NBER Reporter 2015 Number 1: Research Summary
New Perspectives on the First Wave of Globalization
Christopher M. Meissner
The first "Great Wave of Globalization," during the late 19th and early 20th centuries, witnessed a historically unprecedented rise in spatial economic integration. Between 1850 and 1913, transportation costs plummeted, information flows accelerated, tariffs fell, trade treaties such as free trade agreements with unconditional most-favored-nation clauses and treaty ports proliferated, and empires expanded. In addition, a set of global financial intermediaries flourished, migrants flowed to previously unsettled regions in unprecedented numbers, and economic and political stability was largely the norm.
Unsurprisingly, many commodity prices converged and the export share of total production increased dramatically, dou-bling or tripling in many small, open economies between 1850 and 1914. In addition, new markets opened up to international trade and previously unavailable varieties of goods became accessible. Patterns of specialization and production processes were transformed. All of these forces significantly affected the living standards of those participating. Modern economic growth, meaning sustained rises in the standard of living, became the new norm. Social and political transformations also accompanied this episode of great integration.
My research, in collaboration with Michael Huberman, David Jacks, Dan Liu, Dennis Novy, and Kim Oosterlinck, seeks to shed further light on the causes and consequences of the international trade boom between 1870 and 1914. How much did trade costs actually fall in this period of globalization? What fraction of the rise in trade flows can be explained by the decline in trade costs? What was the relative contribution of geography, policy, and technology in explaining the first wave of globalization? What impact did trade costs and trade integration have on welfare and then on institutional and policy outcomes such as labor standards or the level of democracy?
To help answer these questions we have digitized and compiled a large amount of historical data from national data sources covering bilateral trade flows, GDP, gross production, and many other geographic and policy variables. Comprehensive bilateral trade data were recorded in the 19th century by national authorities and colonial powers, since a large fraction of government revenue came from taxes on international trade. Moreover, as I will detail below, not only can we make use of aggregate bilateral trade data, but economic historians are now able to rely on bilateral, product-level trade flows which provide greater granularity and deeper insight into the mechanics of the first wave of globalization. While research is only just beginning as regards the latter, these data will allow us to gain a greater understanding of forces driving globalization and its connections to economic growth, both in industrial leaders and their followers. Such questions potentially have great relevance today both to developing countries and to leading countries that are being strongly affected by globalization. This brief survey discusses what emerges when we combine these data sets and analyze them with the help of trade theory and modern empirical methods.
Trade Costs and the Determinants of GlobalizationTrade costs can be broadly defined as the resource costs of shipping and trading commodities across international borders. When such trade is costly, foreign demand for domestic goods is assumed to be lower than it would be in the absence of such costs. What role did these costs play in explaining the growth of international trade and the types of goods traded during the first globalization? Especially important is understanding how much trade costs mattered relative to other determinants, such as economic growth and comparative advantage.
Previous work in economic history has emphasized the rapid decline in transportation costs and the fall in tariffs.1 However, a number of other trade costs mattered over this period, and not all of them followed the same path as real transportation costs. My collaborators and I have built up a number of historical datasets that allow us to track the evolution and impact over time of trade costs other than transportation and tariffs. For instance, my research with J. Ernesto López-Córdova covering 28 countries between 1870 and 1910 uses a gravity model of bilateral trade flows.2 We find that when two nations adopted the gold standard, trade was higher by 15 percent, on average, relative to non-adopters. Monetary unions, political alliances, lan-guage, and trade treaties also affected the direction of trade.
Many other factors determine trade costs, and very often these are unobservable or impossible to measure in any conven-tional sense. In this case, a structural approach to trade costs can be taken. Jacks, Novy, and I measure trade costs as the scaled difference between domestic and international trade flows.3 The structural approach provides a measure of trade costs in terms of a tariff equivalent, and it is often referred to as the Head-Ries measure. This measure is quite general and is consistent with nearly all leading theoretical models of trade. Our data for the U.S., U.K., and France and their major trading partners between 1870 and 1913 show that trade costs fell at a rate of about 0.3 percent per year, which is significantly slower than the decline in average maritime freight rates of 2 percent per year. Our explanation for this is, first, that our all-encompassing trade-cost measure captures many other frictions which were slower to decline than freight rates. These include border frictions, legal and cultural barriers to trade, and significant rises in tariffs during the period. Another crucial aspect to highlight is that interna-tional integration can only rise when the relative costs of engaging in international trade fall. During this period, the railroad and many other domestic infrastructure projects promoted internal as much as international integration.
