The below paper claims its model has accurate out-of-sample forecasting of systemic conflict propensity with a 10-year lead time. I.e., we can know right now if we're heading to WW3, potentially. The key variable is trade fractionalization. I.e.., we're probably heading to WW3 unless we can shore up the trade system very quickly. That makes Mark Carney arguably the most important world actor right now.
I've done analysis since and it's scary (and sad) how well that model performed quantitatively (against e.g. Uppsala data). We're expanding the approach now, maybe something will come of it. The gist is that monetary stabilization -- which also affects trade fractionalization -- is also important (something Carney also understands).
Both models integrate conflict and economy (rather than splitting in IPE and Security separately) into a dynamic model that does *not* assume observational independence, but instead takes structure seriously as a relevant "variable". Brandon Kinne also does this in his work on co-evolving networks.
I agree that taking old ideas off the shelf is a good way forward. Jervis is great. Mostly we just need to try. Journals don't care. So nobody tries. It's never going to be perfect, but we could probably do a lot better if we actually tried (instead of spending all our time on survey experiments), and can refine the models as new data comes. There are no institutional incentives to do this, however, so it isn't done.
The core problem is that IR as a discipline is premised on equilibrium rather than evolution. The "debates" are about equilibrium conditions. You don't need to forecast if you are building equilibrium models, because in equilibrium things don't change.
The journal system in academia is well documented as a failed institution since, what, the 1990s? Maybe someone should build a replacement. In math, there's ArXiV, but you still need a replacement for the curators who say "this is worth reading".
Most interesting, but more useful academically than prescriptively, as it seems that none of the big powers driving the current transition in IR are much interested in “good” foreign policy decision making.
"I really do think that climate change is an incredibly complex challenge that has multiple dimensions, where uncertainty is often quite high, and the stakes of decision-making are quite high". We were addressing that - until Trumpkopf arrived on the scene and screwed it all up.
I do agree. After all, Trump won after asking for $1 billion in donations from oil billionaires—while Houston was undergoing a fierce rainstorm. I’m in the middle of reading Mr. Farrell’s article, and so far I’m not getting a sense of theory of power and inequality (which is the oeuvre of the Marxists and world-system theorists and left-populist muckrakers) in it. There’s bound to be some sort of applicability in the paper to it
I'm much less schooled in IR (did some stuff in IPE), but I don't find your take convincing. Theory is not about prediction but levers. Of course, theory offers context(s), but uncertainty at that level is the norm. I don't work with a general theory of econ either.
In particular, I think the portrayal of climate mitigation is unhelpful. Granted, we are living in a pseudo-reality constituted by the widespread notion that what we have to do is manage an energy transition. This is an illusion for multiple reasons, as I showed in my climate book (Alligators). Really, what we have to do is forcibly limit the burning of fossil fuels, ideally through annually declining production permits, auctioning them off, and returning the revenue to the public. (Pay the increased price of energy to ourselves.) The rest -- adapting to such a regime -- is beyond the power of public management and requires enormous freedom of substitution. Subsidies to green energy production can help, but mostly this will be a complex societal adaptation. The political problem is that a big chunk of the capital stock built on the expectation of continued cheap and abundant energy will have to be written off. That's where political economy comes in. Global interdependence introduces further complexity, but there are tools like-minded countries can use for coordination. (Also in the book.)
Honestly, I despair at the increasing likelihood of a climate armageddon, due partly to the intransigence of the far right and partly to the misguided effort of green-minded people to make it all go down easily -- and avoid a showdown with capital.
Coda: the climate formula everyone ought to know is that the inverse of the years to exhaustion of a remaining carbon budget is the sustainable rate of emission reduction. 20 years? 5%. This is the benchmark that tells us how we're doing in avoiding disaster, not how much noncarbon energy we're using. We've got more wind and solar than ever, while emissions accumulate relentlessly.
