Evan Kalikow Dr.

Casper PL SC 300H 12/15/10 Research Design Paper Research Question: My research question is How can the Power Transition Theory explain intrastate conflicts involving non-state actors? As interstate war has decreased in prominence among major powers in the past few decades, it is important for theories that deal primarily with interstate conflict, such as Power Transition Theory, to be applied to intrastate conflict in order to remain relevant. There has not been much research on this topic, and my contributions could be applied to the conflicts going on in modern day Afghanistan or in several African countries. Literature Review: Throughout reviewing relevant literature, I found a few scholarly works that will help guide me through my research. The first of which is a paper written by J.M. DiCicco and J.S. Levy entitled Power Shifts and Problem Shifts: The Evolution of the Power Transition Research Program. DiCicco and Levy traced the evolution of Power Transition Theory research from its origins in 1958 to 1999, the year of the articles publication (DiCicco, 675-676). Their main takeaway was that future Power Transition Theory research must develop concepts and measurements for status quo dissatisfaction and better explain the causal steps leading up to war in order to be successful (DiCicco, 702-703). These lessons will be invaluable for my research. Dr. Lemke is the primary author who has researched Power Transition Theory in the capacity in which I intend to conduct research. His 2002 book, Regions of War and Peace, breaks the standard power transition model into a multiple hierarchy model, identifying four key regions and testing the theory within each of them. Power was measured through GDP,

population, and military strength, and status quo dissatisfaction is measured with a buildup of military spending (Lemke 2, 102). Lemke concluded that Power Transition Theory can explain conflicts within the multiple hierarchy model, showing that the theory can be applied to smaller models (Lemke 2, 109). In 2008, Lemke wrote an article, entitled Power Politics and Wars without States, in which he attempted to apply Power Transition Theory to intrastate conflicts in South America during the 19th century. He took about twenty autonomous political entities (APEs) from the era and defined their power based on population and their status quo dissatisfaction based upon their attitudes towards Buenos Aires efforts at unification (Lemke 1, 776). In the end, Lemke found that the Power Transition Theory applies not only among major powers in interstate conflict, but with regional powers in intrastate conflict as well (Lemke 1, 785). There has also been some research regarding economic impacts in wars in China during the twentieth century. In his 2003 article entitled "Power Shifts and Economic Development: When Will China Overtake the USA?", Emilio Casetti looks at economic activity in China and how that can relate to power shifts and power transitions. Casetti uses economic measures like GDP and GNP, growth rates, and development levels to measure power, comparing Chinas growth with the relative stagnation of other countries such as France, UK, Germany, Italy, and Japan (Casetti, 661-663). It was found that, through analysis over the past 50 years, China will overtake the United States power-wise within the next 50 years (Casetti, 773). Although this article deals with interstate conflict, it is still a valuable showing of the impact of economic development in China starting at the end of the warlord era and will be useful in my measurements of power.

Another important aspect that has been researched is the variation among regions during the warlord era. This is examined by Elizabeth Remick in her 2002 paper The Significance of Variation in Local States: The Case of Twentieth Century China. This article examines local state variation, particularly in China during the early and late twentieth century. Remick compares local states in two time periods: 1927-1937 and 1980-1994. Using economic measures like taxation, Remick argues that local state variation occurs during both time periods, a point that Remick believes is often overlooked (Remick, 415). This study shows that the different local states in 20th century China acted differently from one another, an invaluable point for my research for intrastate conflict during that time and in that region. David Carter provides more support regarding territorial disputes in his 2010 article entitled The Strategy of Territorial Conflict. This article attempts to measure territorial conflict by setting up a territorial dispute game, wherein states challenging a standing territory are given points based on if they engaged in conflict and, if so, if the target state decided to consolidate their power to improve their standing. Cases such as the Sino-Vietnamese conflict in the late 1970s and eighteenth-century Prussia were used to illustrate this concept (Carter, 2-5). Carter eventually found that consolidation by the target state was more likely and more effective when the territory in question had strategic location (Carter, 13). Although the examples used were all interstate conflicts, the broader findings can easily be applied to intrastate war, especially if the non-state actors have easily-defined borders. Power transitions and territorial conflicts are fused together in S. Swaminathans 1999 article, Time, Power, and Democratic Transitions. Swaminathan examines when democratic transitions are most likely to occur in this article. Using game theory, Swaminathan takes 17 Latin American countries from 1950-1990 and determines that democratic transitions occur

