Social Revolution,the State,and War

Social Revolutions, the State, and War

We begin by reviewing the theory of revolution developed in Theda Skocpol’s (1979) States and Social Revolutions. Her theory has utility for our discussion for two reasons. First, her approach is state-centric and the ability of countries to wage war is predicated on national capabilities. Second, she locates the origins of social revolution to a substantial degree in the international system, in the inability of traditional states to compete effectively with more powerful neighbors. Thus, her theory is one that focuses on the impact of revolution on the state and the place of the state in the world in terms of its ability to compete globally.

Skocpol’s work focuses on the three great social revolutions that dominate the literature on this subject—the French, Russian, and Chinese revolutions. There have been of course a larger number of social revolutions, which we utilize when we turn to our statistical analyses. Within comparative politics and sociology, there has been a debate over whether Skocpol’s theories, and more generally theories generated by these three essential cases, are relevant to these other revolutions (Geddes 2003; Foran 1992; Goodwin 2001). We test whether revolutions, generally, have an impact on war-making capabilities and outcomes.

“Social” revolutions introduce large-scale changes that go beyond the mere turnover of political power. According to Skocpol, “Social revolutions are rapid, basic transformations of a society’s state and class structures; and they are accompanied and in part carried through by class-based revolts from below” (1979, 4-5). They thus overturn existing patterns of rule and social structures quickly and replace them with new patterns of order. Skocpol examines how revolutions ultimately reconstitute state power in a new form, which exceeds the ability of prerevolutionary regimes to control and direct society.

Skocpol’s theory of revolution is not one that rests primarily on a relative assessment of discontent in society, but instead gauges the state’s capacity to keep subaltern classes under the control of the state and the ruling class. Revolution only occurs when the repressive apparatuses of the state are sufficiently weakened to the point where the lower classes can disregard the commands of the state and its rulers. Only then can discontent and rebellion overturn the ancien régime. The precipitant for such events, however, lies in the international system. The three countries studied in States and Social Revolutions all faced challenges from “more economically developed military competitors” (1979, 41). Further, all three revolutions occurred in great powers in which the state was not fully modern, and in which the aristocracy, rather than a modern state bureaucracy, played the dominant role in maintaining the political and social order.

Skocpol’s discussion draws heavily on Gerschenkron’s (1962) ideas concerning the “relative backwardness” of “late developers.” When faced with challenges from outside that threaten their power, the rulers of old regimes undertook modernizing reforms to enhance the power of the state and its ability to contend with external challenges. For France, the problem lay in its global struggle with a more economically formidable England. For Russia, it was a series of martial and diplomatic setbacks—the Crimean War, the Congress of Berlin, and the Russo-Japanese War—which culminated in its disastrous campaign in World War I. While Russia did manage to initiate industrialization, its growth and modernization was slower than that of the other powers with whom it would come into conflict. Both the Germans and the Japanese were more successful in their own attempts to “catch up” internationally by successful revolutions from above. Finally, the Ch’ing Dynasty in China was unable to cope with extraterritorial encroachments on coastal China by a range of external powers. The Chinese nationalists who overthrew the Ch’ing, in turn, had trouble consolidating their hold over the country due to rebellious warlords and, under Chiang Kai-shek, were spectacularly unsuccessful in repelling Japanese encroachment and extensive conquest in Northern China in the 1930s.

Skocpol’s analysis of the impact of revolution acknowledges that postrevolutionary France, Russia, and China emerged as stronger states and that this had an impact on the global system. Her detailed discussion of the impact of revolution on the state in her book, however, is limited to the domestic sphere (1979). She does address the question of how revolution changes the potential military capabilities of states in an article nine years later. The piece reflects on how more recent revolutions (Mexico, Cuba, Bolivia, Nicaragua, Vietnam, and Iran) have been more successful in resisting the encroachments of great powers. While the article is suggestive, Skocpol concludes with a call, “The full exploration of these notions will require more precise theorizing and more systematic comparative research, going beyond the historical illustrations offered here” (1988, 168).

We take up this challenge to study the question with more precise theoretical propositions and more systematic empirical evidence. Our reading of Skocpol leads us to argue that one important way that social revolution should have an impact on the international system is through enhancing a state’s war-making capacity. Skocpol, like Huntington (1968) before her, sees revolutions as modernizing political events that simultaneously increase the number of participants in the political system and erect new institutions to incorporate these populations. Great social revolutions lead to the replacement of agrarian empires that have personalistic and not fully encompassing bureaucracies with “a centralized, mass incorporating state, with enhanced great-power potential in the international arena” (Skocpol 1979, 162).

