Has Mali's junta-led strategy of substituting Russian mercenaries and anti-French nationalism for Western counterterrorism partnerships been discredited by JNIM's advance on Bamako and the loss of Kidal, or is it too early to judge against the baseline of prior Western failures?

Round 1 Response: The Baseline Problem Undermines Advocate A's Entire Framework

The Weakest Point in Advocate A's Argument: The Missing Counterfactual

Advocate A's argument rests on a critical logical flaw that undermines every specific data point they cite: they measure the junta's performance against an implicit standard of "success" without establishing what the Western-backed baseline actually produced over a decade of deployment.

This is not a minor methodological quibble. It is the entire evidentiary frame. If the pre-junta security architecture also failed to hold territory, also failed to stop JNIM's geographic expansion, and also failed to prevent civilian displacement — then citing those same outcomes under the junta does not demonstrate that the junta's strategy has failed; it demonstrates that Mali faces an extraordinarily difficult insurgency that has defeated multiple strategic approaches. Those are empirically different claims.

Let me be specific about what the baseline actually shows:

Operation Barkhane ran from 2014 to 2022. During that period, ACLED data shows JNIM attack frequency also increased year-over-year, consistently, from 2017 through 2021. The geographic expansion of JNIM toward southern Mali and toward Bamako's supply corridors — which Advocate A presents as evidence of junta failure — was already documented and underway before the junta expelled French forces. The UN Panel of Experts on Mali's 2021 report (pre-expulsion) explicitly flagged JNIM's expanding southern reach as an accelerating trend. Advocate A is attributing to the junta a deterioration that predates the junta's security choices.

Kidal specifically: The CMA held Kidal against the Malian state from 2012 onward — through the entire Barkhane period, through MINUSMA's decade-long presence, through multiple negotiated frameworks including the 2015 Algiers Accord. French and UN forces did not retake Kidal. FAMa did enter Kidal in November 2023, which — whatever its subsequent fragility — represents something Western-backed forces demonstrably did not accomplish in eleven years. Advocate A acknowledges FAMa entered the city but frames contested control as failure. That standard, applied consistently, would also condemn every Western-backed operation in the region.

---

My Strongest Evidence Points

Evidence Point 1: The 2015 Algiers Accord Was the Western-Backed Framework — and It Institutionalized Separatist Control of Kidal

The peace architecture that Western partners constructed, and that France and MINUSMA underwrote, was the Algiers Accord. That agreement effectively froze CMA control over northern territories, including Kidal, as a precondition for negotiations. The Institute for Security Studies documented in multiple reports (2016-2020) that this framework gave the CMA — a coalition that includes groups with documented ties to arms trafficking — a de facto governance veto in the north while providing no mechanism for restoring Malian state authority.

This is the "durable outcome" that Western-directed counterterrorism produced after a decade: a negotiated partition that the Malian state never accepted as legitimate, that a majority of Malian political opinion (per Afrobarometer 2021 survey data showing 67% support for the coups in urban Mali) rejected, and that demonstrably failed to contain JNIM's expansion anyway. The junta's attempt to reassert sovereignty over Kidal — however imperfectly executed — is a response to a failed Western-designed framework, not a departure from a successful one.

Evidence Point 2: ACLED Data Requires Contextual Decomposition, Not Headline Citation

Advocate A cites ACLED attack frequency as a straightforward indicator of junta failure. The picture is more complicated. ACLED's own analysts (notably in their January 2024 briefing on the Sahel) distinguish between attack frequency and territorial control. A counterinsurgency force that is actively contesting territory will typically produce higher short-term ACLED event counts than a force that has ceded rural areas and retreated to garrison duty — because more engagements occur when forces are in the field.

Barkhane's final years showed declining ACLED engagement numbers in some zones — not because JNIM was suppressed, but because French forces pulled back from active rural patrolling ahead of withdrawal. A spike in recorded conflict events under FAMa/Africa Corps operations is at least partially consistent with more offensive activity, not less control. I am not asserting this is the correct interpretation — the data is genuinely mixed and ACLED itself flags methodological limitations in conflict-zone reporting. But Advocate A presents contested, multivalent data as unambiguous confirmation of a narrative. That is not empirically rigorous.

Evidence Point 3: The Sovereignty and Partnership Diversification Argument Operates on a Longer Timeline Than 24 Months

Advocate A's framing treats this as a completed empirical test. It is not. The junta expelled Barkhane in 2022, MINUSMA in 2023, and began Africa Corps integration while simultaneously pursuing diplomatic reorientation toward the Alliance of Sahel States (with Burkina Faso and Niger). These are structural changes to Mali's security architecture — the kind of changes that the RAND nation-building literature Advocate A cites acknowledges require years, not months, to assess.

