Beyond Fear: Why defense acquisition reform needs a more honest diagnosis

The system's critics may be subject to the same behavioral forces they describe

A recent War on the Rocks article made a forceful and well-documented case that fear functions as "architecture" in the defense acquisition system, not merely a cultural disposition but a structural operating logic that shapes decisions at every level ^[2]. The author’s argument, consistent with a growing body of practitioner commentary, is that acquisition professionals are not resistant to change; they are responding rationally to a system that punishes visible failure while rendering invisible the cost of delay. The diagnosis is largely correct. The treatment, however, stops short, and may reproduce the very pathology it describes.

The broader reform narrative tends toward a common conclusion: the system needs to be faster, more commercial, more agile, more willing to accept risk. Innovation organizations like the Defense Innovation Unit, AFWERX, and Kessel Run are held up as demonstrations of what is possible. The persistence of slow, conservative procurement is attributed to the system's failure to learn from them at scale.

Our article seeks to advance a more uncomfortable argument. The reform narrative itself may be subject to the same fear it identifies in the system, specifically, a reluctance to acknowledge that defense-oriented innovation organizations are frequently imperfect, often overhyped relative to systematic outcomes, and selectively cited as evidence of possibility rather than examined as evidence of limitation. More fundamentally, the reform critique rests on a category error: it implicitly treats defense procurement as a deficient version of commercial procurement rather than as a categorically different enterprise with structural characteristics that constrain streamlining in ways that no amount of cultural change or authority expansion will resolve.

What the reform literature gets right

The behavioral diagnosis in the reform literature is well supported. Kahneman and Tversky’s ^[3] prospect theory demonstrates that when decision makers face asymmetric consequences, they tend to prefer options that minimize potential losses rather than maximize gains. Empirical estimates in the prospect theory literature commonly find that losses are weighted roughly two to three times as strongly as equivalent gains, a phenomenon known as loss aversion. Bellé et al. ^[4] confirm that these cognitive biases manifest systematically among public employees across policy and management domains, extending prospect theory's predictions into the specific context of public administration. This asymmetry is not a personal failing. It is a rational response to the actual consequences the system assigns.

The Government Accountability Office's 2025 assessments confirm the scope of the problem: major defense acquisition programs now take almost 12 years to deliver an initial capability, DOD plans to invest nearly $2.4 trillion in 106 of its costliest weapon programs, and weapon systems acquisition has remained on GAO's High Risk List for over two decades ^[5]. These are not outcomes a well-functioning system produces. Something is structurally wrong.

Romzek and Dubnick's ^[6] accountability framework, derived from their analysis of the Challenger disaster in Public Administration Review, explains why. Public organizations managing multiple simultaneous accountability demands (bureaucratic, legal, professional, and political) default to the most conservative interpretation satisfying all principals at once when those demands conflict. Koppell termed this "Multiple Accountabilities Disorder" in Public Administration Review ^[7] Weaver in the Journal of Public Policy ^[8] identified the underlying mechanism: political actors, and by extension bureaucratic actors, are motivated primarily by the desire to avoid blame rather than claim credit, because potential losses register more powerfully than equivalent gains.

These theoretical frameworks, and the practitioner accounts that draw on them, correctly identify fear as a structural product of the acquisition system's incentive architecture. Every actor, including program managers, legal counsel, senior leaders, budget analysts, oversight bodies, and acquisition professionals, is behaving rationally given the consequences they face. The collective outcome is irrational. The individual decisions are not.

So far, so good. The reform literature earns its diagnosis. Where it becomes less reliable is in what it proposes as evidence of the alternative.

The innovation organization problem

The War on the Rocks piece and others point to organizations like the Defense Innovation Unit, AFWERX, Kessel Run, and the Joint AI Center as demonstrations of what the system can produce when incentives temporarily align. They are held up as proof that the authorities, the talent, and the capability exist, and that only systemic fear prevents their replication at scale. However, this framing deserves more scrutiny than it typically receives.

