“We [Houthis] promise the Saudi regime that the next operation will be wider and more painful if the blockade and aggression continues,” a Houthi spokesperson stated.
Despite the Houthis, a rebel group fighting Abdrabbuh Mansur Hadi’s government, admitting responsibility behind the Saudi oil attack, Saudi Arabia has still implicated Iran as the perpetrator. However, Iran maintains its position that it was not involved in the recent attack on Saudi Arabian facilities and directed blame for the escalation on Saudi Arabia due to their continued campaign in Yemen.
While the statements between Saudi Arabia and Iran are just a part of state competition dynamics, the Houthis’ demands necessitate a need for further inquiry. The Houthis’ rhetoric suggests that their operations will yield great devastation should attacks continue, but they are willing to halt further attacks in exchange for a ceasefire by Saudi Arabia and its coalition allies. In other words, this is an opportunity to achieve peace and end the five year Yemeni Civil War.
The following questions are the centerpiece of this analysis: Are the campaigns by the coalition allies actually working in lessening Houthi control and retaliation, or do they encourage the Houthis to rebel more often than they already would have? How does the data and analysis inform the coalition allies’ decision on taking the Houthis’ offer, and what are the implications of this?
However, should the coalition accept the Houthi offer? To unravel this central question, a panel regression model based on time-series and cross-sectional datasets of the number of coalition allies’ campaigns and Houthi retaliation in Yemen was conducted to measure the association between coalition aggression and Houthi retaliation.
Hypotheses:
There are three scenarios that this analysis may present. Each scenario yields its own unique discussion and answer to the central question of whether or not coalition forces should take the Houthis’ demands for a ceasefire.
Hypothesis One: The relationship between coalition aggression (independent) and Houthi retaliation (dependent) is positive.
Hypothesis Two: The relationship between coalition aggression (independent) and Houthi retaliation (dependent) is negative.
Hypothesis Three: There is no relationship between coalition aggression (independent) and Houthi retaliation (dependent).
If the first hypothesis is supported by the data, then it can be assumed that the number of Houthi retaliations are increasing as the coalition forces continue to increase their aggression via drone strikes and airstrikes. With an increasing number of retaliations, it may suggest greater lethality and more casualties and injuries in the war. Therefore, in light of the Houthi demands, it becomes evermore prudent to abide by the demands of a ceasefire in Yemen by the Houthis in an effort to reduce the bloodshed.
If the second hypothesis is supported, the coalition forces may have the advantage over the Houthis. While the coalition forces may seemingly have a leverage of more drone strikes and airstrikes in such negotiations, the Houthis may also cite the fact that there are mass civilian casualties because of the coalition’s leverage. If negotiations are to happen, it is important the coalition allies understand that their leverage comes at the expense of continued civilian casualties. Leverages only work insofar as their costs are beared in mind. Therefore, it is exceedingly prudent that peaceful negotiations and standards should occur.
If the third hypothesis is supported, one can expect that coalition campaigns have little to no effect on Houthi retaliation. If that’s the case, it would be extremely prudent for the coalition allies to agree to the Houthis’ demand, lest the high civilian death toll and the failure to attain peace in Yemen.
Data Methodology:
This analysis derives data from Armed Conflict Location & Event Data project , The Bureau of Investigative Journalism, and the Global Terrorism Database. Because the primary mode of coalition aggression is conducted through drone strikes, this analysis will focus on number of strikes as the primary independent variable. This analysis will also consider the number of Houthis killed in drone strikes as a secondary independent variable to measure the lethality of coalition aggression. Both variables are used in conjunction via a panel regression model where the dependent variable is the number of Houthi retaliations.
This analysis took a time-series of years between 2015 and 2017, as some datasets did not have adequate observations in 2018 and 2019 to pool from. The independent and dependent variables were then summed by month. Each month represents a cross-sectional data point in which the dependent and independent variables are associated with that particular point of time. However, because this article is examining the relationship between the dependent and independent variables over a time-series, a panel regression analysis is required rather than a typical multiple linear regression model.
Because this analysis examines drone strikes, there is a quasi-randomness. Drone strikes are subject to “equipment failure, operator errors or weather”, which causes variability in the times a drone may strike a target. Further, many drones are used purely for intelligence and are unable to fire. Drone operators must then wait for a drone with attack capabilities to strike, and in the time it takes to do so, the target has enough time to escape. These factors all influence when a drone operator takes a strike, making the variability of drone strikes particularly high. Additionally, in an effort to mitigate confounding variables, a fixed effect estimation was utilized in this panel regression.
