Bharath Gopalaswamy and Martin Senn at the great blog “C x I” have posted a paper by Robert Schmucker and Markus Schiller titled The DPRK Missile Show – A Comedy in (Currently) Eight Acts.
They elegantly lay out the case that North Korea’s ballistic missile program remains dependent on foreign, likely Russian, assistance:
There is a different explanation that is much simpler – a connection to Russian institutions. All of the North Korean missiles were procured from Russia or at least realized with foreign support. Some, as Scud B, might come from old stocks, single remainders of old Soviet prototypes certainly were among them, and others might still be in production. A guided North Korean licensed production of simpler components can also not be excluded. In any case, the indigenous contributions of the DPRK are small at best. It is not said, though, that the Russian government or the leadership of the institutions in question know of this: Much happens in dark alleys, as was illus-trated by the example of the Gharbiya gyros for Iraq.
The DPRK will of course try to reverse engineer parts and components, and it will try to acquire the capabilities for indigenous development and production. Due to this, single engine tests should be observable, not only to demonstrate indigenous activi-ties, but also to learn and to slowly increase the DPRK’s competence on the missile sector.
But in the public opinion, this explanation is wrong, because – well, because it cannot be right. Because there is a well established view of North Korea that is also con-firmed by defectors: The rockets are secretly designed, tested and produced in huge underground facilities, and these efforts are directed by an evil and megalomaniac villain who threatens the free world with his missiles.
How to best counter this type of threat should be known from the movies – just call James Bond.
This has been a long-standing position of Schmucker, which was initially met with skpeticism in the US intelligence community. Here is how Bradley Graham described it in Hit to Kill:
Much of the speculation [about the 1998 Taepodong launch] centered on Chinese assistance with the satellite. Some analysts figured tat Russia had a hand in helping with the booster and staging. One of the more radical theories along this line, put forward by Robert Schmucker, a German professor specializing in aeronautics who had served as a United Nations inspector in Iraq, held that the North Korean missiles were not really North Korean at all. They were Russian, secretly build with Russian components and the active and ongoing help of some errant Russian scientists inside North Korea.
Schmucker argued that a rogue team of Russian missile scientists — thrown out of work atfter the collapse of the Soviet Union — may have moved to North Korea. And there, for profit or glory or both, they directed the North Korean program, with the North Koreans themselves doing little more than putting the pieces together. Schmucker’s evidence for this was that North Korea had performed few if any important missile tasks independently. Its missile assembly lines were built with Russian help, and designs for the Scud C and Nodong were derived from Soviet missile programs in the 1960s. In Schmucker’s view, Russia was using North Korea to hide the origin of the seller and get new customers. This theory was regarded skeptically by US government analysts (“Don’t shortchange North Korea’s indigenous capabilities,” one senior CIA official said), but it did pique the interest of some American specialists outside the administration.
I have to admit that I have gone from skeptic to “piqued interest” as I have been developing a small effort on DPRK missiles. I will be eager to hear what many of you have to say.
Perhaps it’s just the translation, but Schmucker and Schiller, in their paper, often seem to resort to sarcasm in place of more intellectually satisfying forms of argument. They needn’t bother, as their argument is catching on. Daniel Pinkston’s history of the DPRK missile program cites it approvingly. So does the technical appendix of the East-West Institute’s Iran missile threat assessment, by Prof. Ted Postol.
(Myself, I would like so see some footnotes.)
Their conclusion: “All of the North Korean missiles were procured from Russia or at least realized with foreign support.”
The first argument (“procured from Russia”) strikes me as clearly erroneous. The alternative (“realized with foreign support”) strikes me as correct, although the exact extent is unclear.
There’s little doubt that North Korea has a ballistic missile industry, and does not simply receive and re-ship missiles from abroad.
First, if North Korea can ship a complete Scud factory to Libya, accompanied by documentation and various personal effects in Korean writing — not in Russian — that means they can build complete Scud factories. (If the production lines had come from Russia, not North Korea, also they probably wouldn’t have incorporated so much Japanese technology.) North Korea also set up Nodong factories in Iran and Pakistan. We’ve seen video evidence for the Iranian factory right here on this blog. A.Q. Khan, Pervez Musharraf, and the late Binazir Bhutto, who collectively didn’t agree on much, all have described the technology transfer to Pakistan.
Second, if many hundreds or a few thousands of missiles had arrived in North Korea from Russia in the last decade or two, I think this fact would have been detected and come to public attention by now.
Third, there is a missile factory in or around Pyongyang; a Burmese military delegation visited there in 2008, according to the trip report, which was acquired (with photographs!) by a Burmese opposition group shortly thereafter. For more on North Korea’s missile infrastructure, see NTI’s backgrounder.
That may not satisfy everyone, but when it comes to North Korea, this is about as strong as the evidence ever gets.
Incidentally, this is also the view of the U.S. intelligence community, which reports that North Korea “continues to procure needed raw materials and components from various foreign sources to support its missile industry.”
So what can we salvage of Schmucker’s and Schiller’s argument? Actually, a fair bit, which suggests it has been useful to make the argument. There are good reasons to think that Russian experts were involved in North Korean missile program in the early 1990s, at a minimum.
First, there is the Makeyev incident they mention. Daniel Sneider and others have written about this incident.
Second, the sketch of a Nodong engine in a Russian rocketry textbook from the early 2000s suggests that one of the authors (or perhaps a close colleague or two) may have been involved in the development program. For comparison, see the account of Russian rocket scientists in Iran in the 1990s, described in The Dead Hand.
(Schmucker takes the textbook illustration as evidence that the Nodong secretly existed in Russia in the 1950s, but doesn’t explain why he thinks this.)
Third, there is the close resemblance between North Korean missiles and their Russian predecessors, although I’m not as confident as Schmucker and Schiller that the North Koreans have made no innovations of their own along the way.
It may be an interesting addition to point out that those “US government analysts” so skeptical of Schmucker’s theories were working for the Clinton administration (Bradley Graham’s ‘Hit to Kill’ was first published in 2001, right?).
