and that are not reimbursed by an employer. Data collected on expenditures made on out-of-town trips include lodging, food, alcoholic beverages, intercity train fares, intercity bus fares, local transportation, taxi fares, airline fares, ship fares, auto rental, truck rental, rental of campers and other vehicles, boat and trailer rental, recreation expenses and other entertainment, participant sports, movie and other admissions, admissions to sports events, gasoline and oil purchases, and parking and tolls. For additional information on this survey, see U.S. Department of Labor, Bureau of Labor Statistics, BLS Handbook of Methods, "Consumer Expenditures and Income" (Washington DC: U.S. Government Printing Office, September 1992): 170-175, and U.S. Department of Labor, Bureau of Labor Statistics, Consumer Expenditure Survey, 2992-93, Bulletin 2462 (Washington, DC: U.S. Government Printing Office, September 1995): 1-13, 224-234:

(17.) See E. Raphael Branch, "The Consumer Expenditure Survey: A Comparative Analysis,"Monthly Labor Review (December 1994): 47-55; Jack E. Triplett, "Measuring Consumption: The Post-1973 Slowdown and the Research Issues" Federal Reserve Bank of St. Louis Review (May/June 1997): 15-22; and Constance F. Citro and Robert T. Michael, ed., Measuring Poverty, A New Approach (Washington, DC: National Academy Press, 1995): 85-88, 392-95.

(18.) Triplett, 16.

(19.) This factor was calculated by (1) computing the average ratio of the CEX estimate to the PCE estimate for all corresponding expenditure categories; (2) identifying the expenditure categories with a CEX-to-PCE ratio that was less than the average ratio; (3) computing the average CEX-to-PCE ratio for the expenditure categories; and (4) dividing the average CEX-to-PCE ratio from (1) by the average CEX-to-PCE ratio from (3).

(20.) The TIA estimate was the product of the number of travelers staying in hotels by state (monthly survey of 20,000) times an average cost for three meals in each state, summed for all states. The D.K. Shifflet estimates were based on a monthly survey sample of 25,000.

(21.) This ratio equaled 0.35 for the D.K. Shifflet and Associates survey and 0.57 for the In-Flight Survey; see footnote 11.

(22.) The 1992 American Express Survey of Business Travel Management, unpublished.

(23.) BEA adjusts the four-digit BLS data to extend coverage to partially covered industries, such as membership organizations, and to excluded industries, such as railroads. Other adjustments include the addition of excluded nonprofits as well as misreporting adjustments. (Source: Bureau of Economic Analysis, U.S. Department of Commerce, State Personal Income, 1929-93 (Washington, DC: U.S. Government Printing Office, 1995): M-9--M-13.)

(24.) See footnote 7.

(25.) If these durables were treated as investment, the services of these assets would be included in GDP. For estimates of the stock of consumer durables that would be used to estimate these services, see U.S. Department of Commerce, Bureau of Economic Analysis, Fixed Reproducible Tangible Wealth in the United States, 1925-96, CD-ROM (Washington, DC: Bureau of Economic Analysis, 1998). For the treatment of owner-occupied housing in the I-O accounts and the NIPA's, see U.S. Department of Commerce, Bureau of Economic Analysis, Personal Consumption Expenditures, Methodology Paper Series MP-6 (Washington, DC: U.S. Government Printing Office, 1990): 8.

Acknowledgments

The U.S. Travel and Tourism Satellite Accounts for 1992 were prepared by the staff of the Bureau of Economic Analysis under the direction of Sumiye Okubo, Associate Director for Industry Accounts. The project was initiated by J. Steven Landefeld, Director of nut,, Robert Parker, Chief Statistician of BEA, and Helen N. Marano, Director of Tourism Development, Tourism Industries Office, International Trade Administration, U.S. Department of Commerce.

At BEA, Mark A. Planting, Branch Chief, Industry Economics Division, developed the framework for the production, demand, and value-added accounts, and the estimates of these accounts. David I. Kass developed the method for estimating tourism expenditures on eating and drinking places, the framework for measuring tourism employment and compensation of employees, and the estimates for the accounts for employment and compensation of employees; he also drafted the sections of the article describing tourism employment and compensation of employees. John Turner assisted in the data processing of the accounts. Michael Mann from the Balance of Payments Division provided assistance on evaluating data from the In-Flight Survey, and Clint McCully from the National Income and Wealth Division, on evaluating the data from the Consumer Expenditures Survey.

RELATED ARTICLE: Satellite Accounts

BEA has developed several satellite accounts for transportation services, environment and mineral resources, and research and development. Satellite accounts are frameworks designed to expand the analytical capacity of the national accounts without overburdening them or interfering with their general-purpose orientation. In this role, satellite accounts organize information in an internally consistent way that suits the particular analytical focus at hand, yet they maintain links to the existing national accounts. Further, because they supplement the existing accounts, rather than replace them, they can serve as a laboratory for economic accounting in that they provide room for conceptual development and methodological refinement.

Satellite accounts can add detail or-other information about a particular aspect of the economy to that in the existing accounts; for instance, they can integrate monetary and physical data. They can arrange information differently, perhaps by cutting across sectors to assemble information on both intermediate and final consumption. For example, a satellite account can assemble business expenditures on training--treated as intermediate consumption in the existing accounts--and education-related expenditures by households and government to analyze the role of education in the economy. They can use a classification other than that used in the existing accounts. For example, they can identify expenditures on "research in education" as part of research expenditures even though they are included in education expenditures in the national accounts.

The terminology and concepts associated with satellite accounts reflect the experiences of several countries that have constructed them, largely on an ad hoc basis, for fields such as health, education, agriculture, research and development, and transportation. The System of National Accounts 1993, which presents the newly revised international accounting guidelines, includes a chapter that provides a general framework for satellite accounts and demonstrates how that framework can be used for some of the fields in which such accounts would be most useful. This chapter represents, in a real sense, the coming of age of satellite accounts as an analytical tool.