The verification regime of the New START Treaty has again come under fire in recent weeks. As Marshall Billingslea, President Trump’s new special envoy for arms control, said in a recent Washington Times interview, “The Obama administration negotiated a very weak verification regime. It really has very little of what the original START treaty contained and has significant loopholes in the way verification is physically conducted, which the Russians have been exploiting. So those behaviors also have to stop.”
This comment is odd, first because Trump administration officials have repeatedly acknowledged the security benefits of New START. Also, not a month earlier, the State Department released its annual compliance report, in which the entire US government, including the White House, agreed that Russia is in full compliance with New START.
The argument that Billingslea made also popped up during the treaty’s ratification debate in 2010: New START is inadequate because it does not replicate the verification regime of the 1994 Strategic Arms Reduction Treaty (START).
Indeed, the New START verification regime does not replicate all the measures contained in the earlier agreement, and for very good reason: New START made good use of what worked in previous treaties, but it did not perpetuate problems encountered in implementing those treaties. New START contains detailed, streamlined procedures that make inspections reliable in confirming information that the Russians provide to the United States, and, of course, vice versa. It is precisely this carefully crafted verification regime that has made the treaty, which limits each nation’s strategic nuclear arsenals at no more than 1,550 deployed nuclear warheads and 700 delivery platforms, so effective and trustworthy.
The purpose of verification. To get our arms around this debate over New START’s verification procedures, a couple of basic points need to be understood from the outset:
First, the verification regime of any arms control treaty must be effective. That means, according to a definition laid down long ago by Paul Nitze, that the monitoring procedures must be good enough to detect cheating in time to do something about it—before the cheating party gains a strategic advantage. In other words, verification regimes must not tempt either side to try an illicit treaty break-out.
Second, the verification regime must be designed to monitor the obligations of a particular treaty; the information gathered is limited to that goal. It is not the same as intelligence collection, where we want to maximize the amount of information gathered about the other party’s weapon systems.
Third, there is no “one size fits all” or “best and only” way to go about ensuring that both sides are living up to a treaty. Sometimes measures that are needed in a particular treaty no longer make sense when they are fast-forwarded to a new treaty that goes about limiting nuclear weapons in a different way. Governments also should have the freedom to decide when improvements to verification must be made.
Early on in the New START negotiations, those of us on the US team came to the conclusion that we could improve knowledge of actual warhead deployments if we could dispense with the counting rules that had dominated earlier arms control treaties, including START. The counting rule for ICBMs, for example, designated each missile of a type to have a certain number of warheads, determined by the maximum number that had been observed in flight testing. In that way, the Soviet SS-18 heavy missile was determined to have 10 warheads, although its carrying capacity is much greater—by some reckoning 14 or even more.
This approach came to penalize the US submarine-launched ballistic missiles (SLBMs), which according to the START counting rule could carry eight warheads each. In fact, the United States had decided to download the SLBM force, so actual warhead deployments on each missile were sometimes no more than four. As a result, the START database showed higher numbers of US submarine warheads than actually existed on the missiles: Our SLBMs were being over-counted. We wanted to fix that problem—a matter important to the Defense Department and the US Navy.
In other words, we needed to ensure that the verification measures were tailored to the limits of the new treaty, and did not simply replicate START in a way that had no purpose for this treaty.
Here was where the great telemetry debate began. The sharing of telemetry data, essentially flight test data for missiles, had an important function under START, because it was used to confirm the viability of the counting rules: a missile was considered to carry the maximum number of reentry vehicles for warheads with which it had been tested. The way each side confirmed that maximum number was to exchange telemetry data about flight tests under detailed START procedures.
In New START, we discarded the counting rules in favor of confirming declared warheads on the front of missiles through reciprocal inspections; in fact, we did not need telemetry measures to confirm compliance with the warhead limits in the new treaty. Nevertheless, the importance of telemetry was hotly debated during the course of the negotiations and afterwards, during the ratification fight for New START in the Senate. In the end, we did agree to telemetry measures with the Russians. They became an important confidence-building aspect of the treaty during its implementation, when both sides exchange telemetry data on an annual basis, on a select number of missile flight tests.
