Forces of Darkness: Territorial Control & Resource Allocation
start the webinar. The Pearson Institute: Here we go. The Pearson Institute: Good morning, everyone Thank you so much for joining us today for our February edition of the Pearson lunch and learn we're excited to have Scott gates with us. The Pearson Institute: From prio we are going to give it just a couple of minutes and let some more folks log on before we get started with today's program again, I thank you for being here and I hope you will join us in future for more Pearson institute events.
The Pearson Institute: Okay we'll go ahead and get started folks will be joining us i'm sure, but again, thank you for being here for the February edition of our lunch and learn series. The Pearson Institute: We have Scott gates with here at the invitation of our institute director James a Robinson very excited to have you and excited for this talk. The Pearson Institute: And we will start the program off with a introduction from our Pearson fellow Sabrina fields and then from there we'll hear from Scott gates himself Thank you all so much. Sabrina Fields : Good morning, everyone, my name is Sabrina fields i'm a first year NPP student at Harris, and a fellow here with the Pearson Institute. Sabrina Fields : When I learned I would be spending a good portion of my first year here at the University of Chicago virtually I know it requires some adjustment of my expectations. Sabrina Fields : I expected to miss out on the face to face interaction with my peers and my instructors, but what I didn't expect was the silver lining of our extended global reach.
Sabrina Fields : This year at the Pearson institute we've heard from experts speaking with us from Sierra Leone Columbia Syria cutter and today, all the way from Norway. Sabrina Fields : I look forward to when we can safely hold events in person i'm very excited that we have the opportunity today to hear from a renowned expert from one of the world's premier Peace Research institute's joining us from Oslo. Sabrina Fields : Start Dr Scott gates is currently a research professor at the Peace Research Institute Oslo, also known as. Sabrina Fields : A professor of political science at the University of Oslo and I guess researcher at the Center of equality, social organization and performance at the University of Oslo. Sabrina Fields : He formerly served as Director of career Center for the study of civil war which was designated as a Norwegian Center of excellence, funded by the Research Council of Norway for a 10 year period from 2000 to 2012.
Sabrina Fields : He received his PhD in political science from the University of Michigan. Sabrina Fields : Alongside john brown he is the Co recipient of the 2014 Herbert Simon Award for research on bureaucratic politics, he is also a member of the Norwegian Academy of science and letters and their. Sabrina Fields : Royal in our region society of science and letters his current research includes applied game theory analysis international relations theory. Sabrina Fields : and national political economy formal models of bureaucracy and economic modeling. Sabrina Fields : He applies these methods to issues ranging from armed conflict and civil war in the Middle East and South Asia to social media recruitment by extremist groups. Sabrina Fields : Today he joins us to discuss the strategic dynamics of territorial control and research allocation during civil war, please join me in welcoming Dr Scott gates.
Scott Gates : Alright, thanks for a nice introduction i'm happy to be here, I actually wish I was in Chicago, but we have very cold day here in Oslo, but no snow. Scott Gates : i'm going to share my screen and. Scott Gates : Go so there. Scott Gates : I want i'm going to do is present a paper called forces of darkness Islamic state armed conflict and resource allocation it's co written. Scott Gates : By siri roost also at prio and Chris Butler who is at the University of new Mexico and, in addition to pre oh i'm also at the University of Oslo as Sabrina mentioned.
Oh. Scott Gates : There so as an overview what I wanted to do today is i'm going to look at the strategic dynamics of territorial control and resource allocation during civil war. Scott Gates : And, in particular, what i'm going to do is i'm going to look at the basically the rise of Islamic state in Syria and in Iraq. Scott Gates : Islamic states also called dash icicle ISIS i'm going to mostly just refer to it as Islamic state or is just for keeping things simple and that's the ISIS flag.
Scott Gates : Just as a means of overview there's not too much systematic formal or quantitative analysis, what happens during conflict most analysis is what Why do wars happen what causes war and what happens after war. Scott Gates : But there's more and more working on what happens during war and a lot of that speak in on looking at the duration of war and. Scott Gates : Evan flow and battle deaths and things like that, and what i'm going to do here is going to focus particularly on the ground game what goes on territorial. Scott Gates : i'm going to employ game theoretic modeling and the model i'm going to borrow from and develop is contest success function technology, and from that model i'm going to divide a set of hypotheses, which all test statistically. Scott Gates : And in particular i'm going to use data on nighttime light you missions, and these are satellite images and then we're going to take those images and projected onto. Scott Gates : A data set um a GIs data set called prio grid, which basically takes the whole world and maps it into grid cells that are half a degree.
Scott Gates : squared and then we then project different data onto each grid cell This allows us to examine the local aspects of conflict and a sub national level of analysis. Scott Gates : And the other thing that we're going to do here is there's a big problem in studying economic activity and the strategic logic of attacks and control of territory. Scott Gates : And that is that the relationship is incredibly indulgent eyes, so the arrows are going to and from the dependent variable and to help address the problem of indulgent at. Scott Gates : we're going to use a course and matching approach that i'll talk about later that creates a quasi experimental design to help us account for this indulgent entity. Scott Gates : So the research questions that i'm going to look at our how to rebel groups allocate the resources for fighting and productivity. Scott Gates : And how did these factors affect the strategic dynamics of armed conflict in the location of battles.
