Jeffrey LewisIs the Musudan Real?

The other day, I posted an entry at 38North outlining three scenarios for North Korea — that North Korea would eventually test a Musudan, that China has stayed Kim Jong Un’s hand for the moment and that North Korean politics have aligned against a test.

Two readers — Markus Schiller and Anon O’Moose — wrote in to observe that the piece would have been stronger had I considered a fourth scenario: what if the Musudan is not real at all?

I happent to think the Musudan is real for reasons I will explain, but I admit the piece would have been stronger had I considered the alternative possibility.  Consider this compensation for that oversight.

Markus Schiller has agues in his monograph, Characterizing the North Korean Missile Threat, that North Korea probably does not have any operational intermediate range-ballistic missiles based on the SS-N-6.  Markus believes that the North Korean may never have tested an indigenously manufactured missile:

The best-supported hypothesis (i.e., the one with the lowest inconsistency score) is the “Bluff” hypothesis. According to this hypothesis, in its testing, North Korea has launched Soviet/Russian-made missiles (that are proven but old designs) to maximize the appearance of performance, but may never have tested missiles from its own production—any such indigenous missiles cannot have noteworthy reliability or accuracy

As a result, North Korea according to Markus is unlikely to have any operational SS-N-6 missiles.

Even if rumors of transfers of longer-range SS-N-6 missiles in the 1990s are true, these SS-N-6 missiles would have far  outrun their service life and are likely not operational anymore. (Contrary to other Soviet missiles, SS-N-6 missiles were  fueled in the factory and sealed, resulting in limited service life.)

In the scenario based on the “bluff” hypothesis, North Korea takes the display mockups from the parade, then deploys them to the coast with real Scuds, then enjoys watching us freak out about it before eventually putting them away. Since “the main purpose of the North Korean missile program is … to gain strategic leverage in foreign politics,” the deployment was a ruse.

He’s right — even if I don’t think the scenario is terribly likely — the article would be stronger for considering the possibility.  (My entry on the subject is here; Markus added his views in the comments.) Many of my colleagues agree with Markus.

On the other hand, the US intelligence community appears to believe that North Korea can produce Scud and Nodong missiles and that the Musudan is real. The 2009 Ballistic and Cruise Missile Threat, 2010 Ballistic Missile Defense Review and 2013 Report to Congress on Military and Security Developments Involving the DPRK all express the view the the DPRK’s missile program, including the Musudan, is real — as does a leaked 2009 cable containing a paper the US distributed to other MTCR states.

I”ll be the first person to suggest we ought to  scrutinize IC assessments based on open source information.  In general, my approach is to try to understand the judgements of the US intelligence community, the basis for these judgements and any areas of disagreement.  Absent some compelling information in my possession, I will default to those judgements.

In this case, I just don’t think the available information is conclusive enough to reject the view of the IC that the DPRK missile program is real. In particular, I am mindful of the fact that the US intelligence community has had the opportunity to examine Scuds sourced to both the Soviet Union and North Korea, with the latter coming from Libya and probably the UAE.  DIAC has a Scud in the rotunda, for Pete’s sake. They should know what these things look like. More seriously, the Missile Defense Agency has conducted intercept tests against “foreign military assets” — anywhere from Scuds to components like foreign reentry vehicles.  GAO even include a picture in their list of targets (third from the left.)

And then there are the many components, materials and production equipment intercepted on DPRK ships.  It would seem to be straightforward enough to assess whether these items represented a North Korean missile program or Soviet overstock.  The fact the IC continues to conclude the missiles are of North Korean origin is persuasive enough for me.

I am mindful, as Markus will no doubt point out, that some of the items we see in Iran or have seized do bear Cyrillic markings. So, obviously there is prima facie evidence to take his view seriously and respectfully.  I just don’t find it sufficient.  Interesting, though.  Very interesting.

I’ll leave you with some passages from that cable distributed to MTCR states:

North Korea’s ballistic missile program started in the 1980s, when it reverse-engineered Soviet-made 300 km-range Scud B SRBMs acquired from Egypt. In return for these systems, North Korea assisted Egypt’s efforts to domestically produce Scuds. Building on this success, Pyongyang began designing the 500 km-range Scud C in the mid-1980s. These Scuds have been exported to customers in the Middle East and are deployed in North Korea. Given their 20 years experience working with Scud technology, North Korea is able to design and produce extended-range variants of the Scud, capable of delivering payloads of over 500 kg to ranges up to 1,000 km.


