China's Militarized Interstate Dispute Behaviour 1949 1992: A

China's Militarized Interstate DisputeBehaviour 1949-1992: A First Cut at the Data*

Alastair lain Johnston

China's military exercises in the Taiwan Strait in 1995-96 accentuatedconcerns among states in the Asia-Pacific region about what kind of greatpower China will become if its economic growth continues at presentrates and if its domestic political system does not change appreciably. Inmost respects many Chinese internal post-mortems on the crisis werequite similar to those in other states: coercive diplomacy led to anincrease in voter support for Beijing's nemesis, Lee Teng-hui, and itincreased worries among surrounding states about how China mighthandle bilateral disputes with them; but it also showed just how seriouslythe Chinese regime takes threats to the related interests of territorialintegrity and domestic legitimacy.1

The question is: to what extent might China's recent coercive diplo-macy be a harbinger of future approaches to resolving regional conflictsof interest, that is, to what extent is the Taiwan issue an aberration, aunique case that has less relevance for forecasting China's conflictmanagement behaviour? Thus far there is little consensus about what theTaiwan crisis meant, indeed on what China's rise as a great power meansfor regional and global peace and development. The debate in the UnitedStates over the implications of "China's rise" - while there are those withquite set opinions - is just starting. This debate, however, lacks sophisti-cated analytical frameworks, research and evidence, in part because theChina question has moved so quickly and suddenly on to the official andacademic radar screen, and in part because Chinese foreign policy is aseverely understudied subfield of international relations.

This study seeks to contribute to this discussion by concentratingon patterns in Chinese conflict behaviour and crisis management since1949. This does not presuppose that future behaviour can always bepredicted by past behaviour, though this is often a good place to start. Itdepends on what variables change. Economic development strategies andspeed are obviously variables that have changed dramatically. China'sparticipation rates in international economic, social, environmental andsecurity institutions have increased in the last 15 years. The interdepen-dence literature would argue that the economic costs of potential conflict

*The author gratefully acknowledges the Olin Institute of Strategic Studies for helping fundthis project. Thanks also to Tom Christensen, Karl Eikenberry, Paul Godwin, MichaelPilsbury and Stephen Rosen for their comments and criticisms.

1. This was the assessment of a number of foreign policy analysts with whom I spoke inBeijing and Shanghai during a research trip to China from January to July 1996. The officialline was that the relative public silence from states in the region indicated basic agreementwith China that Taiwanese "flexible diplomacy" had gone far enough and was threateningregional stability. Internally, however, many analysts concluded that the only positiveoutcome, from China's perspective, was that it had increased the credibility of future threatsto use force.

© The China Quarterly, 1998

The China Quarterly

ought to be expected to have a more constraining effect on Chinesebehaviour as the Chinese economy has become more dependent onforeign investment and markets for growth, and concomitantly, as theParty's legitimacy rests increasingly on supplying these economicbenefits. The literature on how international institutions and norms affectstates suggests that China's behaviour will be increasingly restricted ortransformed by the rules of these institutions.

On the other hand, the nature of China's political system has changedmuch less dramatically: the regime continues to promote nationalistdiscourses in domestic and foreign policy. Hyper-sovereignty values arestill a central driver of Chinese foreign policy. A realpolitik strategicculture still colours the world-views of many of China's senior securitypolicy decision makers, a world-view in which military force is apotentially useful tool, among others, for the pursuit of traditional powerand prestige maximizing national interests in a competitive and relativelydangerous world.2 There is still a large number of unresolved, or ambigu-ously resolved, territorial disputes between China and its neighbours.There is still a relatively large gap between the status that China desiresand the status that other states are willing to bestow. And it is still notclear that China's decision makers fully understand the concept of thesecurity dilemma - where a defensive action taken by one status quoactor is interpreted as threatening by another; the second actor then takeswhat it believes are defensive counteractions that, in turn, are interpretedby the first actor. The result is an interactive spiral of insecurity.3

The tension today between external economic, institutional and norma-tive constraints on Chinese behaviour on the one hand and realpolitikconcepts of "majorpowerhood" internalized by China's leaders on theother has never been more acute in the history of the People's Republic.This makes straight-line projections from historical patterns in conflictbehaviour somewhat risky. Those who doubt that "interdependence" hasconstrained Chinese behaviour point to the recent Taiwan crisis as anexample. Between 18 and 30 per cent of China's exports go to theAmerican market and almost $30 billion of Taiwanese money is currently

2. This is not a unique world view among the major powers. China's version of sovereigntyconverges in many places with that espoused by Gaullists in France and Republicans andisolationists in the United States. The reason the U.S. has not ratified international treatiesgoverning the rights of women and children, the reason why Jessie Helms opposed Americanratification of the treaty on genocide and the Chemical Weapons Convention, the reason whyBob Dole once claimed that UN peacekeeping activities were "out of control," and the reasonthe U.S. is in arrears in its financial obligations to the UN is precisely the fear in some quartersin the U.S. that international institutions and obligations impinge on its sovereignty andautonomy.

3. In one conversation I had with a highly placed strategic analyst in PLA in July 1996he concluded that the concept of a security dilemma was probably not well understood at thetop. Thus they are less sensitive to the interactive, and possibly counterproductive effects ofChina's military modernization programme. The same, of course, could be said aboutAmerican and Soviet leaderships through most of the Cold War, with the exception perhapsof Gorbachev. The Clinton administration's drive to develop and deploy ballistic missiledefences that might also undermine China's fragile deterrent is another example of inattentionto unintended consequences (Chinese nuclear modernization) that could, in the end, reduceAmerican security.

China's Militarized Interstate Dispute Behaviour 1949-1992

invested in China, yet Chinese leaders risked a major confrontation withAmerican naval power over Taiwanese flexible diplomacy. This doeslittle to buttress the argument that "trading state" major powers are lessbelligerent than traditional territorially acquisitive major powers. AnAsian Wall Street Journal article argued during the missile crisis in thespring of 1996 that China's military exercises turned interdependencearguments on their head. For proponents of this kind of argument,China's past behaviour in conflicts and crises is crucial for forecastingfuture behaviour.

The counter-argument would look at the counter-factuals. Would therehave been such a careful delineation of the boundaries of the PLAmilitary exercises, such careful signalling to American and Taiwaneseofficials, such continuing normalcy in all other aspects of Sino-U.S. andChina-Taiwan economic, cultural and political relationships in the midstof a very heavy concentration of modern firepower if these economiclinkages had not existed? Some analysts I spoke to in China at the timeof the Taiwan crisis acknowledged that regional economic and politicsleaders from South-East China were not as enthusiastic about the use offorce as the PLA apparently was, for obvious economic reasons. Forproponents of this kind of argument, China's past conflict managementbehaviour is less likely to determine future behaviour, except for issuesthat clearly affect territorial integrity and political legitimacy. Whicheverargument is more accurate, however, a systematic understanding ofpatterns of conflict and crisis management would be a helpful addition tothe evolving debate about the meaning of China's rise.

The focus of this study is on Chinese conflict and crisis managementbehaviour. Neither of these two concepts is particularly well defined inthe international relations literature. Steve Chan, who provided one of theearlier, systematic analyses of Chinese behaviour, defined conflict man-agement as "deliberate policy programs designed to monitor the actionsand intentions of one's actual or potential adversaries, to deter them fromcarrying out unwanted policies, to engage in military conflict with themshould deterrence fail and to control and resolve such conflicts lest theythreaten more important domestic and foreign policy goals."4 This is avery broad definition: it includes day-to-day security analysis, political,diplomatic and military signalling, the actual use of force, and thetermination of the use of force. In effect, it covers all five nodes in arough conceptualization of the stages of conflict behaviour by states (seeFigure I).5 Chan's own study, however, focuses empirically on five casesof the use of force (such as situations that emerge after deterrencefailure), that is, behaviour that occurs after node three. This selection

4. Steve Chan, "Chinese conflict calculus and behavior: assessments from a perspectiveof conflict management," World Politics, No. 2 (1978), p. 391.

5. Figure 1 provides a simplification of the major decision stages that foreign policy decisionmakers face in conflicts with other states. Once a conflict has emerged, it may develop to alevel of acuteness where it constitutes a crisis. Once in a crisis, decision makers must decidewhether or not to use force. Once they have decided to use force, they must then decide whatlevel and spatial/temporal scope of force to use. A particular scope of force may then changethe crisis situation such that escalation is required or termination of the crisis becomes possible.

4 The China Quarterly

Figure 1: Conceptualization of a Crisis

escalation.

noconflict

makes it difficult to determine the conditions under which, when in acrisis, China decides to use force in the first place.

