Introduction

1970 is when the "third wave" of democracy is known to have started, and the 1990s is when the fourth wave of democracy hit the world. It is quite fascinating to see the shifts in countries that may have been dictatorships or authoritarian regimes before and see their political systems transform into democracies, often times presidential democracies.

Juan Linz, a political scientist, has an article that many reference when thinking about presidentialism, called the "Perils of Presidentialism." This article is what inspired my interest in examining presidentialism further, as Linz examines how features of presidential democracies may cause democratic backsliding. He mainly discuses the role of institutions in democratic backsliding and political fragmentation, such as how polarization is often more common in presidential elections and how that impacts political gridlock. He also discusses the stability of presidential leaders and how time limits and fixed terms impact policy for decades.

The breakdown of presidentialism is also quite relevant because of the United States' former president, Donald Trump. During President Trump's term, he showcased many of the reasons that Linz discussed in his article. For one, the attack on the nation's capitol on January 6th, 2021 was a clear example of a president using their powerful power to rouse a mob and attack a government. Democratic breakdown can also be examined further under former President Trump, with him running for presidency without having prior experience in the political system and no "allegiance" to the political system. President Trump also rallied up extremists and used harmful rhetoric to rally citizens against the system and diminish trust in the same system he was working for.

Thus, I sought to examine how presidentialism as a system impacts democratic backsliding, whilst comparing it with "non-presidential" systems, like assembly-elected and parliamentary systems of democracy.

Related Studies

Many prior studies have been conducted on the impact of presidentialism on democratic backsliding, such as the Annual Review of Political Science examining democratic backsliding. He analyzes Latin American countries, saying that countries that emerge from military dictatorships are far less likely to survive than presidential systems that emerge otherwise. Many of the presidential systems that emerged within the third and fourth wave of democracy are less likely to survive if they have a military or authoritarian legacy. This ties into my project because one of my rival hypotheses was the impact of military / authoritarian legacy on democratic backsliding.

There are other studies that examine the impact of polarization on democratic backsliding, examining countries around the world that experienced backsliding between 2000 and 2020. The study found that polarization rose in all of the countries that had some level of backsliding, including countries like Brazil and the United States, both of which are presidential democracies. It also directly examined countries that experienced backsliding that were not presidential democracies. This study also examines the breakdown of democratic institutions, such as when they are attacked and undermined. This was examined in the context of autonomy of electoral authorities as one example, which the US was a victim of between 2016 and 2017.

Our course book, How Democracies Die, also examines the democratic breakdown of the United states, focusing on how presidentialism has lead to populist notions arising, as well as forbearance and mutual toleration. These aspects of presidentialism can lead to its demise, because they lead to mistrust in government and institutional undermining, which leads to democratic breakdown.

Another study on the "Perils of Turkish Presidentialism" analyzes the new presidential system taking effect in Turkey as of 2019. This study analyzes how presidentialism in Turkey will create further political instability in the country, no matter who gets elected. It will also increase polarization and the marginalization of minorities, and if Erdogan, the current president, wins, Turkey may become an electoral authoritarian system, according to the authors.

These studies corroborate the need to study presidentialism as a system further, analyzing its success and failures and if it leads to democratic backsliding.

Hypothesis / Hypotheses

My hypothesis is that presidential democracies are more prone to democratic backsliding than other types of democracies. I am expecting to find that presidential democracies face more democratic backsliding than parliamentary and other types of democracies.

I am also examining the impact of GDP per capita of a country and the legacy of military / authoritarian regime in a country and the impact that it has on democratic backsliding.

Research Design

In my study, my independent variable(s) was presidential systems and other types of systems, such as parliamentary systems. My dependent variable was democratic backsliding, categorized by the level of electoral democracy in a country. My unit of analysis was also systems of democracy, because we're measuring the systems in relation to their level of electoral democracy, or backsliding. I denoted presidential countries by 1 and used 0 to account for other types of democracies, such as parliamentary and assembly-elected democracies. I tested these variables by examining countries in Sub-Saharan Africa to see what democracy was looking like in the 90s there. I collected data on the types of democracy by using data from the World Bank. I chose to examine a 10-year period, ranging from 1990 to 2000, to see the change in the level of electoral democracy within a longer period, because a country's system can entirely change within a 10-year timeline, as can their level of democracy. For this, I utilized the Varieties of Democracy dataset, which had a range of data on electoral democracy change through time.

