There are many valid criticisms of the Obama administration’s drone policy and related targeting rules. We have made some of those ourselves (here and here). There are also invalid grounds for criticism. A recent news story in the Washington Times is a case study of the latter.
The headline for the story reads, “Obama War Chiefs Widen Drone Kill Box,” but the content of the story proves no such thing. The Washington Times essentially relied on a single source, a report by the nongovernmental organization, Public Intelligence.
The thesis of the Public Intelligence report (which the newspaper fully embraces) is that the Pentagon has loosened targeting rules for drones by making subtle changes to the language of military guidelines. At issue is a comparison of the 2009 and 2012 versions of the Joint Chiefs of Staff Instruction entitled, “No-Strike and the Collateral Damage Estimation Methodology.”
The changes that the newspaper and NGO identify as evidence of more expansive targeting rules instead demonstrate these organizations’ own misunderstandings of the laws of war and international law more generally. Properly understood, some of the identified changes to the text actually make no meaningful difference for law or policy. Some of the other changes tighten – rather than loosen — the standards set forth in the 2009 document – and thereby narrow the drone kill box. And some of the textual passages that the organizations suggest were completely deleted from the 2012 Instruction, were simply moved to another — more prominent (!) — location in the document.
Public Intelligence produced a chart with three columns identifying the changes and their purported significance. We reproduce the chart below, but add a fourth column explaining the problems with the analysis.
There are important lessons in this exchange for the Great Debate about Transparency in US post-9/11 policy. First, some of the mistakes and conspiratorial assumptions reflected in the Washington Times story and NGO report are, in part, a product of the administration’s lack of transparency. And, indeed, the administration will pay a price—in the form of inflated and mistaken concerns–by not sufficiently informing the public of its legal justifications for its actions. Second (and on the other hand), some of the mistakes that we identify below indicate a highly unfair or improper reading of the administration’s positions after they become public. Such a reception might make any administration more reluctant to share even more information.
Let’s now turn to the evidence presented in the chart:
February 13, 2009
October 12, 2012
|Changes [asserted by Public Intelligence]||Goodman and Jinks’ Analysis|
| The source and method for defining a person, place, or thing as a No-Strike entity is derived primarily from the body of international law collectively known as LOW. The LOW incorporates international treaties and agreements adhered to by the U.S. government, as well as customary international law, into a comprehensive set of guidance and requirements governing the conduct of modern warfare.||The basis for defining a person, place, or thing as a No-Strike entity is derived primarily from the body of international law collectively known as LOW.LOW comprises the international law related to the conduct of hostilities binding on the United States or its individual citizens, including treaties to which the United States is a party and applicable customary international law that regulates the conduct of armed hostilities.||Law of war is now the “basis” for defining no-strike entities rather than the source. The law of war is defined as international law “binding” rather than “adhered to” by the U.S. This includes treaties and “customary international law” that “regulates” rather than “requires” certain actions when conducting military activity.||All the language from the 2009 document that was ostensibly deleted or modified is, in fact, included (near verbatim) in the 2012 document. The more recent directive simply moves this language text from Enclosure C (providing a summary of the no-strike policy) to the introduction of Enclosure B (providing a more detailed categorization of no-strike entities). The changes in the language in the revised Enclosure C are therefore simply cosmetic. Reference to the law of war (LOW) as the “basis” for defining no-strike entities reflects no meaningful change from the “source and method” formulation. The LOW provides the legal “basis” for these targeting principles by providing their legal “source” and the “method” of applying them in specific contexts. It is unclear what the difference is, if any, between rules “adhered to” by the US and those that are “binding.” The formulations appear largely synonymous. If anything, the former could be narrower than the latter. That is, the US could decide not to adhere to (comply with) some of its international obligations. The change to “regulates” from “requirements” is also cosmetic. Both formulations reference that part of customary international law that governs the conduct of hostilities. These rules “regulate” activities in this context by “requiring” certain actions and prohibiting others. Regulation necessarily implies requirements. And it is boilerplate to refer to the LOW as those rules regulating the conduct of hostilities. Moreover, the 2009 version also describes custom as “guidance”–eliminating that term thus strengthens the statement.|
