See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/4901525

E-Commerce: Measurement and Measurement Issues

Article  in  American Economic Review · February 2001

DOI: 10.1257/aer.91.2.318 · Source: RePEc

CITATIONS READS

86 1,834

1 author:

Barbara M. Fraumeni
Central University of Finance and Economics
70 PUBLICATIONS   2,894 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Human capital, for various international agencies, althoug I expect an IZA World piece on GDP to be out soon. View project

Agricultural productivity accounts View project

All content following this page was uploaded by Barbara M. Fraumeni on 27 May 2014.

The user has requested enhancement of the downloaded file.

 E-Commerce: Measurement and Measurement Issues

By BARBARA M. FRAUMENI*

E-commerce is a hot topic; however, little is downs of e-commerce activities are typically
known about the actual size and impact of e- not available. Some forecasters are very opti-
commerce in the United States. In part this is mistic about the future of e-commerce, others
because e-commerce is a recent and rapidly less so. For example, Jupiter Communications
evolving phenomenon, but it is also because the has an optimistic outlook, recently estimating
measurement of e-commerce presents a num- that 42 percent of all B2B trade would be con-
ber of challenges. Some of these challenges ducted over the Internet by 2005 (PSINet, 2000
are essentially unique. Others are similar to, p. 65).
or extensions of, old economy measurement The fact that there are fewer B2C forecasts
challenges. than B2B forecasts probably reflects the exis-
tence of Bureau of the Census (Census) survey
I. E-Commerce Estimates programs that are emphasizing the estimation of
the size of B2C as opposed to B2B (Thomas L.
Many estimates and forecasts of the size of Mesenbourg, Jr., 2000). In the third quarter of
e-commerce exist. As Table 1 demonstrates, 2000, e-commerce retail sales increased 15.3
there is a considerable range between the high- percent from the second quarter but are still less
est and lowest estimates of business-to-business than 1 percent of total retail sales (U.S. Depart-
(B2B) and business-to-consumer (B2C) e-com- ment of Commerce, 2000). There is general
merce. Differences in methodology, coverage, agreement that the size of B2B e-commerce is
and general outlook are several possible reasons significantly larger than the size of B2C e-com-
for the differences in the estimates and fore- merce. As Table 1 indicates, on average, the
casts. It is frequently difficult or impossible to corresponding B2B estimate (e.g., comparing
determine the specific methodology or coverage lowest to lowest for a specific year) is over three
underlying the estimates and forecasts. Without times the B2C estimate.
a clear understanding of the methodology em-
ployed, it is difficult to evaluate the estimates. A II. Is E-Commerce Covered in Aggregate
common concern is the extent of double-count- U.S. Economic Statistics?
ing (e.g., intermediate inputs being counted as
well as the full value of the product produced by E-commerce activities are embedded in the
these inputs). Another concern is the validity of U.S. national accounts, including the GDP ac-
comparisons to gross domestic product (GDP) counts (Fraumeni et al., 1999). Many are trans-
(e.g., the contribution of e-commerce-generated actions formerly done in more traditional ways,
transportation-sector activities to GDP includ- undoubtedly some previously via mail order.
ing the value of the product being shipped, These transactions are picked up by surveys
instead of only the margins that are included in conducted by the Bureau of Labor Statistics
the national income and product accounts (BLS), by the Census, by tax records from the
[NIPA] GDP). Even if coverage differences are Internal Revenue Service (IRS), and by other
known, comparisons are difficult, as break- data sources drawn on by the Bureau of Eco-
nomic Analysis (BEA). However, these activi-
ties could not be separated from other types of
* U.S. Department of Commerce, Bureau of Economic activities (e.g., to estimate e-commerce GDP)
Analysis, Washington, DC 20230 (e-mail: barbara. without special surveys or additional questions
fraumeni@bea.doc.gov). This paper represents the view of on existing surveys.
the author and is not an official position of the Bureau of
Economic Analysis or the U.S. Department of Commerce. I Looking beyond the national accounts to
thank Nicholas Brannick for his assistance in collecting aggregate U.S. economic statistics in general,
e-commerce estimates and forecasts. the consensus is that the vast majority of
318
VOL. 91 NO. 2 E-COMMERCE E-CONOMY 319