The control variables or rival hypotheses that I examined were the GDP per capita of the country and the legacy of military / authoritarian regimes and their impact on democratic backsliding of a country. I examined those by collecting data from the World Bank on the GDP per capita during the year 1995, as it was right in the middle. I also collected data on whether a country had a legacy of authoritarian / military regime and how many years that legacy lasted.

Throughout my study, I noticed that there were missing pieces in my data collection, such as countries that were missing data on GDP or the type of democracy that existed in their country. I also noted that most of the countries I analyzed in my dataset were presidential, with only about 5 non-presidential countries. This is important to note because with only 5 countries to compare levels of electoral democracy to, the average and mean numbers can be slightly off because of the low sample.

Results

Through my study, I found quite fascinating results. As one can see, the histograms above show an overall increase in the level of electoral democracy between 1990 and 2000 in Sub-Saharan African countries, without even controlling for presidentialism. In 1990, most countries' level of democracy ranges between 0.0 and 0.3, and in 2000, the range increases with many countries' level of democracy is between 0.4 - 0.8. This positive trend continues to be showcased even whilst controlling for presidentialism, which is denoted as 1. EDIChange, which is the variable for the change in the level of electoral democracy between 1990 and 2000, displays an overall positive trend in the presidential realm, whereas in the non-presidential realm, it is sort of standstill. Non-presidential, being denoted as 0, has a significantly smaller sample size than presidential countries, which impacts its overall averages and increases between 1990 and 2000.

The evidence for my hypothesis leads me to conclude that I cannot reject my null hypothesis, because as the histograms and the linear regression shows below, there is an overall positive trend in the level of electoral democracy between 1990 and 2000 for Sub-Saharan African countries. The coefficient depicted below for presidential is at 0.124, showing that when going from non-presidential to presidential countries, the level of democracy increases. My p-value was not too high, showing that there is a statistically significant relationship, but the relationship itself was contrary to my original hypothesis. Thus, we fail to reject our null hypothesis, showing that because of my research design and the evidence we found, we cannot sufficiently say that presidentialism and democratic backsliding were related.

In terms of my control variables, I controlled for GDP per capita, which I controlled for in the year 1995. As the coefficient shows us, the relationship between GDP per capita and the level of electoral democracy is a negative one. The other control variable is the opposite, with the coefficient showing that the relationship between legacy of military and authoritarian legacy being positive. However, since both of these relationships have significantly high p-values, we cannot really conclude that the relationship is statistically significant, and thus our control variables are not super relevant in the study overall.

Concluding Thoughts

This project was useful in determining what needs to be analyzed further in terms of democratic backsliding in countries. Having concluded that my null hypothesis could not be rejected and that my evidence did not support my hypothesis, I've learned a few things about why it was so. To start, the choice in time period was flawed; a better time period to examine for democratic backsliding in sub-Saharan African countries would have been 1990 - 2010, because many countries started backsliding after US intervention and in the early 2000s, as illustrated by the studies I referenced earlier. Furthermore, the problem in sampling is a big issue because I realized that my sample held very few non-presidential countries with various democratic trajectories, which made it difficult to accurately find a directional effect for them. It also made it difficult to examine the problem of backsliding in non-presidential countries, as the sample had too few non-presidential countries. Beyond that, analyzing sub-Saharan Africa as a region maybe was not the strongest idea, because between 1990 - 2000, many African countries were 'developing' and presidentialism was perhaps not harmful to their development, but rather, beneficial.

For the future, other scholars and I could better analyze democratic backsliding between 2000 - 2015 because I would be curious to see how presidentialism as a system was faring a decade or so after the "fourth wave" of democracy in the late 1900s. Many US interventions also took place during that time period, such as Operation Enduring Freedom, which started taking place after the September 11th, 2011 attacks on the United States. With these interventions taking place, democratic breakdown would be an interesting variable to analyze and see if and how it endures through intervention and instability. Changing the sample in the case of the countries included in it would also be critical, as examining a region may not be the best idea. Examining countries that meet a certain level of democracy (e.g. 0.5) at the start of the time period would be more useful, as my project included countries that had just shifted to presidential democracies and had large improvements, and countries that had higher levels of electoral democracy at the beginning of the time period, and thus, increased very little during. Examining a variety of countries across the world that had a consistent level of electoral democracy at the start of the time period would be a better way to have more of a controlled sample of countries, as well as more variety with presidential and non-presidential countries.