The City University of New York

Democracy and Economic Development: Modernization Theory RevisitedAuthor(s): Zehra F. AratSource: Comparative Politics, Vol. 21, No. 1 (Oct., 1988), pp. 21-36Published by: Ph.D. Program in Political Science of the City University of New YorkStable URL: http://www.jstor.org/stable/422069 .Accessed: 07/07/2011 14:37

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and youmay use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at .http://www.jstor.org/action/showPublisher?publisherCode=phd. .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact support@jstor.org.

Ph.D. Program in Political Science of the City University of New York and The City University of New York arecollaborating with JSTOR to digitize, preserve and extend access to Comparative Politics.

http://www.jstor.org

Democracy and Economic Development

Modernization Theory Revisited

Zehra F. Arat

Comparative Research on Democracy

Interest in the relationships between society and politics is as old as written history. Ecological theories, suggesting a relation between the geographic location of homelands and the personalities of the residents and their governments, were advanced by scholars such as Aristotle, Cicero, Ibn Khaldun, and Montesquieu and may be viewed as efforts to explain the contextual determinants of political systems. Such literature in the modern era, however, has focused on the conditions of democracy.

Historical studies suggest that modem democracies can occur only under certain conditions of capitalist industrialization. Karl Marx identified the bourgeoisie as the major force behind the emergence of democracy. He argued that the capitalist class used parliamentary systems and democratic mechanisms to capture the control of the state from the traditional elite. Similarly, Moore, in his study of major western democracies, and Soboul, in his analysis of the French Revolution, stressed the role of the middle class and urban bourgeoisie in the transformation of political systems into democracies.' Max Weber marked the importance of Protestantism in the development of western democracies. He considered individualism and a sense of individual responsibility, inherent in the Protestant ethic, as the major conditions for the development of burgher classes and a democratic political culture.

Contemporary writers, following Weber's lead, have searched for cultural requisites and the elements of a "democratic personality." A "civic culture" in Almond and Verba and a "modern" personality in Lerner have been identified as essentials of a participant (democratic) society.2 Tocqueville, in his study of American institutions, pointed out the virtue of voluntary associations as the basis of social pluralism, which in turn nurtures democracy.

However, the sociology of knowledge, which emphasizes the influence of structure and organizational setting on the development of attitudes and behaviors, compels us to direct our attention to the structural basis of developing such a culture or personality. In this regard, aspects of modernization appear to be the common explanatory factors used in the analysis of democracy.

Modernization and Political Democracy

Modernization Theory and Quantitative Analysis Almond and Verba, in their analysis of survey data from five countries, considered education as the most important factor in creating the attitudes and values vital for a participant culture.3 Deutsch and Pye stated the

21

Comparative Politics October 1988

necessity of an integrative system of mass communications to build the social cohesion which is crucial for the reinforcement of democratic procedures.4 On the other hand, Lerner and Lipset suggested a complex of interrelated social and economic conditions as prerequisites for political democracy."

Using survey data from Middle Eastern countries, Lerner identified urbanization, education, and media growth (or communication) as the essential factors for the process of democratic development. He considered urbanization to be a factor stimulating education, which in turn accelerates media growth and eventually democratic development.

Lipset, on the other hand, by conducting comparative research at the aggregate level with emphasis on the socioeconomic characteristics of societies as causal factors of democratization, initiated a new trend in which the focus is shifted from individual to system characteristics. Lipset's argument, through its expansion by subsequent scholars, started a discussion about the impact of socioeconomic development on political democracy.

Lipset compared mean values and ranges of socioeconomic development for four groups of countries: (1) European and English-speaking stable democracies, (2) European and English-speaking unstable democracies and dictatorships, (3) Latin American democracies and unstable dictatorships, and (4) Latin American stable dictatorships. He suggested a positive linear relationship between levels of socioeconomic development and democratic development. His conclusion was that "economic development involving industrialization, urbanization, high educational standards, and a steady increase in the overall wealth of the society, is a basic condition sustaining democracy, it is a mark of efficiency of the total system." He added, however, that "the stability of a given democratic system depends not only on the systems' efficiency in modernization, but also upon effectiveness and legitimacy of the political systems."6 It is, therefore, possible to conclude that, by considering effectiveness and the legitimacy of the system as well as socioeconomic development, Lipset was arguing for socioeconomic development as a necessary but not as a sufficient condition for the establishment and/or maintenance of a democratic political system.

Similarly, Coleman classified seventy-five developing countries as competitive, semicompetitive, or authoritarian on the basis of their democratic political performance in the late 1950s.7 Comparing the mean scores of each group on indicators of economic development, as well as the individual scores of countries in each category, Coleman reached two conclusions.

