Body
In the 21st century, international digital flows have taken the lead role in promoting globalization. The calculations that follow make the strong policy argument that growth in digital trade compensated for the lethargic growth in merchandise trade and business services in recent years, and that governments should be cautious in regulating digital traffic. Yet some governments try to censor or tax digital traffic, and many governments seek to compel localization of digital platforms. Forced localization of digital platforms (servers and analysts) is a bad idea, censorship or taxation of digital traffic is worse, and privacy rules should be narrowly drawn.
Digital flows include everything from entertaining movies and music to instruction in mathematics and engineering to performance data on refineries, factories, and offices. Examples include Netflix movies, Spotify music, and Massive Open Online Courses. More obscure but equally important are the enormous daily data flows between the global subsidiaries of multinational corporations—ExxonMobil, GE, Caterpillar, Apple, Microsoft, JPMorgan Chase, and many others.
For assorted reasons, the growth of conventional trade and foreign direct investment (FDI) has spluttered since 2011. Liberalization of national barriers stalled and populist politicians from President Donald Trump to Prime Minister Narendra Modi revived the import substitution model of yesteryear. The multilateral Doha Development Round is dead, and new regional and bilateral trade agreements sometimes exempt large swaths of trade and investment. For example, the Korea-China Free Trade agreement exempts agriculture and autos; the US-Mexico-Canada Agreement (USMCA) made no headway in liberalizing services, only a token improvement in agriculture, and took a couple of backward steps on North American auto trade. A bright spot amidst the current wave of anti-globalization policies is the tremendous growth of global digital flows. It is therefore worth asking whether digital flows can compensate for lethargic conventional trade and FDI flows in terms of bolstering world economic growth.
The answer to this question is a qualified "yes." In broad macroeconomic terms, it appears that the exponential expansion of digital flows since 2005 has replaced the drag to world GDP resulting from sluggish conventional trade and FDI growth. To be sure, the digital boom is no excuse for failed trade negotiations. Yet the world is better off thanks to dramatic gains in digital communication.
How do we know? In 2016, the McKinsey Global Institute (MGI) constructed an econometric model to assess the contribution of several types of flows to world GDP growth. Merchandise trade and FDI flows were normalized by GDP, while digital flows were normalized by population in the model. MGI estimated "long-run elasticity coefficients" for the GDP impact of growth in normalized merchandise trade, FDI, and digital flows. An elasticity coefficient measures the degree of association, in terms of percentage changes, between an independent variable, such as world merchandise trade flows, and a dependent variable, such as the world GDP. To illustrate, the long-run elasticity coefficient for normalized merchandise trade is 0.05, which suggests that when merchandise trade grows 10 percent faster than world GDP, the boost to world GDP in the long run is 0.5 percent. Importantly, the boost lasts indefinitely, meaning a big long-term payoff from expanded international flows of trade, FDI, or data.
Table 1 shows MGI's estimated coefficients and their statistical significance. The small p-values indicate that the coefficients are statistically different from zero.
Table 1 GDP impact of global flows, using normalized flow values | ||
---|---|---|
Flow variables | Long term elasticities | P-valuesa |
Goods trade | 0.05 | 0.0129 |
FDI | 0.04 | 0.0000 |
Data | 0.02 | 0.0036 |
a. p-value is the probability of reaching an elasticity value at least as different from 0 as the estimated value by pure random sampling variation, assuming that the true elasticity is 0. When the p-value is quite small, as for these estimates, it is highly unlikely that the true elasticity is 0. | ||
Source: McKinsey & Company (2016). |
Making bold assumptions, the analysis here used the MGI coefficients to calculate the contribution of international flows to world GDP at 5-year intervals for the period 1995 to 2015, with projections to 2020. The key assumption is that elasticity coefficients are constant over the entire period, even though the MGI estimates were based on data just for the period 1995 to 2013. Another assumption is that published Cisco data on global internet traffic closely parallels the proprietary cross-border bandwidth data used by MGI to estimate the digital flow elasticity coefficient.
Table 2 presents the calculations. The figures for merchandise trade and FDI flows in the top panel are from United Nations Conference on Trade and Development (UNCTAD) sources, the same as used by MGI to estimate elasticity coefficients. The digital flow figures are from Cisco because the proprietary TeleGeography data used by MGI was not made available for this analysis.