Revolutions constitute a special class of modernizing political events. They transform traditional state structures which placed many countries at a decided disadvantage in the interstate system. In talking about modernity we are not confining ourselves to or even primarily concerned with economic development (something that successful revolutions also promote) but the emergence of modern state structures where none or only the rudiments of them had been in place before. Revolutions replace the decentralized and fragmentary networks of local authority associated with agrarian societies and create bureaucratic polities that for the first time in the history of their nations created comprehensive centralized state power that extended down to the local level and brought all of society directly under state authority.

There are three aspects of social revolutions that we think lead to enhanced war-making capacity and, ultimately, more successful interstate war outcomes. We believe that these factors will contribute to the ability of postrevolutionary regimes to field larger and better equipped militaries. First, there is the incorporation of heretofore excluded social classes into these new, unified state structures. Such processes include the mobilization and incorporation of the lower classes (as well as the middle classes in some cases, e.g., France 1789). Especially in the early charismatic phase of revolutionary state building this will often include a great popular revolutionary and nationalistic fervor translating into broad support for the postrevolutionary regime (Tilly 1993). This postrevolutionary élan should allow the regime relatively wide latitude to implement its preferred policies.

Second, revolutions are led and completed by highly motivated ideologically coherent counterelites who are driven to mobilize society in pursuit of their transformational goals and to overcome past military failures (Skocpol 1979). Jowitt discusses how revolutionary elites are motivated with the intensity of charismatic authority in a Weberian sense (e.g., the belief in a transformative mission), while being devoted to an organizational ethos of fealty to hierarchy and command. This means that revolutionary leaders have at their disposal highly disciplined organizations that pursue their transformational goals with exceptional zeal (Weber 1968; Jowitt 1993).

Third, social revolutions lead to more modern, bureaucratized states (Huntington 1968; Skocpol 1979). By replacing the dispersed power of traditional elites with a fully bureaucratic state, social revolutions provide the regime with an enhanced power to command (Jowitt 1993) and modern state structures capable of extracting a higher level of revenue from society (North 1981; Levi 1989). Given their motivation to improve upon their states’ past military failures and the support of the population, these modernized state structures provide postrevolutionary leaders with an increased ability to mobilize their citizens for military purposes (Skocpol 1988). Postrevolutionary bureaucratization institutionalizes the short-term effects of revolutionary fervor into lasting capabilities. It is for this reason that the impact of revolution on state power persists beyond the initial postrevolutionary mobilization.

In sum, elite coherence, the creation of a modern bureaucratic state, the bringing of new populations under the direct control of the state, and the shared revolutionary ethos of both elites and the population at large should allow postrevolutionary regimes to exercise a greater degree of command and control over society and thus raise larger armies, and extract greater resources. This leads to our first two hypotheses:

Three effects of social revolutions should give postrevolutionary regimes an advantage in achieving successful war outcomes. First, as previously noted, social revolutions modernize and centralize the apparatus of the state. Second, they put in power ideologically disciplined and unified elite with high levels of domestic support who are strongly motivated to overcome poor past performance in the international system. For these two reasons, the militaries of postrevolutionary states should be less likely to suffer from many of the inefficiencies and civil–military conflicts typically associated with other nondemocratic regimes (see Feaver 1999; Reiter and Stam 2002; Biddle and Long 2004). Third, if our hypotheses about postrevolutionary capabilities are correct, social revolutions should also result in larger, better funded militaries. All in all then, social revolutions should yield stronger, more efficient militaries that should perform more successfully on the battlefield than do nonrevolutionary militaries. Conditional on a country’s involvement in an interstate war, our third hypothesis is

The claim that countries with more modern militaries tend to fare better in interstate wars is not new (e.g., Stam 1996). However, the expectation that social revolutions lead to greater success in war is novel. Moreover, our claim is that social revolutions are a special class of modernizing events. This is an argument that goes beyond mere military modernization, but puts stress on the state, the way it is organized and who controls it.

