.

E-Commerce in United States Agriculture
ERS White Paper

E-Commerce Team
William Chambers, Jeffrey Hopkins, Kenneth Nelson (Coordinator),
Janet Perry (Division Liaison), Shirley Pryor, Peter Stenberg, Thomas Worth

May 25, 2001

1


.

E-Commerce in U.S. Agriculture
I.

Introduction

The effect of e-commerce on the agricultural sector, and the implications for farmers and
agribusiness, has for some time been considered a controversial topic, reflected in the dissenting
opinions of critics, proponents, and analysts. Some proponents have said it is a great cost
reducer and economic equalizer providing access for all to information and markets. Some
critics see a mechanism for concentrating information and market access among the powerful
and relegating those left out to the digital desert. Other analysts view e-commerce through the
eyes of a market investor, where efficient market theory predicts a sharing of risks and gains
between the firms developing technological innovations and the farm and agribusiness managers
and households that implement them.
Given the secular decline in the market performance of e-commerce firms over the last year, it

appears obvious that some investors would re-evaluate their decisions had they known what they
know now. However, the outlook for e-commerce in agriculture is still largely driven by the
implications of the technology for management of firms and the implications of the technology
for the well-being of households. The authors’ aims in this paper are to collect and report briefly
on what is known about the type, use, and impacts of e-commerce in U.S. agriculture, both at the
farm and agribusiness levels. We address the hopes of the proponents, the fears of the critics,
and the outlook for further investment. With this, we hope to identify researchable issues related
to e-commerce which match the ERS mission of providing information for a competitive
agriculture.
We begin with a section that places e-commerce within a broader information technology
context (a glossary of terms is included at the end of the document). Then, we draw extensively
on existing literature to discuss the macroeconomic and microeconomic implications of ecommerce for agriculture, using the more formal economics, management, and technology
literature as well as the popular press, which has a great deal to say about economics,
management, and technology. The literature review section is followed by a section that creates

2


.
a typology of e-commerce firms found in the agricultural sector. Next, we discuss a key issue
related to the availability of e-commerce data for the farm and nonfarm sector, concentrating on
trends in business-to-business sales for the economy as a whole and the growing use of ecommerce by traditional agribusinesses. We also show estimates of the gaps in Internet use (the
so-called “digital divide”) in the U.S. and gaps in online purchases and sales among farm
households. Finally, some research ideas that we believe merit further study at ERS are
presented.

II. Defining Information Technology and E-Commerce
The “Information Revolution” has been going on well before the creation of the Internet. It has
inspired lofty rhetoric about how it will change the way we live and conduct business. It has also
created several new concepts and terms. Some of the terms border on slang (“B-to-B”) with

definitions that are still evolving. The purpose of this section is to define and clarify the basic
terms used in the rest of this paper.
A. Information Technology
The broadest term is “information technology” (IT). It encompasses everything that allows us to
electronically gather, generate, store, analyze, distribute, or otherwise use information. Although
the Internet has received the most attention, IT includes other computer technology such as
microchips, monitors, hard drives, and software. It also includes more traditional
telecommunication technologies such as cell phones or fax machines - anything related to the
electronic use of information. More recently, IT has begun to include broadcast technologies
such as cable TV that are offering access to the Internet. It is IT that makes e-commerce
possible.
Although much of IT relates to communication and networks, it also has an important role that is
separate from the exchange of information. For example, laser scanners at supermarket checkout
counters offer advantages over manually typing in the price of each item. Word processing
programs on personal computers offer many advantages over typewriters that are separate from
the ability to transmit word processing files over the Internet. These “non-network” technologies

3


.
are not the focus of this paper, but it is important to note that these may have been an important
source of change in how businesses operate.
B. Network Technology, E-Business, and E-Commerce
Though the focus of this paper is on the effect of e-commerce on agribusiness, it is helpful to
define the technology that makes e-commerce possible. Network technology is not new. The
telephone and the fax have been around a long time, and their effects are well understood. More
relevant to e-commerce are the newer technologies that link computers together.
There are two major types of computer networks that are used for e-commerce; electronic data
interchange (EDI) and the Internet. EDI systems are proprietary networks that enable firms to

send data between remote establishments or to link with other firms and conduct business
transactions. EDI predates the Internet. One advantage of an EDI is that it is private and the
data transmitted through it is secure from the public. EDI systems tend to be more expensive
than the costs of just connecting to the Internet.
While EDI networks are private, the Internet is an electronic network open to the general public.
Any firm or individual with the right equipment may access it. The Internet can be used for
transactions between firms as well as transactions between firms and individuals. Another
network technology, “extranet”, is a hybrid of EDI and the Internet. An extranet uses the
Internet to transfer information, but encodes the information to maintain privacy.
The rapid growth of the Internet has given rise to the concept of electronic commerce, or ecommerce. Popularly, e-commerce is thought of as conducting business on the Internet. The
U.S. Census Bureau, however, has a more precise and sweeping definition of e-commerce as any
transaction completed over a computer-mediated network. This includes any transactions over
the Internet as well as through an EDI system or extranet. It is this definition that guides their
measurement of the volume of e-commerce.
The Census Bureau has also constructed a definition of electronic business, or e-business. It is
any process that a firm or other organization conducts over a computer-mediated network. This
includes anything that would be classified as e-commerce. It also includes more functions than