That is to say: The New START methods for confirming warhead numbers were indeed different from what had been contained in the old START. They were better because they got us a more accurate accounting of warheads.
Monitoring missile production. Another issue that was hotly debated during the New START ratification fight involved what was known, rather cumbersomely, as Perimeter Portal Continuous Monitoring (PPCM) at the Votkinsk Missile Production Plant about 630 miles northeast of Moscow. When START was negotiated in the late 1980s and early 1990s, both the Soviet Union and the United States thought that they would be deploying mobile ground-launched intercontinental ballistic missiles. The two countries would therefore need to be able to distinguish such intercontinental-range missiles from mobile ground-launched intermediate-range missiles, which were banned by the Intermediate-Range Nuclear Forces Treaty (INF).
Both sides agreed to continuous monitoring at production plants, the Soviets at Votkinsk, the Americans at the Thiokol Corporation missile plant in Promontory, Utah. The purpose of the monitoring was to ensure that no intermediate-range missiles were being produced under cover of the intercontinental missile production runs.
At the end of the day, the United States cancelled its mobile missile program, so monitoring at Thiokol was not needed. At Votkinsk, however, monitoring went on from the time START entered into force in 1994. It involved American teams actually living in Votkinsk, the city adjacent to the plant, and continuously monitoring the missiles exiting the plant, according to strict procedures that were worked out in the START verification regime.
It was an expensive program for the United States to implement, and by 2009 it had been in place for almost 20 years. The inspectors were not seeing any new SS-20s, the Russian ground-launched intermediate-range ballistic missile. In the meantime, the verification regime of INF had been wrapped up in 2001; no other on-site monitoring of potential INF missiles was underway.
Because of the Russian INF violation that emerged in 2011, one might argue today that we were wrong to dispense with missile production monitoring in New START. However, the Russians were careful to center their INF-violating program on the 9M729 cruise missile, which is not produced at Votkinsk. The new ballistic missile that potentially is of INF range, the RS-26, was tested to intercontinental range and would therefore be counted and captured in New START. Indeed, the Russians offered to exhibit it as a New START system. Thus, an expensive monitoring program would have continued to operate at Votkinsk for an additional 10 years, with no treaty purpose.
Another important factor in our considerations during New START negotiations involved national technical means of verification, a combination of satellites, ground-based radars, and aircraft through which the United States monitors the Russian nuclear arsenal. The Russians, of course, deploy the same kinds of monitoring systems to keep track of US systems. From the earliest days of negotiated nuclear arms control in the 1970s, noninterference with national technical verification has been a basic principle to which both sides can agree.
“Noninterference” means that the Russians, for example, will not drape nets over their mobile missiles to prevent clear satellite pictures. If they do, they will be called on it.
Noninterference with national technical verification was one of the earliest and easiest points of agreement in the New START negotiations. So US and Russian officials both understood that we would continue to monitor Russian missile flight tests for range, to catch violations. Indeed, that is what happened with our discovery that the Russian 9M729 missile is a violation of the INF Treaty. It would not have been caught by on-site presence at the Votkinsk Missile Production Plant.
Innovating toward better verification. Dispensing with the counting rules for ICBMs and SLBMs was the most important innovation in New START. We thought long and hard about how to confirm the number of warheads that we or the Russians would declare on an individual missile. The basic breakthrough was the ability to physically check and count what was on the front end of each type of missile, which we accomplished through reentry vehicle on-site inspection. Inspectors would be given the opportunity to count objects on the front ends of missiles, which would be opened up but covered with a soft or pliable cover so that objects could be counted without revealing their technical characteristics
One problem was that every object on the front end of a missile is not a nuclear warhead. Both sides deploy penetration aids, chaff dispensers or decoys that are used to spoof missile defenses. They are conventional, not nuclear, objects. To differentiate between them, we came up with an idea: Inspectors would be able to use radiation detection devices to determine that an object was non-nuclear in nature. In other words, an inspector would be able to confirm the number of nuclear objects declared; if there were extras, he would be able to determine that they were not nuclear warheads.