For comparison: the U.S. "surge" in Afghanistan (2009-2011) showed measurable deterioration in security metrics in its first 18 months before any stabilization gains became visible — and even those gains proved temporary. The FM 3-24 framework that Advocate A invokes explicitly cautions against short-cycle assessment of counterinsurgency outcomes. Applying a 24-month window to judge a multi-year structural transition is inconsistent with the very doctrinal literature being cited.

---

Summary

Advocate A has presented real data about a deteriorating security situation. I do not dispute the humanitarian severity. What I dispute is the causal attribution: the claim that these outcomes constitute evidence that the junta's strategy has failed, rather than evidence that Mali's insurgency has defeated every strategy tried against it so far — including the one Advocate A implicitly endorses. The baseline problem is not a technicality. It is the difference between an honest empirical assessment and a politically convenient one.

Round 2: Identifying A's Concessions and Their Implications

Where Advocate A Conceded Ground — And What That Implies

Advocate A made two explicit concessions in Round 2 that deserve precise examination before being minimized.

First concession: A acknowledges that "JNIM's southward geographic expansion was documented and underway before the junta expelled French forces," and accepts that attributing post-2022 JNIM expansion exclusively to junta choices "overstates the causal claim."

Second concession: A accepts that ACLED event frequency data "requires contextual decomposition" and that higher event counts can partially reflect more offensive activity rather than deterioration.

These are not marginal admissions. Together, they hollow out the two central empirical pillars of A's Round 1 argument — the JNIM expansion data and the ACLED frequency data. What A presented in Round 1 as straightforward empirical evidence of junta failure is now acknowledged to be causally ambiguous and methodologically contested.

The logical implication A does not address: if A cannot reliably attribute JNIM's expansion to the junta, and cannot straightforwardly read ACLED data as failure, then the "empirical test" the headline asserts is not complete. A's own concessions undermine the premise of the debate — that early results are "not ambiguous." They are, by A's own admission, considerably more ambiguous than Round 1 claimed.

---

Where A's Round 2 Arguments Fall Short

A's Unfalsifiability Accusation Misrepresents My Position

A argues that my baseline logic makes the junta's strategy "permanently immune to empirical evaluation." This is a rhetorically effective framing, but it mischaracterizes what I actually argued.

I did not claim that no evidence could ever evaluate the junta's strategy. I claimed that 24 months is an insufficient evaluation window for a structural security transition, and that the correct comparison baseline is the Western-backed framework's performance over its actual operational timeline. These are specific, bounded methodological claims — not a blanket unfalsifiability defense.

A specific falsifiable test I would accept: if, by 2027, JNIM attack frequency in central and southern Mali is higher than its 2022 trajectory, and if FAMa demonstrates no increased independent operational capacity, and if civilian displacement continues accelerating beyond pre-junta trend lines — that would constitute meaningful evidence against the junta's thesis. I am willing to define that test. What I reject is A's assertion that we already have the answer after 24 months of a structural transition that A's own cited literature — FM 3-24 — acknowledges requires years to assess.

A's Reading of the Junta's Promise Is Overstated

A argues that Choguel Maïga's September 2021 UN address constitutes a falsifiable promise to perform better than France, and that this is the correct evaluative standard. This requires examination.

Political speeches at the UN General Assembly are not operationalized strategy documents with measurable benchmarks. Maïga's address — which I have reviewed — argued that Mali was being "abandoned" by France mid-operation as a justification for new partnerships. It does not specify a timeline, a metric, or a threshold for success. Treating a diplomatic speech as a precise empirical commitment is selective interpretation. The junta's actual security doctrine, as articulated through the Alliance of Sahel States framework, explicitly frames the project as multi-year sovereign consolidation. Judging it against a speech is not the same as judging it against its own operational terms.

A's Kidal Analysis Contains an Unacknowledged Contradiction

A argues that FAMa entering Kidal kinetically produced "a city they could enter but not govern, hold, or administer" — and that this is worse than the "frozen-but-stable CMA arrangement that previously existed."

This requires A to implicitly defend the Algiers Accord framework as producing stability. But A does not defend it — and cannot, because A's own argument in this round calls it a "different strategic doctrine" without evaluating its outcomes. Let me be specific about what the Algiers Accord's "stable freeze" actually produced:

The 2015 Accord left CMA in de facto control of Kidal while JNIM expanded throughout the region. The ISS documented in its 2019 and 2020 reports that the Accord's northern zones became primary transit corridors for arms trafficking and JNIM recruitment. The "stable" frozen arrangement was simultaneously a permissive environment for the insurgency A claims the West was suppressing. A cannot simultaneously argue that the Algiers framework produced stability and that JNIM's expansion southward predates the junta — those trends coexisted under Western-backed governance. The frozen north was not separate from JNIM's growth; it was adjacent to and arguably enabling of it.