4


.
those that are price transactions, such as logistics, customer support, sending email, or
advertising.
C. E-Commerce Transactions: B-to-B and B-to-C
Firms that do business on the Internet, whether exclusively or part of a broader business,
generally fall into two categories; B-to-B and B-to-C. B-to-B is short for “business-to-business”
and applies to transactions between firms. This includes firms electronically purchasing inputs
from other firms. This also applies to firms that act as intermediaries, or wholesalers, between
firms. The term B-to-C, or “business-to-consumer”, describes firms that sell goods or services to

a final consumer, usually a member of the general public. Firms may be involved in both B-to-B
and B-to-C transactions. The relationship between e-commerce and IT is illustrated in Figure 1.
These two categories do not represent equal shares of e-commerce. According to the Census
Bureau data, ninety percent of e-commerce consists of B-to-B transactions. These transactions
occur on either an EDI system or the Internet. B-to-C transactions occur almost exclusively on
the Internet, where the general public has access.
A well-publicized example of an e-commerce company is the bookseller Amazon.com. It is an
example of a “pure play” company in that it buys and sells only over the Internet. Pure play
companies are frequently referred to as “dot-com’s” because of the suffix of their Internet
addresses. Many traditional, or “brick and mortar”, companies also participate in e-commerce
through websites where they buy and sell commodities. Dot-com companies have gone through
considerable transformation over the last year, and many of them have experienced large stock
value losses.

5


.

Figure 1.
The Relationship between Information Technology E-Business, and E-Commerce

E-Business

Information Technology

Network
Technology

E-Commerce

B-to-B

EDI
Extranet

B-to-C
Internet

6

NonNetwork
Technology


.

III.

Literature Review and Discussion

A. What Are the Major Sources of Information about the Information Highway?
There is little data available on the use or the impact of e-commerce, e-business or the Internet.
Researchers tend to be discouraged when lacking data, thus there is little written in journals.
There is little in the literature about e-commerce in the food and agricultural sector. The
economics literature is more fertile and business schools seem to have taken on the challenge of
understanding e-commerce with even more enthusiasm. This section primarily reviews those
business school articles that have made their way into economic journals. However, most of the
information about the sector must be gathered from magazines and newspapers and talking with
industry people. The Economist periodically has articles on e-commerce and the Internet.
Business Week regularly has a section entitled Ebiz and the New York Times and the Wall Street

Journal have insightful articles on e-commerce and related topics. It is particularly important to
look at the popular press because conditions are changing so quickly, even though much of the
information is anecdotal. Forrester Research specializes in collecting quantitative information
about the Internet although it is not clear how accurate the data are. Because new data must be
collected, current analysis relies on qualitative data collected through interviews and magazine
and newspaper article or purely theoretical analysis. The lack of availability of data has
implications for the direction and scope of ERS analysis of the impact of the Internet.
B. Is There A New E-conomics?
One of the questions asked during the earlier days of e-commerce was: "Is there an economics of
e-commerce?" Borenstein and Saloner quote their colleagues as saying “traditional economic
analysis and tools are of little value in analyzing electronic commerce” (p. 4) This is, of course,
not true, but there are some ideas in economics which are particularly relevant to understanding
the role of e-commerce.
R.H. Coase in “The Nature of the Firm” (1937) described the role of transaction costs in
determining the nature of the firm. Transaction costs are costs that are above and beyond the

7


.
price of the commodities sold. The costs of contracting are one of the major costs of using the
market. According to Coase, transactions costs help determine market structure. When
transaction costs are low, the firm is more likely to purchase whatever is needed from outside the
firm. If the costs are high, the firm is more likely to produce from within.
Understanding transaction costs may be the key to understanding the impact of e-commerce.
Oliver Williamson defined the market and the firm as alternative means of performing certain
transactions. Their relative costs determine which one you use. Human and environmental
factors determine which one is least cost. Thus, uncertainty and opportunities for opportunistic
behavior encourage firms to produce for themselves rather than contract out. (Carlton, Perloff, p
5) The availability of more readily accessible information would decrease the uncertainty and

opportunistic behavior.
Information technology and e-commerce would presumably lower transaction costs, and so
according to Coase and Williamson, will have an important impact on market structure.
Search costs, one component of transaction costs, are important in the economics of the Internet.
Advertising is the method sellers have for providing buyers with product prices. Buyers, on the
other hand, “search” for prices. “Prices change with varying frequency in all markets, and unless
a market is completely centralized, no one will know all the prices which various sellers (or
buyers) quote at any given time. A buyer (or seller) who wishes to ascertain the most favorable
price must canvass various sellers (or buyers) – a phenomenon I shall term search” (Stigler, p.
213). These costs become a part of the buyer’s total costs. The Internet should lower the search
costs to the buyer and well as to the seller.
Shapiro and Varian, in their book Information Rules, discuss some of the economic concepts
particularly relevant to information technology. “Information is costly to produce but cheap to
reproduce” (p. 3) in other words it has high fixed costs and low marginal costs. Information is an
experience good, thus you must experience it to know if you value it. Most goods are experience
goods and advertisers use such things as promotional samples to help consumers experience the
products. Information has the particular characteristic in that it “is an experience good every
time you consume it.” (p. 5) Thus, the media use branding and reputation to overcome that
problem. “Users of information technologies are notoriously subject to switching costs and lock-