when there is relative parity between the two actors. In particular, mass mobilization is examined for the role it plays in determining if an authoritarian government will engage in civil war or negotiate a new status quo (Swaminathan, 178.91). While the study does not deal specifically with China, it does teach valuable lessons in the role of mobilizations in intrastate conflicts regarding power transitions, and it seems to agree with Carters conclusions. James Tongs 1991 book, Disorder Under Heaven: Collective Violence in the Ming Dynasty, is particularly useful for me. In this book, Tong looks at conflicts that occurred during the Ming Dynasty and attempts to explain when and why conflicts would arise during this time period. Tong measures violence based on frequency, extent, severity, and magnitude. He specifically does not include interstate conflict in the study and instead focuses exclusively on intrastate violence in 11 of the 15 provinces (the other four had insufficient data) and found that the second half of the Ming Dynasty was significantly more violent than the first half (Tong, 2526). The most relevant information for my research, though, was his aforementioned measures of violence, which will help me further analyze conflicts during the time period that I chose. I will be primarily basing my research off of Lemkes work, as he is the only author that I have found who has studied the general hypothesis in depth. There has not been much research on intrastate conflicts in China during the warlord era, especially with regards to power politics, and I hope to fill this hole with my own studies. Much of the research done in this region has dealt with how to measure power or violence, and those studies can be applied and modified to fit the time frame that I chose. Theories and Hypotheses: I will be testing the hypothesis that increases in power parity and in status quo dissatisfaction lead to an increased likelihood or magnitude of intrastate conflict, which are the

two independent variables that primarily make up the Power Transition Theory. Power will be measured by population size, military size, territorial control, military spending, and GDP (or revenue, as non-state actors may not have recognized GDPs). In order to measure conflict, fatalities, injuries, and frequency over time will be taken into account. My conflict measurements will be based around the model built by Tong, as mentioned in the literature review section. Status quo dissatisfaction will be measured by relative levels of military spending prior to conflict, public displays of dissatisfaction (riots, protests, etc.), and specific cases (for instance, if a non-state actors primary goal is to prevent the central government from invading Japan, then dissatisfaction would be measured by steps taken by the government to invade Japan). Research Design: For my research, I will be studying cases from the warlord era in China during the earlyto mid-twentieth century. It was a natural choice for region and time period, as I have already done research on non-state actors in the warlord era; additionally, the period has dozens of cases, making it easier to collect and analyze data. I will select cases from both the northern and southern regions of eastern China (where most of the warlords resided), making sure to select cases from my independent variables, including groups with both high and low dissatisfaction levels as well as high and low power parity. My unit of analysis will be conflict, with conflict outset and degree or magnitude of conflict each contributing to the general conflict levels. After careful examination of the time frame, it has become clear that it is unrealistic to do a detailed analysis of every twentieth-century Chinese warlord. There are no existing, comprehensive data sets, and the amount of data available for individual warlords vary strongly. Therefore, my research will be a nested analysis. I will first do a rough, quantitative analysis of each warlord for which some data exist. In this quantitative analysis, the only factors of power

parity that will be taken into account are population size and territorial control, as those data are generally available. Similarly, status quo dissatisfaction will be measured using only the quantitative measures of changes in military spending. For the qualitative portion of my nested analysis, I will select four or five cases of specific warlords. This case selection will be based on available data and diversity of cases (e.g. I will try to select cases from different regions and with different levels of power parity and status quo dissatisfaction). These cases could also be selected based on results gotten from the quantitative analysis, as that will give me a better idea as to what cases should have certain levels of power parity, status quo dissatisfaction, and conflict. In addition to the criteria and variables used in the quantitative section of my research, I will use more detailed variables to measure power parity and status quo dissatisfaction in the case studies, including revenue and military size/spending for power parity as well as public displays of dissatisfaction and case study analysis for status quo dissatisfaction. This will allow me to get a general picture for the conflict climate during the warlord era in addition to a more detailed analysis, and I will be able to determine if there is any discrepancy between the quantitative and qualitative analyses. In order to properly measure conflict and differences between the independent variables, I will put the actors in dyadic pairs. Since most conflict occurred between warlords and the central Chinese government, the dyads will almost certainly be China and the warlord in question. Some warlords did have disputes amongst themselves without significant intervention from the central government, though, and if those cases frequently appear throughout my research, it may be necessary to include warlord-warlord dyads. Feasibility:

