Just finished Lyle Goldstein and William Murray’s “Undersea Dragons: China/’s Maturing Submarine Force,” (International Security, Spring 2004, pp.161-196). It contains what I would call several factual errors. This requires some additional research, but some things are awfully suspicious.

For example, take the paragraphs concerning China/’s plans for the future of its sea-based nuclear deterrent (pp.171-173).

1. Goldstein and Murray claim that the “PLAN continues to operate” the Xia SSBN and that the submarine “has just emerged from a major overhaul.”


The Defense Intelligence Agency concluded in 2002 that China’s single Xia class SSBN is not operational. In 1999, General Eugene Habiger described the submarine operational life as it “went on one cruise and has been essentially in dry dock ever since.” The 2003 and 2004 editions of Chinese Military Power report that China has still not deployed the JL-1 SLBM aboard the Xia SSBN. The 2002 and 2003 edition of Chinese Military Power do mention an overhaul, but that overhaul occurred between 1995-1998—before the above statements that indicate the PLA does not continue to operate the Xia SSBN.

2. They claim that “the first of the Type 094 SSBNs may have been launched in 2003, on schedule to become operational in 2005.”

I suspect the authors have confused two types of submarines. According to Chinese Military Power in 2003, China “launched the lead hull of its next-generation SSN, the Type 093-class, which is expected to enter service by late 2004 or early 2005.” No US intelligence official has made a similar comment about the launch of the 094. Given how quickly the word leaked that the US detected a new Chinese diesel submarine, you have to wonder.

3. Finally they claim that “the JL-2 will reportedly carry either a solitary 1-megaton warhead or alternatively three to eight multiple independent reentry vehicles with nuclear yields of up to 150 kilotons, in addition to penetration aids.”

Chinese Military Power in 1998 reported that the nuclear warhead for the China’s new generation of solid fueled ballistic missiles was completed during 1992-1996. The six tests that completed validation for the design had yields between 50-150 KT range, not 1 MT. (You can take my word for it, do the math yourself or read this.)

The RV, which the National Air Intelligence Center estimated at 470 KG, is too heavy to place more than one on any of China/’s solid fueled ballistic missiles including the JL-2. When asked, point blank, “How many missiles will China be able to place multiple reentry vehicles on?” National Intelligence Officer Robert Walpole responded: “In the near term, it would be about 20 CSS-4s that they have, the big, large ICBMs. The mobile ICBMs are smaller and it would require a very small warhead for them to put multiple RVs on them.” He noted China would have to use the RV designed for the DF-31/JL-2 to accomplish this task.

According to the 2001 National Intelligence Estimate, “pursuit of a multiple RV capability for its mobile ICBMs and SLBMs would encounter significant technical hurdles and would be costly.” [Emphasis in original]

Although some advances in miniaturization could be made without resuming nuclear testing, the intelligence community believes that China’s continued testing moratorium would “impede China from placing multiple warheads on a mobile missile.” Former Los Alamos Director Harold Agnew, who concluded, in reference to allegations of Chinese nuclear espionage, that “No nation would ever stockpile any device based on another nation’s computer codes. … If China doesn’t resume testing, no harm will possibly have been done other than to our egos.”


If the authors were aware that the unclassified intelligence assessments contradict their claims, they should have alerted the reader to that fact and made some attempt to reconcile the competing claims.

This is important: You can find material that will support virtually any point of view about the PLA, which the authors demonstrate when they cite one strategist calling for “a future force of twelve SSBNs and thirty SSNs to augment a fleet of sixty-six conventional submarines.” (DIA expects just a total of “over 20 advanced diesel and third generation nuclear submarines.”)

US intelligence estimates are not, of course, flawless. But they are the proper place to start any analysis for at least three reasons:

  1. The US intelligence community has unparalleled access to national technical means of collection. For example, the intelligence community uses a variety of means to monitor ballistic missile tests (which is how the evidence of Chinese penetration aids was confirmed). There is no comparable unclassified source of such data, unless it is released by the government conducting the test.
  2. Other sources are difficult to assess. Citing a publication, such as Jane’s or IISS, tells us very little about how the authors gathered the information in question or what sort of review it underwent. By contrast, the intelligence community employs a well known methods of collection and review process that can be considered for gaps or bias.
  3. Intelligence assessments can be assessed over time. For example, the intelligence community has tended to exaggerate Chinese ballistic missile deployments, in part because Chinese industrial capacity tends to exceed production. This is useful information when considering estimates about future Chinese deployments.

After we established official estimates, we can then ask whether they make sense or are consistent with other sources of information. But only after

Paul Godwin offered this cautionary note on the wealth of material in interpreting Chinese military doctrine:

It is important to determine the authoritative source of PLA doctrine and strategy. Presumably, the PLA Academy of Military Science (AMS) fulfills this role, as it has since it/’s founding 1958. Nevertheless, with extensive doctrinal materials accessible from military journals and books, the field would be well served if we could separate the wheat from the chaff among the doctrinal assessments evidently available. It is my sense that there is a lot of “chaff” out there.

Spin-offs from the MA and Ph.D. programs now part of the PLA NDU Institute for Strategic Studies, the AMS and other centers of professional military education (PME), including service colleges and research centers, are obvious sources of doctrinal speculation. This not to suggest we should ignore doctrinal assessments that lack authority, for these can provide important indices of where the PLA may be going. However, our research agenda should include distinguishing between what is authoritative and what may be entirely speculative.