Recently, North Korea has developed a new land- mobile intermediate-range ballistic missile (IRBM) derived from the Soviet SS-N-6 (‘Serb’//R-27/4K10) submarine-launched ballistic missile (SLBM), which was designed by the Soviet Union’s Makeyev Design Bureau in the 1960s. This technology represents a substantial advance in North Korea’s liquid propellant technology, as the SS-N-6 had a much more advanced engine and used more energetic propellants-unsymmetrical dimethylhydrazine (UDMH) and nitrogen tetroxide (N2O4)- than those used in Scud-type missiles. The new IRBM is a single-stage missile and may have a range of up to 4,000 km with a 500 kg payload. Moreover, development of the new IRBM is even more disturbing since this more advanced propulsion technology allows North Korea to build even longer-range missiles-or shorter range missiles with greater payload capacity-than would be possible using Scud-type technology


North Korea has developed most of the necessary capability and infrastructure to produce and assemble its ballistic missiles indigenously. Its deployed forces include hundreds of Scud and No Dong missiles, and we expect the new IRBM and Toksa to be fielded in the coming years.


Pyongyang continues to sell ballistic missile- related technology to countries in the Middle East, while seeking to re-engage with former customers in the region. It also is likely pursuing new markets for its missiles, including in regions such as Southeast Asia and Africa. North Korea offers a wide-range of ballistic missile services. North Korea almost certainly is willing to offer any missile design in its inventory for sale to customers interested in complete systems, and can design missiles to meet specific customer needs. For customers with established missile programs or otherwise lacking interest in complete systems, North Korea provides missile refurbishment and technical expertise, ground support equipment and launchers, and production technology. North Korea can also broker precision machine tools and other missile- related raw materials for customers through Pyongyang’s extensive procurement network.



  1. John Schilling (History)

    A related possibility is that the Musudan is real but is not derived from the R-27 aka SS-N-6. North Korea uses the same technology they used for the Nodong, to build a slightly larger missile with slightly better performance but with an external appearance that says to the informed observer, “Stretched SS-N-6; now with 30% more range!”.

    If that’s the case, then a test shot would likely reveal the deception.

    I think there’s enough other evidence of R-27 technology transfer that I don’t consider this explanation highly credible, but it is within the realm of possibility and it would explain the conspicuous absence of any flight tests of any R-27 derived system in North Korea.

    On the other hand, there are persistent reports from US and South Korean intelligence of North Korean ground test activity, and it should be possible for plume spectral analysis to distinguish between the UDMH-fuelled R-27 and any of North Korea’s traditional Kerosene-fuelled engines. So, again, the IC almost certainly knows the truth, and it would be quite perverse for they and the North Koreans to essentially collaborate on deceiving us in this matter.

    • Magpie (History)

      If you had R-27 tech, and you were North Korea, wouldn’t you copy it? It’s the Soyuz of mid-range ballistic missiles. And the only downside isn’t even a downside! Getting your peasants melted by a BFRC thanks to an enemy pre-emptive is pure propaganda gold!

      I still say they were begging to get shot at with their TEL erection, and when they realised they weren’t going to, they took their good missiles back home and had one last go with some short range stuff in the hopes that someone might still be on a hair trigger. No dice, and now they’ll need to pack up, lick their wounded pride, and get ready for next round.

      Kim Mag-Pi’s intel resources in the South clearly suck, so Kim Mag-Pi is taking his bat and his ball and is going home.

      Next time, we’ll try to own a battery in the South.

      (In their defence, this whole Game Of Silly-Buggers played quite well in the Chinese press, so it wasn’t a total loss).

    • Cthippo (History)

      Can you elaborate on the evidence for the R-27 technology transfer? The evidence I’ve seen referenced on this site always seems more than a little sketchy.

      I think that the fact that the Uhna uses kerosene fuel is the best evidence that North Korea either doesn’t have R-27 technology available to them, or else perhaps they have it but decided for whatever reason not to use it in their rockets. Uhna was arguably the most import and and certainly the most visible project in North Korea and literally the whole world was watching to see if it worked. IF they had R-27 technology and IF they felt it worked better than SCUD tech they would have used it in their most critical rocket.

      It seems to me that either North Korea didn’t get ahold of the R-27 tech, or perhaps they decided that it wasn’t enough of an advantage over the existing kerosene based systems and didn’t pursue it. It makes a certain amount of sense, given that they aren’t trying to build storable fuel rockets that have to fit into the hull of a submarine for months at a time, but rather road mounted ones which are relatively easy to fuel and maintain.

    • John Schilling (History)

      Jeffrey has already provided an excellent summary of the Origins of the Musudan IRBM that explains the case for an R-27 heritage.

      To deal with the specific questions asked, first, the Unha probably did use R-27 technology, in the form of (modified) R-27 vernier engines as main propulsion for the upper stage. Same arrangement the Iranians used on the Safir launch vehicle, and for the same reason – the improved performance of an NTO/UDMH engine is most important on the upper stage, but the complete R-27 main engine is kind of overkill for the upper stage of a small satellite launcher. Just the verniers are about right.