Is it possible that said Clinton administration (1993-2001) had a strong incentive to see the north korean ‘missile program’ being misinterpreted as completely and strictly indigenous, as illogical as that may sound to anybody with a technical background?
There are already a couple of comments posted at CxI:
http://www.c-x-i.eu/?p=242#comments
Schmucker makes one major mistake, he completely ignores politics.
He should finally realize that North Korea is not Germany, where engineers are in control of technical developments and where reason rules. In North Korea the leader’s opinion is supreme law. If the dear leader wants a completely new missile he gets one – period. Their currency reform was pure lunacy, every mediochre economy student could have told you. Still it was done and caused devastation. So why should the North Korean regime act in a professional manner when it comes to missiles?
It reminds me of one TV interview Schmucker gave with German television before the Iraq war. He claimed the Al-Samoud was pretty much useless without a WMD warhead. Again he proofed to have a lot of technical knowledge but totally ignored politics. The point was that Saddam wanted to have a new missiles for reasons of prestige whether the whole effort and cost was justified for its questionable military value did not matter.
“But instead of a renewed launch attempt with an improved third stage – the logical
path that every engineer on this planet would go –, the program was cancelled and
the Taepodong 1 was never seen again.”
Correct “every engineer on this planet”. The problem is in North Korea engineers are not the ones responsible for making decisions.
There are other statements I do not agree with:
“It is also strange that Russia silently watches the DPRK cloning and selling Soviet
products, thus earning hundreds of millions of dollars, and doing this without any financial
compensation for the Russians.”
This isn’t strange at all, North Korea has been copying Russian tanks and artillery systems for years without any Russian complaints. China’s copies of pretty much all Russian 1950s and 1960s weapon systems (some of which are still exported like the F-7 MiG-21 derivative) did not spark a lot of Russian complaints either.
“There is not one single proven example for
successfully reverse engineered missiles and rockets”
How about the Ghadr?
By the way, are there any of the photos of the Scuds sent to Yemen available?
Is the Cyrillic lettering on the first Shahab 3s confirmed?
Josh:
We should make a distinction between ‘indigenously developing and producing missiles’ (or ‘reverse engineering’, which simply does not work out the way most people assume) and the existence of ‘missile factories’.
The ~11m long Scud-B, for example, will certainly be disassembled for discrete transport (if you take a closer look at an R-17, then you’ll notice that the main components engine compartment, IRFNA-tank, Kerosene-tank and guidance compartment are attached to each other by easily replaced rivets – specifically for the purpose of disassembly/maintenance; the warhead is dismountable anyways) and later reassembled in what somebody could describe as ‘Scud factory’ (although ‘repair shop’ may be a more fitting description). So, even the existence of a ‘missile industry’ or the help to other countries in establishing (or rather ‘cobble together from what is available’) their own ‘missile factories’ does not at all count against a simple ‘receive and re-ship’ modus operandi. This also devaluates what you wrote about the detection of missile transfer activities from Russia to the DPRK – none of the Scud-B’s main compartments, boxed in unsuspicious packing cases, would have excessive dimensions for rail or ship transport (absolutely unlike the SS-4/Sandal/R-12 and SS-5/Skean/R-14 missiles the Russians tried to secretely ship to Cuba) and thus would be rather difficult to detect for any outsider (e.g. by satellite surveillance). Can you say for certain what kind of ‘goods’ exactly cross the north korean/russian border via rail or ship each day?
And we should not forget that there are simpler and less simple tasks in producing a missile. Assembling the airframe/tank structures from appropriate sheet metal will certainly be within the range of what the North Koreans (or Iranians or Pakistanis etc.) can manage, especially with explicit instruction in this regard by the original manufacturer.
But guidance systems and rocket engines are typically highly complicated (to design and produce), and thus have to be considered beyond the scope of third world (and even many second world) countries. So, most likely (especially in case of larger missiles difficult to smuggle even in disassembled state), those highly complicated, but due to their compactness easily concealable, components will be procured from abroad, while the rest of the missile is locally manufactured from appropriate materials (which also are procured from abroad) by native personnel (trained abroad) on dedicated equipment (also procured from abroad, although a lot of neccessary machinery will be of a ‘general purpose’ type and thus is easily replaced by off-the-shelf items), most likely supervised by personnel from the original manufacturer.
Nonetheless, the resulting missiles will have to be subject to lot acceptance testing (at least one test shot for ~50-100 missiles produced, higher rates for initial batches) – otherwise the resulting missiles will be of doubtful reliability/functionality at best.
Last but not least, there are several technical details (like the general make-up of the rocket engine, or the comparably large volume of the guidance compartment) in case of the Nodong indicating that this missile
a) was designed by the same team(s) (aka Makeyev/Isayev) which also designed the R-17/SS-1c/Scud-B and
b) is a predecessor of the Scud-B – aka must have existed, at least as an elaborated concept, in Russia in the 50ies.
Jochen:
Here’s another hypothesis: some Russian experts taught the North Koreans how to make the Scud family of missiles. And maybe even helped them create the scaled-up version of the Scud-B that we call the Nodong. But however exactly this came about, the North Koreans are capable of making the missiles by themselves. Some components and materials, e.g., maraging steel, they appear to continue to acquire through the black market, along with machine tools and the like. But it does not all come in neatly packed crates from a hidden factory somewhere in Russia.
I think that fits much better with the presence of Japanese-origin materials in a missile shipment from North Korea than your suggestion does. It also fits better with the export of Nodong manufacturing technology from North Korea to Iran and Pakistan.
What is more, at the risk of appealing to authority, I’ll point out that the U.S. IC, whose judgment is cited in my comment above, probably has a very good idea about how and where North Korean missiles are made. Consider this vignette from an profile of the National Air and Space Intelligence Center (NASIC) that appeared a couple of years ago in the Dayton (Ohio) Business Journal:
For more on NASIC, go here.
