I once worked in a large organization and my boss was “rotated” to a new job. He seemed to have fallen into disfavor because he took care of his work fairly quietly and pursued strategies to avoid crisis. The new boss took a different approach. He exaggerated the problems in his reporting, and then “solved” them down the road. The latter was seen as more effective than the former, it would not surprise you to learn. I saw that in other areas as well, where people who found themselves in crisis and somehow got out of the situation were rewarded, while those who never got into a crisis in the first place were seen as mildly incompetent.
I don’t know how you fix a culture like that. At least in business, there’s ultimately the success and profitability of the company to keep things in check. There’s the ebb-and-flow of talented people out of an organization and into new organizations, with the attendant decline and rise of those respective organizations.
Maybe a better solution is to come up with “shadow CDC(s)” to compete with “actual CDC”? What if we had two FDAs? Or two FTCs? Would that be less or more efficient? The military somehow has embodied this approach through the different services, who have to compete and jockey with each other, and who clash on strategies and funding. There are certainly problems with this idea, but it may not be a terrible model, or at least it may not be as terrible as the current model.
I'm tired of the trope that peacetime military leaders are somehow necessarily bad at war, at least in the US. For one, there really isn't a distinction because you become a wartime military leader...by being a good peacetime leader, being promoted, then being in position when the war breaks out. After the war, the wartime leaders become peacetime leaders. But mostly, talented leaders are good at both precisely because they are talented and because the skills are very much overlapping.
"A classic example is in the military, in which The Adverse Game works in peacetime but wartime requires The Favorable. In peacetime, generals are rewarded for filling out reports properly, flattering their superiors, and conforming to established doctrine. In wartime, they have no idea how to win. They lack drive, the ability to inspire followers, and the ability to improvise when the standard doctrine is not working."
Succeeding in the peacetime military requires "drive, the ability to inspire followers, and the ability to improvise when the standard doctrine is not working." Likewise, leaders who demonstrate that during war don't magically lose it when the war ends.
When I was in the Army, soldiers constantly bitched about how the Army cared about "useless" things like customs and courtesies, wear and appearance of their uniform, or the dreaded 2 mile run and not about how good they were at their job. The truth is that the soldiers that sucked at those things *always* sucked at their job. The soldiers good at their job (I was infantry, so this meant being good at war) almost always were good at those things too. Ranger Regiment is fantastic at everything they do, precisely because if they weren't good at garrison things then they would suck at war things and vice versa. So their ranges are immaculately planned and their uniforms (in garrison) are squared away, then they go to war and are equally disciplined and effective.
Generals are different but the same principle applies. Generals manage incredibly complex, large organizations and if they are not good at it, it's really obvious. Fighting a war at the general level requires you to effectively manage an incredibly complex, large organization. Filling out reports properly (something a general never does, btw, they have staffs for that) is very important. More important at war than at peace. Because proper reporting *is how you get the information* needed to "improvise when the standard doctrine is not working."
Most of the common disdain for our generals are due to our politicians giving them stupid missions that the military is not built for, like "nation-building". When it comes to pure maneuver, our generals are very good. The same radical accountability coupled with confident initiative - along with the very good cultural bias against ever saying "no" to a mission and deference to civilian authority - means they suck at getting politicians to not intervene constantly. They also tend to have interventionist foreign policy views like the rest of the Blob, which I think is different than whether or not they can prosecute the war given to them. They guy that complains about not getting promoted because they didn't kiss up usually is just: (1) not as good as he thinks he is, and/or (2) won't stop voicing disagreements *after* the decision has been made.
People just like saying stuff like this because it sounds good and it's easy to conjure up the image of the annoying bureaucrat pitted against the grizzled "war leader". In 2023, after decades of war, in the US those two caricatures *are the same person*. Every single leader in the military above a certain rank is a war leader and if they suck at war, they've probably been weeded out - they get evaluation reports when they're at war too. Officers who were Company Commanders during the invasion of Iraq are now reaching GO ranks, and they've deployed every 2-3 years for twenty years.
