Procurement, Capacity and Soverignity

When your contractors are also your enemy

If state capacity is basically the ability to get things done, what does it mean when non-state actors control that capacity? Then, it becomes a question of sovereignty.

Elon Musk, Dario Amodei, Larry Ellison

Historically, we have assumed that public sector procurement processes involved bland contractors with no real ideology beyond profit-seeking. They might do a bad job, or cheat the government, but not because they were opposed to the government. Indeed, this is part of why governments (and especially conservative governments) turned to contracting: they trusted contractors to be more responsive than civil servants.

But what if that is not true? The Trump administration does not think its true. It is cancelling many more contracts from contractors based in blue states than red states. Most recently, it labeled Antrophic as “woke” for not violating red lines about the use of AI to kill and surveil humans, declaring it a supply chain risk, effectively imposing sanctions on the company from within the US, while cancelling a $200 million contract.

These episodes reflect Trump’s partisanship, paranoia and tendency towards retribution more than any coherent philosophy. Antrophic CEO Dario Amodei isn’t fundamentally opposed to the Trump administration and it is disturbing that a private contractor is the one trying to insist on such values. But the example should be causing Democrats to rethink how they assess procurement.

Gabe Menchana offers a thoughtful critique of those supporting the Anthrophic decision, whose view is: “No government could, would, or should tolerate a private vendor dictating the terms of how it fulfills its core functions.” Henry Farrell drew out some implications for the future:

if the Trump administration actually uses the Defense Production Act or similar measures against Anthropic, it’s going to mark a big shift in the political economy of state-private actor relations in the US…If Elon Musk or Alex Karp are at all capable of sober reflection, they might realize that this change is likely not in their interests, regardless of whether the Trump administration loses or wins the next presidential election.

Maybe you are wondering “how big of a deal is procurement and contracts to current modes of governing?” The answer: a very big deal.

• The federal government spends about twice as much on procurement (about $750 billion according to some estimates) than it spends on federal employees (about $360 billion).

• State and local governments spend about twice as much on procurement as the federal government.

Historically, however, we have given private contractors and the processes that hire and manage them a fraction of the political and analytical attention we direct to government employees. One exception is Paul Verkuil's Outsourcing Sovereignty, which warns of the practical and constitutional problems of relying on private actors performing public tasks, especially in the domain of national security.

The Abundance-themed debate about state capacity has helped to change this, surfacing long-simmering problems. This is new and important! I cannot tell you the number of times I’ve been in conversations with government officials where they might raise concerns about certain contractors but refuse to name them, as if they were Lord Voldemort.¹

I would characterize this discourse as having two characteristics:

• Broad bipartisan agreement that procurement processes are important but broken, undermining state capacity.

• No clear consensus or models about what to do about it. Which is especially important for tech since tech products are a both a major item of procurement spending and also the means by which governments plan to improve their capacity.

Beyond principal agent models

One limitation of the welcome attention to procurement is that it arrives within the standard political economy logic of principal agent theory: Procurement is a market transaction where the agent (the contractor) uses information asymmetry to exploit the principal (government).

The logic of a market approach is that the agent exploits for market gains: to sell more to the principal than they need, to charge more, to sell a shoddy product and so on. If agents have broader ambitions to reshape government policy are market-related: to curry favor, reduce competition, to establish conditions of rent-seeking and monopoly.

Principal agent logic is useful, but framing government procurement only in market terms misses something, which is that procurement with government is not like selling products between two private actors. It involves the building and maintenance of state power to undertake critical public functions.

So an additional way to think about procurement is via the frame of sovereignty. Doing so helps to put the stakes and dilemmas of procurements in more stark terms, perhaps to the point that people whose eyes naturally glaze over when they hear about procurement (most of us, lets be honest) actually pay attention. It does not mean abandoning the logics of principal agent, which I rely on below, but putting them in a broader framework of sovereign power.

So lets consider three specific ways that a sovereignty perspective might reshape our sense of how procurement matters.

1. Sovereignty to do things

At some fundamental level, if a government can’t get core tasks done, can it maintain sovereignty?

Andrew Greenway asks this question, comparing the ennui of weakened state capacity (including, but not limited, to poor procurement practices) to the feeling that drove Brexit:

The other argument for state capacity is to maintain and enhance sovereignty; or more plainly, a feeling of control. As geopolitics becomes increasingly febrile, more careful thought is going into how national governments maintain their ability to exercise power within their domain. The Brexit campaign got ahead of that nagging sense of loss, playing effectively on voters’ sense of decisions being made elsewhere. It crisply tied the loss of control and agency people many felt in an increasingly globalised world directly to a supranational entity who really did have some power to bind Britain’s hands in ways she may not have independently chosen. If you voted for Brexit because you were persuaded by the repatriation of sovereign power - the ability of government to ‘take back control’ - it must be frustrating to find yourself with a state that still does not appear to be in control.