A's "What Is Being Built?" Argument Ignores the Alliance of Sahel States

A asks what the junta is building, then cites the absence of a training pipeline and civil administration in Kidal as evidence that nothing is being constructed. This is too narrow a frame.

The Alliance of Sahel States — Mali, Burkina Faso, Niger — represents a documented institutional development: a mutual defense pact signed in September 2023, a joint military force announced with a combined budget commitment, and a formal withdrawal from ECOWAS in January 2025. These are structural developments in regional security architecture. Whether they will produce superior outcomes is genuinely uncertain — I acknowledge that explicitly. But A's claim that "there is no construction occurring" is demonstrably false at the regional institutional level, even if civil administration in Kidal remains absent.

---

My Remaining Evidence A Has Not Adequately Addressed

On the displacement figures: A argues that new displacement waves in Mopti and Ségou specifically postdate 2022-2023 operations, implying junta-era causation. But Mopti has been a primary conflict zone since 2015 — the UN's own humanitarian reports from 2019 and 2020 document severe displacement in Mopti predating the junta. A is selecting a recent reference point for a region with a continuous displacement crisis. The trend is not new to the junta; the question is whether it is accelerating beyond the pre-existing trajectory, and A has not provided that differential analysis.

On the resource extraction point: A correctly notes I did not engage the mining concession documentation in Round 1. I engage it now. The UN Panel of Experts documentation of Africa Corps personnel near gold concessions is real and warrants scrutiny. However, the same Panel's reports on Mali from 2019 and 2020 documented French-linked companies operating in Malian resource sectors during Barkhane's deployment. Economic entanglement with foreign security partners is not unique to the Russia relationship — it characterized the French relationship as well. The relevant question is whether Africa Corps' extractive orientation uniquely precludes capacity transfer in a way that Barkhane's presence did not. A asserts this but provides no comparative evidence that Barkhane produced superior FAMa capability development. The ISS's own 2021 assessment of FAMa — published before junta expulsion of French forces — described Malian military capacity as severely degraded despite years of French training investment.

---

Summary

A's concessions in Round 2 are more significant than A acknowledges. Accepting that JNIM's expansion predates the junta, and that ACLED data is causally ambiguous, removes the empirical foundation for the headline's claim that we already have decisive early results. What A retains is a reasonable hypothesis — that Wagner's extractive model is structurally incapable of building governance — but this remains a hypothesis, not a demonstrated outcome, particularly given that the Western-backed model also demonstrably failed to build durable governance capacity in FAMa or the Malian state after a decade of trying.

The honest assessment, consistent with A's own concessions, is that Mali's insurgency has defeated multiple strategic frameworks, that the junta's experiment is genuinely ongoing and unresolved, and that the declared empirical verdict is premature. That is not a defense of the junta's choices. It is a demand for the same evidentiary standard A invoked — and then partially abandoned — across two rounds.

Round 3: Steel-Manning A, Why My Position Remains Stronger, and What Would Change My Mind

Steel-Manning Advocate A

The strongest version of A's argument is this: the junta did not merely inherit a difficult situation — they made a specific, public, falsifiable claim that their approach would outperform the Western-backed framework they displaced. Given that JNIM has continued expanding, Kidal produced a contested ruin rather than administered territory, and Africa Corps shows documented structural orientation toward resource extraction rather than capability transfer, the junta has now had sufficient operational time to demonstrate at minimum directional progress toward their promised reversal. The absence of any measurable governance footprint in recovered territory, combined with worsening displacement in central Mali's previously secondary conflict zones, constitutes meaningful evidence against the specific thesis the junta advanced — not merely evidence that Mali's insurgency is hard.

That is a serious, well-constructed argument. I have engaged it seriously across two rounds, and I want to be precise about what it gets right before explaining why it still falls short.

---

Why My Position Remains Stronger Despite A's Best Argument

A's steel-manned position is at its strongest on one specific point: the junta made a directional claim, not merely a "we will try something different" claim, and directional claims are falsifiable on shorter timelines than structural transformation claims. I acknowledge this creates genuine tension for my "insufficient timeline" argument.

However, three specific problems prevent A's argument from closing the case.

First, A's unfalsified prediction argument conflates trajectory ownership with causal responsibility. A argues that the junta promised to reverse a worsening trajectory, the trajectory continued worsening, therefore the prediction is falsified. But this inference requires that the junta's chosen instruments — Africa Corps, AES coordination, expelled MINUSMA — had adequate time and operational conditions to produce a measurable reversal signal. Consider the specific sequence: Barkhane expelled August 2022, MINUSMA withdrawal completed December 2023, AES mutual defense framework signed September 2023. The junta was simultaneously managing three major security architecture transitions while conducting offensive operations. Demanding a measurable reversal signal within 12-18 months of completing that transition is not the same evidentiary standard A applied to Barkhane, which operated continuously for eight years before A's own cited literature flagged its failures. I am not claiming the junta gets unlimited time. I am claiming A has not applied a consistent timeline standard across the two frameworks being compared.

