The U.S. Politico–Military–Industrial Complex

John A. Alic

doi:10.1093/acrefore/9780190228637.013.1870

Summary

The three large military services—Army, Navy, and Air Force—comprise the core of the U.S. politico–military–industrial complex. They dominate decision making on multi-billion dollar weapon systems and the operational concepts these are intended to embody. The armed forces need private firms to realize their visions of new weaponry, since government has limited capacity in engineering design and development and limited production facilities. Running a successful defense business means giving the services what they want, or think they want, whether this makes technical and operational sense or not; thus industry caters to the views of the services, and while it seeks to influence them, does so mostly at the margins.

The political dynamics of the complex take place in two primary domains, only loosely coupled. The first is largely contained within the Defense Department. This is the main arena for conflict and bargaining within and among the services and between the services, individually and collectively, and Pentagon civilians. Most of what happens here stays hidden from outsiders. Service leaders generally seek to resolve disagreements among themselves; the goal, often although not always achieved, is to present a united front to civilian officials and the public at large. The second domain extends to the rest of government, chiefly Congress, with its multiple committees and subcommittees, and the White House, home of the powerful Office of Management and Budget among other sources of policy leverage.

The complex as a whole is an artifact of the Cold War, not greatly changed over the decades. Repeated efforts at restructuring and reform have led to little. The primary reason is that military leaders, senior officers who have reached the topmost ranks after lengthy immersion in generally conservative organizational cultures, usually have the upper hand in bureaucratic struggles. They believe the military’s views on choice of weapons—the views of seasoned professionals—should have precedence over those of civilians, whether Pentagon appointees and their staffs, elected officials, or outside experts. They usually prevail, since few of the political appointees on the civilian side of DoD and in policy-influencing positions elsewhere can command similar authority. If they do not prevail on a particular issue, service leaders expect to outwait their opponents; if they lose one battle over money or some cherished weapon system, they anticipate winning the next.

Bureaucracies and Politics

This article consists of two largely self-contained parts. The first surveys the politico–military–industrial complex (PMIC), with considerable attention to its origins. The second looks in more detail at major acquisition programs, the multiyear or multidecade and multibillion dollar undertakings that result in systems and equipment including aircraft and spacecraft, naval vessels, ground combat vehicles, and guided missiles and other munitions. These programs lie at the center of the PMIC. They offer defense firms lucrative opportunities, pose large risks for the armed forces and the rest of government, and sometimes provoke lasting controversies over strategic issues. James Q. Wilson (1989), pioneering student of organizational behavior in the U.S. government, tells us to view organizational entities from the bottom up, as if each is unique unless and until shown to be part of a class. This article adopts a similar approach.

The political dimensions signaled in the title take two primary forms: those internal to the Department of Defense (DoD) and those that involve other parts of the federal government. The two domains are linked, formally at the top—the president is putative commander-in-chief and with the consent of the Senate appoints high-level Pentagon civilians, the chiefs of staff of the military services, and the chair of the Joint Chiefs of Staff (JCS)—and less formally elsewhere through connections such as temporary attachment of military officers to congressional panels and the National Security Council (NSC).

The discussion relies in considerable part on illustration, example, and anecdote. DoD is very large and comprised of so many disparate elements as to be difficult just to describe, which limits the usefulness of the sorts of analytics associated with economics and other quantitatively inclined social sciences. The Marine Corps, for instance, is administratively part of the Navy; nonetheless, the Commandant of the Corps is a member of the JCS on a par with the Chief of Naval Operations. The Coast Guard, administratively within the Department of Homeland Security, is part of the “Naval Service” along with the Marine Corps and Navy, and may come under the Department of the Navy in times of war. The 2019 Defense Authorization Act established the Space Force; just as the Marine Corps is attached to the Navy, the Space Force is housed within the Department of the Air Force. Its head, the Chief of Space Operations, became the eighth member of the JCS in late 2020.

The most cursory glance at DoD will reveal stark differences between the armed forces and the civilian side of the organization, the Office of the Secretary of Defense (OSD) and the military departments, each with its own secretary. Political appointees occupy the top civilian positions. Some take office knowing little of military affairs and leave before learning much. Military officers enter at the bottom and so long as they meet the expectations of their superiors can expect to rise step by step along structured career paths. The incentives breed conformity and solidarity. Unlike private firms or other parts of government, no one comes into the uniformed military at high levels from outside to shake things up.

Combatant commands organize personnel from the various services into ostensibly unified organizational structures. As of 2020, with U.S. military personnel serving in some 140 countries, there were six geographically defined commands, plus five others established on some other basis (e.g., Cyber Command). Other attributes of the armed forces include independent systems of justice and specialized schools for education and training of both officers and enlisted personnel. When not at war, the U.S. military studies, plans, and trains. Doctrinal commands review, revise, and codify principles and prescriptions for how to fight—guides to action—preparing manuals and materials as a basis for classroom discussion and field exercises. There is hardly anything like this, in its variegated yet self-contained totality, elsewhere in government or in civil society.

Congress, the second pole of the “iron triangle” (Adams, 1981), is easier to grasp—confusingly messy but open and transparent. The third pole, the defense industry, consists of profit-seeking firms that operate much as do their counterparts in other industries, studied by academics and tracked by the business press and financial analysts.