Now, maybe all this is a big bluff; I don’t have any way of knowing for sure. But given that MDA’s public briefing materials hint rather broadly that Scuds are used as target missiles, and that it’s a matter of public record that Libya donated its North Korean-origin Scud-C force to the United States, I tend to believe that the IC would know the answers in this instance.
I remain mystified by the conviction with which you hold your beliefs on this issue. Let’s leave it at that.
Fabian Hinz:
“If the dear leader wants a completely new missile he gets one – period.”
Well, as you noticed yourself, the recent north korean currency reform for example simply didn’t result in something expedient – how can you expect such an approach to work on anything technical, not to mention ‘rocket science’? And yet, ‘north korean’ Scud-Bs and -Cs, and by now also the Nodong-A, work perfectly (aka exactly like russian-produced originals)!
Let’s face it, politicians may be the decision-makers, but if they overrule reason (happens too often anyway!), the only result to realistically expect is total catastrophy!
Additionally, politicians have their motives, too. To impute wasting a lot of scarce resources on non-working single-item multi-stage-missile prototypes to north korean leaders because of sheer madness or narcissm neglects this correlation in a rather arrogant manner. The North Koreans want to achieve something with this, and Schmucker/Schiller offer a plausible explanation: They desperately want to show potential customers that they still have some ace in the hat. Doesn’t matter if all this is only some sort of Potemkin village.
BTW, seems to have worked with the Iranians – they bought the Eunha’s 4-engine-cluster for their Simorgh (and there’s also the possibility that Simorgh is more or less identical with the ‘Taepodong-2’ that was tested in 2006).
Considering the issue of politics, we also should not ignore that maybe it’s our own politicians who prefer us mere mortals to have an, let’s say ‘altered’ view of the world. Sometimes technicians have a less obstructed, more product-related point of view, less cluttered by ‘political correctness’. And sometimes it takes an independent thinker to point out the 800lb gorilla in the room.
“He claimed the Al-Samoud was pretty much useless without a WMD warhead.”
Well, you can say a lot about that particular missile, but not that it was not pretty much useless without a WMD warhead. BTW, may i call your attention to the multi-chamber spin-off-projects (see e.g. Comprehensive Report of the Special Advisor to the DCI on Iraq’s WMD, Volume II – Delivery Systems, p.33) or the fact that Al-Samoud II would have offered the opportunity to use that missile in a modular way to produce larger, longer ranged types with heavier payload in the same way Iraq demonstrated with Scud and Al-Abid several years earlier?
“North Korea has been copying Russian tanks and artillery systems for years without any Russian complaints”
Now, doesn’t this strike you as suspicious? Either the North Koreans are an ‘underground’ universal industrial giant – or those were likewise quietly supplied by the Russians!
“China’s copies of pretty much all Russian 1950s and 1960s weapon systems (some of which are still exported like the F-7 MiG-21 derivative) did not spark a lot of Russian complaints either.”
Then it’s strange how the Russians react to chinese copies of e.g. their Su-27SK, Su-33 or the S-300PMU missile system in recent years. The answer is, as i believe, apparent: The Russians don’t protest because in the other, unprotested cases, there are silent agreements about off-the-record license production, most of them dating back to the days of the Soviet Union. ‘Compensation’ of some sort may also play a decisive role.
“How about the Ghadr?”
Which Ghadr do you mean, the Ghadr-1 (which is a nothing but a modified, not reverse-engineered, Shahab-3/Nodong-A) or the Ghadr-110/Ashura (which is a two-stage solid-fueled design without any RevEng-antetype i’d know of)? Or do you mean the GBU-15-clone (which is a non-propelled electro-optically guided aircraft-glidebomb, not a missile)?
“are there any of the photos of the Scuds sent to Yemen available?”
Check e.g. this out:
!http://www.clker.com/cliparts/5/a/a/0/1257500378199193027bprjlr-md.png!
Unfortunately, the few pictures still available on the ‘net on the So-San-incident are of rather low quality. Schmucker and Schiller have better ones at their hands, though.
“Is the Cyrillic lettering on the first Shahab 3s confirmed?”
I must admit that this is beyond my knowledge at the moment. But as i know Schmucker and Schiller, they wouldn’t write this unless they can back that claim up with photographic evidence.
Regarding the Cyrillic lettering on the first Shahab 3 – as claimed by the authors – I saw their presentation from 2010 – they show a picture of Shahab 3 in its original color scheme of 1998, AND some close up pictures with Cyrillic letters. HOWEVER, these letters could be from a Scud missile. (AKA Shahab 1 or 2).
See:
http://www.security-research.at/bmd/wp-content/schiller_17022010.pdf, page 52.
Thanks, Tal Inbar!
Now i remember – Schmucker once showed me these pictures.
They show indeed a Shahab-3, namely the one paraded/exhibited at the 22.September-parade in 2001 on the single existing 4-axle-MEL.
For another, widely known photo of that particular missile/MEL, see e.g.:
!http://www.armsky.com/yuanchuangzhuangao/UploadFiles_4662/200452219529734.jpg?IM_WIDTH_MAX=710&IM_HEIGHT_MAX=1000!
Note the distinctive mottled camouflage of that missile, which i only ever have seen on Shahab-3s, never any other iranian missile, and the unmistakable ‘zebra’-style paintjob of the MEL.
Josh:
First of all:
“I remain mystified by the conviction with which you hold your beliefs on this issue.”
Dito.
Unless you have any evidence clearly backing up your point of view (‘literature mention’ or ‘higher authority’ does not count – in contrast, e.g. ‘photographic evidence’ of cyrillic markings on ‘north korean produced’ Scuds and Nodongs or ‘retraceable technical similarities’ do count!), absolutely irreproducible to me.
Besides, don’t you realize that what you write and what i write is not that terribly different – only our corresponding interpretation?
In essence, this whole discussion circles around what exactly these ‘missile factories’ do:
design and manufacture missiles, including all of their components, even the most intricate ones, from unspecific raw materials,
or
re-assemble missiles manufactured elsewhere after disassembly for transport/smuggle.