You can definitely make the case that our current crop of generals are not very good based on their track record in the GWOT...but it's not for a lack of war experience. This comment ended up pretty rambly but I'm tired of people throwing out silly surface level arguments like that as if it's a given.
We already have an CA, it's called the IG, EEOC, OSHA, and OSC. We already have a COO, it's called OMB and OPM. And you know what, they suffer the same problem. Incentives matter and their simply is no incentive to do oversight at any level from municipal offices to federal executive agencies by anyone with any of your three branches.
These only get fixed when the public starts to care (i.e. never), qualified immunity is lifted (i.e. never), and we allow the public to criminally, not just civilly, purse charges as the prosecuting party against state agents (i.e. never).
I continue to think that the case of Saddam Hussein and the Iraqi WMD allegation debate in 2001-03 is useful, if only because it flips the partisan/tribal coding. As with Trump and Russia, you had people who had strong institutional and cultural incentives to believe the allegations, and a villain who seemed very much like the sort of person about whom the allegations would be true, and who not only didn't credibly rebut the allegations, but seemed at times to be actively trying to avoid credibly rebutting them. And nobody ever really got held accountable for acting on the basis of the resultant mistaken beliefs.
What system of epistemic rewards and punishments would have kept people from falling prey to such a tempting and reasonable-seeming belief? Bipartisanship is surely not part of the solution-- the consensus that Saddam had WMDs was bipartisan! And any Chief Auditor who was enough of a gadfly to effectively and confidently point out the problems with that consensus would have been a hard person to get through confirmation hearings.
If you ever listen to Jordan Peterson, he talks about how liberals are better at creating companies and intellectual content new companies are often built on while conservatives are better at maintaining. As a conservative, I would think you'd actually prefer a government that made change at glacial pace so maybe I don't understand your point. Anyway, it seems what you are asking for would likely lead to more liberals in positions leading bureaucracy. Be careful what you ask for.
I would reframe "flattering others" as being able to work with others. And I would argue that in leadership roles, whether peace time or war, this is far more important than any battlefield strategist capability. Without it battlefield stratagem is difficult if not impossible to implement. I don't remember the source but it is oft said that C students make the best leaders. Is it true? IDK. Either way, in my experience as an army civilian, I found very few officers above the rank of lieutenant who weren't very capable, especially as leaders. What you might find more compelling is the difference in public speaking between 1 and 3 star generals. One stars were uniformly mediocre, if not out right bad on a stage before large groups. If I were to evaluate based only on public speaking, I'd surely wonder how most got there. Now obviously a few do have good public speaking skills because someone has to become 3 star (but very few). The difference in my experiences was stark. These guys had all the skills and could hold an audience even when they didn't have anything important that needed to be said.
I'm sorry but I have no confidence in an audit mechanism. It is another iteration of the disease. The accountability you seek is actually accountability to nature itself, or anything extrinsic to humanity. In most cases, this is felt most keenly by those unable to protect themselves by buffering with other human bodies. Thus the accountability of leaders in a crisis to the desperate mob. Either the leaders are held accountable, eventually, in this way, or those mobs suffer to the point that even the leaders are no longer able to be buffered, and the crisis becomes a cataclysm. Ironically, the leaders are most rewarded for buffering themselves most, and for making the system more and more rigid.
This is a fundamental, inescapable problem. We need to design systems that fail the leaders more rapidly so that they don't fail everyone completely. Marie Antoinette should have been on Elba rather than Napoleon. It would have saved her head and everyone quite a bit of trouble. But try selling leaders on the prospect of real accountability...
It's time to formalize the reality of a de jure 2 party system in the USA, unlike most other democracies, with less stable parliamentary multi-party democracies.
We need both a Republican controlled and a Democrat controlled set of FBI & DoJ & IRS attorneys, as well as two Inspector Generals with the power to subpoena and recommend the firing of any bureaucrat among the top 3 gov't levels (those not elected nor limited time appointed). The recommendation to fire could only be overridden if a majority of both the Dem and Rep.