This is sort of sovereignty vibes — a sense of inability to do basic tasks — triggers a questioning of the state itself. The Brexit comparison is apt. The Brexit slogan was “take back control.” If you look at the abundance discourse, sovereignty vibes are there, reflected in a sense of loss of ability to provide safe and affordable communities with good public services.

Depending on your political perspective, maybe this sense of helplessness reflects the poor performance of public organizations and justifies more outsourcing. But it is federal policy to contract out if there are viable commercial providers and it has been for decades. Contracting out is a central feature of our current model of governing, just as much as bureaucracy bashing has been a feature of our political rhetoric. Over time, both parties have been willing to grow the contractor state while they limit or cut the number of federal employees. So its impossible to engage in complaints about public services while excluding contractors

And in some cases, private actors blur the role between contractor and rent seeker. The US seems to be unable to build a simple free tax filing team that other countries have. This is not due to a lack of technical capacity: under the Biden administration, IRS proved it was possible with Direct File. In no small part, this is because of private vendors like Intuit who both use their political clout to block IRS from building such a system, and “partnered” with the IRS to deliver a subpar product designed to send most taxpayers to its their for-profit services.

Using a different example: as state governments respond to the onerous demands that Trump’s policies place upon administering safety net services like SNAP and Medicaid, they may find that their ability to do so is limited by private vendors who run those systems. Luke Farrell has labeled this the means-testing industrial complex.

2. Sovereignty as national security

The majority of federal government procurement goes to the military. A hollow state where the state only contracts and never builds is vulnerable to the limits and preferences of private actors, and this point becomes especially salient in times of conflict. ).

The additional wrinkle that other countries face is that they are not just dealing with private vendors, but with the government that can use them as a instrument of political pressure. The US government freely accepts this logic when it comes to other countries, even when it involves non-military companies, like TikTok and Huawei.

But now, the US is the major source of military procurement sovereignty concerns for other countries. The majority of military weaponry for non-US NATO members comes from the US. And these countries are not just buying weapons, but an entire US-controlled support system to maintain these increasingly digital systems.

As NATO members view the US with growing suspicion, they also start to think about their supply systems in the context of sovereignty. Mark Carney has ordered a review of purchasing F-35s from a country where the President has threatened Canadian sovereignty. The Portuguese Minister of Defense said:

This ally of ours, which for decades has always been predictable, may bring limitations in use, maintenance, components, everything that has to do with ensuring the aircraft will be operational.

These sovereignty concerns run both ways. The US is dependent on allies for production of its weapons, and for parts of its immigration enforcement procurement. But it is the actions of the Trump administration that has caused the relationships to fray and these sovereignty concerns to become much more real.

Using a more non-military example, the digital sovereignty movement has made the case that a government that cannot guarantee national control of its digital systems is not truly sovereign. The position of the United States is that other countries cannot be allowed to control digital platforms their citizens use, or establish data sovereignty.

The Government of Canada (GC) has had reason to think about this, and framed it in the following terms:

For the GC, digital sovereignty is defined as the ability of the GC to exercise autonomy over its digital infrastructure, data and intellectual property. It is the capacity to operate effectively and make independent decisions about digital assets, regardless of where technologies are developed, hosted, or supported.

Digital sovereignty relies on the GC’s shared ability to govern, access, and secure its digital systems so that programs and services can continue without interruption. It is a collective responsibility across government to keep those systems reliable, resilient, and available. It is impossible for the GC to obtain a state of complete digital sovereignty, known as digital autonomy, due to the absolute interconnected nature of the digital world.

Dominance of military and digital services, and control of the global financial system gives the United States the ability to dictate terms for now. Declining political stability in the United States, and the naked antagonism of the Trump administration towards its erstwhile allies, means that ability becomes a threat to others.

As the Trump administration switches America from an ethos of “speak softly and carry a big stick” to one that is closer to “post crazy threats and occasionally act upon them” it pushes other countries commit to the hard work of building systems that make them less dependent on the US. This was, in effect, the message of the Mark Carney speech at Davos, but is easier said than done.

3. Sovereignty when the contractor is your ideological enemy

Thus far, the motivations of private actors can be explained by following a) profit maximization or b) the dictates of its government when it engages with other governments.