Crisis management behaviour refers somewhat more narrowly toactions in situations of short duration and intense interaction where thereis a high probability of escalation to violence because the things underthreat (say, territory, strategic policies or regime stability) are consideredto be highly valuable (such as nodes 2 to 5).6 Of course, all nodes are ofinterest in order to understand the nature of Chinese coercive diplomacy.But for the purposes of intellectual manageability, this study concentratesmore on crisis management behaviour: in other words, what kinds ofcrises has China historically found itself in, and once in a crisis ordispute, under what conditions does it resort to what kinds (scope) offorce (nodes 2 to 4). Patterns in China's escalation or termination ofcrises, of course, are critically important to understand as well, but sucha study should come after one that focuses on explaining the resort toforce in the first place.

One critical issue is evidence and sources. Given the obviousdifficulties in acquiring detailed primary materials on Chinese crisismanagement decision making over a large number of cases, my approachis quite eclectic. I am interested in broad patterns in the frequency andscope of crisis management behaviour, as well as in some initial compari-sons with other countries, so as to put Chinese behaviour in context. Alltoo often, consciously or unconsciously, China specialists have treated

6. Jonathan Wilkenfeld, Michael Brecher and Sheila Moser, Crises in the TwentiethCentury, Vol. 2: Handbook of Foreign Policy Crises (New York, 1988). Chinese definitionsof conflict and crisis are somewhat more abstract than those used by Western scholars.Strategic analysts at the National Defence University define a conflict as a situation whereinterstate differences and disparities in interests are handled differently by different states suchthat these different roads to resolution do not lead to unanimity. Under these conditions a crisisis a situation of mutual resistance. See Gao Jinxi and Gu Dexin, Guoji zhanlue xue gailun(Introduction to International Strategic Studies) (Beijing: National Defence University Press,1995), pp. 158-59.


Chinese behaviour as unique and incomparable across time and acrosscountries.

This article explores a relatively new data set on national crisismanagement behaviour, the Militarized Interstate Dispute (MID) data set.This was compiled by the Correlates of War research group at theUniversity of Michigan over the last decade or so, and defines MIDs as"united historical cases in which the threat, display or use of militaryforce short of war by one member is explicitly directed towards thegovernment, official representatives, official forces, property, or territoryof another state,"7 that is, a sequence of causally connected actions wheremilitary force of varying levels is used against other states for politicalpurposes. For China, there are 118 cases of MIDs from 1949 to 1992, theyear for which the data set ends.

There are two key variables of interest in the MID data. The first is thetype of action taken by a disputant. Here the compilers of the data setdeveloped two indicators. One is a five-point categorical hostility scalemeasuring the level of the hostility reached by the disputant in a disputeand ranging from no hostile actions up to interstate war. The other is atypology of 21 kinds of actions ranging from no militarized actionthrough blockades and seizures to the use of nuclear weapons.8 Thesecond main variable of interest is the type of goal in the dispute. TheMID data classifies the disputes into three basic types: "territorial," wherea state is using some level of force primarily to change or defend theterritorial status quo; "policy," where a state is using some level of forceprimarily to change the foreign policies of another; and "regime," wherea state is using some level of force primarily to change the nature of thegovernment of another state.9

The caveats in using aggregate data are obvious and there is no oneinherently better or worse method. The problems of inference in quanti-tative analysis are different from those in qualitative analysis, but they arenot worse. The problems arise mainly with the development of validindicators for some underlying variable whose causal effects are beingtested, or with the sloppy coding of cases. Those who use these data mustto some extent rely on the skill and consistency of those coding the data.But the process of drawing inferences is explicit and, in principle,reproducible, open to confirmation or critique by other scholars. Qualita-tive methods run into problems precisely because of the difficulty in

7. Stuart Bremer, J. David Singer and Dan Jones, "Militarized interstate disputes,1816-1992: rationale, coding rules and statistical findings," mimeo, February 1996, p. 6.

8. See n. 31 for a detailed list of the categories in these two variables.9. There is a certain arbitrariness in this kind of coding. Disputes almost invariably involve

more than one issue and states often have a number of goals when entering a dispute. But,to the extent that the historical documentation is available, it is often fairly obvious how thesegoals are ranked by decision makers. Given the number of MIDs coded for all countries from1815 to 1992 (N = 2042), and given the general credibility of Correlates of War data sets inthe field of quantitative international relations, I think it is fair to assume that codinginstructions were rigorous and coding errors were random across all cases and thus are randomfor the Chinese cases as well (e.g. that they cancel each other out on balance). The processof coding is described in Daniel Jones, "Preliminary user's manual: militarized interstatedisputes" (Ann Arbor: The University of Michigan, Correlates of War Project, 1 July 1991).

The China Quarterly

reproducing, confirming and critiquing the process of inference. Bothkinds of problems require one to be tentative in the presentation offindings. Suffice it to say, the analysis of the MID data here is"triangulated" where necessary with other primary and secondary materi-als on specific cases.

The article is organized as follows. The first section reviews therelatively limited existing literature on Chinese conflict and crisis man-agement to draw out what is thought to be known about patterns inChinese behaviour. The second section provides a wholly inductive,descriptive analysis of patterns in Chinese behaviour found in the MIDdata set. The third section then tests some of the predominant explanatoryhypotheses about crisis management behaviour in international relationstheory literature to see if any sense can be made out of the patterns in thedata.

What is Thought to be Known About China's Crisis Behaviour

The dominant view of Chinese conflict and crisis management behav-iour, whether in Western or Chinese scholarship, is that China hashistorically stressed the limited, political uses of coercive diplomacy, andhas eschewed crusading, offensive wars a outrance. This pattern, it hasbeen argued, characterized imperial China and has persisted into thePeople's Republic. The secondary literature pretty much accepts thenotion that in Chinese strategic thought and practice the use of violencewas a "last resort," to use Fairbank's words.10 This reluctance to resort toforce rested, it is argued, on a low estimation of the efficacy of violencewhich was inherent in the Confucian moral order, and led to the "pacifistbias of the Chinese tradition."11 Different authors put it different ways: astress on psychological warfare over the use of weaponry and firepower;12

a stress on "gaining victory while keeping as much intact as possible bothsocially and materially rather than destroying who or whatever stands inthe way";13 an "anti-militarist bias";14 an emphasis on humans or mindover weaponry and the use of non-violent strategem and deception toovercome "bare strength";15 an "aversion to violence in war";16 the

10. John K. Fairbank, "Varieties of the Chinese military experience," in Frank Kiermanand John K. Fairbank (eds.), Chinese Ways in Warfare (Cambridge MA: Harvard UniversityPress, 1974), p. 7; Thomas Cleary (trans.), Mastering the Art of War: Zhu Getiang and LiuJi's Commentaries on the Classic By Sun Tzu (Boston, 1989), p. 20.

11. Fairbank, "Varieties," pp. 6-7, 11, 25.12. Michael Pillsbury, "Strategic acupuncture," Foreign Policy, No. 41 (1980/81).13. Thomas Cleary (trans.), The Art of War: Sun Tzu (Boston: Shambhala, 1988), p. 19.14. Wei Rulin and Liu Zhongpin, Zhongguojunshi sixiang shi (History of Chinese Military

Thought) (Taipei: Liming Publishing House, 1985), p. 437.15. Georges Tan Eng Bok, "Strategic doctrine," in Gerald Segal and William Tow (eds.),

Chinese Defense Policy (London: Macmillan, 1984), p. 4; Edward S. Boylan, "The Chinesecultural style of warfare," Comparative Strategy, Vol. 3, No. 4 (1982), p. 345.

16. Chong-pin Lin, China's Nuclear Weapons Strategy: Tradition Within Evolution(Lexington MA: Lexington Books, 1988), pp. 31-33.


absence of a notion of ideologically based total war;17 the minimization •of war and the corollary that wars which were not minimized wereethically unjust;18 a "systematic denial of belligerence";19 a view that waris an "aberration";20 and "strategy by strategem"21 as opposed, pre-sumably, to strategy involving the application of violence among othersimilar characterizations.

Scholars have tried to trace the roots of this minimal violence axiom toa number of disparate sources.22 One is Sun Zi's notion of "not fighting andsubduing the enemy" (bu zhan er qu ren zhi bing). Others suggest that itcomes from Lao Zi and his doctrine of using "softness to overcomehardness" (yi mo ke gang). Still others contend that the deprecation ofviolence is rooted in the Confucian-Mencian emphasis on the ruler'scultivation of virtue and good government as the basis for the security andprosperity of the state. When the use of force becomes "unavoidable" then,the literature suggests, Chinese strategic propensities lean to the defensiveand limited use of force. Offensive wars of annihilation were rarely usedhistorically. Chinese political and military leaders generally eschewedexterminating states, occupying territory or killing the people of an enemystate, and prefered security through the enculturation of "barbarians."