The Defense Innovation Unit has made over 450 prototype awards since 2016. GAO reported that 62 of those awards (approximately 14%) transitioned into production Other Transaction Agreements or FAR-based contracts from fiscal years 2016 through 2023 ^[9]. DIU leadership has cited higher internal transition estimates, but the GAO-reported figure is the most methodologically transparent one. More importantly, the transition metric measures movement into a subsequent acquisition phase, not fielded operational capability. Such a distinction matters considerably for evaluating the claim of transformative impact. GAO reported that DIU described its prior activities as having a limited effect due to systemic scaling challenges and gaps in performance assessment and found that DIU itself lacked adequate metrics to assess whether its activities were producing the outcomes it claimed ^[10].

Kessel Run is cited frequently as evidence that government software factories can deliver at commercial speed. The origin story is compelling: the program's estimated cost had grown from $374 million to $745 million and no code had been delivered to the field when a stop-work order was issued in April 2017, after which a new agile approach delivered capability in months ^[11]. What is less frequently cited is that Kessel Run's model worked specifically because organizational incentives were aligned within a single structure. As the program matured and required coordination across separate program offices, contracting commands, and legal authorities, the institutional gravity of conventional practice reasserted itself. Hudson Institute analysis concluded that the approach worked best when development teams were embedded with the operational organizations responsible for delivery ^[12]. However, this is a structural condition that has yet to demonstrate that it be universally replicated across a department of this scale.

AFWERX, the Joint AI Center, and similar organizations all exhibit variations of the same pattern: measurable success within protected environments, limited and inconsistent transfer to the broader enterprise. This is not a failure of the broader enterprise to learn from them. It may be evidence that the conditions producing their successes are not generalizable.

The uncomfortable question the reform literature does not ask is whether innovation organizations are themselves subject to the fear they describe in others. Specifically, do they suffer from the fear of honest self-assessment? Organizations that are publicly positioned as demonstrations of what is possible, that attract resources and personnel on the basis of that positioning, and whose institutional survival depends on continued advocacy for their model face significant institutional incentives to emphasize successes and attribute limitations to external systemic resistance rather than to internal imperfections. This is not a criticism of the individuals involved. It is the same behavioral logic that includes blame avoidance and asymmetric accountability, that the reform literature correctly identifies in the broader system.

If the Defense Innovation Board finds that the department's "incentive structures are broken" ^[13], the honest follow-on question is whether that finding applies to the organizations advocating for reform as readily as it applies to the organizations being reformed.

The category error

The deeper limitation of the reform narrative, however, is not its selective use of evidence about innovation organizations. It is a more fundamental mischaracterization of what defense procurement is.

Commercial procurement is, at its core, a transaction between a buyer and a seller responding to market realities. The buyer defines a need, the market produces options, price signals allocate resources, competition drives efficiency, and accountability is ultimately exercised through commercial outcomes. Either the product works or it does not, the vendor survives or it does not. The system is heterogeneous, certainly, but it is governed by a relatively unified logic: value exchange between parties with aligned interests in an efficient transaction.

Defense procurement is something categorically different. It is a political, legal, and administrative process for allocating public resources toward national security objectives under conditions of radical uncertainty, competing stakeholder interests, and layered accountability to a democratic public. The "buyer" in any given defense acquisition is not a single entity with unified preferences. It is a coalition of program offices, military services, combatant commands, congressional authorizers and appropriators, oversight bodies, industry partners, allied nations, and legal authorities, each with legitimate but often incompatible interests in the outcome.

This structural heterogeneity is not a correctable deficiency. It is an inherent feature of what democratic governance of military capability requires. Senator A authorizes a program because it employs constituents in her district. Senator B funds it at a reduced level because his committee has different budget priorities. The program office defines requirements based on current threat assessments that the intelligence community disputes. Legal counsel applies precedent from bid protest decisions that reflect judicial interpretations of procurement law developed over decades. The Inspector General applies audit standards derived from appropriations law. Industry partners price risk based on their own organizational incentives. Allied nations with interoperability requirements insert constraints through foreign military sales frameworks.