TABLE 1—E-COMMERCE ESTIMATES categories that have grown most rapidly over
a
AND FORECASTS ($BILLIONS)
the post-World War II period (Sherlene K. S.
Lum et al., 2000). With the absence of price
Year Lowest estimate Highest estimate
indexes (mainly in consumer spending on ser-
B2B: vices such as insurance, education, and medical
1998 12 92 care), over 10 percent of consumer spending,
1999 23 114 and almost 25 percent of GDP, is measured
2000 52 406 using either physical inputs as extrapolators
2001 91 717
2002 131 1,510 (mainly labor hours) or as input-cost indexes,
2003 634 2,940 which produce zero or low growth in labor
productivity and often negative growth in multi-
B2C: factor productivity (J. Steven Landefeld and
1998 4 27 Fraumeni, 2001). There are other components,
1999 9 95 such as brokerage services, where real output
2000 7 200
is estimated using partial output measures
Notes: The B2B lowest estimate for 2001 was interpolated that probably do not capture improvements in
between the 2000 and 2002 estimate. service quality associated with information-
Sources: Boston Consulting Group (1999), Goldman Sachs technology innovations. Also, additional work
(1999), B. K. Atrostic (2000), Steven J. Kafka et al. (2000), needs to be undertaken to measure the value
Stacy Lawrence (2000), PSINet (2000), and Julie Sturgeon
(2000). of investments and quality-adjusted prices for
a
Billion ⫽ 109. e-commerce infrastructure, such as servers and
telecommunications equipment.
Free services, particularly information provi-
sion, seem to be a hallmark of the web. There
e-commerce activities are covered (Fraumeni et may be less of a connection between informa-
al., 2000). However, as surveys are conducted tion provision and sales because of the web. For
and sample rotations are carried out on various example, a customer can now obtain flight in-
time schedules, there may be some lag and formation and tickets on the web, over the tele-
differences in survey capture rates. These are phone, or in person at a brick-and-mortar
the same sort of issues that arise for data col- establishment. There are simply more options
lection in dynamic sectors of the economy. It is open, which increases the possibility that sales
not believed that this contributes to significant and information are provided by different
e-commerce under-coverage. agents. In some cases, there may be a shift from
fee for products to free products (e.g., the cur-
III. E-Commerce Measurement Challenges rent day’s newspaper). Advertising revenues,
and Issues including charges to have a link appear on a
web page, represent an important source of rev-
The question of whether e-commerce is in- enue for many e-commerce businesses, accord-
cluded in aggregate U.S. statistics is just a ing to anecdotes. For sites that provide for free
starting point for discussion of e-commerce search engines, content, e-mail or Internet-
measurement. Although e-commerce is included service provision, advertising is frequently the
in the statistics, measuring it is frequently only source of revenue. As is true with radio
problematic. and television, those who enjoy the free services
E-commerce exacerbates preexisting old may be different from those who buy the ad-
economy measurement problems. A number of vertised goods and services. The Census retail
e-commerce-intensive sectors are in difficult- e-commerce surveys do not count free goods
to-measure active or emerging e-commerce and services; therefore it is unclear how these
sectors such as banking, insurance, brokerage activities would be included in a study of e-
services, education, and medical care. In gen- commerce’s contribution to economic growth.
eral, the difficult-to-measure sectors are concen- Between businesses and consumers, who pro-
trated in services and in finance, insurance, and vides what may be changing. For example, if a
real estate (F.I.R.E.) industries, the broad sector consumer makes an e-commerce purchase, the
320 AEA PAPERS AND PROCEEDINGS MAY 2001

