Public Opinion, Democracy, and the Economy: Case Studies from the Southern Cone

Data and Methods

All data is derived from the 2009 Latinobarómetro public opinion survey and covers Argentina, Chile, Paraguay, and Uruguay5. Two independent variables were chosen to measure public evaluations of the economy, one measuring respondents’ evaluations of the national economy, and the other measuring respondents’ evaluations of their personal economic situation6. A total of three variables were used to measure overt support for democracy, and a total of six were employed to measure intrinsic support for democracy. The intrinsic support variables mirror Inglehart and Welzel’s (2003) “self-expression values” of life satisfaction, interpersonal trust, tolerance of diversity, public expression, and liberty and participation.

The operationalization of life satisfaction and interpersonal trust is the same in this study as in Inglehart and Welzel7. Tolerance of diversity is operationalized by tolerance of homosexual neighbors, the same indicator used by Inglehart and Welzel, except that in this study I have only employed this indicator, whereas Inglehart and Welzel combined it with a variable measuring tolerance of neighbors with AIDS to create a composite variable8. The operationalization of public expression in this study is different than that of Inglehart and Welzel, as their indicator (responses to the question of whether respondents had ever signed a petition) was not asked in the Latinobarómetro survey in these countries and years.

Instead, respondents’ answers as to the importance they placed on voting were used as a substitute for the purposes of this study. I believe that this indicator accurately measures what Inglehart and Welzel intended, as it is a form of public participation and expression carried out through an institutional mechanism, just as is signing a petition. In a similar fashion, the variable operationalizing liberty and participation was measured in this study by respondents’ beliefs as to the importance of a free media. This parallels Inglehart and Welzel’s use of respondents’ beliefs as to the importance of protecting the freedom of speech, and was used because Latinobarómetro did not ask this question for these countries and years.

Lastly, a composite variable was created in order to combine all five self-expression values9. This is of particular importance given that, as mentioned previously, Inglehart and Welzel (2006) found that all self-expression values tend to occur together. This composite variable therefore gives the most comprehensive view as to how these self-expression values may be associated with individuals’ evaluations of the economy.

Assuming the hypothesis of the previously mentioned literature that the existence of self-expression values are important to democracy, then the extent to which self-expression values and economic evaluations are associated may suggest important implications for democracy in the region, because a high association would likely indicate that permeation of self-expression values fluctuates with economic conditions. The potential volatility of the latter would then lead to volatility in the former, a condition that is clearly not conducive to democratic consolidation from a political culture perspective.

Analyses were conducted for each of the four countries individually (for each, N=1200), as well as all four countries combined (N=4800). The bivariate statistic gamma (G) was employed in order to measure the degree of association and directionality of the relationship between each of the dependent variables (overt support and intrinsic support for democracy indicators) and both independent variables (evaluations of national economic conditions and personal economic conditions).

All variables employed in this study are either ordinal or dichotomous, making the use of gamma appropriate. The use of gamma for the dichotomous variables is especially valid given that all of the dichotomous variables can be conceptualized ordinally.

Results

The results of the analyses are detailed in Tables 1 to 5:10

Table 1: Argentina

N=1200	Overt Support for Democracy			Intrinsic Support for Democracy (Self-Expression Values)
	Overt Support #1	Overt Support #2	Overt Support #3	Life Satisfaction	Interpersonal Trust	Tolerance of Diversity	Public Self-Expression	Liberty and Participation	Composite Variable
Economic Evaluation (National)	.050	.110**	-.227***	.228***	.302***	.027	.157***	-.078	.212***
Economic Evaluation (Personal)	.124**	.078	-.125**	.484***	.067	.060	.172***	.027	.268***

*=p<.05; **=p<.01; ***=p<.001

Table 2: Chile

N=1200	Overt Support for Democracy			Intrinsic Support for Democracy (Self-Expression Values)
	Overt Support #1	Overt Support #2	Overt Support #3	Life Satisfaction	Interpersonal Trust	Tolerance of Diversity	Public Self-Expression	Liberty and Participation	Composite Variable
Economic Evaluation (National)	.297***	.160***	-.079	.406***	.021	.115	.239***	.032	.275***
Economic Evaluation (Personal)	.226***	.039	-.053	.500***	.045	.096	.200***	.029	.298***

*=p<.05; **=p<.01; ***=p<.001

Table 3: Paraguay

N=1200	Overt Support for Democracy			Intrinsic Support for Democracy (Self-Expression Values)
	Overt Support #1	Overt Support #2	Overt Support #3	Life Satisfaction	Interpersonal Trust	Tolerance of Diversity	Public Self-Expression	Liberty and Participation	Composite Variable
Economic Evaluation (National)	.154***	.208***	-.094*	.273***	.094	.052	.087	-.132***	.095*
Economic Evaluation (Personal)	.082	.166***	-.069	.515***	.085	.079	-.027	-.103*	.154***