Scott Gates : and, more particularly, what should we expect to observe and, by the way of nighttime light emissions in areas attacked by Islamic state and in areas of is control and the regarding the change and control. Scott Gates : What motivated This project was a debate that occurred in 2016 and it was a paper written by Hansen Lewis and. Scott Gates : JEREMY Shapiro that appeared in 2016 and it came out about the same time in in much in reaction to a UN UN report on how strong Islamic state was UN report was extremely hyperbolic and claiming that it was the richest and most powerful rebel group that ever had existed. Scott Gates : I wasn't exactly agree, in agreement with that and the other extreme, was the Hampstead Lewis and Shapiro argument, who claim that Islamic state's strengths were greatly exaggerated. Scott Gates : Their argument was based on nighttime light emissions in which they were looking at the men means of control and who is showing where. Scott Gates : I asked areas were darker and from that darkness they concluded that it was much weaker than even the weakest nation states and was hardly a strong state.
Scott Gates : The problem with that analysis or conclusion is that the logic of contest success functions, we should not expect the asymmetrically weaker party to devote more resources to fighting effort than other users. Scott Gates : So. Scott Gates : here's a little vignette mode as part of the motivation say Islamic state captures some diesel fuel.
Scott Gates : Does Islamic state use the fuel to run generators or their humvees as I show in this picture here similarly if they capture other resources and especially for Islamic State while. Scott Gates : which can be sold to buy weapons, so the question we turn to instead of guns and butter, we look at light or fight now keep in mind Islamic state controlled areas were not part of the electrical grid maybe some parts were but. Scott Gates : A lot of these areas were kept out and it's due to systematic deprivation, especially in Syria, of the Sunni territory.
Scott Gates : Power is generated by generators, like the one I have a picture of here, and even in non conflict zones if you've traveled around in poor countries. Scott Gates : This actually lack of infrastructure is a big problem and a lot of electricity is generated by such diesel generators so basically if you've got diesel fuel, you can put it in the generator like this, or you can put it in a humvee that. Scott Gates : serves as a means of fighting and increasing your fighting capability. Scott Gates : there's some relevant literature and i'm going to ego is typically focus on my own, without a few others in the mix. Scott Gates : One of the first statistical analyses of territorial dimensions of conflict is a article that I wrote with Harvard boo hope. Scott Gates : That appeared in 2002 and we focus on border areas and the interaction between the scope, meeting the size of the area being held, and of conflict zones and the location where they were centered.
Scott Gates : In 2009. Scott Gates : Lula joined of our and I and we looked at geography and the duration of conflict, we again looked at periphery issues projection of force and use of resources, and indeed the model that I developed for that paper i'm going to be returning and showing it in this presentation. Scott Gates : hammond and. Scott Gates : Does some work on strategic location and fighting in particular, focusing on valuables junctions and networks and Korean and start body in.
Scott Gates : Have a really nice article on state capacity and insurgency looking at the territorial aspects of rebellion with regard to the state's capacity and its geographical distribution of its capacity. Scott Gates : And then there's a most recent article in the special issue of the journal of peace, research, for which I am the editor and it's article by Teresa under. Scott Gates : which by the way, one the visual presentation prize for the 2020 volume and she looks at territorial control and civil wars and looks at the types of activities. Scott Gates : Whether it's terrorist or conventional tactics being employed depending on territorial control now let's turn to the model, I wanted to develop.
Scott Gates : So I start with a basic contest success function which basically is saying that the probability of victory, which is designated by this pie. Scott Gates : And that's going to be the capabilities military capabilities of our the rebels minus the capabilities of the government and we simply create a ratio of the capabilities. Scott Gates : For the rebels versus total sum of capabilities for all the parties so military capability depends on on some unspecified combination of. Scott Gates : troop size military budget technological sophistication all the things that we think that are constitute military capability. Scott Gates : And for all levels of military capability power is going to decay, as its projected across distance, this is something that the Economist Kenneth bolding. Scott Gates : tells us about back in 1962 and the rebels capability or Islamic states in this particular.
Scott Gates : example of what we're going to do today can be expressed in terms of this capability for the rebel is going to depend on their location and then their allocation of resources into military military force or the allocation of resources to military and then the location of the fighting. Scott Gates : And, which means the military capability, the rebels is a function of the distance between the site of a battle ax a and the rebels stronghold X are and the military efforts of the rebels and the closer the rebels are to their own base the higher their probability of territorial conquest. Scott Gates : All other things being equal, now we can express and do some transformation, by taking the log of the difference between. Scott Gates : That location of the battle and the rebel base and we can put it in the logistic form which then takes the ratios of military.
Scott Gates : expenditures here and then also looks at the distance now what's nice about this function is that exhibits the same logistic principles as logistic regression does. Scott Gates : Here we have the probability of victory for the rebels being equal to these relationships, which is the ratio of the military expenditures the rebels in the government and the distance. Scott Gates : That the rebels are from the battles site.
Scott Gates : Now, in terms of rebel APP resource allocation rebel resources are mostly allocated, meaning that there's no endogenous allocation of them, but then, once they have those resources they can put it into productive allocation. Scott Gates : Which is in the normal guns and butter kind of model into butter or into guns or fighting and productive effort and. Scott Gates : butter in our case is going to be measured in terms of light emissions, which is going to be productive effort.
