Is It Possible to Punish Acts of Democratic Backsliding?

Let’s take a look to see if the withholding of foreign aid could be a way to punish a president that doesn’t respect the democratic process.

Foreign aid can be a nice gesture to help a country that either experienced some type of disaster, or needs some goods, financial, or labor support. Smaller countries tend to rely on foreign aid to keep their economy afloat. This is the case for Hungary where their government was labeled as a "hybrid regime of electoral autocracy.” This was enough to cause the EU to threaten the foreign aid they were willing to give Hungary. This left Viktor Orban to either bring Hungary back to a democracy or lose a substantial amount of foreign aid.

This shows a clear respect for democracy from the EU, which was willing to punish Viktor Orban for his Executive Aggrandizement. But what if a country has their own personal interest with another country that is not a democracy? The Biden administration threatened to withhold 300 million dollars out of a their 1.3 million dollar foreign aid budget for Egypt’s military because of human rights violations. More specifically, the imprisonment of political enemies, no matter how small. Not only did the Biden administration not withhold the amount they said they would after Egypt showed little to no attempt to fix their human rights violations, but also they made a 2.5 billion dollar weapons deal with Egypt. Egypt suffered no consequences for ignoring the threat of the Biden administration to withhold foreign aid.

This article mentions that electoral accountability alone is not enough to uphold a democracy and that term limits are a neutral necessity for a democracy. While this article mentions that a lack of term limits, and a lack of bipartisan roles in government leads to more polarization of political factions. Based on these research papers, presidential term limits and the abolishment of said limits could be a great way to measure an instance of democratic backsliding.

Going back to my initial question, Is a country that experiences an instance of democratic backsliding likely to receive less foreign aid as a result? I think we’ve gathered enough information to start testing my hypothesis. So let’s start testing!

We have a clear dependent variable (amount of foreign aid received) and independent variable. (attempted or successful abolishment of presidential term limits) These were tested by first randomly selecting 20 countries that experienced the form of executive aggrandizement mentioned above, and 20 countries where presidential term limits were respected. I then looked at a four year period. For backsliding countries I looked at the two years prior to, and the two years after the backsliding occurred. As for countries that respected presidential term limits, I took the first four years of the sitting president’s term and split it into the first two years, and the latter two years. After that I found the data for how much foreign aid each country received during those years. Averaged them by the prior two years and latter two years, then accounted for the percent difference between these two groups.

Next I found the average income for each year that I want to test for, and then split those years in the same way, then accounted for the percent difference between them. This is to control for whether increased or decreased average income per capita would affect the amount of foreign aid a country receives. The last variable I wanted to control for was displaced persons within a country during this time frame. There was no data set available with the information I needed, so I had to search the internet to find whether or not there were persons displaced for each year I was measuring for. Since there was no data set, I made this a binary variable where if any persons were displaced, it’s a yes, (1) and if there weren’t, it’s a no. (0) This means I did not account for how many persons were displaced, which is the main reason a country might receive more foreign aid than usual. This is important to point out, so we know how meaningful my measure of displaced persons is, and by that, I mean it could’ve been more meaningful if I had access to the appropriate data.

When conducting a T test, we find that the p-value is 0.04, and the T statistic is 2.15. I would call this statistically significant. We can reject the null hypothesis.

Now we've found that there is a statistical significance in the relationship between allocation of foreign aid and whether or not a country experienced backsliding. So let's try to control for something. The most common type of foreign aid is meant to promote development and reduce poverty. It's plausible that as a country becomes wealthier, or more poor, foreign aid allotment could decrease or increase accordingly. So, the question is, does my data remain significant if we control for the average income per capita?

After controlling for average income per capita, I found that the p-value remains almost the same, which can be rounded up to 0.04. The T statistic remains relatively the same too, at -2.2. It would appear my hypothesis is still significantly significant even after controlling for average income per capita. Great!

Let’s control for one more variable before we feel confident in saying there really is statistical significance between foreign aid and democratic backsliding. Another significant determiner in deciding where foreign aid goes is something that would cause people to be uprooted from their homes, such as natural disasters or areas becoming too unsafe to live in caused by other means. As mentioned above, my data for displaced persons is a binary variable that does not take into account how many persons were displaced.

The p-value for instances of backsliding after controlling for displaced persons is 0.044. A slight increase, but still low enough for me to believe there is statistical significance. The T statistic is -2.1. There still seems to be statistical significance! But what if we control for both average income per capita and displaced persons at the same time as controlling for backsliding?

While controlling for everything we wanted to test for, the p-value and the T statistic don’t change by much. P-value is 0.04, T statistic is -2.2. So even when controlling for everything, we can reject the null hypothesis. There is statistical significance for the amount of foreign aid a country receives when looking at whether or not a country backslid. The income and displaced people variables didn’t seem to change the outcome of my hypothesis by much.

As it turns out, my hypothesis was right! There is statistical significance in the relationship between democratic backsliding and foreign aid allotment. Splitting foreign aid allotment between backslid and non-backslid countries showed a strong relationship in how much foreign aid was allotted. Even when controlling for average income per capita and whether or not persons were displaced, the data suggest that there is statistical significance. Therefore I feel confident in saying foreign aid allotment is one way that a country may punish another country that experiences backsliding.