*=p<.05; **=p<.01; ***=p<.001

Table 4: Uruguay

N=1200	Overt Support for Democracy			Intrinsic Support for Democracy (Self-Expression Values)
	Overt Support #1	Overt Support #2	Overt Support #3	Life Satisfaction	Interpersonal Trust	Tolerance of Diversity	Public Self-Expression	Liberty and Participation	Composite Variable
Economic Evaluation (National)	.331***	.071	-159**	.273***	.168***	.076	.314***	.031	.267***
Economic Evaluation (Personal)	.339***	.033	-.036	.513***	.143**	-.016	.325***	-.047	.294***

*=p<.05; **=p<.01; ***=p<.001

Table 5: All Countries

N=4800	Overt Support for Democracy			Intrinsic Support for Democracy (Self-Expression Values)
	Overt Support #1	Overt Support #2	Overt Support #3	Life Satisfaction	Interpersonal Trust	Tolerance of Diversity	Public Self-Expression	Liberty and Participation	Composite Variable
Economic Evaluation (National)	.245***	.226***	-.226***	.289***	.212***	.130***	.244***	-.010	.254***
Economic Evaluation (Personal)	.181***	.084***	-.061*	.510***	.119***	.061*	.180***	-.015	.262***

*=p<.05; **=p<.01; ***=p<.001

The first important note to make regarding the results is that none of the overt support or intrinsic support variables reached 0.6, the commonly accepted threshold for determining a strong association11. Indeed, the highest associations across all countries occurred between life satisfaction and respondents’ evaluations of their personal economic situation, ranging quite close across all samples from .484 in Argentina to .515 in Paraguay. This association is not surprising, as a correlation between life satisfaction and, essentially, how much money one has, certainly possesses face validity. While only qualifying as a moderate association, this finding does suggest (especially given its consistency across all countries sampled) that the extent to which individuals are satisfied economically may also affect how satisfied they are with their lives more generally, and this life satisfaction is an important component to democratic culture if we accept Inglehart and Welzel’s findings. However, the association between evaluations of the national economy and life satisfaction is not nearly as cohesive, ranging from a low of .228 in Argentina to a high of .406 in Chile, and an average of .289 in all four countries combined. This suggests that life satisfaction is more greatly associated with personal economic conditions than national economic conditions, and that the relationship between this latter variable and life satisfaction is generally weak.

A focus on the overt support variables reveals a few interesting trends. First, there is a great degree of volatility regarding the degree to which each of the three overt support variables are associated with the two economic variables. In Argentina, the overt support #3 variable had the greatest degree of association of all other variables, and in fact reduced more than twice as much error in predicting evaluations of the national economy than did the next highest ranking variable12.

Nevertheless, its association remains a weak one. In Chile and Uruguay, the overt support #1 variable was the strongest predictor of both the independent variables, even reaching a moderate level of association in Uruguay. In these countries, the overt support #3 variable reduced prediction error only by trivial amounts, often even failing to achieve statistical significance. Further complicating one’s ability to find a coherent trend is that in Paraguay, the overt support #2 variable achieved the greatest association with both independent variables, yet was often statistically insignificant in the other countries sampled.

Returning to the results for the intrinsic support for democracy variables (that is, the self-expression values), very few consistent associations are present, let alone strong ones, except for the previously mentioned association between life satisfaction and personal economic evaluation. One consistency is that the coefficients for the tolerance of diversity variable were very small and failed to reach statistical significance in all countries. It did reach statistical significance when all countries’ data were pooled together, although this is likely due the very high number of cases in the sample, and even here the association is very weak. The liberty and participation variable also was consistently insignificant, except for in Paraguay, although here the association is again only very slight.

Interpersonal trust coefficients are highly volatile among the samples. Values range from being statistically insignificant in Paraguay and Chile, to weak positive associations in Uruguay, to a (albeit barely) moderate, positive association (.302) in Argentina when linked with evaluations of the national economy. However, when this variable was compared with evaluations of personal economic conditions in Argentina, its value plummeted dramatically and became statistically insignificant. The sample of all countries pooled together shows a weak association with interpersonal trust and both independent variables, with its association with national economic evaluations being a bit stronger than with personal economic evaluations.

Public self-expression was also very volatile across all countries, ranging from statistical insignificance in Paraguay, to weak associations in Argentina and Chile, to moderate associations in Uruguay with both independent variables. In all of these cases, values were similar for associations with both of the independent variables.

The composite variable, measuring all five self-expression values simultaneously, is of particular importance given that Inglehart and Welzel (2006) found that all self-expression values tend to occur together. The results indicate that the composite variable reduced error in predicting both independent variables by between about 21 and 30 percent, with the exception of Paraguay, where error was only reduced by 9.5 (when crossed with national economic evaluations) and 15.4 percent (personal economic evaluations). In all instances, the association was slightly stronger with personal economic evaluations than with national economic evaluations.

However, these findings suggest that self-expression values are only weakly associated with evaluations of economic conditions (national and personal), as none of the coefficients passed the 0.3 threshold for declaring a moderate association. At the same time, these results also suggest that self-expression values, when pooled together under a single variable, are more consistently associated with national and personal economic evaluations across countries than are indicators for overt support for democracy.

However, while more consistent, the strength of this association is not greater on average than is the strength of the association between overt support indicators and economic evaluations. Indeed, the coefficients for the highest overt support democracy variable for a given country were generally very close to the coefficients for composite self-expression values variable.Continued on Next Page »