| No-Strike entities are those designated by the appropriate authority upon which kinetic or non-kinetic operations are prohibited to avoid violating international law, conventions, or agreements, or damaging relations with coalition partners and indigenous populations. The infliction of unnecessary suffering or damage to civilian persons or property that is excessive in relation to the concrete and direct military advantage anticipated is inconsistent with international law and is contrary to DOD policy outlined in this document and in references a and b.||No-Strike entities are those designated by the appropriate authority upon which kinetic or non-kinetic operations are prohibited to avoid violating international law, or agreements, or damaging relations with coalition partners and indigenous populations.||The new version removes “conventions” from a list of things to avoid violating. An entire sentence prohibiting the infliction of “unnecessary suffering or damage to civilian persons or property” is removed, as well as the statement that this is “inconsistent with international law” and “contrary to DOD policy.”||The purportedly deleted language is, once again, included (near verbatim) in the 2012 document. The more recent directive again simply moves the language text from Enclosure C (providing a summary of the no-strike policy) to the introduction of Enclosure B (providing a more detailed categorization of no-strike entities).The deletion of “conventions” is meaningless given that the term “international law” includes both treaties and applicable customary international law. And deleting the reference to “unnecessary suffering or damage” has no significance since this language simply describes, at a high level of generality, precisely what international law requires—as the detailed categorization of no-strike entities in Enclosure B makes clear.|
| The LOW requires all military personnel to take reasonable precautions to ensure that only military objectives are targeted and to ensure that civilian or noncombatant objects are not made the object of attack.||The LOW requires all military personnel to take reasonable precautions to ensure that only military objectives are targeted and to avoid targeting (i.e., attacking) civilian or noncombatant persons or objects.||A requirement to “ensure” that civilians are not the subject of attacks is changed to an admonishment to “avoid targeting” civilians.||The change is inconsequential. The 2009 document, using slightly clunky language that tracks the LOW prohibition, directs commanders to “ensure” that civilians are not made “the object of an attack”—that is, to “avoid” intentionally targeting civilians. The revised language is essentially nothing more than a restatement of the LOW prohibition on intentional targeting of civilians. Other parts of the 2012 document (pp. 35 & 71) also show that the word “ensure” is used interchangeably with “avoid” and that the prohibition is strict (e.g., p. 35 “The LOW requires reasonable precautions to ensure only legitimate military objects are targeted. The LOW also requires combatants to refrain from intentionally targeting civilian or noncombatant personnel or facilities.”).|
| However, in these circumstances the commander must weigh the anticipated loss of life, damage to property, or other negative effects incidental to the attack versus the military advantage expected to be gained by the attack. In making the decision, commanders must consider the military necessity for attacking the target, proportionality of the means planned for target engagement, and reasonableness within the framework of operational objectives.||However, in these circumstances the commander must weigh the anticipated loss of life, damage to property, or other negative effects incidental to the attack versus the military advantage expected to be gained by the attack. The anticipated injury or loss of civilian or noncombatant life, damage to civilian or noncombatant property, or any combination thereof, incidental to attacks must not be excessive in relation to the concrete and direct military advantage expected to be gained.||When making the decision to attack a location where collateral damage is likely, the commander is no longer told to consider “proportionality” and “reasonableness within the framework of objectives”, but is instead required to ensure that loss of civilian or noncombatant life or property is “not excessive” in relation to the “concrete and direct military advantage expected to be gained” by the attack.||The revised language reflects no change in policy–and the claim that it involves a substantial loosening of restrictions reflects a troubling lack of knowledge of LOW. The 2012 language drops the imprecise description of the LOW “proportionality” test in favor of the very language used in the First Additional Protocol to the Geneva Conventions to operationalize the principle. In effect, the 2012 documents drops the bare references to “proportionality” and “reasonableness” in favor of a more concrete explanation of what these terms mean.|
| By limiting unnecessary suffering and disproportionate damage, the No-Strike process will accelerate recovery in post-conflict operations and minimize operational limitations routinely imposed as a result of international sensitivities over the humanitarian impacts of military operations.||By mitigating human suffering and property damage, the No-Strike process will accelerate recovery in post-conflict operations and minimize operational limitations routinely imposed as a result of international sensitivities over the humanitarian impacts of military operations.||A commander’s “limiting” of “unnecessary” suffering is now only a requirement to “mitigate human suffering” and property damage.||More needs to be said to show a substantive difference, let alone a loosening of standards, in comparing an obligation to “limit unnecessary” suffering or damage and an obligation to “mitigate” suffering or damage. Indeed, the latter could be more restrictive, because it may not require a showing that the suffering or damage was militarily unnecessary. The better view is that the terms here are interchangeable. Indeed, the earlier (2009) document includes several references to “mitigate” and “mitigation” as the operative term. Also, recall also that Enclosure B of the 2012 directive states flatly that the infliction of unnecessary suffering or damage is prohibited.|