To sum up, we argue that social revolutions serve as a mechanism by which countries discard traditional state structures, develop into a fully encompassing bureaucratic state, fully subject the population at large—including the rebellious subaltern classes—to the authority of the state, allow for greater, more efficient extraction of resources, and empower a hierarchical and disciplined elite motivated to overcome past military failures. Accordingly, postrevolutionary states should be able to field larger, better financed militaries that are able to achieve more successful interstate war outcomes compared to their nonrevolutionary counterparts. ¹

The “Revolutions” of 1989

Among the events routinely classified as revolutions, there is one subset of cases that differs from the classical revolutions that have played such a large role in our theorization of their impact on war-making capacity and war outcomes. These are the revolutions that brought an end to Soviet-type systems in East Central Europe in 1989. British historian Timothy Garton Ash (1990) labeled these events “refolutions” because as processes of change they seemed to fall between reform and revolution (1990). Like great reforms they were (largely) peaceful and often negotiated. Like revolutions, they were rapid and brought about radical changes in patterns of rule, economic allocation, and control of property. This tension is reflected in how some political scientists treat these events as democratic transitions (Linz and Stepan 1996; Elster 1996), whereas others use the language of the study of revolutions, movements, and social conflict (Bauman 1994; Ekiert and Kubik 1999; Beissinger 2002).

The events of 1989 differ from other twentieth-century revolutions in two distinct ways. First, the events associated with the end of the Soviet Bloc did not enable underdeveloped traditional regimes to make a multidimensional jump to political modernity. The countries of Central and Eastern Europe, if they had not already embarked on modern forms of rule and economy, either were transformed by the Russian Revolution or the adoption of Soviet-type institutions in the wake of World War II. Thus, the revolutions of 1989 are different from earlier revolutions because they replaced one form of modernity (Soviet-type systems) with another—liberal democratic rule and a market economy. Second, the post-Soviet revolutions in our sample produced democratic regimes. This is in marked contrast to other twentieth-century social revolutions, all of whom replaced one form of nondemocratic regime with another (see appendix).

These differences suggest that the post-Soviet revolutions should not have the same effect on war-making capabilities and war outcomes as other social revolutions. First, as the countries of Eastern and Central Europe already had modern encompassing bureaucratic states, post-Soviet revolutions would not have had the same modernizing effect on those states as social revolutions that occurred in countries with more traditional state structures. For instance, we would not expect the Polish revolution of 1989 to bring previously unincorporated populations under the control of the state, extract untapped resources to enhance the capacities of the state, or to substantively expand and professionalize the state to the same extent that the Russian Revolution of 1917 did.

Second, because the revolutions of 1989 ended in democracy, we expect these states to adopt the more modest military recruitment and spending patterns of other democracies. Therefore, postcommunist regimes should place fewer men under arms and have lower levels of military spending than they did as part of the Soviet bloc (Goldsmith 2003; Fordham and Walker 2005). Additionally, we would expect the post-Soviet states to be more successful in interstate wars outcomes than nonrevolutionary states would be but because of the greater selectivity and various war-fighting advantages attributed to democracies. ² From the preceding analysis, we expect that the states that have experienced post-Soviet revolutions to behave like other democracies rather than the other revolutionary states in our sample.

The Cases of France, Russia, and China as a Natural Experiment

Before subjecting our hypotheses to statistical analysis, we investigate the military performances of the three states which figure centrally in the study of social revolutions before and after their own revolutions. We treat these longitudinal observations of the prerevolutionary and postrevolutionary military fortunes of France, Russia, and China as a quasi-natural experiment which establishes the plausibility of our argument (Shadish, Cook, and Campbell 2001).

The historical record shows that postrevolutionary France, Russia, and China achieved superior outcomes in the type of interstate conflicts that helped to undermine their prerevolutionary predecessors. The most plausible change that explains these different war outcomes before and after the revolution is the transformation of the state brought about by the revolution itself. Though this historical evidence is not sufficient in and of itself to validate our claims fully, it adds robustness to the statistical findings that we present in the following section.

In the case of prerevolutionary France, the Bourbon monarchy was unable either to contend with England globally for control of colonies or to wage successful land campaigns on the European mainland. France’s one unambiguously successful campaign of the epoch, that to assist England’s North American colonies in the Revolutionary War, was an act of fiscal suicide that not only broke the bank but also the whole ancien régime. It led to the calling of the Estates General, and, eventually, the storming of the Bastille. Compare this to Napoleon’s campaigns and the fact that the French army got all the way to Moscow. The ability of the postrevolutionary regime to raise troops on the basis of levee-en-masse equipped Napoleon with substantially larger armies composed of citizen–soldiers who fought with conviction, used flexible (not set piece) tactics, and foraged off the land (rather than remaining tied to supply trains). This was a new, more effective type of army that made the small professional armies of the eighteenth century obsolete and changed the face of warfare (Skocpol 1988, 153).

