Market Lessons From The Internet

Lessons About Markets from the

Internet

Glenn Ellison and Sara Fisher Ellison

Many of us have grown used to, tired of, and finally downright skeptical of

claims of the transformative powers of the Internet. It was to usher in

the New Economy, but we seem mostly to have the Old. It would

transform retail, but Toys 'R' Us has outlasted EToys. Frictionless commerce would

be the norm, but plenty of friction still exists.

The Internet was also claimed to require a whole new economics with all new

laws. While this, too, was very far from the truth--existing theories have mostly

done quite well--the Internet has had a substantial effect on economic thought. In

this paper, we discuss some ways in which the Internet has affected how economists

think about markets.

That the Internet has affected economic thought is perhaps surprising. Some

economic events, like the Industrial Revolution and the inflation of the mid-1970s

and early 1980s, have had a major impact on the way economics is done. But other

major events, like the development of the interstate highway system and the 1987

stock market crash, seem not to have made much of a difference. For this reason,

we begin in the next section with a brief discussion of what about the Internet

seems to be making it important to economic research.

We then turn to our main task and discuss two specific topics that received a

lot of attention in the popular press and that have been viewed differently by

y Glenn Ellison is Professor of Economics and Sara Fisher Ellison is Senior Lecturer in

Economics, both at the Massachusetts Institute of Technology, Cambridge, Massachusetts. The

paper was written while the authors were visiting the Institute for Advanced Study, Princeton,

New Jersey, during 2003-2004, the former as a Member and the latter as the Richard B. Fisher

Member. Glenn Ellison is also a Research Associate, National Bureau of Economic Research,

Cambridge, Massachusetts.

Journal of Economic Perspectives--Volume 19, Number 2--Spring 2005--Pages 139-158

economists since the Internet boom: how the Internet was to create online

marketplaces and whether the Internet would usher in "frictionless commerce." In

each case we describe how economists would have thought about these topics circa

1999, what actually happened and what lessons about the operation of markets have

been learned.

The Internet and Economic Research

The Internet has affected economic research along a number of dimensions.

Whenever a new phenomenon arises, especially one that receives so much attention

in the popular press, it will attract study by economists. We think that this fact alone

cannot account for the number of economists who have worked on the Internet

and the impact this research has had. In this section, we discuss briefly why the

Internet has helped economists develop new insights into the operation of markets.

First, the Internet has provided researchers with the opportunity to study how

markets function in novel and extreme circumstances. A vivid example is that with

the growth of the Internet, we suddenly have markets with essentially no search

costs. Price-search websites like Dealtime, Pricewatch and Shopper allow customers

wanting to purchase a particular book, a particular name-brand digital camera or

even a flower arrangement to receive a sorted and annotated list of price quotes

from dozens of retailers.

Why is this useful for research? Economists have long known that markets are

affected by consumer search costs, and empirical studies have tried to compare

markets with higher and lower search costs. It is difficult, however, to find examples

of two very similar markets where only search costs vary significantly. There are also

some striking theoretical insights--like Diamond's (1971) argument that there can

be a great discontinuity between markets with small incremental search costs and

those with no incremental search costs--that one cannot test by looking at markets

with low, medium and high search costs. Observing what happens when a

parameter--here the incremental search costs--takes on an extreme value can be

very informative.

Second, the Internet has provided a number of interesting natural experiments

for researchers to exploit. For example, in traditional retail markets, "identical"

products offered by different retailers were always differentiated by location,

customer service and idiosyncratic consumer preferences for a given store. With the

development of e-retail, products suddenly became differentiated by just the last

two of those. Comparing traditional and online retailers can thus

...