8


.
in: Once you have chosen a technology or a format for keeping information, switching can be
very expensive” (p. 11). Network externalities are very important in the information industries
and they exhibit positive feedback. Fax machines are a good example. Sharp produced low cost
fax machines in 1984 and sold 80,000 of them. (Gladwell, p. 12) In 1987 a million machines
were sold and two million were sold the following year. Gladwell described the phenomenon as
a tipping point. “Nineteen eighty-seven was the fax machine tipping point.” (Gladwell, p. 12)

A critical mass of ownership is achieved after which it makes sense to buy a fax machine.
Nineteen ninety-eight was the tipping point for cell phones in the U.S. (Gladwell, p. 12 )
C. What Has Been the Impact on the Economy?
Borenstein and Saloner (p. 4) detail how the Internet creates value. “The Internet creates value
by vastly lowering the cost of transferring many types of information…. In cases where the
product itself is information, the potential for value creation is enormous.” The Internet has the
following advantages that may give some clues as to which industries will be most affected. The
Internet provides very detailed content; it allows for communication at any time thus possibly
having a more profound effect across time zones. Finally, the Internet allows flexibility in
dealing with information that allows more personal service (p. 5). This may not be the same type
of service that is provided by face to face contact, but it may be less costly and also may provide
personal services to a wider group of people. (p. 5) A good example of the expansion of
personal services is the purchase of clothing over the Internet. If you provide some companies
with your body measurements they will show you how the clothing might look on you. This
certainly does not replace a personal visit to the store, but it does provide a much larger group of
people the option of trying on the clothes-- even if it is virtually.
Jorgenson and Stiroh (2000), wrote that information technology has had a tremendous impact on
the economy. However, they argue that the impact has been in the manufacturing sector because
of the production and sales of computers and associated equipment. They also reported that
there is no indication that information technology has had a positive impact on the service sector
(where computer use would be expected to contribute to growth). They attributed this lack of
impact partially to the difficulty in measuring productivity in the service sector.

9


.
Brynjolfsson and Hitt make the point that the impact of information technology is difficult to
measure because it is combined with changes in “work practices, strategy, and products and
services.” “These complementary investments, and the resulting assets, maybe as much as an

order of magnitude larger than the investments in the computer technology itself…However,
they go largely uncounted in national accounts...”(p. 45).
More recent analysis by Martin Bailey and Robert Lawrence indicates that there is clear evidence
that information technology has had an impact on productivity in the service sector (p. 1). They
state that they have not proven that information technology is boosting productivity in the service
sector but their evidence is consistent with that view (p. 4). The Internet has in the past and will
in the future continue to lower the costs of communication (p. 8).
D. What Is the Impact on the Structure of Industries and on Competition?
Internet use is apparently having an impact on the structure of industries, both encouraging the
disintegration of the value chain, and in other cases encouraging concentration. “Companies are
outsourcing parts of their value chain to concentrate on core competencies – extending the
benefits of comparative advantage.” (Bailey and Lawrence, p. 5) An example of this according to
Bailey and Lawrence is research and development, which is being outsourced by some larger
companies. Thus new research and development companies are small and are increasing in
numbers (p. 6). Kagan and Conklin suggest that rapid adoption of information technology will
likely have a negative impact on non-metropolitan community banks and will lead to a loss in
market share. (p. 93) Borenstein and Saloner indicate that even though many Internet activities
appear to encourage competition, many markets are “characterized by … increasing market
concentration”(p. 9). Bailey and Lawrence’s conclusion is that “there is a new economy” if one
means that there “has been a wave of innovation, much of it tied to IT driving greatly improved
economic performance in this expansion, affecting old and new firms”(p. 8).
Aoki uses a theoretical model to examine “the impact of an oligopolistic upstream electronic
market on upstream and downstream prices. (Aoki, Abstract) Aoki’s work has the advantage of
not requiring any data.

10


.
Price dispersion is an indicator of “ignorance” in the market (Stigler). The Internet should

lower search costs and encourage a more competitive market, lower prices and less price
dispersion. Industry studies do not indicate that these are necessarily true. In the insurance
industry, Brown and Goolsbee find that, “a 10 percent increase in the share of individuals in a
group using the Internet reduces average insurance prices for a group by as much as 5 percent”
(Brown and Goolsbee, Abstract). Prices for those not using the Internet do not change. They go
on to say this increased price dispersion is consistent with the literature that would indicate that if
a small group of people have access to more information than others; the price dispersion would
increase (p. 17).
Morton, Zettlemeyer and Risso examine Internet automobile referral services and their impact on
dealer profits and auto prices and find that the, “Internet has changed the product market
behavior of dealerships. While the Internet performs the expected roles of reducing search costs
and serving consumers’ information needs, it also is changing firm pricing behavior in an
important sector of the economy”(p. 25). Customers who are referred to various dealerships
through the Internet save, on average, 2 percent per car, about $450. Internet sales also appear to
have an impact on price dispersion. Price dispersion declines according to the amount the
dealership has sold through Autobytel.com. Dealership profits also increased. The costs of
dealing with customers on-line are less than face to face (p. 26).
Clay, Krishnan and Wolff examined book sales over the Internet from data collected between
August 1999 to January 2000. They test whether low cost price information on the Internet will
force price and price dispersion to fall; and whether firms will seek to differentiate their products
(p. 3). Firms prefer to differentiate their product in order to increase their profits. However, the
Internet, by providing more readily available price information, should drive down prices. They
find “no evidence that prices or price dispersion are converging” (p. 4) but do find evidence that
firms attempt to differentiate their product (p. 5).
Other industry studies examine related issues. Goolsbee examines competition for computer
purchases between the Internet and retail sales. Goolsbee finds that the retail price for the
computer has a significant impact on whether or not a person is willing to buy on-line (p. 3).
Gertner and Stillman examine whether or not vertical integration is an impediment to change and

