The ongoing negotiations in Afghanistan are now officially one year old, but they have little to show for it. Secretary of State Mike Pompeo appointed Zalmay Khalilzad the US Special Representative for Afghanistan Reconciliation in September 2018. Khalilzad was on the cusp of a political breakthrough by late August 2019, indicated by what would have been a watershed visit with the Taliban at Camp David. However, President Trump abruptly cancelled the meeting, blaming a Taliban attack that killed an American soldier on September 5.
For the time being, violence has spoiled the peace process. Indeed, Taliban attacks against Afghan forces and civilians have persisted throughout the negotiations, even increasing in recent months. This raises the question: does the onset of a peace process impact the level of insurgent violence? Although still in its preliminary phase, the Afghanistan case provides an opportunity to explore this question and gain insight into the Taliban’s strategic use of violence.
Literature on the subject suggests that there is a correlation. Höglund and Darby both observe that violence often increases after the onset of negotiations. Darby explains that mutual distrust and attempts to maximize bargaining positions lead to more military offensives. Likewise, Höglund observes that violence during a peace process is often intentional, “well planned and timed.” She cites Pillar, who argues that as a deal draws closer, parties seek to strengthen their positions, which leads to increased violence.
Data provides a fuller picture, allowing for comparison of violence before and during the peace process. Khalilzad’s appointment marks the transition point between these two phases because it publicly signaled that the U.S. was serious about a conflict resolution. The Armed Conflict Location & Event Data Project (ACLED) provides data on Taliban attacks. Metrics for violence include number of attacks, number of kills, and lethality (kills per attack).
Figure 1 shows a clear increase in total kills perpetrated by the Taliban against coalition forces. These totals correlate strongly with the substantial increase in incidents: more attacks lead to more kills. The data shows that the Taliban largely target government forces over civilians, reflecting their ongoing effort to reduce government control. There is also a noticeable increase starting in January 2019. Interestingly, this jump occurred weeks after news broke that President Trump directed the Pentagon to begin planning troop withdrawals. Additionally, Khalilzad met with Taliban representatives in January, and they mapped out the framework for a peace deal. In response to a potential American exit, the data indicates that the Taliban, emboldened by the likelihood of a deal, increasingly stepped up the death tolls of their attacks.
Figure 2 measures the lethality of Taliban attacks and tells a more nuanced story. Lethality is measured by calculating the average kills per attack. In contrast to Figure 1, lethality against government forces noticeably decreased after January 2019. This is a puzzling difference, and the explanation for this opposite correlation is difficult to determine without further examination. On the other hand, lethality against civilians has risen in recent weeks. This increase is due to high-profile suicide attacks and bombings, which have grabbed newspaper headlines lately.
What does this tell us about the Taliban? It reinforces the view that the Taliban have pursued a “fight and talk” strategy in order to maximize their strategic advantage against both the U.S. and the Afghan government. With President Trump ready to withdraw and Khalilzad eager for a deal, the Taliban may reduce their attacks temporarily in order to bring the U.S. back to the table and reach an agreement. However, some still fear further intensification of violence. What is clear is that the start of negotiations marked a new phase in the war—perhaps the beginning of the end—and the Taliban adjusted their strategy accordingly.
James Suber is a student fellow with the Center for Security Policy Studies. He is a second-year graduate student in the International Security MA at the Schar School of Politics and Government at George Mason University.
 Kristine Höglund, Peace Negotiations in the Shadow of Violence, International Negotiation Series, v. 6 (Leiden ; Boston: Martinus Nijhoff Publishers, 2008), 3; John Darby, The Effects of Violence on Peace Processes (Washington, DC: United States Institute of Peace Press, 2001), 38.
 Darby, The Effects of Violence on Peace Processes, 39.
 Kristine Höglund, “Violence — Catalyst or Obstacle to Conflict Resolution?” (Department of Peace and Conflict Research, Uppsala University, September 26, 2001), 12.
 Höglund, 12.
 One extreme outlier was removed for clarity of graph: the lethality against Civilians was 55 during 2018 week 4.