My esteemed colleague Sebastian Krapohl has just published a major book on regional integration. The major argument is that regional integration in the global „South“ follows a different logic than integration in the „North“, i.e. in the EU. I´m very happy that I could contribute to this book with a chapter on trade networks that corroborates this argument and shows the very different trade networks underlying, e.g. ASEAN, MERCOSUR and SADC.
In my course on policy networks, I today discussed and re-analyzed the famous article „Robust Action and the Rise of the Medici“ by Padgett and Ansell. We tried to find network visualizations that test (or illustrate) the claims made in the article, e.g.
„The Medici party was an extraordinarily centralized, and simple „star“ or „spoke“ network system.“ (p.1278)
One central claim is that
„The Medici did not marry those with whom they lived.“ (p.1281)
My students suggested that it would be nice to test this claim using actual geographical data. And thanks to the geographic mapping feature of visone, it worked out really well. I used wikipedia and http://www.mapcoordinates.net/ to find the coordinates of the headquarters of the main Florentine families (it was quite easy, if a family is called Strozzi, it is likely that there still is a „Palazzo Strozzi“), and merged these coordinate data to the visone network data.
Then I used mapping – coordinates – geographic, and here it is:
In the background is a map of Florence. The Medici and their supporters are green, their opponents purple, and those families with split loyalties are green. There is surely room for improvement, but we can see how the Medici (living „up north“) have their marriage ties spread out through the city, but not to those living close by (the Pucci, Ginori and Bischeri, for example).
Clever, clever Lorenzo de Medici…
On tuesday, I will participate in the brown bag session „Social Network Analysis – Ever expanding frontiers?“ at the ECPR Winter 2016 Winter School in Methods and Techniques. Together with Philip Leifeld and Kai Fischbach I will discuss the current state of play in the development of network analysis methods. I´m still not sure what my main discussion points will be, but probably along the line of: We may have sophisticated network methods, but the quality of network data has not kept pace with the development of methods.
Case in point: Trade data. I worked a bit with trade data myself, and one should think that interstate trade is the most well-understood, problem-free, and easily available type of network data. Dyadic trade is a standard variable in research on policy diffusion, the liberal peace, or regional integration.
Still, what happens if you want to know how much Swaziland exports to the EU? The answer depends very much on whether you ask Swaziland or the EU (data are from the SADC trade database and the World Bank WITS database):
Similarly, how much does Botswana export to South Africa…depends on who you ask:
Barbieri, Keshk and Pollins outline some reasons why importer-reported and exporter-reported trade data may not coincide, but the discrepancies here are huge. One could rightly argue that most trade data are more reliable (in the sense of: importer and exporter reports agree), but still: Subsahara Africa is an interesting world region for many reasons, and if we simply use the trade data at hand, we may be in trouble.
So, the message in a nutshell (or in a brown bag): Let´s try to complement sophisticated network methods with equally sophisticated (and valid and reliable) network data.
As always: Thanks for the smooth stata figure schemes goes to Daniel Bischof.
Or at least: Good news, Stata users.
I have just stumbled upon the nwcommands package for Stata. That seems to me to be one of the best innovations for Stata I have seen in years. Now I can visualize the networks inherent in my diffusion data using Stata, without having to transfer the data to pajek or visone. The ability to handle network data was a major reason to use R instead of Stata, because R has excellent network data packages…but now, let´s see what the new Stata package can do.
In 2013, the article „Different Paths of Regional Integration: Trade Networks and Regional Institution-Building in Europe, Southeast Asia and Southern Africa“ by Sebastian Krapohl and me will be published in the Journal of Common Market Studies. Major argument of the article is that the economic structures underlying regional integration projects strongly determine the institutional development of the project.
Update: The article has been published!
The article contains visualizations of trade networks underlying the EU, ASEAN and SADC. However, for publication on paper, we had to restrict the visualizations to a small number of time points. If you are interested in the networks for all time points, you can watch or download movies of our dynamic network visualizations (we used the dynamic layout algorithms offered by visone). You may watch the movies in your browser, but we recommend downloading them for better quality (you find the download button at the top right corner). Alternatively, you can download jpegs for all time points as zip-files.
EU: Movie / jpegs
ASEAN: Movie / jpegs
SADC: Movie / jpegs
As a robustness check for our analysis of ASEAN, we have also visualized how the ASEAN trade networks look like if we conceive of the ASEAN+3 countries (China, Japan, South Korea) as members of the regional integration project. This makes the ASEAN networks more similar to the economic structures underlying, e.g., the EU.
ASEAN (+3 aggregated): Movie / jpegs
Additionally, we have created movies and graphs visualizing the trade networks underlying NAFTA and MERCOSUR.
NAFTA: Movie / jpegs
MERCOSUR: Movie / jpegs