There's a more fundamental problem. The defining premise of internatioanl relations as a discipline is that there are actors called nations and the relations between these actors determine world affairs. The different schools just have different models of the motives of these actors. That's no longer true - in most conflicts between governments, the domestic opposition hopes for the government to be defeated. And the gradual rise of the EU as a confederal international actor complicates things further.
💯- History and psychology are much more useful to explain events and catastrophes driven by leaders like Putin, Xi, Khamenei, MBS, Trump, Erdogan etc. than IR, which seems as useful as homeopathy.
Have been re-reading Richard White's "Railroaded" as an analogy to what's going on with AI development. White recounts the continuing struggle the transcontinental railroad developers had to acquire finance - money that they acquired under the excuse of building & maintaining the railroads but which in great part went instead into their personal fortunes. Blatant corruption of Congress, officials, international financiers & investors, epic waste, social & environmental disruption & destruction... A gripping tale but hard to take because of the horrors described.
The obvious conclusion is that, once again, the artery running through everything is finance & the institutions set up to both facilitate & control it.
Welp. Now what do we do. Watching with keen interest what happens in China under socialism with Chinese characteristics. Was the jerking of Jack Ma's chain a highly significant moment?
"Scholars like Hong and Page argue that groups with highly diverse viewpoints are better able to capture the contours of complex systems than groups with greater expertise but more similarity across their members’ understanding."
Immediately called to mind Philip Tetlock's "Expert Political Judgment: How Good Is It? How Can We Know?"
Up early and as usual checked my Inbox first thing and there you were.
Immediately started reading but had to get up for a coffee refill and during the couple of minutes that took I was struck by the realization that I was EXCITED, yes, EXCITED, to get back to reading your post which continued to thrill until the end!
" Plural Perspectives" indeed. What a lovely concept Jane Flegal: you go girl!
I took a course on systems analysis from Dana (Donella, a name she hated) Meadows in 1978. I cannot say I was her best student, but to all of us she was gentle and supportive.
I had the opportunity to play with World3, the program she brought us from MIT. Clearly, from my attempts, it was challenging to obtain a set of parameters that would not set the whole world crashing.
There was one sustainable solution the model supported, and is well described in “ Limits to Growth,”but this world hasn’t followed these recommendations these past fifty years.
What we have done in these decades was clearly delineated in the “Business as Usual” assumptions. This genie does not get put back in the bottle. There is no instant turnaround for us.
Thanks for the insider tip on Dana vs. Donella! Didn't know. From the 'Basics' section of her book "Thinking in Systems",
"A change in purpose changes a system profoundly, even if every element and interconnection remains the same."
"Changing relationships usually changes system behavior. The elements, the parts of systems we are most likely to notice, are often (not always) least important in defining
the unique characteristics of the system—unless changing an element also
results in changing relationships or purpose.
Changing just one leader at the top—from a Brezhnev to a Gorbachev, or
from a Carter to a Reagan—may or may not turn an entire nation in a new
direction, though its land, factories, and hundreds of millions of people
remain exactly the same. A leader can make that land and those factories
and people play a different game with new rules, or can direct the play
toward a new purpose.
And conversely, because land, factories, and people are long-lived, slowly
changing, physical elements of the system, there is a limit to the rate at
which any leader can turn the direction of a nation."
Until there's a change in intent and purpose (which, on a local level, is what Mamdani -- and perhaps other more centrist and less loquacious politicians [Michelle Wu?] -- seems to signal ...at least rhetorically), it's hard to imagine outcomes changing.
The move from testing futures against each other to generating incompatible but plausible ones feels right when feedback loops make prediction unreliable. It also suggests the real payoff is institutional: policy systems that can pivot, rather than ones betting on a single forecast. How would you tell, in practice, when a system has become adaptive enough rather than just hedging in every direction?
@Henry Farrell - Nit about "rotational velocity of the late Hans Morgenthau’s cadaver increases by another few centimeters per second"
The line is amazing, the units are mismatched.
