July 12, 2010 | Lecture on Arms Control and Nonproliferation
Abstract: The Obama Administration is asserting that the New START arms control treaty with Russia has a “robust” verification regime, and that it is effectively verifiable. But it is certainly much less verifiable than the original START. The U.S. will know significantly less about current and future Russian missiles under New START, and the Russians will be able to do much to advance and expand their strategic forces. The consequences of circumvention or cheating are more dangerous and destabilizing in the absence of robust U.S. missile defenses. Former Assistant Secretary of State for Verification, Compliance, and Implementation Paula DeSutter explains how the devil is in the details.
One downside to what I view as the Bush Administration’s more realistic view of arms control—which is that it is only worth pursuing arms control if it serves our national security interests—is that far fewer people are familiar with the intricacies of arms control concepts and jargon, especially when it comes to United States and Soviet, and then Russian, strategic arms control. The American people have not been exposed to these issues for some time. The Moscow Treaty doesn’t really count, in my view, because there was little public debate prior to ratification.
The lessened familiarity with the terminology and concepts that we arms control geeks talk about means that when discussing these issues, we will lose our audience if we speak the way we speak among ourselves. We really need to focus on commonsense concepts and explaining issues to our great-aunt or our cousin or any of the other fine Americans across America, who care about this great nation, about our security, and about the safety of our generations to come.
The American public—blessedly—are not arms control experts. They will not read the New START treaty and its annexes article by article. But they have common sense and do not want to be spoken over or down-to by a bunch of D.C. snobs, which is how they are likely to see those of us from “inside the Beltway.” They expect their Senators and Representatives to do due diligence. They expect that their tax dollars are paying them to look this over, explain the issues to them in commonsense terms, and then cast their votes to make sure that their security and their interests are protected.
We are training a new generation of rising experts, of people who are willing to focus on these issues and who care about issues that your friends back home and at high school reunions probably don’t want to talk about. It’s really important work and its not always fun. When you become a true certified geek, you will think this is really fun. So when you hit that moment, go for it! May you build on the good things we’ve done and fix the things we didn’t do as well as we thought we had!
Basic Questions on Verification
Everyone here should have read The Heritage Foundation’s recent Backgrounder, “New START: Potemkin Village Verification” on New START’s verification weaknesses. If you haven’t, I strongly recommend it. It hits virtually all of the major points that are wrong with verification of the treaty. I personally think it was a bit optimistic, if that tells you anything.
So what I wanted to do today is to talk about some of the general concepts and hopefully the commonsense evaluations that should guide assessments of verification. We know that there is no perfect verification. We are talking about sovereign nations. We are talking about whether verification is good enough, given the risks. How do we evaluate that? The United States seeks to answer two questions: What is the degree of verifiability, and is verification effective?
The Degree of Verifiability
First, we need to consider the proposed limitations and the clarity of the language by which the limitations are expressed, which must be weighed against our ability to detect noncompliance in a timely fashion, using both our own national means and methods of verification and possible treaty-mandated or agreed-upon cooperative measures. This assessment must include an evaluation of the means verified parties have of denying the United States the ability to detect noncompliance. Language clarity is an often underestimated element of verifiability, because to assess compliance, verifiers must be able to determine if a detected action is permitted or prohibited.
Effectiveness of Verification
Degree of verifiability must be then be weighed against a broader set of criteria to determine whether verification can be considered to be effective. Such “effectiveness” judgments are informed by a broader context, including: the compliance history of the parties to the potential agreement; the risks associated with noncompliance; the difficulty of responding to deny violators the potential benefits of their violations; and the impact of constraints imposed on U.S. freedom of action, particularly given the risk of undetected cheating prior to a “breakout” from a regime.
Why do effectiveness judgments have to be informed by larger context, like the compliance history of the party with whom you are reaching the agreement? It is common sense. If you are reaching an agreement with the Brits, you are not going to be as concerned as if you are reaching an agreement with the North Koreans, the Iranians, or, let’s say, the Russians—who have violated every agreement we have ever had with them. Russia is currently not complying by policy, and they’ve said so, with regard to the Treaty on Conventional Forces in Europe.