Because history can reveal much that had earlier been obscure—and because the PMIC has changed relatively little since the end of the Cold War, arguably because change happens only when provoked by something close to crisis—a good deal of the discussion that follows looks back to the decades in which the PMIC as it existed in the first two decades of the 21st century took shape.

Spending

Being concerned with politics, this article is much concerned with money, the glue that binds the disparate pieces of the PMIC. As Table 1 shows, spending related in one way or another to national security will exceed $1 trillion in fiscal 2021, some 22% of the federal total. By itself DoD accounts for half of all discretionary spending, funds appropriated annually as opposed to legally mandated outlays, including social security and interest on the national debt.

As the notes to Table 1 suggest, the meanings attached to national security are elastic and a number of agencies outside DoD have claims on the national security budget: The penumbra of the PMIC is expansive. Most obviously, Table 1 does not include any portion of spending by the National Aeronautics and Space Administration (NASA), because no accounting is available that would permit even rough estimates of the civil–military split. If not tied as closely to the military as in the past, NASA nonetheless remains an adjunct of the PMIC. When established in 1958, it absorbed numerous programs run by one or another of the services along with those of the National Advisory Committee for Aeronautics, which had supported both military and civil aviation since the 1920s. Many military space programs later migrated back to the services, and DoD’s spending on space has exceeded that of NASA since 1982 (Chadwick et al., 2011, p. 225). Still, as in aviation, there remains much overlap in terms of technology and sometimes in fielded systems as well; the Space Shuttle, conspicuously, was sized for military payloads (Hays, 2006).

Defense spending after adjustment for inflation has risen far above the levels of the second half of the 1950s, when the Cold War arms race was gaining momentum and the PMIC solidifying (Figure 1). Yet the world hardly seems as dangerous a place for the United States today as then, when the Soviet Union was expanding its nuclear stockpile and developing intercontinental missiles. The DoD budget in Dwight D. Eisenhower’s last year as president, 1960, came to $398 billion in 2012 dollars, little more than three-fifths the 2020 figure of $648 billion, likewise in 2012 dollars (OMB, 2020, Table 6.1). Moreover, current spending exceeds that in 1989, the final year of the Reagan administration’s defense buildup, by some 14%. The subsequent peace dividend, as it was called, lasted roughly a decade; outlays then jumped following the September 2001 terrorist attacks on New York and Washington and ensuing wars in Afghanistan and Iraq. As the U.S. commitment to these conflicts waned, so did the budget. Yet in 2017 it began to rise once more. The new bulge demanded a new explanation. This was found in needs for “Modernizing capabilities in the air, maritime, and land domains to enhance lethality” (DoD, 2019b, pp. 1–3) and support “the irreversible implementation of the [2018 National Defense Strategy] … to prepare for a potential future, high-end fight” (DoD, 2020a, pp. 1–3). Given that the United States in 2019 accounted for an estimated 38% of global military spending (Tian et al., 2020), it has sometimes seemed as if the nation was preparing to fight the rest of the world.

Open in new tab

Figure 1. Real (inflation-adjusted) U.S. Defense spending, 1940–2020.

Note: Budget function 050 or equivalent; fiscal year transition period (1976–1977) omitted; 2020 estimated.

Source: OMB (2020, Table 6.1).

On average, a little over half of DoD spending—typically 50%–55%—flows to some 70,000 business firms for purchase of goods and services (Peters, 2020). Private companies run cafeterias and clean offices, maintain computer systems and aircraft, and supply everything from small arms to spacecraft and nuclear-powered aircraft carriers. Much of the remainder of the budget covers salaries and benefits for DoD’s 775,000 civilian employees and 1.44 million active-duty military personnel (DoD, 2020g, p. 291).

Money buys technology, and DoD pays a lot for stealthy bombers like the forthcoming B-21. Technology enters through the acquisition accounts, budget categories that include both procurement and R&D (research and development), or in Pentagon terminology RDT&E (research, development, test, and evaluation). Intelligence, too, has become big business, in large part because of the costs of spacecraft, their payloads, and the terrestrial systems that scan, sort, and interpret data and images.¹ In addition to RDT&E on new systems, defense agencies fund a wide range of less targeted work. Pentagon managers have long understood that “The real difference in performance between a weapon system and its predecessor was usually not the consequence of one, two, or three scientific advances or technological capabilities but was the synergistic effect of 100, 200, or 300 advances, each of which alone was relatively insignificant” (DoD, 1969, p. xviii). To take one example, DoD has been channeling funds to artificial intelligence (AI) research for decades. While defense firms conduct nearly all the engineering work on systems slated for procurement, universities undertake a good deal of more fundamental research (see Leslie, 1993, among much else on universities and their linkages with defense agencies and defense spending).

Eisenhower and the Military–Industrial Complex

A five-star Army general before winning the 1952 election, Eisenhower became “the principal architect of the modern institutional Presidency” (Hoxie, 1983, p. 589). The PMIC took on its lasting contours during his administration and Eisenhower gave the military–industrial complex its name in his much-remarked January 1961 farewell address, which now reads as something of a mea culpa.

