Vincent paper

Martin Journal of Peace Research

No. 2 (1996)

The Relative Importance of

Power, Economic Development, and Political

System During the Middle of the Cold War

By Dr. Jack E. Vincent

Borah Professor

University of Idaho

Abstract

To empirically answer the question of the relative importance of power, economic development and political system in explaining state behavior during the cold war, Distance Theory was used on the WEIS data collection for the years 1966-68, 1969-73, 1974-78. All behavioral acts were weighted and summed to produce three time-based dyadic conflict vectors. Attribute distances were then correlated with the vectors to produce predictive models for each state. Sums of squares analysis on these models indicated that the relative importance of the predictors (in rank order) appears to be power base, economic development and political system. If an assumption of causal linkages between these factors and conflict is made, the expected effect of changes in power base should be roughly equal to the combined effect of economic development and political system changes on conflict behavior. It is important to know if this balance continues in the post cold war period. It appears to have direct relevance to the current debate over the wisdom of altering political systems in a democratic direction to promote peace.

Introduction

Scholars do not agree on what "drives" the international relations system, in the sense of weighting the relative importance of several potentially important factors, such as power, economic development and political system. This was particularly true during the cold war where a set of power configurations and relations occurred which may not be repeated for the foreseeable future. We now have an opportunity to evaluate whether the importance of such factors as power, economic development and political system, were unique to cold war era or may tend to prevail in a similar way in new systemic arrangements, as in the present pattern, or in still other patterns likely to develop in the future. The purpose of the present paper is to set the stage for such an evaluation by focusing on the "middle period" of the cold war (the years from 1966 to 1978) were data is rich enough to attempt such an evaluation.

To some, following the school of realism, power is the main driver and explains most of what occurs (Morgenthau, 1972, Schuman, 1969, Organski 1958). This school tends to be concerned with the balance of power, arms race, etc.

Others, however, suggest that economics or economic development is the fundamental force behind behavior exhibited in the system. This view has historically been associated with Marxism, but, more generally, with economic analysis (Marx and Engels, 1932, Lenin, 1939, Hobson, 1965, Ebenstein, 1972, Dougherty and Pfaltzgraff 1981). In Marxism, for example, the existence of classes, one owning the means of production and the other class producing goods becomes the major driving force of history and explains war. [It should be clear from the outset that this project is not attempting to directly test the theories of Marx, but rather, it attempts to assess the relative importance of economic development compared to political system and power base.]

Still others have focused on the political system as the key variable. Idealism, for example, has been associated with the idea that totalitarian systems tend to act one way and democratic systems another (Cook and Moos, 1952, 1954, Tannenbaum, 1955). Elaborations of this viewpoint were developed through and during the cold war era (Byrnes, 1947, Spanier, 1960, Graebner, 1960, Crabb, 1972). More recently, a great deal of research has been published suggesting that democracy may be a key factor relevant to peace, because democracies do not tend to use force against one another and are more peaceful domestically. See for example Kilgour (1991), Maoz and Russett (1991), Lake (1992), Dixon (1993), Dixon (1994) Rummel (1993) and Morgan and Campbell (1991).

A Method to Evaluate Relative Importance

Is there any way to approach this problem of comparative relevance in a systematic fashion? Through much of the cold war, the problem was usually not viewed as solvable in empirical terms. Most of the debate was framed in the terms of the impressions of one set of the scholars hold as opposed to others. Since then, it is much easier to explore such questions in terms of systematic empirical tests, given the development of large scale data banks and easier access to powerful multivariate methods. Therefore, it was decided to evaluate the relative importance of the above factors, such as economic development vs. power, by applying Distance Theory (Vincent, 1976) to a carefully constructed data base, the WEIS collection (McClelland, 1978).

Although this way of approaching the study of conflict (Distance Theory) in the international system is called a "theory", it could as easily be called a "framework". That is, Distance Theory provides the operational steps to determine the relationship between national attribute differences and national behavior but it does not deductively assert what the predicted direction of the relationships will be. Procedurally, the researcher: (1) obtains measurements on attributes (this defines a-space, or attribute space), (2) computes the absolute difference of each state from every other state on every variable to create raw distance space), (3) factor analyses raw distance space to create d-space, or the factor scores of distance space, [Thus, on a distance dimension, such as economic development distance, the d-space factor score for a dyad, such as the US-USSR, indicates whether that dyad tends to be similar or dissimilar on the variables grouped on that dimension], and, (4) correlates the factor scores of d-space with dyadic behavior (such as the conflict flowing from the US to the USSR) to determine the predictability of dyadic behavior from attribute distances.