(1) The major hypothesis that economic development and competitiveness are positively correlated is validated when countries are grouped into major differentiating categories of competitiveness and when mean scores of economic development are employed; but (2) the hypothesis is weakened by negative correlations found when the economic scores and relative competitiveness of individual countries are considered. To this should be added the caveat that economic modernization constitutes only one dimension of the ensemble of determinants shaping political institutions and behavior in the countries with which we are concerned [developing countries].8

More sophisticated measures and methods were utilized to test the Lipset and Coleman hypothesis arguing that socioeconomic development and the level of democracy of the system are positively related. Authors developed scale measurements of democracy that are

22

Zehra F. Arat

more refined than the common dichotomous perception of democracy. Correlation and regression analyses by Cutright, Cutright and Wiley, Smith, and Coulter provided empirical support for a positive linear relationship between levels of economic development and democracy in a system.9 The linearity of the relationship has been questioned by the "threshold phenomenon" argument of Neubauer and Jackman.'0 These two authors found no significant relationship between the two properties for highly developed countries. However, the implicitly evolutionary thesis suggests that, at least at the lower stages of economic development, increases in economic development level lead to increases in the level of democracy. This contention has not been effectively challenged, and the thesis has dominated the discussions of political democracy for the past thirty years.

On the other hand, it seems fair to question whether such linearity is validly demonstrated even for the less developed countries. The countries which may be classified as middle range developed, in particular, display a conspicuous fluctuation in the level of democracy they have experienced.II

The authors who suggested a linear or curvilinear relationship between socioeconomic development and democratization have based their arguments on the findings from cross-national regression analysis of several countries at single time points or, more commonly, separate periods. The high correlation between the two factors and the significant regression coefficients displayed by such data have seemed to support their evolutionary hypothesis.

A Critique of Modernization Theory and the Evolutionary Thesis The validity of modernization theory and its evolutionary thesis of democracy needs to be verified with improved data and measurement. For this purpose answers to the following questions will be sought.

A. 1. Considering a large number of countries at different levels of economic development and democratic performance, can we find a systematic relationship between the levels of economic development and democracy?

2. If there is a systematic relationship between the two characteristics, what is its form'? Specifically, is it linear or curvilinear?

3. Does this relationship hold for different time points? B. 1. Considering each country for a number of years, do we see a systematic

relationship? 2. If there is a systematic relationship, is it the same as the one obtained from the study of

a number of countries at one time point? The three questions in part A require cross-sectional analysis of countries at one time

point and the repetition of the same analysis at multiple time points. The questions in part B refer to the analysis of each country over time. Thus, if what has been suggested by the modernization theorists holds, we should find a statistically significant relationship between socioeconomic development and democracy--linear or curvilinear. The same relationship will also occur at different time points. Moreover, the individual countries should display the same relationship over time. That is, if the linear argument holds, most of the countries should display increasing levels of democracy with increasing levels of socioeconomic development. However, if the curvilinear hypothesis holds, most of the developing countries

23

Comparative Politics October 1988

should manifest a linear relationship, but the developed countries should not suggest any significant relationship, or at least it should be weaker.

To test for these arguments the index scores of democracy are regressed on an indicator of socioeconomic development.

A New Measure of Democracy Although the authors mentioned in relation to the "linear/curvilinear" discussion, as well as some others,'2 developed some sophisticated measures of democracy, their scales involve problems which can be recapitulated as follows.

(1) They are cumulative in nature, that is, they tend to measure the democratic performance of political systems over a time period (one or more decades) or around one time point (e.g., circa 1960, circa 1965), which tends to iron out the wrinkles within the time periods and prevents any meaningful analysis of change. (2) Usually no distinction is made between colonial rule and national government, and this neglects a significant historical transformation. (3) Most of the studies exclude a considerable number of countries and limit themselves to regions or to certain categories of development levels and economic systems. (4) Some ambiguous indicators with unclear implications, such as voter turn-out, are utilized as components of democracy scales. (5) Most studies rely on data drawn from the 1960s, a few being from around 1969. The late 1960s and 1970s, however, include interesting cases, since many countries gained their independence during this period.

Considering the limitations of present measures of democracy, the author developed an alternative measure for a group of independent countries for the years from 1948 to 1977. This measure focuses on three components of popular control: availability of political participation, competitiveness, and civil liberties.'3

The participation component includes measures of the extent to which popular will can be reflected in decision-making institutions, which are for the most part in the legislative and executive branches of government.14 If the effective executive, which "refers to the individual who exercises primary influence in shaping major decisions affecting the nation's internal and external affairs," is elected, the country is assigned one point.

The legislative item in the participation component is composed of three different indicators.

1. Legislative Selection: If the legislative body of the government is elected, that is, the "members of the lower house in a bicameral system are selected by means of either direct or indirect popular election," two points are assigned. If this body is not elected, meaning that the "legislators are selected by the effective executive, or by means of heredity or ascription," only one point is assigned. No point is assigned if there is no legislature.