The truth is that today the elite Democrats control most top gov't positions, and are leading to the abuse of law - too tough against many non-elite Republicans, too easy against many top Democrats.
Similarly, an 8 or 10 year term limit on bureaucrats would increase turnover and provide more opportunity for accountability. Despite Handle's likely correct prediction that this rule could be gamed into becoming gamed and more negative, it's a huge immediate improvement and good for the mid-term. With dark horse Vivek talking about it, the Overton window might move to allow it and some Rep governor might try it on some state bureaucrats - and we'd see how well it works, or not. (TX? FL? AR - S. H. Sanders?)
I can also see more gov't reports being summarized as well as checked for consistency by AIs. There was never a need for the "top people" to be top bureaucrats. We need top people making strategy. We need honest, rule-following & rule-making careful people to follow the rules - with bad results leading to changing the rules followed to get the bad result.
On Freddie's idea, as also mentioned in comments, that LLMs are not "true general intelligence", I completely agree - yet also strongly disagree. The LLMs are evolving to simulate intelligence. There will be good books written with bot help, at first, and later almost completely by bots, with human curating and editing. Even blog posts as well written as Freddie does.
How to tell the difference between real and simulated intelligence? But the LLMs are not so scary, as they become better servants - despite the "now the servant is the master" fears and, likely true in many respects.
I'll remain less scared as software bots make more and better sw bots - but start getting scared as physical bots make more physical bots. Which seems most likely in the context of a war. A drone war; soon a robot war; then an ai-drone & robot war. I live in Slovakia, next to the big European war in Ukraine. The bots are becoming better "enemy" killers daily.
Again, we can change a few words in one of Arnold's passages and apply it to the private sector, as follows:
"[Corporate] bureaucracies suffer from the same adverse pressures. The people who get promoted tend to be people who stick to the party line. They never question superiors. They distrust creative thinkers. Their goal for power is stronger than their goal to serve [customers and shareholders]. When they get to the top, they value people like themselves. The [corporation] gets captured by the players of the Adverse Game."
One might argue that the markets will ferret out inefficient corporate bureaucracies, but that can take a long time and have adverse side effects. Consider the recent SVB bank failure as a case in point. It can take corporate raiders such as Carl Icahn years to displace inefficient management. And some corporations muddle along for decades without fundamental change, witness GM in the '80's, '90's and beyond, Japanese zaibatsu, and others. Japanese corporations are well known for their shareholder-unfriendly behavior.
The case for adverse selection of leaders in the military and civilian bureaucracies flows from your assumption that the goal of these organizations is to serve the public. If instead you judge the goals of these organizations based on how they behave and what they actually do (for example, eliciting submission from the public), then the incentives facing career managers are properly aligned with the goals of the organization.
Perhaps a federal sunset law could put the chief auditor in the position of determining the dates of sunset reviews for various agencies and laws, and preparing the report to Congress with recommendations.
Arnold's reform (COO + CA) would be an efficient way to make big government more accountable, unless (as MikeW notes in the comments) adverse selection delivers the wrong COO and CA.
Compare some traditional libertarian approaches to the problems of big government (problems enumerated by Arnold):
• Pursue innovations in *exit* options; e.g., deregulation, rollback of licensing, decentralization, judicial decisions for individual liberty, "starve the beast," "choice" (e.g., school choice), counter-majoritarian checks, technology for autonomy (e.g., crypto?), even secession?
• Experiment with more accountable forms of voting (and elective representation) in government; for example, quadratic voting, and unbundling of decisions and services.
On a more serious note, another excellent post. It brought somehow to my mind Stalin and Trotsky: The former played methodically the Adverse Game, while the latter shined at the Favorable Game and stuck to it till it was too late. A very rough sketch, but I think it kind of fits your Adverse/ Favorable paradigm.
