On the warehouse of unexercised options, and what happens when someone buys it.
Imagine the warehouse is visible.
Imagine, specifically, that Amazon has bought it.
The warehouse in question is the American state, treated as an inventory of held capabilities rather than as a set of current activities. Most of what the state contains at any given moment is not being used. The sat in a drawer for seventy years before March 2020. The has been drawn down meaningfully a handful of times in fifty years. The mRNA research apparatus received grant money for four decades on a bet that paid off in eleven months. The civil defense infrastructure built to evacuate cities ahead of Soviet ICBMs now runs wildfire evacuations in California, repurposed for an exercise condition its designers never imagined. These are held options. They have strike prices in the form of institutional conditions required for exercise. They have premiums paid continuously in the form of workforce, budgets, and political attention. They can expire. Most of them will never be exercised. Some of them are, occasionally, worth more than everything around them combined.
This is a portfolio. It has a structure. It has properties that are not reducible to the sum of its parts.
Now imagine it is owned by Amazon.
The premise is a thought experiment, not a prediction. Amazon is not going to buy the American state. But something structurally similar is happening, and has been happening, and the thought experiment is a way of seeing what the process actually does. The logic of Amazon's warehouse operations (which is also the logic of AWS, of Whole Foods Market after the 2017 acquisition, of MGM after 2022, and of dozens of smaller acquisitions in the Amazon portfolio) is the logic of making illegible operations legible, metricizing what was previously tacit, automating what can be automated, extracting rent from accumulated infrastructure, and optimizing observable outputs while letting unobservable ones decay. This logic is well-documented, consistent, and effective at what it does. It has produced extraordinary operational capabilities and substantial shareholder value. It has also, in each of its applications, transformed the acquired institution in specific ways that were not visible to the people running it before the acquisition.
The question is what this logic produces when applied to state capability.
The answer is not a hypothetical. It is a description of forces already at work, named by a company that has perfected their application in other domains. The essay's move is to make the acquisition explicit, walk through the playbook one step at a time, and report what comes into view. The warehouse becomes visible not as an object of nostalgia or defense but as an object of analysis. Some of what the acquirer would keep is worth keeping. Some of what the acquirer would cut is not worth defending. And some of what the acquirer would cut, precisely because it does not show up in any measurable performance metric, is what the state exists to hold.
The distinction between these three categories is the subject of what follows.
The playbook is not malicious. It is operational. Every move Amazon makes in its own warehouses is a response to specific incentives that work. Move more units through the system. Reduce variance in output. Compress cost per unit. Extract margin where the competitive structure permits. Over three decades those incentives have produced a mature operational language with six legible moves, each teachable and defensible in quarterly reports.
Applied to the warehouse, the playbook does what it was designed to do. It finds the places where cost is continuous and output is discontinuous, and it converts the first into the second. It does this efficiently. The resulting state is optimized along the dimensions the playbook measures, which is why the acquisition is attractive to people who believe government should run more like a business. The state that comes out the other side is not broken. It is optimized. What it no longer holds is a different matter, not visible in the operational reports the acquirer files.
What follows is the playbook, move by move, as it lands on the portfolio.
In Amazon's fulfillment centers, every task has a measure. Pick rate, stow rate, time off task, units per hour. Workers see their own metrics and know where they fall in the distribution. Management sees the whole distribution and acts on the tails. Bottom-ranked workers are systematically removed. The system does not need to form judgments about individual performance because the metrics do the judging.
Applied to , the move appears reasonable. Every federal employee should have clear performance expectations. Every agency should track output. Taxpayers deserve accountability. No one disputes this framing, because the framing is not the argument.
The argument is about what cannot be measured on quarterly timescales. A reviewer at the spending three years building the regulatory science expertise to evaluate a new class of biologics produces no measurable quarterly output during those three years. A rotating through a four-year tour in a mid-priority country builds relationships whose value is only legible when something breaks. An maintaining surveillance for a disease that has not emerged produces nothing measurable until the disease emerges. A curator at the , maintaining accessions that have not been requested this decade, produces nothing legible on any quarterly metric at all.
Each of these people, under the metric regime, ranks in the bottom of their cohort. Each is identified by the system as a performance concern. The system does not need to decide that their work is valueless. It only needs to decide that their work is unmeasurable, and unmeasurable in the metric regime is equivalent to valueless.
The cuts produce savings that show up in the next budget cycle. The capabilities that were being held, often for decades, show up only when exercise conditions arrive. Exercise conditions tend to arrive years after the cuts, under different political actors, and are routinely attributed to the conditions themselves rather than to the prior degradation of the response capability.
Amazon's warehouse roles are designed for rapid onboarding. A new hire can be productive within hours. The system does not depend on accumulated judgment in any individual role because the roles themselves are designed to require minimal judgment. Turnover is tolerated because replacement is frictionless.
Applied to state capability, the move attacks the directly. Career service is expensive precisely because it accumulates judgment that is not codifiable. The who knows which counterpart in the host country's finance ministry can actually deliver on a commitment. The antitrust attorney who has litigated enough mergers to recognize a consent decree that will not hold. The audit specialist who can see the structure of a tax shelter from the first filing. None of this expertise was in the job description. It was accumulated over decades of assignments, interactions, and institutional memory.
The playbook reduces these roles to specifications that can be filled by contractors rotating on short cycles. The work technically continues. Submissions are processed, reports are filed, meetings are held. What disappears is the tacit pattern recognition that made the work function. The replacement workforce performs adequately for tasks that resemble historical tasks and fails silently for tasks whose exercise conditions are novel.
The loses area expertise because rotation cycles are shortened and language training is cut to what can be justified by current deployment. The antitrust workforce loses the senior attorneys who knew which cases would hold up in court because compensation structures make private practice dominant for experienced litigators. The loses the epidemiologists who worked prior outbreaks because detail assignments to state health departments no longer count toward promotion. In each case the observable output metrics can be maintained or even improved. The underlying holding is gone.
Kiva robots replaced human walking between warehouse shelves. The automation was specific and legible. It did one thing well, and the thing it did was the most codifiable component of the previous human role. Each round of automation picks off the next most codifiable component.
Applied to state capability, automation is sold as reform. AI systems trained on historical agency data can adjudicate benefit claims faster and more consistently than human reviewers. Document processing can be automated. Routine inspections can be drone-conducted. Case assignment can be algorithmically optimized. All of this is technically achievable and all of it produces measurable efficiency gains.
The move's limit is what automation can encode. The systems perform adequately on tasks that resemble their training data. They perform unpredictably on tasks whose conditions fall outside the training corpus, which is to say the exact tasks for which the capability was being held. A disability claims system trained on historical claims processes historical-type claims well. Novel claims, edge cases, situations where the claimant's actual circumstances do not fit the intake forms, produce either denials or referrals to the residual human workforce, which has now been cut below the threshold where referral processing is tractable.
The state automation story is in its early chapters. What is visible so far is that automated systems are good at current operations and poor at novel ones. The capability to handle the novel is specifically what held. That capability is what the automation replaces.
Amazon acquired Whole Foods substantially for the real estate and the distribution network. Grocery retail was secondary. The move is general. Where infrastructure has been built over decades by predecessor organizations, the acquirer monetizes what the builders left behind.
Applied to state capability, the opportunities are substantial. The is sixty thousand scientists and tens of billions of dollars in specialized facilities, built with public money over seventy years. The archive is a century of continuous atmospheric and oceanographic data, impossible to reconstitute if lost. The special collections contain items that exist nowhere else. Each represents accumulated value that can, under acquisition logic, be converted to current-period revenue.