For other reasons too, Eisenhower’s tenure as president marks a watershed. Over much of the 1950s the politics of military policy and defense spending were more tangled than before or since. As summarized in Box A, numerous technological families were advancing rapidly—electronic systems, jet propulsion, nuclear weapons and nuclear submarines, rockets and missiles—with disruptive consequences for both fighting wars and deterring them. These developments added fuel to ongoing disputes among the services, surfacing during World War II and flaring afterward, as evidenced by the 1949 “Revolt of the Admirals” (Caraley, 1966). The clashes, accompanied by friction, fog, and confusion, some of this deliberate, centered on roles and missions—roles being more broadly conceived (sea control), while missions refer to specific tasks (antisubmarine warfare). Roles and missions heavily condition budget shares. Who will do what in war with which new weapons and the budget allocations they deliver? How, for instance, should responsibility for close air support—close, that is, to the frontlines of ground combat—be split between the Air Force and Army? With jet propulsion becoming the norm for tactical aircraft, limiting their low-speed, low-altitude maneuverability, hence ability to loiter over battlefields and target enemy forces? And with gas turbine-powered helicopters on the horizon, promising weapons-carrying capabilities much superior to piston-engined helicopters? Such questions remain in dispute: the Army worries that F-35s will be less capable in close support than the F-16s to be replaced and that the Air Force does not much care; the Air Force touts the F-35 and argues, as it has for decades, that the Army’s attack helicopters will be deathtraps in the face of any reasonably capable opposing force.

Box A. Military Technologies: Episodic Evolution

Punctuated equilibrium—the idea that technological families advance incrementally for often considerable periods until some radical development emerges, the jet engine or intercontinental ballistic missile (ICBM)—has become a familiar trope in studies of innovation (Mokyr, 1990). For military technologies, the 1950s brought the first full flowering of numerous such bursts of rapid advance, most of them seeded in the preceding decade (see, e.g., Alic, 2007, and citations therein). Earlier episodes—gunpowder, the machine gun, steam-powered warships—had come with little clustering, a pattern that began to change with widespread mechanization around the time of World War I.

During World War II, the federal government directed more funding to radar than any other technological family, leading to applications ranging from surveillance over trackless reaches of air and sea to proximity fuzes for munitions. German V-2s prefigured ballistic missiles of longer range and greater accuracy. Design calculations for nuclear warheads became the first task for the first U.S. electronic digital computer. Transistors emerged in 1949, integrated circuits in 1959–1960, and Moore’s law then took over, with a doubling in capability every two years or so—a classic example of ongoing incremental innovation with implications for everything from computers themselves to ICBMs, smart weapons, war games, and now-ubiquitous forms of intelligence gathering and surveillance.

Only when combined in systems—electronic warfare suites, fly-by-wire flight controls for stealthy but aerodynamically unstable aircraft like the F-117—would the full potential of this generation of technologies be revealed. Specially modified B-29s dropped the atomic bombs that devastated Hiroshima and Nagasaki. If not a complex system in later senses of the term, the B-29 was far more complicated than its predecessors, and by the time production ended had cost more, around $3 billion (Vander Meulen, 1995, p. 100), than the bomb-building Manhattan Project itself, which absorbed some $2.2 billion (Buck, 1983, p. 34). By the end of the 1950s, ICBMs brought a more telling example of systemic complexity, since reduction to practice required integration of light and compact warheads, powerful yet reliable rocket motors, high-temperature materials for heat-resisting ablative nose cones, and intricate electromechanical guidance and control componentry. In its turn, GPS depended on constellations of orbiting satellites, atomic clocks, and techniques for code-division multiplexing—one reason that, like most system-level innovations, much time elapsed between conception and realization, in this case three decades (Getting, 1993). In another example, work on pilotless aircraft began during the World War I era but advanced only slowly until sophisticated mathematical models of aerodynamics and flight control combined with powerful digital processors made possible recent generations of unmanned aerial vehicles. Until about 2010, computational capacity constrained AI applications; it took denser ICs, some of these designed for specialized tasks such as image processing in gaming for machine-learning AIs to move from research to applications. As this suggests, technological change since the Cold War has once again begun to cascade, with implications that have yet to be fully plumbed.

One of the attributes of complex systems is that no one person, or handful of people, can understand their functioning in detail: grappling with system-level behavior becomes a collective task for specialists with expertise in multiple domains. Even then, understanding may be partial. Machine learning takes off from human-written code, in principle comprehensible and explainable. Once self-learning begins, however, the code changes itself, spontaneously—that is what “machine learning” means. The changes cannot be anticipated and it may well be impossible even in hindsight to say why the system did what it did (JASON, 2017). Many years ago Admiral Chester Nimitz told Congress, “no weapon that is effective and efficient has ever been outlawed” (Nimitz, 1949, p. 346). As he noted, although poison gas had spread terror among unprepared troops during World War I, the next war saw no gas attacks; military decision makers had concluded that even if dispersal could be controlled, gases would not be disabling or lethal against well-trained and well-equipped opponents. The likely consequence for AIs, one that policymakers have mostly talked around: if DoD concludes that killer robots lacking a trigger-pulling “human in the loop” promise battlefield advantages—which since they will be able to act at superhuman speed seems likely—the armed forces will get such weapons (Boulanin & Verbruggen, 2017; Watts, 2016).