Another alleged element of Chinese strategic thought and behaviour isa stress on limited wars, that is, military behaviour constrained bywell-defined spatial and temporal restrictions on violence. China has beenrelatively successful, it is argued, in using limited amounts of force, inco-ordination with diplomatic tools, to pursue clearly defined, limitedpolitical aims. Grand strategic behaviour in the post-1949 period has beencharacterized by defence at or only slightly beyond "the gates," a relativerather than zero-sum concept of victory, pacification and deterrence, anon-zero sum concept of conflict which reduces pressures to escalate, andan ability to preserve a strict hierarchy of political goals in the midst ofconflict, among other traits.23

17. Peter Alexis Boodberg, "The art of war in Ancient China: a study based upon the'Dialogues of Li Duke of Wei'," Ph.D. dissertation, University of California, Berkeley, 1930,pp. xii-xiv.

18. Cleary, Mastering the Art of War, p. 20.19. Mark Edward Lewis, Sanctioned Violence in Early China (Albany: State University

of New York Press, 1990), p. 103.20. Jonathan Adelman and Shih Chih-yu, Symbolic War: The Chinese Use of Force,

1840-1980 (Taipei: Institute of International Relations, National Chengchi University,1993), p. 31.

21. Howard L. Boorman and Scott A. Boorman, "Strategy and national psychology inChina," The Annals, No. 370 (1967), p. 152; Wang Jiandong, Sunzi bingfa sixiang tixijingjie(A Clarification of the Structure of Thinking in Sun Zi's Art of War) (Taipei, 1976), p. 77.

22. For details on the scholarship on traditions of Chinese strategic thought see AlastairIain Johnston, Cultural Realism: Strategic Culture and Grand Strategy in Chinese History(Princeton: Princeton University Press, 1995), pp. 63-65, 117-123.

23. See Chan, "Chinese conflict calculus"; Allen Whiting, The Chinese Calculus ofDeterrence, (Ann Arbor: University of Michigan Press, 1975); Davis Bobrow, "Peking'smilitary calculus," World Politics, Vol. 16, No. 2 (1964), pp. 287-301; Paul Godwin,"Soldiers and statesmen in conflict: Chinese defense and foreign policies in the 1980s," inSamuel S. Kim (ed.), China and the World: Chinese Foreign Policy in the Post-Mao Era(Boulder: Westview Press, 1984), pp. 215-234; Georges Tan Eng Bok, "Strategic doctrine,"in Segal and Tow, Chinese Defense Policy, pp. 3-17; Chong-pin Lin, China's Nuclear

The China Quarterly

There are very few dissenting voices to this characterization of Chinesestrategic traditions.24 Yet new research based on heretofore unavailableprimary materials from the 1950s suggest that China's dispute behaviourhas in some cases been higher risk, more militarized and less connectedto specific limited political demands than was once believed. It has, insome cases, also been less than successful, exposing China to unexpect-edly severe security threats.25 In the spirit of re-examination that thesenew secondary works embody, below is a first cut at some new data thatit is hoped will allow scholars to think more comparatively across timeand space about China's dispute management behaviour.

Descriptive Analysis of Chinese Crisis/Dispute Management Behaviour

It might be best to begin with some of the simplest findings from theMID data set. How frequently has China become involved in militarizeddisputes, relative to other major powers? That is, how dispute-prone hasChina been?26 To what level of hostility or violence has it resorted when

footnote continued

Weapons Strategy; Boorman and Boorman, "Strategy and national psychology in China," pp.143-155; Scott A. Boorman, The Protracted Game: A Weich'i Interpretation of MaoistRevolutionary Strategy (New York: Oxford University Press, 1969), and "Deception inChinese strategy," in William Whitson (ed.), The Military and Political Power in China inthe 1970s (New York: Praeger, 1973), pp. 313-337; and Adelman and Shih Chih-yu, SymbolicWar. They all, in various ways, point to the Chinese exercise of military force as an exampleof a strategic tradition capable of maintaining a rational balance between limited political endsand limited, and generally defensive, military means. Contemporary Chinese analysis of thePRC's strategic behaviour essentially stresses the same characteristics. Indeed, it embodiesan intensely ethno-racialist stereotyping of the Chinese approaches to conflict. It is a politicalmantra in most open Chinese analyses that the Chinese people are a uniquely peace-lovingpeople; that China has historically rarely invaded other states (the exceptions being when itwas ruled by ethnically non-Chinese, such as the Mongol Yuan dynasty or the Manchu Qingdynasty); that the PRC has never occupied "one inch" of another state's territory; that the PRChas never invaded another state except to "teach it a lesson" as part of a just counterattackagainst prior aggression; and that the PRC prefers to use political rather than military meansto resolve disputes.

24. For dissenting views on traditional China see Anthony William Sariti, "A note onforeign policy decision-making in the Northern Sung," Sung Research Newsletter, No. 8(1973), p. 5; Carl-A. Seyschab, "The 36 strategems: orthodoxy against heterodoxy," in C.-A.Seyschab, A. Sievers and S. Synkewicz (eds.), Society, Culture and Patterns of Behavior(Bonn: Horlemann, 1990); and Johnston, Cultural Realism. For a dissenting view of PRCcrisis behaviour see Gerald Segal, Defending China (London: Oxford University Press, 1985).Segal looks at nine cases of the Chinese use of force from 1949 to 1985 and argues that therewere no obvious patterns, that China demonstrated strategic and tactical flexibility and awillingness to use whatever amount of force necessary to achieve a wide range of politicalends. One problem with the study is that it "selects on the dependent variable": that is, byfocusing only on crises in which force was used it is hard to determine the kinds of conditionsthat led or didn't lead to the use of force in the first place. This also makes it difficult to compareChina's dispute behaviour with that of other states.

25. See Thomas J. Christensen, Useful Adversaries: Grand Strategy, Domestic Mobiliza-tion and Sino-American Conflict 1947-1958 (Princeton: Princeton University Press, 1996);Chen Jian, China's Road to the Korean War (New York: Columbia University Press, 1994).

26. I want to emphasize that dispute proneness says nothing in particular about intentions,that is, it does not equal "aggressiveness," nor does it say anything about whether or not Chinainitiated these disputes. It is simply a measure of the number of disputes that China has becomeinvolved in, regardless of which state started the dispute. Thus these data should not beinterpreted as "China is the second most aggressive major power in the international system"or something such as that.


Table 1: Comparative Frequency of MIDs during the Cold War

State

U.S.ChinaUKIndiaUSSRFrance

Period

1946-921949-921946-921947-921946-921946-92

MIDs/year

3.932.741.891.871.720.94

Table 2: Comparative Frequency of MIDs

State

USSRChinaIndiaU.S.UKFrance

Period

1918-19921949-19921947-19921815-19921815-19921815-1992

MIDs/year

3.222.741.871.751.440.94

in a dispute? What kinds of force has it tended to use? What have beenthe primary political goals in these disputes?

Frequency. To begin with, China has been involved in an average of2.74 new MIDs per year from the founding of the People's Republic to1992.27 This statistic ranks China as the second most dispute-prone stateof the major powers in the post-Second World War period, behind theUnited States (see Table 1). If one looks at dispute proneness for themajor powers over their history as states, China still ranks second, thoughthis time behind the Soviet Union (see Table 2). Disaggregating further- into decade-long periods - China was the most dispute-prone majorpower in the 1950s, though the United States was a very close second(see Figure 2).28 But for the remaining decades China was consistently thesecond most dispute-prone state until 1990 (India's level of disputeproneness reaches that of China for the 1980s).29 The gap betweenChinese dispute proneness and that of the first place holder - the UnitedStates - increases over time.

27. This datum refers to the number of MIDs that start in that particular year, not the totalnumber that start and are ongoing from earlier years.

28. The vertical axis is the total number of disputes per country in each decade.29. The average number of MIDs per year drops dramatically in the 1990s to 0.3 for the

PRC, the lowest of all the major powers plus India. France replaces China as the second mostdispute-prone major power from 1990 to 1993 after the U.S. with an average of 2 MIDs peryear, compared to 2.6 for the U.S. Since there are only three disputes in this decade for China,however, these data should not be given too much weight as indicative of trends for the restof the decade.


