Globalization & eCommerce: Quantifying Cross-border Integration

by Ben Lorica (last updated Nov/2011)

Thomas Friedman's runaway bestseller, The World is Flat, popularized the notion that we already live in an (economically) integrated world. Friedman relied on a series of anecdotes to support his thesis. Armed with some data, Pankaj Ghemawat reckons that in many areas the level of integration is partial at best and closer to zero in some cases.
But Friedman's assertions are simply the latest in a series of exaggerated visions that also include the "end of history" and the "convergence of tastes." .. In truth, the world is not nearly as connected as these writers would have us believe. Despite talk of a new, wired world where information, ideas, money, and people can move around the planet faster than ever before, just a fraction of what we consider globalization actually exists. The portrait that emerges from a hard look at the way companies, people, and states interact is a world that's only beginning to realize the potential of true global integration. And what these trend's backers won't tell you is that globalization's future is more fragile than you know.
One can quibble with Ghemawat's metrics but at least he relies on data to justify his claims. I reproduce his bar graph below and document sources for estimates of each metric1. Despite the low levels of integration exhibited in the metrics below, I'll close this note with a series of charts that capture why I remain bullish on cross-border eCommerce.

Internationalization Levels
[ Source: World 3.0, figure 2-1 ]

Financial Measures

Data from the World Bank was used to estimate cross-border ownership of government debt ( ~ 35%). Based on surveys, cross-border ownership of bank deposits is about 25%. A 2007 study estimated that foreign investors owned about 20% of stock market equity. Finally a 2008 study estimated that between 15-20% of Venture Capital funds were deployed in countries outside those of the investing funds' home countries.


Foreign students account for about 2% of all university students worlwide, while first generation immigrants are about 3% of the world's population. I suspect that if one includes all legal and authorized immigrants (commonly referred to as the foreign-born population), the number would be slightly higher. The foreign-born population in an immigrant-friendly country like the US is around 13%. In comparison the foreign born population in China and India are about 0.5% and 0.3% respectively (see my Immigration by the numbers page for details).


Foreign Direct Investment (FDI) refers to the purchase of physical assets (e.g., buildings, land) or operations (e.g., management, joint-venture, transfer of technology and expertise) in another country, but excludes investments in the purchase of shares. Over the past few years FDI has averaged only about 10% -- in other words 90% of all fixed investments is from domestic sources.

Ghemawat adjusts global exports to better reflect actual "trade intensity". The classic example used by economists is the iPod. It turns out that

China is just the final assembly platform for 400+ components from East Asia and elsewhere, and adds only a few dollars of value, 1% -2% of the retail selling price of $299. Most of the value added ... goes to American companies and workers, with Apple alone pocketing almost half ... Yet every iPod sold in the US is recorded as contributing $150 to the US trade deficit with China. And since virtually all the components are shipped ... several times across national borders, the total trade officially recorded around that sale is a multiple of the $150 number!
It may well be that deflating trade figures to account for value added instead of revenues might be difficult or controversial (the WTO is meager to make the switch to value added). But on the surface it seems more sensible to measure trade intensity in this way. Ghemawat estimates that exports account "for only 20% of all the value produced in the world (GDP)."


A 2010 Pew Survey esitmates that 21% of news coverage in the U.S. is devoted to international news. Ethan Zuckerman examined the top 50 online news sites in 30 countries and found that "almost every country reads all but 5% of its news from domestic sources".

Only about 1% of physical mail crosses national borders, and less than 2% of telephone calling minutes are spent on international calls. Using data from Cisco and TeleGeography Research, Ghemawat estimates that between 2006 and 2008 about 17-18% of Internet traffic "was routed across a national border". Given that recent studies have shown that Netflix accounted for about a third of "peak downstream traffic", I'm not that surprised that less than a fifth of Internet traffic moves across national borders. As I explain in the next section, what should grab the attention of Internet site owners is the growing importance of non-domestic users.

Cross-border eCommerce

A recent factoid I came across brought home the importance of non-U.S. users. Despite having only a few employees outside the US, as of Oct/2011 about half of the usage of the mobile social network Foursquare is outside the US!

As the Internet audience continues to grow globally, a greater share of users for US sites comes from other countries. In the chart below I took comScore's list of Top U.S. sites and estimated the non-US audience using one of Mary Meeker's slides. In particular note that eCommerce oriented sites (like Amazon, ebay and Apple) draw a lots of visitors and sales from outside the U.S. While many of these sales may not technically be cross-border in nature, these are still instances of U.S. retailers expanding to accomodate the faster growing overseas market.

