Pakistan, the Islamabad Memorandum and the gap between making and keeping an agreement, June–September 2026
ABSTRACT
This article examines why the Islamabad Memorandum, mediated by Pakistan between the United States and Iran in June 2026, lost its function within weeks, and how Pakistan’s position changed by September. It draws on Kydd’s (2003) model of mediator bias and credibility and Beardsley’s (2008) argument that mediated agreements are fragile because of time inconsistency. The evidence is the memorandum’s terms, the collapse chronology, Pakistan’s Security Council vote, its Saudi defence alignment and a US Treasury designation of 29 September. Washington declared the deal over on 7 July, Iran suspended its commitments on 18 July, and the 60-day period lapsed on 17 August. Pakistan had enough access to bring the parties together but lacked the instruments and party-specific credibility to hold them to the text.
1. Producing an agreement, failing to sustain it
The Islamabad Memorandum was the most concrete diplomatic framework for ending the war that began on 28 February 2026. It provided a 60-day negotiating period, lifting of the US blockade within 30 days and Iranian mine clearance in Hormuz (Shamim, 2026). President Trump declared it over on 7 July. The research question is why Pakistani mediation could produce an agreement but could not sustain it. The claim is that Pakistan could provide access and a venue but lacked instruments to secure compliance after signature and credibility specific to either party; a Saudi defence rapprochement and a US designation of a Pakistan-based firm in September deepened that limit. The scope is Pakistan’s role in June–September 2026.
2. Bias, credibility and time inconsistency
Two questions are separable: how a mediator brings parties to agreement, and why the agreement lasts or fails. Kydd (2003) treats the first as information transmission: a mediator is valuable when its messages about the other side’s resolve are believable. An unbiased mediator that wants peace above all recommends concession in every circumstance, so its messages carry no information. A mediator close to one party is believable when it tells that party the other side is resolved and it should settle, because it would send that message only when concession was necessary. Credibility is thus a relation to a particular party, not a general attribute.
Beardsley (2008) addresses the second question. Mediation raises the likelihood of a formal agreement, but such agreements are more fragile because they rest on temporary incentives and pressure from the mediator, and the motive to comply weakens once these lapse. Beardsley defines this as a time-inconsistency problem: a commitment rational at signature ceases to be rational at implementation. The two models therefore separate the capacity to produce an agreement from the capacity to sustain it.
3. Architecture and collapse of the memorandum
Al Jazeera and Brussels Signal date the signing to 17 June 2026, Yeni Şafak to 18 June (Brussels Signal, 2026; Shamim, 2026; Yeni Şafak, 2026). The design matches the structure Beardsley describes. The memorandum set 60 days for a broader peace agreement, envisaged $300 billion for reconstruction and economic development, and linked lifting the blockade within 30 days to Iranian mine clearance as reciprocal steps (Shamim, 2026). Spreading heavy obligations over time gave each side reason to wait for the other’s move, and the source describes no mechanism for monitoring breaches or guaranteeing performance.
The collapse was rapid and cannot be reduced to one date. The Ever Lovely was struck on 25 June, the United States targeted Iranian coastal facilities on 27 June, and on 7 July Trump declared the agreement over (Shamim, 2026). On 18 July Iran’s Deputy Foreign Minister Gharibabadi announced the suspension of all commitments under the memorandum (ProPakistani, 2026). The 60-day period expired on 17 August (Shamim, 2026). The three dates correspond to three kinds of collapse: a US political withdrawal, a formal Iranian suspension and the lapse of the legal period.
The sequence supports Beardsley’s model. A commitment rational for each side at signature was questioned at the first incident and abandoned before implementation began. The sources record no Pakistani intervention returning the parties to the text, which indicates that the mediator held no usable incentive or sanction after signature.
4. Pakistan’s position: access without specific credibility
Pakistan’s record demonstrates access: it brokered the 8 April ceasefire and hosted talks lasting more than 20 hours, and its mediation continues alongside Qatar’s (Yeni Şafak, 2026). On 31 March it presented with China a five-point initiative for the safe passage of civilian and commercial ships through Hormuz (Baird Maritime, 2026). With Colombia, it abstained on the 7 April Security Council draft on Hormuz navigation that China and Russia vetoed (UN News, 2026).