Figure 2: Relative Frequency of MIDs per Decade

1950s 1960s 1970s

decade

1980s

Figure 3: Chinese MID Frequencies per Five Year Period

25

N

10

O)

ooIT)

4m

cocoO)inO)

COCO

4co

cor»O)COa>

CO

4cocoa>

O)

0000

4coa>

five year period

Figures 3 and 4 break these data further into frequencies per five-yearperiod and cumulative frequencies, respectively. Both figures show, forinstance, that China was involved in most of its MIDs prior to the end ofthe Cultural Revolution (the 1969-73 period). The rate of increase in thecumulative sum of disputes (Figure 4) slows after around 1969. About 66per cent of Chinese MIDs occurred before the 1969-73 period. The twomost dispute-prone five year periods were 1954-58 (an average of 5.5MIDs began in each year in this period) and 1964-68. While the yearlyfrequency of MIDs declined fairly dramatically after the Cultural Revol-ution, it remained relatively constant from the mid-1970s to the late


Figure 4: Cumulative Sum of Chinese MIDs

120

100

11

I I I I I I I I I I I I I 1 I I I I I

year

1980s. In other words, the Dengist period until the late 1980s was notnoticeably less MID-prone than the late Maoist period, though it was lessdispute-prone than most of the rest of the Maoist period.

Type. As for the primary political goals in China's disputes, the largestportion of Chinese MIDs were classified by the Correlates of War teamas "territorial" (type 1) (49 per cent), with those coded "policy" (type 2)the second most frequent (42.3 per cent) and those coded "regime"

Figure 5: Frequency of Chinese MIDs by Dispute Type per Five YearPeriod

CO

m

1949-

co

1954-

coCO

1959-

coco

1964-

five

CO

1969-

year

COf -

1974-

period

CO

co

1979-

CO

co

1984-

O)

1989-


Figure 6: Average Hostility Score per MID in Each Five Year Period

(type 3) the least frequent (7 per cent).30 The lion's share of territorialMIDs (41 per cent) occurred in the first ten years of the People'sRepublic of China (PRC) (see Figure 5). As suggested below, these dataare consistent with the hypothesis that new states will be more sensitiveabout establishing territorial control, and thus one should expect theprevalence of territorial MIDs earlier in the regime's history.

Force levels. The MID data use two indicators to represent the levelsof force reached by a state in a dispute: a five-point hostility level scaleand a scale comprising 21 categories of action.31 Using hostility levelsfirst, the most violent periods were at the beginning of the regime(because of the Korean War) and in the mid-1980s (because of ongoingconflicts with Vietnam).32 The average hostility level per dispute in the

30. Most of those coded "regime" involved disputes with Taiwan in the mid-1960s. Thusone could plausibly recode these as territorial disputes, as the Taiwan issue was in some sensea dispute over sovereign control of the island. Moreover, Chinese uses of force against Taiwanwere not designed to overthrow the Kuomintang since the PLA was simply incapable of doingso. Finally, Chinese military conflicts with KMT forces would not have occurred had Chinanot claimed Taiwan as Chinese territory. The number of disputes coded "regime" is small,however, and does not change the analysis dramatically.

31. The hostility levels are: no militarized response (1), threat of force (2), display of force(3), use of force (4), interstate war (5). The categories of actions are: threat to use force (1),to blockade (2), to occupy territory (3), to declare war (4), to use nuclear weapons (5), alert(6), mobilization (7), show of troops (8), show of ships (9), show of planes (10), fortificationof border (11), nuclear alert (12), border violation (13), blockade (14), occupation of territory(15), seizure of material or personnel (16), clash (17), other use of military force (18),declaration of war (19), tactical use of nuclear weapons (20), interstate war (21).

32. Although these are categorical data and technically ought not to be averaged, since theydo reflect an interval-like increase in violence I did not see much danger in averaging them.


Figure 7: Level of Hostility Reached by China in Relation to the LevelsReached by its Adversaries in the Same MID

13

0

SAME HIGHER

relative hostility

LOWER

1949-53 period was around level 4 (actions entailing "the use of force"),while the average level in the 1984-88 period was just under 4 (seeFigure 6). The lowest average hostility level was in the 1969-73 periodwhen China - essentially recovering from the Cultural Revolution andopening a strategic relationship with the United States - was in one of itsleast dispute-prone periods. Average hostility levels per dispute droppedto just over level 2 (actions entailing the "threat to use force"). With theexception of the Cultural Revolution period - when Chinese attentionswere focused on internal political struggle - there is a relatively constantaverage hostility in Chinese disputes throughout the entire post-1949period.

In comparative terms the average level of hostility reached by the PRCper dispute was greater than that of India, France, Britain, the UnitedStates and the Soviet Union from 1945 onwards. China's average hostil-ity score in the post-1949 period was 3.71, statistically significantlyhigher than the average among these powers (3.53).33 The next highestmeans were France, India, the Soviet Union, Britain and the UnitedStates, with only the United States being statistically significantly lowerthan the mean.34

In terms of the 21 categories of militarized action, as a proportion ofall codable disputes China's top three choices of action were "clashes"

33. A difference of means test was significant at the p = 0.012 level, meaning that China'shigher-than-average hostility levels is unlikely to be random.

34. The difference of means test was significant at the p = 0.0 level, meaning that theUnited States' lower-than-average difference in average hostility levels is unlikely to berandom.


Figure 8: Level of Hostility Reached by China in Relation to LevelReached by Selected Adversaries in Same MID

PRC relative hostility levels

HSAME

• HIGHER

• LJCWER

USA USR India SRV

state

(26.3 per cent), "other force" (17.8 per cent) and "seizure" (11 per cent).In most of its disputes China also tended to escalate to a level of force atthe same level of hostility as that used by its adversaries in the samedispute (Figure 7).35 In about 20 per cent of its disputes China wouldescalate to a higher level of hostility than its adversary, and in less than20 per cent of cases to a lower level. Disaggregating this pattern,however, one finds some variation: in nearly half its disputes with theUnited States, China tended to escalate to a higher level of hostility thanthe United States did in the same dispute (Figure 8).

The five-point hostility scale is useful as a first cut at examiningChina's escalation patterns. But it is still a fairly blunt measure ofhostility: level 4 (the use of force), for example, cannot differentiatebetween limited exchanges of fire across a border and the more seriousacts of seizing a piece of territory or declaring war. A more fine-tunedviolence score can be constructed by multiplying the action code numberand the hostility level number.36 This creates an interval-type scale thatcaptures the exponential nature of increasingly violent uses of force indisputes. The average violence score per MID for the PRC was 61.34,again statistically significantly higher than the average for the other statesincluded in the analysis. The sample mean was 53.085. The next highestmeans were France (58), India (57.3), the Soviet Union (49.88), Britain

35. Per cent refers to the percentage of China's MIDs in which China escalated to a levelsimilar to, higher than, or less than its opponent.

36. The scale captures the vast difference between very low levels of threat to use force(e.g. a threat to blockade would have a violence score of 4 (action code 2 X hostility level 2))versus the use of weapons of mass destruction in all-out interstate war (action code21 X hostility level 5 = violence level 105). The actual numerical boundaries or values of thismetric have no intrinsic meaning other than that they allow one to differentiate statisticallyand more accurately between very low levels and very high levels of violence. This allowsmore fine-tuned comparisons within and across states in their dispute behaviour.


Figure 9: Major Powers' Mean Violence Score per MID

100-

80.

oo60.

I 40.o

20.

China France India sample USR UKmean

US

15

(47.65) and the United States (44.35). Again only the United States meanis statistically lower than the sample mean (see Figure 9).37

These statistics, in combination with the MID frequency statistics,suggest that until 1992, although China was not the most dispute-prone ofthe major powers, it was the most violence-prone. That is, once in amilitarized dispute, on average China tended to resort to a higher level offorce than other major powers.

MID type and violence levels. The data suggest that China wasrelatively consistent in the kinds of actions it took regardless of the typeof conflict. There is little correlation between dispute type and hostilitylevel (see Figure 10). In about 80 per cent of MIDs coded as policy andregime-related disputes, China resorted to hostility level 4 (the use offorce). It was slightly less likely to resort to force in territorial disputes,but still in about 65 per cent of these cases it escalated to hostility level4. It did not go to interstate war over regime/government MIDs, meaningthat it never resorted to war to overthrow a regime in another state.

This pattern holds as well when one uses the interval-scale violencescores as an indicator for level of force used in a dispute. As Figure 11indicates, there is not as much variation in average violence scores acrossdispute types. China's violence scores tend to cluster together at about thesame level across all three types of conflict. France's violence profile isthe most similar to that of China's. This contrasts with most of the other

37. For China, t value = 3.23 and is significant at the p = 0.002 level, meaning that China'sscore deviates upward from the mean in a way that is unlikely to be random. For the U.S.,t value = - 4.1 and is significant at the p = 0.000 level, meaning that the U.S. score deviatesbelow the mean in a way that is unlikely to be random.


