BACKGROUND: GROWTH AND CRISIS

The J-curve focuses on a period of strong growth, followed by a sharp economic downturn. While Latin America's past has been littered with periods of either economic decline or slow growth coupled with dictatorship, instability, and democratic breakdowns, the years since the new millennium began have been boom years for most countries in the region. In sharp contrast to the 1990s, when growth averaged only 1.7% in the seven largest countries, which collectively account for 91% of the region's GNP, it increased only to 2% by 2003. Beginning in 2004, however, growth was steadily above 6% annually. Inflation, long a serious problem in Latin America, declined from an annual average of 35% in 1991 to about 5% beginning in 2003. During this period, real exchange rates appreciated by nearly 40% and stock markets prices quadrupled (Izquierdo and Talvi Reference Izquierdo and Talvi2009, 5–6).

As of this writing, it is impossible to say how severe the current crisis will turn out to be or how strong its impact on Latin America will eventually become. Yet, this much is already clear: “The triple shock in external drivers—industrial country recession, a severe drop in commodity prices and terms of trade and FP [financial precarization] … has had stark implications for LAC's [Latin America and Caribbean] growth forecasts” (Izquierdo and Talvi Reference Izquierdo and Talvi2009, 16). Macroeconomic stabilization will depend to a great degree on how quickly recovery will occur in the advanced industrial nations. If recovery resumes in 2009, the Inter-American Development Bank predicts that Latin American countries will be able to weather the storm. On the other hand, if the crisis is protracted in the advanced industrial world, the negative economic outlook will likely endure for years to come.

The social costs of the economic crisis, however, are already deep. Deteriorating standards of living as a consequence of the decline in remittances and increasing unemployment are prominent. Indeed, falloffs in remittances have been observed in many high-migration countries (SELA 2009), resulting in a reversal of more than a decade of steady growth. Moreover, according to the latest predictions, as a consequence of the economic crisis, between 2.3 and 3.2 million more people will likely lose their jobs in the Latin American region in 2009, joining the 15.9 million who were already unemployed there in 2008 (ILO 2009). Thus, it is reasonable to explore whether growing economic hardship will affect citizens' views on their lives, politics, and democracy.

THE STATE OF MIND OF LATIN AMERICANS ON THE EVE OF THE CRISIS

The survey data that were gathered on the eve of the crisis allow us to develop some informed speculation as to how the crisis, depending of course on its severity and how it is managed in each country, might affect the state of mind of the population in Latin America. We assess how important economic conditions are, at the individual and national levels, for shaping core attitudes linked to democracy. Our goal is to envision what the likely consequences of failing to cope with the economic crisis might be for democracy. In particular, we examine whether deteriorating personal economic conditions and poorly performing national economies are likely to produce more discontented citizens and more individuals with less confidence that national problems can be solved effectively through democratic institutions and within the bounds of liberal democracy.

Specifically, we explore whether bad economic times and disappointment with how the state deals with the crisis might translate into more citizens being dissatisfied with their lives and skeptical that elections are effective in securing change in their country. If those predictions are borne out, then we further hypothesize that citizens are increasingly likely to oppose representative democracy. Finally, we then speculate that disaffected citizens opposing representative democracy may also favor authoritarian alternatives to solve national problems such as the current economic crisis. In this article, we seek to provide answers to these questions.

Methodologically, because we are interested in assessing how the personal and national economy affects citizens' attitudes, we relied on multilevel modeling techniques. Hence, unlike the conventional approach used in comparative politics of considering either individuals or nations as units of analysis, by employing the newer trend of multilevel analysis, we simultaneously took into account how individual economic traits and characteristics of the national economy influence their state of mind. To do this, at the individual level, we assessed the effects of objective and subjective measures of economic well being. Economic literature and evidence from Latin America in particular indicate that economic well being is multidimensional (Gasparini et al. Reference Gasparini, Escudero, Marchionni and Olivieri2008), suggesting that more than one measure is needed to explore its political effects. For example, Davies suggests that subjective assessments of personal economy rather than objective well being are what matters the most for political discontent; however, it is far from clear in the literature whether being poor and/or feeling poor are what matter the most for the formation of citizens' political views. For this reason, we took into account in our statistical analysis two objective measures of individual economic well being: the level of household wealth and the respondent's occupational status. The wealth index was constructed on the basis of a list of household-asset indicators in the survey using Principal Component Analysis (PCA) and a quintiles-of-wealth variable was then computed.Footnote ¹ Respondent occupational status was determined by three indicator variables: whether the person is employed, unemployed (i.e., actively looking for a job), or out of the labor force (i.e., student, homemaker, retired, or permanently disabled). Individuals' subjective economic well being was measured by an item in the survey that reads as follows: the salary that you receive and total family income is (1) good enough for you, you can save from it; (2) just enough for you, so that you do not have major problems; (3) not enough for you, you are stretched; or (4) not enough for you, you are having a hard time. Other things being equal, we expected that individuals at the bottom of the economic ladder (i.e., those falling in the first quintiles of wealth), the unemployed, and those self-reporting economic stress would show more negative attitudes toward democracy.