Stalin died two months into Eisenhower’s presidency, and a second set of issues, centered on the import of new technologies and systems for larger questions of national security at a time of much uncertainty in international affairs, stretched far beyond the Pentagon. The Kremlin, if opaque, had been something of a known quantity under Stalin’s rule; no one knew what might follow. Intelligence was poor, abysmally so until orbiting satellites began returning images of the Soviet interior to reveal military installations including missile launch sites and test ranges. Military predilections for worst-case scenarios and bureaucratic pressures on intelligence agencies led to exaggerated estimates of Soviet military capabilities and intentions (Karber & Combs, 1998; Rovner, 2011); if much of this was little more than speculation, it fed into ongoing debates over U.S. policy.

Eisenhower’s task, then, was not so much to map the way ahead—an impossible task, given the uncertainties—but to move the nation, and the Pentagon, with its internal schisms, into the unknowable future (Box B). He did so exercising firm control over policy. As president, Eisenhower “made all the vital decisions and firmly enforced them,” in the words of national security advisor Gordon Gray (Smith, 2012, p. 569). The choices made in the 1950s, some of these during the Truman administration but chiefly under Eisenhower, and the nature of the PMIC that emerged, had consequences that in later years would come to seem nearly irreversible.

Box B. Eisenhower and Military Affairs

When Eisenhower took office, the U.S. nuclear stockpile consisted solely of atomic bombs. In late 1952 the Atomic Energy Commission (AEC) demonstrated nuclear fusion, and in August 1953 the Soviets carried out a test of their own. These were proofs of principle: usable hydrogen warheads were several years away, and more time still would pass before the full ramifications sank in. Like most of those in the Pentagon and State Department, the president thought war with the Soviet Union unlikely. Determined to keep taxes low and the federal budget in balance—for Eisenhower these were near-obsessions—he opted to rely heavily on nuclear weapons. An early document from the NSC encapsulated the president’s views: “A vital factor in the long-term survival of the free world is the maintenance by the United States of a sound, strong economy. For the United States to continue a high rate of Federal spending in excess of Federal income, at a time of heavy taxation, will weaken and might eventually destroy that economy” (FRUS, 1952–1954, p. 307). This was not, as might be thought today, meaningless posturing: NSC 149/1 would not be declassified until 1990.

At first only the Air Force, with its modified B-29s and lumbering B-36s, fewer in number, could deliver atomic bombs. The Army and Navy sought to end the Air Force monopoly on nuclear weapons by pressing the AEC for warheads tailored to their respective missions and warfighting domains. The congressional Joint Committee on Atomic Energy (JCAE) joined in, urging the AEC to design and build more, and more varied, weapons. Soon the weapons laboratories were producing atomic artillery projectiles, land mines, and warheads that could be fitted to short-range rockets and missiles for the Army; the Navy got gravity bombs suited to carrier-based aircraft along with atomic torpedoes and depth charges (Schwartz, 1998, pp. 87–91). The bombs were vital for Navy interests. At the end of the Korean War the U.S. fleet numbered some 1,100 vessels, whereas the Soviet Union, historically a land power, had neither means nor motives for a blue-water navy. By carving out a role for carrier-based strike aircraft to supplement to the big bombers of the Strategic Air Command (SAC), the Navy managed to preserve the core of its armada (Miller, 2001). Later, its missile-carrying Polaris submarines sealed the Navy’s place alongside land-based bombers and ICBMs in the nuclear triad.

All this was costly, even if nuclear warheads themselves were not, once the necessary production facilities had been built—a task largely completed during the Truman administration. As early as 1951, JCAE chair Senator Brien McMahon stated that “If we mass-produce this weapon … the cost of a single atomic bomb will become less than the cost of a single tank” (Williamson & Rearden, 1993, p. 151). He was right, and to Eisenhower, who could see no reason why the Kremlin would risk war, a forbidding nuclear stockpile promised to protect the nation’s budget as well as the peace. Yet the overall policy made less and less sense as the decade advanced, bringing thermonuclear weapons, ICBMs, and the specter of unfathomable overkill.

When Eisenhower entered the White House, the nation’s stockpile consisted of 1,500 fission bombs. When he left, it numbered 20,000 warheads, many of these fusion devices. The average yield (explosive power) had risen from about 35 kilotons (TNT equivalent; the bombs dropped on Japan were in the range of 15–20 kilotons) to 1,000 kilotons (DOE, 2002, p. D-1). By the end of the decade SAC fielded some 1,000 jet-propelled B-47s and more than 600 intercontinental B-52s, the first Polaris submarines lurked in the Norwegian Sea, the initial installment on what would become 1,000-plus land-based ICBMs were launch-ready, and early warning radars in Alaska and Canada had begun feeding electronic communications to command centers in the lower United States. This was where the money went. Never in the 1950s did AEC spending reach more than about 5% of DoD outlays (Budget of the United States Government, various years, various pages).