Of course, the truth may lie somewhere inbetween, as there are several possible intermediate levels between those two extremes, more on this below.
“Some components and materials, e.g., maraging steel, they appear to continue to acquire through the black market, along with machine tools and the like. But it does not all come in neatly packed crates from a hidden factory somewhere in Russia.”
How else would you expect “some components” (like e.g. the missile’s guidance gyros, or a pre-produced rocket engine) to come in other than in “neatly packed crates”? Where else would you expect definitely russian-designed equipment to come from except Russia?
As i wrote before, airframe/tank structures are fairly simple (in fact i’d expect every better-equipped and -staffed backyard repair shop to be able to reproduce something like this according to given specifications from raw materials of appropriate quality, especially if an explicit construction manual and personnel training on this issue has been provided), as are conventional HE-warheads – all other aspects, beyond the scope of those countries, will have to be dealt with by means of procurement.
Nonetheless, missiles manufactured in this way will have to be subject to lot acceptance testing (at least one test shot for ~50-100 missiles produced, higher rates [up to 100% of early-production missiles] for initial batches) – and we simply don’t see anything like this happen in the DPRK.
No testing, but nominal performance (of the original)/high reliability/high confidence in the ‘manufactured’ missiles – no indigenous production (not to speak of ‘reverse engineering’!), simple as that.
“I think that fits much better with the presence of Japanese-origin materials in a missile shipment from North Korea than your suggestion does.”
If you re-read my above comment, then you’ll notice that i wrote “procured from abroad” in respect to materials or production equipment (where i explicitly alluded to ‘off-the-shelf-items’ in case of general-purpose machinery). That not neccessarily excludes procurement from other sources than the original manufacturer. But the more things you change in the original manufacturing process, the more unpredictable/unreliable the outcome will be – thus again the issue of extensive testing – something we simply don’t see in case of the DPRK! (Gee, didn’t i write that somewhere before?)
Last but not least, considering NASIC: Those will be subject to politics, political correctness and ‘issues of erring on the right side’ to cover their tracks. They also always have the option to simply not publicize their real findings, for ‘matters of national security’. On the issue of politics – see my answer to ‘Fabian Hinz’. Maybe it’s our own politicians who don’t want us to know the truth for some reason.
Thank you everybody for picking up our line of thoughts for discussion. We take every argument as a serious contribution that we factor into our theories.
There is indeed a bit of sarcasm to be found in our essay, but this is merely a result of working on this issue for too many years. To be clear: We truly understand that other people have other views on this matter, we totally respect them, and we mean no offense in any way. Actually, even though Robert had done rocket development programs and reverse engineering projects himself, he still held the common beliefs about third world reverse engineering when he joined the UN inspections in Iraq in the nineties. It was a long way for us to arrive at our current beliefs, and we still challenge our own theories almost every day. But currently, our laid out perspective seems to be the only consistent (!) explanation. For our taste, there are too many inconsistencies in the commonly accepted views about third world missiles.
Several aspects were addressed already in the discussion here, with good arguments on both sides. We will only comment on a few of them, but, as said, we take all of them into consideration. Thank you for the food for thinking.
The common view of Reverse Engineering (RE) can definitely be ruled out by now. We fully understand, though, that it will take a lot of arguments, discussions and additional proofs to convince the community of this. We are working on this.
However, to make a point, think of the several reports of “poorly manufactured parts” or parts of “poor quality” that were produced in Iran or North Korea. Even Russian sources confirm this. Or better, think of the RD-180 production (the Atlas V engine) that was to be transitioned to the U.S. – more than ten years later, they are still produced in Russia (actually, this has more to do with licensed production because of Russian assistance – at RE, no assistance is given!). Or think of this: Are the TELs that can be observed in NK also reverse engineered? They are definitely not the Chinese version, but look exactly like the Russian MAZ in every detail. If NK is such a genius in RE, why not copy refrigerators, cars, helicopters, TVs, computers, cell phones or milling machines and start earning some money on the global market?
So, if we assume that RE is to be ruled out, the missiles are either from licensed production or they were produced elsewhere (not necessarily at present days). If they are produced with license, why are there no lot acceptance tests, no firing table tests, no quality assurance tests? Because they were ordered not to test their products?
This inevitably leads to the mentioned I.C. issue. The I.C. surely does great work, but we should never underestimate the problem of both acquisition of information and evaluation of it. As open source, we cannot know what the I.C. really knows, or even what it really assumes. They may think they know the truth, and they may actually know the truth. But remember the familiar cases in which the I.C. was obviously wrong in its published assessments. And never forget that politicians and I.C. have to assume worst case scenarios in order to avoid unpleasant surprises.
We do not know the truth, but for us, the only consistent explanation currently is the theory that we laid out in our essay.
Please always keep in mind that, as Jochen Schischka pointed out, political issues cannot overrule shortcomings of technical, physical or other nature. Else, we should already know the football world champion of 2010: North Korea, of course, because the dear leader wants (sorry – we could not resist to insert this sentence!).
Jochen, let’s review:
You think that the only possible place that Russian-designed missiles could come from is Russia. Is that a fair representation?
This would imply that every Scud or Scud derivative everywhere was actually built in Russia, including the ones that we naively suppose to have been made in North Korea, Iran, Pakistan, Egypt, Syria, and perhaps elsewhere. While all of this goes on in competition with Russia’s official missile export industry, it is tolerated or even quietly encouraged. But the rest of us have somehow never caught on. Correct?
This also implies that non-Russians are not in a position to master Russian technology from the 1950s, let alone transfer it to others themselves. Yes?
It seems to me that there are too many cases of countries transferring missile technology to other countries, or even informal networks of experts transferring complex precision-engineered technologies (notably gas centrifuges) to be forced to resort to such a perspective.
Just consider what the purpose of the journey of the Makeyev design bureau experts to Pyongyang would have been: to instruct North Koreans in how to make a new missile type. Otherwise, it’s hard to explain why so many people from a design bureau would have been headed over there, isn’t it?