There is a third motivation: c) the contractor has its own ideological agenda. That is to say, the contractor has their own view of politics that goes beyond their business interests, and will use their government position to pursue those goals.

One way in which we see this happen is contractors explicitly align with one political party to the degree they can no longer be viewed as good faith actors for the other. Another way is contractors who operate independent of the state, because they have amassed so much power.

The obvious example here is Elon Musk. Musk is an explicitly white nationalist political force now, aligned with Republicans, but implementing his politics by supporting the far right in at least 18 countries around the world. Within the US, he was Trump’s largest donor during the 2024 campaign, and used his position in government to dismantle parts of government that progressives have championed, often in legally questionable ways.

The U.S. government has developed a multi-layered dependence on Musk’s companies, particularly in aerospace, national security, and digital infrastructure. If it wants to send astronauts to space, or bring them home, it increasingly must turn to Musk. The DOD depends upon Musk for both Starlink for internet service, and for military-use satellites. Grok has been embedded into government, including the Pentagon. Most recently, HHS is sending users to Grok as part of its Make America Healthy Again campaign.

This seems fine for Republicans, who are happy to subsidize Musk and use US diplomatic and economic power to push foreign governments to unconditionally purchase Musk’s products and accept Musk’s platforms that are structured to upend the politics of their countries.

But what about the next Democratic President? The scale of Musk’s power makes it hard to dislodge him from government, even as the extreme nature of his views make it unsustainable to treat him as a neutral contractor.

Musk is just one example. A huge chunk of Silicon Valley bowed down to Trump, ranging between the begrudging, the opportunistic, and the enthusiastic. Professor Francesca Bria has referred to the emerging: “Authoritarian Stack“— “a network of firms, funds, and political actors turning core state functions into private platforms.” She points out that the key actors in the Silicon Valley right — including Anduril, Palantir and venture capital funders like AndreessenHorowitz — are closely connected and working with one another. They also, crucially, share an ideology that places them as empire-makers able to compete with sovereign states.

Another example: Larry Ellison joined phone calls about overturning the 2020 election, and Trump has enabled him and his son to expand their media empire via TikTok, Paramount (including CBS News) and now Warner (including CNN). Also: HBO, Nickelodeon, MTV, Comedy Central, BET, TNT.

The financial backing for the Ellison family’s consolidation of media power comes largely from the fortune Larry Ellison made with Oracle, which provides the backbone of cloud computing and data management for the US government and military. Mikey Dickerson, the first administrator of the US Digital Service noted on LinkedIn:

We got used to seeing a giant, unnecessary Oracle deployment stapled onto the side of every government program, burning millions of tax dollars in O&M [Operations and Maintenance] per year. If you asked why, you got eye rolls and basically "if you don't put in a few million dollars for Oracle, they will protest and sue infinitely many times and stop the program dead.

With Oracle, the government is funding a business closely aligned with one political party, which has shown a willingness to use newly acquired media to tilt the playing field toward their preferred regime.

How do you run a government, when the government is run by your enemy?

Remember that the critique of “the deep state” is that civil servants were using their position to favor one party over the other. The logic is that we need a relatively neutral workforce that follows the law rather than its own preferences. Shouldn’t this logic should also apply to contractors, who are more brazen in melding their political preferences and organizational powers?

Why does this matter? Going back to principal agent theory, there is an underlying assumption that government always holds power over the contractor. Can’t a new government just swap out one contractor for another? For some services, yes. It’s unlikely that the pop-up firms with ties to Kristi Noem will do so well winning DHS contracts in the future now that she is gone. Simple grafters are easily replaced. But in other cases, such as military or tech procurement, replacement is very costly for a number of reasons.

Complex contracts are often for multi-year products that are not easily abandoned without great costs. For digital services, in particular, the contractor can build lock-in in a variety of ways: contracts that are expensive to break, learned helplessness on the government side, and the stacking of tech dependencies that makes it difficult to withdraw one system without affecting the other.

For moguls like Musk and Ellison, the costs of pushing out tech vendors may be even harder. They are massive political donors, and they can turn their media platforms against the government of the day.

If nothing else, one lesson has to be that debates about outsourcing, or increasingly insourcing, have to be based on considerations that are broader than a reluctance to hire new government employees, or imagined efficiencies. Some of these tasks that vendors perform are not so complex that skilled government digital teams could not perform them.

The reality is that government bureaucrats are far more accountable than embedded vendors, and building a nonpartisan and ethical state capacity is in the interest for anyone who wants to govern. But that requires a willingness to build excellence in government, and to recognize the risks of having government run by people who often oppose government.

1. It’s often about Deloitte.