I am writing this up because a few people I talked with, in my view, have a slightly wrong model of agency decision-making in the federal government. In particular, some people think that as ICE becomes less popular, you should expect a retreat from their least popular activities, policies, and decisions.
I think the folk theory is pretty reasonable in many circumstances.
Agencies have substantial uncertainty over how the public will view various activities (in part because it’s frequently impossible to predict in advance, as views will crystallize based on the particulars of a semi-random incident and then be applied to an entire category).
In this theory of the world, while agencies will occasionally do something very unpopular if it’s in pursuit of their core mission, support is broadly useful for all activities, so you get something like this.
But, of course, agencies don’t necessarily know in advance where a given activity will fall, and so will sometimes need to walk back their decisions.
In addition to the uncertainty factor, like most things in life, public support has declining marginal value. If you’ve got 70% support, moving to 75% is nice. If you’ve got 49% support, moving to 54% is vital. This is somewhat complicated by a need for at least reasonably bipartisan support: 52% from one party and 56% from the other is much better than 30/78.
Furthermore, Congress exercises oversight on agencies, and I think the folk theory, quite reasonably, expects sharper and more critical Congressional oversight on more embattled and politically vulnerable agencies.
But what if you assume that it’s going to become impossible to complete the mission?
There’s a related problem in corporate finance. Under normal conditions for a company that is doing well, the company will seek to repay the people it owes money to, even though the mission is to send money to shareholders. Legally speaking they get money before shareholders in event of bankruptcy, so why not send them the money they’re owed?
However, shortly before bankruptcy, companies are understood to have a duty to maximize returns for their shareholders. Which means that, if you can, the day before you declare bankruptcy there is some desire to sell everything the company owns, send out all that money as dividends, and then show up to the bondholders with empty pockets. There are various legal and covenantal restrictions on what companies can get away with here, but the point I want to make is this: if you know that you’re going to die tomorrow no matter what, the optimal actions to achieve your goals can look very different from when you have a longer time horizon.
A DHS funding bill was just passed in the House with support from every single Republican except Thomas Massie. If the Senate passes a funding bill, until September 30th 2026, at least, DHS leadership is largely free to do as they please, so long as it pleases the President.
If ICE’s support is bankrupt among democrats, such that DHS leadership and domestic policy advisors whose goals are closely tied to ICE activity feel that moderating for the sake of public support isn’t worthwhile, I expect to see a rapid increase in activities in the mission-critical/unpopular box above. Declaring that you no longer need warrants signed by a member of the judicial branch to break someone’s front door down, for example, would be convenient for many parts of law enforcement, not just ICE. DEA isn’t going to declare it, at least partially because they want to have support from both parties.
The bankruptcy model is not the only factor in play. DHS leadership and Stephen Miller report to the President, and if President Trump feels that ICE actions are causing sufficient political problems for him, he may order different policies. There may be pressure from other law enforcement agencies that seek to practice “policing by consent”, and fear that ICE actions could interfere with that for all law enforcement. But I think there is a substantial chance that ICE responds to declining popularity with increased urgency and willingness to sacrifice public/bipartisan support to accomplish the mission.
What the mission of ICE, and more broadly Noem’s DHS, is is up for debate, and almost certainly internally contested. I don’t think I can pass an ITT for all of them reliably, so I’ll refrain from trying to describe them.
One of the effects of the massive hiring spree ICE has gone on is to help leadership shift organizational culture, which can be extremely slow to change when hiring is at a slower pace. New hires in a slow-hire world tend to be acculturated into the existing norms, while hiring at this rate enables the new hires to bond over the factors that brought them in, and bring a new culture to the organization. Having a very short training period, reportedly just 47 days, will also help the new hires resist any old norms, practices, and expectations.
Quantitative predictions:
The agent(s) who shot Alex Pretti will face credible non-ICE investigation by the federal government during Trump’s time in office: 10%
On January 27th, in my judgement, there has not been an explicit “backing down” by senior administration officials from the claim that Pretti approached agents with a handgun: 90%. “We’re investigating the matter” does not count, nor does silence on the topic or refusing to answer questions on it.
I am writing this up because a few people I talked with, in my view, have a slightly wrong model of agency decision-making in the federal government. In particular, some people think that as ICE becomes less popular, you should expect a retreat from their least popular activities, policies, and decisions.
