July 1, 2025 • 39 mins
In this episode of Horns of a Dilemma, co-hosts Ryan Vest and Sheena Chestnut Greitens interview scholars Nicholas Anderson and Daryl Press about their article, "Lost Seoul? Assessing Pyongyang’s Other Deterrent," featured in Volume 8, Issue 3 of the Texas National Security Review. Anderson and Press challenge the conventional wisdom that North Korean artillery could devastate Seoul during a conflict. They discuss the origins of their research, the methodology behind their military campaign analysis, and the important implications for policymakers, including the unexpected resilience of urban areas and the crucial steps South Korea has taken to mitigate potential damage. Tune in for an insightful discussion that reevaluates the artillery threat on the Korean Peninsula and explores broader security implications.
Read the article: https://tnsr.org/2025/06/lost-seoul-assessing-pyongyangs-other-deterrent/
For additional information on the model, underlying data, and other supplementary materials, please visit the online appendix at https://doi.org/10.7910/DVN/86HBGS.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Ryan Vest (00:05):
Welcome to Horns of a Dilemma, the podcast of the Texas
National Security Review at theUniversity of Texas at Austin.
I'm Ryan Vest, Executive Editorof TNSR, and I'm here with our
Editor-in-Chief, Sheena Chestnut Greitens.
We're pleased to have joining ustoday, two distinguished scholars in
security studies, Nicholas Anderson andDaryl Press, authors of "Lost Seoul?
(00:26):
Assessing Pyongyang’s OtherDeterrent," which is featured in
Volume 8, Issue 3 of the journal.
Nick is an assistant professor ofPolitical Science and International
Affairs at the George WashingtonUniversity, and Daryl is Professor
of Government and faculty Directorof the Davidson Institute for
Global Security at Dartmouth.
Gentlemen, welcome to Horns of a Dilemma.
It's great to have you on the show.
Nick Anderson (00:46):
Thanks for having us.
Daryl Press (00:47):
Thank you.
Ryan Vest (00:47):
Thank you.
Sheena Chestnut Greitens (00:48):
To get us started, I wonder if you could
just tell us a little bit about yourmotivation for writing the article.
There's a widespread assumption for aslong as I've been going to South Korea,
that North Korean artillery capabilitycould be devastating for Seoul at any
point should a conflict break out on theKorean Peninsula, and your piece pretty
directly challenges that narrative.
(01:09):
But I'm curious how you cameto do research on this topic.
What motivated you to even startasking questions about whether
that conventional wisdom is true?
Could you tell us about that?
Nick Anderson (01:19):
Sure.
I'll start with my half of it, becausewe actually started separately.
This started as a term paper in BarryPosen's legendary US military power
class, which I wasn't a student at MITbut I audited the course when I was a
fellow at the Belfer Center at Harvard.
And I had the same sort of assumptionthat you did, Sheena, going in that the
North Korean artillery is this sort ofconventional Trump card, very difficult
(01:41):
to deal with if you're South Korea orif you're the US forces in South Korea.
Especially a challenge because of theproximity of Seoul to the border and the
population density and the vulnerability.
And so, it started asa paper in the course.
And so I guess the lesson is forany graduate students listening,
hang on to your term papers, younever know where they'll end up.
But it started as a paper in the courseand, I finished up graduate school and
(02:01):
had this draft and found out that Darylwas working on something similar or
had been working on something similar.
And so I reached out and we sortof joined forces at that point
to conduct the analysis together.
Daryl Press (02:11):
I would add and say, the paper that we wrote together actually
has a double origin in Barry's coursein the sense that 25 years before
Nick took that course, I took thatcourse, TA'd it for a couple years,
and then I just completely stole it.
I ripped it off and I made a completeduplicate carbon copy of that
course, which I've been teachingfor the past 25 years at Dartmouth.
(02:32):
And one of the papers thatundergraduates have been writing.
Over and over year after year isthis question assessing the threat of
North Korean artillery against Seoul.
And yeah, they foundslightly different things.
They set it up different ways, butas I learned more and more about the
topic by advising them and mentoringthem on this paper, I started moving
(02:54):
to the point where I decided I wasgonna write on this topic, and that's
when I got this email from a graduatestudent I'd never heard of before.
This peppy, eager Nick Anderson, who hadwritten a terrific first draft and we
decided that we'd work on this together.
the paper really starts both from Nick'sdirection and from my direction, from
Barry's class, which has, quite clearly,influenced us both, but also just
(03:16):
influenced a generation or two generationsof scholars to work on topics like these.
Sheena Chestnut Greitens (03:20):
Can I ask a follow up question, which is, why
do you think that narrative has beenso persistent in policy and media and
military circles for such a long time?
What is it that has led people to beconvinced that this would be so deadly?
Nick Anderson (03:34):
I guess I would say a couple of things.
I think , on the one hand, as I mentioned,the vast number of artillery pieces that
North Korea has, their proximity to Seoul,Seoul's population density, it seems
like this combination that's very kindof deadly basically to put it plainly.
The other thing is that I thinkthere's like a sort of bias that when.
Things like you know, sort ofattacks on populated centers
(03:57):
have devastating effects.
They tend to be reported more.
They tend to stick in our minds more.
And we sort of remember those, theysort of stand out in terms of, our
memory and how we associate theeffects of these kinds of weapons.
What we don't think about and don'toften notice is that many of these
things happen and there's, thankfullyno fatalities, no casualties.
(04:17):
And so I think it's a kind of acombination of those two things.
Daryl Press (04:21):
Yeah, I would just build on that and say the
conventional wisdom is so intuitive.
I mean, we all know large caliberartillery is so destructive, and
we've all seen pictures from WorldWar II of Stalingrad, or the cities
in Germany that were just destroyedsome of them by bombing and some
of them by conventional artillery.
So I've had the experience that you'vehad, Sheena, walking around Seoul.
(04:43):
And you can imagine if a fewthousand shells fell that are so
destructive in this really denselypopulated space, how could they
not kill thousands and thousands ofpeople, tens of thousands of people.
And so the conventional wisdom has thispower of just intuition on its side.
And so, I'd say at least the firsttime I look at it, I don't actually
know this from Nick, but the firsttime I looked at it, my strong
(05:06):
assumption was that the conventionalwisdom would be roughly correct.
What did you think when you started Nick?
Nick Anderson (05:11):
Yeah, I, I would say not only.
Was that my assumption, the firstdraft of this paper that I wrote in, in
Barry's class, you sort of reinforced theconventional wisdom because of assumptions
I made going in, which maybe we'lltalk about later in the conversation.
But because of the assumptions I madegoing in, in that paper, yeah, the result
did mirror the conventional wisdom.
It was sort of tens of thousands,if I remember correctly.
Daryl Press (05:32):
I gotta say I didn't remember that.
I wasn't, yeah, settingyou up there, Nick.
Ryan Vest (05:37):
So you built a detailed open source model to estimate what an artillery
barrage on Seoul might actually look like.
Can you walk us through thebasic structure of that model
and what our audience needs tounderstand about your methodology?
Nick Anderson (05:48):
I'll say that what we're doing in this paper is what's
known as military campaign analysis.
That's sort of an approach tounderstanding hypothetical military
operations or military operationsthat haven't happened, and it's a way
to narrow the bounds of uncertaintyand get kind of a grasp of how they
might go if they happen in practice.
And often what's used is basic math, basicquantitative reasoning along with kind
(06:10):
of modeling of military operations in areally kind of basic and abstract way.
So that's what we're doing in the paper.
The model itself is in Microsoft Excel,it's posted on Harvard Dataverse,
for those who are interested.
You can go look at it, playwith it, these kinds of things.
But it basically has threekind of core components.
The first is the artillery sort ofshelling component of the model.
So this is the North Korean artilleryforces firing shells into Seoul.
(06:34):
That's piece number one.
And that's the main thing thatkind of generates our outcome, our
primary outcome of interest in thepaper, which is fatalities in Seoul.
The second component of the model isthe response by the Combined Forces
Command by the US and South Korea.
So they're basicallyusing their artillery.
It's what's known as counter battery fire.
They're using their artilleryto strike North Korean
(06:55):
artillery that's firing at them.
And then the third component of themodel is civil defense in South Korea.
What is the public doing in this moment?
What are the people of Seoul doing,and mainly this is about how quickly
they can get to the shelters thatare scattered throughout the city.
What this does is drives down thepopulation density over time, right?
(07:17):
People are sort of out on the street andin their homes and in buildings, and they
gradually make their way underground.
And this drives down the populationdensity and therefore makes
each strike less effective.
Daryl Press (07:27):
If I were one of your readers listening to Nick talk, my
reactions would be number one is,wow, you explained that clearly.
And my second reaction would be, man, Idon't think I believe your results at all.
I don't even know what your resultsare, I don't believe them because
you just described a model with threedistinct parts—each of which undoubtedly
has a bunch of variables underlyingthese— and it's gonna be sufficiently
(07:50):
complex that God knows what you'regonna find in any given circumstance.
And we all know war's unpredictable.
So I'd say, I'm glad youcould explain that clearly.
I'm glad you have a nice model, but I'mnot sure if I'm gonna believe anything.