In Kydd’s framework this describes a mediator with no specific closeness to either party. The interpretation advanced here is that the abstention and the joint initiative with China make Pakistan a less credible messenger to Washington that Iran is resolved, since in Kydd’s model credibility depends on closeness to the recipient. No source shows comparable Pakistani leverage over Iran. The result is a channel acceptable to each party but carrying discriminating information to neither: sufficient for a ceasefire and a venue, insufficient to bring parties back after breaches.
Two September developments complicated this position. The first is Saudi defence alignment. Al Jazeera reports that Saudi Arabia, Pakistan and Türkiye signed a joint defence agreement in Mecca on 7 August; the official text has not been verified (Al Jazeera, 2026). On 18–19 September Pakistan’s military spokesman, Lieutenant General Ahmed Sharif Chaudhry, said Pakistan would defend Saudi Arabia in every respect (Al Jazeera, 2026), although according to Semafor Pakistan and Türkiye had not committed troops as of 25 September (Gonzalez, 2026). Iran’s conditions for reopening the strait include ending the war on all fronts, Lebanon included (Brussels Signal, 2026). Pakistan thus openly backs the Saudi side on a front Iran wants closed. Under Kydd’s model this may earn credibility in Riyadh while weakening Pakistan’s claim to impartiality in Tehran.
The second is the sanctions designation of 29 September. The US Treasury designated ten individuals and entities in a network procuring ballistic missiles and drones for Iran’s Ministry of Defence, including Pakistan-based Cavalier Dynamics and Cavalier companies registered in Saudi Arabia and Türkiye (U.S. Department of the Treasury, 2026). The action does not target the Pakistani government, which had issued no response that could be found, but Washington has publicly named Pakistani territory as a link in Iran’s military supply chain. By Kydd’s criterion, this does not raise Pakistan’s credibility with Washington.
5. Alternative explanations considered
An alternative explanation locates the collapse in the parties’ preferences rather than the mediator’s attributes. A senior Iranian official told Reuters that Tehran had sent its latest proposal to Washington through mediators and could open the strait to Gulf shipping within seven days if its demands were met (Slattery et al., 2026); the core demands are that the United States reduce military pressure and lift the blockade of Iranian ports (Schuler, 2026). On 22 September Trump said he expected a deal shortly after the election (Schuler, 2026). On this reading the parties, not the mediator, set the timetable.
This identifies a real factor but not the full record. Preferences are the object of mediation, and Beardsley’s model asks whether they shift after signature and whether the mediator can offset the shift. China on 16 September urged the parties to return to the Islamabad Memorandum quickly (Carroll, 2026), and Prime Minister Sharif told the UN on 24 September that Pakistan was “absolutely determined” to reduce tensions (Yeni Şafak, 2026). US envoys Witkoff and Kushner held indirect talks with an Iranian delegation in New York on 22 September (Brussels Signal, 2026). The text survives as a reference point while the negotiation moves outside Pakistan’s framework.
6. What the case shows about the theory
The case confirms Beardsley’s distinction and adds an observation on Kydd. Pakistan’s even-handedness helped at the agreement stage and hurt at the enforcement stage, because it carried no specific leverage over either side. Its September tilt towards Riyadh did not convert into leverage over Washington, the party that declared the deal over. Bias helps a mediator only when it runs towards the party that must be persuaded.
7. Outlook: Türkiye, Saudi Arabia and the World
The US midterms are expected to set the pace of US–Iran negotiation in October 2026. If Trump’s statement holds, October will select a framework rather than conclude a deal, with the Islamabad text and Iran’s seven-day proposal as rival starting points. If the Islamabad text is used, Pakistan stays at the table as the framework’s owner; if a new text is adopted, it becomes one channel among several.
Türkiye shares part of Pakistan’s dilemma. NTV reported on 15 June that President Erdoğan welcomed the US–Iran peace agreement and acknowledged Pakistan’s mediation and the diplomatic support of Qatar and Saudi Arabia (NTV, 2026); the report predates the 17–18 June signing dates given elsewhere. Türkiye is party to the Mecca framework reported by Al Jazeera, and the 29 September designation listed Cavalier companies registered in Türkiye. Should Ankara seek to mediate, the Pakistani case indicates that defence proximity to Riyadh may cost credibility in Tehran.
For Saudi Arabia, Pakistan’s defence pledge and mediating role may prove incompatible: a troop deployment would raise Pakistan’s contribution to Saudi security and lower its value in the US–Iran channel. As of 25 September no such decision had been made. Globally, the 2027 lesson for any post-election agreement concerns design: Beardsley’s finding calls for a guarantee mechanism monitoring implementation and answering breaches, and Kydd’s for a combined channel of mediators with different affinities, such as Pakistan and Qatar, each credible to a different party.