You also have to examine the risks associated with noncompliance. There are some agreements where it does not really matter very much if somebody cheats because the agreement is not very important, does not much constrain the United States, and the consequences for national security are limited if there is cheating, although violations can be politically significant. Another part of the assessment should be that if the other side is cheating, how difficult is it to respond either by changing your national programs and policies to redress the imbalance created by noncompliance or to bring that country back into compliance? Iran is the perfect example of why that can be so difficult. It’s Fred Iklé’s “After Detection—What?” argument.
The best-case scenario from a national security standpoint is an agreement which has a high degree of verifiability, is reached with a good treaty partner with a record of compliance as scrupulous as our own, with clearly understood and readily implementable sanctions for noncompliance, but which does not constrain the United States’ freedom of action in pursuing unilateral measures to secure the nation.
The worst-case scenario from a national security standpoint is an agreement with a low degree of verifiability, with parties with a history of intentional noncompliance, that significantly constrains U.S. freedom of action, and with only a low capability to deny a violator the benefits of its violation and restore at least the level of security that existed prior to the agreement.
Such a worst-case scenario would be compounded if the ineffectiveness of verification was poorly understood, since this would inevitably lead to a false sense of security. Other tools and approaches to address the threat thought to be addressed by the agreement are unlikely to be pursued with the rigor and urgency that might be called for. In such situations an agreement can therefore damage, rather than enhance, national security and international stability.
Let us ask a question: Is any treaty better than no treaty? Does it matter what the treaty says?
Secretary of State Clinton echoed statements which have been made by others, including former Secretary of Defense James Schlesinger, that having negotiated and signed the treaty, if we don’t ratify it, we will: a) have no strategic nuclear agreement with Russia, and b) lose credibility.
Senator Kerry basically said the world would end without it, and that with it proliferation would end and peace would break out. Senator Lugar was more moderate: “In my judgment, the question before us is not whether we should have a strategic nuclear arms agreement with Russia, but, rather, whether the New START Treaty’s provisions meet our objectives….”
The Senate can: a) give its advice and consent; b) give advice and consent with reservations and express its concern and interpretation of what it gave its advice and consent to; c) refuse to give its advice and consent.
However, verification is only one of the issues. Does it really matter so long as we are getting some data? Should verification be weaker or stronger as numbers go down?
The Administration is asserting that this treaty has a “robust” verification regime, and essentially that it is effectively verifiable. But it is certainly much less verifiable than the original START. The Administration has argued that less verification is needed now. To me, at lower numbers the consequences of circumvention or cheating are more dangerous and destabilizing, particularly in the absence of robust missile defenses.
Monitoring Warhead Numbers
Secretary of State Clinton said in her testimony on June 17, that “for the first time ever, we will be monitoring the actual numbers of warheads on deployed strategic missiles.” Sounds pretty good. Now, remember how to make commonsense arguments, how to explain this, how Senators are going to explain this to their constituents.
The verification measures in the New START treaty add nothing to what was there before in the original START treaty. They are using the original START Reentry Vehicle On-Site Inspection regime, complete with all of the same shrouds and covers that were used during the original START, some of which we found to violate the Treaty because we couldn’t confirm the number of Reentry Vehicles (RVs). And those are all still permitted. But they say we are going to use the Radiation Detection Equipment (that was negotiated for the New START) to confirm that an object that appears to be a possible RV is non-nuclear. That, they say, solves the problem. Maybe, maybe not.
I was not a big fan of the attribution regime for missiles because it didn’t always take into account the full capacity that we knew Russian and Soviet ICBMs had—but there’s nothing in the treaty that says the Russians may have only a certain number of Reentry Vehicles on any particular type of missile or any particular missile. And you can change the number of warheads on any given missile. So you do your RV onsite inspection and you discover that missile X in that silo has six Reentry Vehicles when you thought it had only a couple of warheads. So what? That tells you nothing about any other missile in the inventory.
Now, there can be either intentional or unintentional noncompliance. The original START treaty was very complicated, and so we anticipated some unintentional noncompliance. But we also saw significant intentional noncompliance. With the New START treaty, I think it is true that there will be fewer violations. Why? Not because the Russians have changed their compliance approach, but because there is no reason for them to bother cheating. Cheating, especially on any significant scale in the New START treaty, if you find it, it is equivalent to the Russians simply being ill-mannered. So getting back to RV onsite inspection, in my view there is no basis for the Administration’s statement that there is something new there.