The extraction mechanisms are varied. Facility use fees. Data licensing. Research service contracts. Deaccessioning. Premium access tiers for what was previously public. Each individual monetization is defensible in isolation. Aggregated, they convert state infrastructure from public capability to revenue stream, and in the process alter the conditions under which the infrastructure can be maintained. A national laboratory whose survival depends on industry contracts becomes a contract research operation. A weather service whose revenue comes from premium subscribers develops the data products its subscribers want.
The infrastructure persists in physical form. What has changed is what the infrastructure is for.
Amazon's third-party marketplace is the company's highest-margin business. Sellers compete with one another on price. Amazon retains platform authority, sets the rules of participation, extracts rent from the competition, and takes no responsibility for the products. The suppliers bear risk. The platform captures margin.
Applied to state capability, the move is already substantially operational. Medicare converts to a platform on which private insurance plans compete. Higher education regulation converts to an accreditation-based platform on which institutions compete. Defense procurement converts to a platform on which prime contractors compete. In each case, state authority is retained, and the substantive work is performed by private suppliers competing on price.
The move preserves the authorities it depends on. Regulatory power is more valuable to a platform than to an operator. The holder of tariff authority, sanctions authority under , environmental permitting authority under the Clean Air Act, can extract compliance rents, grant exemptions to favored suppliers, and direct supplier behavior without doing the substantive work. The statutory authorities in the warehouse are thus preserved with care. It is the capabilities to use them well, to design sanctions that target intended threats, to issue permits that actually protect watersheds, to impose tariffs that achieve coherent industrial objectives, that are cut. The platform retains the authority. The authority, in the absence of the capability to exercise it well, becomes a rent-extraction instrument.
The deepest move, and the one the others are in service of. The playbook does not decide to let unobservable outputs decay. It decides to optimize observable ones. The decay follows, because resources are finite and attention directed at what is measured is attention withdrawn from what is not.
The move has the appearance of neutral improvement. Faster benefit processing, shorter regulatory review times, more cases cleared, more audits completed, more permits issued, more contracts let. Every one of these is genuinely improvable, and improvement in each is legible as success. The accountability regime rewards improvement on measured outputs.
What decays is everything that was not measured. The thoroughness of the regulatory review that was cleared faster. The quality of the audit that was completed faster. The fit of the benefit adjudication that was rendered faster. The long-tail expertise that does not appear in current output but underlies future output. The relationships that constitute diplomatic capability but do not show up in cable counts. The institutional memory that allows response to novel events but does not appear in routine operations.
Most of what the state holds is in this second category. Most of what mattered about the state was never in the metrics. This is not an oversight. It is the nature of held options. They are held precisely because their value cannot be realized in current periods, and the inability to realize value in current periods is identical to invisibility in period-by-period metric systems.
The acquirer is not concealing anything. The decay is publicly reported in the sense that the cuts are documented. What is not reported is what the cuts foreclose, because the foreclosed futures have not happened yet and do not exist in the form of observable data. They exist only as the set of exercise conditions that will later find the warehouse either stocked or empty.
Apply each move to see which options it hits. Tap any dot for the full analysis.
The playbook is a view from above. It sees the warehouse as a portfolio of assets, the holders as a cost center, and the cuts as optimization. From inside the warehouse, the work looks different, and the difference matters for reasons that have less to do with sympathy than with the limits of the metric regime itself.
Consider four holders, drawn from the inventory and composited from what the work actually looks like.
A curator at one of the sites in Iowa has spent twenty-three years maintaining the crop genetic repository. Her daily work consists of regeneration cycles (growing out a small number of accessions each year so the seed stock does not decline in viability), characterization (recording the morphological and genetic traits of accessions whose significance may not be apparent for decades), and taxonomic judgment (deciding when an accession is distinct enough from others to warrant separate preservation). In a typical week, she is contacted by perhaps two researchers who need specific accessions. Most of the collection she maintains has not been requested by any researcher since she arrived. She could not tell you, because no one can, which accessions will matter in forty years. Her job is to maintain the option.
A in his second year of a three-year tour in Ecuador has developed, through a combination of language study, sustained professional engagement, and a considerable volume of coffee-table conversation, working relationships with perhaps a hundred mid-level officials across the Ecuadorian government and civil society. None of these relationships has produced anything a cable could report as a deliverable. Most of them will go unused during his tour. Their value, if any ever materializes, will appear during a crisis that could occur at any point in the next thirty years and will by then be run by someone else in his job, relying on what he and his predecessors built.
An has spent her first eighteen months learning the scientific and regulatory terrain of the therapeutic class she has been assigned. Her supervisor has told her it will be three to four years before her work is independently reliable, and perhaps a decade before she can anticipate which failure modes of new submissions are characteristic of which sponsors. During this development period she produces work that her senior colleagues must review closely. The productivity metrics register her as below average for the office. She is, in fact, doing exactly what the office was designed to produce.
A in Atlanta has been reviewing surveillance reports for enteric disease outbreaks for eleven years. In most weeks the signal is background noise. Her value to the agency and to the country resides in what she would recognize on the week the signal stopped being noise, a recognition that depends on her having spent the preceding years calibrating to the noise. She does not know in advance which week that will be. Neither does anyone else.
In each case the work has three properties that distinguish it from the work the playbook can see. It is preventive rather than productive. Its outputs are diffuse, and they accrue to beneficiaries who are not the payer. Its value is established only retrospectively, at moments of exercise that the worker does not control and may not witness.
These three properties have a literature. The literature has been developed over the last forty years, mostly outside public administration, largely in feminist economics. What it describes is care labor.
Nancy Folbre's work on the economics of care begins from the observation that conventional economic accounting cannot register the value of work whose output is a public good whose beneficiaries are not the purchasers. Child care, elder care, the domestic labor that produces and maintains the workforce itself: all of this work creates value that markets cannot price. Folbre's argument is structural rather than moral. The problem is not that caregivers are undervalued emotionally or that their contributions go unappreciated. The problem is that the metric regime by which economic value is recognized is conceptually unable to register what caregivers produce. The regime is doing what it was designed to do. What it was designed to do excludes the work.
Marilyn Waring developed an adjacent argument in If Women Counted (1988), tracing the exclusion of care work from GDP calculations to specific methodological decisions made during the design of national income accounts in the 1940s and 1950s. Waring's point was that the accounts were not wrong about what they chose to measure. They were wrong about what was left out, and the exclusion was not accidental. It followed directly from the conceptual framework. Work is what produces marketable output. Caring for a child does not produce marketable output. Therefore caring for a child is not work. The framework was coherent. It was also systematically wrong about what sustained the economic activity the framework claimed to describe.
Silvia Federici's work on social reproduction pushes the argument further, describing how the care sector is politically disorganizable: the workers are isolated in households or in feminized professions, their outputs are not measured, their value is invisible until the moment of its absence, and no concentrated beneficiary exists to defend their conditions. The labor is essential. It is also politically orphaned. The orphaning is not an accident of organizing failure. It is a structural property of work whose value is diffuse and preventive rather than concentrated and productive.
The argument generalizes. Anywhere work has these properties, the same consequences follow. The metric regime cannot see it. No concentrated beneficiary will defend it. Exercise of the underlying capability, when it comes, is often credited to someone else. The workers themselves tend to understand what they are doing while being unable to make the case for it in the terms the accounting regime recognizes.
Folbre, Waring, and Federici were describing care work specifically, but the structural analysis is not confined to care. The state holds, through its career workforce, a capability that has the same properties. The curator at the germplasm repository is doing preventive and prospective work whose beneficiaries are unknown; the is maintaining relationships that produce public goods under crisis conditions that may not arrive; the and the are accumulating pattern recognition whose value is only legible at moments of failure that better-functioning agencies work to prevent. The workers can see their work. The metric regime cannot. The acquirer, applying the metric regime, cuts.