Figure 10: Chinese MID Hostility Level by Dispute Type (Type1 = Territory; Type 2 = Policy; Type 3 = Regime)

type 1 %type 2 %

• type 3 %

use offorce

hostility level

Figure 11: Comparative Violence Scores by Dispute Type

120 i

China India

comparative cases where there is more skewing. The United States, forinstance, was far more likely to act more violently in policy-related MIDsthan in territorial MIDs. The Soviet Union was far more likely to actmore violently in regime-related MIDs, as was India (though in India'scase the N = 1). As the data indicate, however, China was more likely toact more violently in territorial and policy MIDs (types 1 and 2) than theother states compared here.

If one refines the analysis and looks at the 21-category action scale(Figure 12), the most common form of action taken by China in itsterritorial disputes was overwhelmingly the military "clash" (48 per centof cases). This was followed well behind by "other force" (9.7 per centof cases) and border violations (also 9.7 per cent). In MIDs relating to


Figure 12: Chinese MID Action Level by Dispute Type

17

action level

efforts to change other state's policies, China's behaviour tended to bedistributed more evenly among seizures (31 per cent), clashes (24 percent) and "other force" (21 per cent). Finally, in MIDs aimed at changingother state's governments, the most common form of violence was "otherforce" (60 per cent) followed by border violations (20 per cent) andclashes (20 per cent). Since most of these MIDs were conflicts withTaiwan, "other force" probably refers to things such as artillery barrages,one of the key forms of violence China used in the various disputes withTaiwan over the islands off the East China coast.

The data suggest, then, that in general China has been more dispute-prone than many other major powers (except the United States); it hasalso been more likely than most other major powers to resort to higherlevels of force when in a militarized dispute, regardless of the type ofdispute. While the first 15 years of its existence were more dispute-pronethan the subsequent 25, there has been a fairly constant level of hostilityand violence across Chinese MIDs up to the end of the 1980s. Whileaverage violence scores were quite consistent across all types of disputes,China tended to resort to the highest scale of military action ("clashes")in territorial disputes.

Explanations

There is a range of plausible hypotheses that might explain the generalfeatures of China's dispute management behaviour. Below are analyses ofsome of the standard explanations in the international relations literatureto see how well they account for the patterns in the data, and to suggestwhich explanations can probably be put aside and which are worthpursuing further.


Figure 13: Frequency of Chinese MIDs and Domestic Protests

1

1

I

1

N

6

4 -

2 -

o -8 -

6 -

4 -

2 -A

Kc1949

\l V

1951

1953

1955

1957

l\/\•i

1959

MIDs/yearprotests

hh. A l \A A

A A^/\ /\ WV A/ vft/\_7"\^

i i i / V t i V i i i i i i

1961

1963

1965

1967 •

1969

1971 -

1973 -

1975 -

1977 -

1979 *

1981 -

year

Diversionary theories of conflict and crisis behaviour. One generalhypothesis is that states will deliberately create or escalate externalconflicts, disputes and crises for domestic political purposes. As Levy hasshown, there are a number of related variations on this theme:38 externalcrises will be manufactured to divert domestic attention from the policyfailures of the regime; crises will be used to consolidate the legitimacy ofthe regime as it portrays itself as the main bulwark against a dangerousexternal adversary; crises will be used to justify the repression of internalopposition; external crises will be used to justify the mobilization ofresources for political or economic strategies.

How well do various diversionary explanations work in the Chinacase? Christensen has argued convincingly that the Jinmen and Mazucrisis of 1958 was designed primarily to create an atmosphere of threatwith which to mobilize the population for the radical Great Leap Forwardexperiment.39 In the aggregate, however, the data do not appear to fit withmost of the diversionary war hypotheses.

For one thing, there is no relationship between domestic unrest andChina's use of force externally. As a first cut, the frequency of protestswas used as an indicator of domestic unrest;40 as the graph of annualfrequencies in MIDs and protests shows (Figure 13), the frequency ofprotests basically picks up as the frequency of MIDs declines from thelate 1960s onwards. This is not the pattern one would expect if diversion-ary uses of force were being designed to suppress domestic unrest. It ispossible that the MIDs in the 1950s and 1960s were successful in

38. Jack S. Levy, "The diversionary theory of war: a critique," in Manus I. Midlarsky (ed.),The Handbook of War Studies (Boston: Unwin Hyman, 1989).

39. Christensen, Useful Adversaries, ch. 6.40. Protest data were taken from World Handbook of Political and Social Indicators III,

1948-1982. The data in the graph stop in 1982 because this is where World Handbook datastop.

China's Militarized Interstate Dispute Behaviour 1949-1992 19

Table 3: Effect of Domestic Unrest on Cumulative MID Violence Scoreper Year (OLS regression)

Variable Beta t Sign

Sanctions ( t - 1 ) 2.02 2.73 0.01Riots ( t - 1 ) - 3 . 6 6 - 2 . 0 1 0.05Protests ( t - 1 ) -12 .12 - 2 . 6 0.02Executions ( t - 1 ) - 1.8E-04 - 2 . 1 1 0.04Constant 165.57 7.5 0.00

Notes:R2 = 38; adjusted R2 = 0.29; standard error = 88.8; F = 4.13;

sign = 0.01.

diverting attention and thus protests occurred with less frequency in theseyears. But even so, one would at least expect to see some rise in protestsprior to increases in disputes during the 1950s and 1960s.

The diversionary hypothesis is further undermined by a test for anyrelationship between the amount of domestic unrest and the amount ofexternally directed violence. The theory would suggest, at its simplest,that an increase in domestic unrest in one period should lead to anincrease in the level of conflict China is engaged in with other states. Totest this hypothesis a simple linear regression model was constructed. Thedependent variable in the model is the cumulative yearly amount ofviolence (the sum of the violence scores for all disputes in that year). Theindependent variables are riots, protests, executions and sanctions. Thefrequency of riots and protests captures popular expression of opposition.The frequency of executions and sanctions can be used as a proxy for theregime's concerns about domestic opposition. Together, these indepen-dent variables capture domestic unrest.41

The results of the test are in Table 3. Except for sanctions, theindicators of domestic turmoil are negatively related to an increase inviolence. In other words, contrary to a hypothesis about the diversionaryuse of violence, in the China case an increase in domestic unrest leads toa decrease in the yearly amount of MID-related violence.42 This suggestsa "pre-occupation" model is more appropriate than a diversionary model,that is, a preoccupation with domestic social disorder is associated witha lower level of externally directed violence. If one uses the same data totest the reverse hypothesis, namely that external violence in one periodleads to the decline of domestic unrest in subsequent periods-a"suppression model" - there is no statistically significant relationship

41. The independent variables are lagged by one year, as one might expect if conflict isbeing used to divert domestic attention. The data come from the World Handbook of Politicaland Social Indicators.

42. The adjusted R2 = 0.29. That is, almost 30% of the variance in the dependent variableis accounted for by the independent variables. All the independent variables are statisticallysignificant below the standard 0.05 level.


between either MID frequencies or cumulative yearly amounts of viol-ence on the one hand and indicators of domestic unrest on the other.

The foreign policy of revolutionary states. A second set of hypothesesdraws from work on the foreign policies of revolutionary states. Herethere are two basic arguments. The first is that revolutionary states tendto be led, initially at least, by people with messianic or manichean worldviews. Whether for purposes of mobilizing (or terrorizing) populations, orjustifying militarized revolutionary behaviour with militant symbols andmyths, revolutionary movements tend to understand the external world asone full of adversaries and threats. Foreign policy will tend to reflect thisstrategic culture. This world view will push revolutionary states into morefrequent crises and escalation.

The second argument is somewhat the reverse: because revolutionarystates at least appear to threaten regional and at times the global politicaland economic status quo, there are pre-emptive pressures on status quostates to deal with this potential threat early before the revolutionarystate(s) becomes too powerful. This will lead to counter-revolutionaryalliances and foreign policies among surrounding states, pulling revol-utionary states into more conflicts with other states on average.43

In the Chinese case, then, there ought to be a correlation betweenperiods of heightened revolutionary militancy on the one hand andconflict involvement and/or levels of violence on the other. There is someevidence - based on newly available Russian and Chinese documents- that Mao's decision to enter the Korean War and to support thenationalist revolution in northern Vietnam in the 1950s was motivated bya desire to expand the influence of Communist parties in East Asia.44

Scholars have also argued, of course, that American involvement inVietnam-a source of many Chinese MIDs in the mid-1960s-was inpart driven by a fear of Chinese expansionism and by a desire to provewrong the Maoist model of rural revolution. These examples wouldsuggest both a "push" and a "pull" correlation between revolutionarymilitancy and several kinds of militarized disputes.