| Failure to observe these obligations could result in disproportionate negative effects on civilians and noncombatants and be considered a LOW violation. Furthermore, U.S. leadership and military could be subject to global criticism, which could adversely impact military objectives, alliances, partnerships, or national goals.||Failure to observe these obligations would be considered a LOW violation. Furthermore, the United States could be subject to global criticism, which could adversely impact military objectives, alliances, partnerships, or national goals.||A statement that collateral damage could “result in disproportionate negative effects on civilians and noncombatants” is now no longer included. Global criticism in such a circumstance would also be leveled at the U.S. generally rather than the “U.S. leadership and military.”||These changes are inconsequential, if not more restrictive. The statement that failure to observe targeting obligations “would” (rather than “could”) be considered a LOW violation is more precise–and arguably more restrictive. Deleting the statement that collateral damage could “result in disproportionate negative effects on civilians and noncombatants” is of no moment since this claim is beyond obvious. The statement that the “United States” could face global criticism for LOW violations by its forces is a simpler and more accurate way to express the idea that compliance with targeting rules has broad strategic implications.|
| The CDM is a balance of science and art that produces the best judgment of potential damage to collateral concerns.||The CDM is a balance of science and art that produces a conservative characterization of the risk of collateral damage for commanders and decision makers.||The collateral damage estimation (CDE) methodology (CDM) has changed from a “best judgment of potential damage” to a “conservative characterization of the risk of collateral damage”.||This change appears, at first blush, an important shift in policy on the importance and reliability of the CDM in no-strike decisions. Review of the full text of Enclosure D (concerning CDM) in both documents makes clear that this alteration in language does not reflect a change in policy. Both documents state flatly: that the CDM produces only an estimate; that this estimate is unavoidably inexact; that the CDM suffers from several specific limitations and exclusions that tend to underestimate collateral damage; and that this estimate is only one factor among many that commanders should consider in targeting decisions.Moreover, if the 2012 language effectuates any change, it would presumably narrow the kill box by explicitly reminding commanders that the CMD produces a lower bound (a “conservative” estimate) in evaluating the risk of civilian casualties.|
| However, the science is inherently limited by the quantity and reliability of collected and analyzed weapons effects data and target information.||However, the science is inherently limited by the quantity and reliability of collected and analyzed weapons effects data, weapon delivery uncertainties, and target information.||The new version inserts “weapon delivery uncertainties” into a list of factors affecting collateral damage estimation.||This change is inconsequential. Including the reference to “weapon delivery uncertainties” in this passage more fully accounts for the range of CDM’s limitations explicitly referenced elsewhere in Enclosure D of both documents. (See, e.g., D-4 of the 2009 directive: “The CDM does not account for weapon malfunctions, operational delivery errors, or altered delivery tactics based on operator judgment. The CDM assumes weapons will function as designed and will be delivered to achieve the desired effects.”).|
| All of these sources contain some degree of inherent error and uncertainty.||All of these sources contain some degree of inherent variability.||Sources of information that feed into the collateral damage estimation process are no longer subject to some degree of “error and uncertainty”, but are instead subject to “variability.”||The terms used seem largely synonymous. Additionally, the full directives detail, in virtually identical language, several specific ways in which the sources produce inexact estimates.
| CDM is merely an estimate to assist a commander in the decision making process relying on informed data and sound judgment.||Ultimately CDE is an estimative process to help inform a commander’s decision making.||An entire sentence is changed to reflect that attempting to predict collateral damage is an “estimative process” rather than an “estimate” to “help inform the commander’s decision making.” A description of this process as “relying on informed data and sound judgment” is also removed.||It is not clear what the difference is between these formulations. And, it is plain that both directives consider the CDM a useful, even if flawed tool, to assist the commander in the “estimative process.” We would agree that deleting a reminder that the CDM is derived from “informed data and sound judgment” could somewhat soften the strength of this statement. However, those terms could be read to modify “a commander in the decision making process.” Removal of the terms may have thus been the best approach to simplifying this language.|
|||The CDM does not account for secondary explosions. Collateral damage due to secondary explosions (i.e., weapons cache or fuel tanks for military equipment) cannot be consistently measured or predicted. Commanders should remain cognizant of any additional risk due to secondary explosions.||A passage is inserted indicating the collateral damage estimation methodology (CDM) does not apply to secondary explosions from gas and fuel tanks on the ground.||The 2012 directive does, indeed, clarify that the CDM does not account for secondary explosions. It is not clear that this reflects any change in policy. Instead, it presumably just further specifies the limitations and exclusions of the CDM in generating precise estimates. Moreover, it could be read as an additional reminder to commanders that they must still take these factors into account because the CDM fails to include these secondary effects in its estimate.|