This is the bridge: the state holding workforce is performing, at the scale of the federal government, a labor that is structurally analogous to care labor. The analogy is not metaphorical. The structural features are identical: preventive output, diffuse beneficiaries, retrospective valuation, political disorganization. What the feminist economists diagnosed in the household has been replicated in the . The has the advantage of formal employment and the protections that accompany it. It has the same disadvantage in the accounting regime. The accounting regime, when applied with full rigor, is the acquisition playbook.
What makes the analogy more than typological is the question of transmission. Care labor produces, among its other outputs, the adult workers of the next generation. The labor is at least partly about the making of future workers. State holding labor, too, is about the making of successors. The germplasm curator does not only maintain accessions; she trains the person who will maintain them after she retires. The does not only develop relationships; he develops the junior officer who will run the country desk in fifteen years. The does not only review surveillance data; she teaches the fellow assigned to her branch how to read it.
The literature on communities of practice, developed by Etienne Wenger, Jean Lave, and others, describes how specialized competence transmits across generations in domains where the knowledge is substantially tacit. What transmits is not rules but judgment. The novice does not learn the rules for identifying an anomalous surveillance signal; she learns, through sustained apprenticeship, to recognize one when it appears. What she is recognizing is not in any document. It is in her supervisor's accumulated experience with previous signals, transmitted through sustained co-presence across years of work.
Harry Collins' work on tacit knowledge, and Donald MacKenzie's studies of how specialized technical communities sustain themselves, describe what happens when transmission is interrupted. The knowledge does not persist in documentation, because documentation was never its carrier. When the community of practice contracts below the threshold for sustained apprenticeship, the knowledge disappears. MacKenzie's case study of nuclear weapons design knowledge, which became substantially unrecoverable after the end of atmospheric testing truncated the apprenticeship pathway, generalizes. Communities of practice are physical arrangements of human co-presence over time. They cannot be reconstituted from their artifacts.
The acquisition playbook targets exactly the conditions under which such communities can sustain themselves. It contracts out apprenticeship. It rotates personnel before the ten-year horizon at which judgment becomes independently reliable. It replaces mentored career development with at-will contracting. It optimizes for the observable work product of current operations and treats the invisible work of producing the next cohort as slack to be cut.
What the acquisition does is not malicious. It is consistent. It applies the same metrical logic across all domains and produces the same effects everywhere it is applied. Care work in households was cut from the accounts in the 1940s. Care-analogous work in the state is being cut from the accounts now.
One implication of the structural analogy deserves separate statement. The holders of an option and the exercisers of that option are usually not the same people. They are often separated by decades.
The scientists who, between 1958 and 2018, maintained American capability across six decades of minimal exercise are, for the most part, not the ones who are currently cashing in the option. The people who did the work at are dead or retired. The people who held 's institutional knowledge through the 1970s are mostly out of the field. The people currently running and the programs are often the intellectual grandchildren of the first generation of holders, having received the tacit knowledge through two or three rounds of apprenticeship.
The pattern generalizes across the warehouse. Katalin Karikó did the mRNA work in the 1990s and 2000s, when no one could say what it was for. The people who exercised the option in 2020 were mostly different people, trained at different institutions, operating under different incentives. The program officers who kept the funding alive across two decades of obscurity were different people again, none of whom received the credit for the 2020 vaccines. The scientific papers attributed the vaccines to the 2020 work. The vaccines, actually, were substantially attributable to the 1990s work, the program officers' decisions to fund it, and the decades of institutional knowledge at the relevant universities that made the 2020 scale-up possible.
The credit misallocation is systematic. Exercise is legible. Holding is not. Exercisers are visible individuals with stories that the press can tell. Holders are dispersed across generations, mostly anonymous, often retired or dead by the time exercise occurs. The accounting regime that credits exercisers and ignores holders is the same regime that pays care workers less than it pays the workers who are produced by their care. The regime is coherent on its own terms. It is also wrong about what sustains the capabilities it takes credit for.
The consequence is not that exercisers are villains. The consequence is that a system that rewards only exercisers, and that treats the conditions of holding as costs to be minimized, will predictably produce a generation of exercisers with less to exercise. The cuts are made by the generation that inherited the full warehouse. The effects are borne by the generation that inherits the cut warehouse. The exchange is not a negotiation between those two generations. It is a decision made by the first about what the second will find.
The section closes where it opened. The acquirer is not concealing the existence of this workforce. The workforce is publicly enumerated, the budget lines are published, the job descriptions are in the Federal Register. What the acquirer cannot do is fit this labor into the accounting framework that the playbook runs on, for the same reason that national income accounts cannot fit domestic care work into GDP. The mismatch is not an empirical oversight. It is a conceptual feature of the accounting framework.
When the playbook measures the output of a germplasm curator and finds nothing on quarterly timescales, the playbook is not making a mistake. It is operating as designed. The design is simply unable to represent the labor of holding, because the labor of holding produces no quarterly output, which is the point. Labor whose purpose is to maintain optionality produces nothing until exercise conditions arrive. Labor whose value is preventive produces nothing by producing, precisely, nothing happening.
This is the structural condition under which the state's capacity to hold is being cut. The cuts are legible as efficiency on the accounting regime the acquirer operates by. They are cuts to a labor that the accounting regime never registered in the first place. The metric regime and the acquisition are, finally, the same thing.
One case is worth pausing on because it shows the full shape of holding compressed into a single institutional history.
In 1965, an Atlas-Agena rocket carried into low Earth orbit. was the first and, as of this writing, only American fission reactor ever operated in space. It ran for forty-three days before a voltage regulator failure ended the mission. The voltage regulator failure was unrelated to the reactor, which performed as designed. The United States had, in the same year, flown both its first and most recent reactor.[^1]
What followed was not cancellation. It was holding. Between 1965 and 2018, across eleven presidencies and roughly sixty years, the American program did not fly another reactor. It also did not disappear. continued as a thermal propulsion effort until 1972. sustained a reactor design capability through the 1980s. kept the option alive into the 2000s. and its flight-testable variant carried it into the 2010s, with a successful ground demonstration at the Nevada National Security Site in 2018. Through each of these programs, a small community of scientists and engineers at Idaho National Laboratory, Los Alamos, Oak Ridge, NASA Glenn, and a handful of other sites maintained the specific capability to design, build, fuel, and operationally deploy a nuclear reactor intended for space.
The holding was expensive. It was also largely invisible. The program names changed; the continuity of personnel, method, and institutional knowledge did not. Webb scientists trained Webb scientists. The apprenticeship chains were unbroken across six decades. In 2026, the National Space Technology Memorandum 3 directed NASA to field a lunar surface reactor by 2030, which is to say, directed the current generation to cash in an option that had been held, continuously and at real cost, by three preceding generations of their professional ancestors. The option's strike price was unknowable in 1965. The premium was paid, decade after decade, without any specific exercise condition in view. The exercise is now happening, and it is happening because the holding was done.
This is the warehouse in microcosm. Six decades of holding labor, performed by a community small enough that a reasonably complete list of the relevant personnel could be assembled, none of whom receive credit for the current exercise and most of whom will not be present for it. A capability whose value was unknowable during the holding period and is now, suddenly, central to a presidential directive. A transmission of tacit knowledge across professional generations whose continuity could not have been secured by documents alone. A strike price set by conditions that could not have been anticipated. An exercise that will be attributed, when it succeeds, to the current programs and the current administration, not to the long work that made both possible.