Yet the relationship is more ambiguous than this. During the heightof Maoist revolutionary fervour in the Cultural Revolution the frequencyand average violence scores of China's MIDs were generally much lowerthan before or after this period. If one uses a simpler indicator ofthe presence of revolutionary ideology - say, the duration of Mao'srule - one finds a fairly dramatic variation in the frequency and violencelevels of disputes within the Maoist period. Moreover, there is relatively

43. For a helpful summary of various push and pull hypotheses linking revolutionary statesto war proneness, see Stephen M. Walt, "Revolution and war," World Politics, Vol. 23, No.3 (April 1992), pp. 323-333. Walt adheres to the pull arguments.

44. See the arguments made by Chen Jian, China's Road to the Korean War (New York:Columbia University Press, 1994); Chen Jian, "China's involvement in the Vietnam War,1964-1969," The China Quarterly, No. 142 (June 1995), pp. 356-387. Walt's coding ofChinese intervention in the Korean War - as an instance where a revolutionary state wasdragged reluctantly into a conflict started by others - is suspect in the light of the new data.Thus his dismissal of the revolutionary push argument is less convincing.


little variation in the frequency and average violence of disputes acrossthe late Maoist period to the last years of the Dengist period. Yet Deng'sChina has been identified with the wholesale abandonment of most ofMao's ideological legacies.

As an alternative test, one might plausibly hypothesize that revolution-ary states tend to militarize their societies and economies: the moremilitarized the society, the more militant the politics, and thus the morelikely the state will become involved in MIDs, whether for push or forpull reasons. Is this the case in China? A series of simple bivariate linearregression analyses indicates the relationship is in fact ambiguous. Forthis test the share of China's total world power capabilities comprisingmilitary expenditures and military personnel were used as indicators ofmilitarization.45 That is, the greater China's share of power attributable toits military expenditures and military personnel, the more militarized theeconomy and society; and the more militarized the economy and society,the more militant the polity. The more militant the polity, the greaterChina's dispute proneness.

As it turns out, however, this relationship is not especially strong.There is no positive relationship between the increasing share of totalChinese power capabilities devoted to military expenditures on the onehand and MID involvement on the other (adjusted R2 = 0, p = 0.9). Noris there a statistically significant relationship between the proportion ofChinese power capabilities comprising military personnel and disputeproneness (adjusted R2 = 0.04, p = 0.12).

However, when disaggregated into disputes with the Soviets andAmericans, there is a relatively strong positive relationship in the case ofChina's MIDs with the Soviet Union, but no relationship in the case ofMIDs with the United States.46 This finding fits the nature of China'sconflict with the Soviet Union. While ideological competition betweenCommunism and capitalism was obviously an element in Chinese con-ceptions of Sino—U.S. conflict in the 1950s, the geopolitical conflicts ofinterest were perhaps more prominent in the Chinese calculus. Americanpower threatened Chinese domestic stability and territory, whether inKorea, Taiwan or Vietnam. Thus Maoist militancy was probably not the

45. The data for share of world power come from the Correlates of War Capabilities DataSet (World Base). The indicators of militarization are "per cent of per cent" data. TheCorrelates of War data on a state's total share of power capabilities uses a basket of sixindicators (military expenditures, military personnel, iron and steel production, energyconsumption, urban population, and total population). Each indicator comprises a share of thetotal capabilities of a state. Per cent of per cent refers to the proportion that each indicatorcomprises of the total basket, with the total basket being a proportion of total worldcapabilities. Thus as the per cent of per cent figure for military expenditures or militarypersonnel increases this simply means that these indicators comprise a larger portion of thetotal basket of capabilities. This changing portion shows the degree to which a state's powerrelies more or less on military expenditures and military personnel. 1 use an increasing portionof power comprised of military expenditures as a surrogate for the militarization of theeconomy. I use an increasing portion of power comprised of military personnel as a surrogatefor the militarization of society.

46. For military expenditures the adjusted R2 = 0.24, p = 0.001; and for military personnelthe adjusted R2 = 0.08, p = 0.05. The figures for the U.S. case are not statistically significant.


driving element in China's conflicts with the United States in the 1950sand 1960s. The ideological component of the Sino-Soviet dispute, how-ever, was prominent throughout. Territorial and geopolitical competitionwere products, to some extent, of the collapse in shared identities asMarxist-Leninist states. The Chinese accusations of Soviet betrayal ofsocialism and Soviet charges that Maoists were out to subvert theSoviet-led socialist bloc came in the mid-1960s to the mid-1970s, periodsof ideological fervour and/or militarization within China. These were alsoperiods when China's militarized disputes and border clashes with theSoviets were most intense. Thus the different findings in China's disputeswith the Americans and with the Soviets may reflect different core issuesin the conflicts with each superpower.

Foreign policy balancing. Another possible explanation for patterns inChina's dispute behaviour draws from neo-realist theory. Neo-realistsmight argue that China's militarized behaviour reflects the frictionsgenerated by its efforts to balance against the dominant player(s) in theinternational system, given the threat posed to the security and autonomyof China from a system dominated by another power. As long as Chinais a weak major power in a bipolar system, its primary security concernsshould come from the two poles - the United States or Soviet Union.Thus most of China's disputes should be with one of the two contendersfor system hegemony, depending on which is predominant.

The evidence, on the whole, does not buttress this kind of balancingargument. To test this hypothesis a series of simple linear regressionequations were run testing the relationship between superpower power onthe one hand and various features of Chinese dispute behaviour on theother. There is certainly a positive and significant relationship betweenthe frequency of China's disputes with the United States and the UnitedStates' share of global military expenditures (adjusted R2 = 0.12,p = 0.02).47 This would be consistent with a neo-realist balancing argu-ment. As American global power peaked in the 1950s and 1960s, ChineseMIDs with the United States also peaked. As the American share ofworld military power declined in the mid to late 1970s and through the1980s, China had no MIDs with the United States. This was, of course,the period of Sino-U.S. anti-Soviet collaboration. One would expect,then, that as the Soviet share of world power increased, China wouldbegin to balance against the Soviets and one should expect to see asimilar relationship between Chinese MIDs and Soviet power. However,the results of the regression analysis for the Soviet case shows norelationship between changes in the Soviet share of world power and the

47. Neorealists are, unfortunately, unclear about how to measure power. But in order tobe consistent with their materialist ontology, they should concede that for research purposespower can be measured using material capabilities, military expenditures for instance. Thusfor the purposes of testing this balancing hypothesis, I use Soviet and American shares ofworld military expenditures as the independent variables. If one were to use some indicatorfor "perceptions" of power, then one would not be making a neorealist argument.


frequency of Chinese MIDs with the Soviet Union (adjusted R2 = - 0.03,p = 0.97).

If one looks at the proportion of China's MIDs with the superpowers(assuming that as a major power China is compelled to balance againstpotential systemic hegemonies and thus will be less pre-occupied withconflicts with smaller states), the findings are similarly problematic forneo-realist theory. There is a weak, though significant, positive relation-ship between change in American share of world power and the portionof China's total MIDs that are with the United States (adjusted R2 = 0.08,p = 0.05). There is no statistically significant relationship in the Sino-Soviet case (adjusted R2 = 0.01, p = 0.4), however.

Figure 14: Cumulative Sum of Chinese MIDs (Total MIDs and MIDswith Superpowers)

140,

• Running Sum of MIDs w/USAA Running Sum of MIDs w/USR• Running Sum of total PRC MIDs

1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995

year

Indeed, over time, disputes with the superpowers as a portion ofChina's total disputes has declined even as the international systemremained bipolar. Figure 14 indicates that the rate of cumulation ofdisputes with the superpowers slows relative to the overall cumulationrate from the early 1970s onwards. That is, the superpowers have notbeen the only interlocutors in Chinese MIDs. If one of China's basicstrategic decision rules were to balance primarily against the dominantstate(s) in the international system, one would expect that Chinese MIDswith the two superpowers - as a portion of its total MIDs - should berelatively constant. Yet it is clear from Figure 14 that from around 1980to 1992 this portion drops to near zero.


