No (Process) Risk, No Reward
Design a process that lets you win
In our launch piece, we argued that government is far too risk-averse. Much of that risk aversion manifests in the way we design processes.
Policymakers are often too preoccupied with designing procedures that are immune to criticism, leading to needless rigidity and bureaucratic torpor. Typically, this preoccupation centers on either oversight risk (What if an inspector general or congressional committee starts an investigation?) or litigation risk (What if we face legal challenges under the APA, NEPA, or other statutes?). In turn, initiatives fail because they underweight the ultimate risk: not delivering a result.
CHIPS succeeded in large part because we recognized that delivering results required taking on some process risk. We gave ourselves more flexibility to evaluate and engage with applicants than a typical government program allows, while keeping the process consistent and fair. Although we took oversight and litigation risk seriously, we knew that the greatest danger lay in failing to meet the mission. In this piece, I’ll revisit two implementation decisions that forced us to choose between breaking procedure and embracing it, distilling some general principles for effectively managing process risk to achieve program goals.
Inertia is lethal
The federal government tends toward inertia.
The first source of inertia is bureaucracy. This challenge has private sector analogs — I imagine that the bureaucracy at the Commerce Department somewhat mirrors that of a legacy health insurance company. But in government, the large bureaucracy of an individual agency like the Commerce Department is itself nested within the even larger bureaucracy of the federal government.
And operating within that huge bureaucracy is uniquely challenging. At CHIPS, we fielded regular incoming from a glut of White House players,1 as well as other federal agencies,2 each with different, sometimes competing priorities. And each at some point made interventions that threatened to derail our timelines — sometimes reasonably and sometimes less so, often raising issues that were critical to their missions but less central to ours. The White House mediated by installing a deputies-level3 CHIPS coordinator to help manage the bureaucratic onslaught. But it was a lot to handle.
The challenge of working with a mass of bureaucratic stakeholders is compounded by laws, regulations, and procedural norms that intentionally slow things down (again, sometimes reasonably and sometimes less so). Almost every operational step we took — from hiring to engaging contractors to executing funding decisions — came with process requirements that would be unrecognizable outside government. We’ll dig into many of those procedural constraints throughout Factory Settings, but suffice it to say that while those speed bumps can mitigate risk, they also create a powerful bias toward inertia.
The mix of bureaucratic and procedural constraints threaten to get you stuck in the mud, and getting stuck in the mud will kill you. Weeks of delay turn into months; applicants and other stakeholders get frustrated; the Hill starts to call; the press jumps in; your team loses morale; recruitment gets harder; the spiral accelerates. As program director, my job was first and foremost to keep pushing the cart through the mud. A slower or more constrained approach might have mitigated some oversight or litigation risk, but it would have exposed us to the biggest risk of all: failure to get the job done.
Decision #1: Abandoning the grant conclave
With so many bureaucratic and procedural constraints outside your direct control, it’s critical to get the ones within it right. We started without a playbook at CHIPS, so we were essentially building the process from scratch, borrowing from both government and private sector analogs where applicable. Many of our early briefings with Secretary Raimondo were spent debating the basic architecture of the program. We were operating in a fog of uncertainty: we didn’t know how many — or which — companies would apply, or the scale of the projects we’d be asked to evaluate.
To build the CHIPS program, we had to design dozens (if not hundreds) of smaller processes. In each one of those processes, we had to decide how to calibrate process risk: how to prioritize and evaluate applications; how to size awards; the sequence an applicant would follow from initial submission to final approval; the government’s internal review flow; our internal governance mechanisms; how to integrate statutorily required environmental reviews; and much more. But one early decision shaped everything that followed: the rules of the road for engaging with industry.
In a typical grant process, interaction with applicants is tightly controlled. Applicants submit their materials, and the government retreats behind closed doors, only re-emerging with a funding decision, as if from a chapel with white smoke. While the government is in that chapel, engagement with applicants is minimal. That practice stems partly from the Administrative Procedure Act’s prohibition on “arbitrary and capricious” actions: if Applicant A gets ten meetings and Applicant B gets three, Applicant B might argue in litigation that they received unfair treatment.