The case is also an argument for why the warehouse analysis is necessary. If the exercise succeeds, the credit allocation will be wrong. If the exercise fails, the attribution of failure will also be wrong: not to the current program's execution but to the thinning of the holding that preceded it. Either way, the accounting regime that credits only the exercisers will be unable to describe what actually happened. The warehouse frame offers a correction, not by sentimentalizing the holders but by naming what they did as the specific work it was: not building, not maintaining, not preserving, but holding. Holding is the work of keeping an option live across conditions under which exercise is not yet warranted, and its character as work is not visible from inside the accounting regime that treats it as idle cost.
The warehouse exists because certain institutional conditions made holding possible. The conditions are not natural. They were constructed, mostly in the late nineteenth and mid-twentieth centuries, through specific legal and administrative choices that could have been made differently. Each of them is now under pressure, from directions that are sometimes coordinated and often not. The pressure is not the subject of this section. The conditions are.
Six conditions are load-bearing. They are not the only conditions under which holding is possible, but they are the ones on which American state holding has historically depended, and they are the ones whose current erosion bears directly on the portfolio.
The was established in 1883 by the Pendleton Act, in response to a spoils system that had, by the late nineteenth century, produced administrative dysfunction serious enough to generate cross-partisan reform support. The Act's core move was to make removal from covered positions dependent on cause rather than on political preference. The move was extended across the twentieth century through the Hatch Act, the Lloyd-La Follette Act, the Classification Act of 1949, and the Civil Service Reform Act of 1978, each of which added procedural protections against removal for reasons unrelated to job performance.
The protections were not designed to benefit individual civil servants. They were designed to solve an institutional problem: how to maintain, across changes of administration, a workforce whose value depended on accumulated expertise that no individual administration had built and none could replicate from scratch. The protections were the mechanism by which the state could hold expertise across the time horizons required for the expertise to be useful. Tenure was, in the argument of this essay, the legal infrastructure of the option portfolio.
The condition is now contested in ways it has not been contested since the Pendleton Act. Proposals to reclassify substantial portions of the as at-will, to restructure removal procedures so that cause can be invoked more easily, or to replace career positions with term appointments serving at the pleasure of the appointing authority, have moved from the periphery of administrative reform discourse to the center. The arguments for the changes are the arguments that were made against Pendleton in 1883: that elected officials should be able to direct the workforce they are elected to lead, that tenure produces rigidity and unresponsiveness, that the public interest is served by responsiveness rather than by protection.
The arguments are not unreasonable on their own terms. They are, as they were in 1883, arguments against the conditions that make holding possible. A workforce subject to dismissal on political grounds cannot hold options whose exercise conditions outlast the current administration. The holders become, in the aggregate, short-horizon actors, regardless of their individual preferences. Their incentive is to produce visible output during the current administration, because their employment depends on that administration's satisfaction. Holding labor, which produces no visible output during any current period, becomes the labor most vulnerable to cuts.
Congress appropriates annually, but the programs appropriations fund operate on longer horizons. Research programs at fund projects that run for four or five years and contribute to scientific programs that run for decades. Weapons programs at DoD extend across multiple budget cycles. Diplomatic investments yield returns, if they yield returns at all, over the course of a generation. Option portfolios require, for their maintenance, some degree of stability in the resource flows that fund the holding.
The stability has historically been produced through a combination of mechanisms. Authorizing committees in Congress, whose members often serve across multiple administrations, provide continuity in program direction. Appropriations subcommittees, through established relationships with agency officials, smooth year-to-year fluctuations. The budget process, even when it works poorly, produces continuing resolutions that sustain baseline operations. Multi-year authorizations commit future Congresses to baselines that current ones establish. The federal research system, the diplomatic corps, the regulatory agencies, and the have all been sustained, across political cycles, by this combination of formal and informal mechanisms.
The stability is not a law of nature. It depends on congressional practices that have themselves eroded. The appropriations process increasingly operates through continuing resolutions rather than regular order. Authorizing committees have lost influence relative to appropriations committees, and appropriations committees have lost influence relative to leadership-driven deals. The regularity of annual appropriations, much less multi-year commitments, has become uncertain. Agencies increasingly operate under the assumption that funding may be disrupted, rescinded, or redirected mid-year.
The consequence for option holding is direct. An option whose maintenance requires continuous investment of a particular kind cannot be reliably held if the investment is subject to short-horizon disruption. The germplasm curator cannot run the regeneration cycle if this year's appropriation arrives in April. The epidemiologist cannot maintain the international surveillance network if the foreign service staffing that supports it is cut in a continuing resolution dispute. The research program officer cannot commit to a five-year grant if next year's appropriation may require the grant to be rescinded. Appropriations instability degrades holding even before any specific cuts are made, because the expectation of instability itself changes what can be held.
Inspector general offices were established by the Inspector General Act of 1978 in response to executive branch scandals whose scope would have been difficult to establish through normal oversight. The offices were given investigative independence, protection from removal, and direct reporting lines to Congress. Similar independent oversight capacities exist in the Government Accountability Office, in agency ombudsmen, in statutorily independent commissions, and in the federal judiciary's supervisory role over executive action.
The oversight capacity matters for option holding because holding is the category of state activity most vulnerable to undetected erosion. Active operations produce failures that are visible. Held options decay invisibly. The difference between a maintained stockpile and a depleted one, between a staffed regulatory workforce and a hollowed one, between an international commitment that is honored and one that has become performative, is often not visible from outside without the kind of investigative access that oversight offices provide.
The condition is currently under pressure. Inspector general appointments have become politicized in ways that were rare before the 2010s. Oversight findings are increasingly met with retaliation against the offices producing them. Budgets for independent oversight have been cut even as the scope of executive action requiring oversight has expanded. The Government Accountability Office continues to produce high-quality work; the federal response to its findings has become increasingly indifferent. The federal judiciary remains a partial check, but its reach is limited by standing requirements and by doctrines that narrow the circumstances under which executive action can be reviewed.
Without independent oversight, the degradation of held options becomes detectable only at the moment of exercise. The is adequate until the outbreak arrives and it is not. The regulatory workforce is adequate until the novel situation arrives and it is not. The oversight apparatus was designed to make degradation visible before exercise conditions arrived. That function is now substantially impaired.
The statutory authorities in the warehouse are held, in large part, under the expectation that they will not be exercised except under conditions that match their original design. The is on the books; presidents have overwhelmingly declined to invoke it. The is on the books; for most of the twentieth century, it was dormant. is used routinely against foreign actors; its application to domestic economic policy was, until recently, outside the range of conventional invocation.
The norm of restraint is not in the statutes. The statutes permit wide invocation. The norm is in the accumulated practice of administrations, across both parties, that chose not to use authorities beyond the conditions for which they had been established, and in the expectation, among civil servants and Congress and the judiciary, that such restraint would continue.
The norm is degrading. Authorities that were dormant are being invoked. Statutory provisions that were understood to apply only in wartime are being applied in peacetime. The strike price of held options, in the metaphor, is being reset downward, not by amendment of the statutes but by changed practice of invocation.
The degradation is significant for the warehouse argument in two ways. First, options that were held with concentrated-benefit exercise conditions in view (wartime, national emergency, catastrophic threat) are now being exercised under peacetime conditions, which changes what the option was for. Second, and more important for the prospective dimension, the norm of restraint was itself a condition of holding. Authorities that can be invoked freely against any opposition tend to generate organized opposition to their existence. The norm of restraint protected the authorities from the kinds of statutory rollback they might otherwise have faced. As the norm degrades, the authorities themselves become politically contested in ways they had not been before.
The state's option portfolio exists across time horizons that exceed any single Congress. Holding requires that some institution, somewhere in the federal government, pay sustained attention to the portfolio: to whether the holdings are adequate, to whether the conditions under which they might be exercised are changing, to whether the investments required for their maintenance are being made. Historically this attention was distributed across authorizing committees, oversight subcommittees, CBO and CRS, and the committee staff whose tenure often exceeded that of members.