For more details and a comparison with other approaches, see Vincent (1976, 1977). Basically, the Distance Theory approach allows the investigation of the rich field of dyadic behavior (such as US-> China behavior) as opposed to Attribute Theory involving monadic behavior (such as total US behavior, total Chinese behavior, total Cuban behavior etc.). Correlations using Distance Theory are correlations of dyadic attribute distances with dyadic behavior while correlations of Attribute Theory are of attribute scores (such as GNP per capita for the US) with total behavior (such as all the conflict sent by the US to all others in the system). Since Distance Theory is much richer in detail (16,256 dyads are considered), it is applied here. If Attribute Theory were applied it would only deal with 128 monads.

That is, such an application will enable us to compare the relative importance of the above enumerated factors (power, economic development, political system), in rich dyadic detail, using all measurable states in the system (with sufficient attribute data and behavior data).

Applying Distance Theory

To apply Distance Theory, it is necessary to sample enough variables to define the many characteristics of states. The World Handbook of Political and Social Indicators, (Taylor and Hudson, 1971) and the Cross-Polity Time Series, (Banks, 1971) lend themselves to this task and center the data well concerning the time period selected. 144 variables can be identified which have less than 10% missing information (missing data was estimated by simple regression using the best correlated variable in the collection) and which tend to span most of the identifiable characteristics of states. Following the operational rules of Distance Theory, d-space was constructed (formed) by computing the absolute distance of every state to every other state, generating a distance matrix of 16,256 (dyads) by 144 (variables). A factor analysis (see Appendix A for more details) of the distance matrix then "located" the dyads in the dyadic distance space through the computation of the d-space factor scores. Examination of the resulting factor scores tells us, the propensities of any particular dyadic pair. For example, a dyad with a positive score on a distance dimension, such as economic development distance, indicates that the dyadic partners tend to be dissimilar on the variables that load heavily on that dimension. For dyads that have negative scores on the same dimensions, however, the interpretation simply reverses. Such dyads with negative scores are more similar, than the average. In general, factor names provide the key for understanding the analysis and dyads at the top of a dimension tend to have the opposite characteristics from dyads at the bottom of the dimension. All 16,256 dyads were located, by their factor scores, on each of the 33 dimensions resulting from the analysis, but only the power, economic and political dimensions were used, given the purpose of this study. [The factor names, of course, are assigned by the investigator. The assignment here is consistent with other research where similar factors were employed.

Defining Dyadic Behavior

The WEIS data bank was then used to score conflict behavior in three separate time periods, 1966-68, 1969-73 and 1974-78. This was done in order to assess the relative stability of results as we move across time. Each number in the WEIS collection indicates whether or not a behavioral act has taken place between two actors, such as a protest sent from the US to France. Different acts, however, clearly indicate different degrees of conflict. That is, to sum such raw frequencies does not make sense for this type of analysis. To weight the variables by their varying degree of importance, thirty judges (professors and students)) were asked to rank each verb category in the WEIS collection in terms of intensity on a five-point scale. The results follow:

Table 1 here

Since the standard error for the judges was small (average = .20), similar ratings for other sets of judges ought to produce fairly similar results. In view of this, the judges weights were used to sum all of the acts, in the time periods considered, and the three resulting conflict vectors are each 16,256 x 1. [Goldstein has demonstrated that this manner of weighting WEIS verbs produces very similar results to weighting "events" as an alternative (Goldstein, 1992). Goldstein states "The replications do more to corroborate than to challenge the validity of past studies using the Vincent scale, because results were quite similar using the more detailed new scale" (1992: 382).]

Correlation Results

D-space factor scores were then correlated (regression) with the three vectors generating a predictive model for each dyadic set. For example, the US scored -.32 on power distance in the period 1966-68, indicating that the US tends to export more conflict to nations close to it than nations far from it on power base. That is, this correlation was produced when the US dyadic factor scores (measuring US distance from the USSR, the US distance from China etc.) were correlated with the US conflict with the USSR, the US conflict with China etc.