2. Legislative Effectiveness: One point is assigned if the legislature is largely ineffective; two points are assigned if it is partially effective; and three points are assigned if it is effective. No points are assigned for the cases of no legislature. In situations where the "legislative activity is essentially of a 'rubber stamp' character," or "domestic turmoil has made the implementation of legislation impossible," or the "effective executive has prevented the legislature from meeting, or otherwise substantially impeded the exercise of its functions," the legislature is considered to be ineffective. In a situation in which the "effective executive's power substantially outweighs, but does not completely dominate that of the legislature," the legislature is considered to be partially effective. The legislative process is defined as effective when there is "significant governmental autonomy by the

24

Zehra F. Arat

legislature, including, typically, substantial authority in regard to taxation and disbursement and the power to override executive vetoes of legislation."

3. Competitiveness of the Nomination Procedure: Again, no points are assigned in the absence of a legislature. One point is assigned if the nomination procedure for the legislature is essentially noncompetitive; two points are assigned if the nomination procedure is competitive. The categories of essentially noncompetitive and competitive nomination procedures reflect differences between the degrees to which the nominations are determined in a public arena. The competitive category refers to a process in which the public has an opportunity to influence the options. In countries which have essentially noncompetitive nomination procedures, although competition is allowed, the options are mostly predefined and the choices are dictated by a dominant organ which ultimately provides a single slate of nominees.

Thus, one plus the product of these three scores provides us with the score for the legislative part of the participation component. The sum of the legislative and executive components yields the final score for the participation component.

For measurement of the competitiveness of the political system, two characteristics are used, party legitimacy and party competitiveness.

1. Party Legitimacy: If there are no political parties or if all but a dominant party and satellites are excluded, one point is assigned. If some parties are allowed but there is significant exclusion of parties, two points are assigned. If only some "extremist" parties are excluded, three points are assigned. And if no parties are excluded from the system, four points are assigned.

2. Party Competitiveness: If the largest party held less than 70 percent of the total votes in the most recent national elections, two points are assigned. If there were no elections, or if the largest party received 70 percent or more of the votes, only one point is assigned.

The sum of these two items provides us with the final competitiveness score. The third major component, which refers to civil and political liberties, or conversely the

coerciveness of the government, is the most complex of all in terms of operationalization. The standardized scores for the number of government sanctions, which are the "actions taken by authorities to neutralize, suppress, or eliminate a perceived threat to the security of the government, the regime or the state itself,"15 within a year are regressed on the standardized scores of the number of social unrest events encountered that year, such as antigovernment demonstrations, assassinations, guerrilla warfare, riots, general strikes, and deaths from domestic violence.'6 The residual values from this regression analysis provide a measure of government coerciveness.17 All governments have coercive power, and they use it to maintain their systems. Since we have no explicit theoretical guidelines for the optimum or reasonable levels of coerciveness, an empirical strategy is utilized. By regressing the number of sanctions employed on the social unrest indicators for each year, we can estimate a line of optimum coerciveness for the real world. If a country employs more sanctions than the others given equal levels of unrest, we can describe it as a more coercive government, but if it employs fewer coercive actions, it can be ranked as less coercive. Then the difference between actual and predicted (estimated) values, the residuals, can provide us with a measure of coerciveness, or suppression of liberties.

Finally, by subtracting the degree of coerciveness'8 from the sum of the other two components we obtain an additive scale to measure the level of democracy.

25

Comparative Politics October 1988

This scale of democracy is sensitive to annual changes experienced by a country, and the scores of democracy range from 1.55 to 19.91 for a total of 1985 cases (64 to 130 countries for the years from 1948 to 1977). Moreover, this measure of democracy, which marks the annual changes in this property, displays the short-term fluctuations experienced by several countries on this property. Among the 130 countries for which data are available, some countries appear to manifest extreme change, both up and down. While Peru displays the greatest positive change, 15.81 points from 1962 to 1963, Chile displays the most pronounced decline, - 16.90 from 1972 to 1973. High increases took place also in Turkey from 1960 to 1961, with a rise of 14.98; in Venezuela from 1958 to 1959, with a rise of 14.88; in Spain from 1976 to 1977, with a rise of 13.27; and in Burma from 1959 to 1960, with a rise of 13.17. Severe declines, on the other hand, are experienced by Uruguay from 1972 to 1973, with a decline of - 16.85 points; Greece from 1966 to 1967, with a decline of - 16.68 points; Peru from 1967 to 1968, with a decline of - 16.04 points; and Burma from 1958 to 1959, with a decline of - 15.99 points.

Energy consumption per capita has been preferred as the indicator of socioeconomic development both for its common use by the other authors, especially by the modernization theorists, and for its relative superiority over the other possible indicators.'9

Data Analysis First the linear and curvilinear hypotheses are tested for all the countries with available data for thirty different time points. The index values of all countries are regressed on energy consumption per capita (ENCON) to test for the linear hypothesis for each year separately for the time period 1948-1977.

