Divided government in Latin America

ABSTRACT

The literature on divided government focuses on the relationship between divided government and its size and has found that governments are limited when divided government exists. Fewer studies have examined the conditions that will generate a divided government. Theoretical research argues that voters want to moderate power. This study attempts to determine if the same factors that predict a divided government in the US explain divided government in Latin America. We use a continuous measure of divided government in 15 Latin American countries from 1981–2010. We find that countries in LA with midterm elections moderate power during concurrent elections when inflation is increasing. However, they are more likely to vote in a unified government during midterm elections when there are poor economic conditions. Countries with no midterm elections, moderate power when the economy is doing well but are more likely to choose a unified government when inflation and unemployment are increasing.

KEYWORDS:

• Divided government
• limited government
• Latin America
• midterm elections

JEL CLASSIFICATION:

• D72
• E62
• F55

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ For a more detailed literature review of US divided government see Calcagno and Lopez (2012).

² We are studying what economic variables affect presidential and legislative electoral outcomes so we limit the sample to democracies as defined by Przeworski et al. (2000). One thing to note is that our 15 countries have had less experience with democracy relative to the United States. What makes our question about the factors that move a Latin American country more or less toward divided government interesting is that most of these countries have less than 30 years of experience with democracy, as most of them are part of the third wave of democratization (Huntington 1993).

³ Within the Latin American context, there is some concern regarding the relative strength of the legislature vis-à-vis the executive branch. Cox and Morgenstern (2001) note that in many Latin American countries where presidents have a lot of power, presidents still take into consideration the reaction of the legislature when forming policies. This is due to the fact that while these legislatures are weaker than in the US, they still have the power to change and/or reject proposals.

⁴ During part of the sample (1997–2004), Chile conducted midterm elections and is included as a midterm country during those years.

⁵ We also estimate our model using a dichotomous variable and find the results to be consistent. These results are available from the authors upon request.

⁶ For details on how this measure is calculated, refer to Calcagno and Lopez (2012).

⁷ We include a linear time trend to capture region-wide trends that occurred during our period of analysis such as the policy changes that came about from countries dealing with the debt crisis of the 1980s.

⁸ Since our sample of countries is relatively small, we employ a jackknife analysis of our samples to determine if any one country might be driving our results. The findings are similar to those reported in Table 2. However, none of the coefficients estimated for the midterm countries is consistently significant throughout the 6 iterations suggesting that for the smaller subset of midterm countries that the results may be driven by one country – Argentina.

⁹ Jackknife results for non-midterm countries show consistently significant results.

Additional information

Funding

This work was supported by the Center for Public Choice & Market Process.