  • Amazon: For the nine months ending 9/30/2011, 45% of net sales came from outside North America.

  • ebay: For the third quarter of 2011, 62% of gross revenue (excluding vehicles) came from outside the U.S.

  • Apple: For the fourth quarter of 2011, 65% of gross revenue came from outside the Americas.
  • comScore Top U.S. Properties: Sept/2011
    [ Source: comScore ]

    An earlier comScore report provides detailed estimates of the geographic location of users for a list of large retail and social networking sites. In June/2011 Amazon and ebay drew 65% of their visitors from outside North America, while Apple drew an even higher share (68%). In comparison HP and Walmart drew 55% and 17% of their users from outside North America.

    Regional Audience Composition Analysis of Select Retail & Auction Sites: June/2011
    [ Source: comScore ]

    In the past many e-tailers found cross-border online daunting. Many e-tailers didn't offer their products to non-domestic users, or handled such transactions on a case-by-case basis. I've noticed a few trends that make venturing into cross-border eCommerce more doable:

  • Growth in Non-domestic Audience: As documented above, even services (like Foursquare) that aren't targeting non-domestic users are gaining traction outside their home countries.

  • Machine Translation: While localization is preferrable, nowawadays users interested in purchasing from foreign language sites have tools like Google Translate available.

  • Search: For over a year Google has allowed users to include "translated pages" in search results. At some point, I imagine product search tools will (routinely?) have cross-border content appear in their organic results.

  • (Mobile) Payments: There is so much innovation in the payments space, both from startups and established companies like Paypal and Visa/Mastercard/AmEx. For online retailers this translates to more users accustomed and qualified to use many of the standard payment options.

  • International Trade Organizations: For years the EU2 has worked to increase cross-border eCommerce across its member states. Recent EU studies (see here and here for recent reports) have shown that consumers benefit by saving money and by having access to products not available in their home country. Currently most online purchases within the EU are done domestically. In 2010, 57% of all Internet users in the EU purchased at least once from an e-tailer. And while 51% purchased from a domestic e-tailer, only 12% conducted a cross-border online purchase within the EU (7% purchased cross-border from a non-EU e-tailer). Put another way, 12% means that in 2010, 1 in 8 of all EU Internet users purchased from an e-tailer based in another EU member country.

    Online Consumers in the EU27
    Units = Percentage Share of Total EU27 Internet Users
    [ Source: Consumers at home in the single market, figure 4 ]

    Along with the EU's ongoing initiatives to create a single (online) market, there is a noticeable change in consumers attitudes within its member countries. In 2010, 47% of those surveyed expressed NO interest in cross-border online purchases, down from 57% in 2006.

  • Shipping and Logistics: For physical goods the primary obstacle remains shipping. Online purchasers have gotten used to expedited shipping and online tools for tracking their packages. Traditional shippers like UPS/FedEx and DHL make sense for higher value items ("overseas" shipping runs around $20 per pound or more). In addition actual costs frequently include additional duties (customs tax, VAT) that aren't communicated properly to consumers. (In the shipping industry this is referred to as the Total Landed Cost). A startup called Yakit (note: I'm an advisor to Yakit) lowers the cost of (cross-border) shipping, provides online package tracking, and handles the necessary calculations to arrive at total landed costs. For each package Yakit uses routing and optimization algorithms to discover the best combinations of shippers and routes. For e-tailers a web service like Yakit can be esaily integrated and quickly made available to users, as another low cost, domestic & cross-border, shipping option.

    There might be instances when your service model might not be replicable due to shipping. E-tailers like Zappos have struggled to maintain their high-touch service model, primarily due to vagaries of cross-border shipping.

    Related resources:
  • Mobile Cellular Subscriptions & Smartphone Usage by Country

  • U.S. Legal & Unauthorized Immigration by the Numbers

  • (1) Choice of metrics was inspired by the 2007 Foreign Policy Globalization Survey.

    (2) While the EU is clearly in trouble at the moment, their work in cross-border eCommerce continues.  
    Back to Resources page.

    NOTE: Reproduction & reuse allowed under Creative Commons Attribution.    Creative Commons Attribution