Unique Identifiers (UIDs)
But, they say, we have another verification achievement: Unique Identifiers (UIDs). Secretary of Defense Gates said at the May 18 hearing that: “Unique identifiers, for the first time, will be assigned to each ICBM, SLBM, and nuclear-capable heavy bomber, allowing us to track the disposition and patterns of operation of accountable systems throughout their life cycles.” That sounds pretty good. Again, with verification the devil is always in the details.
Senator Isakson asked Admiral Mullen about whether the unique identifiers would be visibly or technically detectable: “Is that going to be like a transponder from an airplane?” Admiral Mullen said: “I think some may know. I don’t. It is very clear that it was going to be visible and verifiable and every single weapon would have it.” He later clarified that “actually, the UIDs are mechanical, they’re not technically detectable.”
The Inspection annex says that unique identifiers “shall be applied by the inspected Party, using its own technology…. Such a unique identifier shall not be changed. Each Party shall determine for itself the size of the unique identifier.” Inspectors are supposed to be able to check them. So you can’t see them through National Technical Means (NTMs), you have very few inspections, and even though the inspected party is not supposed to change the unique identifier, how would you know? Plus, there will be no way to know if there are duplicates. Has anybody ever painted a room? It is easy to change paint. Maybe I am underestimating the ability of this system to work. So, I think of this as “verification by paint, or in a good case, nail polish.” This contributes in this regime, as near as I can tell, virtually nothing.
Numbers of Inspections
The number of inspections, as I mentioned, went down. There are 18 inspections permissible per year (10 are the “Type 1 Inspections” of deployed bases and 8 are “Type 2 Inspections” of non-deployed facilities and items). Secretary of Defense Gates said we can do a data inspection and Reentry Vehicle On-Site Inspections at the same inspection. The Administration, in the Questions and Answers, said that since under the old START “there are 73 facilities that we inspected...under this treaty, there are only 27. And, in fact, based on the number of inspections—18—there are almost twice as many inspections per facility per year than under the previous treaty.” Maybe. So, we have onsite inspector numbers going down, the rights of the inspectors to go to critical places much less. So, NTM takes on a more important value in this.
Robust National Technical Means (NTMs)
Secretary of Defense Gates said: “The Treaty provides for noninterference with national technical means of verification, such as reconnaissance satellites, ground stations, and ships. This provides us with an independent method of gathering information that can assist in validating data declarations.” Okay, sounds pretty good. What’s the problem?
First, the problem is that our NTM infrastructure is, shall we say, broken. We do not have the independent satellite capabilities to be able to achieve the level of contribution to verification that we had in the Intermediate Nuclear Forces (INF) treaty or in the START treaty. Second, the treaty permits concealment activities—you can’t interfere with NTM, but you can conceal activities at ICBM bases. After I read that I called people and said, can that be true?
Telemetry Not Necessary
The other thing that happened has to do with telemetry. Secretary of Defense Gates said that “while telemetry is not needed to verify the provisions of this Treaty, the terms, nonetheless, call for the exchange of telemetry on up to five launches per year per side.” The other side decides which systems they want to give you telemetry for! Admiral Mullen added later: “the telemetry needs of this Treaty are different from the telemetry needs we had in the past. And we really don’t need telemetry for the kind of verification that we need for this Treaty that we had before, to include the ability to understand the weight of a missile, when we didn’t know what was actually inside it.”
These are fundamental misunderstandings of what we use telemetry for! You want to know the throw-weight so you can figure out how many RVs a missile can carry and how many RVs it is being tested to carry. You want to know how many RVs they are putting on the next generation of missiles, and how many on the SS-27. You need to know that. So, since they say they are monitoring the actual number of deployed RVs, they ought to want to know carriage capacity. Since the testing party decides which flights it will broadcast, you can’t expect to get telemetry and interpretive data for any flight tests of new systems. It appears that the only telemetry for which data is exchanged is the telemetry that is broadcast. Encapsulated data isn’t shared. It also appears that only the data from the first stage will be broadcasted.
