The Electronic World Treaty Index

“International Law is to law what Professional Wrestling is to wrestling,” or so claimed the commentator Stephen Budiansky.  The political scientist George Downs and his co-authors once claimed that “most treaties require states to make only modest departures from what they would have done in the absence of an agreement.”  In other words, there exists a long-standing debate in which some scholars view international law as virtually meaningless or as mere “scraps of paper,” to borrow the oft-cited phrase attributed to Bismarck.

Is this perception true?  Presently, we have no way of systematically answering this and related questions.  Why?  Because scholars do not possess the population of international treaties and, without this population, it is not possible to draw a truly representative sample on which to study.  This theme is echoed in the recent working paper by Miles & Posner  (2008) where the authors note: “[M]any empirical regularities of treaty-making remain almost entirely unknown. Basic facts, such as which types of treaties countries most commonly enter and which types of countries engage in the most treaty-making –- let alone what are the consequences of treatymaking –- have not been established.”

Overcoming these limitations is the goal of the Electronic World Treaty Index (www.worldtreatyindex.com).

Existing Treaty Databases

Scholars have often sliced off and studied portions of the body of international legal agreements.  The Trade Agreement Dataset (discussed here) lists all trade agreements involving reciprocal concessions on tariffs or tariff-equivalents (such as import quotas) from 1815 to 1914. The Continent of International Law (COIL) database, first introduced in Koremenos (2005), draws a sample of treaties from the list of treaties reported in the United Nations Treaty Series (UNTS) with the goal of coding these treaties’ provisions. The Alliance Treaty Obligations and Provisions (ATOP) dataset compiled by Leeds, Ritter, Mitchell, and Long (2002), is a comprehensive dataset of all known military alliance treaties formed between 1815 and the present.  The Multilateral Agreement and Treaty Record Set (MARTS), developed by Denemark and Hoffman, includes 6976 multilateral treaties signed between 1595 and 1995.

Though these projects have coded the makeup of various treaties, they are of limited scope. They either focus on a specific treaty topic, contain a random (or not so random) sub-sample of specific treaty categories, or limit their compilation to just multilateral agreements. Truly gaining a perspective on the development and dynamics of global cooperation via treaty making requires a broader and more comprehensive treaty compilation. The World Treaty Index, as first developed by Rohn (1983), avoids these limitations.

Rohn’s World Treaty Index (WTI)

In the 1960s, Peter Rohn of the University of Washington began a National Science Foundation-funded project of developing a comprehensive database of metadata describing international treaties. The first edition of the World Treaty Index was published in 1974, with a second edition published in 1983 (and reprinted in 1997). Though the effort originally focused on just coding registered treaties found in the League of Nations Treaty Series (LTS) and the United Nations Treaty Series (UNTS), Rohn soon expanded the project to include thousands of unregistered treaties. Forty percent of the treaties in the 1974 edition, and forty-five percent of the treaties in the 1983 edition, were unregistered treaties. In total, the first edition of the WTI, which covered the 1920 to 1970 time period, included metadata for 4,834 treaties from the LTS; 9,832 treaties from the UNTS; and 7,000 treaties identified through national treaty lists and collections. The second edition incorporated metadata for treaties from 1900 to 1920, newly discovered agreements from 1920 to 1970, and treaties from 1970 through 1979 for a total of 44,000 treaties. The inclusion of metadata for such unregistered treaties justifies characterizing the World Treaty Index as a valuable, if not definitive list of international treaties.

Updating and Digitizing the WTI

Since the publication of these volumes, the WTI has been moved to electronic format and the coding of treaties has continued. However, as described in Pearson (2001), the coding of treaties since 1990 has been limited to the UNTS and, starting in the late 1980s, the WTI also stopped inputting information about treaty annexes.

Starting in 2010, the World Treaty Index entered a new phase with a dedicated URL (http://worldtreatyindex.com/) and a complete site redesign. The site is now maintained by Paul Poast, Michael Bommarito, and Daniel Martin Katz.  Under their stewardship, the WTI has undergone several updates.  First, metadata on all pre-1945 treaties was added to the electronic WTI.  Second, treaties for 1998 and 1999 were added.  Third, the multilateral treaties were added.  When complete, the WTI will provide a comprehensive database of metadata for over 75,000 treaties formed during the twentieth century.

Using the World Treaty Index Website

A user arriving at the World Treaty Index homepage (www.worldtreatyindex.com) will encounter the searchable interface pictured below:

The user can choose to search treaties by country name, by the widely used Correlates of War treaty code (www.correlatesofwar.org), by a specific topic (see below), or by signing date.