Analysis and Discussion:
Figure 1: Parameter Estimates from Panel Regression Model
Based on the panel regression model, there seems to be a very weak positive association (0.0078) between the number of Houthis killed by coalition strikes and the number of Houthi retaliations. However, the most glaring issue is the high p-value not being significant at a standard α level of 0.05.
Moreover, the number of coalition strikes has a weak negative association (-0.0181) as it relates to the number of Houthi retaliations. Furthermore, it is statistically significant at α = 0.05. Such a negative association indicates that with an increase of coalition strikes, the number of Houthi retaliations actually goes down. Despite a negative association, further discussion is needed to reconcile the two independent variables and which hypothesis is supported.
Interpreting the p-values further is essential to understand which hypothesis seems to be more aptly supported. In regression models, any p-value calculated is meant to reject or fail to reject a null hypothesis that assumes the parameter of each variable is zero, that is, no relationship. In order to determine the meaning of the p-value, this analysis will operate at the standard alpha level of 0.05.
Since the number of Houthis killed by coalition strikes has such a high p-value (0.7599), it can be fairly concluded that this analysis has failed to reject a null hypothesis of a non-relationship. As such, the number of Houthis killed by drone strikes is not an appropriate predictor when looking at the association of changes in the dependent variable. On the other hand, though, the low p-value of the number of coalition strikes (0.0145) indicates that there should be a rejection of a non-relationship between number of drone strikes and Houthi retaliation at the 0.05 level. Moving forward, then, the number of coalition strikes is perhaps a greater predictor of the number of Houthi retaliations.
Figure 2: Number of Houthi Retaliations and Coalition Strikes over time (2015 to 2017)
Figure 2 represents a time-series of the sum of coalition strikes and Houthi retaliations as they correspond to an individual month in a given year. As the graph indicates, there seems to be an adequate amount of evidence over time, coupled with the panel regression, that illustrates a decrease in Houthi retaliation as the number of coalition strikes goes up. This is most evident in October of 2015. Inversely, with fewer coalition strikes comes an increase of Houthi retaliation, as shown in April of 2016. The association of higher coalition strikes and less Houthi retaliations continues into 2017, though not as pronounced. Regardless, it seems that when plotted over time, an increase of coalition strikes is associated with fewer Houthi retaliations.
With the conjunction of the evidence presented, then, the second hypothesis is most aptly supported. The negative association denotes the idea that coalition forces have leverage when it comes to negotiations of a ceasefire in Yemen. That is, if Houthis choose to renege on any kind of negotiation in favor of further violence, the United States and Saudi Arabia (along with other coalition members) may simply increase the number of drone strikes and airstrikes. However, just because this is the leverage the coalition forces hold does not mean they should opt to use it.
Because the coalition allies might think their leverage is enough to stop Houthi attacks, a full consideration on coming to the negotiating table with the Houthis may be discounted. But this is a flawed argument. This kind of logic and consistent force in Yemen by coalition allies implies that the costs of the war are perhaps low in exchange for lasting peace once the Houthis are eliminated. As a result, perhaps the costs of coming to the negotiating table with the Houthis may be too high. Put differently, the coalition allies may fear the Houthis will renege if concessions are not met to their standard, and therefore peace will be unsustainable. Yet, the costs associated with continued war are far greater than the costs of not meeting at the negotiating table. The coalition allies must understand that their leverage holds little weight if it is associated with high costs, especially when juxtaposed to the alternative of negotiations.
This analysis has not investigated the relationship of civilian casualties on either side and whether or not that has an effect on Houthi retaliation. As such, this analysis misses a larger part of Yemen that is cited for one of the greatest modern day collateral damages in the war against terrorism. The coalition allies bear severe costs of this war, that is, civilian deaths. There have been over 8,000 estimated civilian deaths in the duration of this war, followed by an additional 14 million who are at risk of starvation and disease. Though, it does not seem that the coalition allies have taken their own leverages and costs with great precaution.
The rising death toll is cause for reconsideration of counter-terrorism policies in Yemen. The escalation of war will only produce greater civilian casualties, and while the coalition allies may believe their leverages are high and costs are low, the fighting will only continue. Should fighting continue and no such attempts at negotiations are made, this becomes counterproductive to the goals of why the U.S. and other coalition allies are fighting a war against terror. It is absolutely essential that coalition allies take the Houthis’ concessions seriously by meeting at the negotiating table, lest to prevent further casualties. It is equally as important for the coalition allies to recognize that their leverage of drone strikes resulting in fewer Houthi retaliations is not the be-and-end-all due to the severely high costs associated with continued drone strikes.
Featured Image Source: Houthi Rebels Cheering