Table 2 Contributions of merchandise trade, FDI, and digital flows to world GDP | |||||||
---|---|---|---|---|---|---|---|
Flows, billion current US dollars and petabytes per month | |||||||
1990 | 1995 | 2000 | 2005 | 2010 | 2015 | 2020c | |
Merchandise trade, imports + exports | 7,105 | 10,411 | 13,107 | 21,277 | 30,728 | 33,235 | 40,436 |
FDI, inflow + outflow | 449 | 698 | 2,522 | 1,782 | 2,746 | 3,543 | 3,543 |
Digital flowa | 2,962 | 20,151 | 72,521 | 228,411 | |||
Normalized 5-year growth rates, percent changeb | |||||||
Merchandise trade, imports + exports | 8.3 | 16.5 | 14.6 | 4.1 | –4.4 | –2.4 | |
FDI, inflow + outflow | 15.0 | 234.3 | –50.1 | 11.1 | 14.0 | –19.8 | |
Digital flow | 539.6 | 239.2 | 198.3 | ||||
Effect on world GDP, percent | |||||||
Merchandise trade, imports + exports | 0.4 | 0.8 | 0.7 | 0.2 | –0.2 | –0.1 | |
FDI, inflow + outflow | 0.6 | 9.4 | –2.0 | 0.4 | 0.6 | –0.8 | |
Digital flow | 10.8 | 4.8 | 4.0 | ||||
Total contribution of flows to world GDP | 1.0 | 10.2 | –1.3 | 11.4 | 5.1 | 3.1 | |
Contribution to world GDP, billion current US dollars | |||||||
Merchandise trade, imports + exports | 129 | 278 | 347 | 137 | –165 | –110 | |
FDI, inflow + outflow | 187 | 3,149 | –954 | 293 | 420 | –735 | |
Digital flow | 7,124 | 3,573 | 3,691 | ||||
Total contribution of flows to world GDP | 316 | 3,427 | –607 | 7,554 | 3,828 | 2,846 | |
World GDP, billion current US dollars | 22,985 | 31,102 | 33,601 | 47,602 | 66,010 | 74,696 | 93,080 |
Nominal 5-year growth rate of world GDP (percent) | 35.3 | 8.0 | 41.7 | 38.7 | 13.2 | 24.6 | |
World population, in millions | 5,331 | 5,751 | 6,145 | 6,542 | 6,958 | 7,383 | 7,795 |
a. Global internet traffic, measured in petabyte per month. | |||||||
b. Merchandise trade and FDI are normalized by dividing flows by world GDP; data flow is normalized by dividing flows by world population. | |||||||
c. Projected values. Estimates of world GDP and population in 2020 are from World Economic Outlook (October 2018) by the International Monetary Fund and 2018 Revision of World Urbanization Prospects, respectively. Estimate of global internet traffic in 2020 is from Cisco (2017). The projected annual compound growth rate of world GDP from 2016 to 2020 is 5.3 percent. With growing protectionism, it is assumed that merchandise trade grows at a compound rate of 4 percent annually. Global FDI flows decreased by 5.7 percent in 2016 and again by 14.4 percent in 2017. The calculations assumes that 2020 will regain the 2015 level. | |||||||
Sources: Merchandise trade, FDI, world GDP, and world population come from the United Nations Conference on Trade and Development; data on digital flows are from Cisco's Visual Networking Index. |
Growth rates of trade, FDI, and digital flows over 5-year intervals starting in 1990 are shown in the second panel of table 2. Multiplying growth rates by the long-term elasticity coefficients estimated by MGI yields each component's percentage contribution to global GDP, shown in the third panel.
The fourth panel expresses the contribution of each flow in dollar values. The fifth and final panel gives world GDP in current dollars, the 5-year nominal growth rate of world GDP, and world population.
Results presented in the third and fourth panels suggest that, since 2005, digital flows made by far the largest contribution to world GDP growth. In fact, the estimated digital contribution is so large that its magnitude can be seriously questioned. But even if the long-run elasticity coefficient is only half the MGI estimate, 0.01 rather than 0.02, the contribution of digital flows exceeds the combined contribution of merchandise trade and FDI flows. Moreover, since 2005, digital flows have contributed a significant share of world GDP growth.
It is bad enough that post-21st century liberalization of goods and services trade has been sporadic, and that foreign direct investment is being more strictly screened by many countries. It would be truly unfortunate if anti-globalization sentiment retarded the growth of digital traffic. Reflecting that simple conclusion, the United States has mounted an effort in the World Trade Organization to preserve the freedom of cross-border digital traffic. For the moment, China has very different views on the issue. An agreement on rule making regarding digital trade could be an outstanding outcome of the wide-ranging economic debate between Presidents Donald Trump and Xi Jinping.
References
Cisco. 2017. Cisco Visual Networking Index: Forecast and Methodology, 2016–2021 (accessed on November 28, 2018).
McKinsey Global Institute. 2016. Digital Globalization: The New Era of Global Flows. McKinsey & Company.
United Nations Department of Economic and Social Affairs. 2018. 2018 Revision of World Urbanization Prospects (accessed on November 16, 2018).