2020 Fiscal Year Research-status Report
Creative Destruction in the International State System
| Project/Area Number |
19K01605
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
Weese Eric 東京大学, 社会科学研究所, 准教授 (50777844)
|
|---|
| Project Period (FY) |
2019-04-01 – 2022-03-31
|
| Keywords | Balance of Power / War / Country dyads / Coalition formation |
|---|
| Outline of Annual Research Achievements |
We made substantial revisions to the first paper linked with this project ("Creative Destruction in the European State System: 1000-1850") and submitted it to the American Economic Review. Unfortunately it was rejected after split reviews from the referees. We intend to revise the paper substantially before submitting it to another journal.
We started research on the second paper linked with this project. Currently, empirical research that examines wars between countries generally models those wars using a model of pairwise wars, where each potential pairwise war either occurs or does not. This dyadic approach is easy to estimate using a probit model; however, there is no potential to model how sides emerge in each war. In this new paper we aim to develop a simple model (based on coalition formation models) that allows us to examine whether the two sides in historical European wars were "more evenly balanced" than we would expect by random chance, and whether the most powerful countries are more often involved in wars than would be expected.
We have finished data collection and cleaning for this project: we now have a dataset of wars in Europe over the past 1000 years, linked with the countries that existed in Europe at that time. Work on the econometric model is ongoing.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
For our completed working paper ("Creative Destruction in the European State System"), a difficulty at this point is that the paper is now at the maximum length for journal submission, and so any work to address referee comments would likely require deleting material from the paper. This makes it more difficult to address the negative comments we have received. However, we aim to submit the paper to a new journal in the coming year.
For our new project on warfare in Europe over the past 1000 years, a major difficulty is developing an appropriate estimation strategy to produce parameter estimates for wars that involve more than two countries (one on each side). No standard estimation procedure seems likely to work in this case, as even a simple model of warfare would make it possible for there to be multiple equilibria (i.e. different wars) that could occur with the same observed characteristics and idiosyncratic shocks. We attempted to address this using a maximum score style estimator, but the approximation we were basing this estimator on appears to have substantial problems with bias even with the relatively infrequent and small wars that we observe in our actual data.
|
| Strategy for Future Research Activity |
We need to finalize the estimation strategy over the next year. One possibility might be to use a simulated method of moments approach, taking advantage of the fact that not very many wars appear in our data, and so it may be computationally feasible to simulate what wars our model predicts would occur at various parameter values. Once we have a working estimation strategy, we can then produce parameter estimates describing what configurations of countries are more likely to go to war, and whether powerful countries are more likely to be "ganged up on" by their weaker rivals.
Although there is a substantial literature using dyadic regression to look at pairwise wars, we observe many wars involving more than one country on each side, and thus an estimation strategy that could work with this sort of data would be a substantial contribution to the literature.
Separately, we will finalize a method of ranking countries based on what wars they win. This provides us with an important input to our model of why certain wars occur while others do not: in order to see whether the most powerful country is involved in more wars than would be expected, we must have some sort of index in order to determine what country is actually most powerful. For the early period in our data (1000-1500 AD), there are very few data sources available, and so we must work only with boundary changes to determine which countries are more powerful than which other countries. We intend to base our method on looking at which countries lost territory as a result of wars, and which countries gained it.
|
| Causes of Carryover |
No travel was possible due to corona.
|
Research Products
(2 results)