I think the folk theory is pretty reasonable in many circumstances.
Agencies need support to continue doing their work. People hired at an agency generally support continuing the activity of the agency and think it’s a good thing to do. People hired by the FBI want to make the world a better place, bring some measure of relief to victims of crime, be part of an elite team, and uphold the Constitution. DHS hires people who want to destroy the flood, defend your culture, and deport tens of millions of US citizens. Agency ultimate goals are a mix of expanding power and budget and accomplishing what they understand the mission to be.
Agencies have substantial uncertainty over how the public will view various activities (in part because it’s frequently impossible to predict in advance, as views will crystallize based on the particulars of a semi-random incident and then be applied to an entire category).
In this theory of the world, while agencies will occasionally do something very unpopular if it’s in pursuit of their core mission, support is broadly useful for all activities, so you get something like this.
But, of course, agencies don’t necessarily know in advance where a given activity will fall, and so will sometimes need to walk back their decisions.
In addition to the uncertainty factor, like most things in life, public support has declining marginal value. If you’ve got 70% support, moving to 75% is nice. If you’ve got 49% support, moving to 54% is vital. This is somewhat complicated by a need for at least reasonably bipartisan support: 52% from one party and 56% from the other is much better than 30/78.
Furthermore, Congress exercises oversight on agencies, and I think the folk theory, quite reasonably, expects sharper and more critical Congressional oversight on more embattled and politically vulnerable agencies.
But what if you assume that it’s going to become impossible to complete the mission?
There’s a related problem in corporate finance. Under normal conditions for a company that is doing well, the company will seek to repay the people it owes money to, even though the mission is to send money to shareholders. Legally speaking they get money before shareholders in event of bankruptcy, so why not send them the money they’re owed?
However, shortly before bankruptcy, companies are understood to have a duty to maximize returns for their shareholders. Which means that, if you can, the day before you declare bankruptcy there is some desire to sell everything the company owns, send out all that money as dividends, and then show up to the bondholders with empty pockets. There are various legal and covenantal restrictions on what companies can get away with here, but the point I want to make is this: if you know that you’re going to die tomorrow no matter what, the optimal actions to achieve your goals can look very different from when you have a longer time horizon.
A DHS funding bill was just passed in the House with support from every single Republican except Thomas Massie. If the Senate passes a funding bill, until September 30th 2026, at least, DHS leadership is largely free to do as they please, so long as it pleases the President.
If ICE’s support is bankrupt among democrats, such that DHS leadership and domestic policy advisors whose goals are closely tied to ICE activity feel that moderating for the sake of public support isn’t worthwhile, I expect to see a rapid increase in activities in the mission-critical/unpopular box above. Declaring that you no longer need warrants signed by a member of the judicial branch to break someone’s front door down, for example, would be convenient for many parts of law enforcement, not just ICE. DEA isn’t going to declare it, at least partially because they want to have support from both parties.
The bankruptcy model is not the only factor in play. DHS leadership and Stephen Miller report to the President, and if President Trump feels that ICE actions are causing sufficient political problems for him, he may order different policies. There may be pressure from other law enforcement agencies that seek to practice “policing by consent”, and fear that ICE actions could interfere with that for all law enforcement. But I think there is a substantial chance that ICE responds to declining popularity with increased urgency and willingness to sacrifice public/bipartisan support to accomplish the mission.
What the mission of ICE, and more broadly Noem’s DHS, is is up for debate, and almost certainly internally contested. I don’t think I can pass an ITT for all of them reliably, so I’ll refrain from trying to describe them.
One of the effects of the massive hiring spree ICE has gone on is to help leadership shift organizational culture, which can be extremely slow to change when hiring is at a slower pace. New hires in a slow-hire world tend to be acculturated into the existing norms, while hiring at this rate enables the new hires to bond over the factors that brought them in, and bring a new culture to the organization. Having a very short training period, reportedly just 47 days, will also help the new hires resist any old norms, practices, and expectations.
Quantitative predictions:
The agent(s) who shot Alex Pretti will face credible non-ICE investigation by the federal government during Trump’s time in office: 10%
On January 27th, in my judgement, there has not been an explicit “backing down” by senior administration officials from the claim that Pretti approached agents with a handgun: 90%. “We’re investigating the matter” does not count, nor does silence on the topic or refusing to answer questions on it.
On February 27th, ““: 75%