purchase is sent electronically (e.g., software) one else, an electronic book perhaps cannot be
or delivered to the home (e.g., most goods). lent. Is an electronic ticket, which may or may
Formerly, except in the case of most mail- and not be bought on the web, the same as a paper
telephone-order purchases, the consumer drove ticket? If a customer loses a paper ticket, they
to a brick-and-mortar facility to pick up the have to pay a penalty to have it replaced, but
good. The full e-commerce home-delivery trans- they can normally use it on another airline if
portation cost is included in the output of the their original flight was canceled.
U.S. economy; the transportation-margin com- Finally, e-commerce may blur the distinction
ponent of it is included in GDP. When the between goods and services. Is an on-line news-
consumer picks up the good himself, a much paper a good or a service? A delivered paper
smaller transportation-related cost appears in newspaper is a good, but an on-line newspaper
output and GDP, including items such as con- might be categorized as a service.
sumer expenditures for gasoline, motor-vehicle
depreciation, and so forth. In addition, the con- IV. E-Commerce, the Digital Economy,
sumer is now typing in the order and billing E-Business, and E-Business Processes
information on the e-commerce site. Formerly
salesmen performed these functions and their How are e-commerce and the e-business pro-
wages were reflected in product prices. There cesses changing the way businesses do busi-
are a number of activity shifts between the ness? One of the challenges is to determine the
nonmarket (not in the national accounts) sector potential impact of process changes. A typical
and the market (in the national accounts) sector. set of economic data does not record directly
It is not clear that the overall level of activity how businesses do business; rather, the data just
has changed at all, yet output and GDP may be record transactions indirectly and incompletely
higher. by the inputs used in production. The digital
Time savings (or loss) and the convenience revolution may be having an impact on indus-
(or inconvenience) of e-commerce are not cap- trial composition and structure not only because
tured in the national accounts. Often the asser- some industries may grow at the expense of
tion is made that these aspects should not be in others, but also because some businesses who
the national accounts, yet to a limited extent, take a greater advantage of process changes,
they are already included. When computers be- including supply-chain changes, may have a
come faster and easier to use, at least some of competitive advantage over others.1 We may
this is reflected in the quality-adjusted computer not have a clear picture of what has occurred
price indexes. It is a matter of degree. Unless a for a number of years. The BEA input– output
satellite household production account is con- matrices, which are the source for the industry
structed, most time savings (or losses) from accounts underlying the GDP accounts, are
e-commerce and other aspects of the digital constructed only every five years because
economy will not be measured. It is unlikely they depend upon the quinquennial censuses
that a household production account would in- (Fraumeni et al., 1999).2 Relative to the pace of
corporate measures of convenience. Therefore, the web, five years is an eternity. In addition,
perhaps an important element of e-commerce e-commerce may open up new opportunities for
would be missed, unless convenience is cap- consumers and businesses in the domestic and
tured by quality adjustments. global economy, changing the geographic dis-
A general question which encompasses the tribution, reach, and distribution channels of
provision, time, and convenience aspects of e- businesses. E-commerce and the digital econ-
commerce goods and services as well as other omy may be facilitating the global economy.
features is whether e-commerce goods and ser- Many believe that changes in business pro-
vices are really the same as brick-and-mortar
goods and services? Are there quality differ- 1
ences between, say, a paper book brought on the See Luis Garicano and Steven N. Kaplan (2000) for a
discussion of possible transaction-cost savings from B2B.
web, a paper book bought in a bookstore, and an 2
The BEA input– output accounts also appear with a lag.
electronic book bought either on the web or in a For example, the 1992 input– output accounts were released
bookstore? A paper book can be lent to some- by BEA in late 1997.
VOL. 91 NO. 2 E-COMMERCE E-CONOMY 321