No streamlining initiative resolves these competing interests. No authority expansion eliminates the legitimate accountability demands that each of these stakeholders represents. No cultural change makes the political economy of defense procurement behave like a commercial transaction, because it is not a commercial transaction and was never designed to be one.

Kelman (1990), in Procurement and Public Management: The Fear of Discretion ^[14], identified this tension decades ago: the design of public procurement systems reflects a fundamental choice between discretion, which enables better outcomes but creates risks of abuse, and rule-boundedness, which limits abuse but constrains judgment. The American procurement system has chosen rule-boundedness more aggressively than most comparable democracies, and not without reason. The procurement scandals, fraud cases, and political controversies that periodically animate congressional action are not invented concerns. They are real events that reflect real vulnerabilities in a system where enormous resources are allocated by officials exercising judgment on behalf of a public that cannot directly observe their decisions.

The reform literature tends to treat the rule-bound accountability architecture as an obstacle to be removed. The more honest framing is that it represents a negotiated response that is imperfect, certainly, but not arbitrary. It has done so to genuine governance challenges that would not disappear in a more agile system, they would simply manifest differently.

A more honest reform agenda

None of this is an argument for the status quo. The behavioral evidence is clear that the current system's incentive architecture produces outcomes that serve neither mission effectiveness nor public value. The GAO data are damning. The behavioral economics literature explains precisely why capable people produce bad collective outcomes in this environment.

But a reform agenda built on the premise that defense procurement should approximate commercial procurement, or that innovation organizations have demonstrated a replicable model for doing so, is likely to produce the same result as the previous thirty years of reform efforts: a new layer of authorities that are inconsistently used, a new cohort of innovation organizations that succeed in protected environments and struggle at scale, and a workforce that continues to respond rationally to incentive structures that the reforms did not actually change.

A more durable reform agenda would begin from a more honest foundation. It would acknowledge that defense procurement's complexity, multi-stakeholder character, political accountability requirements, and heterogeneous mission demands are structural features, not correctable deficiencies. It would design accountability mechanisms appropriate to that reality. It would distinguish, as aviation's just culture framework does, between honest errors made under genuine uncertainty, at-risk behavior that requires coaching, and willful misconduct that warrants discipline ^[15]. It would measure what it is actually trying to achieve: not process compliance but mission outcomes, not the existence of flexible authorities but their effective use, not the existence of innovation organizations but the systematic impact of their outputs.

It would also require that advocates for reform apply to themselves the same analytical standards they apply to the system. Innovation organizations that claim transformation should be evaluated against systematic evidence of transformation, not against selective case studies. Transition rates should be measured end-to-end, through operational deployment, not at the next acquisition milestone. The conditions under which innovative approaches succeeded should be examined as rigorously as the conditions under which conventional approaches failed.

A reformed accountability architecture must also reckon with a dimension the fear literature rarely addresses directly: what might be called “top cover” and its proper scope. Top cover is commonly understood as leadership authorization, specifically the grant of authority and institutional backing that enables professionals to move out on bold decisions without fear of being labeled reckless. The War on the Rocks author rightly identifies this as essential. But top cover understood only as permission to act boldly is an incomplete and potentially distorting frame. Warranted contracting officers and other acquisition professionals with fiduciary responsibilities serve multiple principals simultaneously: the mission owner, yes, but also the public treasury, the integrity of the procurement system, applicable law, and the taxpayers whose trust the system is obligated to protect. A contracting officer who declines to execute a transaction she believes violates the law, conflicts with regulation, or compromises her professional obligations is not exhibiting risk aversion. She is exercising precisely the professional judgment her warrant authority is designed to protect. The fear dynamic cuts in both directions. Reform advocates rightly worry about the professional who is too cautious to approve a reasonable transaction and is paralyzed by hypothetical protest risk. But the accountability architecture must equally protect the professional who is too principled to approve an unreasonable one, and who faces institutional pressure, career consequences, or the informal label of obstructionist for doing so. A governance culture that treats only forward-leaning action as courageous, and that treats caution exercised in defense of law, equity, or stakeholder integrity as a personal failing, has not escaped the fear problem. It has merely reversed its polarity. Top cover, properly understood, must confer the professional power to discern: to say yes when yes serves all the principals the acquisition system is obligated to serve, and to say no when no is the more responsible answer. Leadership that provides the former without protecting the latter has not addressed the system’s fear architecture. It has simply redirected it.