8. Conclusion
Pakistan could bring the parties to a common text but could not keep it alive. The United States declared it over on 7 July, Iran suspended it on 18 July and it lapsed on 17 August, a sequence consistent with Beardsley’s thesis that mediated agreements are fragile because of time inconsistency. Pakistan’s UN abstention and joint initiative with China kept it from holding Kydd’s party-specific credibility with Washington. In September its defence support for Riyadh weakened its impartiality in Tehran, and the US designation tied Pakistani territory to Iran’s procurement network. Pakistani mediation did not close the gap between producing and sustaining an agreement, and September widened it.
SOURCES
Al Jazeera. (2026, 19 September). Turkiye backs Saudi Arabia’s security amid escalating Houthi attacks: FM. https://www.aljazeera.com/news/2026/9/19/turkiye-backs-saudi-arabias-security-amid-escalating-houthi-attacks-fm
Baird Maritime. (2026, 31 March). China and Pakistan call for calm in the Strait of Hormuz. https://www.bairdmaritime.com/security/china-and-pakistan-call-for-calm-in-the-strait-of-hormuz
Beardsley, K. (2008). Agreement without peace? International mediation and time inconsistency problems. American Journal of Political Science, 52(4), 723–740. https://doi.org/10.1111/j.1540-5907.2008.00339.x
Brussels Signal. (2026, 23 September). Kallas presses Tehran for toll-free passage through Hormuz. https://brusselssignal.eu/2026/09/kallas-presses-tehran-for-toll-free-passage-through-hormuz/
Carroll, R. (2026, 16 September). China presses Iran, US to reopen Hormuz as disruptions threaten oil supply. Al-Monitor. https://www.al-monitor.com/originals/2026/09/china-presses-iran-us-reopen-hormuz-disruptions-threaten-oil-supply
Gonzalez, J. (2026, 25 September). Saudi Arabia intercepts Houthi missiles. Semafor. https://www.semafor.com/article/09/25/2026/saudi-arabia-intercepts-houthi-missiles
Kydd, A. (2003). Which side are you on? Bias, credibility, and mediation. American Journal of Political Science, 47(4), 597–611. https://doi.org/10.1111/1540-5907.00042
NTV. (2026, 15 June). Cumhurbaşkanı Erdoğan: Barış anlaşması memnuniyet verici [President Erdoğan: The peace agreement is welcome]. https://www.ntv.com.tr/gundem/cumhurbaskani-erdogan-baris-anlasmasi-memnuniyet-verici-1728625
ProPakistani. (2026, 18 July). Iran suspends Islamabad MoU. https://propakistani.pk/2026/07/18/iran-suspends-islamabad-mou/
Schuler, M. (2026, 22 September). Trump calls for Iran’s economic isolation at UN as Tehran offers Hormuz reopening. gCaptain. https://gcaptain.com/trump-calls-for-irans-economic-isolation-at-un-as-tehran-offers-hormuz-reopening/
Shamim, S. (2026, 17 August). US-Iran Memorandum of Understanding expires: How and why it fell apart. Al Jazeera. https://www.aljazeera.com/amp/news/2026/8/17/us-iran-memorandum-of-understanding-expires-how-and-why-it-fell-apart
Slattery, G., Hafezi, P., & Slodkowski, A. (2026, 23 September). Trump threatens to annihilate Iran as diplomats push for deal at UN. Al-Monitor. https://www.al-monitor.com/originals/2026/09/tehran-hints-hormuz-talks-us-leaders-gather-un
U.S. Department of the Treasury. (2026, 29 September). Operation Economic Outcast takes down Iranian military procurement networks. https://home.treasury.gov/news/press-releases/sb0637/
UN News. (2026, 7 April). Security Council: Russia and China veto resolution on Strait of Hormuz. GlobalSecurity.org. https://www.globalsecurity.org/military/library/news/2026/04/mil-260407-unnews01.htm
Yeni Şafak. (2026, 24 September). Pakistan ‘absolutely determined’ to advance US-Iran de-escalation. https://en.yenisafak.com/world/pakistan-determined-to-advance-us-iran-peace-talks-3723670
Furkan Mahmut Tameroglu