But oversight risk also looms large. The more back-and-forth with applicants, the more likely some email or text — stripped of context or read uncharitably — ends up as fodder for an inspector general or a congressional committee, and ultimately for the Wall Street Journal or New York Times. To eliminate those risks, agencies often limit engagement across the board.
For us, that model had considerable shortcomings. We were dealing with massive information asymmetries while standing up a program to negotiate with companies in one of the world’s most dynamic and competitive industries. Sustained, iterative engagement was the only way to understand the companies, the industry structure, and the strategic context.
On the financial and commercial side, we had built a team with deep market expertise to dissect project-level economics — work that requires frequent (sometimes daily) information exchange. We were also making strategy calls that couldn’t be reduced to static checklists: How much investment should go to leading-edge logic versus memory? When are we comfortable relying on allies, and when do we need domestic capacity? If China is pouring resources into a particular segment, does that make a project less attractive (because the economics are tougher) or more (because we may need to meet fire with fire)? What’s the value of supporting American companies versus onshoring foreign investment? Moreover, we were making these commercial and strategic judgments in an environment where the facts on the ground were continually shifting, requiring regular dialogue to understand those shifts and adapt our approach accordingly.
And then there was the simple fact that we were heading into negotiations — pressing companies to do more for less. We were asking firms to make capital commitments totaling tens of billions of dollars and take on execution risk in an industry with brutal cyclicality. That kind of negotiation can’t be done from the conclave. It requires continuous engagement to pressure-test assumptions, surface concerns early, and build mutual credibility and trust. Without that back-and-forth, we had no shot at getting to outcomes that both protected taxpayers and aligned with our strategic objectives.
Our implementation strategy also hinged on working with the broader industry. Beyond engaging the applicant chipmakers, we also wanted to engage with their major customers: Apple, Nvidia, AMD, Qualcomm, Alphabet, Amazon, Dell, HP, and others. We hoped to (a) get them to signal demand for US chip manufacturing and (b) learn all we could about the industry to lessen the information asymmetry between us and our applicants. Once again, regular customer interactions introduced risks of inconsistency that could be wielded against us in court. But they were indispensable to our mission.
Our desired approach provoked extensive discussions with the Department’s (excellent) legal team. While we worked to structure our engagement framework, we had to bar the financial and industry veterans we’d hired into the Investments Office from making contact with their industry connections — even though their industry experience was the reason they’d been hired in the first place.
Ultimately, though, we worked with the lawyers to create the flexibility the program needed. Two decisions were key:
First, we designed a highly iterative application process. No chapel, no white smoke. We stayed in constant dialogue with applicants at every stage of the process, continuously exchanging information and meeting both staff and executives.
Second, we continuously engaged with major chip customers, encouraging them to signal that they wanted to buy domestically produced chips. And we became an information hub — continuously triangulating, validating claims, and generating insights that were essential to implementation. If an applicant claimed they were close to landing a big customer, we didn’t take it at face value; we called the customer to check.
There were real risks to this strategy. Ongoing dialogue created more oversight exposure. We were never going to treat every applicant identically, even if we were committed to treating them fairly. But we would not — and could not — have succeeded without the situational awareness that came from deep engagement with applicants and the broader industry.
Decision #2: Running an open and competitive process
The first example shows how we embraced risk and reshaped the process to give ourselves a better shot at success. This example is different: a deliberate decision to embrace process, even at the cost of meaningful delay.
Due to a crucial intervention by Commerce staff during the legislative drafting, the statute granted us Other Transaction Authority, or OTA. We’ll have lots to say about OTA throughout this series, but, generally, it gives the government the ability to contract much more flexibly and creatively, unencumbered by onerous grant or acquisition regulations. Instead of running an open and competitive application process, we could have used OTA to just go out and start cutting deals.