Similarly, the presence of Cyrillic letters on an Iranian missile is intriguing, but should not come as a huge shock. The training of Iranian missile engineers in Russia and the assistance of Russian experts (and Chinese experts, and North Korean experts) in the Iranian missile program has been reported for years, and seems like an adequate explanation.
Robert and Markus:
Thanks for putting on an appearance.
I feel you are presenting us with a false choice when you write:
I don’t claim enough expertise to know whether reverse-engineering is truly out of the question. But let’s just say you’re right about that point. Does that really mean that the only possibilities are licensed production or foreign production?
Perhaps it depends on what you mean by “licensed production,” but it seems to me that you are omitting a third and very important possibility: that the North Koreans learned how to build Scud and Scud-derivative missiles not simply in collaboration with Egypt, but also with help from other foreign experts, most likely from the USSR, either with or without state sanction.
(Or is that indeed what you mean by “licensed production”?)
I do not understand why you do not seem prepared to give this idea due consideration. Granted, it is not so exciting or counter-intuitive as your preferred idea, but it is there. Certainly, it fits the general pattern of technology transfers mentioned above, and does a better job of explaining why North Korea has acquired components and raw materials from Japanese sources in the past. So I take it pretty seriously.
It seems to me that you have made too much of North Korea’s unusually abbreviated test record. They’ve had their share of failures, after all. Nor should we assume that our understanding of the test record is complete.
In any case, assuming that this is not some kind of semantic issue, I would certainly like to know why you seem to give no consideration to the hypothesis that North Korea acquired the technology with assistance from Russian experts.
Josh:
For us, reverse engineering is definitely out of the question. We understand that this hypothesis needs sufficient justification, but for now, let us assume this is true. We plan to publish something on reverse engineering soon, and hopefully, this will be the extensive rationale that we should give here but cannot give here. Thank you for not making RE the subject of this discussion, at least for now.
Regarding the issue of licensed production: You are right, this is exactly what we meant – learning by doing, instructed by the original equipment manufacturer. Sorry, we might still trip over some semantic issues.
To be honest, we considered this option in great detail. And again, some inconsistencies surfaced.
What we call licensed production is a very common approach. The original manufacturer transfers know how, experts, documents, procedures and/or much more to another institution which ultimately gains the capability to produce the desired object (several types and levels of licensed production are possible, but let us not go into the details). In turn, license fee is paid.
The thing is: We know of a large number of well documented transfers of production. And, at least in the high tech sector, none of them went as anyone would imagine.
Think of the F-104 Starfighter production in Europe, for example, of the immense manpower that was required, of the problems, the timeframes and the results. Think of car factories in other countries and their problems (interesting enough, as far as we know from German car companies, all important parts – even including metal sheets – are produced in one place and shipped abroad for integration). We know of stories about Siemens that had severe problems in welding their streetcars – not in a 3rd World country, but in the USA. And, as already mentioned, Pratt & Whitney Rocketdyne still has problems producing the Russian RD-180 engine in the USA, even with help from Energomash. Robert worked in a licensed production project, and it was a nightmare.
And once you get a production line running in a foreign country, you cannot expect that it produces with the same standards as the original line. It takes time and experience. You have to have excellent quality assurance, and, most important, you have to check the product for failures. In rocketry, this can be done only with flight tests, especially when you deliver your products to a country that is in war and launches your product in the field (Scud B Iran-Iraq). And every 10 or 20 or 50 rockets, you have to pick one out of production and see if everything still works fine, or if you produce crap because one single worker “optimized” something on his own, for example. Think of the continuous Trident launches, of the Topol launches and so on.
But let us forget the test flight argument for now. There are two even more interesting aspects.
Why would Russia, sitting on a significant stockpile of existent Scuds and other missiles (for sure during the 1980s/1990s), take the efforts to destroy these weapons and simultaneously create a new production line for them in North Korea? This makes no sense, with financial considerations not the least of all.
And, even if Scud B production was instructed to the Norks, and even if they managed to produce them in sufficient quality – how could they develop AND produce even more sophisticated systems (Scud C) and systems with more performance (Nodong) only by knowing how to produce Scud B? Can we soon expect Spartanburg, for example, to start producing cars more sophisticated than the BMW X5 or X6? And South Carolina is not North Korea! (Even though this comparison is flawed, of course, it should be clear where we are aiming.)
Of course, the North Koreans will have tried to become indigenous, producing parts and more, just like the Iraqis. Perhaps they even successfully produce simple parts and airframes. But it seems unlikely that they succeed where (all) others failed completely.
We hope that you can slowly share at least some of our skepticism about the established hypothesis of an impressive indigenous North Korean missile program.
Anyway, the main objective is to identify proliferation and prevent it, and this can be done best by knowing the truth. Let’s find out.
Josh:
Please re-read my previous comments on this thread. I explicitly did not rule out the possibility of partial production several times. But, as Schmucker/Schiller already adverted to in their last comment, without actual flight testing on a credible scale (i’d expect an initial series of 5-10 shots until the kinks are worked out of that production line, then ~1 for every 5 manufactured missiles for some time, then ~1/10 and so on until that production-line can be considered mature – and even then, a sustained rate of ~1/50-100 is a minimum requirement for quality assurance), this has to be considered as suspicious. And don’t forget that the North Koreans reportedly have at least two different Scud-types (8K14/SS-1c/Scud-B and 9M77/SS-1d/Scud-C) ‘in production’ – which would also double the number of neccessary flight tests. As Schmucker/Schiller point out (did you take a look at the pdf-file Tal Inbar was so kind to put a link to on this thread?), there are at the moment only ~12-14 flight tests in the DPRK of Scud-type systems of any kind known – over a period of 26 years. This would be, under rather sanguine optimum conditions, only consistent with a maximum overall production of ~30 missiles. How many of the remaining ~16 missiles did the North Koreans export to Iran, Libya and Syria again?