cesses may be the truly revolutionary impact of fore, with home and electronic delivery common.
e-commerce and the digital economy. Manufacturers, such as Dell, also can more readily
There are additional measurement problems sell directly to the customer, bypassing wholesal-
because B2B transactions are intermediate trans- ers and retailers. The pace of innovation seems to
actions that do not show up as GDP. If one be accelerating, facilitating e-commerce and in-
wants to know the impact of B2B and B2C creasing the need for quicker incorporation of
e-commerce one needs to measure the impact new products. Information on production workers
of both on real GDP and productivity. For is reported monthly; the advent of e-commerce
example, are the prices of consumer goods and the digital economy increases the need for
and services falling, and are there greater monthly reporting on all workers, including non-
efficiencies as a result of the e-commerce rev- production workers and the self-employed. New
olution, including B2B, and the associated high- data needs include regular reports on new forms of
technology infrastructure? For goods, many compensation, new sources of statistics on exports
of the efficiencies of the new economy are and imports, information to facilitate construction
likely to be captured in the estimates. However, of price and output indexes for difficult-to-
for services the absence of adequate price data measure sectors and e-commerce infrastructure,
makes it difficult, if not impossible, for mea- and incorporation of virtual stores into point-of-
sures to reflect higher measured output and pro- purchase surveys. Stock options represent a po-
ductivity arising from the new activities and tentially important new form of compensation
technologies. that anecdotally is more commonly received by
e-commerce and digital-economy business em-
V. Intangibles ployees.3 As e-commerce facilitates commerce
across international boundaries, information on
Measuring intangibles is another old-econ- exports and imports may be reduced, as there are
omy problem which may be exacerbated by no customs agents on the web and e-commerce
e-commerce and the digital economy. The web country of origin and destination may be uncer-
and the Internet have facilitated the speed and tain. As prices and quality of e-commerce goods
ease with which knowledge and information are and services may not be the same as those ob-
transmitted. There is no question that knowl- tained in brick-and-mortar outlets, both clicks and
edge and information are important, but difficult bricks should be included in point-of-purchase
to quantify as inputs. Other types of intangible surveys. These are just a few of the additional data
capital that are frequently mentioned include needs suggested by the growth of e-commerce.
intellectual, organizational, institutional, and
reputational assets. Although some progress has VII. Conclusion
been made in measuring R&D and human-
capital investments, these are usually not inte- As a practical matter, it does not seem feasi-
grated into a set of national accounts. Very little ble to separate e-commerce activities from other
is known about how to quantify intangibles, activities. Even without attempting a separation,
yet they seem to be a particularly important e-commerce presents a number of measurement
source of market valuations for new-economy challenges and issues. As is true in any period
firms, including those involved in e-commerce of rapid change, the measurement community
(Margaret M. Blair and Steven M. H. Wallman, needs to rise to the occasion and devote the
2000). additional resources to address these issues. Al-
though at this point e-commerce may represent
VI. Data Needs a relatively small percentage of both B2C and
B2B, its future size and impact on e-business
A by-product of e-commerce is the need for processes and economic growth may be large.
more timely and new data. More timely data
would be useful with respect to distribution
methods, changes in point-of-sale, new products, 3
BLS recently conducted a pilot survey on the incidence
and compensation for all workers. E-commerce is of stock options. For more information, see Bureau of Labor
changing distribution methods, as mentioned be- Statistics (2000).
 322 AEA PAPERS AND PROCEEDINGS MAY 2001