Scott Gates : And so we basically are going to look at the decision that goes back to my motivating case with the capturing little vignette about diesel fuel is do they spend it on guns or butter or is it fight or light. Scott Gates : Now we can also think about how there's going to be the allocation by both the government and the rebels, and this is going to constitute if we only have those two groups and society. Scott Gates : And the rebels controlling some territory and government controlling some territory there's going to be a total income for the country that will be.
Scott Gates : The some with some technological some technological multiplier, in turn, we can look at the rebels share and that's going to be the proportion of. Scott Gates : Total and similar for the government now if we differentiate equations six, which is expressed here. Scott Gates : What we see is an interesting phenomenon, and that is the marginal payoff of productive activity is directly related to fighting and as Alpha and. Scott Gates : epsilon go to zero, so does the partial derivatives now, in contrast, the marginal payoffs for military capability, by taking the. Scott Gates : Derivatives here remains positive and even when resources are declining as the rebel resources approaches zero a weaker opponent, will continue to fight, as long as they can, in other words.
Scott Gates : The rebels are going to take an increasingly move resources from productive activities towards fighting capability, they are going to devote more and more resources to fighting the relatively weaker, they are would given the government string. Scott Gates : As long as they can see keep fighting keep fighting but what's going to expected in this relationship from the modeling is, we should expect in an asymmetric relationship. Scott Gates : Islamic state will dim the lights and they will pour their resources into military fighting capability, we can come up with. Scott Gates : A proposition that derives straight out of this modeling, and that is conflict take actors will make allocation decisions between fighting and productive effort. Scott Gates : In response to the allocation decisions of other conflict actors strategic at calculations these allocation decisions absolute resources matter less. Scott Gates : than actors resources relative to the resources of its opponent, so that goes back to the formal modeling it's going to be the relative ratio.
Scott Gates : Of those expenditures that's going to be important and the proposition formally is as rebel resources decrease relative to government resources. Scott Gates : The rebel group invest more of its resources in fighting up to a point where all rebel resources are used for fighting. Scott Gates : The corollary, is that as rebel resources decrease relative to government resources, the rebel group invest less of its resources in productive effort, up to a point where no proto state that would be a rebel group controlling territory. Scott Gates : has nothing left for productive effort. Scott Gates : We can take translate these propositions into hypotheses that i'm going to test with with data and we're going to look at nighttime late emissions. Scott Gates : As a measure of productive effort and the first hypothesis one eight is the territory attacked by Islamic state will have greater nighttime light emissions than territory controlled by Islamic state.
Scott Gates : In other words, Islamic state will aim its military capability towards conquest of territory that is rich in resources. Scott Gates : Moreover, one be hypothesis is territory controlled by Islamic state will have less nighttime light emission than territory not controlled, because once. Scott Gates : Islamic State takes over territory it's going to divert the resources away from productive activities, meaning less light dimmer light projection not as much nighttime light emissions and they will be using that fuel to fight and not light. Scott Gates : So, in terms of our analysis, there were three phases of war with Islamic state.
Scott Gates : There was the rise, the period of control and the fall i'm going to focus mostly on the rise and control the fall of I asked comes after interventions from Turkey, Russia and the US. Scott Gates : And problem with doing our analysis, then, is the Russian and American bombings at a new dimension that we haven't figured out a way to handle that yeah. Scott Gates : So the focus here is going to be on the period of rise and control in Iraq and Syria, the rise of Islamic state from 2013 to 2016. Scott Gates : Now, what we do is we download satellite imagery like this, and here is Iraq that's Baghdad. Scott Gates : Here is.
Scott Gates : This is Damascus, Syria and basically along here on the Euphrates river is where you see this is Islamic state. Scott Gates : Control territory and then up here towards Kurdish El toro cherries, the This is another area, so we basically the Kurds. Scott Gates : Are along the northern borders of both Syria and Iraq and just at the south of them were the Islamic state. Scott Gates : controlled areas, what we do is we take this nighttime little lady mission and we projected entrepreneurial grid and program that, as I mentioned before, is a. Scott Gates : divides the whole world into the cell structure of one to one half degree by one half degree and then we're able to digitalize the measures, this is extremely time consuming, this is really takes a couple days actually to download this on a normal.
Scott Gates : laptop and then digitalized and then what we did was we took maps created by the Institute for the study of civil war or institute for studying work. Scott Gates : And they did detailed maps i'm showing where attacks occurred, where Islamic state control was and then where there was a broader level of support. Scott Gates : We take those that data and we intern digitalize it, and so we have and we digitalize it in a much more refined way we take the quarter of a prio grid cell. Scott Gates : And then we put in those are the areas, the cells that are being attacked, these are cells under control, and these are controls of support, we did some robustness checks, but mostly what i'm gonna do is i'm going to focus on control and attack by Islamic state.
Scott Gates : So the unit of analysis is going to be the quarter cell of a prio grid. Scott Gates : The treatment is going to be in a quasi experimental sense is going to be a quarter cell grid that experienced Islamic state attack or is. Scott Gates : As experienced Islamic state authority or control the experimental control is going to be that should say no Islamic state attack and no Islamic state authority. Scott Gates : And what we do with this notions of treatment and control is we match them and we sort the matched Q cells, one that is treated and one that's not so we're able to make a comparison. Scott Gates : And it's very important to doing this that the matching our creed treatment, the war and what what went on during the war should have no. Scott Gates : bearing on the matching itself it occurs before, and in this way we kind of set up a quasi experiment it's not a natural experiment and it's not a controlled experiment, because the treatment is not randomly assigned.