I know exactly what you’re about to say – what if we didn’t detect all of those tests? Quite right, but i hope you understand that at estimated production numbers of many hundred Scuds (reportedly over 500 Scud-Cs alone!) of any type, a rather excessive (i’d even say incredible, as in ‘unlikely’) number of tests would have remained undetected. Also consider that these testing events apparently don’t follow any continuity (like ‘twice a year’ or something like that) or decline over time down to a stable minimum number, as would be expected in case of continuous production – they seem to happen in surges (remember 2006?). Then, strangely enough, there’s no test at all for many years. And if you say “well, then they simply don’t produce uninterrupted” – bad idea. In this case, they’d start at zero again and again and again – and even more testing activity should be expected. To rest is to rust.
Please consider: what if we did detect at least the majority of north korean test launches?
And i already know what you’ll criticize next – what if your testing numbers are flawed? Well, then take a look at open-source real-world examples like the russian R-1/SS-1a/Scunner (particularly interesting, since that would be a case more or less directly comparable to what the North Koreans allegedly did). Or ask anybody with operating experience in the current U.S. missile production. Or generally anybody with experience in license production of more complex technical devices (remember the teething troubles Spartanburg had with the BMW X5? Cars aren’t rocket science…). You don’t have to believe me. Or Schmucker. Or Schiller. Find out yourself.
Please consider: what if those minimum testing numbers are in a realistic range?
Now for some more direct answering to your last comment:
“You think that the only possible place that Russian-designed missiles could come from is Russia. Is that a fair representation?”
Yes and no. I don’t deny the possibility of license production. I’m highly skeptical about ‘reverse engineering’, though. And remember: NO TESTING ON A CREDIBLE SCALE – NO PRODUCTION.
“This would imply that every Scud or Scud derivative everywhere was actually built in Russia, including the ones that we naively suppose to have been made in North Korea, Iran, Pakistan, Egypt, Syria, and perhaps elsewhere.”
Yes, that sums it up pretty well. Although i’d exclude ‘Scud derivative’ to some degree. Consider e.g. what the Iraqis did with their Scud-Bs delivered from Russia – initially they cut up the tanks of three missiles to create two stretched S80/Al-Hussains. Later, they even achieved a 1:1 conversion rate with self-produced tank inserts. Remember what i wrote before about airframe/tank structures? Nonetheless, these modifications of the basic Scud-B arguably had a rather dubious track record.
“While all of this goes on in competition with Russia’s official missile export industry, it is tolerated or even quietly encouraged.”
Which ‘official missile export industry’ do you mean? Iskander-E? Well, that one is limited according to the MTCR-guidelines, and not in blatant breach of the INF-treaty. And is certainly not openly exported to precarious countries. The same can’t be said about the majority of missiles the DPRK ‘produces’ and exports. I can only guess about the “tolerated”-part, though. And “quietly encouraged” certainly is an exaggeration.
“This also implies that non-Russians are not in a position to master Russian technology from the 1950s, let alone transfer it to others themselves. Yes?”
No, not per se. But this presupposes a qualified receiver (the U.S. for example, as industrially capable as they are, obviously don’t fill that requirement in respect to russian rocket engines…and i find it somewhat hard to believe that a country which is a dark spot on night shots from space and whose population starves can single-handedly outdo an established industrial giant). And once again: NO TESTING ON A CREDIBLE SCALE – NO PRODUCTION.
“It seems to me that there are too many cases of countries transferring missile technology to other countries, or even informal networks of experts transferring complex precision-engineered technologies (notably gas centrifuges)…”
1.) Could you please name palpable examples “of countries transferring missile technology to other countries” other than DPRK-related? (Nah…at the moment i’m not in the mood to write something about flight testing again…)
2.) “Notably gas centrifuges” – those are thoroughly tested before being actually fed with UF6, aren’t they? What about experimental ‘pilot cascades’? Isn’t there some destructive lot acceptance testing, too?
“Otherwise, it’s hard to explain why so many people from a design bureau would have been headed over there, isn’t it?”
Don’t forget about all the other projects the North Koreans (and Iranians and Pakistanis and Syrians…) have been tinkering with (Paektusan-1/Taepodong-A, Paektusan-2?/Eunha-1?/2006-shot, Eunha-2/Taepodong-B?, ER-Scud, Nodong-B…). Nonetheless i think that you somewhat misinterpret the involved numbers. 64 (as in the 1992-incident at Moscow) hardly is a large number in this context. Heck, even ten times that would be a rather small number for full-fledged missile production. Also note that this ‘russian exodus’ only started after the DPRK apparently started ‘production’ of Scud-B and Scud-C.
“The training of Iranian missile engineers in Russia and the assistance of Russian experts (and Chinese experts, and North Korean experts) in the Iranian missile program has been reported for years, and seems like an adequate explanation.”
Then why do we see not a single chinese or north korean character on iranian missiles (or even on north korean ones)? And why are newer iranian missiles like the Safir IRILV labeled in english (just like most of their combat aircraft)??? British experts helping Iran…or what?
P.S.: Jeffrey, i apologize for the length of this post. I hope you understand that this is a concern of major importance to me.
Jochen, You repeatedly write :“NO TESTING ON A CREDIBLE SCALE – NO PRODUCTION.”
But WHAT IS a CREDIBLE scale? Are the standards acceptable in Germany (vis a vis Shmucker and Schiller) are the same that applies to Iran, NK or Russia?
Iran sees the Shahab 3 as operational – and the success rate is very good. Ir seems that in major drills they just fire them for fun…
We discussed the issue of testing somewhere in this blog a while ago – and the bottom lime was that a Shahab 3 or Ghauri launch is in fact re-validation and test of the Nodong.
One last comment – there are some cases when a country possesses ballistic missiles – sees them as operational and does not conduct tests – ore conducts even less tests than NK. I can think of two examples.
Tal Inbar:
Please (re?-)read my last comment (directed at Josh) on this thread. I explicitly explained what i’d expect to be a credible scale in case of missile production.