What I would say in response is, thething that led me to think that Nick and
I were really onto something, and by ontosomething I mean that the likely level
(08:11):
of fatalities in Seoul would be terrible,but one to two orders of magnitude
less than the conventional wisdom.
The reason that led me to think thatthat's very likely to be true is not the
point estimates we reached, but ratherthe sensitivity analysis and our analysis
of how those point estimates changeas we vary a wide range of variables.
(08:35):
And the thing that convinced me andI think convinced Nick was when we
took our initial results and we ranliterally thousands and thousands of
simulations to try to figure out howthese results that we found varied
across a wide range of assumptions.
And the answer is they do vary, butthey cluster in a range of between
(08:58):
800 to a thousand dead on the verylow end to about 5,000 dead at the
high end, and they don't get anywherenear these 20,000, 30,000, a hundred
thousand numbers you frequently hear.
So what I would urge your readerto be thinking is that we're not
standing up here telling you we knowhow such a war would go, other than
(09:18):
to say it would be chaotic and itwould be terrible and it would be a
disaster and lots of people would die.
What we are saying is that across awide range of assumptions and across
a wide range of model specifications,that the numbers don't change very much,
and they cluster around ranges thatare way out of the norm of what you
(09:40):
read in the newspaper, what you hearresponsible American officials discuss.
Sheena Chestnut Greitens (09:44):
So one of the things that I think is interesting
is that your results really are,as you said a minute ago, an order
of magnitude or two lower thana lot of the existing estimates.
And that's for a conflict thatstays in some ways limited, right?
I think many of our readers probablywill have heard the much bigger numbers
that relate to a sort of more total waron the peninsula where you commonly will
(10:08):
hear the phrase, "a million dead and abillion dollars worth of damage." And
so I think to readers, the idea thatthere's a set of estimates for a more
limited conflict that is this much loweris really surprising and interesting.
And I wondered if you could talk alittle bit more about that sensitivity
analysis, because you're right.
(10:30):
You didn't just run one simulation andyou're not reporting the results of
just one scenario or one setup that hasyour favorite set of assumptions, right?
You not only incorporate the abilityof the combined forces counter battery
weapons and the ability of Seoul residentsto seek shelter, but you also identify
three different types of conflict scenariothat you run these simulations in, right?
(10:54):
There's a surprise scenario, a crisisscenario, and then a preemption scenario.
When I picture conflict on theKorean peninsula, these different
scenarios kind of help me lay out inmy mind how we get into a conflict
and therefore what it is that you'reactually measuring when you look at
the impact of North Korean artillery.
So can you tell us a little bit abouthow you developed those three scenarios
(11:15):
and how much they matter for your model?
Nick Anderson (11:17):
Yeah, sure.
So as you said, Sheena, there's threebasic scenarios that we envision:
surprise, crisis, and preemption.
The surprise scenario is one in whichsort of out of the blue the North
Koreans unleash masses of amounts ofartillery on Seoul with sort of no
warning and no sort of expectationthat this is going to occur.
Not the most plausible, maybe notthe most realistic scenario, but
(11:40):
one that does kind of come up inpolicy discussion and concern.
And so we wanted to sort of try it out.
The main thing that differs hereis like how much time we wait until
the US and ROK start responding.
That's like the big differencein terms of model terms.
That's how that differs.
So that's the sort of first scenario.
Second scenario is crisis.
This is probably in many ways thekind of most realistic scenario.
(12:01):
This is a case in which we imagine there'ssome sort of intense crisis happening.
Both sides are sort of expecting thatviolence may break out at any moment.
And so in this case, South Koreansand the United States are ready to
respond, and so they respond much morequickly than in the first scenario.
Third scenario is preemption.
This is a scenario in which we envisionthere's an expectation on the part
(12:22):
of the ROK Army and the US that NorthKorea is getting ready to launch some
sort of attack using its artillery.
And the South Koreans andthe US collectively decide to
preempt and to attack first.
And they do so not just with counterbattery fire, but also with strikes from
naval platforms and from air platforms.
So those are the three basic scenarios,as you said, each one we look at
(12:44):
in two different kind of versions.
There's the nominal version, whichmakes the assumption that the North
Koreans perform effectively as fastas their weapons will allow them to.
So they're sort of like a normalmilitary or a pretty adept military
that kind of goes through the cycleof setting up, firing, breaking down,
and relocating and reloading, andthen going through that cycle again.
(13:04):
So that's the sort of nominal scenario.
We also have what we consider to be amore realistic scenario, where it's like,
not necessarily the most adept military,everything takes them a little bit longer.
Their munitions don'tdetonate all the time.
Which we know in practice is a fact.
We've seen this in many caseswhen North Korea uses artillery.
So those are the sort ofsix basic scenarios that we
present in our main results.
And just to kind of put it on thetable so listeners can understand, the
(13:27):
estimated fatalities across these sixscenarios is, on the low end, roughly 700.
And that's the sort ofrealistic preemption scenario.
And then on the very high end,roughly 4,600, and that's the
surprise nominal scenario.
Daryl Press (13:41):
Again, from a listener standpoint, they might say, well, you
have no idea how a war's gonna start.
You have no idea what'sgonna come beforehand.
And we would say, you'reexactly right, nobody does.
And so what Nick just described is kindof our method for, I don't wanna say
throwing the kitchen sink at it, butjust trying to conceive the wide range
of ways this thing could kick off andwhat that might mean for the lag in
(14:03):
response time before South Korea andthe United States start responding.
And as he just described, we do as youwould expect, generate a significant range
of results from, I think he just said, 750fatalities to 4,500 or 4,800 fatalities.
And that's reassuring to me that themodel did what it should have done.
(14:23):
But on the other hand, while that isterrible —while that is a hundred or
200 times worse than what Tel Aviv isexperiencing right now, so it's terrible
— at the same time, it's nothing compared tothe "North Korea can destroy South Korea
with conventional forces," concern that isjust repeatedly asserted as kind of an end
(14:48):
of the conversation observation about themilitary balance on the Korean Peninsula.
What we found is, yes, there areproblems on the Korean peninsula.
But this is not the gravest one.
Nick Anderson (14:58):
I will just say one more thing, Sheena, 'cause you had asked
initially about the sensitivity analysis.
So just to also make surelisteners know what we did here.
Basically the model has a whole bunchof inputs in the Excel model, which
people can see if they go and look at it.
For sensitivity analysis,we did two basic things.
One was take every single kind of keyinput variable in the model and shift
it from its lowest plausible value toits highest plausible value to get a
(15:21):
sense of what that does to outcomes.
This was an effort to just see,do our results rest really heavily
on any individual key variables.
And the answer is that some are moreinfluential than others, but none of
them shift it to anything like theexpectation of the conventional wisdom.
The other thing we did was simulations.
What we did in these cases wasnot change one variable at a time,
but change all variables at once.
(15:43):
Microsoft Excel allows you todefine ranges for each variable.
Randomly select a value from thatrange, plug it into the model for all of
them, run the model, save the results.
And it does that 10,000times for each version.
And what that gives us is this verybroad range of potential outcomes.
(16:04):
And it allows us to see are thereunusual combinations of inputs here
that would give us extreme outcomes.
So, like you could imagine a day whereNorth Korea has more artillery ready
than we expect, more of their munitionsdetonate than we expect, the US and
ROK forces take longer to respond.
So that's one version.
And we can see that in the simulationdata or on the opposite end the
(16:26):
North Koreans munitions are notdetonating at a regular rate.
They're much lower,and so on and so forth.
So the point of the simulations is togive you that broad range of what are
the sort of extreme versions of thisthat we can imagine based on things
that we can't conceive of easily, butthe computer allows us to do for us.
Sheena Chestnut Greitens (16:43):
Can I follow up on one other factor that seems
like it might be important or that youidentify actually as being important
in your results, and that's justthat Seoul's ability to protect its
citizens, to get them to shelter quicklyseems to be an important factor in
the ranges that your models generate.
And so I wondered if you couldjust talk a little bit about what's
(17:03):
the story there and how much of adifference does the effort that Seoul
has put into civil defense make?
Daryl Press (17:10):
Yeah.
If I jumped in on this, I think at thecore there are three reasons that the
model turns up that our answers areso different than everyone else's, and
two of them are about what you justsaid, Sheena, about, in a way, Seoul's
ability to protect its civilians.
One of them is, and it's the one that'smost general and applies all around
the world, is that cities in generalprotect their civilians pretty darn well.
(17:35):
That when we think about the largedestructive radius of artillery shells
or aerial drop bombs, we kind of thinkabout them in a forest or think about them
in an open field, but dense cities arebasically jungles of reinforced concrete.
They're filled with cars and wallsand all sorts of things that deflect
(17:55):
pressure waves and absorb shrapnel.
And so one of the shocking things wefound, not even from the model, but
from the historical research we didon historical cases in which people
used aerial bombs or artillery againstcivilians in cities, is that you actually
had to use on average multiple artilleryshells to kill a single person in a city.
(18:20):
And it's simply becausecities protect people.
And so that's kind of result number one.
I have to say.
It's a result which we're seeing playout again in the Iranian ballistic
missile attack on Israeli cities.
So answer number one is what we discoveredis cities just are a hard place to kill
lots of civilians with high explosives.
