The first rule of studying international trade is that no single rule, statistic or equation is capable of adequately explaining either current or future trade patterns. Trade is too intricate and too variable, its shape naturally adjusting to conflict, politics, disruptive technologies and so on. This makes reliable data difficult to collect, and the presentation of trade data, like any statistic, can be manipulated to support almost any argument. Anecdotal evidence can complete the picture, or even paint a different one, and thus is just as important as what the numbers say. Ultimately, even the most elegant system of trade-related econometrics must judge itself not on the cleverness of its equations but on whether it describes reality accurately.
Despite the inherent difficulties of this work, the ability to precisely evaluate current and future global trade patterns is more important than ever. Protectionism is back in the news. In truth, protectionism never went away, and lest we think this is only a U.S.-driven process, we need only remind ourselves that Canada and the U.K. intimated last week that they were considering certain protectionist measures unrelated to U.S. moves. The international trade system since 1945 has been deeply marked by a veneer of free trade, but no one would assert that the progression toward freer trade is linear, an inevitable end of economic history.
In recognition of the need to do this important work and to make our analysis more accessible and transparent, GPF is exploring new quantitative ways to analyze trade relationships. We are beginning with a fairly unambitious goal: We want to establish a way to quantify the relative dependency in a bilateral trade relationship over a single commodity. The basis of our approach comes from a 1991 analysis by GPF founder and chairman George Friedman in relation to Japan’s dependence on imports. This approach is by no means sacrosanct, nor do we think it is a perfect representation of trade dependency. We do, however, think it is a useful tool for starting a more holistic conversation about how trade works – a conversation sorely missing from the present media environment.
This approach depends on establishing a relationship between three variables. The first variable is the importance of the commodity or good to the importer. Take, for example, Canadian imports of U.S. steel. We need to know how much steel Canada imports from the U.S. and how much it imports from the world overall. Dividing the Canada-U.S. import relationship by Canada’s total imports gives a sense of Canadian reliance on the U.S. for that commodity. Note that this is measured by physical weight (or in the case of a finished good, number of units), because that is more important to the importer than price.
The second variable is the importance of the trade of the commodity or good to the exporter. For the exporter, the monetary value of the trade is more important than the quantity – after all, the goal of the exporter is to make money, while the goal of the importer is to acquire a commodity or good. The second step, therefore, is to measure the importer’s payment for the import in relation to the total amount the exporter derives from that commodity or good in general. To use the example above, that would mean determining the amount Canada paid the U.S. for steel, and then dividing that figure by total payments for all U.S. steel exports.
Once these two relationships are established, a final variable is determined: whether the importer could easily find another source for the commodity or good in question. For this, we return to weight or unit measurements, and we compare the exporter’s production of the commodity or good (in this example, steel) to the total production of said commodity or good in the world. This last step is in many ways the most crucial. If the total availability of the commodity or good indicates that it would be fairly easy to find alternative sources, what might have looked like a highly dependent relationship for the importer can become a highly dependent relationship for the exporter.
By dividing the ratios established in the first and second variables, and then multiplying that figure by the ratio established in the third variable, we are left with a figure that gives a relative sense of dependency. If the figure is greater than 1, the importer is vulnerable; if it is less than 1, the exporter is vulnerable. The further the figure gets from 1 in either direction, the greater the dependency. For example, the vulnerability coefficient for Japanese imports of Australian coal in 1991 was 13.7, indicating that Japan was extremely dependent on Australian coal. (The figure has since dropped to 3.69, suggesting a trend of decreasing, if stubborn, dependence.) Conversely, Japan’s vulnerability coefficient for oil from the United Arab Emirates was just 0.01 in 1991, suggesting that Japan, which is almost entirely dependent on imports for its oil, nevertheless had the upper hand in the oil trade relationship with the UAE.
This way of looking at dependence comes replete with flaws, as does any mathematical approach. It does not, for instance, account for the ability of the importer to produce for itself the commodity or good in question. It is also less valuable the broader the definition is of the commodity or good. Steel is a commodity, but there are many types of niche steel products. Canada may not exhibit a statistical dependency on U.S. steel imports overall, but perhaps it does depend on American coiled stainless steel wire, which would be very important in determining the direction of a trade negotiation. This means that to grasp the nuances of dependency requires highly focused analysis of specific commodities about which it will often be difficult or impossible to acquire the requisite data to make a calculation.
GPF has always defined itself as an organization whose research is based on both qualitative and quantitative methodologies. For those to whom data is a religious principle, this has at times provoked negative reactions. They believe “qualitative” is a dirty word for lacking rigor. But blind attachment to data is equally lacking in rigor. We have always tried to find a medium between these two poles, both trusting what we see while developing mathematically driven ways to simplify a complex issue like bilateral trade.
We will use the approach above increasingly in the coming months. Indeed, our first analysis on this issue will appear later this afternoon. As always, we welcome your feedback – especially from those who see solutions to the flaws inherent in this approach. One does not need an economics degree to engage in this kind of work; real-world observations can be just as valid as attempts to describe reality through equations. This is not a knock on economists but a testament to our readers, who have a great deal of both to offer. We invite you to join us in better understanding international trade as old patterns begin to transform.