For my quantitative analysis, the data do exist for my cases; however, each non-state actor has a varying degree of data available, and my selection for case studies will be partially dependent on how much data are available. To acquire my data, I have found several sources. One source of data that I will be using for my research is the International Population Census Bibliography. This bibliography includes sources of data for over 40 countries. Included in this bibliography are resources for Chinese census data from 1912, 1928, and 1940, all of which will be relevant for my research. Through this bibliography, I was able to find the Chinese census data, which include population size (including total population size, age size, and gender size) and government offices held. These data were originally collected in order to provide a better idea for the population size and breakdown of China, which complements my purpose of using the data (International Population Census Bibliography, 10-1 - 10-4). The International Population Census Bibliography will be of tremendous use to me. The population data that I have already found through it will help me determine Chinas relative power during the warlord era. Additionally, the government offices held and number of officers may provide some insight into the size and changes of Chinas military during this time, which will help me code both power and status quo dissatisfaction. In addition to the International Population Census Bibliography, I will be utilizing various historical accounts of the era in order to get more precise information on the warlords. Most notably, I will be using Edward McCords 1993 book, The Power of the Gun: The Emergence of Modern Chinese Warlordism. This book provides an historical account of the warlord era, highlighting both general trends in the era and going into detail with specific warlords. For instance, McCord provides a good deal of information about the territorial and

population reach of Lu Rongting, one of the most powerful warlords, who rose to power in 1916 (McCord, 228-229). Since this source is a written record, it will be necessary for me to extract the relevant information (population/territory size, military size, revenue, military spending changes in time, conflict). The most difficult part of that will be getting uniform data for each warlord that I use. I will be using several historical sources, including Arthur Waldrons From War to Nationalism, His-Sheng ChIs Warlord Politics in China, and David Bonavias Chinas Warlords. Each of these sources covers different timeframes (from 1911 to about 1940), different warlords, and they each have differing focuses, from general history to military history. Because of this, finding uniformity in data will prove to be a challenge. However, these sources each have extensive bibliographies and works cited pages, so it will be relatively easy to find the origins of such data. While the data I have found are quite useful, there are still missing data that will need to be collected in order for my research to be accurate and meaningful. The biggest missing pieces regard revenue and GDP of both warlords and the official Chinese government. The data sources I have used so far have not addressed financial or economic history in much detail, and it is a crucial component of my measures of power. The data found in the International Population Census Bibliography are quite reliable, as they are taken from Chinese census data. Since the censuses were measuring population size, which is what I am also measuring, they also have high validity. Their measures of government offices and titles will have less validity, as I will be using those to estimate the size or importance of the military, but I still believe those data to be indicative of military size and stature.

Similarly, the written record data, such as the data found in McCords book, are relatively reliable, with population size, territory size, and conflict data that can be found on repeated trials. Because these written records have varying purposes, from describing the regions history to talking about military strategies, their measures will have mixed validity when applied to my research. As a researcher, it will be my responsibility to ensure that the data sets I code are both reliable and valid. Since the data I have found are raw numbers and not part of existing data sets, I will need to code said data in a manner that ensures that future researchers can repeat my findings and yield the same results. It will also be integral that I guarantee validity in my measures. I have put a great deal of thought into what data I would need to measure power and status quo dissatisfaction, and I still firmly believe that the data for population size, military size, territory control, and revenue/GDP to be accurate measures of power, with changes in military spending and public displays of dissatisfaction to accurately measure status quo dissatisfaction. My data sources have both their advantages and their drawbacks. Since the census data are raw, they are not constrained by researcher biases and are not subject to the judgment calls that researchers must make when creating data sets. The census data also come officially from the Chinese government, which adds to their integrity. The same measures are used across each of the censuses, so there is a good deal of uniformity. Additionally, the data are all in English, meaning that I can easily access them without the need of a translator or translation tool. There are, however, some disadvantages to the census data. The data are limited in that they only give numbers regarding population sizes and demographic sizes. There are no financial data relevant to my research in the census. The census also only measures data for the entirety of China, so it provides little to no assistance regarding warlords, which were not officially