(18:41):
Number two is, as you articulated,Seoul in particular, because they
face this threat, has gone togreat lengths to try to protect its
people from these kinds of attacks.
And there are lots and lotsof shelters all over the city.
There's a metro line which providesshelter to people who are anywhere
near it on the surface of the streets.
And then even tall buildings.
You know, the interior stairwellsof tall buildings are almost always,
(19:05):
they're basically vertical sheltersmade up of reinforced concrete, and
they're a terrific place to be if you'rein a disastrous situation like this.
So Seoul's development of civildefense capabilities and civil defense
plans also have greatly reducedthe vulnerability of its people.
Even though the people understandablydon't always take so seriously the
actual alarms, once the weaponsstart detonating, they quickly will.
(19:29):
The last thing is the South Koreangovernment, the South Korean military
gets a lot of credit, which is they'vebeen spending a great deal of time and a
great deal of money working this problemand working this problem by buying the
kinds of artillery systems that aredesigned to do counter battery fire,
to buy the radar systems to exercisethe command and control techniques.
(19:51):
Make no mistake, this is somethingthat they need to keep working on.
This is something that is very,very difficult to execute.
And the more they take seriously that job,the more lives they will save in Seoul.
But I really think it's those threethings put together: it's the inherent
protection provided by cities, it'sthe civil defense preparations that
South Korea has made, and it's theserious investment in the kind of
(20:13):
artillery and air systems that would benecessary for a counter battery fire.
That's why our results areone to two orders of magnitude
less than everybody else's.
Sheena Chestnut Greitens (20:21):
So you mentioned what's happening in the Middle East
right now, and so I can't resist askingfor your thoughts on this, even though
it's a little bit outside the scope ofyour article on the Korean Peninsula.
But for listeners, we are recordingthis in mid-June of 2025, at a moment
when conflict between Israel and Iranhas sharply escalated, and we have
(20:43):
seen the implication of the kinds ofattacks that may be similar to the
things that your article examines.
And so I wondered if you could tell us alittle bit about your thinking about, is
the contemporary conflict and the exchangebetween Israel and Iran that we're seeing
today the kind of thing that we can drawlessons from your article to help us
(21:03):
understand what we're seeing in the MiddleEast today, or is it just too different?
How should we think about yourresults in light of what's going
on today in the Middle East?
Daryl Press (21:13):
I think the horrible conflict going on right now between Israel and Iran
really highlights in a way our two mainfindings as opposed to the implications.
Finding number one is, we havebetter data, I think as of June
17th on both the number of attackson Israeli cities than we do on the
(21:35):
number of attacks on Iranian cities.
I think we have better data on the numberof fatalities in Israeli cities as well.
So I'll use that as kind ofthe source of the comparison.
And I would say the two things that comeloud and clear from this Iran-Israel
exchange, and particularly the Iranianballistic missile attacks on Israel,
are number one (21:51):
you don't have to kill that many people to cause a disaster
and to cause a situation that there'simmense amount of suffering and horror.
And at this point, approximately20 to 25 Israelis have been
killed in these attacks.
And yet I think the Israeli population,though steadfast is reeling a bit and
(22:11):
the pictures of Tel Aviv are horrifying.
And so I would just point out that whenNick and I say the attacks on Seoul are
not nearly as lethal as one believes, andNorth Korea does not have nearly as good
a capability, killing a thousand people inSeoul would probably mean injuring 5,000
people in Seoul or 10,000, and it wouldmean causing maybe that billion dollars
(22:32):
of damage that you just talked about.
So conclusion number one from what'shappening in the Middle East right now
with respect to civilian casualties,is even a relatively small number of
civilian fatalities might correspondwith a complete catastrophe, a
complete disaster, and a real shock.
Number two is, once again, you'reseeing this weird counterintuitive
finding (22:48):
that you can set off very large explosives in cities over and
over and over again and not kill nearlyas many people as one might expect.
At this point in the war, Iranhas fired about 350 ballistic
missiles at Israel, approximately.
It seems that about 35 of them, roughly,have detonated somewhere in Israel, mostly
(23:10):
around urban areas, and with 35 ballisticmissiles, they've managed to kill between
20 and 25 people on the ground, lessthan one person per ballistic missile.
And to compare this to the NorthKorea case, I'd point out that each of
these ballistic missiles has about 500kilogram-size whereas the very large yield
guns that North Korea would be firinghave about seven kilograms of explosive.
(23:33):
So it's a difference of a function of 70.
Now, because lethality andconventional explosives generally
increases with the cube root.
70 times the explosive is usuallyabout four times the lethal radius,
or 16 times the lethal area.
But basically, it falls eerily intothese exact ranges that Nick and
(23:54):
I found from the historical data.
So to me, what's striking fromthis is A: make no mistake there's
a war on the Korean peninsula andNorth Korea does this to South
Korea, this is a terrible war crime.
It's a catastrophe for South Korea,and we should be thinking and praying
and caring about our friends and lovedones in Seoul 'cause it's a disaster.
(24:15):
But also make no mistake that NorthKorea doesn't have either a war-winning
or a stalemating weapon, and in fact,the greatest thing that we should all
be worried about, besides the fatalitiesin Seoul, is that if North Korea
ever makes this disastrous mistake,that in reality it's their military,
including their artillery, which islikely to crumble almost immediately.
(24:39):
And the question we should alwaysbe worried about is, and then what?
What's North Korea's next move?
Ryan Vest (24:43):
So with the model, you've come up with these ranges of casualties,
and as we talked about what's going onin Iran and Israel today, this really
kind of stood out to me that you cameup with on the low end, around 700
for a preemptive strike versus closerto 5,000 for an unprepared strike.
Based on what we've seen, what kind oflessons should policymakers draw from both
(25:04):
what they're seeing in the Middle Eastnow, but really from your article, when
they look at these differences in numbers?
Certainly any loss of life is horrific,but 700 is a whole lot better than 4,700.
Is that something that theyshould be thinking about?
Nick Anderson (25:17):
As Daryl has already said, it's important that we don't
come across as sort of callous inconducting this kind of research and
talking about this number of fatalities.
It is, regardless of any ofthese numbers, it is horrific.
I think the thing that's most importantto kind of take away is, and as we
try and sort of emphasize in theintroduction of the article: the
expectations really are on the order ofkind of tens or hundreds of thousands.
(25:38):
And what we're finding is sort of,high hundreds to low thousands.
And we've even made quite a few sortof pretty conservative, pretty sort
of, North Korea-favorable assumptionsgoing in to sort of stack the deck
a little bit against our arguments.
And even then the findings are pretty low.
So there's a number of implicationsthat come from the project and I'll talk
about the first couple and, I'll letDaryl talk about some of the other ones.
(26:01):
The first, and I think importantone is that the US and the ROK's
ability to protect Seoul is greaterthan we've commonly thought.
They've made the right investments,they're sort of training with the
right materials, their militaries havedeveloped to a point where this is a
problem that is manageable— difficult,requires a lot of commitment— but
is ultimately a manageable problem.
And so the kind of, defeatism thatI think has defined a lot of the
(26:25):
discussion around the artillery threatto Seoul is maybe unwarranted, or
less warranted than many have thought.
Second important implication is that , ifwe are correct and if the North Koreans
know this, they should be deterredfrom engaging in this kind of behavior.
They should be much less likelyto engage in a massive artillery
bombardment of the South.
(26:45):
And this is also great news, right,for the United States and South Korea?
Conventional deterrence willkind of help keep the peace.
And I think that's a second,positive implication of the research.
Daryl Press (26:57):
Nick always does this.
Nick always gives the good newsstory and then he leaves me with
the bad news story, but fair enough.
I do think there's also two worrisomeconclusions that come from this.
And, The first one is that if there is acrisis, if there's a serious crisis and
it looks like war hangs in the balance,we will have all, of course all be hoping
that war can be avoided, that there'llbe a way for the diplomats to find a way
(27:21):
out of the conflict, but South Korea andUS military leaders and political leaders
will know that if we were to strike first,rather than waiting for North Korea to
cross over the threshold, that we couldprobably reduce the subsequent fatalities
in South Korea by maybe a factor of four.
What was it, about 750to a couple thousand.
(27:44):
And so they're gonna be stuck withthis really nasty dilemma where on
one hand they want to see peace, buton the other hand, if it looks like
war is likely, they want to go first.
And as we know from history, those kindsof dynamics sometimes lead to a war, which
in retrospect could have been avoided.
(28:04):
And so the preemption story itself isa mixture of good news and bad news:
good news that there's another tool,another arrow in the quiver that could
reduce further losses to South Korea.
The bad news is that might bethe useful tool that gets us
into a war that none of us want.
The other piece of bad news is theoverarching argument here that we're
(28:24):
making is in our minds, not just aboutthe artillery balance in Korea, but really
about the balance of power in Korea.
And as Nick kind of suggested earlier,experts who've watched the Korea for years
have known that the North Korean militaryin general is not what it appears.
It's a lot weaker.
It poses a much lower threatthan it appears on paper because
(28:47):
it's equipment is old, it doesn'ttrain very much, et cetera.
But North Korea's two Trump cardshave always been the artillery and the
threat in Seoul and the nuclear weapons.