Now you can do some monitoring on your own but that is why these telemetry protocols were created. And what the Administration has done by shortening and constricting the telemetry protocol demonstrates a fundamental lack of understanding of the importance of telemetry and the telemetry exchanges for verification.
One interesting side note, I personally fought hard during the Bush Administration to get the full START Telemetry Protocol into the treaty, with a cut-out only for missile defense tests (interceptors and target vehicles). The Pentagon didn’t want any of the Protocol in, and the intelligence community didn’t want the missile defense exemption. They fought for months against the missile defense exemption because they were concerned they would lose too much data. Eventually the policy, having taken note of the intelligence community’s concerns, was that we would insist on the full START Telemetry Protocol with the exception for missile defense. In New START, the Obama Administration has agreed to significantly less, but apparently the intelligence community is going along with it. It does not make much sense to me and is very disturbing.
What the Russians, In the Course of Just Being Russians, Could Do to Get Around the Numbers
Article III, Paragraph 6 provides that if all the ICBM or Submarine-launched Ballistic Missile (SLBM) launchers for a certain missile type are eliminated, all those missiles of that type go out of accountability. It means that the Russians could eliminate the launchers for a type of missile, taking it out of treaty accountability, but retain a capability to launch them from a different launcher.
Eliminated systems can only be inspected or observed by NTMs within a certain timeframe. It would be a simple matter to “eliminate” items at several locations, thereby precluding inspection of all “eliminated” systems. If the U.S. inspectors declare an elimination to be inadequate, the Russians can argue it was different from the others.
What is even more disturbing is the fact that the only use of converted launchers or missiles that is precluded by the treaty is the use for missile defense. Anti-satellite weapons, Space Launch, marketing abroad like they did the INF Transporter-Erector-Launchers are not prohibited.
The Protocol, Part 3, Section 1, Paragraph 2, sets the standard for elimination. Items are to be rendered “inoperable, precluding their use for original purpose.” A flat tire would come close to making a Transporter-Erector-Launcher inoperable and precluding its use for its original purpose.
Moreover, according to Paragraph 4, if the Russians propose a conversion or elimination procedure the other side believes is inadequate, all the Russians have to do is do a demonstration under the Bilateral Consultative Commission. Whether or not the other party is convinced about adequacy, those procedures can be used to take systems out of accountability.
Last but not least is the issue of heavy bombers. Russian Prime Minister Putin has called for production of two new heavy bombers. Since they only count as one warhead, this is a good deal for the Russians. Moreover, there are no provisions for determining the number of weapons on deployed heavy bombers. Given that the bombers are counted as one nuclear warhead, this makes no sense.
So, we will know significantly less about current and future Russian missiles under New START. We are not going to gain much through identifiers; we are not going to gain much for treaty purposes through the Reentry Vehicle onsite inspections; the telemetry protocol has been gutted. And so what are we left with? Not that much. So it is true that even with a fantastically tight verification regime you are going to have the possibility of violations. We knew in the START treaty—it was very technical and complex and you had to go back and forth—that there would be at least technical violations.
What is true, I believe, of the New START treaty, again on the upside, is that there will be very few violations because we have made it virtually impossible for the United States to collect the type of information and weigh it against a solid treaty with solid language that would permit us to say they are cheating. The Russians can do so much under this treaty to advance and expand their strategic forces over the length of the New START treaty and our ability to determine whether or not they are doing that and whether it violates the treaty is very, very low. The degree of verifiability if very low.
I would assert that an assessment that says it is effectively verifiable would be incorrect. It is also true that the Senate can give its advice and consent to an unverifiable treaty.
—The Honorable Paula A. DeSutter served as Assistant Secretary of State for Verification, Compliance, and Implementation from 2002 to 2009. This publication is based on her prepared remarks at a Heritage Foundation panel, “A Good or Bad START?”
The New START Working Group, “New START: Potemkin Village Verification,” Heritage Foundation Backgrounder No. 2428, June 25, 2010, at http://www.heritage.org/Research/Reports/2010/06/New-START-Potemkin-Village-Verification.
This goes back to the July 2009 agreement on nuclear weapons reductions between the United States and Russia.