Rotational velocity is about how fast an angular quantity is changing, typically radians or degrees. Assuming fixed axis rotation, of course. Centimeters would make sense for a the translational (i.e., non-angular) velocity at a given radius that a particular point is moving, but the rigid body's motion itself would be in angles. So something like a few more degrees per second. Except that's likely too large, but degrees are typically not made smaller using metric notation (e.g., micro degrees) and instead use degrees / minutes / seconds. So, likely, the most accurate restate would be another few seconds per second, with a footnote explaining the choice of units.
I wrote a program called Serious Scenarios in the early 1990s, based on cross-impact research by Selwyn Enzer and urging by Wayne Boucher, both colleagues of mine at The Center for Futures Research at the University of Southern California. It was written in Visual Basic for easy understanding. Its purpose was to generate a series of possible future trends and events, based on data and estimates of existing trends and events.
The core problem, of course, was the accuracy of the basic set of events and trends. The success of the program was not its ability to show dominant future trends but to demonstrate how little we actually knew about existing events and trends. As your article implies.
I quit improving the program in the face of far greater large number crunching capacity of later computers. Yet the believability constraints persist.
I contend that any model should be grounded in scenarios that start with considering governmental strategic initiatives based on two fundamental areas, incentives and regulations (carrots and sticks.) Human nature, at its core, is most strongly influenced by the opportunity to seize financial advantage. Essentially, government policies either manipulate domestic and international financial pursuits through incentives or regulations. The complexity comes from not just competing goals but a lack of continuity among how nations choose to achieve those goals. And there are also areas where there is an absence of any structured strategic approach - perhaps best defined as untethered free markets. But, that’s just the basis for analysis. It needs to be applied to the myriad of financial theaters, including energy, trade, and mineral and agricultural commodities. That’s complexity upon complexity many times over. But, at least it is grounded in what truly drives behavior.
The below paper claims its model has accurate out-of-sample forecasting of systemic conflict propensity with a 10-year lead time. I.e., we can know right now if we're heading to WW3, potentially. The key variable is trade fractionalization. I.e.., we're probably heading to WW3 unless we can shore up the trade system very quickly. That makes Mark Carney arguably the most important world actor right now.
https://www.pnas.org/doi/abs/10.1073/pnas.1509423112
Ore Koren and I inadvertently forecast current conflict levels 5 years ago with the model published here:
https://link.springer.com/article/10.1007/s12571-022-01300-0
I've done analysis since and it's scary (and sad) how well that model performed quantitatively (against e.g. Uppsala data). We're expanding the approach now, maybe something will come of it. The gist is that monetary stabilization -- which also affects trade fractionalization -- is also important (something Carney also understands).
Both models integrate conflict and economy (rather than splitting in IPE and Security separately) into a dynamic model that does *not* assume observational independence, but instead takes structure seriously as a relevant "variable". Brandon Kinne also does this in his work on co-evolving networks.
I agree that taking old ideas off the shelf is a good way forward. Jervis is great. Mostly we just need to try. Journals don't care. So nobody tries. It's never going to be perfect, but we could probably do a lot better if we actually tried (instead of spending all our time on survey experiments), and can refine the models as new data comes. There are no institutional incentives to do this, however, so it isn't done.
The core problem is that IR as a discipline is premised on equilibrium rather than evolution. The "debates" are about equilibrium conditions. You don't need to forecast if you are building equilibrium models, because in equilibrium things don't change.
The journal system in academia is well documented as a failed institution since, what, the 1990s? Maybe someone should build a replacement. In math, there's ArXiV, but you still need a replacement for the curators who say "this is worth reading".
Most interesting, but more useful academically than prescriptively, as it seems that none of the big powers driving the current transition in IR are much interested in “good” foreign policy decision making.
"I really do think that climate change is an incredibly complex challenge that has multiple dimensions, where uncertainty is often quite high, and the stakes of decision-making are quite high". We were addressing that - until Trumpkopf arrived on the scene and screwed it all up.