11



.
whether or not an apparel store decided to sell on line as an example of changes. They find that
“vertically integrated specialty retailers tended to start on-line sales sooner.” (Gertner and
Stillman, Abstract) Barber and Odean discuss the impact of on-line trading and show that it has
enabled newer and smaller firms to challenge the old guard brokerage firms. They speculate on
the impact of more and more people unfamiliar with the functioning of the market trading
without guidance from a broker. Autor examines the labor market and has three conclusions:
•

The Internet will change the way employee-employer matches are made,

•

Labor … will increasingly be delivered over the Internet not on-site

•

The demand for labor will become less dependent upon local conditions. (Autor, p. 1)

E. Are the Dot-Coms Viable Businesses?
Recent data indicate that venture capital firms and market investors have lost faith in the dotcoms ( And, the recent stock market drop indicates that
the investing public also lost enthusiasm for dot-coms. Numerous newspaper and magazine
articles identify the major issue. The dot-coms are having trouble identifying a business model
that allows them to earn a profit. Without a source of profit, many firms are going out of
business. Dot-coms have earned funds by allowing advertising. However, the very technology
which allows the firms to identify which customers are most likely to buy which product also
allows the advertisers to identify the hits on the advertising banner and the amount of purchases
resulting from the hit. The advertisers have been able to document that the advertising has not

been generating much revenue.
Auction sites such as eBay have been able to charge a commission on each sale. The sales price
and the seller and buyer are easy to identify and eBay has developed a method which allows the
buyer and seller to verify the reliability of the other. Some spokespeople for the fruit and
vegetable industry indicate that the margins from on-line auctions are so small that it is
impossible to pay a commission to the on-line auction dot-com. Auctions have another
problem. “Some companies have close confidential relationships with their buyers and their
sellers. They want to improve those, not shift to open or anonymous transactions that might
require them to disclose pricing or other proprietary information” (Clark, p. B8).
12


.
Model N has developed software that allows firms to negotiate with each other on price and
quantity. The Chicago-based meat and poultry exchange called Provision X was built using
Model N’s software and has as its clients Cargill and Tyson Foods. The software allows these
firms to complete their transactions in a timely fashion (Clay). They are asking firms to pay
subscriptions in order to use their software.
F. What Are E-Commerce And E-Business’ Impacts On Global Trade?
Freund and Weinhold conclude, “the Internet stimulates trade.” Using a gravity model and
recent trade and Internet data, they conclude that their data support the following predictions
from theory:
•

Trade should expand because sunk costs associated with trade are reduced.

•

The development of global markets via the Internet implies that historical linkages should be
less important.


•

Countries with the fewest past trade linkage (most likely the poor countries) have the most to
gain from the Internet. (Freund, page 24.)

G. How Will Adoption of E-Commerce Affect Production Efficiency in Agriculture?

The economy-wide effect of e-commerce on production efficiency is the object of many current
debates. We interpret the question of the effect of e-commerce on productivity and growth as one
couched within a broader debate over the effect of IT in general on growth rates over time for an
economy. This question is most often answered in aggregate through estimation of the effects of
technology on the rates of growth or on factor productivity. The general tone of those who feel
that growth paths have been fundamentally altered, via increases in factor productivity, is that
there is a “new economy” that is capable of sustained growth. Economic theories of sustained
growth differ from the neoclassical growth models of Robert Solow (1956) that predict steady
state growth. New growth theory predicts that sustainable growth levels are the result either
from spillovers from endogenized technological change (attributed to Paul Romer 1990) or from

13


.
human capital investments (attributed to Robert Lucas 1988). Both of these explanations are
given an airing in the intuitive claims that information and communication technologies have
resulted in efficiency gains throughout the economy.

Those who look to answer the question of whether some fundamental changes in the economy
have occurred usually look for evidence in measures of efficiency. Conventional measures up to
the 1990’s showed little effect of information and communications technology on efficiency,

leading Robert Solow to declare that these effects are “seen everywhere except in the
government statistics” (quoted in David, 1990). The government statistic of interest to many
economists was labor productivity, because increases in labor productivity potentially allow for
an increase in output without increasing wages, and is therefore non-inflationary. Labor
productivity growth was relatively constant from 1975 throughout the first half of the 1990’s at
1.4% per year, but doubled to nearly 3% throughout the latter part of the 1990’s.

The Economist (May 12, 2001) summarized recent interpretations of these findings in the context
of a measured downturn in labor productivity in the first quarter of 2001. The downturn
suggested to some that the increase in productivity from the mid-1990’s was tied to cyclical
growth, rather than structural growth that would be anti-inflationary. Gordon (2000) is the
champion of the skeptics, believing that only about 2% of the growth in labor productivity was
structural. On the other side are estimates by the Federal Reserve (Oliner and Sichel 2000) and
OECD showing structural causes behind almost all of the productivity gains.