Second, A's Kidal-as-ungoverned-ruin argument, while factually grounded, omits a critical counterfactual mechanism. A argues that the Algiers framework produced an "imperfect frozen governance arrangement" while FAMa's kinetic operation produced an "ungoverned ruin" — and that on A's own terms about ungoverned space enabling JNIM, the junta's outcome is worse. This is A's most sophisticated argument in Round 3, and I want to engage it precisely. The frozen CMA arrangement did not merely coexist with JNIM expansion — as I documented in Round 2 using ISS reporting, the northern corridors under CMA de facto control became primary JNIM recruitment and arms transit routes. The "stable freeze" was not neutral ungoverned space; it was actively administered by actors with documented trafficking ties that benefited JNIM's logistics. FAMa's disruption of that arrangement, even if it has not yet produced administered governance, disrupted a specific logistical architecture that was enabling JNIM. Whether that disruption net-positive or net-negative for JNIM's operational capacity over the medium term remains genuinely unresolved — A presents one causal pathway (ungoverned ruins enable JNIM) while ignoring the competing pathway (CMA-administered corridors enabled JNIM more systematically). The evidence does not clearly resolve which effect dominates.

Third, A's Round 3 falsifiability criteria are actually more consistent with my position than A's, because they implicitly acknowledge we do not yet have the answer. A specifies four evidentiary tests — FAMa holding contiguous territory by 2026, UN Panel documenting capability increase by 2027, JNIM frequency declining below 2022 baseline by 2026-2027, and comparative resource entanglement analysis. Every one of these tests is set in the future. A is, in Round 3, conceding that the current evidence base does not meet the evidentiary standards A's own framework requires for a decisive verdict. The headline under debate asserts "early results suggest the security vacuum is widening, not closing" — and A's own future-dated falsifiability criteria implicitly acknowledge that the evidence for that claim is insufficient to close the debate. That is the position I have been arguing since Round 1.

---

What Specific Evidence Would Convince Me That A's Position Is Correct

I will be as concrete as A was, because this question deserves a genuine answer.

1. FAMa-independent operational failure, documented by 2026. If the UN Panel of Experts or a credible equivalent (SIPRI, ISS) documents by 2026 that FAMa cannot conduct battalion-level offensive operations without Africa Corps co-deployment — specifically, that attempted independent operations systematically fail to initiate or collapse within 48 hours — that would constitute strong evidence that the dependency transfer critique is correct and that no capability accumulation is occurring. This is falsifiable in both directions: it is the specific operational test of whether something is being built.

2. AES joint force operational nullity by 2026. The Alliance of Sahel States announced a joint unified force in 2024. If, by end of 2026, that force has conducted zero documented joint operations — and independent security analysts confirm it exists only on paper — that would substantially weaken my "structural institutions being built" argument and support A's "nothing is being constructed" claim.

3. JNIM territorial administration expanding into new urban centers in central Mali by 2026. Not attack frequency — A and I both agree that metric is ambiguous. If JNIM establishes documented administrative presence (taxation, dispute resolution, governance functions) in towns within 200km of Bamako that were outside its administrative zone in 2022, that would represent a qualitative capability expansion that goes beyond trend continuation into genuine threshold-crossing. The ACLED "non-violent" events dataset and reporting from organizations like Crisis Group tracking JNIM governance activity would be the relevant sources.

4. Comparative resource entanglement evidence favoring A's asymmetric dependency claim. If a credible analysis — by ISS, SIPRI, or a peer-reviewed source — documents that Africa Corps' gold concession revenue extraction is quantitatively larger as a share of Mali's extractive economy than French security-linked economic entanglement during Barkhane, and that this differential correlates with lower training investment per FAMa personnel, that would support A's structural dependency argument rather than my "symmetric entanglement" counterpoint.

What would not change my mind: continued citation of displacement figures and JNIM attack frequencies without the differential trend analysis that distinguishes junta-era acceleration from the pre-existing trajectory — because A conceded in Round 2 that the causal attribution problem is real. Showing that bad things continue to happen in Mali, absent evidence that they are happening worse than the pre-junta trend after adjusting for the transition period's structural disruption, does not meet the evidentiary standard A's own methodology requires.

The honest verdict at Round 3 is the one A's own falsifiability criteria inadvertently confirm: the empirical test is ongoing, the baseline problem is real, and declaring a decisive result in either direction at this stage is premature. That is not a defense of the junta. It is a demand for the intellectual honesty the debate's own framing claimed to apply.