As policy, these decisions were muddled at best, incoherent at worst. Under constant pressure from the Pentagon and the JCAE, the AEC churned out growing numbers of warheads. The Air Force, given more bombs, asked for and got more aircraft to carry them. By the second half of the 1950s SAC’s blueprint for “general war” specified over 1,100 targets, some 820 in the Soviet Union and the rest in Warsaw Pact countries, China, and North Korea; many of these targets were to be hit with multiple warheads (Wellerstein, 2016). In addition, large numbers of smaller tactical or battlefield nuclear weapons had been shipped to Europe as backstops for the forces of the North Atlantic Treaty Organization. Yet U.S. atomic bombs had not deterred North Korea or China in 1950, the Soviet Union was building its own nuclear arsenal, repeated simulations and field exercises showed that millions of noncombatants would die should even small tactical weapons be used in densely populated Europe, and uncontrollable escalation to global strategic exchange would anyway likely follow (Carter, 2015, p. 241; more generally see Craig, 1998).

A Sketch of the PMIC

Before World War II the Army and Navy kept private businesses at arm’s length unless surge capacity was needed to meet wartime demand. Since then the armed forces have relied on industry for both R&D and production. While DoD has 60-plus research and engineering laboratories, with minor exceptions these have little capability for system-level design and development. The government still owns some production capacity, including aircraft factories built during World War II, but except for three ordnance and ammunition plants all have been given over for operation by contractors, along with the great majority of maintenance and repair facilities (DoD, 2018b, p. 97).

The Korean War solidified the shift to near-total reliance on private industry. The unexpected new conflict delivered a shock that transformed U.S. military policy, and for this reason the first half of the 1950s marks the nearest thing to a well-defined origin for today’s PMIC (Box C). Procurement declined following the 1953 armistice in Korea, but to nothing like the levels of the late 1940s. And even as procurement fell, RDT&E continued to climb as the armed forces searched for technology-based “force multipliers” to offset numerical disadvantages of the sort experienced in Korea, the Pentagon sending an ongoing flow of dollars to contractors for exploratory research, design studies, and prototype development (Alic, 2007, pp. 31–35). By the end of the 1950s, DoD had funded RDT&E on over 120 new aircraft programs alone (Lorell, 2003, p. 80). Only the National Nuclear Security Administration (NNSA) continues to resemble the old arsenal-like acquisition system, and then only in part. Not only is NNSA a civilian agency, housed since the 1970s in the Department of Energy, but from the establishment of the AEC in 1946 contractors have managed the government-owned nuclear weapons complex.

Box C. Toward the PMIC: Slow Start, Fast Finish

Before World War II the United States had nothing resembling a defense industry. In the latter part of the 19th century Congress, in a departure from peacetime practice, directed a portion of shipbuilding for the “New Navy” to private yards (Alic, 2014). Until then, leaving aside the Civil War, internal supply bureaus had designed and produced almost all durable goods needed by the armed forces (Alic, 2007, pp. 35–43). On occasion the bureaus might purchase rights to proprietary designs, as for “patent arms,” or turn to specialized suppliers for items such as precision instruments, but these were exceptions until World War I brought fast-paced advances in aircraft and submarines, in which the United States lagged the major European combatants. As the nation moved closer to joining the war, no one in Washington could bring much technical understanding to designs for either. Private firms were the only recourse. The infant U.S. industry proved unable even to replicate British and French aircraft satisfactorily, while U.S. submarines proved greatly inferior to those of Germany.

Neither then nor later did either service try in a committed way to develop internal capabilities in aviation-related technologies (Alic, 2007, pp. 56–59). The Navy did edge into aircraft design and production, but the motivation was in large part to create a yardstick against which to measure the cost claims of private firms (Trimble, 1990). The Army’s air arm remained even more dependent on industry; as late as the mid-1930s a single engineer staffed the office responsible for all Army Air Corps bombers (Kelsey, 1982, p. 43). Accustomed to designing surface warships down to small details, at the time of World War I Navy personnel lacked the technical knowledge and experience to do so for submarines. Instead, the service set down general specifications and contracted with private shipbuilders for detailed design and construction (Weir, 1991). The results were unsatisfactory and the Navy made a fresh start in the 1920s by copying German technology before setting an independent course.

Preparation for World War I had begun far too late. Mobilization was at best improvisatory, at worst chaotic (for an overview, see the first chapter in Wilson, 2016). The experience left a deep imprint on military planners and they set out to ensure that any future buildup would be more orderly. Resources were the obstacle: Congress so starved the services for funds that as late as 1940 appropriations for agricultural research exceeded those for military research (NSF, 1963, p. 163). With few exceptions, then, the United States once again went to war with weapons that were often inferior to those of adversaries and allies alike. Yet U.S. industry could manufacture aircraft, ships, tanks, and guns in unmatched numbers, and did so.

The production buildup got underway in the late 1930s with U.S. manufacturers filling orders from Britain and France, then accelerated, fueled by massive injections of public funds. Federal agencies paid for some two-thirds of World War II investment in new plant and equipment, most of it operated by private firms under contract (see Wilson, 2016, for a concise and balanced treatment of World War II mobilization, the subject of a vast literature of widely varying quality).

