Why the World Isn’t Flat

Globalization has bound people, countries, and markets closer than ever, rendering national borders relics of a bygone era -- or so we're told. But a close look at the data reveals a world that's just a fraction as integrated as the one we thought we knew. In fact, more than 90 percent of all phone calls, Web traffic, and investment is local. What’s more, even this small level of globalization could still slip away.


Ideas will spread faster, leaping borders. Poor countries will have immediate access to information that was once restricted to the industrial world and traveled only slowly, if at all, beyond it. Entire electorates will learn things that once only a few bureaucrats knew. Small companies will offer services that previously only giants could provide. In all these ways, the communications revolution is profoundly democratic and liberating, leveling the imbalance between large and small, rich and poor.” The global vision that Frances Cairncross predicted in her Death of Distance appears to be upon us. We seem to live in a world that is no longer a collection of isolated, “local” nations, effectively separated by high tariff walls, poor communications networks, and mutual suspicion. It’s a world that, if you believe the most prominent proponents of globalization, is increasingly wired, informed, and, well, “flat.”

It’s an attractive idea. And if publishing trends are any indication, globalization is more than just a powerful economic and political transformation; it’s a booming cottage industry. According to the U.S. Library of Congress’s catalog, in the 1990s, about 500 books were published on globalization. Between 2000 and 2004, there were more than 4,000. In fact, between the mid-1990s and 2003, the rate of increase in globalization-related titles more than doubled every 18 months.

Amid all this clutter, several books on the subject have managed to attract significant attention. During a recent TV interview, the first question I was asked — quite earnestly — was why I still thought the world was round. The interviewer was referring of course to the thesis of New York Times columnist Thomas L. Friedman’s bestselling book The World Is Flat. Friedman asserts that 10 forces — most of which enable connectivity and collaboration at a distance — are “flattening” the Earth and leveling a playing field of global competitiveness, the likes of which the world has never before seen.

It sounds compelling enough. 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.” Some writers in this vein view globalization as a good thing — an escape from the ancient tribal rifts that have divided humans, or an opportunity to sell the same thing to everyone on Earth. Others lament its cancerous spread, a process at the end of which everyone will be eating the same fast food. Their arguments are mostly characterized by emotional rather than cerebral appeals, a reliance on prophecy, semiotic arousal (that is, treating everything as a sign), a focus on technology as the driver of change, an emphasis on education that creates “new” people, and perhaps above all, a clamor for attention. But they all have one thing in common: They’re wrong.

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.


The few cities that dominate international financial activity — Frankfurt, Hong Kong, London, New York — are at the height of modern global integration; which is to say, they are all relatively well connected with one another. But when you examine the numbers, the picture is one of extreme connectivity at the local level, not a flat world. What do such statistics reveal? Most types of economic activity that could be conducted either within or across borders turn out to still be quite domestically concentrated.

One favorite mantra from globalization champions is how “investment knows no boundaries.” But how much of all the capital being invested around the world is conducted by companies outside of their home countries? The fact is, the total amount of the world’s capital formation that is generated from foreign direct investment (FDI) has been less than 10 percent for the last three years for which data are available (2003–05). In other words, more than 90 percent of the fixed investment around the world is still domestic. And though merger waves can push the ratio higher, it has never reached 20 percent. In a thoroughly globalized environment, one would expect this number to be much higher — about 90 percent, by my calculation. And FDI isn’t an odd or unrepresentative example.

The levels of internationalization associated with cross-border migration, telephone calls, management research and education, private charitable giving, patenting, stock investment, and trade, as a fraction of gross domestic product (GDP), all stand much closer to 10 percent than 100 percent. The biggest exception in absolute terms — the trade-to-GDP ratio shown at the bottom of the chart — recedes most of the way back down toward 20 percent if you adjust for certain kinds of double-counting. So if someone asked me to guess the internationalization level of some activity about which I had no particular information, I would guess it to be much closer to 10 percent — the average for the nine categories of data in the chart — than to 100 percent. I call this the “10 Percent Presumption.”

More broadly, these and other data on cross-border integration suggest a semiglobalized world, in which neither the bridges nor the barriers between countries can be ignored. From this perspective, the most astonishing aspect of various writings on globalization is the extent of exaggeration involved. In short, the levels of internationalization in the world today are roughly an order of magnitude lower than those implied by globalization proponents.