The Russian experience is also supportive of the idea that revolution has tangible military benefits. The Russian economy and the Imperial Army both collapsed in 1917, setting the stage for the seizure of power by the Bolsheviks following the failure of the Provisional Government. The fledgling Bolshevik, authorities, in turn, were forced to sign the humiliating Treaty of Brest-Litovsk which ceded control of large swaths of Poland, Ukraine, and the Baltic region to the Central Powers. Compare this to the thirty years following the Russian Revolution. The Red Army defeated the German army on the Eastern Front of World War II, despite suffering extremely heavy losses. Then, following the conclusion of hostilities, the Soviet state established de facto control over Europe east of the Oder–Neisse line running all the way to Trieste. The size of the Russian military and the resources spent on it in World War I and World War II offer further support for our claims. At its peak, prerevolutionary Russia fielded a military of nearly 11 million men during World War I. While this is certainly a sizable army, the Soviet Union put approximately 12.5 million men in the field in the final year of World War II, about 13.6 percent more men. The difference is starker in terms of military spending. In constant 1995 US dollars, prerevolutionary Russia spent an average of $36 billion dollars per year on its military during World War I. The Soviet Union spent 68 percent more per year (about $63.5 billion) during World War II. ³

Finally, the Chinese case also lends credibility to our argument. Both the failing Empire and the fledgling nationalist state that replaced it had insufficient capacity to fend off foreign encroachments. In contrast, the postrevolutionary People’s Liberation Army (PLA) was able to fight the forces of the United Nations to a standstill in neighboring Korea in the early 1950s. Looking at the resources spent on the Chinese military during World War II and the Korean War, the postrevolutionary advantage is striking. In constant 1995 US dollars, the prerevolutionary Chinese government spent, on average, approximately $1.7 billion dollars per year on its military during World War II. In sharp contrast, the People’s Republic of China spent just over $14.2 billion per year on the military throughout the Korean War, an increase of approximately 750 percent. The sizes of the prerevolutionary and postrevolutionary Chinese militaries seemingly run counter to our expectations but in actuality reflect very different circumstances and are thus hard to directly compare. During World War II, the Chinese army peaked at approximately 4.7 million men in 1945. The postrevolutionary, People’s Liberation Army reached a maximum of about four million men during the Korean War in 1950. The contexts of these two numbers though are markedly different. The combined forces of the Nationalist and Communists during WWII included a substantial number of irregulars, whereas the postwar PLA was a conventional fighting force. And furthermore the resolution of the civil war meant that a substantial military force was evacuated by the nationalists from the mainland to Formosa and is not taken into account as part of the Chinese military. ⁴

Overall, these longitudinal comparisons of France, Russia, and China support our argument. In all three cases, prerevolutionary states fared abysmally in conflict with international rivals. Within a relatively short time, however, states transformed by social revolutions faced similar military challenges and achieved much better results. Having demonstrated the plausibility of our hypotheses using the examples of the three most important social revolutions, we now turn to a more systematic testing of our predictions.

Research Design

Our hypotheses are tested on two data sets, each generated using EUGene (Bennett and Stam 2000, version 3.201). Our capabilities hypotheses are tested on a country-year, time-series cross-sectional data set covering all states in the international system. The temporal domain of our analyses covers the period from 1900 until 2001 for our test of Hypothesis 1 and, due to data constraints on our dependent variable, 1900 to 1998 for our analysis of Hypothesis 2. Our claim that postrevolutionary states should achieve greater success in interstate wars is tested on a directed-dyad-year data set covering all countries in the interstate system from 1900 to 2001.

We use directed-dyads for three primary reasons. First, dyadic analysis allows us to incorporate characteristics about the relationship between belligerents that affect war outcomes in a more straightforward manner than is possible with a state-year approach (e.g., relative capabilities, and contiguity). The dyadic approach is particularly useful given our attempt to correct for the nonrandom sample selection of countries into the population of interstate wars with a selection model (see the following discussion). Second, compared to a nondirected setup, a directed-dyadic approach allows us to more easily analyze cases in which two postrevolutionary regimes fought against one another in an interstate war (China vs. Vietnam in 1978 and 1986 and Vietnam vs. Cambodia in 1975). Third, a directed-dyadic approach allows us to identify a state’s role in an interstate war (initiator or target) and the context(s) under which social revolutions lead to more successful war outcomes, something not readily accomplished with an analysis of nondirected dyads.