H. How Will E-Commerce Affect the Distribution of Well-Being Across Farms and Farm
Households?

Much of the debate about whether a “new economy” exists revolves around productivity
measures that are highly aggregated. For example, much of the debate cited above revolves
around the measures calculated for the entire non-farm economy. Casual observers might reach
the conclusion that if researchers are unable to tell plausible stories about the effect of IT at the
macroeconomic level, e-commerce might not be very relevant at more micro levels.

14


.
At one level, this may be correct, because if productivity is not affected then there is little
opportunity for the competitive advantage of individual firms to be affected, and little reason to

suspect that there would exist any winners or losers as a result of technological innovations.
This offers scant cover for those who might want to ignore the effect of e-commerce, however.
In the first place, the effects of IT on the economy have been measured and determined to be
positive and any remaining debate concerns the magnitude of the impact. More significantly, in a
competitive industry such as agriculture even small changes in productivity can affect the
performance of firms relative to each other, particularly when the technology is adapted nonuniformly across the sector. Under this scenario, despite only marginal changes in aggregate
productivity, winners and losers can still be identified. Variables that are fixed in the short run,
such as farm size and location and the demographics of farms and families that farm, as well as
their financial or human capital endowments, are frequent candidates for identifying winners and
losers. In production agriculture, these identifying characteristics are commonly referred to as
the structure of farming.
One way to come to grips with how e-commerce and IT in general might impact the structure of
the production agriculture sector is to look at the process of the diffusion of innovations. The
diffusion of innovations is important because technology diffusion is often assumed path
dependent, i.e. the way that technology evolves is dependent on the process itself, leading to
many possible points of equilibrium. This is distinguished from path-independent diffusion,
where a unique, global equilibrium can be identified from the outset, perhaps determined from
the fundamental characteristics of the technology.
Two theories of diffusion, the epidemic theory and the threshold theory, are relevant in the case
of agricultural e-commerce. Epidemic theories are based on the premise that information drives
diffusion. Under epidemic theories, the value of the innovation is clear at the start and the greater
the value of the innovation the more “contagious” the disease is to non-adopters. Highly valuable
innovations diffuse very quickly. According to epidemic theories of diffusion, market-driven
rates of adoption may be sub-optimal, and adoption may be subsidized or the government may
take a role in speeding the transmission of information from adopters to non-adopters. This is
the historical role played by USDA through land-grant universities. Epidemic theories of

15



.
diffusion assume that the potential group of adopters is relatively similar and the benefits of
adoption are known at the outset.

Threshold theories of diffusion, in contrast, consider the net benefits of adoption to be highly
variable across the pool of potential adopters, and rank order the adopters according to net
benefits using probit econometric models (David 1999). This ordering is often correlated with
variables that define industry structure, such as firm size. Technology adoption studies in
agriculture often find farm size and other structural characteristics, such as household
demographics and financial and human capital stocks, to be important determinants of adoption.
However, it is the interdependence of technology adoption and size that is potentially the more
pervasive effect, as producers who adopt a technology find that adoption allows them to increase
the size of their operations. Where adoption and size are determined jointly, the structure of the
industry can become endogenous to the diffusion process. For example, expansion of firm size
by adopters, if accompanied by declines in output prices, could work to the disadvantage of nonadopters, who must reduce their output as a group. In this scenario, the size of the production
agriculture sector could become endogenous to the diffusion process, as non-adopters are
expelled from the market. Farmers looking to speed up their rate of adoption, but not willing to,
unable to, expand in order to make adoption profitable can look at either sharing the technology
or renting it (Lissoni 2000). For example, smaller farmers could fund an e-commerce venture
through a cooperative or otherwise share the technology among many smaller farmers.
Alternatively, they could rent the technology, buying e-commerce services from a larger
provider.

Contrary to the scenario where the benefits of adoption are positively correlated with farm size,
there are those who believe that e-commerce can help small producers willing to seek out and
produce for niche markets, relative to larger farms that tend to be more closely aligned to the
agricultural sector hierarchy (Buhr 2000). In fact, many niche marketers are thriving as a result
of e-commerce, as markets are opened through the Internet without incurring the high costs
associated through traditional marketing procedures. Recent evidence on adoption rates of ecommerce among agricultural producers has indicated some correlation with farm size


16


.
(measured by sales volume) as adopters have been better educated and younger than nonadopters (Morehart and Hopkins 2000). Because this is a common pattern seen with other
studies of technology adoption in agriculture, it may be indicator of willingness to undertake an
activity with an uncertain outcome, rather than either ex ante or ex post association with size.

I. Can Policymakers Do Anything to Influence the Rate of Adoption?

The possibility of path dependence is often an essential component of most arguments of a role
for government in establishing a technology policy (Jaffe, Newell, and Stavins 2000). Moreover,
because policy objectives in agriculture frequently have pre-conceived notions with regard to
agricultural structure (such as preserving the family farm) there is often a desire to pursue
general policy objectives through technology policy. Technology policy is usually pursued by
subsidizing adoption or by promoting the free flow of information. In the case of the Internet,
like many forms of technology used in agriculture, government had a clear primary research role
as well.