And what we're basically findingis that that artillery is as much
of a paper tiger as the rest of theconventional North Korean military.
And as a result, if God forbid, we findourselves in a serious war on the Korean
(29:11):
peninsula, very soon into that war,the leaders in Pyongyang are likely to
see a conventional military force oftheirs that's collapsing, an artillery
force that's been disarmed, and they'regonna have the nasty choice of either
accepting the fate that Saddam Husseinfaced, or being pressured to find
(29:32):
some way of escalating and coercing anend to the war with nuclear weapons.
And so, like all these things,there's good news, there's bad
news, and they're often intertwined.
North Korea's profound weakness,not only in its main conventional
forces, but also in itsconventional artillery is good news.
The alternative would be worse news.
(29:52):
But do understand that with thatgood news comes with quite serious
pressures on North Korea to escalateto the nuclear level, should war come.
Sheena Chestnut Greitens (30:01):
You know, one of the things that I think is
really important about this articleis that it identifies a set of risks
that are just different from what theconventional wisdom commonly focuses on.
And this is one of the things that Ireally liked the first time I read this
piece, and one of the reasons why wethought it was important, and we're
glad to have the chance to publish itin the Texas National Security Review
(30:23):
is that, I don't in any way in readingthe article or listening to you talk
about this, take away that you'recavalier about the costs of an artillery
exchange between the two Koreas.
It's much more that actually mostpeople in hearing, well, the North
Korean military is conventionallyweaker than outside analysts think
would think, wow, that's great.
(30:44):
That's great news.
And, there is a part of that that's good,although having interviewed some former
members of the North Korean militarywho were hungry during their military
service, there's also real human costthat sort of undergirds that military
weakness that has its own just sortof human level tragedy and cost to it.
But, what I think is really importantis this part that you're emphasizing
(31:09):
now, which is that something thatwe might regard as a positive, North
Korea's conventional military beingweak and less able to threaten Seoul or
the citizens of South Korea, Americanresidents, or international residents
of South Korea than we might think—that all sounds good— but there's a set
(31:29):
of risks even to the good news, right?
North Korea in the past couple yearshas for the first time explicitly
laid out conditions under which itmight use nuclear weapons first.
And so, one of the things that I reallyappreciate is the tying in of this
conventional threat, the artilleryrisks and costs of an artillery
exchange with the risks of nuclearescalation on a peninsula that has long
(31:53):
been one of the world's most seriousflashpoints, so I guess I wanted to
ask if there's one thing or a fewthings that you wanted policymakers
to take away from this article?
What would you want them to walkaway knowing or taking back to their
teams to think about and work on?
Nick Anderson (32:09):
I'll take a stab at that first.
So I think probably if there's kind of oneoverarching, policy relevant finding of
the article, it's that the North Koreanconventional military is weaker in all
of its aspects than we have thought.
And I think a lot of that has beenunderstood, but artillery has stood
out as this part that was particularlystrong or threatening still, and we're
(32:31):
finding that's no longer the case.
And if that's true, then a lot of whatwe're seeing, in terms of what North Korea
is doing maybe makes more sense, right?
The sort of doubling down on thenuclear capability under Kim Jong-Un.
It's really taken off since he tookpower in the 2010s, and as you said,
this change in North Korea's doctrine in2022 to allow the first use of nuclear
(32:54):
weapons— maybe they would've donethat anyway— but now it's out there.
It's sort of explicit in cases wherethe regime or leadership is threatened.
Trading away millions of artillery shellsto Russia in Ukraine in exchange, as
far as we understand, for technologythat will aid the nuclear program,
pushing really, really hard towards,an ICBM capability and the ability to
(33:16):
strike the continental United States.
All of this makes a little bitmore sense if you understand
that they don't have as strong aconventional deterrent as we thought.
And so, going forward, I thinkNorth Korea is gonna just be in more
situations in which it will relymore heavily on nuclear weapons.
Hopefully that doesn't includeuse, but sort of this is like in
some ways the kind of last leg.
(33:37):
And if that's true, the odds of themabandoning this, as long as relations
remain as they are now are very, very low.
The expectation that they may givethese up someday for some sort of
deal or grand bargain, I think theresults of our research suggests
that that's unlikely to happen.
Daryl Press (33:51):
I agree with everything Nick just said.
I would add two things.
If I was thinking, what are the twoflashing red to-dos that come out
of our research for policymakers?
I'd say the first one is largelyfor South Korean audience, for
South Korean military audience inparticular, which is to say your
society has already done the hard part.
Your society has already found the moneyand found the will to buy the weapon
(34:14):
systems that are necessary for you todo this really difficult thing, namely
protect Seoul, which is nestled rightup next to the North Korean border,
and yet you now have the equipment todo it, but the equipment is part one.
Because the mission is really difficult.
It's about rapidlyidentifying incoming shells.
It's figuring out the origin of thoseshells, distributing targets in a chaotic
(34:38):
environment very quickly to shooterstargeting, doing bomb damage assessment.
None of that is simple, and especiallyin the chaos of war, that therefore what
South Korea really has to do now is doubledown and triple down in the training,
and kind of the soft work that has tohappen underneath this— about trying
(35:00):
to practice doing this in a chaotic,difficult circumstance and getting
as good as you possibly can becausethey have to succeed the first time.
And so I'd say the firstthing is congratulations.
South Korea's bought the stuffthat will let them defend Seoul.
Now they have to do the hardwork of actually preparing to
execute a difficult mission.
The second thing is a takeaway both forUS and South Korean leaders, leaders
(35:23):
at the political level, leaders atthe military level, which is this
challenge of intra war deterrence.
In Ukraine we're wrestling withhow much can we push back Russia in
territory they've seized before youstart getting toward Russian red
lines and there are views all overthe map about this and big debates.
On the Korean Peninsula, it mightbe much, much quicker and much,
(35:45):
much more dire, which is if thingsgo the way that Nick and I expect.
If there is, God forbid, a majorconventional war in the Korean
peninsula, we really do not thinkthe North Korean military is gonna
survive as a functioning force for, Idon't know, more than a week, 10 days.
And very rapidly, the North Koreangovernment is gonna be facing life and
death decisions about how they make usstop, and they don't have a whole lot
(36:11):
of tools other than nuclear weapons.
So I think what this means forUS and South Korean leaders is
we have to be thinking really,really hard now about A: how we're
gonna fight that conventional war.
Sure, defend Seoul fromincoming artillery.
What are the things in our normalmilitary toolkit that we are
(36:32):
gonna do and not gonna do for thepurpose of preventing escalation?
How are we gonna message that war?
What are we gonna be telling North Koreanleaders and other observers real-time
about what our objectives are and are not?
And fundamentally, how we're gonnacome to grips, not just as the United
States, but as an alliance with how muchdamage we want to inflict and how much
(36:55):
success we want to achieve in this warto balance the operational objective
of defending South Korea and thwartingan attack, and the strategic objective
of all of us getting out of this warwithout a significant nuclear exchange.
And so I think both the training,training, training conventional
(37:18):
job is really job one forthe South Korean army today.
And the thinking about intra warescalation is job number one for
strategists at the US strategiccommand, at CFC, at USFK and in the
South Korean military and leadership.
And they are flashing red topics of justas great importance as the dangers we
(37:42):
face in the Taiwan Strait or in EasternEurope, or now in the Middle East.
Sheena Chestnut Greitens (37:47):
I found that point actually really striking.
I'm glad you brought up the compressionof time that decision makers will have
to think about these escalation risks.
And that intra war deterrence aspect whenit's compressed in time, I think could
be, as you said, a flashing red light.
So I'm glad you brought that in, andI'm glad you phrased it that way.
Daryl Press (38:06):
That's great, and hey, I don't know if this belongs in this podcast
or not, but it is so nice have TNSR andits particular voice in terms of the
topics that it covers as an outlet for notjust Nick and me, but for our community,
all of us, and to be a place that'sopen to publishing campaign analyses.
I mean, there aren't very manyplaces out there that see this as
(38:29):
an important part of either foreignpolicy analysis or scholarly research.
And so, having TNSR as a placewho sees the importance and the
contributions of that style of workis a tremendous benefit to the field.
Sheena Chestnut Greitens (38:43):
Well, I think your article's a great example of why
we at TNSR find it valuable and whywe wanted to put it out there as a
contribution for the security studiesscholarly community, but also for the
American and international strategycommunity and the community that thinks
about these risks and these problemson a daily basis as professionals
and as practitioners, so we'rereally delighted to have you both on.
(39:06):
Thanks so much for joining ustoday to talk about your article.
Congrats on having it out and, again,really pleased to feature it in TNSR.
Thanks for being with us today.
Nick Anderson (39:14):
Thanks for having us.
Daryl Press (39:15):
Thank you.
Ryan Vest (39:17):
Thanks for joining us on Horns of a Dilemma from the
Texas National Security Review.
Our guests today have beenNicholas Anderson and Daryl Press,
authors of "Lost Seoul (39:24):
Assessing Pyongyang's Other Deterrent,"
which as always can be accessedfor free on our website, tnsr.org.
If you enjoyed this episode,be sure to subscribe and leave
a review wherever you listen.