Congressional capacity for sustained attention has declined. Committee staffing has been cut relative to the executive branch it oversees. The institutional knowledge held by long-tenured committee staff has thinned, as staff increasingly move to K Street or to agency positions before accumulating the decades of oversight experience that earlier generations of staff brought to their work. CRS and CBO continue to produce high-quality analytical work; the number of members who consult it has declined. The capacity to hold Congress itself accountable for holding the warehouse has weakened.
The consequence is a drift of attention toward whatever produces floor votes, leadership attention, and media coverage. Held options produce none of these unless they become objects of controversy. The , the tsunami warning network, the quarantine station system, and analogous holdings draw no sustained congressional attention in the ordinary course of legislation. Their condition is visible only when they fail or when their budgets come up for specific contestation. Between those moments, they are held or not held depending on what happens inside agencies, largely outside congressional view.
The final condition is the most diffuse and the most important. For the state to hold options that can be exercised when conditions warrant, the exercise itself must be politically available. A state that lacks legitimacy as a competent actor may hold the option in statute and maintain the workforce to exercise it, but find itself unable to exercise because the public will not accept the exercise as legitimate.
The condition is not the same as popular approval of any particular administration. It is the more durable condition by which the citizenry of a large and diverse country recognizes the state as capable of acting in the common interest, at least sometimes, under at least some conditions, on at least some problems. When this recognition is widely held, the state can exercise held options under crisis conditions and secure cooperation. When it is not, the options are nominally held but functionally unavailable.
The condition has eroded across roughly five decades, through mechanisms that are complex and not the subject of this essay. What is relevant here is that the erosion is itself a depletion of the warehouse, and one that operates independently of any specific cut to any specific program. A public health authority whose instructions will not be followed cannot exercise quarantine authority in any meaningful sense, whatever the says. A regulatory workforce whose rules will be routinely defied and rarely enforced is, for practical purposes, not holding the authority the rulebook describes. The legitimacy condition is the condition under which all other holdings become exercisable, and its erosion renders other holdings, regardless of their formal robustness, increasingly nominal.
The six conditions do not degrade spontaneously. They degrade under specific pressures, most of which are not part of any coordinated program to degrade them. The short-horizon incentives that drive the acquisition playbook also drive the pressures on the conditions themselves. An executive seeking to move faster on priorities has an interest in weakening tenure protections. A Congress under pressure to produce legislative wins has an interest in reducing the friction created by oversight. A media environment that rewards controversy has little patience for the boring conditions that sustain holding. An electorate that has lost confidence in the state has no political investment in defending the conditions under which the state can act.
The acquisition this essay describes is therefore not an event that happens to the warehouse from outside the accounting regime that governs its operation. It is the same accounting regime, applied not only to the portfolio of options but also to the institutional conditions under which options can be held. The regime optimizes for what is observable and cuts what is not. The conditions, like the labor they support, are not observable on quarterly timescales. They are observable only when they fail.
They are now failing. The failure is not a single event that a reader could point to. It is a coordinated drift across six institutional dimensions, each of which is being pulled in the direction of shorter horizons, greater exposure to political pressure, and reduced capacity for sustained attention. The drift is aggregate. Each individual movement is defensible on its own terms. The aggregate is a state increasingly unable to hold options across the time horizons for which the options were built.
This is the condition under which the acquisition is already operating. The warehouse is not being sold to a new owner. It is being evaluated by a logic that has colonized the institutional terrain on which holding was previously possible. The logic does not announce itself. It operates through ordinary administrative decisions, made by actors with no particular intent to disable the warehouse, applying the metric regime that has become the default language of institutional evaluation. The disabling is the aggregate effect of their individually unremarkable choices.
In the summer of 2023, Katalin Karikó received the Nobel Prize in Physiology or Medicine, shared with Drew Weissman, for discoveries that enabled the development of mRNA vaccines against COVID-19. The citation noted their foundational work on modified nucleosides, conducted primarily between 1997 and 2005 at the University of Pennsylvania. The award ceremony made much of the contrast between the eventual scale of the discovery's impact and the obscurity of the work during its development. Karikó had been demoted at Penn. Her grant applications had been repeatedly rejected. She had considered leaving science. The prize, the ceremony suggested, was a vindication.
The Karikó story is often told as a story about individual perseverance against institutional failure. A more accurate telling would be a story about institutional success in spite of individual adversity, because what made the 2020 vaccines possible was not Karikó's persistence alone. It was an funding apparatus that, across two decades, sustained basic research on mRNA without any specific exercise condition in view. It was a university system that continued to employ Karikó, even after demotion, in a laboratory with the instruments and adjacent colleagues her work required. It was a biotech sector sustained by federal research infrastructure that developed the delivery technologies Karikó's mRNA needed to reach cells. It was an regulatory science workforce with the institutional memory to evaluate a novel modality under emergency conditions. It was a and an Operation Warp Speed that could commit government procurement to vaccines whose efficacy was not yet proven. It was, above all, an ecology of holders whose work was invisible during the holding period and remained invisible even at the moment of exercise, because the accounting regime that allocated the credit could see only the exercisers.
Karikó was the exerciser who became visible. The holders were everyone else, across decades, whose work she could not have done without.
The asymmetry has a structure that generalizes beyond any single case.
Exercise is temporally concentrated. It occurs at a specific moment, often under crisis conditions, involving a small and legible cast of actors. The moment produces a narrative that the media can tell, a set of decisions that can be attributed, and a result that can be measured. Exercise is what journalism is organized to report, what history is organized to chronicle, and what political systems are organized to reward.
Holding is temporally distributed. It occurs across decades, involves large and dispersed casts of actors, and produces no narrative that can be told in the present tense. The germplasm curator's twenty-three years, the 's four-year tour, the epidemiologist's eleven years of surveillance review, the program officer's decade of grant decisions: none of this produces events that can be reported as they happen. It produces conditions that become visible only at moments of exercise, by which time the conditions have often been constructed over periods that substantially exceed any single career, much less any single news cycle.
The asymmetry between concentrated exercise and distributed holding creates systematic credit misallocation. Exercisers get credit not because of any bias in favor of exercisers, but because the accounting regime cannot represent distributed holding in forms the regime can process. Karikó got the Nobel because prizes are awarded to named individuals. The program officers who kept the funding alive got no prize because there is no prize for program officers, and no prize could be awarded because their work was collective, distributed, and constituted by many small decisions rather than by any identifiable breakthrough.
This is not a bias that can be corrected by awarding more prizes. The problem is not that holders are under-recognized. The problem is that the category "holder" does not exist in the accounting regime, and the absence of the category is what allows the regime to treat holding as idle cost rather than as work.
One could respond that the credit misallocation is an aesthetic problem, not a substantive one. The holders, on this view, are not owed prizes, and the exercisers' fame does no material harm to the holders or to the capabilities they maintain.
The response is wrong in a specific way. The credit misallocation is not merely aesthetic. It has three substantive consequences for the warehouse.
First, it shapes the career incentives of the next generation of potential holders. A young scientist observing the Karikó story does not learn that basic research is important. She learns that basic research is low-status, low-paid, and precarious, and that the path to recognition runs through translational work, commercial applications, and the kind of exercise that can be attributed to identifiable individuals. The rational response is to pursue the exercise-facing work. Over a generation, this rational response produces a workforce structured toward exercise and away from holding. The holding positions go unfilled, not because they were cut in any given year's budget, but because the candidates no longer exist to fill them.