(1) Level of Democracyi = a + b (ENCON)i + ui

For the curvilinear hypothesis, two different models are employed for two possible curvilinear relationships. First, the log-linear model suggests that the level of democracy increases with economic development until a certain level of economic development is reached and then further economic development does not have a significant effect on the level of democracy. This is tested by regressing the scores of democracy on the logarithmic transformation of energy consumption per capita values.

(2) Level of Democracyi = a + b (InENCON)i + ui

Next, a second degree polynomial model suggests that the level of democracy increases with economic development until a certain level of economic development is reached but it starts to decline with further economic development. This is tested by regressing the index scores of democracy on energy consumption per capita and the squared value of energy consumption per capita.

(3) Level of Democracyi = a + b, (ENCON)i + b2 (ENCON)i, + ui

A comparison of the results of these three regression models enables us to evaluate the strength of each hypothesis. If the linear hypothesis of the early modernization theorists holds, the regression coefficient (slope = b) should be positive and significant and the

26

Zehra F. Arat

coefficient of determination (R2) should be high and significant. If the revised modernization theory holds, the log-linear model should yield a positive and significant regression coefficient (b) and a better fit to data than the linear model, where the better fit to data is captured by a relatively higher and significant R2. The third model, a slightly different curvilinear model, needs a positive regression coefficient for energy consumption per capita

(b,) and a significant negative regression coefficient (b2) for its squared value, which refers

to the decline after a certain level of economic development is reached. The regression results of cross-national analysis for thirty different years confirms the

findings of Jackman and Neubauer. The log-linear model (2) provided a better fit to the data for almost all years. The average R2 for thirty years is .27 for this model as opposed to .25 for the polynomial model and .20 for the linear model. This finding indicates that at a given time point relatively more developed countries tend to experience more democratic systems but the level of democracy does not reflect any relation for some countries which are located at highest levels of the development measure. Moreover, this relationship appears to be stable over time.20

As can be seen in Table 1, the same relationship is observed when the scores of democracy for all countries for all of the available years are regressed on socioeconomic development. However, it should be noted that such an analysis of pooled data (longitudinal and cross-sectional data combined) treats each year for each country as a separate observation. It does not capture the pattern followed by an individual country over time.

Thus, to check if this curvilinear relationship holds when the relationship is studied for each country, scores of democracy for each country from eight to thirty years21 are employed for the same three regression models.

Table 1 The Relationship between the Levels of Economic Development and Democracy: Comparison of Linear and Curvilinear Regression Models by Pooled Data (N = 2850)

Regression Coefficients

MODEL ENECON ENECON2 InENECON R2 F-Ratio

Linear .001* ... ... .18* 643.29

Second Degree Polynomial .003* -.00* ... .23* 430.94

Log-linear ... ... 1.95* .26* 1024.92

Significant at .05 level.

Note: Some coefficients are listed as (.00) or (-.00) if the first three digits after the decimal point are zero. This strategy is pursued for its convenience in reporting. Such values indicate that the coefficients are quite small in magnitude, but not necessarily equal to zero. In fact, as indicated, some of them are significantly different from zero, in a statistical sense.

27

Comparative Politics October 1988

Contrary to the results from the cross-national analyses, the findings from the longitudinal application of data to the three models suggest more complex relationships. Although the short period of time with available data prevents us from deriving any comprehensive conclusions, the patterns manifested by each country can be sorted under seven major categories, as in Table 2, to display the variety of relationships as illustrated in Figure 1.

The no relationship category includes the countries which display no systematic relationship in terms of the three models tested, at .05 significance level.

Countries are included in one of the linear categories if there is a significant positive or negative linear relationship (positive-linear and negative-linear respectively) observed between economic development and democracy and neither of the curvilinear models provides a better fit to the data (a significant increase in the explained variance, that is, a significantly high R2). While the positive-linear category includes the countries which become more democratic over time with increasing levels of socioeconomic development, the negative-linear category refers to the change in opposite direction where socioeconomic development appears with decline in the level of democracy.

Positive-curvilinear and negative-curvilinear categories refer to the countries where the log-linear model provides the best fit to the data. The positive-curvilinear category covers the countries which display an increase in the level of democracy with increasing levels of economic development until a certain level of economic development is reached and no relationship after that economic development level. Similarly, the negative-curvilinear category refers to continuous decline in the level of democracy with increasing economic development up to a certain level of economic development.