Scott Gates : Islamic state as a strategic actor and it controls it made a decision as to where to attack and what areas to control. Scott Gates : it's a strategic actor, but we can treat these non random assignments, and in a quasi experimental manner by matching in a pre with pretreatment data. Scott Gates : So what we want to do is want to get matching control group as closely matched to the treatment group.
Scott Gates : We use a technology called course and exact matching, which is a powerful monotonic imbalance reducing matching method that accounts for potentially confounding influence of pretreatment factors, and it was developed by Gary king and co authors. Scott Gates : And what we did was we match on covariance to predict either Islamic attack Islamic state attack or Islamic state territorial authority and. Scott Gates : What we do, then, is we estimate unbiased causal effects by ensuring that the control group, those are the grid sales, remember that are not attacked or not controlled by Islamic state. Scott Gates : And those are matched to the treated cells, those are the ones that are been attacked or controlled, and we want to get them as close as possible. Scott Gates : So what we did was we sort them in these three in terms of these three variables in terms of their pre war nighttime light emissions in terms of the log of the distance from the queue sell to the nearest main road. Scott Gates : And that's going to be either to Baghdad or to Damascus, depending on whether they're in Iraq or in Syria, and these are calculated by prio grid calculations and, in particular by relating to an article by tollefson and Strand and boo hug.
Scott Gates : And then we look, and this is really, really important because it's an most important resource that we're talking about and that's oil production in the queue cell, so this is mostly going to be the inherent. Scott Gates : oil reserves that are going to be in a queue cell, and this is going to be incredibly important for understanding and note it's the pretreatment so we're matching Q cells that would have similar capable capacities in terms of each of these variables. Scott Gates : Now the nice thing is, we create these matrices and where we get.
Scott Gates : It is imbalanced but that's where the course and matching helps us and that is, we get 9000 total that are no attack and then Islamic state attacks about 1500. Scott Gates : And we get a matched set and unmatched we have many more unmatched. Scott Gates : Here, but proportionally actually we have fewer here, and then, when we look at control, we actually have almost none drop out for Islamic state control, but we have many we have many more that drop out in terms of the the control the not treated.
Scott Gates : sample. Scott Gates : And then, what we run is a course and exact matching fixed effects ordinary least square model so it's a regression but it's using and taking into account these matches and with effects defects. Scott Gates : We run two tests, the first is looking at attacks and what we find is, we have a number of control variables we looked at suny controlled areas distance to the capital and population. Scott Gates : None of those are statistically significant, the only factors that are statistically significant is. Scott Gates : The with nighttime white emissions is an attack, so the brighter the cell, the more likely that it will be attacked and cells that are close, so this is a spatial lag what we do is we employ.
Scott Gates : What is called a kings and queen like structure, and this is borrowed from chess where a king or queen can move in any move. Scott Gates : diagonally or horizontally or vertically on the grid, and so what we did was we played around with different leg structures, a one cell. Scott Gates : lag structure would be a kings and then queen would be multiple and what we find is we get very high level statistical significance for this meaning basically that. Scott Gates : Close to Islamic states grains they're going to attack the quote the cells that have the greatest amount of the greatest light emissions indicating resource rich once the we look at their control and what we find is a dimming occurs. Scott Gates : And we also did some subsequent analysis looking at change in status and we find confirmation here that a change in status leads to reduction in light time light. Scott Gates : Light emissions when Islamic State takes control, and we also see a effect the specialty related I can show this in the coefficient plots and what we see is in control clear affects here, this does not touch the zero line but it's quite you can see quite a.
Scott Gates : Large confidence interval on light emissions of the cells that are attacked. Scott Gates : This is probably due to the fact that we have an estimated strategic value of certain cells that might not be just. Scott Gates : they're being attacked for their energy or for their resources, and then we see again the spatial leg playing a very big role here. Scott Gates : In summary, the areas controlled by Islamic state has significantly less nighttime late late emissions than other areas.
Scott Gates : And areas not experiencing a change in Islamic state control has significantly more nighttime light emissions than other areas. Scott Gates : and energy resources and control territory where diverted away from lighting presumably towards fighting capacity. Scott Gates : We can't measure fighting a capacity, and so we have to look at only in terms of the reduction in evidence of productive use of resources, moreover, areas attacked by Islamic state where resource rich, especially relative to other.
Scott Gates : cells and the pattern of attacks the spatial lags indicates a strategic design to maximize resource capture and then conclusion, these patterns fitter theoretical expectations, as expressed in propositions one and. Scott Gates : The corollary actually not to regarding resources us by rebel groups so while Islamic state may have been a wealthy insurgent group. Scott Gates : The combined forces arrayed against it were wealthier, it was an asymmetrically weaker group and by our formal model, we should have expected it to dim the lights and ports resources into fighting capacity.