Even so, the Cold War defense industry and PMIC did not spring full-blown from World War II. Once Japan surrendered, defense agencies canceled contracts by the thousands, and the Truman administration imposed stringent ceilings on new spending, holding to these despite the protests of military leaders (Condit, 1996). Employment fell precipitously, with jobs in aircraft production, the largest of all wartime industries, dropping from a 1943 peak of over 1.3 million to fewer than 250,000 in 1946 (Modley & Cawley, 1953, p. 48). Nor was there much money for new technologies, leaving aside jet propulsion and nuclear warheads, of too obvious significance to ignore; R&D even on guided missiles declined by more than half from 1946 to 1947 (Lonnquest & Winkler, 1996, p. 19). It took the Korean War to revive the defense industry.

The now-famous document known as NSC 68, a forceful case for much greater military spending, went to the White House in early 1950, months after the first Soviet atomic bomb test and Mao Zedong’s consolidation of power in China. In June, North Korean invaded its neighbor to the south, catching U.S. forces in the region woefully unprepared. By December, Chinese forces flooding down the peninsula had pushed the U.S. Eighth Army into disorganized retreat, and Truman approved a fourth and final version of NSC 68. Since declassification in 1975, a good deal has been written about its significance (Young, 2013). Here the point is simply that, along with the bloody fighting in Korea, NSC 68 set the seal on the Pentagon’s search for military advantage through technology.

The Korean War put idle production capacity back to work and profits back on defense industry books. The DoD budget quadrupled, and RDT&E rose even more (DoD, 2020g, p. 80). With money again flowing, automakers and other firms that had produced military goods during World War II returned, sometimes establishing entire new divisions to take on defense work, and companies such as Raytheon, a pioneer in radar, expanded, as did newer entrants including the semiconductor pioneer Texas Instruments. By the end of the decade, top-secret CORONA satellites had returned the first photographs of the Soviet interior, and in 1965 the first commercial satellite, Early Bird, launched; in that year more money flowed to activities directed at space and spacecraft (military and civilian) than to tactical and strategic rockets and missiles (Hincks et al., 1965, p. 7). The aircraft industry had become the aerospace industry.

Acquisition

All acquisition dollars do not matter equally for policy. Defense agencies handle routine purchases much as the Forest Service, for instance, does in buying pickup trucks and gasoline to fuel them. By contrast, money flows out by the billions for Major Defense Acquisition Programs (MDAPs) such as the Navy’s Ford-class aircraft carriers and the tri-service F-35 Joint Strike Fighter. Defined as programs with RDT&E expenditures of at least $480 million or procurement expenditures of $2.79 billion (both expressed in 2014 dollars), for decades DoD has had 80 or more MDAPs underway. (Peck & Scherer [1962] and Scherer [1964] are essential early studies of acquisition, still relevant.) Some are all-new, others modifications. Aegis missile defense systems, for example, have been updated almost continuously since first deployed in the early 1980s, and DoD is requesting another $1.7 billion in 2021, $1 billion of this for RDT&E (DoD, 2020d, p. 4)

DoD’s fiscal 2021 budget request includes 88 MDAPs, totaling $89 billion (DoD, 2020d, p. 3). Each year the Government Accountability Office (GAO, a congressional agency) surveys those for which DoD provides annual reports to Congress (there are various exceptions, as for tightly classified “black” programs). In its latest review GAO states that the 85 programs examined, plus 8 planned future MDAPs, comprise a “portfolio” valued at more than $1.8 trillion, based on Pentagon estimates (likely to be low) of expenditures expected over the acquisition cycle of each (GAO, 2020). This is a better indicator of MDAP significance than yearly spending.

MDAP outcomes, despite ongoing reorganizations, policy shifts, and other reform efforts, have not improved much over the years. Budget and schedule overruns, often large, and unsatisfactory functional performance once systems and equipment reach the field, remain endemic. DoD began preparing annual MDAP reports half a century ago, and GAO (which has been renamed several times) began analyzing them (GAO, 1970; see also GAO, 1988). Over the years, the agency has developed a standardized template for evaluating MDAPs that draws in part on managerial practices found in commercial business firms, and the language in its reports has changed accordingly. Yet the overall thrust of recent findings remains remarkably similar to GAO’s earliest assessments. Not-for-profit federally funded research and development centers including the Institute for Defense Analyses and RAND also study acquisition, though not so regularly as GAO, reaching generally similar conclusions (see, e.g., Light et al., 2017; McNicol & Wu, 2014).

DoD buys services as well as goods in large volume. Private firms maintain equipment in war zones as well as domestically, provide base security, foreign military training, and a good deal of space-based surveillance. During the Vietnam War, the on-the-ground ratio of military to contract personnel had been about 5:1; in Iraq, contractors were present in about the same numbers as U.S. military personnel (CBO, 2008).

The Defense Industry

Some 2000 subcontractors and suppliers feed parts, components, and subassemblies to Newport News Shipbuilding, the only yard able to build nuclear-powered aircraft carriers such as the Ford class (O’Rourke, 2019). Named for President Gerald R. Ford, the Navy is buying four of these at some $13 billion apiece. Supply chains funnel in everything from commodities such as welding rod and paint to nuclear power reactors and catapults to sling F-35s into the air.