Countries are classified as curvilinear positive-negative or curvilinear negative-positive if the second degree polynomial model provides a better fit in comparison to the others. The

Table 2 The Relationship between the Levels of Economic Development and Democracy by Country

No-Relationship Linear Curvilinear(Log-Linear) Curvilinear(Polynomial) Positive Negative Positive Negative Positive-Negative Negative-Positive

Afghanistan Korea, Rep. Guatemala Guyana El Salvador Cameroon Australia Austria Albania Malasia Mongolia Jamaica Egypt Chad Bulgaria Benin Algeria Malta Paraguay Madagascar Kuwait France Canada Brazil Argentina Mauritania Portugal Mali Gambia Congo Burma Barbados Mexico Romania Netherlands Iceland Dominican Rep. China Belgium Moracco Senegal Niger Laos Kampuchea Central Africa Benin New Zealand Spain Norway Liberia Honduras Colombia Burundi Pakistan Venezuela Panama Luxemburg India Cuba Chile Saudi Arabia Rwanda Malawi Ireland Cyprus Costa Rica Sri Lanka South Africa Mauritius Jordan Ecuador Czoslovakia Sudan Nigeria Korea, PR. Greece Denmark Sweden Nepal Haiti Ethiopia Switzerland Nicaragua Iraq Finland Tanzania Peru Lebanon Gabon Thailand Philippines Libya Germany, DR. Trinidad Upper Volta Poland Germany, FR. Tunusia Uruguay Singapore Ghana Turkey Yemen, Arab Rep. Somalia Guine Uganda Yugoslavia Syria Hungary U.S.A. United Kingdom Indonasia Vietnam, PR. USSR Iran Vietnam, DR. Israel Yemen, DR. Italy Zaire Ivory Coast Zambia Japan Zimbabwe

28

Zehra F. Arat

Figure 1 Some Possible Relationships between the Levels of Democracy and Economic Development Level

positive

negative

Economic Development a-Linear Relationship

positive

negative

Economic Development b-Curvilinear Relationship (log-linear)

positive-negative

negative-positive

Economic Development c-Curvilinear Relationship (polynomial)

29

Comparative Politics October 1988

positive-negative category covers the countries where an increase in the level of democracy is observed with increasing levels of economic development until a certain level of economic development is reached and then further increase in economic development appears with a decline in the level of democracy. The negative-positive category refers to the reverse of this relationship.

If the evolutionary thesis suggested by the modernization theorists is to hold, most of the "developing countries" should appear in the positive-linear category, and most of the "developed" countries should appear in the No-relationship category. However, as seen in Table 2, only a few countries fit the models suggested by modernization theory: out of 120 countries only eight (even though all can be considered as developing countries) experience increasing levels of democracy with increasing levels of economic development, and the developed countries with the rest of the developing countries are spread in the other categories, each of which suggests a different pattern of relationship.

On the basis of these findings it can be concluded that increasing levels of economic development do not necessarily lead to higher levels of democracy, even for the less developed countries. Such a relationship, which has been observed through the study of cross-national data, suggests only that at a given time point the countries with more democratic political systems happen to be the ones which are economically more developed. As Rustow emphasized in his critique of the "Lipset-Cutright genre," "correlation is not the same as causation."22

Developing countries, on the other hand, do not display a linear relationship, but rather more complex patterns or no systematic relationship at all. In fact, most of these countries, especially the ones which are located in the middle of the development axis, experience higher levels of instability, a continuous back and forth shift, on the scale of democracy. Huntington classifies these unstable political systems in his "cyclical model" of democratization and divides them into two groups on the basis of their regularity in oscillation between two forms of systems. In countries such as Peru, Ecuador, Bolivia, Argentina, Ghana, and Nigeria, where the oscillation between despotism and democracy is quite regular, "the alternation of democracy and despotism is the political system;" in others, "the shift from a stable despotism to a stable democracy is a change in political systems."23

Democratic Instability and Economic Development

A measure of "instability" in regard to democracy, based on the scale of democracy discussed above, was developed to identify patterns of oscillation. For each country, the sum of the absolute values of the annual change in democracy scores was divided by the number of years in the time period covered (which varies for each country according to the date of independence and data availability) to obtain an average score of "change in the level of democracy." This calculation follows the logic that, the more frequent and larger the shifts a country displays on the scale of democracy, the higher "democratic instability" score it will get. Out of 124 countries included here, Sudan (3.15), Pakistan (2.67), Burma (2.53), Argentina (2.32), and Peru (2.16) composed the most unstable group; while Austria (.03), Switzerland (.04), Saudi Arabia (.048), Sweden (.05), and New Zealand (.06)

30

Zehra F. Arat

appeared to be the most stable countries, where the mean value for instability is .87 (see Table 3). It should be noted, however, that the instability score does not measure political instability in general but measures instability specifically in regard to the level of democracy. That is, countries which rank high on this scale experience higher rates of shifts between relatively more authoritarian and democratic systems.