The DoD’s recent MP Materials deal exercised a similarly expansive and flexible authority to move quickly and nimbly. The agreement provides a combination of price floors, demand guarantees, loan financing, and equity investment to support domestic rare earth production. Far from getting bogged down by process, the DoD pretty much just kicked up a discussion with the company and got the deal done. Compare that to our program — we:
Drafted and issued a Notice of Funding Opportunity (NOFO) (6 months)
Waited to receive our largest applications (4-6 months)
Put those applications through our evaluation process, sized award across the portfolio, and made preliminary offers (2-4 months for initial deals, which we condensed over time)
Negotiated and executed preliminary term sheets (2-4 months initially, again condensed over time)
Conducted due diligence, negotiated final terms, and executed final awards (4-6 months)
Would CHIPS have been better served by a model more akin to the MP Materials deal? Given the grind of a cumbersome process, some on our team probably think so. But I believe we played our cards right, as our process unlocked several advantages:
Because our competitive process was actually competitive, it created a “Hunger Games” dynamic within the industry. Applicants knew they were competing for limited funds, compelling them to put their best foot forward in their applications — pledging more investment, more advanced technologies, enhanced national security applications, new R&D, and so on. We may have been able to recreate that dynamic in a non-competitive process, but might not have been as successful.
An open application process gave us a reasonably complete picture of potential demand for our funds early on. We weren’t just trying to attract a single company, or even a single segment of the industry: we wanted to catalyze investment across industry categories (leading-edge logic, memory, current and mature, and compound semiconductors), as well as up and down the supply chain (materials, equipment, and advanced packaging). Our process allowed us to size individual awards with a well-informed sense of how the overall portfolio would take shape. It also let us develop a targeted “go-to-market” strategy for the industry segments that we identified as priorities in our Vision for Success, but that were not yet represented in our application pipeline — something we did effectively in the case of advanced packaging.
Because we gave applicants time to apply and received a range of applications within a concentrated timeframe, the competitive process helped us avoid a notorious pitfall of industrial policy — picking winners and losers. Given the pressure to move quickly, it’s easy to imagine making a premature bet that we’d end up regretting had we not followed a competitive process; in writing this, I have in mind a specific early potential investment that could have gone that way. But as we worked through the competitive process, we settled on a diversified portfolio of global players making major investments in the US, including all five companies at the leading-edge (TSMC, Samsung, Intel, Micron, and SK Hynix). That’s a big difference from the MP Materials deal, which was a massive brute-force subsidy to a single player. Hopefully we will see the administration roll out a broader portfolio strategy on rare earths over time.
Finally, CHIPS was highly scrutinized, and I do think that running a competitive process gave us political legitimacy. Our NOFO laid out a clear process, and we followed it rigorously. Throughout the program, we could have been derailed by political criticism and pressure — both from companies trying to exert political influence to get more funds and from companies activating political support to overturn a rejection. But we remained committed to our process, and I believe that the transparency and rigor gave us protection.
This is not to say that an open and competitive process always suits industrial policy initiatives. In fact, most other countries deploy semiconductor incentives through programs that look more like MP Materials and less like CHIPS. And there were certainly elements of our process we could have streamlined or improved. But I do believe that in our case the decision to run an open and competitive process was the right one — even if it put us on a longer timeline to execution.
No risk, no reward
In financial assistance programs, there’s a natural impulse to design seemingly unimpeachable review and reward processes. But that approach creates two predictable traps. You might (a) lock yourself into a process that slows you down or yields suboptimal results, or (b) define a process upfront that later compels you to color outside the lines to get the job done. Option (a) risks stalling momentum; option (b) invites even greater oversight and litigation risk.
But not all process is bad. In both the public and private sector, strong processes can enable efficient, high-quality outcomes. And while the many ways the government handicaps itself as an economic actor truly drive me crazy, it would be a mistake to take that frustration to its logical extreme. Government is different: it has a legal and moral obligation to treat people fairly and to make decisions with rigor. Those commitments to fairness and rigor form the basis of political legitimacy, which in turn helps to preserve outcomes over time. But you have to design a process that lets you achieve those outcomes in the first place.
The Chief of Staff’s, Counsel’s, and Communications offices, the National Economic Council, National Security Council, Office of Science and Technology Policy, Domestic Policy Council, Office of Management and Budget, and the Council on Environmental Quality
The Department of Defense, the Office of the Director of National Intelligence, the Central Intelligence Agency, the Department of Labor, the Environmental Protection Agency, and the Department of Justice
One tier below the principals who report to the President
 | A guest post by
|