Results

We find states that have undergone a social revolution have greater war-making capabilities and are more likely to achieve positive interstate war outcomes than are nonrevolutionary states. As expected, we also find that the revolutions of 1989 do not exhibit the same enhanced war-making capabilities as more standard revolutions. We first discuss the tests of our war-making capabilities hypotheses and then turn to our analysis of the link between postrevolutionary states and interstate war outcomes.

Social Revolutions and War-Making Capabilities

Consistent with Hypotheses 1 and 2, states that have experienced a social revolution have larger militaries and allocate more of their economic resources to the military than do nonrevolutionary states. ¹³ Further, the revolutions of 1989 led to a decrease in national war-making capabilities. We first report a set of summary statistics that support these conclusions before presenting a set of multivariate analyses of the relationship between social revolutions and war-making capabilities. The raw data indicate that the effect of a social revolution on a nation’s population under arms is dependent on whether the revolution was associated with the fall of the Soviet-type regimes. Specifically, traditional social revolutions were associated with an average 39 percent increase in the proportion of a state’s population serving in the military (from 0.65 percent to 0.90 percent) while the post-Soviet revolutions were accompanied with an average 44 percent smaller military (from 1.53 percent to 0.86 percent). The importance of context for the effect of social revolutions on a state’s war-making capabilities is even greater when we focus on the proportion of GDP a state allocates to its military. The traditional social revolutions, on average, were accompanied with a 264 percent increase in a state’s annual defense burden (from 1.12 percent to 4.07 percent). At the same time, the social revolutions of the post-Soviet states were associated, on average, with a 65 percent reduction in the percentage of GDP spent on the military (from 3.3 percent to 1.15 percent).

We present our multivariate analyses of the relationship between revolutions and war-making capacity in Table 1. Models 1 and 2 estimate a state’s Population under Arms, and models 3 and 4 estimate a nation’s annual Defense Burden. Models 1 and 3 estimate the dependent variables as a function of only a state’s revolutionary status and interstate war, while models 2 and 4 control for Democracy, Economic Development, and Years Since War. ¹⁴

OPEN IN VIEWER

Table 1.

Social Revolutions and War-Making Capabilities

	Model 1	Model 2	Model 3	Model 4
Postrevolution	0.07 * (1.71)	0.09 * (1.87)	0.60 ** (2.89)	0.36 * (1.78)
Post-soviet revolution	−0.27 ** (−2.98)	−0.21 ** (−2.21)	−1.10 ** (−2.65)	−0.49 (−1.20)
Interstate war	0.59 ** (20.75)	0.62 ** (20.31)	1.72 ** (14.06)	1.91 ** (15.60)
Democracy		−0.01 ** (−4.15)		−0.03 ** (−5.05)
Economic development		−5.33E−07 (−0.77)		−5.33E−06 † (−1.64)
Years since war		−1.87E−05 (−0.05)		−3.57E-03 ** (2.12)
Population under arms t −1	0.65 ** (90.50)	0.65 ** (86.28)
Defense burden t −1			0.76 ** (109.44)	0.82 ** (117.67)
Constant	0.23 ** (27.22)	0.25 ** (15.33)	0.46 ** (13.77)	0.31 ** (4.98)
Overall R 2	0.68	0.69	0.76	0.80
F statistic	2,454.31	1,305.19	3,319.91	2,268.38
Probability > F	<0.01	<0.01	<0.01	<0.01
Observations	9,633	8,531	8,266	7,487

Note: Dependent Variable: Models 1 & 2: Population under Arms; Models 3 & 4: Defense Burden.

Temporal domain: Models 1 & 2: 1900−2001; Models 3 & 4: 1900−1998. t-scores in parentheses.

One-tailed significance: ^† p < .1. ^* p < .05. ^** p < .01.

The results in Table 1 strongly support our predictions. From model 1, we see that social revolutions were, on average, associated with a 0.07 percentage point increase in a state’s Population under Arms. Post-Soviet revolutions, however, were associated with a 0.27 percentage point reduction in the portion of the total population serving in the military. While these figures may seem substantively small, our data show that countries, on average, had only 0.77 percent of their population serving in the military in a given year during the period from 1900 to 2001. These results are essentially the same when we control for the potentially confounding variables Democracy, Economic Development, and Years since War: postrevolutionary states increase their Population Under Arms by 0.09 of a percentage point while post-Soviet countries saw a 0.21 percentage point decrease. In terms of our control variables, countries at war have significantly larger militaries and the size of a country’s military is decreasing in its level of democracy. We find no significant relationship between Economic Development and Years Since War and the portion of a state’s citizens serving in the military.