Scott Gates : In other words, Islamic state could not afford to invest much and productive effort which meant dimming the lights resources were diverted to keep fighting. Scott Gates : And later after the periods of our analysis, Turkey, Russia and the US intervene in Syria and Islamic state simply lack the resources to main control. Scott Gates : maintain control over their territory and now what we have is a few little pockets of territory and mostly what we have is activity in terrorist cells that are. Scott Gates : organized by Islamic state and Islamic state is still has a presence in Nigeria in Libya in Afghanistan so it's mostly what i'm talking about here is Islamic state in Iraq and Syria. Scott Gates : future work, what I want to do with this is, I really want to develop a dynamic theory of territory and armed conflict. Scott Gates : That means applying this type of analysis to cases where rebel groups controlled territory, like the LTE in Sri Lanka, or the mouse and the Paul.
Scott Gates : And then I would also like to do is a geo spatial analysis comparing the rule of territories against different types of warfare Islamic state. Scott Gates : was trying to establish the caliphate and it was controlling territory and was forced to engage in more conventional forms of warfare. Scott Gates : What it would be interesting, is to look at territorial control for a guerrilla based group and to explore the role of territory across different types of military technologies well that's the end of the presentation i'm open for questions and thanks for your attention. Alex Carr: Thank you so much. Alex Carr: For your presentation will now turn towards the q&a portion from our virtual audience so just as a reminder, if you have a question Please submit it via the Q amp a button at the bottom of your zoom window.
Alex Carr: When you submit your question, please indicate whether you would like to come on screen and ask your question via video or, if you would prefer to ask the audio only. Alex Carr: Or, I can ask the question on your behalf, so we will now turn towards our first question from our very own Pearson institute director James Robinson. James Robinson: Okay, can you hear me now. James Robinson: yeah.
James Robinson: hi Jim how are you. Good. Scott Gates : Thank you you're muted. James Robinson: By yes okay i'm here i'm here okay great Thank you very much for coming and nice to see you so so so I have yeah I had some thoughts and it's not not my it's not my area of expertise, but but.
James Robinson: And I like the mechanism, you know I like I like that I like the idea I like I like the mechanism, but I suppose. James Robinson: You know there's other mechanisms like the model, if I understood correctly, the model sort of focuses on this particular mechanism, which has this sort of counter intuitive prediction which you can look at in the data, but I suppose you could say you know, and if you're going to. James Robinson: Use some bigger theory, you know as you as your sort of suggesting then you'll want to look at other mechanisms, I suppose it that you know the immediate thing people would think about is the sort of. James Robinson: hearts and mind type mechanism, you know so you're you're you know you're taking over territory you're trying to consolidate territory. James Robinson: So, so you want to provide some sort of public goods or services so that people kind of think oh yeah it's not so bad being run by Danish or something like that so that's not in your model. Scott Gates : right but it can't be.
James Robinson: It could be in your model so and it's probably in some more general model, you have so I was wondering. James Robinson: You know okay so so so mad, but maybe there's some sort of heterogeneity here that you could exploit meaning. James Robinson: You know in some contexts, the mechanism your your model is focusing on is kind of really dominant. James Robinson: And in other contexts, the hearts and minds thing become more dominant so I was thinking, I was surprised in the regression that sonny.
James Robinson: I don't know quite how you're measuring that you know, but this Sunday, you know, because you know there's this idea, it seems that this is a kind of sunny. James Robinson: Movement in some sense stylish and that they had a difficult time controlling Shia. James Robinson: Areas so so if there's sort of it seems like you know if one had a model of hearts and minds it'd be interesting to think about like do I have to be nicer to sunny people or they sort of inclined to help me anyway or. James Robinson: Do I have to be nicer to shear people, so it could be that there's some sort of heterogeneity you could exploit to sort of, say, you know both of these mechanisms are going on. James Robinson: And I can sort of identify when the hearts and mind thing is going to be dominant because there's some comparative statics which say I really need to kind of be nice when there, she is because she is not really going to support me or maybe not. James Robinson: Again it's not my area so i'm not really i've you know i've my my knowledge is we haven't heard a good PhD student here who's working on this in political science i'm kind of at the graphically.
James Robinson: But i'm not an expert at all I don't know like what's your back so that was my so when you were talking I kept on thinking you know how does this, how does this what how'd you get these different mechanisms into play here. Scott Gates : Well, I think what you say is really intriguing and there are some. Scott Gates : i've actually been wrestling with this, because this is tell us a nice little type story. Scott Gates : And it elegant but it doesn't match up with my kind of broader understanding of counterinsurgency and insurgency in the success which is along you're what you're talking about how hearts and minds. Scott Gates : One and I frankly was surprised that the control variable on Sunni held territory was not coming out as very significant. Scott Gates : What we might need to do, though, in answering that is there might be a temporal dimension, and that is that early on, when they start taking over territory.
Scott Gates : Then they're really focused on those areas and, once you maintain control, out of a heart land of Sunni, then you start expanding and encroaching into more mixed population. Scott Gates : cells or areas of of each country, with more Kurds and coming into Kurdish territory and and things like that that's just one thought as to why that's not coming out statistically significant The other thing is. Scott Gates : In not a not a statistically rigorous of analysis is ours and not digitalized and no quasi experiment or anything but there's an agricultural economics article about that looks at looking at remote sensing in terms of. Scott Gates : Crop use and Islamic State territory improves. Scott Gates : it's growing of crops when they get taken over.