In this context, it is in the first instance irrelevant in which country this happens, be that Germany, Russia, North Korea, South Korea, Iran, USA, Israel, Burkina Faso or any other. These are absolute minimum requirements universally needed for quality assurance of serial produced highly complex technical items – and in case of missiles, the control sample unfortunately can’t be reused/sold after full end-to-end testing, like e.g. a TV set or an aircraft (especially in the latter case, consider that each manufactured specimen is thoroughly flight-tested before delivery to the customer). If these minimum numbers are not met (characteristically, these numbers are not lower, but even much higher for 2. or 3.world countries! Less qualified workers in a disadvantageous environment will also produce with less quality, this should be self-evident), then the typical outcome will be unreliable products (think e.g. of the notorious ‘friday afternoon car’ – without appropriate QA the overwhelming majority, if not all cars would be lemons!) with an at best erratic track record.
Consider for example that the US-american UGM-133 Trident II D5 has been flight-tested over 130 times up to date – at the beginning for development of the missile itself, then primarily in association with the establishment of series production and nowadays primarily for lot acceptance testing (during production), but also for troop training and customer evaluation (which will happen regularly/yearly until that system gets decommissioned).
Please note that this is about ten times the number the North Koreans launched over approximately the same period for the Scud-B and -C (plus perhaps also an additional Scud derivative)! And yet these missiles apparently perform flawless. Isn’t this strange?
??“Iran sees the Shahab 3 as operational – and the success rate is very good. Ir seems that in major drills they just fire them for fun…
We discussed the issue of testing somewhere in this blog a while ago – and the bottom lime was that a Shahab 3 or Ghauri launch is in fact re-validation and test of the Nodong.”??
Absolutely right (i’d assess the number of these shots to be approximately sufficient for minimal troop training and seller-evaluation purposes, but nothing else) – and that is exactly the point which makes licensed or unlicensed production (or even indigenous development!) of this particular missile system in North Korea, Iran or Pakistan highly dubious. Those missiles simply work too perfectly to be cobbled together rough-and-ready by somebody most likely (-> Scud production must be considered questionable, too!) without previous experience on that sector, without a broad industrial base or copious amounts of money to squander (i guess you have an idea how expensive all this is?) – and, particularly suspicious, not only without any credible production-related testing, but also no credible development-related testing activity to begin with!
“One last comment – there are some cases when a country possesses ballistic missiles – sees them as operational and does not conduct tests – ore conducts even less tests than NK. I can think of two examples.”
And i do not dispute this claim (although it would have been nice if you would have clearly named these examples for matters of transparency). But in those cases, these missile systems were simply procured from abroad, including all required vehicles, propellants (if neccessary), infrastructure, troop training etc., not indigenously manufactured.
A good example of this would be the Pershing-1A in west-german service. In this case, even all flight testing for troop training/evaluation purposes (every year from 1966-88, with great steadiness) was done from the McGregor Range in New Mexico.
Another good example would be Scud-B in east-german service. About 90 (in words: NINETY!) of those missiles were launched by the NVA (the east german army) at Kapustin Yar from 1964-88 – even without any indigenous production.
Or was it your intention to allude at cases in which those missile systems simply aren’t operational in reality (sort of ‘Potemkin village’ or, if you prefer that term, ‘breakout capability’ – i think you know what i mean…)?
To sum that up, at the moment the only consistent explanation, as Schmucker/Schiller point out (and i agree with them on this), seems to be that the North Koreans do not produce, not to mention design these missiles themselves – they’re apparently (in respect to their most frequently exported items) only the front office for missiles of obvious (at least to anybody familiar with that type of technology and an according technical background) soviet/russian heritage designed, manufactured and accordingly tested somewhere else, not neccessarily recently.
Draw your own conclusions.
Robert and Markus,
You raise some interesting points that I think are worthy of a careful reply. So in the interests of advancing the conversation, here are my thoughts, point by point.
But what if the seller has low standards, and the buyer has little choice? After all, do we really know what the performance record of Iranian Scud-Bs was ca. 1985-1988?
I think the right answer is, “for the money.” Whether it was a decision taken somewhere within the Soviet state or a rogue operation, one good reason to transfer missile technology to an ally is because they were willing to pay for it.
When you mention the size of the Soviet Scud stockpile and the destruction of that stockpile, are you referring to the INF Treaty? I’m unsure of how many they had ca. 1987, but note that the Red Army sent thousands of Scud missiles to Afghanistan at this time and shortly thereafter, and used them in battle, especially after the official withdrawal. And I’m pretty sure the Scud-B was not affected by INF in any case. Have I misunderstood your point?
Could some also have been sold to North Korea at that time, too? It’s possible, even though the USSR was backing Iraq against Iran, and probably would not have wanted to see weapons flowing to Iran. The further in time we proceed, as North Korea starts selling missiles that would have exceeded INF limits, the harder it is for me to believe they were made anywhere but North Korea.
Why should we assume that the Russians could only show up in North Korea once? Indeed, if we think they went there in the 1980s, then the attempt of a large fraction of the Makeyev bureau to travel to Pyongyang in the 1990s would point to a larger pattern.
Here it’s best to be cautious. Conditions were very different in Russia in the 1990s. It’s possible that state-sanctioned transfers in the 1980s gave way to freelancing in the desperate times of the early 1990s. (Certainly, the authorities did make efforts to stop the Makeyev group.)
But I increasingly do wonder if there wasn’t freelancing earlier, as well. Work on the Nodong would have started around the same time that the signing of INF would have put some folks out of business, come to think of it.
It’s all quite murky.
To be sure, if the Russians had a free hand to build missiles and ship them to North Korea all through the 1990s, I don’t see why so many Makeyev people would have tried to go there.
Here I wonder if you aren’t talking more about RE than technology transfer. If the Israelis could learn to build the Jericho-1 missile from the French, and the South Africans could learn to build an RSA-2 missile that walked and quacked just like an Israeli Jericho-2, well, why couldn’t the North Koreans integrate Soviet technology?