Path dependence is often cited alongside network externalities. Network externalities, the case
where the benefits of a technology are a function of the number of people using the technology,
can potentially be either an opportunity or a cost to society, therefore government could
potentially play a role in helping to minimize social cost. An opportunity from network
externalities arises through standard-setting in order to minimize costly delays and misallocated
resources among the private promotion of many competing standards. Network externalities can
also be negative if another technology is superior from some broader social sense but users are
“locked in” to the existing one. The economic significance of path dependence is difficult to
assess, however, as one must assess which technology is “socially” superior as well as the cost
associated with non-superior technologies.


A less interventionist role may be played by government if it restricts its activities to providing
information to the market. Because information is a public good, there is reason to believe that it
will be naturally under-provided by markets. This information may take the form of helping
producers identify the size of a network and by implication the value to producers of joining the

17


.
network. Alternatively, the information may take the form of identifying bottlenecks in the
diffusion path, and coordinating cooperative or sharing-type alternatives for farmers that allow
the adoption decision to be accelerated. David (2001) suggests a different approach and offers
the following question as a guide: “Can we identify situations in which it is likely that at some
future time individuals really would be better off had another equilibrium been selected from the
beginning?” One action that governments might credibly make is to delay the market from taking
an irreversible decision before sufficient information is gathered. However, in general, we are far
from having a well-established theoretical or empirical basis for when intervention is preferable
to the outcome of an unregulated market outcome, or the form that intervention should take
(Jaffe, Newell, and Stavins 2000).

III. Functions That E-Commerce Firms Perform in the Agricultural Sector
Although e-commerce has promised to re-shape agricultural production and marketing in
profound ways, few promises have come to fruition. In the span of less than five years
thousands of agricultural related web-sites have been formed, representing a variety of business
models and market areas. Although there remains a tremendous amount of potential with
information technology, agricultural e-commerce firms—like dot-coms throughout the
economy—have had limited success thus far. A fundamental issue regarding agricultural ecommerce is that we cannot be sure how it will be used. This makes it difficult to know if or
how it might affect firm structure or agricultural markets. Most firms throughout the ecommerce sector are themselves experimenting with business strategies and because of this it
will be some time before we fully understand how it will affect agriculture.
We are currently in the midst of an important change in the Internet economy that is affecting

agricultural firms in much the same way as other dot-coms. One of the major impacts from the
change is that venture capital is being reduced and firms are under greater pressure to show
profitability. Because of this, companies are changing business models and moving into areas
that they feel will be more profitable. Another possibility is that the dot-com sector could
experience consolidations or failures in the near future.
A significant proportion of agricultural e-commerce falls in the category of (B-to-B). Businessto-business Internet e-commerce came after the business-to-consumer and is supposed to enable
18


.
firms to use the Internet to reduce the costs of doing business. Agricultural e-commerce firms
typically make money from membership fees, banner ads, or commissions. B-to-B, like B-to-C
before it, has thus far proven to be less effective than originally predicted. We have identified
four different functions that e-commerce firms perform in the agricultural sector, and many firms
perform multiple services. These functions are (1) information distribution, (2) input supply, (3)
commodity trading floor, and (4) logistics/supply chain management. Below we provide a
discussion of each.
A. Information Distribution
Information distribution illustrates one of the major strengths of the Internet because it simplifies
the distribution of data and analysis. Many agricultural related e-commerce firms distribute
some sort of information, which often include weather reports, futures market quotes, and market
analysis. These services were provided for free at many of the e-commerce sites we visited
(although this may change as firms struggle to find profitable uses of the Internet). Information
distribution on the Internet has significantly enhanced the level of market information available
(although information overload could become a problem). Even if no future developments occur
(which is unlikely), the Internet will have had a major impact on agriculture.
There are many examples of the Internet being used for information distribution, not necessarily
from for-profit companies. It is common for trade associations, government agencies, or
universities to provide information to the public, and the Internet enhances this activity. For
example, the ERS home page (www.ers.usda.gov) contains much of the analysis and data

developed at Economic Research Service and USDA. The Internet allows users to receive
USDA market and trade reports as soon as they are released. Prior to this, users would receive
the reports through the postal service or by fax.
An example of a private company that provides market information is Oryza (www.oryza.com).
Oryza was established to serve the rice industry, and part of their core business is operating an
electronic rice-trading floor. As a service (particularly to traders), they offer a variety of ricerelated market information that is free of charge. In addition to providing prices—both domestic
and international—Oryza offers a broad variety of reports, which analyze major market events
for the U.S. and key rice regions throughout the world.

19


.
B. Input Supply
Another major category of B-to-B agricultural e-commerce is input supply. E-commerce input
supply firms provide various inputs to farmers including seed, chemicals, parts, and machinery
over the Internet. Their operation is little different than traditional input suppliers except that
they are not bound to a certain geographic area. Farmers are able to search the Internet for the
inputs they need and compare prices between suppliers. Buyers and sellers agree on price,
quantity, and transportation of the desired product.
An example of an e-commerce input supplier is XSAg.com (www.xsag.com). This company
facilitates electronic commerce for buyers and sellers of farm inputs including chemicals, seeds,
parts and equipment, and animal health products. Users can access the web-site and search for
the products they need or display products they wish to sell. A product listing is offered along
with prices. The farmer then selects the best option for his or her needs. In their users
agreement, XSAg.com states that they are not involved in the actual transaction between buyers
and sellers. Because of this, they have no control over the quality, safety, or legality of the items
advertised, the accuracy of the listings, the ability of the sellers to sell items or the ability of
buyers to buy items. The company views itself as a facilitator for a transaction that takes place
between a buyer and a seller. XSAg.com makes money by charging the seller a 3 percent