Since each state’s results can be interpreted in such terms for each dimension, it is apparent that the fundamental research question can be answered by summing the squares (indicating variance explained) for all nations on the dimensions of power base, economic development and political system. This provides a clean interpretation since the larger the squared results, the greater the predicted relevance of a factor.

If the realists are correct, the sum of squares for power base should be relatively large compared to the other two factors, whereas, if selected idealists are correct, the sums of squares for political system should emerge with the largest relative weight. Our relative importance indexes can be defined as SUMi/Sumtotal, where SUMi refers to the sums of squares for the factor in question, such as power base, and SUMtotal refers to the total sum of squares of all three factors and p is power, e is economic development, ps is political system and RI refers to the relative importance index. Performing the required calculations generates: RI(p) = .48 (1966-68), .43 (1969-73), .50 (1974-78); RI(e) = .27 (1966-68), .31 (1969-73), .29 (1974-78); and, RI(s) = .24 (1966-68), .25 (1969-73), .21 (1974-78). [Naturally there is considerable variation in the size of R, the overall correlation and r, the individual correlations, over all runs, since each dyadic set is treated separately. The average R, when all three factors are considered is .41; the average simple correlation for each factor is: power distance =.28, economic development distance =.22 and political distance =.20. In this connection, the simple correlations are virtually equal to beta weights, since the dyadic factor scores are highly orthogonal].

Conclusions

From the above, it may be concluded that power is more important than economic development and economic development is more important than political system across all three time periods. Thus, the views of the realists appear more supported, during the middle of the cold war, over those of economic explanations or the views of idealism applied to political system. Power base, using Distance Theory, emerges as almost twice as important as the other two factors combined. Economic development comes in second but only slightly outscores political system. Does this mean, however, that realism need not take into account the views of the economic theorists or idealists? Certainly not. Although this analysis clearly stresses the importance of the power base as being the most important predictor of the three considered within the system, during the period considered, nevertheless the analysis is also telling us that the combined weight of the other two factors together is, roughly, just as important or slightly more important than power base. This data, then, supports the view that a multi-dimensional explanation is more appropriate than any explanation just focusing on just one of these three factors to the exclusion of the other two, recognizing as stated above, however, the greater importance of power base. This suggests, if the correlation linkages established here may suggest causal linkages between power base, economic development and political system to conflict behavior, then the expected behavioral impact of power base changes should be roughly double the impact of either economic development or political changes. Also, it should be clear that correlation linkages do not necessarily prove causal linkages. An assumption of actual causal linkages, therefore, at this point are speculative, although the theories related to the supposed effects of political system, power base etc., of course, imply that there should be causal linkages. Empirically, then, this study indicates what the rough balance was during the cold war if the casual assumption is either correct or incorrect.

It will be of interest, of course, to discover if the cold war pattern of relative importance holds for the post cold war era. Is it possible that the grim reality of the cold war (which included the threat of instant nuclear annihilation) influenced the system in a way that fostered realism i.e. stimulated behavior to be primarily power driven. In the new post cold war era, is there more chance for the cold war secondary factors of political system or economics to emerge as primary? Data is presently being collected that may allow an assessment in this regard. In the meantime, these findings at least allow comment upon the current debate concerning the importance of promoting democratic political systems to promote peace, cited above. Much depends on the relative importance of political system in the current era. If the cold war pattern still does hold, the effect of such political system changes may be less than proponents may wish for, given the relatively greater importance of power base and economic development found here.

Summary

Tables

Table 1

WEIS Conflict Classifications

and Weights

Conflict

1. REJECT 2.6

2. ACCUSE 2.9

3. PROTEST 2.4

4. DENY 2.2

5. DEMAND 3.4

6. WARN 3.3

7. THREATEN 3.9

8. DEMONSTRATE 3.0

9. REDUCE RELATIONSHIP 3.5

10. EXPEL 3.8

11. SEIZE 4.5

12. FORCE 4.7

Appendix A

The Construction of D-Space

D-space is constructed by factor analyzing (Principal Components, Kaiser Varimax Rotation) dyadic absolute distances on 144 attribute variables and computing dyadic factor scores. To simplify factor interpretation, loadings were grouped and sorted in terms of magnitude of the three factors used. In the case of the first factor, power base distance, total gross national product in millions of U.S. dollars loads .91, total educational expenditures in million of U.S. dollars loads .88 and so forth. Names, such as power base distance, of course, are arbitrarily assigned but are consistent with similar research using attribute dimensions. See (Vincent, 1976, 1977).