You can find more of our work attnsr.org and in the show notes.
Today's episode was produced byTNSR Digital and Technical Manager
Jordan Morning and made possibleby The University of Texas System.
(39:48):
This is Ryan Vest andSheena Chestnut Greitens.
Thanks for listening.
Welcome to Horns of a Dilemma, the podcast of the Texas
National Security Review at theUniversity of Texas at Austin.
I'm Ryan Vest, Executive Editorof TNSR, and I'm here with our
Editor-in-Chief, Sheena Chestnut Greitens.
We're pleased to have joining ustoday, two distinguished scholars in
security studies, Nicholas Anderson andDaryl Press, authors of "Lost Seoul?
(00:26):
Assessing Pyongyang’s OtherDeterrent," which is featured in
Volume 8, Issue 3 of the journal.
Nick is an assistant professor ofPolitical Science and International
Affairs at the George WashingtonUniversity, and Daryl is Professor
of Government and faculty Directorof the Davidson Institute for
Global Security at Dartmouth.
Gentlemen, welcome to Horns of a Dilemma.
It's great to have you on the show.
Nick Anderson (00:46):
Thanks for having us.
Daryl Press (00:47):
Thank you.
Ryan Vest (00:47):
Thank you.
Sheena Chestnut Greitens (00:48):
To get us started, I wonder if you could
just tell us a little bit about yourmotivation for writing the article.
There's a widespread assumption for aslong as I've been going to South Korea,
that North Korean artillery capabilitycould be devastating for Seoul at any
point should a conflict break out on theKorean Peninsula, and your piece pretty
directly challenges that narrative.
(01:09):
But I'm curious how you cameto do research on this topic.
What motivated you to even startasking questions about whether
that conventional wisdom is true?
Could you tell us about that?
Nick Anderson (01:19):
Sure.
I'll start with my half of it, becausewe actually started separately.
This started as a term paper in BarryPosen's legendary US military power
class, which I wasn't a student at MITbut I audited the course when I was a
fellow at the Belfer Center at Harvard.
And I had the same sort of assumptionthat you did, Sheena, going in that the
North Korean artillery is this sort ofconventional Trump card, very difficult
(01:41):
to deal with if you're South Korea orif you're the US forces in South Korea.
Especially a challenge because of theproximity of Seoul to the border and the
population density and the vulnerability.
And so, it started asa paper in the course.
And so I guess the lesson is forany graduate students listening,
hang on to your term papers, younever know where they'll end up.
But it started as a paper in the courseand, I finished up graduate school and
(02:01):
had this draft and found out that Darylwas working on something similar or
had been working on something similar.
And so I reached out and we sortof joined forces at that point
to conduct the analysis together.
Daryl Press (02:11):
I would add and say, the paper that we wrote together actually
has a double origin in Barry's coursein the sense that 25 years before
Nick took that course, I took thatcourse, TA'd it for a couple years,
and then I just completely stole it.
I ripped it off and I made a completeduplicate carbon copy of that
course, which I've been teachingfor the past 25 years at Dartmouth.
(02:32):
And one of the papers thatundergraduates have been writing.
Over and over year after year isthis question assessing the threat of
North Korean artillery against Seoul.
And yeah, they foundslightly different things.
They set it up different ways, butas I learned more and more about the
topic by advising them and mentoringthem on this paper, I started moving
(02:54):
to the point where I decided I wasgonna write on this topic, and that's
when I got this email from a graduatestudent I'd never heard of before.
This peppy, eager Nick Anderson, who hadwritten a terrific first draft and we
decided that we'd work on this together.
the paper really starts both from Nick'sdirection and from my direction, from
Barry's class, which has, quite clearly,influenced us both, but also just
(03:16):
influenced a generation or two generationsof scholars to work on topics like these.
Sheena Chestnut Greitens (03:20):
Can I ask a follow up question, which is, why
do you think that narrative has beenso persistent in policy and media and
military circles for such a long time?
What is it that has led people to beconvinced that this would be so deadly?
Nick Anderson (03:34):
I guess I would say a couple of things.
I think , on the one hand, as I mentioned,the vast number of artillery pieces that
North Korea has, their proximity to Seoul,Seoul's population density, it seems
like this combination that's very kindof deadly basically to put it plainly.
The other thing is that I thinkthere's like a sort of bias that when.
Things like you know, sort ofattacks on populated centers
(03:57):
have devastating effects.
They tend to be reported more.
They tend to stick in our minds more.
And we sort of remember those, theysort of stand out in terms of, our
memory and how we associate theeffects of these kinds of weapons.
What we don't think about and don'toften notice is that many of these
things happen and there's, thankfullyno fatalities, no casualties.
(04:17):
And so I think it's a kind of acombination of those two things.
Daryl Press (04:21):
Yeah, I would just build on that and say the
conventional wisdom is so intuitive.
I mean, we all know large caliberartillery is so destructive, and
we've all seen pictures from WorldWar II of Stalingrad, or the cities
in Germany that were just destroyedsome of them by bombing and some
of them by conventional artillery.
So I've had the experience that you'vehad, Sheena, walking around Seoul.
(04:43):
And you can imagine if a fewthousand shells fell that are so
destructive in this really denselypopulated space, how could they
not kill thousands and thousands ofpeople, tens of thousands of people.
And so the conventional wisdom has thispower of just intuition on its side.
And so, I'd say at least the firsttime I look at it, I don't actually
know this from Nick, but the firsttime I looked at it, my strong
(05:06):
assumption was that the conventionalwisdom would be roughly correct.
What did you think when you started Nick?
Nick Anderson (05:11):
Yeah, I, I would say not only.
Was that my assumption, the firstdraft of this paper that I wrote in, in
Barry's class, you sort of reinforced theconventional wisdom because of assumptions
I made going in, which maybe we'lltalk about later in the conversation.
But because of the assumptions I madegoing in, in that paper, yeah, the result
did mirror the conventional wisdom.
It was sort of tens of thousands,if I remember correctly.
Daryl Press (05:32):
I gotta say I didn't remember that.
I wasn't, yeah, settingyou up there, Nick.
Ryan Vest (05:37):
So you built a detailed open source model to estimate what an artillery
barrage on Seoul might actually look like.
Can you walk us through thebasic structure of that model
and what our audience needs tounderstand about your methodology?
Nick Anderson (05:48):
I'll say that what we're doing in this paper is what's
known as military campaign analysis.
That's sort of an approach tounderstanding hypothetical military
operations or military operationsthat haven't happened, and it's a way
to narrow the bounds of uncertaintyand get kind of a grasp of how they
might go if they happen in practice.
And often what's used is basic math, basicquantitative reasoning along with kind
(06:10):
of modeling of military operations in areally kind of basic and abstract way.
So that's what we're doing in the paper.
The model itself is in Microsoft Excel,it's posted on Harvard Dataverse,
for those who are interested.
You can go look at it, playwith it, these kinds of things.
But it basically has threekind of core components.
The first is the artillery sort ofshelling component of the model.
So this is the North Korean artilleryforces firing shells into Seoul.
(06:34):
That's piece number one.
And that's the main thing thatkind of generates our outcome, our
primary outcome of interest in thepaper, which is fatalities in Seoul.
The second component of the model isthe response by the Combined Forces
Command by the US and South Korea.
So they're basicallyusing their artillery.
It's what's known as counter battery fire.
They're using their artilleryto strike North Korean
(06:55):
artillery that's firing at them.
And then the third component of themodel is civil defense in South Korea.
What is the public doing in this moment?
What are the people of Seoul doing,and mainly this is about how quickly
they can get to the shelters thatare scattered throughout the city.
What this does is drives down thepopulation density over time, right?
(07:17):
People are sort of out on the street andin their homes and in buildings, and they
gradually make their way underground.
And this drives down the populationdensity and therefore makes
each strike less effective.
Daryl Press (07:27):
If I were one of your readers listening to Nick talk, my
reactions would be number one is,wow, you explained that clearly.
And my second reaction would be, man, Idon't think I believe your results at all.
I don't even know what your resultsare, I don't believe them because
you just described a model with threedistinct parts—each of which undoubtedly
has a bunch of variables underlyingthese— and it's gonna be sufficiently
(07:50):
complex that God knows what you'regonna find in any given circumstance.
And we all know war's unpredictable.
So I'd say, I'm glad youcould explain that clearly.
I'm glad you have a nice model, but I'mnot sure if I'm gonna believe anything.
What I would say in response is, thething that led me to think that Nick and
I were really onto something, and by ontosomething I mean that the likely level
(08:11):
of fatalities in Seoul would be terrible,but one to two orders of magnitude
less than the conventional wisdom.
The reason that led me to think thatthat's very likely to be true is not the
point estimates we reached, but ratherthe sensitivity analysis and our analysis
of how those point estimates changeas we vary a wide range of variables.
(08:35):
And the thing that convinced me andI think convinced Nick was when we
took our initial results and we ranliterally thousands and thousands of
simulations to try to figure out howthese results that we found varied
across a wide range of assumptions.
And the answer is they do vary, butthey cluster in a range of between
(08:58):
800 to a thousand dead on the verylow end to about 5,000 dead at the
high end, and they don't get anywherenear these 20,000, 30,000, a hundred
thousand numbers you frequently hear.