We also studied the effects of the decline in overall trade costs between specific pairs of countries. We analyzed 130 unique country pairs covering approximately 70 percent of global exports and 68 percent of world GDP in the period 1870 to 2000.4 Using our methodology, we show how to decompose the growth in trade between two factors: trade-cost changes and economic growth. We find some differences across major countries like the U.S., France, and the U.K. and between different periods. For instance, while trade-cost declines explain about 60 percent of trade growth between 1870 and 1913, they only explain 30 percent in the period 19502000. In each period, the remainder of trade growth appears to be coming from economic growth. Thus, while we experienced roughly equal increases in global trade flows during the two waves of globalization, the drivers of growth in overall trade in 18701913 and in 19502000 appear to have been quite different.
The Margins of Trade and the First Wave of Globalization
Recently my collaborators and I have begun to use disaggregated historical trade statistics to understand better the un-derlying dynamics of globalization and its impact on local economies. Using newly digitized bilateral, product-level trade data for Belgium, a typical industrializing, small, open economy between 1870 and 1913, we illustrate that globalization in the 19th century had a very important "extensive margin." 5 While the existing literature on pre-1914 globalization has emphasized a "great specialization," this characterization fails to take into account that a significant fraction of the growth of trade was due to the export of new goods and the opening up of new markets. Significant amounts of the observed trade flows were also in fact already intra-industry. This observation leads us to believe that then, as now, firm-level heterogeneity and trade costs mattered.
We first decompose the growth of Belgian manufacturing exports into an intensive margin (old products and old countries) and an extensive margin (new goods and new countries). Between 1880 and 1910 about 58 percent of the growth in the value of exports was accounted for by the appearance and growth of exports of new goods. In this case, 45 percent of the growth is attributable to the intensive margin or products that were already being shipped in 1870. A small set of exports was discontinued, acting to reduce trade by about 3 percent less than would otherwise have been the case.
We are also able to track the evolution and impact of a number of trade costs, some of which acted as "fixed" costs to exporting, and some of which acted as "variable" trade costs. We find evidence that diplomatic representation, colonial ties with other leading nations, and absence of a common language acted to alter the fixed costs of trade, implying that these factors helped generate - or limited, in the case of trade with colonies of the great powers - export success in new goods, such as tramways and other high quality/high value-added manufactures.
We also find other evidence consistent with the idea that firm-level heterogeneity was important in the first period of globalization. Gravity regressions by product or industry reveal a range of elasticities with respect to observable trade costs that depend on the type of good and industry. This is consistent with the predictions of modern models of trade with heterogeneous firms. A final finding is that, as fixed costs fell and presumably as new firms found it profitable to enter export markets, many industries experienced relatively slow productivity growth as low-productivity entrants were now able to survive. This was especially true in older, more-established industries. Although Belgium experienced a rise in productivity, overall productivity growth between 1870 and 1910 was much slower than we would expect in the midst of such an unprecedented trade boom, and it was much lower than productivity growth in the new-goods sectors. Since the former made up for a greater share of total output than the latter, overall productivity growth was muted despite falling trade costs. We ascribe this finding to "negative" selection effects.
The Impact of International Trade: Welfare, Institutions and Policies
Trade costs also directly affected welfare and institutional outcomes of interest in the first wave of globalization.6 Market potential, essentially the global demand for a country's output, is limited by trade costs and hence by the level of integration. Many studies covering the past few decades have found that higher market potential is strongly related to higher income per person. In this context, Liu and I study the importance of market potential in the late 19th and early 20th centuries.7 This work addresses an important and long-running historical debate about how the U.S. overtook Great Britain in productivity leader-ship in the late 19th century.
The economic history and economic growth literature has often attributed this event to the outsized U.S. domestic market. We find however that theory-based empirical measures of market size (i.e., market potential) for the U.S. are not significantly larger than they were for Great Britain, France, or Germany circa 1900. To be sure, international borders greatly reduced the leading European nations' trade, such that they faced an effective 60 percent ad valorem-equivalent tariff on their exports. At the same time, their domestic markets were dense and well-connected via water routes and extensive internal infrastructure, including roads, canals and railroads. We conduct a counterfactual simulation within a general equilibrium model of trade and find that had some of the smallest economies of the time (such as Belgium, Canada, Denmark, and Switzerland) been able to sell into global markets without facing international borders, their real per capita incomes could have risen to the levels attained by the U.S.