For example, one could query for all treaties signed by the United States between 1970 and 1980.

Once the user selects the desired search options, a dedicated page for the query will open.  This page provides three pieces of information.  First, the top of the page provides summary statistics regarding the search.  It lists the total number of treaties that meet the search criteria and provides a visual illustration of the distribution of the requested treaties over the requested time period.  This visual illustration could be useful for researchers wishing to identify trends.  For example, the figure shows a high level of treaty signings in the early 1970s relative to the late 1970s.  Exploring this trend may be of interest to the user.

Second, there is an option that allows the user to download metadata for the requested treaties as a .csv file.  We use the .csv file type as it is supported by various software packages including Excel, R, Stata, SPSS, and SAS.

Third, the page provides the WTI entry for each treaty meeting the search criteria. Each treaty receives one entry in the WTI, and each entry records information about the treaty in several fields. The first field is the Treaty Serial Number, which begins with a number indicating the treaty source, followed by a 0, and then the treaty code found in that particular source document. For instance, if a treaty is drawn from the UNTS, the serial number begins with the number 2 (indicating UNTS), then a 0, and then the UNTS code for that treaty. The second field is the Primary Source Document, which simply provides an abbreviation of the source document from which the treaty is drawn. For example, the United Nations Treaty Series code is UNTS. The third and fourth fields indicate the volume and page number of the Primary Source Document in which the treaty is found. For example, if the treaty is drawn from the UNTS, then the volume field records the UNTS volume number, and the page field records the page of that particular volume.

The next two fields record the parties of the treaty if the treaty is bilateral or unilateral. We are working on incorporating multilateral agreements into a single query.  In our current database schema, we separate these queries from queries for bilateral treaties by allowing for bulk download of metadata for all multilateral agreements. (Click here for the page offering bulk access to metadata for multilateral agreements.) Following the results presented in Poast (2010), we would encourage scholars to carefully consider how best to combine datasets of metadata for multilateral and bilateral treaties.

Fields eight, nine, and ten record, respectively, the date the treaty was signed, the date the treaty went into force, and the date it was registered with the United Nations (if it is a UNTS treaty).

The final three fields record information on the treaty content itself. The topic field records a code for each treaty indicating its substantive issue category (the complete topic list is available on the website). Rohn had developed 90 different treaty topics, over 9 broad categories. The broad categories are: Diplomacy (primarily security issues), Welfare (primarily health and labor standards), Economics (primarily trade and financial cooperation), Aid (primarily humanitarian and development aid), Transportation (primarily addressing rights to air space or water ways), Culture (primarily educational and scientific collaboration), Resources (primarily natural resource management), and Administration (primarily consulate matters or visas).

The title field indicates the type of agreement, such as convention, protocol, etc. (A list of the treaty title categories used by the WTI are also available on the website). Finally, the headnote field records the title of the treaty, so as to quickly provide the researcher with specific details on that particular treaty’s content.

Thus, for example, the bilateral treaty between the United States and Egypt signed on November 3, 1949 has a title of “agreement”, a headnote of “Financing Educational Exchange Program”, and a topic of 4EDUC.

For additional information on these fields please visit: http://worldtreatyindex.com/fields.html

Coming Soon

Our broad goal is provide support for the rigorous exploration of cooperation and conflict in the international system.  In that vein, we are currently exploring a variety additional features that we believe would help aid end users:

(1) Updating the WTI for coverage through 2011
(2) Adding information on treaty terminations and renegotiations
(3) Developing a better interface for Multilateral Agreements
(4) Developing direct access to treaty text (via Hein Online or other third-party providers)

Paul Poast is currently a Ph.D. Candidate in Political Science at the University of Michigan.  In the Fall 2011, he will be an Assistant Professor of Political Science at Rutgers University.

Daniel Martin Katz is currently a Ph.D. Candidate in Political Science and Public Policy at the University of Michigan.  He is also a Fellow in Empirical Legal Studies at the University of Michigan Law School.  In the Fall 2011, he will be an Assistant Professor of Law at the Michigan State University College of Law.

Michael J. Bommarito II is currently a Ph.D. Pre-Candidate in Political Science at the University of Michigan.  He also has a Masters in Financial Engineering from the University of Michigan.

VoxPopuLII is edited by Judith Pratt. Robert Richards is Editor in Chief.