So what I would urge your readerto be thinking is that we're not
standing up here telling you we knowhow such a war would go, other than
(09:18):
to say it would be chaotic and itwould be terrible and it would be a
disaster and lots of people would die.
What we are saying is that across awide range of assumptions and across
a wide range of model specifications,that the numbers don't change very much,
and they cluster around ranges thatare way out of the norm of what you
(09:40):
read in the newspaper, what you hearresponsible American officials discuss.
Sheena Chestnut Greitens (09:44):
So one of the things that I think is interesting
is that your results really are,as you said a minute ago, an order
of magnitude or two lower thana lot of the existing estimates.
And that's for a conflict thatstays in some ways limited, right?
I think many of our readers probablywill have heard the much bigger numbers
that relate to a sort of more total waron the peninsula where you commonly will
(10:08):
hear the phrase, "a million dead and abillion dollars worth of damage." And
so I think to readers, the idea thatthere's a set of estimates for a more
limited conflict that is this much loweris really surprising and interesting.
And I wondered if you could talk alittle bit more about that sensitivity
analysis, because you're right.
(10:30):
You didn't just run one simulation andyou're not reporting the results of
just one scenario or one setup that hasyour favorite set of assumptions, right?
You not only incorporate the abilityof the combined forces counter battery
weapons and the ability of Seoul residentsto seek shelter, but you also identify
three different types of conflict scenariothat you run these simulations in, right?
(10:54):
There's a surprise scenario, a crisisscenario, and then a preemption scenario.
When I picture conflict on theKorean peninsula, these different
scenarios kind of help me lay out inmy mind how we get into a conflict
and therefore what it is that you'reactually measuring when you look at
the impact of North Korean artillery.
So can you tell us a little bit abouthow you developed those three scenarios
(11:15):
and how much they matter for your model?
Nick Anderson (11:17):
Yeah, sure.
So as you said, Sheena, there's threebasic scenarios that we envision:
surprise, crisis, and preemption.
The surprise scenario is one in whichsort of out of the blue the North
Koreans unleash masses of amounts ofartillery on Seoul with sort of no
warning and no sort of expectationthat this is going to occur.
Not the most plausible, maybe notthe most realistic scenario, but
(11:40):
one that does kind of come up inpolicy discussion and concern.
And so we wanted to sort of try it out.
The main thing that differs hereis like how much time we wait until
the US and ROK start responding.
That's like the big differencein terms of model terms.
That's how that differs.
So that's the sort of first scenario.
Second scenario is crisis.
This is probably in many ways thekind of most realistic scenario.
(12:01):
This is a case in which we imagine there'ssome sort of intense crisis happening.
Both sides are sort of expecting thatviolence may break out at any moment.
And so in this case, South Koreansand the United States are ready to
respond, and so they respond much morequickly than in the first scenario.
Third scenario is preemption.
This is a scenario in which we envisionthere's an expectation on the part
(12:22):
of the ROK Army and the US that NorthKorea is getting ready to launch some
sort of attack using its artillery.
And the South Koreans andthe US collectively decide to
preempt and to attack first.
And they do so not just with counterbattery fire, but also with strikes from
naval platforms and from air platforms.
So those are the three basic scenarios,as you said, each one we look at
(12:44):
in two different kind of versions.
There's the nominal version, whichmakes the assumption that the North
Koreans perform effectively as fastas their weapons will allow them to.
So they're sort of like a normalmilitary or a pretty adept military
that kind of goes through the cycleof setting up, firing, breaking down,
and relocating and reloading, andthen going through that cycle again.
(13:04):
So that's the sort of nominal scenario.
We also have what we consider to be amore realistic scenario, where it's like,
not necessarily the most adept military,everything takes them a little bit longer.
Their munitions don'tdetonate all the time.
Which we know in practice is a fact.
We've seen this in many caseswhen North Korea uses artillery.
So those are the sort ofsix basic scenarios that we
present in our main results.
And just to kind of put it on thetable so listeners can understand, the
(13:27):
estimated fatalities across these sixscenarios is, on the low end, roughly 700.
And that's the sort ofrealistic preemption scenario.
And then on the very high end,roughly 4,600, and that's the
surprise nominal scenario.
Daryl Press (13:41):
Again, from a listener standpoint, they might say, well, you
have no idea how a war's gonna start.
You have no idea what'sgonna come beforehand.
And we would say, you'reexactly right, nobody does.
And so what Nick just described is kindof our method for, I don't wanna say
throwing the kitchen sink at it, butjust trying to conceive the wide range
of ways this thing could kick off andwhat that might mean for the lag in
(14:03):
response time before South Korea andthe United States start responding.
And as he just described, we do as youwould expect, generate a significant range
of results from, I think he just said, 750fatalities to 4,500 or 4,800 fatalities.
And that's reassuring to me that themodel did what it should have done.
(14:23):
But on the other hand, while that isterrible —while that is a hundred or
200 times worse than what Tel Aviv isexperiencing right now, so it's terrible
— at the same time, it's nothing compared tothe "North Korea can destroy South Korea
with conventional forces," concern that isjust repeatedly asserted as kind of an end
(14:48):
of the conversation observation about themilitary balance on the Korean Peninsula.
What we found is, yes, there areproblems on the Korean peninsula.
But this is not the gravest one.
Nick Anderson (14:58):
I will just say one more thing, Sheena, 'cause you had asked
initially about the sensitivity analysis.
So just to also make surelisteners know what we did here.
Basically the model has a whole bunchof inputs in the Excel model, which
people can see if they go and look at it.
For sensitivity analysis,we did two basic things.
One was take every single kind of keyinput variable in the model and shift
it from its lowest plausible value toits highest plausible value to get a
(15:21):
sense of what that does to outcomes.
This was an effort to just see,do our results rest really heavily
on any individual key variables.
And the answer is that some are moreinfluential than others, but none of
them shift it to anything like theexpectation of the conventional wisdom.
The other thing we did was simulations.
What we did in these cases wasnot change one variable at a time,
but change all variables at once.
(15:43):
Microsoft Excel allows you todefine ranges for each variable.
Randomly select a value from thatrange, plug it into the model for all of
them, run the model, save the results.
And it does that 10,000times for each version.
And what that gives us is this verybroad range of potential outcomes.
(16:04):
And it allows us to see are thereunusual combinations of inputs here
that would give us extreme outcomes.
So, like you could imagine a day whereNorth Korea has more artillery ready
than we expect, more of their munitionsdetonate than we expect, the US and
ROK forces take longer to respond.
So that's one version.
And we can see that in the simulationdata or on the opposite end the
(16:26):
North Koreans munitions are notdetonating at a regular rate.
They're much lower,and so on and so forth.
So the point of the simulations is togive you that broad range of what are
the sort of extreme versions of thisthat we can imagine based on things
that we can't conceive of easily, butthe computer allows us to do for us.
Sheena Chestnut Greitens (16:43):
Can I follow up on one other factor that seems
like it might be important or that youidentify actually as being important
in your results, and that's justthat Seoul's ability to protect its
citizens, to get them to shelter quicklyseems to be an important factor in
the ranges that your models generate.
And so I wondered if you couldjust talk a little bit about what's
(17:03):
the story there and how much of adifference does the effort that Seoul
has put into civil defense make?
Daryl Press (17:10):
Yeah.
If I jumped in on this, I think at thecore there are three reasons that the
model turns up that our answers areso different than everyone else's, and
two of them are about what you justsaid, Sheena, about, in a way, Seoul's
ability to protect its civilians.
One of them is, and it's the one that'smost general and applies all around
the world, is that cities in generalprotect their civilians pretty darn well.
(17:35):
That when we think about the largedestructive radius of artillery shells
or aerial drop bombs, we kind of thinkabout them in a forest or think about them
in an open field, but dense cities arebasically jungles of reinforced concrete.
They're filled with cars and wallsand all sorts of things that deflect
(17:55):
pressure waves and absorb shrapnel.
And so one of the shocking things wefound, not even from the model, but
from the historical research we didon historical cases in which people
used aerial bombs or artillery againstcivilians in cities, is that you actually
had to use on average multiple artilleryshells to kill a single person in a city.
(18:20):
And it's simply becausecities protect people.
And so that's kind of result number one.
I have to say.
It's a result which we're seeing playout again in the Iranian ballistic
missile attack on Israeli cities.
So answer number one is what we discoveredis cities just are a hard place to kill
lots of civilians with high explosives.
(18:41):
Number two is, as you articulated,Seoul in particular, because they
face this threat, has gone togreat lengths to try to protect its
people from these kinds of attacks.
And there are lots and lotsof shelters all over the city.
There's a metro line which providesshelter to people who are anywhere
near it on the surface of the streets.
And then even tall buildings.
You know, the interior stairwellsof tall buildings are almost always,
(19:05):
they're basically vertical sheltersmade up of reinforced concrete, and
they're a terrific place to be if you'rein a disastrous situation like this.
So Seoul's development of civildefense capabilities and civil defense
plans also have greatly reducedthe vulnerability of its people.
Even though the people understandablydon't always take so seriously the
actual alarms, once the weaponsstart detonating, they quickly will.