Second, it shapes political support for the conditions of holding. When exercise is credited to exercisers, the political constituency for holding becomes the exercisers, who have already exercised. This is a poor constituency. The program officers who funded Karikó's work in the 1990s have retired. They do not lobby Congress for the basic research budget. Karikó, having won the Nobel, is not well-positioned to advocate for the funding stream that supported her work, because the story by which she became celebrated was a story about her individual achievement, not about the institutional system that sustained her. The people with standing to defend the conditions of holding are the holders themselves, who are either still holding (and therefore invisible) or retired (and therefore absent from current political contestation). The result is a pattern by which the visible beneficiaries of successful holding are systematically unavailable as advocates for the conditions that made their visibility possible.
Third, and most consequentially, it shapes the evaluative framework by which decisions about holding are made. When the accounting regime cannot see holding, decisions about whether to continue holding are made on the basis of what the regime can see, which is exercise. Programs that produce visible exercises recently are funded. Programs that have not produced a visible exercise in the current period are cut. The decision is not made by anyone who decides that holding is unimportant; it is made by the aggregate of individual budgetary decisions, each of which applies the criteria the regime supplies, which are criteria that cannot register holding.
This is the deep structural consequence of the asymmetry. It is not that the holders are insulted. It is that the institution's capacity to evaluate its own portfolio is systematically biased toward cuts on exactly the kind of work that the portfolio requires for its continued existence.
There is a related problem visible from the other side.
The exerciser who has exercised successfully is often the person best positioned to understand what was required to make the exercise possible. Karikó knows, better than anyone, what the two decades of support gave her. The generals who conducted successful military operations know, better than anyone, what the peacetime readiness investments gave them. The epidemiologists who detected an outbreak early know, better than anyone, what the years of surveillance maintenance gave them.
These exercisers are, paradoxically, often poorly positioned to advocate for the conditions of future holding. The Nobel laureate's public profile is now built on her individual achievement; returning to the theme of institutional support reads, to the audiences she now addresses, as false modesty or as deflection. The successful general is expected to speak about operational excellence, not about the less glamorous investments in training, maintenance, and readiness that operational excellence requires. The disease detective who caught the outbreak is celebrated for her alertness, not asked about the twenty years of surveillance apparatus she was embedded in.
The result is a pattern by which the exercisers whose insight would be most useful for sustaining holding are discouraged, by the incentive structure of their own visibility, from describing what they actually required. Some exercisers push back against this pattern. Karikó herself has been explicit about her debt to the and to the University of Pennsylvania colleagues who kept her laboratory running through difficult years. Most exercisers, most of the time, tell the story the audience expects, which is a story about individual achievement.
The accounting regime thereby receives, from its most credible witnesses, testimony that confirms its existing framework. The framework is one in which exercise is attributable to exercisers. The testimony confirms that exercisers do in fact do the exercising. The holding that made the exercising possible drops out of the testimony because the framework cannot accept it, and the exercisers, understanding the framework's constraints, do not force the issue.
The case, sketched in Section IV, illustrates the asymmetry in a slightly different form. If the 2030 lunar surface reactor is successfully deployed, the credit will be allocated to the current NASA administrator, to the DoE Secretary, to the specific programs currently in execution, and to the administration under which the deployment occurred. The sixty years of holding, across eleven presidencies, will figure in the story as background, if at all.
The exercisers cannot credit the holders for a reason that is structurally different from the Karikó case: most of the holders are dead. The engineers retired in the 1970s and 1980s. The generation is largely gone. The 1980s team is in retirement. The 2000s team is mostly out of the field. The tacit knowledge that reached the current program did so through professional transmission, across chains of apprenticeship, from people who can no longer be thanked because they can no longer be reached.
This is the extreme case of the asymmetry. When holding spans generations, the exercisers and the holders cannot meet. The credit allocation, whatever form it takes, cannot reach the people to whom it would, under honest accounting, be owed. The warehouse is held by the dead and exercised by the living, and the accounting regime that cannot see holding is also the regime that cannot credit the dead. The dead do not vote, do not lobby, do not testify, and do not appear in any budget document. Their work, if it is to be credited at all, is credited by successors who choose to name it, which in the current accounting regime is a gesture of grace rather than an institutional requirement.
The section closes with the forward-looking implication. If the accounting regime is systematically blind to holding, and if decisions about the warehouse are made on the regime's criteria, the warehouse will be depleted not by any specific villain but by the aggregate effect of many small decisions, each of which is defensible on the regime's terms.
The Karikó case is read as a success. From the perspective of the warehouse, it is also a warning. The mRNA option, held across two decades of obscurity, was successfully exercised in 2020. The next option of similar character, requiring similar conditions for its maintenance, will be held only if those conditions are maintained. The evidence of the last decade is that they are not being maintained, which means that the next Karikó, whoever she is and whatever she is working on, is less likely to have the infrastructure that Karikó had. Her work, when the next exercise condition arrives, may not have been done, because the apparatus that would have sustained it was dismantled in the period between her predecessor's exercise and her own.
The asymmetry, in short, does not merely misallocate credit for past exercises. It forecloses future ones. The warehouse is not only the current stock of held options. It is the set of conditions under which future options can be written and sustained. Those conditions are what the acquisition is cutting, and the cut is legible as optimization only because the accounting regime cannot see what the optimization forecloses.
The essay so far has treated the warehouse descriptively. It has catalogued what is held, named the labor that holds it, described the conditions that make holding possible, and traced the acquisition logic that is currently evaluating the holdings by its own criteria. The next move is to propose a different evaluative frame.
The proposal is not that the acquisition's criteria are wrong in themselves. They are the correct criteria for evaluating an operational business, which is what they were designed for. The proposal is that they are the wrong criteria for evaluating a state's held capabilities, because a state is not an operational business, and a warehouse of unexercised options is not a portfolio of operational assets. The evaluative frame needs to match the object.
What follows is the outline of such a frame. It is not a complete framework, because a complete framework would require empirical work this essay does not undertake. It is a set of dimensions along which any specific held option, and the portfolio as a whole, can be evaluated in ways the acquisition's criteria cannot.
Six dimensions carry most of the evaluative work. They map onto the real options vocabulary from Section I but have been specified for the state case.
*Strike conditions.* What exercise conditions is the option designed to meet? The question is neither trivial nor merely descriptive. Many held options were designed for exercise conditions that are unlikely to recur, while being unavailable for exercise conditions that may actually arrive. The , designed for a prolonged conventional land war in Europe, is poorly configured for the contemporary industrial vulnerabilities it could in principle address. The quarantine authorities, designed for the era before jet travel, have been stretched to cover contemporary pandemic response with institutional infrastructure that was not built for it. Strike condition evaluation asks whether the option, as currently held, could be exercised under the conditions that actually confront the state, not only under the conditions envisioned at the time of the holding.
*Strike legibility.* How clearly can exercise conditions be recognized in advance? Some options have high-legibility strike conditions. The is designed to be invoked when armed resistance to federal authority reaches a specifiable threshold. The is designed to be drawn down when oil supply disruption reaches a specifiable scale. Other options have ambiguous strike conditions that require judgment about whether the conditions have been met. The judgment is itself part of the option's value. A held option whose strike conditions can be determined only by the exerciser gives the exerciser substantial discretion. This discretion is sometimes a feature and sometimes a vulnerability; the essay's argument is that it should be evaluated explicitly rather than taken for granted.
*Premium.* What does holding the option cost, in the full sense of premium that includes workforce, infrastructure, appropriations, political capital, and institutional memory? Most state accounting focuses on the appropriations line, which is the most legible component and typically the smallest. The other components, especially the workforce and institutional memory costs, dominate the true premium but do not appear in the budget documents. A serious evaluative frame would require reporting of the full premium, which in many cases would reveal that the cost is substantially lower than the appropriation suggests (the option is being held partly by the workforce's willingness to do the holding labor at below-market compensation) or substantially higher (the political capital required to maintain the option is being depleted faster than the appropriation line suggests).