I once worked in a large organization and my boss was “rotated” to a new job. He seemed to have fallen into disfavor because he took care of his work fairly quietly and pursued strategies to avoid crisis. The new boss took a different approach. He exaggerated the problems in his reporting, and then “solved” them down the road. The latter was seen as more effective than the former, it would not surprise you to learn. I saw that in other areas as well, where people who found themselves in crisis and somehow got out of the situation were rewarded, while those who never got into a crisis in the first place were seen as mildly incompetent.
I don’t know how you fix a culture like that. At least in business, there’s ultimately the success and profitability of the company to keep things in check. There’s the ebb-and-flow of talented people out of an organization and into new organizations, with the attendant decline and rise of those respective organizations.
Maybe a better solution is to come up with “shadow CDC(s)” to compete with “actual CDC”? What if we had two FDAs? Or two FTCs? Would that be less or more efficient? The military somehow has embodied this approach through the different services, who have to compete and jockey with each other, and who clash on strategies and funding. There are certainly problems with this idea, but it may not be a terrible model, or at least it may not be as terrible as the current model.
Wouldn't the Chief Auditor and Chief Operating Officer be subject to the same adverse selection problems?
I'm tired of the trope that peacetime military leaders are somehow necessarily bad at war, at least in the US. For one, there really isn't a distinction because you become a wartime military leader...by being a good peacetime leader, being promoted, then being in position when the war breaks out. After the war, the wartime leaders become peacetime leaders. But mostly, talented leaders are good at both precisely because they are talented and because the skills are very much overlapping.
"A classic example is in the military, in which The Adverse Game works in peacetime but wartime requires The Favorable. In peacetime, generals are rewarded for filling out reports properly, flattering their superiors, and conforming to established doctrine. In wartime, they have no idea how to win. They lack drive, the ability to inspire followers, and the ability to improvise when the standard doctrine is not working."
Succeeding in the peacetime military requires "drive, the ability to inspire followers, and the ability to improvise when the standard doctrine is not working." Likewise, leaders who demonstrate that during war don't magically lose it when the war ends.
When I was in the Army, soldiers constantly bitched about how the Army cared about "useless" things like customs and courtesies, wear and appearance of their uniform, or the dreaded 2 mile run and not about how good they were at their job. The truth is that the soldiers that sucked at those things *always* sucked at their job. The soldiers good at their job (I was infantry, so this meant being good at war) almost always were good at those things too. Ranger Regiment is fantastic at everything they do, precisely because if they weren't good at garrison things then they would suck at war things and vice versa. So their ranges are immaculately planned and their uniforms (in garrison) are squared away, then they go to war and are equally disciplined and effective.
Generals are different but the same principle applies. Generals manage incredibly complex, large organizations and if they are not good at it, it's really obvious. Fighting a war at the general level requires you to effectively manage an incredibly complex, large organization. Filling out reports properly (something a general never does, btw, they have staffs for that) is very important. More important at war than at peace. Because proper reporting *is how you get the information* needed to "improvise when the standard doctrine is not working."
Most of the common disdain for our generals are due to our politicians giving them stupid missions that the military is not built for, like "nation-building". When it comes to pure maneuver, our generals are very good. The same radical accountability coupled with confident initiative - along with the very good cultural bias against ever saying "no" to a mission and deference to civilian authority - means they suck at getting politicians to not intervene constantly. They also tend to have interventionist foreign policy views like the rest of the Blob, which I think is different than whether or not they can prosecute the war given to them. They guy that complains about not getting promoted because they didn't kiss up usually is just: (1) not as good as he thinks he is, and/or (2) won't stop voicing disagreements *after* the decision has been made.
People just like saying stuff like this because it sounds good and it's easy to conjure up the image of the annoying bureaucrat pitted against the grizzled "war leader". In 2023, after decades of war, in the US those two caricatures *are the same person*. Every single leader in the military above a certain rank is a war leader and if they suck at war, they've probably been weeded out - they get evaluation reports when they're at war too. Officers who were Company Commanders during the invasion of Iraq are now reaching GO ranks, and they've deployed every 2-3 years for twenty years.