Figure 15: Cumulative Sum of Chinese MIDs by Dispute Type per FiveYear Period

35 "

30 •

25 -

20 "

1 5 •

10 "

5 (

n »I

c1

94

9-5

3)

/y

X

1954

-58

/

X

1959

-63

X19

64-6

8

five

X

1969

-73

year

>—

X

I

1974

-78

period

X

1979

-83

^ ^

^ ~

X

1984

-88

—1

)

JI

C

1989

-91

J

—O—territory—•—policy

X regime/gov

Territorial consolidation hypotheses. As John Vasquez notes, territorialdisputes tend to be one of the most common features of interstate wars.48

There are at least three arguments about territory that suggest why Chinashould be especially prone to disputes. First, disputes over land areessentially zero sum. Contiguous states, therefore, are more likely to fightwith each other than non-contiguous ones, controlling for distance andtechnologies of power projection. It follows, then, that states with moreborders will be more dispute-prone than those with fewer borders, onaverage. The PRC has had land and ocean borders with anywhere from15 to 21 states during its existence as a state, more than most. Secondly,new states are likely to be more sensitive to territorial issues because theterritorial integrity of the state will be central to establishing the legiti-macy of the regime. Thirdly, prospect theory might apply here as well:states are more likely to take risks defending what they have thanacquiring what they seek to gain. Together, these three arguments suggestthat China, as a new state with lots of borders, will be more risk acceptantabout consolidating its territorial integrity. If this proposition is generallyright, then one ought to find that most of China's disputes in the earlyyears of the regime were territorial and that the largest portion of allterritorial disputes occurred in the earlier years of the regime.

The data are roughly consistent with these hypotheses. Almost half ofChina's MIDs were related to territorial issues, and 41 per cent of theseterritorial MIDs occurred in the first ten years of the regime. In addition,the cumulative sum of different MID types (Figure 15) indicates the rateof increase in territorial disputes is faster than other kinds of disputes in

48. John Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993); JohnVasquez, "Why do neighbors fight: proximity, interaction, territoriality," Journal of PeaceResearch, Vol. 32, No. 2 (August 1995), pp. 277-294.


Figure 16: Cumulative Sum of Soviet MIDs by Dispute Type per FiveYear Period

25

oo coT- CMcn en

five year period

the first five-year period of the regime. Finally, there is a significantpositive correlation between the yearly percentage of borders thatremained unresolved on the one hand and the yearly cumulative violencescore on the other (adjusted R2 = 0.19, p = 0.003).49 Put differently, as thepercentage of unresolved borders declined, the yearly cumulative MIDviolence decreased.

Similar patterns show up in other states. For Soviet MIDs, for instance,in the first five-year period after the founding of the state in 1918territorial disputes comprised the largest portion of all MIDs. Cumula-tively, territorial MIDs were also the predominant type of dispute in thefirst three five-year periods (Figure 16). In India (Figure 17), territorialdisputes comprised an even larger proportion of its early MIDs than ineither the Chinese or the Soviet cases. As in China, and to some extentthe Soviet Union, the rate of increase in Indian territorial MIDs peakedin the first two to three five-year periods after the founding of the state.

Status gaps and foreign policy. A fifth explanation draws on statusinconsistency theory.50 Here the argument would be that in periods where

49. The percentage of unresolved borders was determined by the year in which a formalborder treaty with another state was signed. Thus as the percentage of borders determined bya treaty increased, the percentage of unresolved borders declined.

50. See Yaacov Vertzberger, The World in Their Minds: Information Processing,Cognition and Perception in Foreign Policy Decisionmaking (Stanford: Stanford UniversityPress, 1990), pp. 282-295.


Figure 17: Cumulative Sum of Indian MIDs by Dispute Type per FiveYear Period

25 -

20

15 -

10

5

0 ft

five year period

the perceived gap between the status desired by an actor and the statusascribed or bestowed by other actors was largest one should expect to seea higher frequency of conflictual behaviour. The argument is a basicfrustration-aggression one: status inconsistent states will tend to believethat the reason for the lack of "respect" accorded to them is a result oftheir insufficient material power and their insufficient willingness todemonstrate this power. That the strategies chosen to close the status gapshould be coercive implies that states have internalized realpolitik worldviews where relative material power is equated with relative status. Thatis, the more powerful they are the more status they believe will beascribed to them. It is known that Mao accepted this correlation and therealpolitik causal argument that supported it. Indeed, in Maoist Chinastatus inconsistency was a major issue in the Sino-Soviet dispute. By thelate 1950s, Mao was utterly disillusioned with the perceived lack ofrespect that China's interests were being accorded in Moscow. Therecords of his conversations with Khrushchev in 1958 when the Russianleader tried to sell him a proposal to set up a joint Sino-Soviet navy andto allow the Soviets to build a submarine communications base in Chinaindicated how little Mao believed the Soviets were acting in a spirit ofmutual equality. Christensen has found evidence that Mao's manufactur-ing of the Jinmen-Mazu crisis in the summer of 1958 was in large partmotivated by a desire to use an external danger to mobilize the Chinesepeople to increase industrial output which could be then turned into


military power. Once this had been done, Mao remarked, the Americans(and Soviets) would have to take Chinese interests into account.51

One should therefore expect that periods where MID frequencies orhostility and violence levels are highest should be precisely those periodswhere Chinese leaders are most dissatisfied with their international status.How one might assess this dissatisfaction with status quo, however, is notobvious. As an initial effort to develop a measure, the Correlates of Warfigures for national share of material capabilities were used.52 As thisfigure declines it can be assumed Chinese leaders are more dissatisfiedwith their international status and are therefore more conflict-prone. Asthe figure increases the leaders ought to be more satisfied with China'srelative status. A bivariate OLS regression indicates that there is indeeda significant negative relationship between change in China's relativematerial power on the one hand and the frequency of MIDs (adjustedR2 = 0.1, p = 0.03) and the cumulative amount of MID violence per year(adjusted R2 = 0.16, p = 0.009) on the other.

These findings would suggest, then, that China's dispute proneness ispositively related to the gap in relative power between China and othermajor powers in the system. As this gap has closed, as Chinese leadershave become less dissatisfied with extant status inconsistencies, the fre-quency of MIDs has declined. This suggests that China either has tendednot to exploit its growing share of world power to pursue interestscoercively, or that other states have tended not to challenge China as itspower has grown, or both. It is more likely to be the former, however,since this negative relationship between the growth of China's relativepower and its MID involvement holds even more strongly in China'srelations with the United States (adjusted R2 = 0.19, p = 0.005). This is arelationship where absolute American power has always exceeded that ofChina. Thus even as Chinese power has grown relative to that of othermajor powers it has become less involved in MIDs with the United States,even though American power has always been sufficient to be undeterredby China. That is, it is probably not the case that as China's relative powerhas grown the United States has been less willing or able to challengeChina. Rather, even as China's relative power has increased it has beenless likely to challenge the United States. This would be consistent witha status inconsistency explanation for China's dispute involvement.

Conclusion

This study has provided a preliminary analysis of new data on Chinesedispute behaviour. In comparative terms, during the Cold War period

51. "Minutes, conversation between Mao Zedong and Ambassador Yudin, 22 July 1958,"in Zhang Shuguang and Chen Jian, "The emerging disputes between Beijing and Moscow,"Cold War International History Project Bulletin, Nos. 6-7 (1995/96), pp. 155-59. See alsoChristensen, Useful Adversaries.

52. Here I used China's share of the total power of all major powers, rather than all statesin the system. My reasoning was simple: China tended to compare itself with other majorpowers. In principle, therefore, its relative power position among major powers should mattermore to its perception of status than its relative power position among all other states, largeand small.


China was more dispute-prone than most other major powers except forthe United States. China also tended to resort to higher levels of violencein disputes than did other major powers and India. In other words, whileChina was not the most likely of the major powers to fall into crises andmilitarized conflicts, it was more likely to use a higher level of violencethan other states in such a dispute. These data do not sit comfortably withmuch of the existing Western or Chinese literature on the uses of forcein PRC history. I do not claim that the findings should thus supersede allprevious scholarship on the question. I have used very different data anddifferent analytical methods with different problems of inference. Thereis still a need to carefully calibrate the findings in the qualitative andquantitative literature: there is plenty of room for studies that explorecases where the different data and methods either confirm or questioneach other. But these data do suggest that there should be some scepti-cism towards the conventional wisdom. They also raise some interestinghypotheses about the centrality of a sensitivity to territorial integrity andinternational status in Chinese dispute management behaviour, hypoth-eses that deserve to be put on the research agenda of Chinese foreignpolicy studies.53

Before advocates for the neo-containment of China abuse these data foralarmist and ethnocentric claims about China's aggressive intentions,however, the tests of various explanations for patterns in the dataunderscore two important caveats. First, the largest portion of Chinesedispute behaviour has involved territorial issues and the consolidation oflong-standing territorial claims. This partially explains why the annualfrequency of Chinese militarized disputes has declined over time: as theregime consolidated power domestically, it also tried to assert controlover its boundaries, either through negotiated border settlement or force.This should not be surprising: control over its borders and a demonstratedcapacity to ensure territorial integrity are critical pillars for the internaland external legitimacy of a sovereign state.