Many of the items purchased by or for DoD are unique. Some companies build their entire businesses around these sorts of products, such as fuzes for munitions and radiation-hardened microchips. Others supply commercial and defense firms alike. Most of the 200 U.S. companies making printed circuit boards, for example, sell some portion of their output into defense supply chains (DOC, 2017). These supply chains are international in two respects. Final systems incorporate many imported parts and components: The bulk of the world’s printed circuit boards are produced outside the United States, as are many of the electrical and electronic components plugged into them. Second, a number of firms with substantial presence in U.S. defense markets are based abroad, including BAE Systems, one of the larger DoD contractors. The defense industry as a whole, then, is something of a virtual sector, and not all that large. DoD spending on goods has recently come to around $150 billion annually, a figure that corresponds to roughly two-and-one-half percent of U.S. gross manufacturing output.

The structure of the industry has altered a good deal since World War II, driven in part by technological change and in part by policy. Aircraft companies had to expand and diversify their engineering staffs and production capabilities to move into aerospace and electronics. As early as 1953, Convair, itself the product of a World War II merger between two aircraft manufacturers, combined with Electric Boat, long-time builder of submarines, to form General Dynamics. At the end of that decade the Textron conglomerate purchased Bell Aircraft. Two major combinations followed in 1967: McDonnell and Douglas, long-established aircraft firms named after their respective founders, merged to form McDonnell Douglas, and Rockwell absorbed North American Aviation. Privatization of services meanwhile began to attract nontraditional suppliers.

With the end of the Cold War and the decline in DoD spending shown earlier in figure 1, defense firms cut their payrolls: Between 1987 and 1996 defense-related employment fell by two-fifths.² Something of a merger wave accompanied retrenchment. In 1992, General Dynamics sold its aircraft division to Lockheed, which later merged with Martin Marietta to become Lockheed Martin. Two years later Northrup and Grumman combined. Boeing purchased the defense and aerospace operations of Rockwell in 1996 and absorbed McDonnell Douglas the following year. In the latest such deal, the merger of United Technologies and Raytheon closed in April 2020.

Consolidation among prime contractors and reports of exits by lower-tier firms in search of better business opportunities have triggered periodic expressions of anxiety. Fewer prime contractors, it is said, promise still higher price tags for goods and services sold to the military. Yet head-to-head competition as seen in commercial markets has never been that common in defense, and post–Cold War consolidation seems to have had only minor effects on acquisition costs, through the turn of the century at any rate (Carril & Duggan, 2020; Hensel, 2010). Import dependency has also excited worry. This goes back to the 1960s and 1970s, when the steel, machine tool, and semiconductor industries, among others, faced stiffer competition from foreign rivals, with some U.S. firms exiting and others narrowing their product lines in search of viable niches (see, e.g., Alic et al., 1992, pp. 257–278 and 350–354).

Some of the warnings reflect little more than special pleading by affected firms and their political allies. In any case, DoD gives considerable attention to the industrial and technology base; planners have long understood that if supply chains do not win wars, shortages of fuel and ammunition can mean losing battles. Although recent Pentagon assessments note worrisome trends, in general they portray a buoyant industry (DoD, 2020i), with the larger defense firms, at least, more profitable than their commercial counterparts (Wang & San Miguel, 2012). Washington also has powerful tools such as the Defense Production Act, passed at the time of the Korean War, to support essential supply chain links, while Congress has several times gone so far as to bail out contractors threatened by bankruptcy, as it did with loan guarantees for Lockheed in 1971 and ten years later for Chrysler, at the time manufacturing M-1 tanks for the Army. More recent worries have centered on whether DoD, so often derogated for bureaucratic sluggishness and an innovation-strangling acquisition rulebook, can take advantage of fast-paced technical advances driven more by commercial than military demand (Box D).

Box D. Rapid Innovation

The Pentagon warns that “state competitors and non-state actors” have access to “advanced computing, ‘big data’ analytics, artificial intelligence, autonomy, robotics, directed energy, hypersonics, and biotechnology” and that “to fight and win the wars of the future … will require changes to industry culture, investment sources, and protection across the National Security Innovation Base” (DoD, 2018c, p. 3). These are not so much new concerns as variations on familiar themes, and DoD has sought repeatedly to speed innovation by short-circuiting normal acquisition procedures or otherwise enhancing access to emerging technologies.

The initiatives go back at least to the early 1950s, when the Air Force set up an office known as Big Safari “to manage special airborne reconnaissance platforms that spied on communist states” (Ehrhard, 2010, p. 6). Tight classification served as justification for skirting many of the formalized steps in acquisition. More recently, defense agencies have turned to “other transactions authorities,” originally authorized in 1958 to help NASA expedite response to Sputnik and later extended to DoD. These provisions permit contracts to be written “generally exempt from federal procurement laws and regulations” (Schwartz & Peters, 2019, summary page; also see DoD, 2019f). Annual expenditures under Other Transaction Authorities reached more $2 billion during the Trump administration. Spread over more than 100 programs, most contracts have been small and hardly anything is known about outcomes.

Policies for exploiting commercial technologies have been harder to realize, as shown by a succession of efforts in microelectronics. Federal spending did much to build Silicon Valley. DoD and NASA placed the first big orders for integrated circuit chips (for ICBM guidance and the Apollo moon landing, respectively). With rising output and experience, chip costs fell and manufacturers slashed prices, opening lucrative markets for industrial and consumer products. Defense markets were small and specialized by comparison, and semiconductor firms largely abandoned them (Alic et al., 1992, pp. 255–282). Ever since, the Pentagon has tried to pull these firms back into its orbit, with only limited success. This was one objective of the Strategic Computing Program, run by the (Defense) Advanced Research Projects Agency, DARPA, in the 1980s (Roland with Shiman, 2002).