*Decay rate.* How quickly does the option degrade without active maintenance? This is the dimension the acquisition playbook most systematically mis-estimates. A regulatory workforce at baseline competence looks similar, on most observable metrics, to one that is slowly losing its senior personnel. A stockpile of medical countermeasures at full stock looks similar, on inventory reports, to one whose rotation cycle has been extended to the point where the stock is effectively expired. A diplomatic relationship at full depth looks similar, on routine communication logs, to one that has been hollowed out by successive disengagements. Decay rate estimation requires asking what would become visible if the option were exercised under demanding conditions, which is a different question from asking what the option looks like during routine operation.
*Switchability.* Can the option be exercised under novel conditions that differ from those it was designed for? This is the dimension that matters most for long-horizon holding, because the exercise conditions that ultimately arrive are rarely the ones originally envisioned. The , designed for Korean War industrial mobilization, was successfully exercised for COVID-19 pandemic response seventy years later. The mRNA funding, designed for therapeutic applications that were never clearly specified, was successfully exercised for an emergent coronavirus pandemic. High-switchability options are robust to uncertainty about future exercise conditions. Low-switchability options require that the conditions the holders anticipated actually arrive. Switchability is partly a property of the option itself and partly a property of the workforce holding it: a workforce with deep general expertise can exercise an option under novel conditions; a workforce optimized for the specific conditions originally envisioned cannot.
*Reconstitution difficulty.* If the option is cut and later needed, how hard is it to rebuild? This is the dimension that makes the holding decision irreversible in ways the acquisition playbook does not acknowledge. Some options can be reconstituted quickly: a statutory authority that has lapsed can be reenacted; a stockpile that has been drawn down can be refilled; a workforce that has been cut can be rehired, albeit with loss of institutional memory. Other options cannot be reconstituted: an extirpated germplasm accession is permanently lost; a dispersed community of practice whose members have retired and died cannot be reassembled; a diplomatic relationship that has been betrayed cannot be restored on the same terms. The distinction between reconstitutable and non-reconstitutable options is fundamental to the evaluation. Non-reconstitutable options should be held, when they are held at all, with premiums that match their irreversibility. Cutting them should require a standard of evidence that the acquisition playbook does not apply.
The dimensions do not produce automatic rankings, but they sharpen the questions that can be asked of any specific holding. A few examples from the inventory.
The has high strike legibility, moderate premium, low decay rate, moderate switchability, and moderate reconstitution difficulty. It is a relatively well-designed option on the whole. The acquisition playbook that would convert it to a commodity trading position is, on the evaluative frame proposed here, making a defensible decision under some assumptions about future oil supply risk and an indefensible decision under others. The reasonable debate is about the assumptions; the playbook does not engage that debate because the playbook does not represent the dimensions on which the debate would occur.
The has low strike legibility, substantial premium, high decay rate, high switchability, and impossible reconstitution. This is a near-ideal option by the evaluative frame proposed here: the switchability is extreme (accessions can be exercised against exercise conditions that cannot be anticipated in advance), and the reconstitution difficulty is absolute (lost accessions cannot be recovered). The playbook that cuts the is, on this frame, making a categorically wrong decision, because the cost of being wrong is infinite and the savings from the cut are trivial. The playbook cannot see this because the playbook has no representation of non-reconstitutable options; all losses are, in the playbook's accounting, recoverable.
The has moderate strike legibility (exercise is continuous at baseline and surges under specific conditions), major premium, moderate decay rate (individual positions can be cut faster than institutional knowledge can be recovered), high switchability, and generational reconstitution difficulty. It is the single highest-stakes option in the warehouse, because most of the other options depend on it. Cuts to the are cuts to the holding infrastructure for most of the rest of the portfolio. The playbook treats the as a workforce cost; the evaluative frame treats it as the platform on which holding happens.
The six dimensions are not a complete apparatus for portfolio evaluation. They are the outline of one. A more complete apparatus would specify the conditions under which each dimension can be estimated, the evidence required for the estimate, the time horizons over which the estimates should be projected, and the procedures by which estimates across dimensions would be combined. All of this is work that could, in principle, be done. It has not been done, primarily because the institutional actors who could do it are the same actors whose holding labor is being cut, and who have neither the time nor the protection to develop the evaluative apparatus that would defend their work.
The frame, even in outline, allows three moves that the acquisition playbook does not.
First, it allows evaluation of cuts on their own terms. When a specific cut is proposed, the frame asks: what is the reconstitution difficulty of the option being cut? What is its switchability? What are the strike conditions it is being held against, and how plausible are those conditions? The questions are specific enough to be answered, and the answers often reveal that cuts appearing efficient under the playbook's criteria are categorically wrong under the frame's criteria.
Second, it allows evaluation of the portfolio as a portfolio. A state that holds many high-switchability, high-reconstitution-difficulty options is in a different condition from a state that holds the same number of low-switchability, low-reconstitution-difficulty ones. The first is robust to uncertainty about future exercise conditions; the second is robust only to the specific conditions it was designed for. The portfolio evaluation treats the warehouse as a system rather than as an inventory of independent items, and notices that the acquisition tends to cut the high-switchability, high-reconstitution-difficulty items preferentially, which is the worst possible pattern from the portfolio's point of view.
Third, it allows evaluation of the conditions that support holding. If the is the platform on which most holding occurs, then cuts to the are cuts to the conditions of the portfolio itself. The frame makes this visible, where the playbook does not. The playbook treats the cut as a line-item savings; the frame treats it as a structural change that degrades every other item in the portfolio.
The frame is, in addition to its diagnostic uses, the outline of a research program. Each dimension is an empirical question that could be investigated. Strike legibility across the portfolio could be mapped. Decay rates could be estimated from the historical record of analogous cuts. Switchability could be assessed through case studies of how specific options have been exercised under conditions that differ from those originally envisioned. Reconstitution difficulty could be characterized through comparative studies of successful and failed institutional rebuilding.
None of this work is impossible. Much of it has been done piecemeal, across public administration, science and technology studies, and organizational theory. The work has not been assembled into a coherent evaluative apparatus for state portfolios, partly because the concept "state portfolio of held options" has not been named as such. The essay's claim is that the concept is a useful one, that the dimensions outlined here are a reasonable starting point for developing the apparatus, and that the development is worth doing.
The worth depends on what the apparatus would be used for. If it is used to defend holdings that are in fact poorly designed, it serves no useful function; every status quo can be defended by selective application of evaluative criteria. If it is used to evaluate holdings honestly, including the many that are badly designed, poorly maintained, or sustained by inertia rather than by ongoing justification, it becomes an instrument for the kind of portfolio management the warehouse has never received. Some options should be cut. Some should be expanded. Some should be restructured. The proposal is not that the current warehouse is sacred. It is that the warehouse cannot be evaluated well by a logic that cannot see what it contains.
The frame, if adopted, would change the evaluation of specific holdings in specific directions. Options with high reconstitution difficulty would be more protected; options with high switchability would be more valued; the would be recognized as the platform condition for the portfolio rather than as a line-item cost; the conditions supporting holding would be treated as part of the portfolio itself rather than as ambient institutional furniture.
None of this would require abandoning efficiency as a consideration. Efficiency is a legitimate criterion for operational work, and much of what the state does is operational work that should be evaluated operationally. The frame's claim is narrower: that the held-options portion of state activity, which is the subject of this essay, should be evaluated by criteria that can represent its specific character, and that the criteria currently dominant cannot.