Considering the first dimension, knowing that absolute distances were computed on these variables and that all the variables positively correlate with the factor, we can make the following interpretation; dyads that are close (small /i-j/) in respect to their power base distance are grouped at the bottom of the dimension and dyads that are dissimilar in respect to their power base distance are grouped at the top of the dimension. The variable list indicates what the primary variables are, in respect to these factor score assignments. Distances, in respect to power base, then, are highly patterned. By knowing a dyad’s factor score on this dimension, we can predict distance for the dyad in respect to a number of variables. Thus, if a dyad has a high factor score on the power dimension, it tends to be distance in respect to total gross national product, total educational expenditures, total energy consumption, etc. On the other hand, if a dyad has a small factor score on this dimension, the dyadic partners tend to be close in respect to total gross national product, total educational expenditures, total energy consumption, etc.

Considering the second dimension, it can be seen that almost all of these variables are per capital economic in nature. Again, the interpretation is relatively simple, all dyads with large factor scores tend to be distant in respect to the listed variables while dyads with small factor scores tend to be similar.

Considering the third dimension, since these are political variables and the loadings are positive, dyads with large factor scores tend to be dissimilar in respect to these variables, and those with small factor scores similar.

Since these three factors, of the 33 dimensions of d-space, are of primary concern in this study, they alone are presented here. Also, because the loading lists tend to be quite long only the largest loaders (.50 or larger) are given here to save space. The loadings given, however, should provide adequate orientation.