It is worth asking whether institutional and policy changes in the late 19th century were related to the first wave of globalization. First, observe that, from the middle of the 19th century, many countries dramatically extended the franchise, thereby increasing the level of ostensible democracy. A similar trend coincided with the more-recent wave of globalization, as the number and share of democracies in the world rose dramatically from the 1960s. Open-economy models of institutional change highlight that if trade induces a more-even distribution of income - say, as labor benefits from an increase in global demand - then greater democracy could result.8 We use an instrumental-variables strategy inspired by Jeffrey Frankel and David Romer to see whether, in the first wave of globalization in particular, exposure to trade flows, might have had a causal impact on democracy.9 There is little evidence that it did.10 However, in the late 20th century, we find that there was a statistically significant and positive relationship between these variables that was strengthened when the middle class benefitted from globalization, much as theory predicts.
Like modern economists and policy makers, authorities in the late 19th century wondered whether the technological changes affecting the integration of global markets would lead to intense labor market competition and a race to the bottom in terms of social policy. Bismarck, amongst others in Europe and the U.S., worried that domestic producers would be negatively impacted by radical changes to the social welfare state - such as the child labor laws, limits on working hours, and other labor standards which they were instituting. Despite the pessimism, labor standards were implemented in many countries. Strikingly, the data clearly show that a number of leading countries heavily exposed to international trade vigorously and enthusiastically adopted new labor regulations. Our research shows that country pairs that traded extensively with one another were more likely to adopt the labor standards of their trading partners.11 Evidently, trade can be used as a lever for better social protection, and globalization does not always promote a race to the bottom.
Recent research using new trade data and theory-based methodology has advanced our understanding of the causes and consequences of the first wave of globalization. Future work will provide new evidence based on recently digitized bilateral, product level trade for the United States, and this should shed further light on the industry-level growth impact of the first wave of globalization in an important industrializer. Together with similar datasets that are currently being processed by researchers around the world for China, France, Germany, Italy, Japan, Switzerland, and the U.K., a new view, or at least a greatly enhanced vision, of 19th century globalization is sure to emerge. Countries did not compete and grow based only on their factor endowments. Like today, producers and consumers gained from access to new finished and intermediate goods and higher-quality varieties of already existing goods, such that the welfare gains from trade strongly contributed to rises in living standards during the first wave of globalization.
2. J. E. López Córdova and C. M. Meissner, "Exchange Rate Regimes and International Trade: Evidence from the Classical Gold Standard Era, 18701913," American Economic Review, 93(1), 2003, pp. 344-53. ↩
3. D. S. Jacks, C. M. Meissner, and D. Novy "Trade Costs in the First Wave of Globalization," NBER Working Paper No. 12602, October 2006, and Explorations in Economic History, 47(2), 2010, pp. 127-41. ↩
4. D. S. Jacks, C. M. Meissner, and D. Novy "Trade Booms, Trade Busts and Trade Costs," NBER Working Paper No. 15267, August 2009, and the Journal of International Economics, 83(2), 2011, pp. 185-201. ↩
5. M. Huberman, C. M. Meissner, and K. Oosterlinck, "Technology and Geography in the Second Industrial Revolution: New Evidence from the Margins of Trade," NBER Working Paper No. 20851, January 2015. ↩
6. For a comprehensive survey of the historical literature on globalization and growth see: C.M. Meissner, "Growth from Globalization? A View from the Very Long-Run," in P. Aghion and S. N. Durlauf, eds., Handbook of Economic Growth, vol. 2B, Oxford, United Kingdom: Elsevier Press, 2014, pp. 1033-69. ↩
9. J. A. Frankel and D. Romer, "Does Trade Cause Growth?" American Economic Review, 89(3), 1999, pp. 379-99, and "Trade and Growth: An Empirical Investigation," NBER Working Paper No. 5476, March 1996. ↩
11. M. Huberman and C. M. Meissner, "Riding the Wave of Trade: Explaining the Rise of Labor Regulation in the Golden Age of Globalization," NBER Working Paper No. 15374, September 2009, and the Journal of Economic History, 70(3), 2010, pp. 657-85.↩