I do agree. After all, Trump won after asking for $1 billion in donations from oil billionaires—while Houston was undergoing a fierce rainstorm. I’m in the middle of reading Mr. Farrell’s article, and so far I’m not getting a sense of theory of power and inequality (which is the oeuvre of the Marxists and world-system theorists and left-populist muckrakers) in it. There’s bound to be some sort of applicability in the paper to it
I'm much less schooled in IR (did some stuff in IPE), but I don't find your take convincing. Theory is not about prediction but levers. Of course, theory offers context(s), but uncertainty at that level is the norm. I don't work with a general theory of econ either.
In particular, I think the portrayal of climate mitigation is unhelpful. Granted, we are living in a pseudo-reality constituted by the widespread notion that what we have to do is manage an energy transition. This is an illusion for multiple reasons, as I showed in my climate book (Alligators). Really, what we have to do is forcibly limit the burning of fossil fuels, ideally through annually declining production permits, auctioning them off, and returning the revenue to the public. (Pay the increased price of energy to ourselves.) The rest -- adapting to such a regime -- is beyond the power of public management and requires enormous freedom of substitution. Subsidies to green energy production can help, but mostly this will be a complex societal adaptation. The political problem is that a big chunk of the capital stock built on the expectation of continued cheap and abundant energy will have to be written off. That's where political economy comes in. Global interdependence introduces further complexity, but there are tools like-minded countries can use for coordination. (Also in the book.)
Honestly, I despair at the increasing likelihood of a climate armageddon, due partly to the intransigence of the far right and partly to the misguided effort of green-minded people to make it all go down easily -- and avoid a showdown with capital.
Coda: the climate formula everyone ought to know is that the inverse of the years to exhaustion of a remaining carbon budget is the sustainable rate of emission reduction. 20 years? 5%. This is the benchmark that tells us how we're doing in avoiding disaster, not how much noncarbon energy we're using. We've got more wind and solar than ever, while emissions accumulate relentlessly.
There's a more fundamental problem. The defining premise of internatioanl relations as a discipline is that there are actors called nations and the relations between these actors determine world affairs. The different schools just have different models of the motives of these actors. That's no longer true - in most conflicts between governments, the domestic opposition hopes for the government to be defeated. And the gradual rise of the EU as a confederal international actor complicates things further.
💯- History and psychology are much more useful to explain events and catastrophes driven by leaders like Putin, Xi, Khamenei, MBS, Trump, Erdogan etc. than IR, which seems as useful as homeopathy.
Have been re-reading Richard White's "Railroaded" as an analogy to what's going on with AI development. White recounts the continuing struggle the transcontinental railroad developers had to acquire finance - money that they acquired under the excuse of building & maintaining the railroads but which in great part went instead into their personal fortunes. Blatant corruption of Congress, officials, international financiers & investors, epic waste, social & environmental disruption & destruction... A gripping tale but hard to take because of the horrors described.
The obvious conclusion is that, once again, the artery running through everything is finance & the institutions set up to both facilitate & control it.
Welp. Now what do we do. Watching with keen interest what happens in China under socialism with Chinese characteristics. Was the jerking of Jack Ma's chain a highly significant moment?
"Scholars like Hong and Page argue that groups with highly diverse viewpoints are better able to capture the contours of complex systems than groups with greater expertise but more similarity across their members’ understanding."
Immediately called to mind Philip Tetlock's "Expert Political Judgment: How Good Is It? How Can We Know?"
https://www.amazon.com/gp/product/B00C4UT1A4/ref=kinw_myk_ro_title
Foxes FTW ;^)
Up early and as usual checked my Inbox first thing and there you were.
Immediately started reading but had to get up for a coffee refill and during the couple of minutes that took I was struck by the realization that I was EXCITED, yes, EXCITED, to get back to reading your post which continued to thrill until the end!
" Plural Perspectives" indeed. What a lovely concept Jane Flegal: you go girl!
And, of course, you too HF, thank you.
I took a course on systems analysis from Dana (Donella, a name she hated) Meadows in 1978. I cannot say I was her best student, but to all of us she was gentle and supportive.
I had the opportunity to play with World3, the program she brought us from MIT. Clearly, from my attempts, it was challenging to obtain a set of parameters that would not set the whole world crashing.
