The assessment in the first part of this analysis leads to the assumption that the Dhaka attack was a manifestation of a new, hybridised terror dynamic in Bangladesh that involves at least some degree of international involvement, in this case the Islamic State (IS). This part of the analysis provides a framework analysis of two possible scenarios or types of IS diffusion in Bangladesh. 

Both models are based on a new prototype of Islamist terrorism that is a cross-fertilisation of local and global jihadist networks, working together in a mutually beneficial arrangement.

Type I: Returnee Networks 

Under the first type, IS-trained returnees run autonomous sleeper cells in Bangladesh, which propagandise, radicalise, and recruit under direct or indirect supervision from IS’ central command in Syria/Iraq. This model assumes the relative ease for returnees in slipping back into their homeland without raising suspicion – a belief reasserted by several close IS observers. The native cells procure arms from local and regional cross-border channels, in connivance with like-minded subcontinental groups. For recruitment and tactical support, they rely on existing domestic outfits like Jama’at ul-Mujahideen Bangladesh (JMB) and al Qaeda affiliated Ansar ul-Islam, which have ready-to-use pools of cadres and functional networks for logistical support. 

In this model, JMB takes organisational lead, given its highest strike capacity and organisational strength amongst all, in addition to its pledged allegiance to the IS. The cells regularly receive inputs, supervision, advice, and funds from the IS in Syria/Iraq and/or its extended networks in the region. This is a self-perpetuating model, wherein returnees routinely maintain a steady stream of local emigrants to Syria/Iraq. 

The IS cells execute the attacks in a customised style, tailor-made for the local demographic profile while maintaining a bare minimum of tactical sophistication. The usage of assault rifles, IEDs, and suicide-style raid tactics, while taking care not to kill Muslims, could be attributed to this model. The Kishoreganj Eid attack was largely un-customised, but remained unclaimed (a signature IS tactic for Muslim societies). 

This type is closer to al-Hanif’s narrative presented in Dabiq (the IS propoganda magazine), and implies that the IS orchestrated the Dhaka and Kishoreganj attacks. 

Type II: Regrouped/Affiliated Networks

The second type comprises of a regrouped set of purely domestic groups like JMB, Ansar, and Hizb ul-Tahrir, pledging allegiance to the caliphate to restore their own credibility. All three groups have shown at least some degree of localised organisational influence in the last two years. This prototype is characterised by a franchise terror module, wherein various domestic groups reorient themselves into a single operational network that draws ideological legitimacy from the IS. Given JMB’s reported allegiance to the IS, this regrouping is initiated and led by JMB, although other entities have fairly equal tactical and ideological control. This is notwithstanding the fact that the domestic groups have critical ideological and tactical differences within themselves. 

This means that the IS command structures – central or extended – have no direct control over the operational aspects of terror attacks in Bangladesh, although lateral assistance is highly probable. IS-trained returnees from Syria/Iraq are automatically absorbed into these networks, bringing conventional IS tactics to the local domain. Furthermore, some advisory collaboration with the IS’ regional cells to choose the right offensive tactics, locating the funding sources, and arranging mobilisation networks cannot be ruled out. Such an arrangement of affiliation serves to ensure the globalisation of local jihad, while providing the larger group a springboard for expansion and international media attention. 

This type is closer to the government’s narrative, and denotes that the IS inspired the Dhaka attack.

This prototype explains the fusion of local and transnational elements in the Dhaka and Kishoreganj attacks, including the mixof sophisticated and crude weapons. It partially explains the target filtering method in Gulshan, which could be due to the prevalence of Ansar’s al Qaeda-based tactics. Most crucially, it could explain Ansar’s warning tweet before the attack. 

Limits of these Models

There are certain critical indicators in the Gulshan attack that neither of the models explains convincingly. 

First, it is hard to conclude why the attackers chose to selectively execute only non-Muslims. It could fit into both types. Within Type I, it could be attributed to IS’ newly-emerging customised offensive design, in which it slightly modifies its terror tactics according to the local demographic context. Within Type II, it could be attributed to the operational and tactical control exercised by al Qaeda affiliate Ansar. In either case, local support is the deciding factor.

Second, the above models are insufficient in explaining the elite recruit pool. While it is closer to Type I (secondary involvement of JMB), it could easily fit into Type II, wherein a radicalised wolf-pack joins local cells. Early this year, Bangladeshi intelligence agencies warned of the formation of JMB sleeper cells with “highly educated members and technology experts” in the north of the country.

Hence, although at this point it cannot be affirmatively concluded if IS networks are operational in Bangladesh, this analysis evinces that a lateral, if not direct, engagement of the group in Bangladesh is a strong possibility. The indications are telling enough for the government to draw out a more holistic counter-terrorist design that does not blatantly overlook international influence and/or active engagement.