      However, the R-27 engine is far too small for the first stage of an Unha-scale launcher, and unlike the Nodong engine it can’t readily be clustered. The R-27 engine is a submerged design, that has to be tightly integrated with the propellant tank and is really supposed to be on the centerline of the tank. As the tank only has one centerline, that lets you have one R-27 engine per stage unless you plan to do a whole lot of redesign work. Since engine performance is least important on the first stage, and since they already had the adequate and easily clusterable Nodong engine, they went with the simple solution.

      North Korea probably could have gone either way for the second stage of the Unha, and it surprised many of us that this also appears to have used IRFNA/kerosene Scud-heritage propulsion. That’s a marginal case; the Scud engine is less efficient but had a thrust level better suited for the satellite launch mission.

      Which brings us to North Korea copying R-27 tech and adapting it to their own needs. They probably can’t do that. Even if they have a complete set of blueprints for the system, courtesy of some rogue Isayev and Makayev engineers, turning blueprints into a production line that cranks out reliable hardware is damnably hard work. Usually harder than the R&D process that came up with the blueprints in the first place, which makes spy-movie plots about “stealing the blueprints” generally ludicrous.

      North Korea has never been seen to build a production line, of any sort, that could reliably build hardware as sophisticated as the R-27’s Isayev 4D10 engine. That they apparently do have the simpler Scud and Nodong engines in local production is an impressive enough achivement; I’ve seen extremely talented Western technicans stumble and fall trying to do approximately the same thing under much better circumstances.

      Possibly the DPRK has worked a miracle of reverse-engineering here, something that would be almost unprecedented in the field. More likely, if they got hold of the hundred or so surplus R-27 missiles that are not fully accounted for from the cold war days, that’s all that they have. They can cut them up and use the parts in whatever combination they like, but they can’t have external-mount clusterable R-27 engines for the Unha first stage or reduced-scale R-27 engines for the upper stages.

    • Cthippo (History)

      Thanks for the reply, John

    • j_kies (History)


      While a sophisticated well-mixed amine propellant missile like the Titan has a distinct plume from things like hydrocarbon-LOX propellant combinations (and well different from Solids) these differences tend to be observable in ‘up close and personal view’. As mixing or fuel-oxidizer ratios become off-ideal (or very modest contaminants are included in fuels) the differences between amine propellant plumes and other combinations become less clear.

      So if the ‘IC’ knows the truth, its not clear what means it applies to ‘know’ that as remote sensing may not provide unique signatures. Perversity is not needed, mere limits to sensing and errors in extrapolation / analysis can suffice for anyone to create errors in estimates.

  2. flamesInTheDesert (History)

    Yeah,and maybe their tanks are made of cardboard.Come on mate silly times over.

  3. Keve (History)

    If South Korea and Japan taking over electronics and shipping industry are any possible sign of reality, I would be surprised to see North Korean missiles capability surpassing everyone else’s. Which country had more experience building missiles and exporting missiles in last 2 decades than North Korea? North Korea had access to more recent updated information, light alloy material, and technology than any other nations which still fields missiles build back in the 80’s(maybe) and 90’s. This reminds me of lesson from a story “Rabbit and Turtle Race”, also a phrase “If you can do it, than someone else figure it out to make it better”. “If you can walk, there is someone who can run; if you can run, there is someone who can fly.” “Never be too confident of yourself, and never under estimate your opponent, if you want to stay ahead”. Experts’ conversation should not be about “what North Korea cannot do”, but “what North Korea achieved and likely already have achieved based on speed of how fast USA made progress on missile technology in the past”: “if I can do it, someone else can do it better.”—historical fact of world history.

  4. z (History)

    What is known precisely about North Korea’s relationship with Pakistan and the Pakistan ballistic missile program ?

    It used to be reported that Pakistan gained its ballistic missiles and technology from NK and now they get their tech from china , is this still thought to be true ?

    If Pakistan got its missile manufacturing technology from NK ,and they still use it and test it ,what engines do they use ? are we sure that there is no ongoing tech relationship between Pakistan and NK ?
    Pakistan could now be quietly passing on information they now get from China and from their tests to NK .

    I find it odd that there are so many commentators that always come out and say that NK ballistic missiles don’t work or are fakes because there have been so few medium/long range tests from the Korean peninsular , and at the same time completely ignore NK relationships with other countries , especially Pakistan .
    If Iran and Pakistan are flight testing half a dozen ballistic missiles every year that were designed and built by NK engineers then there might be little need to flight test missiles from Korea .
    It might be that North Korean missile tech has more successful tests in the last 15years than the United States .

    certainly more successful tests than France .