It sounds easy, but figuring out the process required complex and difficult discussions. For example, how pliable should the soft cover be that is used during the inspections? There could be no possibility that a warhead was hiding under a cover that did not properly display all of the objects underneath. What kind of radiation detector could be used? Both sides were resolved that the detection equipment should be able to determine that an object is non-nuclear, but not discover any design secrets in the adjacent warheads. Many technical discussions ensued, leading to detailed procedures so that each side could check and accept the other’s detection equipment.
The United States and Russia also had to determine the selection process for missile inspections. Clearly, neither country would want inspectors to have access to every single missile in a unit. Such extensive access would prolong an inspection to such a point that base operations would be seriously impeded. Foreign inspectors on site for so long would also potentially endanger secrets that each side wants to keep. In the end, we went with usual inspection practice, wherein the inspecting party gets to select at random the missile to be inspected.
Re-entry vehicle on-site inspection has now been tested in practice during New START implementation, and it has allowed the United States to get away from the problem of over-counting warheads on submarine-launched ballistic missiles. It has also opened the door to the further possibility of warhead inspections in future. Earlier arms control treaties had focused on counting delivery vehicles and launch systems—missiles, submarines, bombers—because they were large objects that could be easily counted, even from outer space. Warheads were constrained by dint of their association with their delivery vehicles—the counting rules. Once the delivery vehicles were eliminated, the warheads went into storage and so were no longer considered a threat.
Such an approach also meant that we could not, or did not try, to distinguish conventional from nuclear missiles. Getting direct access to the missiles to determine if they were nuclear or conventional was considered too sensitive. For that reason, the INF Treaty, although it has nuclear in its name, banned all ground-launched missiles between 500 and 5,500 kilometers in range—conventional and nuclear. Moving away from the counting rules and focusing on confirming what is actually on the front end of missiles was a big change from the past, an innovation that opens up new opportunities for future arms control agreements. In particular, banning or limiting nuclear warheads while letting conventionally armed missiles continue to be deployed becomes an option.
In the same way, it opens up new opportunities for limiting non-strategic nuclear warheads, since we have now gained practical experience monitoring non-nuclear objects that are “cohabitating” with nuclear warheads. The techniques and monitoring equipment that have been used in New START will have applications in discerning nuclear from non-nuclear objects in other settings. For example, a country engaging in an non-strategic nuclear arms control regime could use procedures born of the New START experience to prove that a storage facility was not housing nuclear warheads. Non-strategic nuclear weapons limitations have been the Holy Grail of the arms control community for a long time; re-entry vehicle on-site inspection in New START has begun to make such limits more feasible and practical.
Innovation No. 2: Inspection system overhaul. Another aspect of New START verification that has been hotly contested is the number and types of inspections it requires. START had a wide array—12 different types adding up to 28 total inspections. They were divided into categories: 15 data update inspections, 10 reentry vehicle inspections, three inspections of formerly declared facilities. Suspect sites could also be inspected, but it would have to be under the data update quota.
From the US military’s perspective, the START inspection regime was far from perfect. Nuclear base commanders were particularly concerned about the impact on their operating tempo. Bases essentially have to shut down operations while the foreign inspectors are on site. Sometimes the shut-down can stretch for days as the site is prepared, the inspectors arrive and carry out their monitoring work, and a post-inspection assessment is done after they leave. Such delays are expensive propositions, especially when one inspection follows hard on the heels of the previous one, as was permitted under START. The delays were especially hard on the strategic submarine force, which operates on strict schedules for at-sea deployments.
After 15 years of START inspections, the US government, especially the Defense Department, had come to the conclusion that some of the START inspections could be combined, consolidated, or eliminated, to make New START inspections simpler, less costly, and less of a burden on nuclear base operations. Language of this kind was inscribed in the July 2009 summit statement that Presidents Obama and Medvedev signed in Moscow. It immediately aroused controversy, with US critics arguing hard that the United States was weakening the strict START verification regime, letting the Russians get away with less stringent measures and endangering US security.