The results in Table 1 also support our expectation that social revolutions lead states to allocate more of their economic resources to the military. Accounting for a state’s involvement in an interstate war, model 3 indicates that, on average, postrevolutionary states spent 0.60 percentage points more of their GDP on the military while the post-Soviet countries saw their defense burdens decrease by 1.1 percentage points of their GDP. After accounting for all of our control variables, postrevolutionary regimes are estimated to allocate 0.36 of a percentage point more of their GDP to the military while post-Soviet countries decreased their defense burdens, on average, by 0.49 of a percentage point. Note that the expected decrease in the defense burdens of post-Soviet states is statistically insignificant in model 4. ¹⁵ Consistent with previous research, we find that interstate war is associated with significantly higher defense burdens (Goldsmith 2003) and that a state’s military spending is decreasing in its level of democracy (Fordham and Walker 2005), its level of economic development, and the number of years since it last fought a war (Bueno de Mesquita et al. 2004).

The results reported in Table 1 lend strong statistical support to our argument that traditional social revolutions lead to an increase in a state’s war-making capabilities. To demonstrate how these relationships translate to a postrevolutionary advantage on the battlefield, Figure 1 presents the predicted Population under Arms and Defense Burden of nonrevolutionary and postrevolutionary states during an interstate war based on postestimation simulations of models 2 and 4, respectively. ¹⁶

OPEN IN VIEWER

Figure 1.

Social revolutions and wartime capabilities

Figure 1 makes clear that, compared to states that have not experienced a social revolution, postrevolutionary states send proportionally larger and better funded militaries to war. Focusing on panels A and B, respectively, we see that postrevolutionary regimes are expected to place approximately 13.5 percent more of their population under arms (1.53 percent vs. 1.35 percent) and allocate approximately 42.7 percent more of their GDP to military spending (5.60 percent vs. 3.92 percent) than did nonrevolutionary, nondemocracies during interstate wars. Looking at the second row in Figure 1, we see that these substantive differences in the wartime population under arms (Panel C) and defense burdens (Panel D) of postrevolutionary and prerevolutionary regimes are statistically significant. In sum, our analyses strongly support our claims that social revolutions are associated with a subsequent increase in war-making capabilities. ¹⁷ We now examine whether postrevolutionary regimes were able to translate these capability advantages into successful interstate war outcomes.

Social Revolutions and Interstate War Outcomes

Table 2 presents our estimates of the relationship between social revolutions and interstate war outcomes. One note regarding interpretation is worth making before we discuss our results. While coefficients on variables associated with characteristics of state A in a dyad (e.g., Postrevolutionary_A ) are interpreted in the standard way, a positive coefficient on a variable corresponding to a characteristic of state B (e.g., Postrevolutionary_B ) indicates higher values of that variable are associated with less positive war outcomes for the target. This is because higher values of War Outcome correspond to more positive outcomes for the initiating side and less positive outcomes for the targeted side in an interstate war.

OPEN IN VIEWER

Table 2.

Social Revolutions and Interstate War Outcomes, 1900–2001

	Model 5	Model 6
	(z statistic) β	(z statistic) β
DV: War Outcome
PostrevolutionaryA	0.70† (2.80)	0.71† (2.68)
PostrevolutionaryB	−0.68† (−2.17)	−0.66† (−1.92)
DemocracyA	0.07† (5.14)	0.07† (4.80)
DemocracyB	−0.06† (−4.81)	−0.06† (−4.81)
Relative capabilities	−0.23 (−0.79)	−0.22 (−0.69)
Inverse Mill’s ratio (λ)		0.30 (0.97)
DV: War Initiation
Postrevolutionary A		0.11† (1.97)
Postrevolutionary B		0.005 (0.07)
Democracy A		0.002 (0.65)
Democracy B		−0.001 (−0.28)
Relative Capabilities		−0.09 (−0.71)
Contiguity		0.15† (15.66)
Time		−1.35E−04 (−0.02)
Time2		2.00E−04 (0.67)
Time 3		−3.75E−06 (−1.26)
Constant		3.55† (−56.58)
Cutpoint (α1)	−0.90† (−5.53)	0.19 (0.17)
Cutpoint (α2)	−0.13† (−0.87)	0.97 (0.84)
War Outcome
Observations	245	245
Log Likelihood	−215.99	−215.23
χ2	63.25	57.51
Probability > χ2	<0.01	<0.01
War Initiation
Observations		1,014,524
Log-psuedolikelihood		−2,163.90
χ2		301.04
Probability > χ2		<0.01

Two-tailed significance: ^† p < .01.