There was one sustainable solution the model supported, and is well described in “ Limits to Growth,”but this world hasn’t followed these recommendations these past fifty years.
What we have done in these decades was clearly delineated in the “Business as Usual” assumptions. This genie does not get put back in the bottle. There is no instant turnaround for us.
Thanks for the insider tip on Dana vs. Donella! Didn't know. From the 'Basics' section of her book "Thinking in Systems",
"A change in purpose changes a system profoundly, even if every element and interconnection remains the same."
"Changing relationships usually changes system behavior. The elements, the parts of systems we are most likely to notice, are often (not always) least important in defining
the unique characteristics of the system—unless changing an element also
results in changing relationships or purpose.
Changing just one leader at the top—from a Brezhnev to a Gorbachev, or
from a Carter to a Reagan—may or may not turn an entire nation in a new
direction, though its land, factories, and hundreds of millions of people
remain exactly the same. A leader can make that land and those factories
and people play a different game with new rules, or can direct the play
toward a new purpose.
And conversely, because land, factories, and people are long-lived, slowly
changing, physical elements of the system, there is a limit to the rate at
which any leader can turn the direction of a nation."
Until there's a change in intent and purpose (which, on a local level, is what Mamdani -- and perhaps other more centrist and less loquacious politicians [Michelle Wu?] -- seems to signal ...at least rhetorically), it's hard to imagine outcomes changing.
The move from testing futures against each other to generating incompatible but plausible ones feels right when feedback loops make prediction unreliable. It also suggests the real payoff is institutional: policy systems that can pivot, rather than ones betting on a single forecast. How would you tell, in practice, when a system has become adaptive enough rather than just hedging in every direction?
@Henry Farrell - Nit about "rotational velocity of the late Hans Morgenthau’s cadaver increases by another few centimeters per second"
The line is amazing, the units are mismatched.
Rotational velocity is about how fast an angular quantity is changing, typically radians or degrees. Assuming fixed axis rotation, of course. Centimeters would make sense for a the translational (i.e., non-angular) velocity at a given radius that a particular point is moving, but the rigid body's motion itself would be in angles. So something like a few more degrees per second. Except that's likely too large, but degrees are typically not made smaller using metric notation (e.g., micro degrees) and instead use degrees / minutes / seconds. So, likely, the most accurate restate would be another few seconds per second, with a footnote explaining the choice of units.
thanks for nitpicking and explanation - I half wondered whether I was screwing it up, but figured what the hell, it was a loose metaphor ...
I wrote a program called Serious Scenarios in the early 1990s, based on cross-impact research by Selwyn Enzer and urging by Wayne Boucher, both colleagues of mine at The Center for Futures Research at the University of Southern California. It was written in Visual Basic for easy understanding. Its purpose was to generate a series of possible future trends and events, based on data and estimates of existing trends and events.
The core problem, of course, was the accuracy of the basic set of events and trends. The success of the program was not its ability to show dominant future trends but to demonstrate how little we actually knew about existing events and trends. As your article implies.
I quit improving the program in the face of far greater large number crunching capacity of later computers. Yet the believability constraints persist.
I contend that any model should be grounded in scenarios that start with considering governmental strategic initiatives based on two fundamental areas, incentives and regulations (carrots and sticks.) Human nature, at its core, is most strongly influenced by the opportunity to seize financial advantage. Essentially, government policies either manipulate domestic and international financial pursuits through incentives or regulations. The complexity comes from not just competing goals but a lack of continuity among how nations choose to achieve those goals. And there are also areas where there is an absence of any structured strategic approach - perhaps best defined as untethered free markets. But, that’s just the basis for analysis. It needs to be applied to the myriad of financial theaters, including energy, trade, and mineral and agricultural commodities. That’s complexity upon complexity many times over. But, at least it is grounded in what truly drives behavior.
It is such a privilege to have the opportunity to read your essays for free.
I think this is the right approach for a wide range of communities!
A little Herbert Simon with a dash of modern control theory would go a long way...