Table 3 The List of Countries and Their Instability Scores (127)

Austria .03 Lebanon .55 Laos 1.08 Switzerland .04 Ethiopia .56 Cameroun 1.08 Saudi Arabia .05 Costa Rica .57 Germany, Dem. R. 1.10 Sweden .05 Zambia .57 Cuba 1.10 New Zealand .06 UK .58 Panama 1.11 Ireland .07 Tanzania .59 El Salvador 1.14 Liberia .09 Kampuchea .59 Dominican Rep. 1.15

Belgium .11 India .60 Jordan 1.17 Albania .12 Finland .60 Philippines 1.19

Japan .13 Iceland .61 Uruguay 1.20

Italy .13 Denmark .61 Colombia 1.20

Korea, Peop. R. .14 Germany, Fed. R. .62 Vietnam, South 1.22 USSR .18 Tunisia .64 Botswana 1.30

Yemen, Arab R. .18 Spain .65 Madagascar 1.30 Mexico .19 South Africa .65 Lesotho 1.30 France .19 Egypt .65 Syria 1.36

Bulgaria .25 Australia .66 Guyana 1.38

Senegal .26 Canada .67 Jamaica 1.40 Mali .29 Romania .67 Upper Volta 1.41

Paraguay .29 Somalia .69 Brazil 1.41

Libya .30 Ecuador .69 Yemen, PDR. 1.46

Niger .31 Hungary .73 Korea, Rep. 1.46 Poland .32 Israel .74 Chile 1.46 Burundi .33 Maldives .75 Barbados 1.47 USA .36 Moracco .76 Uganda 1.50

Yugoslavia .36 Sri Lanka .79 Nigeria 1.54 Malawi .38 Ivory Coast .81 Malaya 1.55

Afghanistan .38 Mauritus .81 Indonesia 1.56

Iraq .39 Portugal .81 Gambia 1.57

Mongolia .39 Venezuela .82 Malta 1.57

Nepal .40 Czechoslovakia .83 Guatemala 1.71 Zimbabwe .40 Nicaragua .84 Thailand 1.77 Benin .44 Zaire .87 Bolivia 1.79

Norway .45 Vietnam, Peop. R. .88 Turkey 1.81 Rwanda .46 Haiti .88 Honduras 1.89

Algeria .47 Chad .92 Greece 1.91 Guine .48 Singapore .94 Ghana 2.03 China .48 Congo .94 Trinidad 2.04

Luxemburg .48 Central Africa .96 Peru 2.16

Cyprus .51 Kuwait 1.00 Argentina 2.32 Iran .53 Gabon 1.01 Burma 2.53 Netherland .54 Mauritania 1.03 Pakistan 2.67

Sudan 3.15

31

Comparative Politics October 1988

To test the hypothesis that the middle range developed countries (MDCs) experience higher levels of democratic instability than the least developed (LDC) and advanced developed (ADC) countries, a regression model which captures the curvilinearity of the relationship between economic development and instability is utilized. The instability scores for all of the countries listed in Table 3-except Botswana, Lesotho, and Maldives, excluded due to lack of data on energy consumption-are regressed on the logarithmic values of their average energy consumption per capita (AVENCON) and the squared value of this indicator of economic development. Logarithmic transformation of average energy consumption values is employed to smooth the data which are highly (positively) skewed.

Instability, = a + b) (InAVENCON), + b, (InAVENCON),2 + ui

In the second degree polynomial model, if the regression coefficient b, is positive but b2 is negative and they are both significant, we can say that there is a curvilinear relationship between the instability of the political system and the development level of the country. Economic development appears with democratic instability until a certain level is reached, and further development appears with higher levels of stability in regard to the level of democracy of the political system.

The results of the analysis are reported in Table 4. The findings display that an evolutionary thesis based on the analysis of cross-national data for a fixed time point is misleading. Our findings support the hypothesis that MDCs experience more frequent and/or larger shifts between more and less democratic systems as opposed to LDCs and ADCs.

In fact, if some of the extreme cases seen in Figure 2, which is the plot of instability scores on the logarithmic values of average energy consumption per capita, are dropped, the

Table 4 Regression of Democratic Instability Scores on Economic Development

Regression Coefficients

VARIABLE MODEL I, N=124 MODEL II, N=118 (including all cases) (excluding extreme cases)

Intercept .35 -.86

In AVENCON .55 .76

In AVENCON2 -.05 -.07

R2 .12 .19

Significant at .01 level.

Significant at .01 level.

32

Zehra F. Arat

Figure 2 The Plot of Democratic Instability Scores on Economic Development Level

3.00

2.50 ?

Trinidad

- 2.00 0

0 •

1.50 Upper Volta 0 0 0 0

* * . 0. * Kuwait

1.0000

.75 0 * *

.50 * 0 * Senegal . a

.25- Liberia Albania

2 3 4 5 6 7 8 9 Economic Development

model provides a better fit to the data. When six cases-Kuwait, Trinidad, Upper Volta, Senegal, Albania, and Liberia-are excluded from the analysis, the model explains 19 percent of the variance in democratic instability scores, which is 7 percent more than the variance explained by the model with all cases. It is reasonable to eliminate these cases, since Kuwait and Trinidad have unexpectedly high instability scores for their economic development level. Considering their performance on other socioeconomic indicators, we can argue that these two countries are misplaced on the economic development paradigm by our measure. On the other hand, Upper Volta, as an LDC, experiences unexpectedly high instability. Senegal, Liberia, and Albania, as MDCs, experience extremely low levels of democratic instability and compose another group of extreme cases which require special treatment.