Model 5: robust standard errors clustered on directed dyad.

Model 6: robust standard errors clustered on directed dyad in War Initiation equation and bootstrapped standard errors (1,000 replications) in War Outcome equation.

Model 5 in Table 2 reports the results of our ordered probit analysis of interstate war outcomes. The statistically significant coefficients on Postrevolutionary_A and Postrevolutionary_B , positive and negative, respectively, indicate that postrevolutionary regimes were more likely to achieve positive interstate war outcomes than nonrevolutionary regimes regardless of their role in the war. Regarding the control variables in the model, we find that, compared to relatively authoritarian regimes, more democratic initiators and targets are more likely to achieve positive interstate war outcomes than are nondemocratic initiators and targets, a result consistent with previous scholarship (Reiter and Stam 2002). We find a negative, but insignificant, relationship between state A’s share of the total capabilities in the dyad and successful war outcomes.

While model 5 appears to support our claim that postrevolutionary states are more likely to succeed in interstate wars, the nonrandom, strategically censored population of war participants potentially represents a barrier to making valid substantive inferences from our initial analysis of war outcomes. To check for this possibility, model 6 reports the results of our analysis of interstate war outcomes using a two-step selection model with an ordered probit outcome equation. Two things in particular are worth highlighting about model 6. First, and most importantly, the coefficient on λ (the Inverse Mills’ Ratio) is statistically insignificant. This indicates that the selection (i.e., War Initiation) and outcome (i.e., War Outcome) equations are statistically unrelated to one another and inferences drawn from the parameters estimated in model 5 are not tainted by selection bias. We therefore use model 5 to calculate the substantive effect of social revolutions on interstate war outcomes below. The second point worth highlighting about is that the parameter estimates yielded by our selection model are almost identical to those estimated by our standard ordered probit, with differences across the two models of 0.01 and 0.02 on the coefficients for Postrevolutionary_A and Postrevolutionary_B . Thus, even after accounting for the selection process, our analysis indicates that states that have undergone a social revolution are more likely to achieve positive interstate war outcomes than are states that have not experienced a revolution.

With respect to the results of the war initiation equation of model 6, we find that postrevolutionary states are no more or less likely to be targeted in an interstate war than are nonrevolutionary states, but they are, on average, more likely to initiate an interstate war. This finding would seem to run counter to the idea derived from the bargaining literature that the stronger military yielded by social revolutions would allow postrevolutionary regimes to avoid wars by intimidating potential adversaries. We also find that, compared to nondemocracies, democratic states are no more or less likely to initiate or be targeted in interstate wars. ¹⁸ Our results indicate that the relative distribution of capabilities between a pair of countries is not significantly related to the probability of war initiation. ¹⁹ In that Relative Capabilities controls for differences in military strength, our analysis indicates that the success of revolutionary states on the battlefield go beyond their enhanced capabilities to mobilize personnel and resources. We also find that there is a positive relationship between contiguity and the probability of war. Additionally, the statistically insignificant coefficients on our cubic polynomials of the number of years since a dyad has been at war suggest that the other explanatory variables account for any duration dependence present in the data.

Returning to our analysis of war outcomes, the results presented in Table 2 tell us relatively little about whether postrevolutionary states are more likely to achieve successful war outcomes in any particular situation. This is because the statistical significance of a parameter estimated via maximum likelihood is contingent on the values taken on by each of the other explanatory variables in the statistical model (e.g., Cameron and Trivedi 2005). Following King, Tomz, and Wittenberg (2000), we tested Hypothesis 3 by setting the explanatory variables in Model 5 to represent substantively interesting situations and calculating the unit effect of Postrevolutionary_A and Postrevolutionary_B on War Outcome using postestimation simulations. ²⁰ As discussed earlier, twentieth-century social revolutions were exclusively a nondemocratic phenomenon until the revolutions of 1989. We therefore calculated the probability of a nondemocratic state winning, losing, or drawing in an interstate war before and after a social revolution. Among initiators, this was done in the context of a jointly nondemocratic dyad and a nondemocratic state starting a war against a democracy. Assessing Hypothesis 3 among targeted states was done by simulating a jointly nondemocratic dyad and a situation in which a democratic state targeted a nondemocracy. The analyses reported here are based on an equal distribution of capabilities among adversaries. ²¹ The results of these simulations are presented graphically in Figure 2, with panels A and B reporting the results for initiators and panels C and D reporting the results for targets. ²²

OPEN IN VIEWER

Figure 2.