You can definitely make the case that our current crop of generals are not very good based on their track record in the GWOT...but it's not for a lack of war experience. This comment ended up pretty rambly but I'm tired of people throwing out silly surface level arguments like that as if it's a given.
We already have an CA, it's called the IG, EEOC, OSHA, and OSC. We already have a COO, it's called OMB and OPM. And you know what, they suffer the same problem. Incentives matter and their simply is no incentive to do oversight at any level from municipal offices to federal executive agencies by anyone with any of your three branches.
These only get fixed when the public starts to care (i.e. never), qualified immunity is lifted (i.e. never), and we allow the public to criminally, not just civilly, purse charges as the prosecuting party against state agents (i.e. never).
I continue to think that the case of Saddam Hussein and the Iraqi WMD allegation debate in 2001-03 is useful, if only because it flips the partisan/tribal coding. As with Trump and Russia, you had people who had strong institutional and cultural incentives to believe the allegations, and a villain who seemed very much like the sort of person about whom the allegations would be true, and who not only didn't credibly rebut the allegations, but seemed at times to be actively trying to avoid credibly rebutting them. And nobody ever really got held accountable for acting on the basis of the resultant mistaken beliefs.
What system of epistemic rewards and punishments would have kept people from falling prey to such a tempting and reasonable-seeming belief? Bipartisanship is surely not part of the solution-- the consensus that Saddam had WMDs was bipartisan! And any Chief Auditor who was enough of a gadfly to effectively and confidently point out the problems with that consensus would have been a hard person to get through confirmation hearings.
If you ever listen to Jordan Peterson, he talks about how liberals are better at creating companies and intellectual content new companies are often built on while conservatives are better at maintaining. As a conservative, I would think you'd actually prefer a government that made change at glacial pace so maybe I don't understand your point. Anyway, it seems what you are asking for would likely lead to more liberals in positions leading bureaucracy. Be careful what you ask for.
I would reframe "flattering others" as being able to work with others. And I would argue that in leadership roles, whether peace time or war, this is far more important than any battlefield strategist capability. Without it battlefield stratagem is difficult if not impossible to implement. I don't remember the source but it is oft said that C students make the best leaders. Is it true? IDK. Either way, in my experience as an army civilian, I found very few officers above the rank of lieutenant who weren't very capable, especially as leaders. What you might find more compelling is the difference in public speaking between 1 and 3 star generals. One stars were uniformly mediocre, if not out right bad on a stage before large groups. If I were to evaluate based only on public speaking, I'd surely wonder how most got there. Now obviously a few do have good public speaking skills because someone has to become 3 star (but very few). The difference in my experiences was stark. These guys had all the skills and could hold an audience even when they didn't have anything important that needed to be said.
I'm sorry but I have no confidence in an audit mechanism. It is another iteration of the disease. The accountability you seek is actually accountability to nature itself, or anything extrinsic to humanity. In most cases, this is felt most keenly by those unable to protect themselves by buffering with other human bodies. Thus the accountability of leaders in a crisis to the desperate mob. Either the leaders are held accountable, eventually, in this way, or those mobs suffer to the point that even the leaders are no longer able to be buffered, and the crisis becomes a cataclysm. Ironically, the leaders are most rewarded for buffering themselves most, and for making the system more and more rigid.
This is a fundamental, inescapable problem. We need to design systems that fail the leaders more rapidly so that they don't fail everyone completely. Marie Antoinette should have been on Elba rather than Napoleon. It would have saved her head and everyone quite a bit of trouble. But try selling leaders on the prospect of real accountability...
It's time to formalize the reality of a de jure 2 party system in the USA, unlike most other democracies, with less stable parliamentary multi-party democracies.
We need both a Republican controlled and a Democrat controlled set of FBI & DoJ & IRS attorneys, as well as two Inspector Generals with the power to subpoena and recommend the firing of any bureaucrat among the top 3 gov't levels (those not elected nor limited time appointed). The recommendation to fire could only be overridden if a majority of both the Dem and Rep.