recognized by the Chinese government. Furthermore, the census was only taken in three years during the timeframe in which I am interested: 1912, 1928, and 1940. That means that it may prove difficult to find more precise data for specific years in the censuses, besides those three previously mentioned. The written historical records are a mixed bag as well. On the positive side, each historical account that I have encountered so far has been filled to the brim with detail, citations, and variety about the warlord era. Even more importantly, these written accounts provide great context for the warlord era, both in the instances of specific warlords and in the time period in general. Neglecting this kind of context has been the pitfall of many Power Transition Theory scholars, so being able to tap into this insight will be invaluable. Unfortunately, the historical records also have some significant drawbacks. First, like with the census, the data are incomplete, with financial records and revenue being the most notable missing variables. Furthermore, there is little to no uniformity of coding among the sources or even within the sources. My coding data would essentially be based on whatever information the authors deemed necessary to include while describing the historical events. Additionally, these sources tend to only discuss conflict when it happens, not in the negative. However, it is called the warlord era for a reason, so conflict may have to be measured based on intensity and casualties rather than as a binary calculation. I can, however, use the aforementioned extensive bibliographies included in these sources to find the origins of the data and other relevant information. One significant problem I have is that I will need to code the data that I find myself. Coding the data will be a long, involved process, and my case selection will necessarily be limited based on what data I can uniformly code. The fact that this was such a turbulent period in

Chinese history with territorial and financial claims rarely at consistent levels does not help matters. Despite these issues, however, there is still some benefit to be found in coding my own data. I will have complete control over the coding rules, and I will be able to avoid false equivalencies and misinterpreting the purpose of other data sets. In addition, though the sheer volume of the cases that I will be able to use will be limited, I will understand each case very well, which is a privilege that those coding hundreds of case may not be able to enjoy. This is especially true in the case of the warlords that I choose to examine as part of my in-depth case studies. As previously mentioned in the literature review, Power Transition Theory scholars often overlook the contextual nature of conflicts and see them as just numbered. By doing detailed case studies for at least a few cases, I will be able to more clearly understand the causes and effects of each conflict. Possible pitfalls could arise through my measurement of status quo dissatisfaction. Researchers have historically struggled with this, and I fear that it could be possible to misrepresent the dissatisfaction levels through an over- or under-reliance on each of the previously mentioned measures. If dissatisfaction is not properly gauged, then my research question is put into jeopardy. I will need to do further research into my cases and the time period to determine which measures of dissatisfaction make the most sense to pursue. Another potential pitfall regards the dyads that I choose to analyze. Although the vast majority of intrastate conflict in China during the warlord period involved one warlord fighting against the central Chinese government, I will have to be incredibly careful in which dyads I use. If I pair a certain warlord with the government but that warlord only had significant disputes with other warlords, then the data I collect from that analysis will not be as useful as it could be. This will not be as much of a problem for the cases in the qualitative part of my nested analysis, as I

will have a solid understanding of their conflict history, but I will need to make certain that the dyads that I use in the quantitative section are soundly selected.

Sources: Bonavia, David. China's Warlords. Hong Kong: Oxford UP, 1995. Print. Carter, David B. "The Strategy of Territorial Conflict." American Journal of Political Science (2010): 1-19. Print. Casetti, Emilio. "Power Shifts and Economic Development: When Will China Overtake the USA?" Journal of Peace Research 40.6 (2003): 661-75. Print. Ch'i, Hsi-sheng. Warlord Politics in China 1916-1928. Stanford, CA: Stanford UP, 1976. Print. DiCicco, J. M., and J. S. Levy. "Power Shifts and Problem Shifts: The Evolution of the Power Transition Research Program." Journal of Conflict Resolution 43.6 (1999): 675-704. Print. International Population Census Bibliography. Austin: Bureau of Business Research, Graduate School of Business, the University of Texas, 1966. Print. Lemke, Douglas (1). "Power Politics and Wars without States." American Journal of Political Science 52.4 (2008): 774-86. Print. Lemke, Douglas (2). Regions of War and Peace. Cambridge, UK: Cambridge UP, 2002. Print. McCord, Edward Allen. The Power of the Gun: the Emergence of Modern Chinese Warlordism. Berkeley: University of California, 1993. Print. Remick, Elizabeth. "The Significance of Variation in Local States: The Case of Twentieth Century China."Comparative Politics 34.4 (2002): 399-418. Print. Swaminathan, S. "Time, Power, and Democratic Transitions." Journal of Conflict Resolution 43.2 (1999): 178-91. Print. Tong, James. Disorder under Heaven: Collective Violence in the Ming Dynasty. Stanford, CA: Stanford UP, 1991. Print. Waldron, Arthur. From War to Nationalism: China's Turning Point, 1924-1925. Cambridge: Cambridge UP, 1995. Print.