The historical record offers instructive examples of when saying no was the more courageous act. The U.S. Army’s cancellation of the RAH-66 Comanche program in February 2004 stands as one of the clearest illustrations. After 21 years of development, six program restructurings, and nearly $7 billion in expenditures against an anticipated total cost exceeding $39 billion, Army Chief of Staff General Peter Schoomaker and Secretary of the Army Les Brownlee made the decision to terminate the program entirely. The institutional pressure to continue was substantial: contractors in multiple states had financial stakes in the program, the Army had publicly committed to the platform for two decades, and abandoning it meant accepting sunk costs rather than rationalizing them as future investment. Yet the decision was the right one. The Army redirected $14.6 billion earmarked for 121 Comanches toward upgrading and expanding its existing rotary fleet, and the technologies developed during the program were harvested into future platforms. In retrospect, as shifting strategic priorities toward unmanned systems and counterinsurgency operations demonstrated, the Comanche would have consumed the Army’s aviation budget for a platform poorly suited to the conflicts it would actually fight. Andrew Krepinevich, then executive director of the Center for Strategic and Budgetary Assessments, offered a notably clear-eyed assessment at the time: “It was important to the Army, but it wasn’t the crown jewel” ^[16]. That judgment, rendered publicly against a backdrop of contractor interests and institutional momentum, is precisely the kind of professional discernment the acquisition system’s fear architecture works against. No reform agenda is complete if it creates conditions for bold forward action while continuing to make such honest assessments institutionally costly.

Researchers have actually found that public managers are not more risk-averse by disposition than private sector counterparts ^[17]. They show that the problem is the system, not its people. Bozeman and Kingsley, in Public Administration Review ^[18], showed that risk culture varies more within sectors than between them, driven by structural variables rather than individual disposition. The people are capable. The system, correctly understood, is genuinely complex. The reform task is to design accountability structures that are appropriate to that complexity, not to imagine the complexity away.