Those of us on the New START team knew we had a good reason to streamline the inspection regime, but we had to make the case that we could do effective verification in a different way in New START. To accomplish this goal, we decided to set loose a team of weapon system operators from the Defense Department and seasoned inspectors from both Defense and the intelligence community. Their job was to work through the problem and develop a new solution for the inspection regime.
The team produced a dramatic new concept for verification. It consolidated the 12 different types of START inspections into two major types of inspections: inspections for deployed systems would be called Type 1 inspections and non-deployed systems would be Type 2.
In Type 1 inspections, several activities would be carried out at once, including confirming the warheads declared on reentry vehicles and ensuring the data that we had about the presence of Russian missiles at bases were accurate. Essentially, Type 1 inspections would focus on operating bases and the weapon systems present there and would be used to confirm several aspects of their status—numbers of missiles, numbers of warheads, and deployment locations.
By contrast, Type 2 inspections would focus on non-operational bases, i.e., storage and maintenance facilities, or facilities where systems were being converted or eliminated.
The beauty of this concept is that it retained all of the tasks that were present in the 12 types of START inspection, but folded them into only two. For example, START data update inspections, formerly declared facility inspections, and conversion or elimination inspections were all folded into Type 2 inspections. The inspections would take place over a longer period of time to accommodate the multiple tasks involved, but individual facilities would not be inspected so often—repeat inspections at any given base in a single year were no longer possible—thus their operations would not be interrupted so often.
Some of the ideas that the brain trust devised ended up being important endgame issues for the negotiations. For example, our defense and intelligence experts determined that if each launcher and delivery vehicle had a unique identifying number, which would be recorded in the database and tracked through constant notifications, then we would be in an improved position to monitor the Russian ICBM force on a 24/7 basis, and inspectors would have a clearer picture of what missiles should be present when they arrived at an operating base.
The Russians threw up all kinds of objections to the identifying numbers, even to the point of claiming that inscribing numbers on the sides of their missiles would interfere with missile flight. Finally, after exchanges on the matter between the Chairman of the Joint Chiefs of Staff, Adm. Michael Mullen, and the Russian Chief of the General Staff, Gen. Nikolai Makarov, and even between President Obama and President Medvedev, the two sides agreed that the existing serial numbers that each side used to track and account for their weapon systems would be used for this purpose. It was a simple solution, but one that has proven vital to the effectiveness of New START verification.
The message from this experience is a lesson that would serve well in any negotiation: Sometimes it pays dividends to turn the experts loose and let them think through a problem until they come up with a solution that works for all. We were lucky to have an experienced team lead, Ted Warner, who was the senior representative of the Secretary of Defense to the negotiations. He knew a great deal about past arms control regimes but was not so wedded to the old ways that he would not consider how to adjust them to new circumstances.
In addition, the dynamics among the US experts were important: The weapons systems operators knew the operational costs that inspections could impose, and the inspectors knew what they needed to verify that treaty commitments were being respected. The kind of “creative combat” that ensued ensured that the United States got exactly what it needed from New START: effective verification at a reasonable cost to US operational tempo.
The value of New START verification. In the end, the United States got what it wanted in the New START verification regime: streamlined inspection procedures at a sufficient level of detail to be effectively implemented. We made good use of what had been in previous treaties, but we were not trapped into perpetuating problems, such as over-counting of warheads, that were not in the interest of the United States. Most important, we ended up with detailed procedures that make the inspections reliable in confirming information that the Russians provide us. The same, of course, goes for them, since everything is done on a reciprocal basis.
The New START verification regime has a bigger value than the treaty itself: It bodes well for future arms control regimes that focus more on accounting for warheads, nuclear or conventional, than has been possible in the past. For anyone interested in pursuing limitations on non-strategic nuclear warheads, as the current administration is, this should be good news. Not only should we continue to implement the treaty, but we should also be studying the lessons that can be learned from it. Thanks to New START, a long-sought goal of nuclear arms control— controlling and eliminating warheads—is finally within our grasp.