At the country level, we looked at the effects of economic growth and the level of economic development measured by the GDP per capita.Footnote ² Because at the country level our sample size is relatively small (N = 22), in our statistical analysis we tested the effect of these two macroeconomic indicators one at the time. We expected that individuals living in countries that have experienced relatively low economic growth would show more negative attitudes toward democracy. Similarly, we expected individuals living in the less developed countries in the region to express greater political discontent, suggesting that if the current economic crisis results in negative economic growth and consequently in setbacks in national economic development, as will most likely be the case in many countries in Latin America and the Caribbean, then the crisis might lead to an unhealthy political mood for democracy, to use Davies's terminology.

In the following section, we present the results of our multilevel analysis for three measures of citizen state of mind on the eve of the economic crisis: life satisfaction, trust in elections, and opposition to representative democracy. In addition, we test the hypothesis that lower levels of support for representative democracy might lead to higher support for non-democratic alternatives. Statistically significant findings are displayed graphically; full regression results are available from the authors upon request.

Life Satisfaction

The empirical literature suggests that countries with higher levels of life satisfaction make stronger democracies (Inglehart Reference Inglehart1990; Inglehart and Welzel Reference Inglehart and Welzel2005).Footnote ³ Here we analyze the extent to which economic conditions determine how satisfied individuals are with their lives using the following item in the AmericasBarometer: “In general, how satisfied are you with your life? Would you say that you are … ? (1) very satisfied, (2) somewhat satisfied, (3) somewhat dissatisfied, or (4) very dissatisfied?”

The results of our multilevel analysis are summarized graphically in Figure 1.Footnote ⁴ All the independent variables included in the analysis are listed on the vertical axis. The horizontal line set at the zero value serves as a reference point to indicate whether a variable is statistically significant or not. The lines stretching to the right and to the left of each dot, or standardized regression coefficient, correspond to 95% confidence intervals. If confidence intervals do not overlap with the vertical zero line, a result is statistically significant; hence, confidence intervals to the right and left of the vertical zero line depict positive and negative statistically significant effects, respectively. We examine the impact of individual- and country-level economic factors on each of the dependent variables, in this case life satisfaction, while holding constant individuals' years of schooling, age, sex, and the size of their places of residence.

Figure 1.

Life Satisfaction

Non-linear multilevel model with random effects for the intercept (standardized coefficients)

Sources: LAPOP AmericasBarometer and UNDP (Human Development Report 2007/08)

Economic development corresponds to the national GDP per capita

The state of the personal economy and the degree of economic development at the national level strongly determine the extent of individuals' life satisfaction (see Figure 1). At the individual level, we find that both objective and subjective personal economic conditions have an impact on citizens' degree of life satisfaction, although feelings of personal economic insecurity have a stronger effect. Specifically, our findings indicate that the poor and unemployed, but especially those who think that their incomes are either not enough and are stretched or that they are having a hard time, are more dissatisfied with their lives. Moreover, we find that individuals living in the most developed countries in the region are more likely to be satisfied with their lives, ceteris paribus. These individual- and country-level results are consistent with recent worldwide cross-country evidence (IDB 2008). On the whole, our findings suggest that as poverty and unemployment deepen and more people are having a hard time making their ends meet, more individuals will be discontented with their lives in Latin American and Caribbean countries; in addition, our country-level finding shows that, irrespective of one's objective and subjective economic conditions, lower levels of life satisfaction are strongly associated with poor national economic performance.