commission for all transactions (the buyer doesn’t pay a commission). XSAg.com also charges
buyers a $100 membership fee for some of the products they offer, which is refunded if an actual
transaction is made.
E-commerce has the potential to increase the level of competition in the input supply business.
One important change is that customers will be able to more easily see what different suppliers
are charging, and agricultural inputs are generally highly substitutable. Porter argues that some
companies have used the Internet to shift the basis of competition away from quality, features,
and service towards price alone. This may happen in the agricultural input supply sector. The
Internet could reduce a supplier’s ability to charge higher prices because consumers would know
what other firms were charging. Suppliers may differentiate themselves by using a sales force or
providing good customer service. This could lead to a highly competitive situation that is good
for customers, but harmful to supply companies. It has been argued that input suppliers could

20


.
use the Internet to reduce or eliminate their sales staff, and that this would increase firm
profitability. However, a sales force is also a potentially powerful source of competitive
advantage for input supply companies, and may offer consumers valuable services that the
Internet cannot.
Competition could also be damaging to the e-commerce firm itself. If farmers make purchase
decisions based solely on price and didn’t have strong loyalty towards the e-commerce firm, low
barriers to entry would limit profitability. Unless e-commerce firms were able to erect barriers to
entry (such as licensing agreements) or develop other forms of competitive advantage (such as
branding), competition could be fierce and profits low. For that reason, e-commerce firms may
try to erect barriers to entry—such as licensing agreements—to enhance their competitive
position.
C. Commodity Trading Exchange
Electronic commodity trading floors are also an important development in agricultural ecommerce. The main idea with an electronic trading floor is that the company can host a

marketplace and earn a small commission on every transaction. Since volume is large with most
agricultural markets, many firms hope commission sales will lead to significant profits. It is also
argued that electronic transactions could be more liquid and efficient than traditional markets
because geographic boundaries are eliminated and the potential number of users could increase.
Although most agricultural markets (particularly those for grains) are generally considered to be
highly efficient, this innovation could enhance the information flow through agricultural
markets.
An example of a commodity-trading floor is CattleSale.com (www.cattlesale.com), which offers
markets for different types of cattle. Members can search the site for various categories of cattle,
observe prices and engage in transactions. CattleSale.com seeks to increase the seller’s market
exposure while providing buyers a greater selection of cattle. The firm argues that the Internet
can be used to enhance the price discovery process, which is particularly important in the
livestock industry. E-commerce livestock markets (and electronic trading floors in general) are a
highly competitive business. Firms competing with CattleSale.com include E-Merge Interactive
(www.emergeinteractive.com) and Farms.com (www.farms.com) as well as the traditional

21


.
marketing outlets. Also, the barriers to entry to developing a commodity-trading site are very
low, which means that a competitor could easily enter into the Internet business. Interestingly, at
time of writing, the company Web Sales Group was renting the domain name “cattlesales.com”
to anyone who wanted to enter the electronic cattle marketing business.
Another example of an electronic trading floor is AgEx (www.agex.com). AgEx was formed in
1999 and operated trading floors for almonds, dried fruit, pecans, walnuts, juice, processed and
fresh tomatoes, apples, pulses, and rice. The company also arranged electronic auctions
(generally for larger volumes than transacted on their trading floor), and offered information
services for each of their different commodities. The company made money from a membership
fee, commissions, and banner ads. In early 2001, AgEx announced a general restructuring of

their business. They eliminated their trading floors and scaled back the information services they
provided. AgEx continues to host commodity auctions and performs other for fee services.
There are other firms that are having problems with Internet commerce. For example, IBP Inc.,
Cargill Inc. and its Excel meat processing business, Farmland Industries Inc., Smithfield Foods
Inc., Gold Kist Inc. and Tyson Foods Inc. announced in 2000 plans to establish what was
described as the world's giant of agriculture business-to-business e-commerce sites. These six
companies hold large market positions in beef, pork and poultry production, and analysts—
including some who have family farmers and ranchers for clients—hailed the partnership as
something that would improve the competitiveness of meat and poultry production (Feedstuffs).
Soon after the announcement the U.S. Department of Justice (DOJ) confirmed that it would
investigate possible antitrust issues arising from the plan by six major meatpackers to establish
an Internet-based, business-to-business marketplace. At time of writing, this trading site had not
been developed yet.
The key to the success of e-commerce market exchanges (or any market exchange) is a high
level of liquidity. This ensures enough buyers and sellers to facilitate the price discovery
process. The need to generate liquidity explains why there has been such a push to develop
commodity-trading sites quickly as firms play a highly competitive game where there will, at
most, only be a few winners. Because of this, several of the livestock Internet exchanges have
minimum volume requirements. Other commodity exchanges such as the Chicago Board of