Scott Gates : So that's an indication of productive use and My guess is that's probably the biggest part of hearts and minds is keeping people fed in the territory. Scott Gates : it's not as clean as our model, but it would be interesting to further examine the greening of the Islamic State territory after they conquer and the area that it becomes greener. James Robinson: You know that I mean that's extremely interesting because you could imagine that.
James Robinson: You know they have different instruments and maybe you know it's good to provide electricity, but the opportunity cost is incredibly high because you need the fuel to run the humvees. James Robinson: and other things also other ways of kind of raising people's welfare actually more feasible, given the resource constraints I don't know how they're doing that is that irrigation and you know me being very Labor intensive so that people can men, the men. James Robinson: yeah so that would be really interesting to understand the mechanism. Scott Gates : And yeah.
Scott Gates : yeah and it doesn't contradict our set set central result as you point out, because the opportunity cost on the diesel fuel is I gotta run my humvee that's much more important than. Scott Gates : Running my generator, but there are it's not as simple as just the total resources it's a certain kind of resource yeah which is has that fun it's limited fungibility I mean if I get diesel fuel, what can I use it for I can use it to to power. Scott Gates : I guess there's some degree of selling in you know, for a large period of where i'm talking here, there was a lot of illegal sale of oil of oil and they'd capture oil wells, and then then sell it on the black market. James Robinson: I i'm just worried that i'm gonna i'm gonna over exploit my my my hierarchical authority here Alec can ask another question Alex or survey shut up and let someone else asked the question. Alex Carr: We have a few more in the queue, but I think we'd be happy to hear your additional question.
James Robinson: OK, so the other thoughts I had to you know talking of heterogeneity is that for about the last. James Robinson: Eight years I haven't actually ever written anything about it, but hopefully there'll be something soon i've been studying with to my colleagues this. James Robinson: To my Colombian colleagues this power militarism in Colombia, you know so so so so we realized that it, you know when I was on sabbatical in Bogota. James Robinson: And they were this whole transitional justice process was generating a staggering amount of information about how these paramilitary groups operated but, but when I say staggering I mean you know millions of pages of court documents you know, so this is not easy to kind of. James Robinson: kind of triangulate or whatever, but but we're getting close, but one thing that just you know that sort of relevant here I didn't know how how dice was organized but. James Robinson: But you know the Colombian paramilitaries was sort of organized into about 35 different blocks and within the blocks or multiple fronts so there's probably about 250.
James Robinson: fronts paramilitary fronts and there's enormous variation in something like hearts and minds, you know, so this huge variation in the extent to which they use violence. James Robinson: And there's enormous variation you know, some of them build roads and schools and hospitals and you know they were they provided all sorts of public goes and others basically just were like drug. James Robinson: gangs, you know so so so I wonder, you know also whether there's heterogeneity here I guess the idea I don't have dice was organized, but I assume it must have been the must have been quite a lot of decentralized command or strategies and. Scott Gates : tactics, except that they date discovered some or came across some documents. Scott Gates : captured a base that had been held by Islamic state, and it was even including these travel they had appealed for travel reimbursement and they.
Scott Gates : They were. Scott Gates : Rejection letters about how they hadn't weren't going to get their reimbursement, because it was frivolous travel and it wasn't related Islamic state activities, so there was a high degree of. Scott Gates : hierarchy and centralization and Islamic state, and I think that's one of the reasons they were going after Al baghdadi because they wanted this decapitate strategy. Scott Gates : To I think that there was much more centralized in its structure, than in the park. James Robinson: I think yeah although yeah the you know the fault with decentralized mean I was talking about the paramilitaries, but but. Scott Gates : Oh, the paramilitary sorry yeah.
James Robinson: No, no that's true also that the fault we're very decentralized so yeah I see that's interesting, yes, that that's that is a contrast that's that sounds a bit like these examples jake Shapiro has of Al Qaeda the how. Scott Gates : The Rocker time, yes it. James Robinson: Again yeah that is interesting so let's let's let somebody else Thank you let's let somebody else yeah. Scott Gates : So that was good to hear from you. James Robinson: Yes, nice to see you yeah. Alex Carr: Thank you so much, and we're going to hear our next question from one of our current students JESSICA Anderson she's going to come on screen and ask.
Scott Gates : hi. Jessica Anderson: hi sorry I had to be limited to switch the settings okay and but thank you so much for your presentation, it was fascinating my question is. Jessica Anderson: rather more narrowly focused than than the last couple I was just curious if you could expand a little bit on the conclusions you drew from the spatial legs. Jessica Anderson: I don't know if I noted it correctly, but I have that the spatial pattern indicated a strategy of resource capture and I was wondering if the idea is that. Jessica Anderson: The significant effects on the spatial legs suggest that is targeted not just like individual Q cells that were resource rich but places where it was clusters of Q cells that were. Scott Gates : that's right yeah.
Jessica Anderson: How. Scott Gates : And there and they're like in a chessboard. Scott Gates : So there because there's two different ways, you can do spatial legs.
Scott Gates : Either like in our structure which is based on the grid cell structure than it's less kings and Queens kind of structure if you're not locked into a grid cell and say you had it more based on. Scott Gates : villages and hamlet's and communities, then you would look at a network structure that would look at the distances between and would move were like network geographic network analysis that looks at those. Scott Gates : Like nodes and then it would take into account the size of the towns and things like that so there's a richer quality to doing stuff like that, but there's kind of a simple elegance associated with.