(19:29):
The last thing is the South Koreangovernment, the South Korean military
gets a lot of credit, which is they'vebeen spending a great deal of time and a
great deal of money working this problemand working this problem by buying the
kinds of artillery systems that aredesigned to do counter battery fire,
to buy the radar systems to exercisethe command and control techniques.
(19:51):
Make no mistake, this is somethingthat they need to keep working on.
This is something that is very,very difficult to execute.
And the more they take seriously that job,the more lives they will save in Seoul.
But I really think it's those threethings put together: it's the inherent
protection provided by cities, it'sthe civil defense preparations that
South Korea has made, and it's theserious investment in the kind of
(20:13):
artillery and air systems that would benecessary for a counter battery fire.
That's why our results areone to two orders of magnitude
less than everybody else's.
Sheena Chestnut Greitens (20:21):
So you mentioned what's happening in the Middle East
right now, and so I can't resist askingfor your thoughts on this, even though
it's a little bit outside the scope ofyour article on the Korean Peninsula.
But for listeners, we are recordingthis in mid-June of 2025, at a moment
when conflict between Israel and Iranhas sharply escalated, and we have
(20:43):
seen the implication of the kinds ofattacks that may be similar to the
things that your article examines.
And so I wondered if you could tell us alittle bit about your thinking about, is
the contemporary conflict and the exchangebetween Israel and Iran that we're seeing
today the kind of thing that we can drawlessons from your article to help us
(21:03):
understand what we're seeing in the MiddleEast today, or is it just too different?
How should we think about yourresults in light of what's going
on today in the Middle East?
Daryl Press (21:13):
I think the horrible conflict going on right now between Israel and Iran
really highlights in a way our two mainfindings as opposed to the implications.
Finding number one is, we havebetter data, I think as of June
17th on both the number of attackson Israeli cities than we do on the
(21:35):
number of attacks on Iranian cities.
I think we have better data on the numberof fatalities in Israeli cities as well.
So I'll use that as kind ofthe source of the comparison.
And I would say the two things that comeloud and clear from this Iran-Israel
exchange, and particularly the Iranianballistic missile attacks on Israel,
are number one (21:51):
you don't have to kill that many people to cause a disaster
and to cause a situation that there'simmense amount of suffering and horror.
And at this point, approximately20 to 25 Israelis have been
killed in these attacks.
And yet I think the Israeli population,though steadfast is reeling a bit and
(22:11):
the pictures of Tel Aviv are horrifying.
And so I would just point out that whenNick and I say the attacks on Seoul are
not nearly as lethal as one believes, andNorth Korea does not have nearly as good
a capability, killing a thousand people inSeoul would probably mean injuring 5,000
people in Seoul or 10,000, and it wouldmean causing maybe that billion dollars
(22:32):
of damage that you just talked about.
So conclusion number one from what'shappening in the Middle East right now
with respect to civilian casualties,is even a relatively small number of
civilian fatalities might correspondwith a complete catastrophe, a
complete disaster, and a real shock.
Number two is, once again, you'reseeing this weird counterintuitive
finding (22:48):
that you can set off very large explosives in cities over and
over and over again and not kill nearlyas many people as one might expect.
At this point in the war, Iranhas fired about 350 ballistic
missiles at Israel, approximately.
It seems that about 35 of them, roughly,have detonated somewhere in Israel, mostly
(23:10):
around urban areas, and with 35 ballisticmissiles, they've managed to kill between
20 and 25 people on the ground, lessthan one person per ballistic missile.
And to compare this to the NorthKorea case, I'd point out that each of
these ballistic missiles has about 500kilogram-size whereas the very large yield
guns that North Korea would be firinghave about seven kilograms of explosive.
(23:33):
So it's a difference of a function of 70.
Now, because lethality andconventional explosives generally
increases with the cube root.
70 times the explosive is usuallyabout four times the lethal radius,
or 16 times the lethal area.
But basically, it falls eerily intothese exact ranges that Nick and
(23:54):
I found from the historical data.
So to me, what's striking fromthis is A: make no mistake there's
a war on the Korean peninsula andNorth Korea does this to South
Korea, this is a terrible war crime.
It's a catastrophe for South Korea,and we should be thinking and praying
and caring about our friends and lovedones in Seoul 'cause it's a disaster.
(24:15):
But also make no mistake that NorthKorea doesn't have either a war-winning
or a stalemating weapon, and in fact,the greatest thing that we should all
be worried about, besides the fatalitiesin Seoul, is that if North Korea
ever makes this disastrous mistake,that in reality it's their military,
including their artillery, which islikely to crumble almost immediately.
(24:39):
And the question we should alwaysbe worried about is, and then what?
What's North Korea's next move?
Ryan Vest (24:43):
So with the model, you've come up with these ranges of casualties,
and as we talked about what's going onin Iran and Israel today, this really
kind of stood out to me that you cameup with on the low end, around 700
for a preemptive strike versus closerto 5,000 for an unprepared strike.
Based on what we've seen, what kind oflessons should policymakers draw from both
(25:04):
what they're seeing in the Middle Eastnow, but really from your article, when
they look at these differences in numbers?
Certainly any loss of life is horrific,but 700 is a whole lot better than 4,700.
Is that something that theyshould be thinking about?
Nick Anderson (25:17):
As Daryl has already said, it's important that we don't
come across as sort of callous inconducting this kind of research and
talking about this number of fatalities.
It is, regardless of any ofthese numbers, it is horrific.
I think the thing that's most importantto kind of take away is, and as we
try and sort of emphasize in theintroduction of the article: the
expectations really are on the order ofkind of tens or hundreds of thousands.
(25:38):
And what we're finding is sort of,high hundreds to low thousands.
And we've even made quite a few sortof pretty conservative, pretty sort
of, North Korea-favorable assumptionsgoing in to sort of stack the deck
a little bit against our arguments.
And even then the findings are pretty low.
So there's a number of implicationsthat come from the project and I'll talk
about the first couple and, I'll letDaryl talk about some of the other ones.
(26:01):
The first, and I think importantone is that the US and the ROK's
ability to protect Seoul is greaterthan we've commonly thought.
They've made the right investments,they're sort of training with the
right materials, their militaries havedeveloped to a point where this is a
problem that is manageable— difficult,requires a lot of commitment— but
is ultimately a manageable problem.
And so the kind of, defeatism thatI think has defined a lot of the
(26:25):
discussion around the artillery threatto Seoul is maybe unwarranted, or
less warranted than many have thought.
Second important implication is that , ifwe are correct and if the North Koreans
know this, they should be deterredfrom engaging in this kind of behavior.
They should be much less likelyto engage in a massive artillery
bombardment of the South.
(26:45):
And this is also great news, right,for the United States and South Korea?
Conventional deterrence willkind of help keep the peace.
And I think that's a second,positive implication of the research.
Daryl Press (26:57):
Nick always does this.
Nick always gives the good newsstory and then he leaves me with
the bad news story, but fair enough.
I do think there's also two worrisomeconclusions that come from this.
And, The first one is that if there is acrisis, if there's a serious crisis and
it looks like war hangs in the balance,we will have all, of course all be hoping
that war can be avoided, that there'llbe a way for the diplomats to find a way
(27:21):
out of the conflict, but South Korea andUS military leaders and political leaders
will know that if we were to strike first,rather than waiting for North Korea to
cross over the threshold, that we couldprobably reduce the subsequent fatalities
in South Korea by maybe a factor of four.
What was it, about 750to a couple thousand.
(27:44):
And so they're gonna be stuck withthis really nasty dilemma where on
one hand they want to see peace, buton the other hand, if it looks like
war is likely, they want to go first.
And as we know from history, those kindsof dynamics sometimes lead to a war, which
in retrospect could have been avoided.
(28:04):
And so the preemption story itself isa mixture of good news and bad news:
good news that there's another tool,another arrow in the quiver that could
reduce further losses to South Korea.
The bad news is that might bethe useful tool that gets us
into a war that none of us want.
The other piece of bad news is theoverarching argument here that we're
(28:24):
making is in our minds, not just aboutthe artillery balance in Korea, but really
about the balance of power in Korea.
And as Nick kind of suggested earlier,experts who've watched the Korea for years
have known that the North Korean militaryin general is not what it appears.
It's a lot weaker.
It poses a much lower threatthan it appears on paper because
(28:47):
it's equipment is old, it doesn'ttrain very much, et cetera.
But North Korea's two Trump cardshave always been the artillery and the
threat in Seoul and the nuclear weapons.
And what we're basically findingis that that artillery is as much
of a paper tiger as the rest of theconventional North Korean military.
And as a result, if God forbid, we findourselves in a serious war on the Korean
(29:11):
peninsula, very soon into that war,the leaders in Pyongyang are likely to
see a conventional military force oftheirs that's collapsing, an artillery
force that's been disarmed, and they'regonna have the nasty choice of either
accepting the fate that Saddam Husseinfaced, or being pressured to find
(29:32):
some way of escalating and coercing anend to the war with nuclear weapons.
And so, like all these things,there's good news, there's bad
news, and they're often intertwined.
North Korea's profound weakness,not only in its main conventional
forces, but also in itsconventional artillery is good news.