22


.
Trade (CBOT), EUREX, and the Minneapolis Grain Exchange offer forms of electronic trading.
It is also noteworthy that many large exchanges—such as the CBOT and the New York Stock
Exchange—are not-for-profit companies. Many exchanges perform a public good role as
opposed to a profit-making role.
A possible flaw in the electronic exchange model is that they coordinate transactions for highly
standardized commodities. Not only are current markets effective for these kinds of transactions,

but it counters the trend of non-standardization, which is becoming increasingly important
throughout agriculture. Although traditional markets are still important, we are moving away
from a situation where agricultural transactions are made for non-differentiated products and
price is the only concern. Wise and Morrison argue that the current B-to-B exchange model is
flawed in that it fails to recognize that obtaining a commodity simply for the lowest price may
not be in the best interest of a firm. They claim that other factors must be considered including
quality, timing of deliveries, and customization. These same non-price issues are becoming
increasingly important with agricultural transactions.
D. Logistics/Supply Chain Management
Another service that e-commerce firms provide deals with logistics and supply chain
management. Supply chain management has become a prominent buzzword throughout the Bto-B industry, partly because Internet exchanges have not achieved the promise that was
expected of them. An important issue with supply chain management is coordinating the
movement of products (particularly non-standard products) between buyers and sellers.
An example of supply chain management being used in the agricultural sector is E-Markets
(www.e-markets.com). E-Markets offers an information management tool called AgContract,
which is used to manage the logistics of agricultural contracts. This tool is particularly useful
when buyers require non-standard quality characteristics, which cannot be obtained through
traditional market channels. E-markets uses the Internet to develop contracts and identify
producers for different types of commodities. The firm verifies product quality, and tracks
product movement through the supply chain. The Internet is used throughout the process to keep
close contact between the contracting company and producers.

23


.
Using the Internet in this way is an improvement in a firm’s ability to manage supply chains, but
it is not new. Supply chain management has been an important field for some time and there has
been significant progress over the past two decades. For example, such tools as EDI have
worked to improve the linkages between buyers and sellers and enhance supply chain efficiency.

The Internet will clearly be a powerful tool in the management of supply chains because it has
improved the ability of managers to create networks. Still, using the Internet for supply chain
management should be considered a part of an overall improvement in management science
rather than a revolutionary tool that changes everything.
It is also not clear what firms will develop and manage these supply chain tools. A third party
could be used to coordinate between buyers and sellers. However, it may also be that a firm
could avoid this middleman and perform the function themselves. For example, DuPont
Specialty Grains (DSG) develops value-enhanced commodities (such as high-oil corn) for niche
markets (see www.dupontsg.com ). DSG uses an Internet based system called OSCAR to notify
producers of contracting opportunities for value-enhanced crops, and they manage the logistics
of this supply chain themselves.

IV. Evidence of a Divide within U.S. Farm and Nonfarm Households and
Businesses
A. Is There A Digital Divide in The U.S.?
Over the past several years, efforts to determine the existence of and contributing factors to a
digital divide have increased significantly. The use of computers and the Internet by households
has been chronicled by national data gathered between 1994 and August 2000 through the
Current Population Survey computer use supplements. The most recent results were reported in
“Falling Through the Net: Towards Digital Inclusion.” Analysis indicates that the overall rate of
digital inclusion is rapidly increasing and that groups traditionally characterized as digital “havenots” – rural households, people with limited college education, Blacks and Hispanics, women,
and people over 50 – have made particularly rapid gains. Nonetheless, a gap of 18 percentage
points existed between Black and Hispanic households and the national average. About half of
the gap was estimated to be due to differences in education and income between Black and

24


.
Hispanic households and the rest of the population. Another often-discussed divide is the quality

of Internet access for those using the Internet, as measured by the speed of the Internet
connection. Rural and urban households using the Internet, for example, were shown to use
broadband Internet services at different rates, with rural household use at 7.3 percent, and central
city and urban areas at 12.2 percent and 11.8 percent, respectively.
B. How Important Is E-Commerce to Businesses?
The E-stats initiative of the U.S. Census Bureau measures e-commerce activity as part of their
surveys of commercial activity of manufacturers, wholesale and retail trade, and service
industries, and represents 70 percent of all economic activity nationwide (but does not include
the agriculture sector). In 1999, $644 billion in electronic sales from wholesalers and services
was recorded, compared to $114 billion in 1998. The 1998 survey of online sales did not include
manufacturers, so a part of the increase was due to more inclusive measurement of the businessto-business sector. The high use of online orderings in manufacturing (12 percent of all
commercial activity in the sector was done online) was attributed at least in part to the
longstanding use of EDI in the sector. According to Bureau of Census data, 90 percent of all ecommerce was between businesses, with the remainder between businesses and individual
consumers. Forrester Research (Sanders and Tomkin, 2000) predicts that online sales in the
United States will reach $3.2 trillion by 2004, and that global sales will reach $6.8 trillion by
2004.
C. How important is e-commerce to agricultural producers?
Mirroring the efforts of the microdata collected by the Department of Commerce, USDA
(through its Agricultural Resource Management Survey, or ARMS study) has gathered extensive
information on computer, Internet, and e-commerce participation. ARMS is the only survey to
offer a national perspective on the changes in financial condition of U.S. agriculture, and is the
primary vehicle of USDA for collecting data on a broad range of issues about agricultural
resource use and costs. Other surveys collect data on computer and Internet use but not on
resource availability or use (such as National Agricultural Statistics Service survey) or are not
nationally representative (such as Gallup’s poll on e-commerce in agriculture). Morehart and
Hopkins used the 1999 ARMS study to show that while only about four percent of all farms

25