If you buy into the more extreme views of the globalization triumphalists, you would expect to see a world where national borders are irrelevant, and where citizens increasingly view themselves as members of ever broader political entities. True, communications technologies have improved dramatically during the past 100 years. The cost of a three-minute telephone call from New York to London fell from $350 in 1930 to about 40 cents in 1999, and it is now approaching zero for voice-over-Internet telephony. And the Internet itself is just one of many newer forms of connectivity that have progressed several times faster than plain old telephone service. This pace of improvement has inspired excited proclamations about the pace of global integration. But it’s a huge leap to go from predicting such changes to asserting that declining communication costs will obliterate the effects of distance. Although the barriers at borders have declined significantly, they haven’t disappeared.

To see why, consider the Indian software industry — a favorite of Friedman and others. Friedman cites Nandan Nilekani, the CEO of the second-largest such firm, Infosys, as his muse for the notion of a flat world. But what Nilekani has pointed out privately is that while Indian software programmers can now serve the United States from India, access is assured, in part, by U.S. capital being invested — quite literally — in that outcome. In other words, the success of the Indian IT industry is not exempt from political and geographic constraints. The country of origin matters — even for capital, which is often considered stateless.

Or consider the largest Indian software firm, Tata Consultancy Services (TCS). Friedman has written at least two columns in the New York Times on TCS’s Latin American operations: “[I]n today’s world, having an Indian company led by a Hungarian-Uruguayan servicing American banks with Montevidean engineers managed by Indian technologists who have learned to eat Uruguayan veggie is just the new normal,” Friedman writes. Perhaps. But the real question is why the company established those operations in the first place. Having worked as a strategy advisor to TCS since 2000, I can testify that reasons related to the tyranny of time zones, languages, and the need for proximity to clients’ local operations loomed large in that decision. This is a far cry from globalization proponents’ oft-cited world in which geography, language, and distance don’t matter.

Trade flows certainly bear that theory out. Consider Canadian-U.S. trade, the largest bilateral relationship of its kind in the world. In 1988, before the North American Free Trade Agreement (NAFTA) took effect, merchandise trade levels between Canadian provinces — that is, within the country — were estimated to be 20 times as large as their trade with similarly sized and similarly distant U.S. states. In other words, there was a built-in “home bias.” Although NAFTA helped reduce this ratio of domestic to international trade — the home bias — to 10 to 1 by the mid-1990s, it still exceeds 5 to 1 today. And these ratios are just for merchandise; for services, the ratio is still several times larger. Clearly, the borders in our seemingly “borderless world” still matter to most people.

Geographical boundaries are so pervasive, they even extend to cyberspace. If there were one realm in which borders should be rendered meaningless and the globalization proponents should be correct in their overly optimistic models, it should be the Internet. Yet Web traffic within countries and regions has increased far faster than traffic between them. Just as in the real world, Internet links decay with distance. People across the world may be getting more connected, but they aren’t connecting with each other. The average South Korean Web user may be spending several hours a day online — connected to the rest of the world in theory — but he is probably chatting with friends across town and e-mailing family across the country rather than meeting a fellow surfer in Los Angeles. We’re more wired, but no more “global.”

Just look at Google, which boasts of supporting more than 100 languages and, partly as a result, has recently been rated the most globalized Web site. But Google’s operation in Russia (cofounder Sergey Brin’s native country) reaches only 28 percent of the market there, versus 64 percent for the Russian market leader in search services, Yandex, and 53 percent for Rambler.

Indeed, these two local competitors account for 91 percent of the Russian market for online ads linked to Web searches. What has stymied Google’s expansion into the Russian market? The biggest reason is the difficulty of designing a search engine to handle the linguistic complexities of the Russian language. In addition, these local competitors are more in tune with the Russian market, for example, developing payment methods through traditional banks to compensate for the dearth of credit cards. And, though Google has doubled its reach since 2003, it’s had to set up a Moscow office in Russia and hire Russian software engineers, underlining the continued importance of physical location. Even now, borders between countries define — and constrain — our movements more than globalization breaks them down.


If globalization is an inadequate term for the current state of integration, there’s an obvious rejoinder: Even if the world isn’t quite flat today, it will be tomorrow. To respond, we have to look at trends, rather than levels of integration at one point in time. The results are telling. Along a few dimensions, integration reached its all-time high many years ago. For example, rough calculations suggest that the number of long-term international migrants amounted to 3 percent of the world’s population in 1900 — the high-water mark of an earlier era of migration — versus 2.9 percent in 2005.