The alternative would be worse news.
(29:52):
But do understand that with thatgood news comes with quite serious
pressures on North Korea to escalateto the nuclear level, should war come.
Sheena Chestnut Greitens (30:01):
You know, one of the things that I think is
really important about this articleis that it identifies a set of risks
that are just different from what theconventional wisdom commonly focuses on.
And this is one of the things that Ireally liked the first time I read this
piece, and one of the reasons why wethought it was important, and we're
glad to have the chance to publish itin the Texas National Security Review
(30:23):
is that, I don't in any way in readingthe article or listening to you talk
about this, take away that you'recavalier about the costs of an artillery
exchange between the two Koreas.
It's much more that actually mostpeople in hearing, well, the North
Korean military is conventionallyweaker than outside analysts think
would think, wow, that's great.
(30:44):
That's great news.
And, there is a part of that that's good,although having interviewed some former
members of the North Korean militarywho were hungry during their military
service, there's also real human costthat sort of undergirds that military
weakness that has its own just sortof human level tragedy and cost to it.
But, what I think is really importantis this part that you're emphasizing
(31:09):
now, which is that something thatwe might regard as a positive, North
Korea's conventional military beingweak and less able to threaten Seoul or
the citizens of South Korea, Americanresidents, or international residents
of South Korea than we might think—that all sounds good— but there's a set
(31:29):
of risks even to the good news, right?
North Korea in the past couple yearshas for the first time explicitly
laid out conditions under which itmight use nuclear weapons first.
And so, one of the things that I reallyappreciate is the tying in of this
conventional threat, the artilleryrisks and costs of an artillery
exchange with the risks of nuclearescalation on a peninsula that has long
(31:53):
been one of the world's most seriousflashpoints, so I guess I wanted to
ask if there's one thing or a fewthings that you wanted policymakers
to take away from this article?
What would you want them to walkaway knowing or taking back to their
teams to think about and work on?
Nick Anderson (32:09):
I'll take a stab at that first.
So I think probably if there's kind of oneoverarching, policy relevant finding of
the article, it's that the North Koreanconventional military is weaker in all
of its aspects than we have thought.
And I think a lot of that has beenunderstood, but artillery has stood
out as this part that was particularlystrong or threatening still, and we're
(32:31):
finding that's no longer the case.
And if that's true, then a lot of whatwe're seeing, in terms of what North Korea
is doing maybe makes more sense, right?
The sort of doubling down on thenuclear capability under Kim Jong-Un.
It's really taken off since he tookpower in the 2010s, and as you said,
this change in North Korea's doctrine in2022 to allow the first use of nuclear
(32:54):
weapons— maybe they would've donethat anyway— but now it's out there.
It's sort of explicit in cases wherethe regime or leadership is threatened.
Trading away millions of artillery shellsto Russia in Ukraine in exchange, as
far as we understand, for technologythat will aid the nuclear program,
pushing really, really hard towards,an ICBM capability and the ability to
(33:16):
strike the continental United States.
All of this makes a little bitmore sense if you understand
that they don't have as strong aconventional deterrent as we thought.
And so, going forward, I thinkNorth Korea is gonna just be in more
situations in which it will relymore heavily on nuclear weapons.
Hopefully that doesn't includeuse, but sort of this is like in
some ways the kind of last leg.
(33:37):
And if that's true, the odds of themabandoning this, as long as relations
remain as they are now are very, very low.
The expectation that they may givethese up someday for some sort of
deal or grand bargain, I think theresults of our research suggests
that that's unlikely to happen.
Daryl Press (33:51):
I agree with everything Nick just said.
I would add two things.
If I was thinking, what are the twoflashing red to-dos that come out
of our research for policymakers?
I'd say the first one is largelyfor South Korean audience, for
South Korean military audience inparticular, which is to say your
society has already done the hard part.
Your society has already found the moneyand found the will to buy the weapon
(34:14):
systems that are necessary for you todo this really difficult thing, namely
protect Seoul, which is nestled rightup next to the North Korean border,
and yet you now have the equipment todo it, but the equipment is part one.
Because the mission is really difficult.
It's about rapidlyidentifying incoming shells.
It's figuring out the origin of thoseshells, distributing targets in a chaotic
(34:38):
environment very quickly to shooterstargeting, doing bomb damage assessment.
None of that is simple, and especiallyin the chaos of war, that therefore what
South Korea really has to do now is doubledown and triple down in the training,
and kind of the soft work that has tohappen underneath this— about trying
(35:00):
to practice doing this in a chaotic,difficult circumstance and getting
as good as you possibly can becausethey have to succeed the first time.
And so I'd say the firstthing is congratulations.
South Korea's bought the stuffthat will let them defend Seoul.
Now they have to do the hardwork of actually preparing to
execute a difficult mission.
The second thing is a takeaway both forUS and South Korean leaders, leaders
(35:23):
at the political level, leaders atthe military level, which is this
challenge of intra war deterrence.
In Ukraine we're wrestling withhow much can we push back Russia in
territory they've seized before youstart getting toward Russian red
lines and there are views all overthe map about this and big debates.
On the Korean Peninsula, it mightbe much, much quicker and much,
(35:45):
much more dire, which is if thingsgo the way that Nick and I expect.
If there is, God forbid, a majorconventional war in the Korean
peninsula, we really do not thinkthe North Korean military is gonna
survive as a functioning force for, Idon't know, more than a week, 10 days.
And very rapidly, the North Koreangovernment is gonna be facing life and
death decisions about how they make usstop, and they don't have a whole lot
(36:11):
of tools other than nuclear weapons.
So I think what this means forUS and South Korean leaders is
we have to be thinking really,really hard now about A: how we're
gonna fight that conventional war.
Sure, defend Seoul fromincoming artillery.
What are the things in our normalmilitary toolkit that we are
(36:32):
gonna do and not gonna do for thepurpose of preventing escalation?
How are we gonna message that war?
What are we gonna be telling North Koreanleaders and other observers real-time
about what our objectives are and are not?
And fundamentally, how we're gonnacome to grips, not just as the United
States, but as an alliance with how muchdamage we want to inflict and how much
(36:55):
success we want to achieve in this warto balance the operational objective
of defending South Korea and thwartingan attack, and the strategic objective
of all of us getting out of this warwithout a significant nuclear exchange.
And so I think both the training,training, training conventional
(37:18):
job is really job one forthe South Korean army today.
And the thinking about intra warescalation is job number one for
strategists at the US strategiccommand, at CFC, at USFK and in the
South Korean military and leadership.
And they are flashing red topics of justas great importance as the dangers we
(37:42):
face in the Taiwan Strait or in EasternEurope, or now in the Middle East.
Sheena Chestnut Greitens (37:47):
I found that point actually really striking.
I'm glad you brought up the compressionof time that decision makers will have
to think about these escalation risks.
And that intra war deterrence aspect whenit's compressed in time, I think could
be, as you said, a flashing red light.
So I'm glad you brought that in, andI'm glad you phrased it that way.
Daryl Press (38:06):
That's great, and hey, I don't know if this belongs in this podcast
or not, but it is so nice have TNSR andits particular voice in terms of the
topics that it covers as an outlet for notjust Nick and me, but for our community,
all of us, and to be a place that'sopen to publishing campaign analyses.
I mean, there aren't very manyplaces out there that see this as
(38:29):
an important part of either foreignpolicy analysis or scholarly research.
And so, having TNSR as a placewho sees the importance and the
contributions of that style of workis a tremendous benefit to the field.
Sheena Chestnut Greitens (38:43):
Well, I think your article's a great example of why
we at TNSR find it valuable and whywe wanted to put it out there as a
contribution for the security studiesscholarly community, but also for the
American and international strategycommunity and the community that thinks
about these risks and these problemson a daily basis as professionals
and as practitioners, so we'rereally delighted to have you both on.
(39:06):
Thanks so much for joining ustoday to talk about your article.
Congrats on having it out and, again,really pleased to feature it in TNSR.
Thanks for being with us today.
Nick Anderson (39:14):
Thanks for having us.
Daryl Press (39:15):
Thank you.
Ryan Vest (39:17):
Thanks for joining us on Horns of a Dilemma from the
Texas National Security Review.
Our guests today have beenNicholas Anderson and Daryl Press,
authors of "Lost Seoul (39:24):
Assessing Pyongyang's Other Deterrent,"
which as always can be accessedfor free on our website, tnsr.org.
If you enjoyed this episode,be sure to subscribe and leave
a review wherever you listen.
You can find more of our work attnsr.org and in the show notes.
Today's episode was produced byTNSR Digital and Technical Manager
Jordan Morning and made possibleby The University of Texas System.
(39:48):
This is Ryan Vest andSheena Chestnut Greitens.
Thanks for listening.
Popular Podcasts
Fudd Around And Find Out
UConn basketball star Azzi Fudd brings her championship swag to iHeart Women’s Sports with Fudd Around and Find Out, a weekly podcast that takes fans along for the ride as Azzi spends her final year of college trying to reclaim the National Championship and prepare to be a first round WNBA draft pick. Ever wonder what it’s like to be a world-class athlete in the public spotlight while still managing schoolwork, friendships and family time? It’s time to Fudd Around and Find Out!
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.