Calls to action

1. Codify bilateral top cover: Protect the right to say no as explicitly as the right to act boldly. Acquisition executives and legislative principals should direct that leadership protection frameworks, including top cover policies, innovation culture initiatives, and accountability reform efforts, explicitly and symmetrically protect both kinds of professional courage: the courage to approve bold action and the courage to decline action that compromises legal, ethical, or fiduciary obligations. Current reform language predominantly frames boldness as forward motion and caution as failure. This framing must be corrected. Warranted contracting officers charged with protecting the public trust should have institutional pathways, codified in policy and reinforced through leadership behavior, to exercise professional discernment without career penalty. Congress should consider directing the Warfighter Acquisition University and the Office of the Under Secretary of War for Acquisition and Sustainment to develop standardized guidance on what bilateral top cover means operationally, and to include metrics for its presence or absence in workforce climate assessments.
2. Adopt a just culture standard for acquisition decision-making, applied to all outcomes, including inaction and cancellation. The aviation safety community’s just culture framework, which distinguishes honest errors under uncertainty from at-risk behavior requiring coaching from willful misconduct warranting discipline, offers a proven model for disentangling individual judgment from systemic outcome. Defense acquisition executives should direct that program, and contracting reviews apply this framework explicitly, calibrating accountability responses to the quality of the decision process rather than to outcomes alone. Critically, this standard must apply symmetrically to decisions to proceed and decisions to stop. The cancellation of the RAH-66 Comanche after $7 billion and 21 years of development, redirecting resources toward higher-value platforms and emerging technologies, was a model of courageous professional discernment. Congress should examine whether current oversight and inspector general frameworks inadvertently penalize cancellation decisions disproportionate to continuation decisions, creating institutional incentives to persist with failing programs rather than absorb the reputational cost of stopping them. Legislation establishing outcome-neutral accountability review standards for major program terminations would help correct this asymmetry.
3. Require innovation organizations to submit to the same evidential standards they apply to the broader system. Congress and acquisition executives should condition sustained institutional investment in defense innovation organizations on the submission of rigorous, end-to-end outcome data: not prototype counts, not transition-to-next-phase rates, but evidence of fielded operational capability delivered at meaningful scale. Organizations that publicly position themselves as proof of what the system can achieve must be held accountable to that positioning through independent evaluation using transparent, pre-specified metrics. This is not an argument against innovation organizations; it is an argument for applying to them the same intellectual honesty this article calls for across the acquisition enterprise. The fear of honest self-assessment is not unique to program offices and contracting commands. It operates wherever institutional survival depends on a particular narrative about performance. Requiring innovation organizations to publish rigorous longitudinal transition data, independently verified and traced end-to-end through operational deployment, would both strengthen their credibility when the evidence supports the claim and discipline the reform narrative where it has relied on selective evidence. Acquisition executives should establish these standards administratively now; Congress should codify them in the next National Defense Authorization Act.
4. Map the process before you reform the culture: Distinguish structural complexity from institutional sediment. The reform literature’s most persistent failure is treating frustration as a uniform symptom. It is not. Some friction in defense acquisition is inherent, representing the necessary cost of operating a democratic accountability system under radical uncertainty, serving multiple legitimate principals with genuinely incompatible interests. Eliminating that friction is not possible without eliminating the accountability it protects. But a significant portion of what practitioner’s experience as friction is not inherent at all. It is institutional sediment: approvals that were added in response to a specific failure and never revisited; review layers that once served an oversight function, but whose original purpose has been obscured by organizational memory loss; processes that persist because no one has been asked whether they still serve any stakeholder in a balanced way. Treating these two categories as interchangeable is the organizational equivalent of holding one’s breath and hoping the problem resolves itself. Ignoring complexity does not eliminate risk aversion; it masks it until the deferred consequences surface as program failures, protest decisions, or operational gaps that no one can explain by reference to a specific decision point. Acquisition executives should direct systematically process mapping efforts within their organizations, not as a paperwork exercise but as a structured stakeholder inquiry: for each approval gate, review layer, and documentation requirement, ask explicitly who it serves, what harm its absence would produce, and whether the answer still reflects current operational reality rather than historical inertia. Congress should support this by directing the Government Accountability Office to conduct a comparative analysis of approval architectures across major acquisition commands, identifying requirements that cannot be traced to a current statutory mandate, an active oversight need, or a documented operational rationale. The goal is not to eliminate process. It is to ensure that every process in the system is earning its place, serving the warfighter, the taxpayer, and the integrity of the public trust, rather than simply surviving by default.

Conclusion

The War on the Rocks piece and the reform literature it represents have correctly identified fear as a structural product of the defense acquisition system's incentive architecture and correctly argued that changing behavior requires changing incentives rather than exhorting individuals to be bolder. These are important contributions to a conversation that policy and legislative audiences need to have.

But the honest version of that conversation also requires acknowledging what the reform narrative has been reluctant to say: that innovation organizations, however well-intentioned, are not exempt from the institutional incentives they describe; that their successes are real but often context-specific in ways that limit generalization; and that defense procurement's complexity is not a problem to be solved but a feature to be governed.

The acquisition system will not be reformed by simply “making it more commercial”. It will be reformed, if it is reformed at all, by designing governance structures honest about what it actually is: a political and administrative enterprise operating under democratic accountability, managing genuine uncertainty, serving multiple legitimate principals, and making consequential decisions on behalf of a public that deserves both capability and integrity from the system that protects it.

That is a harder problem than fear. But it is the right one.