Scott Gates : The way we did it and the other advantage is exactly what you're talking about is that we're taking we're able to account for these clusters and they're more obvious. Scott Gates : In a in a cell structure, because the you see you're getting lots of the queue cells that are getting attacked at the same time. Scott Gates : And they're not they're not just pinpointing to where the oil wells are there, of course, taking over the broader territory, so it makes sense when you think about what's going on on the ground to. Jessica Anderson: Thank you so much.
Welcome Alex Carr: Thanks JESSICA our next question is going to come from Michael and he's going to ask via audio. Michael Tatone: hi everyone thanks for much for the presentations and very interesting i'm not a student, but I work at the university Chicago urban labs of the Harris school. Michael Tatone: And I just had a question of Professor if I wonder if it's possible or, if you have the data to speak to this to look at how the sort of the tech the control territory differs.
Michael Tatone: If it's close or not regime control territories and sort of thinking of if there's any competition between dies in the regime to the on this hearts and minds question of sort of competing to deliver sort sort of resources or services in these territories. Scott Gates : that's a great question. Scott Gates : Actually, for the most part, one of the reasons for Islamic states success is the both the governments of Iraq and Syria systematically under invest and exhibit more.
Scott Gates : government repression in the Sunni hill territories so generally there's not much of an effort to compete for hearts and minds. Scott Gates : In that way, but your question, I can broaden a little bit, and that has to do with strategic. Scott Gates : Things so in terms of airfields than we saw a lot of battle battles over airfields, and so there were some airfields. Scott Gates : That were close to the Euphrates or tigers rivers and these these sometimes would be periods of long and engaged conflict.
Scott Gates : It might change sides on different sides which actually makes a problem that I didn't talk about for statistical analysis but it's something we've addressed is the changing of a of a cell that's repeatedly being attacked and and and whatnot. Scott Gates : But in that regard we see a pattern of state trying to project its power well away from either Baghdad or from Damascus. Michael Tatone: Thank you very much, thanks. Alex Carr: Thanks Michael we're next going to go to Alejandro romer one of our other current students he's going to come on video. Scott Gates : hi.
Alejandro Roemer: hi professor and thank you, thank you for for this very interesting talk, my question is just regarding the fungibility of. Alejandro Roemer: diesel fuel sorry diesel fuel and i'm wondering if if the Islamic state can sell an expert these diesel, then it will be able to. Alejandro Roemer: invest those well that income in other resources be at something else, then just light or fight right or light and humvees and so, in that case, my question is.
Alejandro Roemer: You wouldn't be able to see them to see that as an outcome by by just looking at other nightlight emissions, because the absence of light at night light emissions. Alejandro Roemer: couldn't mean that these resources are also being used for other productive activity or four providing public goods and services as. Alejandro Roemer: Professor Robinson Michael were mentioning and so basically my my question is that does is the model, accounting for this is it and ludicrous to think that the Islamic state is also exporting some of the soil and redirecting it to other resources, thank you. Scott Gates : that's a good question.
Scott Gates : um, I think, mostly what they were doing so there, whereas capture of this refined diesel there's also capture of of the raw oil resources, the raw oil resources, I can't take that and immediately put it in the diesel. Scott Gates : I can't put in either in my diesel generator and we're in my home the I can't use it as fuel so most of that's going to be sold and then they're going to buy diesel fuel but there's a lot of captured resources through a process capture processing plants and things like that. Scott Gates : Of course, they could sell it, but I think that mostly what you're going to do with fuel is you're going to use it. Scott Gates : In in terms of it, but I think we're making a big assumption here, and I think you're identifying that big assumption, and that is what we're doing is we're only measuring one type of productive output. Scott Gates : Most economy economics analysis, though, that looks at nighttime later missions, has provided some pretty convincing evidence on your shortlist done some really cool stuff on nighttime late emissions and piracy in Somalia new spider man.
Scott Gates : With co author oh Sebastian should they have an article which does systematic analysis of nighttime light emissions and economic activity, looking at Pakistan. Scott Gates : So there's some good evidence that it's nighttime later missions are highly correlated with productivity and productive effort, but it, you know I think you you're identifying a potential it's a really big assumption that we're relying on and that is the lack of productive. Scott Gates : or the lack of evidence of productive, economic output is a sign that you put it into fighting capacity and we we don't it's impossible to know, except that when they took out or we see a general pattern that fits with the model and it fits with the story. Scott Gates : And, but, of course, he will never really hundred percent know it's impossible it's just impossible to know so it's kind of a conjecture I guess. Scott Gates : To be blunt about it, but I think this goes back to Jim robinson's questions, and that is exploring the other kinds of resources and. Scott Gates : Where the tried up with trade offs occur, and what are the opportunity costs across the different kinds, I think, for.
Scott Gates : The diesel fuel, I think our argument is pretty strong and compelling that I think for other kinds of productive activities like the example I was talking about agricultural productivity. Scott Gates : That is probably not going to deter and, in fact, is probably going to enhance your fighting capacity by investing in the food production. Scott Gates : Thanks. Alex Carr: Thanks Alejandro, so we are approaching the end of the hour i'm Professor gates, we can make this next question the last question, or if you have a couple more minutes there's just a couple more questions.