Trust in Elections

We theorize that the current economic crisis is also likely to erode citizens' trust in elections since economic performance has long been seen as impacting political legitimacy (Lipset Reference Lipset1960; Easton Reference Easton1965; Norris Reference Norris1999; Booth and Seligson Reference Booth and Seligson2009). Recent empirical evidence from eight Latin American countries suggests that the performance of democracy, including the health of the economy, is an important determinant of citizens' views on their political system (Booth and Seligson Reference Booth and Seligson2009). In this article, based on survey data gathered in 22 Latin American and Caribbean nations, we evaluate the impact of personal economic conditions and the performance of the national economy on citizens' confidence in the electoral process, a core institution of democratic systems. The corresponding survey item in the LAPOP survey reads as follows: “To what extent do you trust elections?” (based on a 1–7 scale of “not at all” to “a lot”).

Economic conditions at the individual and country level have an important impact on citizens' level of trust in elections (see Figure 2). Our ordered logistic multi-level individual-level findings indicate that irrespective of one's level of wealth, individuals who perceive that their families are struggling to meet their economic needs are less likely to have confidence in the electoral process. In addition, we find that at the national level, economic development and rate of economic growth matter for individuals' level of trust in elections; citizens living in countries that have enjoyed better economic performance are more likely to believe in using institutionalized mechanisms of political participation. We expect, therefore, that trust in elections is likely to decline as individual and national economic conditions worsen as a result of the economic crisis. Our results are particularly worrisome because on the eve of the crisis trust in elections was already weak in the Latin American and Caribbean region, averaging only 4.03 points on a 1 to 7 scale.

Figure 2.

Trust in Elections

Economic development corresponds to the national GDP per capita

Sources: LAPOP AmericasBarometer and UNDP (Human Development Report 2007/08)

National economic growth corresponds to the average annual GDP per capita growth rate for the period 1990–2005

Non-linear multilevel model with random effects for the intercept (standardized coefficients)

Opposition to Representative Democracy

Do the economic conditions at the individual and national level that we have shown to produce lower levels of lack of trust in elections also yield lower support for representative democracy? In order to explore this question, we analyzed the following item in the AmericasBarometer survey: “The people should govern directly and not through elected representatives. How much do you agree or disagree with this statement?” (1–7 scale). As shown in Figure 3, both respondents in the first quintile of wealth as well as unemployed individuals are significantly more likely to oppose representative democracy, allowing us to speculate that if poverty and unemployment worsen as a result of the economic crisis, opposition to representative democracy is likely to increase. Moreover, as we found in the analysis of trust in elections, the level of economic development at the national level also matters for the extent of opposition to democratic representation; regardless of their individual-level characteristics, citizens living in the most developed countries in Latin America and the Caribbean are more likely to prefer representative democracy. The average level of opposition to representative democracy in 2008 before the onset of the crisis, however, was relatively low; on the 1–7 scale, the regional average was only 3.45, although we also find significant differences across countries with some countries expressing somewhat higher average support (with a maximum of 4.0) and others far lower than this regional average (with a minimum of 2.4).

Figure 3.

Opposition to Representative Democracy

Non-linear multilevel model with random effects for the intercept (standardized coefficients)

Sources: LAPOP AmericasBarometer and UNDP (Human Development Report 2007/08)

Economic development corresponds to the national GDP per capita

Opposition to representative democracy is not in and of itself an indication of anti-democratic sentiment because respondents could be thinking favorably of a plebiscitarian form of government. In fact, however, in Latin America and the Caribbean at least, low support for representative democracy is strongly associated with a preference for non-democratic alternatives. In the survey we asked: “There are people who say that we need a strong leader who does not have to be elected. Others say that although things may not work, electoral democracy, or the popular vote, is always best. What do you think?”Footnote ⁵ As Figure 4 shows, other things being equal, as support for direct popular participation increases, the probability of support for a strong leader who does not have to be elected is substantially higher. Specifically, while on average the probability of supporting an unelected leader is only 10% among those who “strongly disagree” with direct popular participation, those who support direct participation are 22% more likely to prefer authoritarianism to democracy.

Figure 4.

Opposition to Representative Democracy as a Predictor of Support for Authoritarianism

Sources: LAPOP AmericasBarometer

Results based on a non-linear multilevel model with random effects for the intercept (control variables included)