Social revolutions and the outcome of interstate wars.

An initiator or target is more (less) likely to obtain a particular war outcome after a social revolution if the unit effect is above (below) the zero line in the four panels of Figure 2. The unit effect is statistically significant at the 95 percent level if the confidence interval does not contain the zero line. Our expectation is that states should be more likely to win and less likely lose an interstate war after experiencing a social revolution. This expectation is borne out for both initiators and targets. In a jointly nondemocratic dyad (Panel A), a postrevolutionary initiator has a 0.26 higher probability of winning and a 0.16 lower probability of losing an interstate war than does a nonrevolutionary initiator. As the confidence interval about both of these predictions does not include the zero line, the differences in the probabilities are statistically significant. Panel B indicates that, compared to a nonrevolutionary initiator, a nondemocratic, postrevolutionary government has a 0.18 higher probability of winning and, equally important, a 0.27 lower probability of losing an interstate war against a democracy. Both of these relationships are statistically significant.

A similar pattern emerges when we consider states that are targeted in interstate wars. Panel C indicates that in a jointly autocratic dyad postrevolutionary targets have a 0.25 higher probability of winning and a 0.23 lower probability of losing a war than do nonrevolutionary targets. The differences in these respective probabilities are statistically significant. Finally, Panel D reports the unit effect of a social revolution on a target’s probabilistic war outcome when a democracy initiates an interstate war against a nondemocracy. Under these circumstances, a postrevolutionary target has a 0.07 higher probability of winning and a 0.18 lower probability of losing an interstate war compared to a nonrevolutionary target. Again, both of these relationships are statistically significant. In sum, the results reported in Figure 2 strongly support our claim that postrevolutionary regimes should fare better in interstate wars than states that have not experienced a social revolution.

Discussion and Conclusion

Motivated by insights drawn from a stream of the comparative politics literature on regimes and regime change, we examined the relationship between postrevolutionary states and war-making capacity and the outcome of interstate wars. Our quantitative analyses, performed on a sample of all states in the international system during the twentieth century, yield three novel findings. First, states that have undergone a social revolution enjoy an advantage in war-making capabilities over nonrevolutionary states. This is true both of the defense burden and population under arms. Second, the post-Soviet revolutions had a qualitatively different effect on a country’s war-making capabilities than did previous, more traditional social revolutions. Specifically, post-Soviet regimes are associated with significantly smaller militaries and lower defense burdens than either nonrevolutionary or Soviet states. Third, we find that postrevolutionary regimes are significantly more successful in interstate wars than are countries that have not experienced a social revolution.

These results hold at least two larger implications. First, this study highlights the utility of using insights from literatures whose epistemology and methods are radically different. Understandably, researchers are most familiar with scholarship closely related to their own research agenda. While work on social revolutions has stimulated streams of research within comparative politics, sociology, and institutional economics (see Goldstone 2003), it has had less impact in qualitative and quantitative international relations.

The application of comparative historical theories of social revolution to issues typically associated with the empirical study of conflict represents an attempt to overcome such disciplinary barriers to knowledge. While in this case a comparative politics literature provided the means to a new finding in the international conflict literature, the flow should not be one way. Comparative historical analysis can benefit mightily from literatures—like contemporary conflict studies—that approach research from a higher level of abstraction, utilize large cross-sectional time-series, and sophisticated quantitative methods. ²³

Second, the analysis presented here identifies a domestic source of international behavior independent of democratic political institutions. As discussed in the introduction, two of the more enduring findings in contemporary political science (the democratic peace and democratic success in interstate war) owe their discovery to scholars of international relations who identified the cause of interstate outcomes with a domestic level variable, regime type. The research on the relationship between regime type and interstate behavior, however, has focused perhaps too exclusively on the role of democracy. Focusing on an unexamined type of nondemocratic regime and regime change, we find that states that have undergone social revolutions exhibit their own unique pattern of interstate conflict behavior. Along with other recent research on the relationship between nondemocracies and conflict (e.g., Peceny, Beer, and Sanchez-Terry 2002; Weeks 2008), our findings demonstrate the utility of looking beyond democracy for domestic sources of interstate behavior.