Conclusion

The two major findings from the analyses in this paper provide sufficient evidence to question the modernization theory, which considers democratic development as an evolutionary phenomenon. First, a longitudinal analysis yields widely varied relationships between levels of socioeconomic development and democracy. It is clear that democracy is not a one-way ladder that countries climb as their economy and social structures develop.

33

Comparative Politics October 1988

Second, shifts on the scale of democracy are observed in a considerable number of countries. They tend to be highest in countries which are located near the middle of the development scale. These countries have experienced high rates of democratic instability within the 1948-1977 period.

Therefore, a positive correlation between economic development and democracy displayed by the cross-sectional data, but not confirmed by the longitudinal data, supports the early argument of Lipset and Coleman. These authors identified economic development as a necessary but not a sufficient condition of democratic development. However, those who attempted to test their hypothesis neglected this major observation, treated economic development as almost the single determinant, and limited their study and discussions to a linear/curvilinear argument. Thus, thirty years later, they left the same question unaddressed and largely unanswered: What might be the other conditions of democracy? Or, as this study points out, which components of middle range development and what other factors might be related to the destabilization of democracy?

During the last decade, however, a new wave of interest has arisen to look beyond modernization and economic development. Collier's study on Latin American countries identified the timing of development as an important factor.24 Bollen developed regression models to study the impact of the size of Protestant population and the extent of state control over the economy, in addition to the timing of development and economic development level.25 Linz and his coauthors made a major contribution when they studied the impact of the efficiency and effectiveness of the system on its breakdown.26 In an effort to explain MDCs' inability to maintain their once promising democratic systems, the author identified inefficiency and ineffectiveness of governments in recognizing the social and economic rights of their citizens, which eventually threatens the system's legitimacy.27 These quantitative and historical studies, combined with theoretically inspired case studies,28 promise more qualified answers to the questions of democratization and instability.

NOTES

1. Barrington Moore, Jr., Social Origins of Dictatorship and Democracy (Boston: Beacon Press, 1966); and Albert Soboul, The French Revolution 1787-1799 (New York: Vintage, 1975). 2. Gabriel A. Almond and Sidney Verba, The Civic Culture (Boston: Little, Brown, 1963); and Daniel Lerner, The

Passing of Traditional Society (New York: Free Press, 1958). 3. Almond and Verba. 4. Karl W. Deutsch, Nationalism and Social Communication: An Inquiry into the Foundations of Nationality, 2nd ed.

(Cambridge, Mass.: M.I.T. Press, 1966), and Lucian W. Pye, Aspects of Political Development (Boston: Little, Brown, 1966). 5. Seymour M. Lipset, "Some Social Requisites of Democracy: Economic Development and Political Legitimacy,"

American Political Science Review, 53 (March 1959), 69-105; and Political Man: The Social Bases of Politics (Garden City: Doubleday, 1960). 6. Lipset, "Some Social Requisites of Democracy," p. 86. 7. James S. Coleman, "Conclusion: The Political Systems of the Developing Area," in G. A. Almond and J. S.

Coleman, eds., The Politics of Developing Areas (Princeton: Princeton University Press, 1960), pp. 532-581. 8. Ibid., p. 544. 9. Philip Coulter, Social Mobilization and Liberal Democracy (Lexington: Lexington Books, 1975); Phillips

Cutright, "National Political Development: Its Measures and Analysis," American Sociological Review, 28 (April 1963), 253-264; Phillips Cutright and James A. Wiley, "Modernization and Political Representation, 1927-1966," Studies in Comparative International Development, 5 (1969), 23-41; Arthur K. Smith, Jr., "Socioeconomic

34

Zehra F. Arat

Development and Political Democracy: A Causal Analysis," Midwest Journal of Political Science, 13 (February 1969), 95-125.

10. Deane E. Neubauer, "Some Social Conditions of Democracy," American Political Science Review, 61 (December 1967), 1002-1009; and Robert W. Jackman, "On the Relation of Economic Development to Democratic Performance," American Journal of Sociology, 17 (August 1973), 611-621.

11. Arend Lijphart, Democracies: Patterns of Majoritarian and Consensus Government in Twenty-One Countries (New Haven: Yale University Press, 1984), p. 38; and Samuel P. Huntington, "Will More Countries Become Democratic?," Political Science Quarterly, 99 (Summer 1984), 193-218.