Contrary to some impressions, North Korea is not a third-world country, but rather a semi-industrialized state that has fallen on very hard times since the mid-1990s. Over the years, the NKs would have had opportunities to train people in the USSR. They even managed to build and operate the Yongbyon nuclear complex, which was no mean feat. They’ve even tested a couple of Pu implosion devices. So let’s not dismiss them too quickly.
PS. Jochen: Please calm down.
Josh:
Thank you for your careful reply. Let us see if we can sufficiently address at least the most important points.
You are right, we do not know the record of Iranian Scuds for sure. However, reports indicate that the launched missiles suffered no frequent or serious failures. It would be important to get reliable data about this. What we do know for sure, though, is that Scud-Bs that were launched at Iranian maneuvers in the 2000s have exactly the same launch acceleration, thrust and early trajectory as the nominal values of the Soviet R-17 – and by saying exactly, we mean exactly. To us, this does not look like low standard replica production. Of course, this might be other missiles than that launched in 1985 to 1988 – but if they are, where do they come from?
Regarding the most obvious reason for Russian support (if we might call it this), you probably are right again: “for the money”. And the best way to earn money would be not to export a production line, but the product itself to keep the customer dependent. This is not an argument, of course. But it sounds plausible.
As far as we know, the size of the Scud stockpile was not referred to in the INF treaty– R-17 was well below the INF limits. But we know that R-17 was continuously produced in at least two factories in the USSR until 1987. Assuming only one or two missiles per week from the beginning to keep the production running, the numbers quickly become impressive, even without the perhaps 2000 launches in Afghanistan and the high numbers of known exports. Anyway, the whole inventory of the Red Army at decommissioning can hardly disappear, and R-17 was not the least important tactical Soviet missile, certainly deployed and/or stored in significant numbers.
The point that you make right after the stockpile is important. We can assume that weapon traders are not thoroughly affected by idealistic thinking, even more when political interests of their governments are involved. It is not our intention to interpret political strategies, but from a weapon trade perspective, wouldn’t it be great to supply both sides of a war simultaneously? Since the USSR officially backed Iran, the only way would have been by using an intermediate – who officially, of course, has never received anything from the Soviets… and once INF was in place, the existing connections might have offered use for other deals, too. Just plain speculation, but it seems consistent with many other parts of the puzzle.
Perhaps these connections, once established and proven, then led to advanced assistance, for example modification of existing hardware or even revival of old programs, thus requiring increased exchange with Russian based institutions in the 1990s. It cannot be ruled out.
At this point, we should again try to make clear our approach. We never invented a certain hypothesis and tried to verify it over the years. Quite to the contrary, we used the approach of falsification. The common claims about the missile programs seemed more and more inconsistent to us the closer we looked at them. After thorough examination, the early reverse engineering hypothesis seemed completely implausible to us. The theory of strong foreign assistance slowly emerged, and currently it is the only consistent explanation that is left for us. Interestingly, single data points and information that appears over time seems to back it more and more, for example the use of the latest generation MAZ543 in North Korea, the Iranian Scud B performance identity and much more.
If there is any other plausible explanation, we are happy to accept it – our hypothesis might well be wrong. But as long as the reverse engineering and indigenous development can be falsified while our theory can not, it seems inevitable for us to see it as the best explanation.
This leads to your final argument. None of us has ever been to North Korea himself, but there are several reasons to think of North Korea as a 3rd World country: The well known satellite imagery of NK at night, impressions one gets from GoogleEarth and other satellite imagery, media reports about the industrial state of NK and the quality of manufactured goods, the estimated GDP (independently estimated by several institutions), and much more. We cannot say much about the NK nuclear program, but as far as we know, Russian and/or Chinese support was involved at Yongbyon. And who can tell for sure if the true situation regarding their nuclear weapon program is much different than that of their “indigenous” missile program? What achievements have we actually seen? But this is a different issue.
(By the way, risking a heated debate about rocketry and nuclear science: Until now, it took every nation considerably less resources, time and efforts to create a nuclear device than it took them to create an operational missile).
Jeffrey:
Please excuse the length of the post – if you prefer and Josh agrees, we can continue our discussion via email.
From what I have been reading the United States should not be buying the RD-180 let alone build them or use them.
Robert and Markus:
If falsification is your standard for the story of North Korea’s missile program as Joseph Bermudez relates it, then I would encourage you to test your own hypothesis by this standard as well.
Some of the evidence that tends to falsify the idea that North Korea’s missiles are made elsewhere by others:
– Evidence for the existence of North Korea’s own missile infrastructure and experts.
– Evidence that North Korea has transferred not just missiles but missile production technology to other countries.
– Evidence that North Korea has sourced missile components and materials from multiple countries, but Japan in particular. (The latest tidbit appears to be the sale of large construction vehicles whose chassis is suitable for use in an IRBM or ICBM TEL.)
In general, North Korea’s missile-related exports have followed a pattern of starting with complete, turnkey missile systems, then moving onto to the transfer of entire production lines, followed by supply of parts and materials. In some cases, the relationship seems to stop after the export of complete systems. In other cases, it progresses through the entire cycle. I hope to publish an article reconstructing this pattern in the next few months.
Josh:
Of course, falsification must be applied also to our considerations. But we should always be careful with claimed “evidence”. If a construction vehicle is suitable for use as a TEL, for example, this does not imply that it this really was its intended use. Always remember the “evidence” for mobile biolabs in Iraq. What do we think that we know and what do we really know?
North Korea is said to have developed (and produces, and offers for export) Scud B, Scud B-PIP, Scud C, Scud D, Nodong, Taepodong 1, Taepodong 2 (Unha-2?) and BM-25 predominantly on its own, with a total of about 10 launches in 23 years (1983 to 2006), which is less than any other nation needs for one single operational missile program. In our eyes, this is the central mystery that has to be solved.
We look forward to your mentioned article!