Scott Gates : We could go a little bit over it's it's it's been fun. Alex Carr: Great thanks, we really appreciate your time, so this next question comes from billy Heller and he would like me to ask it. Alex Carr: On his behalf, so he says, while he doesn't have experienced studying ISIS or dash.
Alex Carr: He does have experience with studying the Taliban and he's curious if you considered waiting types of attacks conducted by ISIS based off. Alex Carr: Of the perceived value or effects of each type of attack If so, was there any intent to attempt to identify into indicators or warnings of ISIS operations such as staging to expand ISIS territory or sees infrastructure utilizing available geospatial or measurement and signatures data. Scott Gates : that's a great question to my knowledge, a. Scott Gates : afghanistan's going to be a lot different because US troops are there. Scott Gates : And us technology and in the period that i'm looking is before the US really got involved in in Syria, it was in Iraq, though.
Scott Gates : But always been a little bit more hands off in terms of what you're talking about no I don't have any information on the particular analysis that you're talking about, but it should be possible retrospectively to do some of it. Scott Gates : it's quite fascinating a the. Scott Gates : Quite really interesting about Islamic state attacks on not overtaking over oil fields, but over taking over a town or a Hamlet is generally.
Scott Gates : These towns are become fortified because the conflict and there's one gate, which is the entry gate that's available kind of for trade and people to come in and out, but all other. Scott Gates : gates to the old city or town are not open, but what Islamic state would do is they'd pack a couple cars, full of explosives and they would drive into the gate and then blow up themselves and. Scott Gates : All of the security guards at the gate and then with heavy back fire people laying down gunfire than their troops, the Islamic state troops would flood into the town and concrete and it was an extremely effective. Scott Gates : Strategy it's kind of emulates kind of medieval European siege warfare kinds of things. Scott Gates : it's kind of an interesting antidote anecdote but it's not something that we really focused on for this paper, but I do think what you're asking is very similar to the paper that I mentioned in the lit review. Scott Gates : The unders paper which was looking at different kinds of attack terrorism versus guerrilla types of warfare and, in our case, what we also have is this combination of.
Scott Gates : mean there's a suicide bomber but it's part of a conventional war tactics all mixed together within extreme effectiveness and it really isn't until us Russian and American bombs start falling and and one can't. Scott Gates : Over emphasize the role of Turkey entering plays a big role in the fall of fall of Islamic state, but I think that's something that there's a lot of possibilities for doing some good work in different conflict zones thanks for the question. Alex Carr: Great Thank you our next question comes from Alejandro albinus and he's going to come on screen and ask you. Alejandro Albanez: Oh. Alejandro Albanez: hi Scott good to see you so my question is related to dissipation before ISIS rose to power, so I believe so, in that case we don't have we don't have information about the territories that they were controlling because they didn't have any right so.
Alejandro Albanez: Right can can we have a user model in terms of well probably these places were connected to the power grid at that time before it to to control. Alejandro Albanez: And then, now we have the situation where both the Iraqi and the Syrian Government and Turkey are taking control of these territories, making it possible for this place to be connected again to the power agree. Alejandro Albanez: Is there a way to use your model. Alejandro Albanez: To do a pre assessment of these places, before I said to control them, and now that they're taking back by these governments. Alejandro Albanez: And if ISIS could take the same strategy to take over these territories again in the future so just like basically thinking about a predictive model using their methodology. Scott Gates : that's a really intriguing I haven't thought about that, but in looking at in terms of a.
Scott Gates : Post conflict cyrus during conflict but employing this methodology for post conflict analysis. Scott Gates : And then also is really would be super interesting in terms of hearts and minds kind of strategy to what degree, are these governments really trying to. Scott Gates : make life better for the inhabitants of those areas that used to be controlled by Islamic state and are or are they just just really draining them of all their resources that would be super intriguing and then that would have implications for the rise of another group.
Scott Gates : So, in many ways. Scott Gates : Islamic state was a challenger to Al Qaeda. Scott Gates : But at the same time kind of an ally kind of a challenger. Scott Gates : During the period i'm talking about here, though it was it was on its own it wasn't ally Yang with anybody, it was taking over vast amounts of territory. Alejandro Albanez: Exactly yes.
Scott Gates : yeah, but I think you're asking a super interesting question because you can do all of this type of stuff and look at post conflict environments anywhere, I mean that's really especially then that has intriguing aspects of policy implications for. Scott Gates : What governments are doing towards rehabilitation rehabilitating those areas that had. Scott Gates : been foreign or occupied by the rebels yeah. Scott Gates : Great question. Scott Gates : yeah I never thought about that at all that's good thanks. The Pearson Institute: hi Scott, thank you all for these wonderful questions God, thank you for taking some extra time over to answer some of those remaining questions.
The Pearson Institute: We again just are so appreciative that you joined us this morning, we are happy to host you and hopefully can host you again in future on campus at some point, but until then everyone be well keep safe, if you are. The Pearson Institute: interested in joining us again for a future lunch and learn our next is February 24 it's women of color advancing peace and security, featuring Ambassador bonnie Jenkins. The Pearson Institute: will be joining us that's February 24 at 6pm and so thank you guys again for being here Scott, thank you, you all have a wonderful day. Scott Gates : Thank you bye bye everybody thanks for the great questions.
Scott Gates : bye bye.