The adoption of the frame would not be a natural process. The accounting regime that the frame challenges is the same regime that has colonized the institutional terrain on which holding was possible. Replacing it would require the reconstruction of the institutional conditions that the acquisition has degraded: protections, oversight capacity, appropriations stability, norms of restraint, sustained congressional attention, and the public legitimacy that underwrites the state's ability to hold at all. The frame cannot substitute for those conditions. It can, at best, name what they are for.
That naming is itself a contribution. A warehouse that is understood as a warehouse can be defended as one. A warehouse that is mistaken for an operational asset will continue to be acquired by logics appropriate to operational assets, with consequences that become visible only when the conditions the warehouse was held against arrive and find the warehouse empty.
Six dimensions the essay proposes for evaluating held options. Select one. The warehouse rearranges. The pattern the acquisition cuts along becomes visible.
The acquisition does not know it is making this trade-off. The accounting regime cannot see the frame it is cutting along. The cuts will be reported as efficiency.
Return to 1995. Katalin Karikó is a research assistant professor at the University of Pennsylvania, working on an idea that has no obvious application. She is trying to get mRNA to enter cells without triggering the immune response that would destroy it. The problem is technical, the literature is thin, and the grant reviewers who evaluate her proposals cannot see what the work is for. Her position at Penn is demoted. Her lab space is reduced. Her salary is cut. She considers, at various points across these years, leaving science. The record of this period is, by her own later account, a record of sustained institutional skepticism met with personal stubbornness.
This is the story as usually told. Karikó as individual scientist pushing against institutional failure. What the essay has been arguing is that this story, while not false, is the wrong story. The right story is that Karikó was held, during those decades, by an infrastructure she could not see and cannot now fully name. The R01 mechanism that funded adjacent laboratories whose instruments and reagents she used. The University of Pennsylvania's tenure system that made it institutionally awkward, even after the demotion, to terminate her entirely. The federal indirect cost recovery that sustained the building she worked in. The postdoc training grants that produced the students who staffed her bench. The library acquisitions that made the journals she needed available in her office. The patent office workforce that processed the applications that eventually mattered. The regulatory science community whose existence made it conceivable that a novel modality could be evaluated under emergency conditions if one ever arrived. The Bayh-Dole framework that let her university hold the intellectual property. None of these was designed to support her specifically. All of them were conditions under which her work could continue when nothing else would have justified its continuation.
The 2020 vaccines were not the product of Karikó's work alone, and they were not the product of her plus Pfizer plus Moderna plus BioNTech plus Operation Warp Speed. They were the product of Karikó and forty years of held infrastructure inside which her work was one of thousands of bets, almost all of which failed, whose losses were absorbed by the institutional apparatus because the apparatus was built for that purpose. The exercise condition arrived in 2020 and found the option in place because the option had been held. The holding was done by people Karikó does not know and cannot name, across roles that were never celebrated, most of whom have now moved on, retired, or died. The Nobel Prize citation did not mention any of them because no citation framework does.
The acquisition that this essay has been describing is, among other things, the dismantling of the infrastructure that held Karikó. The budget is being cut. The indirect cost recovery is being capped. The graduate training pipeline is thinning. The regulatory workforce is being reduced. The career civil servants who wrote and defended this infrastructure across administrations are being reclassified. The foreign scientists whose labor has staffed American research for decades are facing visa environments increasingly hostile to the long-horizon commitments that laboratory work requires. Each of these changes is individually defensible on the accounting regime the acquisition operates by. Aggregated, they describe a state losing the capacity to hold the kind of option Karikó's work turned out to be. The next Karikó is currently somewhere, working on something that has no obvious application, for a salary that does not justify continuation, inside an institution that is being optimized for legible output. Her work may or may not produce anything. The question the essay asks is whether the conditions under which her work could continue, if it is worth continuing, will be there. The evidence of the last decade is that they are being dismantled. The evidence of the coming decade, which cannot yet be read, will tell us what was foreclosed.
Somewhere in the federal workforce this afternoon, a program officer at the National Institutes of Health is reviewing a grant application in a subfield that nobody currently considers important. She is forty-three years old. She took the position eleven years ago after finishing a PhD in a related field, because the job offered her a way to remain close to the science without running a laboratory of her own. Her salary has not kept pace with what she could earn in industry. Her mortgage is in a Maryland suburb she chose to be near the Bethesda campus. Her husband works in the private sector. They have two children.
The application she is reviewing proposes a line of research whose applications are not clear. The applicant is a junior faculty member at a mid-tier state university. The preliminary data are promising but not yet compelling. Reasonable reviewers could disagree about whether to fund the work. The program officer has some discretion, but not unlimited discretion, and her decision will be shaped by budget pressure, program priorities, and the explicit instruction from agency leadership that the institute's portfolio should emphasize translational work with clearer near-term applications.
She is going to recommend against funding. Not because the science is weak but because the applicant is early-career, the line of inquiry is basic rather than applied, and the funding environment has contracted to the point where marginal applications cannot be sustained. The applicant will receive the rejection notice in three months. He may reapply, or he may not. He is thirty-four. He has a partner who is pregnant with their first child. He has been working on this question for six years. If the rejection is final, he will probably leave academic research.
The question his work was asking may or may not have been important. Nobody will know, because the work is not going to be done. It will not be done in another country either, because the infrastructure for this kind of basic research exists in fewer places than it once did, and because the specific scientific community he was part of is geographically concentrated in the institutions whose funding is being cut. The question will not be answered. The answer, whatever it might have been, will not be available when some future exercise condition arrives that would have depended on it.
This is not a scene from a thought experiment. Some version of this scene is occurring, today, in hundreds of offices across the federal research apparatus, and in the adjacent university laboratories and regulatory workforces and statistical agencies and diplomatic posts where holding labor is currently being evaluated by criteria that cannot see what it produces. The program officer is doing her job. The applicant is accepting a rational decision made under conditions of scarcity. The agency leadership is responding to the budget environment they have been given. Nobody in the scene is a villain. Everyone is acting reasonably within the constraints that have been set for them.
The warehouse is being emptied this way. Not through any specific decision that any specific person could be asked to defend. Through the accumulated effect of a thousand reasonable rejections issued under criteria that were designed for a different kind of institution. The rejections are not announced as cuts to the warehouse. They are announced as prudent stewardship of limited resources. The resources are limited because they have been made limited by prior choices whose full consequences are not yet visible. The consequences will be visible later, in futures that the choices are currently closing.
The essay cannot say which future is being closed, because the foreclosed futures do not announce themselves. They arrive, or fail to arrive, decades after the closure. What the essay can say is that the closure is happening. That it is happening through the accounting regime that cannot see the closure it is producing. That the program officer, the applicant, the agency leadership, and the broader public are all, in different ways, unable to see what is being lost because the loss does not appear in any document they read.
The warehouse held the United States across the twentieth century. What it held was assembled slowly, by people whose names are mostly not recorded, through institutional arrangements whose fragility was not apparent while they were functioning. The arrangements are now under strain. The holding is now being cut. What will be in the warehouse thirty years from now, and what the United States will be able to do with what is there, is being decided right now, in offices like the one I have just described, by people who are not aware that they are making the decision.
The essay ends here, with that program officer still at her desk, still reviewing the application, her afternoon unremarkable, her work part of the quiet and uncelebrated labor by which the warehouse is kept or not kept. She is one of tens of thousands. The warehouse exists because people like her, across the federal government, show up and do work whose value will be apparent only later, if ever. The warehouse is being diminished because the same accounting regime that cannot credit her work cannot register its loss. She will leave her office at five-thirty. She will drive home. She will not know that she has participated, in a small way, in closing a future. Nobody will. That is the nature of the work.
Section IV of this essay treated space nuclear capability as a single compressed case. The companion piece is the long version: six decades of holding labor across eleven presidencies, thirteen programs, and one reactor that flew. It is the case from which this essay's warehouse frame was built.