REFERENCES Goldman Sachs. B2B: 2B or Not 2B, Version

1.1. New York: Goldman Sachs, November
Atrostic, B. K. “Defining and Measuring Elec- 1999.
tronic Commerce: International Perspec- Kafka, Steven J.; Temkin, Bruce D.; Sanders,
tives.” Presentation at the Services Advisory Matthew R.; Sharrard, Jeremy and Brown,
Committee, Statistics Canada, Ottawa, Can- Tobias O. “eMarketplaces Boost B2B Trade.”
ada, 12 May 2000. The Forrester Report (Forrester Research,
Blair, Margaret M. and Wallman, Steven M. H. Cambridge, MA), February 2000.
Unseen wealth. Report of the Brookings Task Landefeld, J. Steven and Fraumeni, Barbara M.
Force on Understanding the Intangible Sources “Measuring the New Economy.” Survey of
of Value. Washington, DC: Brookings Insti- Current Business, March 2001, 81(3), pp.
tution, September 2000. [Available online: 23– 40, [Also available online: 具http://www.
具http://www.brookings.edu/es/research/ bea.doc/gov/bea/pub/0301cont.htm典.]
projects/intangibles/intangibles.htm典.] Lawrence, Stacy. “Behind the Numbers: The
Boston Consulting Group. “New BCG Study Re- Mystery of B-to-B Forecasts Revealed.” The
Evaluates Size, Growth and Importance of Standard, 21 February 2000. [Available on-
Business-to-Business E-Commerce.” Press line: 具http://www.thestandard.com/research/
release, 21 December 1999. [Available on- metrics/display/0,2799,11300,00.html典.]
line: 具http://www.bcg.com/media_center/ Lum, Sherlene K. S.; Moyer, Brian C. and
media_press_release_archive2.asp典.] Yuskavage, Robert E. “Improved Estimates of
Bureau of Labor Statistics. “Pilot Survey of the Gross Product by Industry for 1947–98.” Sur-
Incidence of Stock Options in Private Indus- vey of Current Business, June 2000, 80(6),
try in 1999.” Survey report, Bureau of Labor pp. 24 –54. [Available online: 具http://www.
Statistics, U.S. Department of Labor, Wash- bea.doc.gov/bea/articles/National/NIPAREL/
ington, DC, 11 October 2000. [Available 2000/0600gpi.pdf典.]
online: 具http://www.bls.gov/special.requests/ Mesenbourg, Thomas L., Jr. “Satisfying Emerg-
ocwc/oclt/ncsocs/ncs/ncnr0001.txt典.] ing Data Needs, Measuring Electronic Busi-
Fraumeni, Barbara M.; Lawson, Ann M. and ness.” Unpublished manuscript presented at
Ehemann, G. Christian. “The National Ac- the International Conference on Establish-
counts in a Changing Economy: How BEA ment Surveys, Buffalo, NY, 21 June 2000.
Measures E-Commerce.” Unpublished manu- [Available online: 具http://www.census.gov/
script presented at the Brookings Workshop econ/www/ices.htm典.]
on Measuring E-Commerce, Washington, DC, PSINet. eBusiness, roadmap for the digital age.
24 September 1999. Asburn, VA: Forbes Special Interest Publica-
Fraumeni, Barbara M.; Manser, Marilyn E. and tions, Fall 2000, p. 65.
Mesenbourg, Thomas L., Jr. “Government Sturgeon, Julie. “E-Commerce: Beyond Sales,”
Statistics: E-commerce and the Electronic in Government Services Group, The year in
Economy.” Unpublished manuscript pre- computing 2000. Tampa, FL: Fairncourt
sented at the Federal Economic Statistics Ad- LLC, Fall 2000, pp. 54 – 67.
visory Committee Meeting, Washington, DC, U.S. Department of Commerce. “Retail
15 June 2000. [Available online: 具http://www. E-commerce Sales in Third Quarter 2000 In-
census.gov/econ/www/ebusiness614.htm典.] creased 15.3 Percent from Second Quarter
Garicano, Luis and Kaplan, Steven N. “The 2000, Census Bureau Reports.” U.S. Depart-
Effects of Business-to-Business E-Com- ment of Commerce News (Economics and
merce on Transactions Costs.” National Statistics Administration, U.S. Census Bu-
Bureau of Economic Research (Cambridge, reau), 27 November 2000. [Available on-
MA) Working Paper No. 8017, November line: 具http://www.census.gov/mrts/www/
2000. 20003q.html典.]

View publication stats