The gravity law and Internet

Does the Internet defy Gravity model

International trade

the Law of Gravity? Digital Goods

Analysis of data on Internet activities by US

consumers on non-US websites

Do the trade costs fully account for the effects of

distance on trade?

Methodology

–Introduction– Is physical distance less important when

transportation, time and distribution costs are near

zero?

Relation between trade and digital goods

Methodology Financial transaction: Software, financial information…

Free: Online games, music, pornography….

2 type of

digital Aggregation of digital goods in:

products: Taste-dependent and Non-Taste dependent

Financial

transaction

and Free The analysis will deal with the effect of

distance on taste and non-taste goods

Consumption Theoretically, the Internet reduces the

realized over importance of distance

Data on

internet Internet reduces communication, search

2654 US Internet and transportation costs near zero

users (Shapiro and Varian, 1999) increases trade

(households) from (Freund and Weinhold, 2004) so distance

Dec.27 1999 to should not matter for goods that are

March 31 2000 distributed digitally

Methodology – Choice model for web surfing

− = − + +

•

is the market share of country c in the website category s. It also includes the

• household’s utility of browsing the website. Since US users who visit foreign websites

have always the option of visiting a US website we assume that this choice is “the

outside” choice (c=0)

and are the number of websites from country c and from US in service category s

•

is a vector of observed country and website characteristic, including the distance

•

between the country c and the US

measures how country and website characteristics affect the marginal utility of

• visiting a category s website in country c relative to the marginal utility of visiting a

category s in the US.

Methodology – Choice model for web surfing II

We can write the market share of country c in the website category s as the ratio of

• number of visits to the country c websites in category s and the total number of visits to

0

= =

category s websites in the world: and

Because and do not vary by country within a service category s, they can be