Phase transition in the globalization of trade

Globalization processes interweave economic structures at a worldwide scale, trade playing a central role as one of the elemental channels of interaction among countries. Despite the significance of such phenomena, measuring economic globalization still remains an open problem. More quantitative treatments could improve the understanding of globalization at the same time that help a formal basis for comparative economic history. In this letter, we investigate the time evolution of the statistical properties of bilateral trade imbalances between countries in the trade system. We measure their cumulative probability distribution at different moments in time to discover a sudden transition circa 1960 from a regime where the distribution was always represented by a steady characteristic function to a new state where the distribution dilates as time goes on. This suggests that the rule that was governing the statistical behavior of bilateral trade imbalances until the 60’s abruptly changed to a new form persistent in the last decades. In the new regime, the figures for the different years collapse into a universal master curve when rescaled by the corresponding global gross domestic product value. This coupling points to an increased interdependence of world economies and its onset corresponds in time with the starting of the last globalization wave. Phase transition in the globalization of trade 2 1. Globalization and trade At this point, the beginning of the third millennium in the Gregorian calendar, it seems customarily assumed that several epochs of globalization have occurred to shape the world as we perceive it today. In these periods, a complex series of changes closely intertwined develop into an increasing interdependence and interaction between people and human organizations in disparate locations of the world. These changes, when structural, are in a great part of economic nature, markets becoming natural mediators of globalization forces [1]. In terms of trade [2], there is still controversy whether the last globalization waves, the first roughly identified from 1870 till the beginning of World War I and the second from 1960 to the present, are more different than similar [3] –there is no a complete agreement either whether a third middle wave has occurred [4]. The two waves correspond to processes of decolonization and falls of technical barriers with the corresponding downloads of costs and time expenditures. The first wave was triggered by the Industrial Revolution, with steam power encouraging the expanding of railroad networks and oceanic routes and the telegraph connecting the two sides of the Atlantic. The second came intimately related to the Information Technologies Revolution, communications costs dramatically dropping at the same time that information management capabilities explode. In gross terms, the first globalization burst had to do with lower costs in transportation of materials and goods, while the second deals with exchange of information and ideas. Two different natures of changes that may well produce different impacts. This discussion brings directly to the problem of how to measure globalization. Quantitative approaches could enlighten but have been timid to this moment, most works adopting analytical methodologies. New quantitative ways of studying globalization processes are required to exploit the wealth of information in historical data [5, 6]. One simple step forward consists in complementing the study of aggregated or global values with the statistical analysis of how they are distributed, from where we can obtain not only more detailed but also new information. For instance, a prominent figure when measuring globalization of trade plots the evolution in time of total international trade as a percentage of the global product, computed as the sum of all national gross domestic products (GDPs). As it has been reported by several authors [3, 4, 7], an U-shape pattern emerges, with rise of trade in the two eras of globalization and a major reversal in between, and has been claimed as a common trait which recurs in many other empirical analysis, such as in the plot of global capital flows to GDP ratio or in the correlation between savings and investment [3, 8]. 2. Analyzing the evolution of bilateral trade imbalances Still, further crucial information can be obtained from trade data if, moreover, one draws the evolution in time of the distribution of bilateral trade flows. Here we show that until Phase transition in the globalization of trade 3 188