The truth is that today the elite Democrats control most top gov't positions, and are leading to the abuse of law - too tough against many non-elite Republicans, too easy against many top Democrats.
Similarly, an 8 or 10 year term limit on bureaucrats would increase turnover and provide more opportunity for accountability. Despite Handle's likely correct prediction that this rule could be gamed into becoming gamed and more negative, it's a huge immediate improvement and good for the mid-term. With dark horse Vivek talking about it, the Overton window might move to allow it and some Rep governor might try it on some state bureaucrats - and we'd see how well it works, or not. (TX? FL? AR - S. H. Sanders?)
I can also see more gov't reports being summarized as well as checked for consistency by AIs. There was never a need for the "top people" to be top bureaucrats. We need top people making strategy. We need honest, rule-following & rule-making careful people to follow the rules - with bad results leading to changing the rules followed to get the bad result.
On Freddie's idea, as also mentioned in comments, that LLMs are not "true general intelligence", I completely agree - yet also strongly disagree. The LLMs are evolving to simulate intelligence. There will be good books written with bot help, at first, and later almost completely by bots, with human curating and editing. Even blog posts as well written as Freddie does.
How to tell the difference between real and simulated intelligence? But the LLMs are not so scary, as they become better servants - despite the "now the servant is the master" fears and, likely true in many respects.
I'll remain less scared as software bots make more and better sw bots - but start getting scared as physical bots make more physical bots. Which seems most likely in the context of a war. A drone war; soon a robot war; then an ai-drone & robot war. I live in Slovakia, next to the big European war in Ukraine. The bots are becoming better "enemy" killers daily.
Again, we can change a few words in one of Arnold's passages and apply it to the private sector, as follows:
"[Corporate] bureaucracies suffer from the same adverse pressures. The people who get promoted tend to be people who stick to the party line. They never question superiors. They distrust creative thinkers. Their goal for power is stronger than their goal to serve [customers and shareholders]. When they get to the top, they value people like themselves. The [corporation] gets captured by the players of the Adverse Game."
One might argue that the markets will ferret out inefficient corporate bureaucracies, but that can take a long time and have adverse side effects. Consider the recent SVB bank failure as a case in point. It can take corporate raiders such as Carl Icahn years to displace inefficient management. And some corporations muddle along for decades without fundamental change, witness GM in the '80's, '90's and beyond, Japanese zaibatsu, and others. Japanese corporations are well known for their shareholder-unfriendly behavior.
The case for adverse selection of leaders in the military and civilian bureaucracies flows from your assumption that the goal of these organizations is to serve the public. If instead you judge the goals of these organizations based on how they behave and what they actually do (for example, eliciting submission from the public), then the incentives facing career managers are properly aligned with the goals of the organization.
Perhaps a federal sunset law could put the chief auditor in the position of determining the dates of sunset reviews for various agencies and laws, and preparing the report to Congress with recommendations.
Arnold's reform (COO + CA) would be an efficient way to make big government more accountable, unless (as MikeW notes in the comments) adverse selection delivers the wrong COO and CA.
Compare some traditional libertarian approaches to the problems of big government (problems enumerated by Arnold):
• Pursue innovations in *exit* options; e.g., deregulation, rollback of licensing, decentralization, judicial decisions for individual liberty, "starve the beast," "choice" (e.g., school choice), counter-majoritarian checks, technology for autonomy (e.g., crypto?), even secession?
• Experiment with more accountable forms of voting (and elective representation) in government; for example, quadratic voting, and unbundling of decisions and services.
It's almost as if subverting & preventing popular control of the government leads to breakdowns in confidence, competency, and control.
At least, many of those playing the Adverse Game will hit a wall cuz the Peter Principle (https://en.wikipedia.org/wiki/Peter_principle).
On a more serious note, another excellent post. It brought somehow to my mind Stalin and Trotsky: The former played methodically the Adverse Game, while the latter shined at the Favorable Game and stuck to it till it was too late. A very rough sketch, but I think it kind of fits your Adverse/ Favorable paradigm.