Intelligence and U.S. Foreign Policy: Iraq, 9/11, and Misguided Reform (2011) by Paul Pillar challenges the foundational assumption of Sherman Kent’s framework: that policymakers need and use intelligence assessments to make decisions. Pillar — a former National Intelligence Officer for the Near East and South Asia — argues from decades of experience that the relationship between intelligence and policy is more limited, more contentious, and more structurally constrained than the discipline’s normative model acknowledges.
Core argument
Pillar’s central thesis is that intelligence assessments rarely drive policy decisions. Policymakers make decisions based on political considerations, ideological commitments, bureaucratic pressures, and strategic visions that the intelligence assessment may not address and cannot override. The intelligence community’s most careful estimate may be structurally irrelevant to the decision it is meant to inform — not because the policymaker rejects the assessment but because the decision is being made on criteria the assessment does not engage.
The Iraq WMD case illustrates the argument: the conventional narrative is that bad intelligence led to a bad policy decision (the invasion). Pillar argues the causal arrow ran the other direction — the policy decision was made first, and intelligence was selected, emphasized, and interpreted to support it. This is politicization in its structural form: not crude pressure on analysts to change their conclusions but the selective use of intelligence products to justify decisions already made.
The intelligence-policy disconnect
Pillar identifies several configurations of the intelligence-policy relationship, extending beyond the intelligence-policy disconnect the discipline traditionally recognizes:
- Assessment drives policy (Kent’s ideal) — the policymaker reads the assessment and adjusts the decision accordingly. Pillar argues this is the rarest configuration.
- Policy drives assessment (politicization) — policy preferences shape what the intelligence community produces or how its products are used.
- Assessment and policy are independent — the policymaker makes the decision on criteria the assessment does not address; the assessment is structurally irrelevant.
- Assessment is ritual — the intelligence community produces assessments because it is required to, and the policymaker receives them because it is expected to, but neither side treats the exchange as consequential.
Relevance to the 2026 case
The 2026 Iran war presents Pillar’s thesis in its strongest form: the intelligence community assessed that Iran was not building a nuclear weapon; the policy decision to strike was made regardless, on different criteria (capability rather than intent). This is not the canonical intelligence failure (wrong assessment leads to bad policy) but a configuration Pillar’s framework predicts: the assessment was correct, the policy decision was independent of it, and the intelligence community’s estimative function was structurally irrelevant to the outcome it was designed to influence.
Related concepts
- Intelligence-policy disconnect — the structural problem he analyzes
- Analyst-policymaker relationship — the relationship he argues is more constrained than Kent assumed
- Politicization — the corruption he documents in its structural (not crude) form