    • John Schilling (History)

      The relationship between North Korea and Pakistan’s missile development program is that North Korea apparently provided Pakistan with the technology (i.e. not just blueprints, but direct technical assistance in developing a local production capablity) for the proven North Korean Nodong missile, which Pakistan put into production as the Ghauri and subsequently modified for their own purposes.

      Pakistan has subsequently move on to the Shaheen missile series, a family of solid-fuel missiles that appear to be based on Chinese designs. There probably was Chinese technical assistance in building these weapons, but no smoking-gun proof.

      There does not appear to have been any subsequent technology transfer from Pakistan to North Korea. The missiles North Korea has displayed and deployed, have almost nothing in common with any of Pakistan’s apparently Chinese-sourced missiles. North Korea’s biggest solid-fuel missile is the KN-02, which appears to be a direct copy of a Russian, not Chinese, short-range weapon. If Pakistan, or Iran, have been quietly passing Chinese technology to North Korea, North Korea does not appear to have been doing anything with it.

      And even if the North Koreans were to make use of Chinese missile designs tested in Pakistan, they would still have to test the things themselves. A Shaheen missile built in North Korea is not the same as a Shaheen missile built in Pakistan, even if they are built to the exact same blueprints and even if there are Pakistani experts looking over North Korean shoulders as the thing is being built.

      As I alluded to earlier, there are as many opportunities to screw up and fail completely when turning blueprints into a production line, as there are in turning a concept into a set of blueprints in the first place. No matter how many times the Pakistanis tested their version, the first one North Korea builds will probably be a dud, and they will have to test it to figure out why.

      This, they have not done. Instead, they have shown the world Musudan missiles and KN-08 missiles, which nobody has tested anywhere (possibly excepting an unconfirmed report of a single possible Musudan test in Iran). Either the latest Nork missiles are part of a prolonged development process that hasn’t produced an operational system, or they are the hoax, or they are the result of gross technological optimism by the North Koreans that will probably have them shooting blanks if they wind up in a real war.

      The only long-range missile North Korea has successfully tested is the relatively simple, Scud-derived Nodong, which seems to work tolerably well. The recent flight of the Unha-3 space launch vehicle might at least be counted as a successful technology demonstration, though the Unha-3 itself is not a missile. Even so, that comes to 3-6 successful tests in the past fifteen years for North Korea.

      France, for the record, has conducted ten successful long-range missile tests in the same period. The United States, over a hundred. Nations which are serious about maintaining a reliable strategic missile force, test their weapons quite frequently to verify that the production line and operational processes are still up to the job. These launches are so routine that only their failure or cancellation is newsworthy. So when it still makes the news every time North Korea (or any other country) does test a missile, that should be your indication that – no matter how many foreign partners they have feeding them data – they have at best a marginal domestic capability in the area. And what they haven’t tested at all, very likely won’t work at all.

  5. Magpie (History)

    But does it matter?

    I mean, the Nodong can already hit enough of This Our Planet Earth to give them a solid threat. What new, interesting targets could a Musudan get anyway? Guam? Pfff. I mean, once you can nuke Osaka, you’re pretty much done, right? Tokyo is just icing. A Nodong can almost certainly throw 1000kg at Osaka right now, and might even be in range of Tokyo itself.

    What’s a Musudan FOR, then?

    Kim Mag-Pi has totally reversed his position. Less Musudan, more babes.

  6. panterazero (History)

    Offthread, but tangential: Just sayin’ that Krepon is right, that the comments on “DIAC has a Scud!” have to be one of the geekiest/most insider comment threads I’ve ever seen here.

  7. panterazero (History)

    @John Schilling: Just catching your May 23 comment “the Unha probably did use R-27 technology…” which is freaking brilliant. Kudos.

  8. Arun Vishwanathan (History)

    In light of the above discussion, maybe a report some of my colleagues at the International Strategic and Security Studies Programme (ISSSP) at NIAS, Bangalore wrote on the Unha-3 might be interesting. Link to the full report below.

    Arun Vishwanathan

    • j_kies (History)

      Thanks for the link – read the report and I have a number of issues
      1) The report expects the stages to have very different total burn times than were claimed by North Korea (I recall reading 9:27 or something close in the NK reporting)
      2) The third stage as best reported is not the short burn solid from the TD-1 flight test but may be extremely similar to the Safir second stage
      3) The report does not address the plane change that the NK satellite launch of 12/12/2012 accomplished using the third stage (likely without significant coast or re-ignition capability)

      Other issues have to do with the query of which rocket engines were applied by the NK team to achieve the stated thrust levels.

      Did they revise their model with these new insights and constraints?