D-space, Rotated Factor Loadings

01 Power Base Distance

030 Total GNP (In Millions U.S. Dollars) 91

057 Total Education Expenditure .88

119 Total Energy Consumption in Metric Tons of Coal Equivalent .86

098 Proportion of Total World Trade .85

094 Total Imports .85

036 Total Trade, 1960 (In Millions U.S. Dollars) .83

114 Total Highway Vehicles .82

096 Total Exports .81

110 Total Passenger Cars .79

092 Total Telephones .79

041 Percentage Contribution to World Total of Scientific Authors .78

086 National Government Expenditure .78

084 National Government Revenue .76

060 Total Health Expenditure (In Millions U.S. Dollars) .75

112 Total Commercial Vehicles .75

109 Rail Passenger Kilometers .73

048 Total Number of Physicians .72

080 Population in Cities of 100,000+ .72

105 Total University Enrollment .72

082 Population in Cities of 50,000+ .71

088 Railroad Mileage .70

101 Total Secondary School Enrollment .64

040 Total Scientific and Technical Serials Published .62

107 Total School Enrollment .56

066 Total Internal Security Forces (In Units) .56

099 Total Primary School Enrollment .55

121 Book Production by Titles .54

090 All Mail .54

001 Total Population .49

02 Economic Development Distance

085 National Government Revenue Per Capita .88

031 GNP Per Capita (In U.S. Dollars) .87

087 National Government Expenditure Per Capita .87

115 Total Highway Vehicles Per Capita .86

111 Passenger Cars Per Capita .84

116 Gross Domestic Product Per Capita .84

093 Telephones Per Capita .84

058 Education Expenditure Per Capita (In Million U.S. Dollars) .83

095 Imports Per Capita .81

120 Energy Consumption in Kilograms Per Capita .80

113 Commercial Vehicles Per Capita .80

097 Exports Per Capita .75

015 Radios Per 1,000 Population .75

091 All Mail Per Capita .69

016 Televisions Per 1,000 Population .68

049 Physicians Per 1,000 Population, 1965 .65

017 Cinema Attendance Per Capita, 1960 .64

014 Newspapers Per 1,000 Population .63

019 Unadjusted School Enrollment Ratio .60

061 Health Expenditure Per Capita .60

018 Literacy Rates .58

020 Adjusted School Enrollment Ratio .58

043 Percentage of Gross Domestic Product Originating in Industry .54

122 Book Production by Titles Per Million Population .53

055 Defense Expenditure Per Capita (In U.S. Dollars) .50

03 Political Distance

129 Aggregate Competition Index .87

128 Party Legitimacy .84

127 Competitiveness of Nominating Process .82

077 Legislative Effectiveness .80

126 Size of Legislature / Number of Seats Held by Largest Party .78

064 Electoral Irregularity Score .61

065 Press Freedom Index .58

063 Political Party Fractionalization Based on Number of Seats, 1964 .55

Bibliography

Banks, Arthur S. 1971 Cross-Polity Time Series, Cambridge: MIT Press. 1971).

Cook, Thomas I. and Malcolm Moos "Foreign Policy: The Realism of Idealism" American Political Science Review 46 (1952) pp. 343-356.

Brynes, James F. 1947 Speaking Frankly, New York: Harper and Row.

Cook, Thomas I. and Malcolm Moos. 1954 Power Through Purpose: The Realism of Idealism as a Basis for Foreign Policy, Baltimore: John Hopkins Press.

Crabb, Cecil V. 1972 American Foreign Policy in the Nuclear Age, New York: Harper and Row.

Dixon, William J. "Democracy and the Management of International Conflict" Journal of Conflict Resolution 37( 1993) pp. 42-68.

Dougherty, James E. and Robert L. Plaltzgraff Jr. 1981 Contending Theories of International Relations, New York: Harper and Row.

Ebenstein, William. 1970 Today’s Isms: Communism, Fascism, Capitalism, Socialism, New Jersey: Prentice Hall.

Goldstein Joshua S. "A Conflict-Cooperation Scale for WEIS Events Data" Journal of Conflict Resolution 36 (1992) pp. 369-385.

Graebner, Norman A. 1962 Cold War Diplomacy, Princeton: Van Nostrand.

Hobson, J.A. 1965 Imperialism: A Study, Ann Arbor: University of Michigan Press.

Kilgour, D. Marc. "Domestic Political Structure and War Behavior, a Game-Theoretic Approach" Journal of Conflict Resolution 35 (1991) pp. 266-284.

Lake, David A. "Powerful Pacifists: Democratic States and War," American Political Science Review 86 (1992) PP. 24-39.

Lenin, V.I. 1939 Imperialism: The Highest Stage of Capitalism, New York: International Publishers.

Marx, Karl and Friedrich Engels. 1932 Manifesto of the Communist Party New York: International Publishers.

Maoz, Oeev, and Bruce Russett. Alliance, Contiguity, Wealth and Political Stability: is the Lack of Conflict Among Democracies a Statistical Artifact?, International Interactions 17 (1991) pp. 245-267.

McClelland, Charles. 1978 World Events Interaction Survey, Los Angeles: University Of Southern California.

Morgan, T. Clifton, and Sally Howard Campbell. "Domestic Structure, Decisional Constraints, and War, so Why Kant Democracies Fight?, Journal of Conflict Resolution 35 (1991) pp. 187-211.

Morgenthau, Hans J. 1972 Politics Among Nations New York: Knopf.

Organski, A. F. K. 1958 World Politics New York: Alfred A. Knopf.

Taylor, C. and M. Hudson. 1971 World Handbook of Political and Social Indicators Ann Arbor: University of Michigan.

Tannenbaum, Frank. 1955 The American Tradition in Foreign Policy Norman: University of Oklahoma Press.

Rummel, R.J. 1993 Genocide and Mass Murder: How Many Have Been Killed and Why?, Research Report on Genocide, University of Hawaii.

Schuman, Frederick L. 1969 International Politics New York: McGraw-Hill Book Company.

Spanier, John. 1960 American Foreign Policy Since World War II New York: Praeger.

Vincent Jack E. "Attribute and Attribute Distance Theories of International Behavior: An Overview and the Development of Some Overarching Frameworks" Political Methodology, 4 (1977) pp. 461-91.

Vincent Jack E. "Explorations in D-space: An Application of Deviance Theory", Comparative Political Studies, 9 (1976) pp. 107-35.