12. In addition to the measures by Coleman, Cutright, Neubauer, Cutright and Wiley, Smith, Jackman, and Coulter, other authors developed various measures of democracy. Some major ones are Irma Adelman and Cynthia Taft Morris, "A Conceptualization and Analysis of Political Participation in Underdeveloped Countries," Part 2 of the Final Report to the Agency of International Development, February 12, 1971 (mimeographed); Robert A. Dahl, Polyarchy: Participation and Opposition (New Haven: Yale University Press, 1971); Arthur Banks, "Correlation of Democratic Performance," Comparative Politics, 4 (January 1972), 217-231; Fitzgibbon and Johnson Index as reported by Kenneth F. Johnson, "Research Perspectives on the Revised Fitzgibbon-Johnson Index of the Image of Political Democracy in Latin America, 1945-75," in J. W. Wilkie and K. Ruddle, eds., Quantitative Latin American Studies: Methods and Findings, Statistical Abstract of Latin America, Supplement 6 (Los Angeles: UCLA, 1977), chap. 5; Kenneth Bollen, "Political Democracy and the Timing of Development," American Sociological Review, 44 (August 1979), 572-587; and Kenneth Bollen, "Issues in the Comparative Measurement of Political Democracy," American Sociological Review, 45 (June 1980), 370-390.

13. For a detailed explanation of the operationalization procedure and the actual scores assigned to individual countries, see Zehra F. Arat, "The Viability of Political Democracy in Developing Countries" (Ph.D. diss., State University of New York at Binghamton, 1984).

14. Data referring to the components of participation and competitiveness were obtained from Arthur Banks, The Cross-National Time Series Data Archive, provided by the Center for Social Analysis, State University of New York at Binghamton.

15. Charles L. Taylor and Michael C. Hudson, World Handbook of Political and Social Indicators, 2nd ed. (New Haven: Yale University Press, 1972).

16. Data for social unrest variables, except for death from domestic violence, were obtained from Banks, Cross-National Time Series Data Archive. The numbers for death from domestic violence were derived from Taylor and Hudson, Table 3.4. Data for the 1970s on this variable were kindly released by Charles Taylor upon the request of the author.

17. Raymond Duvall and Michal Shamir, "Indicators from Errors: Cross-National, Time-Series Measures of the Repressive Disposition of Governments," in C. L. Taylor, ed., Indicator Systems for Political, Economic and Social Analysis (Cambridge, Mass.: Oelgeschlager, Gunn, and Hain Publishers, 1980), pp. 105-182.

18. See Arat, chap. 3, for a discussion of the adjustments made before this stage. 19. The pitfalls of using a single indicator of a complex concept like socioeconomic development or modernization

and the changes in the implications of such indicators over time have been discussed and displayed by Arthur Banks, "An Index of Socio-Economic Development 1869-1975," Journal of Politics, 43 (1981), 390-411. However, Andrew J. Sofranko, Michael F. Nolan, and Robert C. Bealer, "Energy Use and Alternative Measures of Societal Modernity," Sociology and Social Research, 59 (July 1975), 301-317, consider energy consumption per capita as a reliable and valid indicator of modernization, defined by Marion Levy in 1966 in terms of the use of inanimate sources of energy. This is a conceptualization which holds valid for a long period in human history. Levy states: "My definition of modernization hinges on the uses of inanimate sources of power and the use of tools to multiply the effect of effort. I conceive each of these two elements as the basis of continuum. A society will be considered more or less modernized to the extent that its members use inanimate sources of power and/or tools to multiply the effects of their efforts. Neither of these elements is either totally absent from or exclusively present in any society." Sofranko, Nolan, and Bealer, p. 302. On the basis of data from seventy-four countries and for five different commonly used indicators of modernization, Sofranko, Nolan, and Bealer conclude that "all five societal measures correlate quite highly with each other and, in an empirical sense, are not highly distinct from one another. The overall significance of the findings is that all of the societal measures, whether they are labeled modernization, societal differentiation, human resource development, or economic development, are apparently not measuring different aspects of societies, but rather some general underlying dimension. Considering data availability, coverage, and accessibility, an argument is presented for use of 'energy consumption' as the measure of societal modernity -development." Ibid., p. 301. The energy consumption data utilized in this study have been obtained from Banks, Cross-National Time Series Data Archive.

20. See Arat, Appendix B, for the results of individual cross-national and longitudinal regression analyses. 21. The number of years included is determined by the independence date of the country and the availability of data. 22. Dankwart A. Rustow, "Transitions to Democracy: Toward a Dynamic Model," Comparative Politics, 2 (April

1970), 337-363. 23. Huntington, p. 210.

35

Comparative Politics October 1988

24. David Collier, "Timing of Economic Growth and Regime Characteristics in Latin America," Comparative Politics, 7 (April 1975), 331-359.

25. Bollen, "Political Democracy and the Timing of Development." 26. Juan J. Linz and Alfred Stepan, The Breakdown of Democratic Regimes (Baltimore: The Johns Hopkins

University Press, 1978). 27. Zehra F. Arat, "Human Rights and Democratic Instability in Developing Countries," Policy Studies Journal, 15

(September 1986), 158-172. See also "The Viability of Political Democracy in Developing Countries," pp. 84-158. 28. See Larry Diamond, Seymour M. Lipset, and Juan Linz, "Developing and Sustaining Democratic Government in

the Third World," paper presented at the 1986 Annual Meeting of the American Political Science Association, Washington, D.C., and several case studies conducted under the same research project.

36