Secondly, the growth of China's relative power capabilities by itselfhas not led to an increase in Chinese dispute proneness. Rather, asChina's share of world power resources has increased - and by extensionas the perceived gap between ascribed and desired international status hasclosed - China has tended not to act in a more confiictual manner. Thefrequency of MIDs, for instance, has not increased appreciably over the1980s even as China's economic and military power has developed. Thissuggests, very tentatively, that excessive fears of a power transition-typeclash between China and the United States, both vying for hegemonicstatus globally or in the Asia-Pacific regional subsystem, may not be

53. There is some work already emerging on patterns in China's diplomacy on territorialissues. See, for instance, Eric Hyer, "The South China Sea disputes: implications of China'searlier territorial settlements," Pacific Affairs, Vol. 68, No. 1 (Spring 1995), pp. 34-54. Seealso Jean-Marc Blanchard, "Borders and Borderlands: an institutional approach to territorialdisputes in Asia," Ph.D. thesis, University of Pennsylvania, Political Science Department,1997. The sources of China's concern with international status and image, and the trade-offsChinese leaders make between maximizing status and maximizing other goals such asdevelopment and security is, however, still a severely neglected topic.


warranted.54 Taken together, these two conclusions suggest that Chinawill be more likely to resort to force - and relatively high levels offorce - when disputes involve territory and occur in periods where theperceived gap between desired and ascribed status is growing or large.

The findings of the study suggest, therefore, that the growth in Chinesepower capabilities does not necessarily portend a more aggressive use ofChinese power, or more dispute involvement. Indeed, there may be less,as long as China's territorial integrity is not challenged and as long as itis accorded sufficient international status (such as involvement in allmajor international institutions and full partnership in major power actionon global order issues).

But the findings do suggest that once in a militarized dispute China willtend to escalate to a relatively high level of force. With doctrinal changesin recent years that stress the offensive, even pre-emptive, use of militarypower,55 and in the absence of alternative forms of crisis management, thepreference may well be to use this force in a militarily offensive mannerand further beyond China's "gates" than in past disputes, even if forpolitically "defensive" purposes. As research on the "cult of the offen-sive" among the major powers in Europe prior to the First World Warimplies, the kind of military doctrine developing in China, when interact-ing with the offensive preferences of the American military, is likely toreduce further the time available for, and the inclination to resort to,diplomatic measures designed to head off military conflict.56 In otherwords, the interactive effect of Chinese and American military doctrinesin any future Sino-U.S. political crisis - say over the Taiwan issue - willcreate stronger incentives for both sides to engage in pre-emption,whether on a local or strategic scale. This process may well underminethe timely search for political solutions. The possibility that Chinese

54. This could change, of course, if there is a sense of rising expectations in China's questfor international status such that relative power share is no longer an accurate indicator ofsubjective assessments of China's status.

55. Interview with Academy of Military Science strategist, Beijing, April 1996. See alsoZhang Jing and Yao Yanjin, Jijifangyu zhanlue qianshuo (An Introduction to the ActiveDefence Strategy) (Beijing, Liberation Army Publishing House, 1985), p. 137; Guan Jixian,Gaojishujubu zhanzheng zhanyi (Campaigns in High Tech Limited Wars) (Beijing: NationalDefence University Press, 1993), pp. 141, 23-24; Deng Xiaoping, "Women de zhanluefanzhen shi jiji fangyu" ("Our strategic policy is active defence"), in Academy of MilitarySciences and Central Documents Research Office (eds.), Deng Xiaoping lun guofang hejundui jianshe (Deng Xiaoping on National Defence and Army Building) (Beijing: Academyof Military Sciences, 1980), p. 98. See also Chen Huiban's discussion of the flexibleapplication of the second-strike principle, with specific emphasis on "exploiting the firstopportunity to defeat the enemy" (xianji zhi ren), "Guanyu xin shiqi zhanlue fangzhen hezhidao yuanze wenti" ("Concerning questions relating to the guiding principles and strategicpolicies of the new period"), in Guofang daxue xuebao (National Defence UniversityJournal), in Fuyin baokan ziliao - junshi (Reproduced Periodical Materials - MilitaryAffairs), No. 3 (1989), p. 28. The phrase is very close to the term for pre-emption -xian fazhi ren - meaning, "setting out first to control/defeat the enemy." See also Nan Li, "The PLA'sevolving warfighting doctrine, strategy, and tactics, 1985-1995: a Chinese perspective," TheChina Quarterly, No. 146 (June 1996), pp. 443-464.

56. See Jack Snyder, "Civil military relations and the cult of the offensive, 1914and 1984,"International Security, Vol. 9, No. 1 (Summer 1984), pp. 108-146, and Stephen Van Evera,"The cult of the offensive and the origins of the First World War," International Security,Vol. 9, No. 1 (Summer 1984), pp. 58-107.


military planners may, in the next century, be forced to deal withAmerican national missile defences that undermine China's fragilenuclear deterrent only increases the incentives for China to pre-empt in acrisis over high value territory and sovereignty issues before suchdefences are credible. American plans for the unilateral deployment ofballistic missile defences threatened to destabilize U.S.-Soviet deterrencein the 1980s by providing the Soviets incentives to strike early in a majorcrisis before facing a degradation in their deterrent. Similarly, there isreason to believe that a world of asymmetrical ballistic missile defenceswill destabilize Sino-U.S. relations in the 21st century. Both sides,therefore, should be engaged in discussions over how to revise theiroffence-dominant doctrines to conform to a more desirable defence-dominant relationship. That is, both sides should be trying to find waysto reduce fears of pre-emption in crises.57

In sum, the development of alternative approaches to conflict anddispute management may well be a critical variable in reducing China'sdispute and violence proneness into the future, and in minimizing thechances of direct Sino-U.S. military conflict. In this respect it is encour-aging that specialists in the Chinese military are beginning to investigatethe principles of non-violent or low violence conflict management tech-niques aimed at preventing insecurity spirals, miscommunication, the"zero-sumization" of disputes, among other pathologies of crises.58 Theconcrete manifestations of these principles - bilateral and multilateralhotlines, joint crisis management centres, doctrinal and operational trans-parency, mutually acceptable and verifiable constraints on military exer-cises and deployed capabilities, among other techniques - ought to behigh on the agenda of all states and multilateral institutions interested in"engaging" China.

57. In the paradoxical world of nuclear deterrence, the existence of survi vable second strikeoffensive capabilities creates a defence-dominant environment because no side can strike firstor move offensively without risking destruction in return. A world with asymmetric strategicdefences, therefore, is not a defence-dominant one. It is, rather, the opposite because it createsthe option for one side to strike first and defend against retaliation. There is no reason to assumethat with the end of the American-Soviet rivalry the basic elements of nuclear deterrencebelieved to operate in the Cold War are inapplicable in the American-Chinese relationship.The U.S. ought to be assisting China to develop an assured second strike minimum deterrencecapability, by providing, for instance, early warning technologies, safety mechanisms forcommand and control, and submarine launched ballistic missile technology in return forverifiable, bilateral and/or multilateral commitments to eschew MIRVing, ballistic missiledefence, and anti-satellite weapons development and deployment. It was conceptuallypossible for the U.S. to have made various offers to the Russians to help in developing mutualBMD capabilities on the grounds this was strategically stabilizing. While the offers - madein both the Reagan and Clinton administrations - were probably disingenuous, it is not beyondthe realm of the reasonable that the U.S. should consider offers to help develop a Sino-U.S.defence-dominant nuclear relationship.

58. Hu Ping, Guoji chongtufenxiyu weijiguanliyanjiu (Research in International ConflictAnalysis and Crisis Management) (Beijing: Military Literature Publishing House, 1993); PanShiying, Xiandai zhanlue sikao (Thoughts on Modem Strategy) (Beijing: World Affairs Press,1992); and Gao Jinxi and Gu Dexin, Introduction to International Strategic Studies. In recentyears China has put some of these principles into practice with a small number of bilateraland multilateral agreements on confidence-building measures with India, Russia, the CentralAsian republics. These agreements include such steps as prior notification of militaryexercises and restrictions on the size of forces that can exercise near borders.

China's Militarized Interstate Dispute Behaviour 1949 1992: A

Documents

Transcript of China's Militarized Interstate Dispute Behaviour 1949 1992: A