Along other dimensions, it’s true that new records are being set. But this growth has happened only relatively recently, and only after long periods of stagnation and reversal. For example, FDI stocks divided by GDP peaked before World War I and didn’t return to that level until the 1990s. Several economists have argued that the most remarkable development over the long term was the declining level of internationalization between the two World Wars. And despite the records being set, the current level of trade intensity falls far short of completeness, as the Canadian-U.S. trade data suggest. In fact, when trade economists look at these figures, they are amazed not at how much trade there is, but how little.

It’s also useful to examine the considerable momentum that globalization proponents attribute to the constellation of policy changes that led many countries — particularly China, India, and the former Soviet Union — to engage more extensively with the international economy. One of the better-researched descriptions of these policy changes and their implications is provided by economists Jeffrey Sachs and Andrew Warner:

“The years between 1970 and 1995, and especially the last decade, have witnessed the most remarkable institutional harmonization and economic integration among nations in world history. While economic integration was increasing throughout the 1970s and 1980s, the extent of integration has come sharply into focus only since the collapse of communism in 1989. In 1995, one dominant global economic system is emerging.”

Yes, such policy openings are important. But to paint them as a sea change is inaccurate at best. Remember the 10 Percent Presumption, and that integration is only beginning. The policies that we fickle humans enact are surprisingly reversible. Thus, Francis Fukuyama’s The End of History, in which liberal democracy and technologically driven capitalism were supposed to have triumphed over other ideologies, seems quite quaint today. In the wake of Sept. 11, 2001, Samuel Huntington’s Clash of Civilizations looks at least a bit more prescient. But even if you stay on the economic plane, as Sachs and Warner mostly do, you quickly see counterevidence to the supposed decisiveness of policy openings. The so-called Washington Consensus around market-friendly policies ran up against the 1997 Asian currency crisis and has since frayed substantially — for example, in the swing toward neopopulism across much of Latin America. In terms of economic outcomes, the number of countries — in Latin America, coastal Africa, and the former Soviet Union — that have dropped out of the “convergence club” (defined in terms of narrowing productivity and structural gaps vis-à-vis the advanced industrialized countries) is at least as impressive as the number of countries that have joined the club. At a multilateral level, the suspension of the Doha round of trade talks in the summer of 2006 — prompting The Economist to run a cover titled “The Future of Globalization” and depicting a beached wreck — is no promising omen. In addition, the recent wave of cross-border mergers and acquisitions seems to be encountering more protectionism, in a broader range of countries, than did the previous wave in the late 1990s.

Of course, given that sentiments in these respects have shifted in the past 10 years or so, there is a fair chance that they may shift yet again in the next decade. The point is, it’s not only possible to turn back the clock on globalization-friendly policies, it’s relatively easy to imagine it happening. Specifically, we have to entertain the possibility that deep international economic integration may be inherently incompatible with national sovereignty — especially given the tendency of voters in many countries, including advanced ones, to support more protectionism, rather than less. As Jeff Immelt, CEO of GE, put it in late 2006, “If you put globalization to a popular vote in the U.S., it would lose.” And even if cross-border integration continues on its upward path, the road from here to there is unlikely to be either smooth or straight. There will be shocks and cycles, in all likelihood, and maybe even another period of stagnation or reversal that will endure for decades. It wouldn’t be unprecedented.

The champions of globalization are describing a world that doesn’t exist. It’s a fine strategy to sell books and even describe a potential environment that may someday exist. Because such episodes of mass delusion tend to be relatively short-lived even when they do achieve broad currency, one might simply be tempted to wait this one out as well. But the stakes are far too high for that. Governments that buy into the flat world are likely to pay too much attention to the “golden straitjacket” that Friedman emphasized in his earlier book, The Lexus and the Olive Tree, which is supposed to ensure that economics matters more and more and politics less and less. Buying into this version of an integrated world — or worse, using it as a basis for policymaking — is not only unproductive. It is dangerous.

Photo Credit: Dan Kitwood/Getty Images

Pankaj Ghemawat is Anselmo Rubiralta professor of global strategy at IESE Business School and Jaime and Josefina Chua Tiampo professor of business administration at Harvard Business School. He is the author of Redefining Global Strategy (Boston: Harvard Business School Press, 2007).