There have also been broader initiatives. In the 1990s, the Pentagon set up the Advanced Concept Demonstration Program to speed adoption of new technologies, including those originating in civilian industries. Congressional legislation in 2003 established the Defense Acquisition Challenge Program with similar objectives. Government-provided venture capital seemed another avenue, with the Central Intelligence Agency starting the not-for-profit venture capital firm known as In-Q-Tel in 1999 (DoD, 2012, pp. 84–87). And in 2015, the Pentagon established its Defense Innovation Unit, with offices now in Austin and Boston as well as Silicon Valley.

Useful as these and other programs may be for particular purposes, they remain small in scale. DARPA, the Pentagon’s storied civilian-run innovation incubator, spends about $3.5 billion annually on contract RDT&E directed in part at systems concepts and in part at research intended to strengthen the technology base in fields now including synthetic biology and neurosciences (DoD, 2020e, pp. D-2 & D-2A). The Defense Innovation Unit gets a tiny fraction of this, recently $25 million to $30 million annually (DoD, 2020b, p. 173).

Appropriators also give DoD many tens of billions of acquisition dollars each year for classified programs. The 2021 budget requests some $26 billion for classified RDT&E (DoD, 2020e, p. IVA) and $61.5 billion for classified procurement (DoD, 2020c, pp. A-26B, N-37B, N-41B, & F-9B). Hardly anything is publically known about how, and how well, this money is spent, much less how much pays for advanced and possibly speculative technologies.

Congress

Legislators have great latent power within the PMIC, but seldom use it decisively. Members of Congress—525 of them—hold a far wider range of views than typically found in the Pentagon or among managers of defense firms. Views also tend to be less stable; they sometimes shift quickly with election results and international events, while few members can claim deep knowledge of military affairs and the number of veterans has drifted downward over the years. The House of Representatives and Senate are disorderly venues in which national security is but one issue among many. Open proceedings invite posturing, but what legislators say in public may bear little resemblance to what they believe or the positions they take when negotiating among themselves or with the executive branch. While senators and representatives often weigh in on big issues, for example by warning of coming threats from China, action is another matter, with lawmakers often preferring delay to going on the record with their votes. Unlike DoD and business firms, there is no well-defined structure of authority and no top-down decision making. When it comes to the PMIC as a whole, then, this is the least predictable major actor. About all that Congress can be counted on to do is to tinker with White House budget requests, hold a multitude of hearings for reasons good and bad, and otherwise dabble in ways that generate complaints of micromanagement from the executive branch (Carter, 2019, pp. 134–142; Gansler, 1989, pp. 101–121; also see Mayer, 1993).

As part of the post–World War II reorganization of defense, Congress combined formerly separate committees that had overseen the War Department (including the Army Air Forces, which had grown enormously) and the Navy Department. In the new structure a single Armed Service committee in the House and another in the Senate authorized military programs, with budget matters the province of subcommittees of the appropriations committees in each chamber. Congress also created the Joint Committee on Atomic Energy, which had considerable prominence at a time when the atom bomb was a mystery shrouded in secrecy, so that members even on the defense committees often looked to the JCAE for guidance. Senator Brien McMahon did much to shape the legislation establishing the AEC and became the first JCAE chair. With others, including longtime member Senator Henry “Scoop” Jackson, McMahon pushed for more powerful atomic bombs in greater numbers and an aggressive hydrogen bomb program (Condit, 1988, pp. 467–473; Jackson, 1953). Otherwise, Congress rarely looked deeply into the defense budget before the 1960s and the war in Vietnam, with exceptions for construction projects such as bases viewed as opportunities to steer dollars and jobs to constituents (Gordon, 1961).

Georgia’s Carl Vinson chaired the House Armed Services Committee for all but two years between 1949 and 1965, and his counterpart in the Senate, Richard Russell, also from Georgia, headed that body’s Armed Services Committee from 1955 to 1969, for a time chairing the defense appropriations subcommittee as well. They gained and held these positions because of seniority rules that also gave near-dictatorial power to committee chairs over matters within the panel’s jurisdiction. When frustrated junior members reinforced in 1974 by a newly elected cohort following President Nixon’s disgrace and resignation finally forced through rules changes, and with personal and committee staffs also expanding, members of the authorizing and appropriating bodies grew more assertive, going beyond top-line approvals to scrutinize a lengthening list of subsidiary items (Lindsay, 1987); eventually Congress was marking up the defense budget on almost a line-by-line basis (Art, 1985; Towell, 2012). Two rough generalizations seem supportable once committee chairs lost control over legislation reaching the floor. First, on major issues such as modernizing the nation’s nuclear weapons, members often seemed to vote their convictions (see, e.g., Mayer, 1991). Otherwise, as on so many other matters, deals get made by compromise, logrolling, and favor-trading (Rundquist et al., 1996).

In other ways, too, Congress inserts itself into defense policy. Over the years, Senator William Proxmire and his staff convened many hearings to critique programs they considered unneeded, wasteful, or mismanaged. Later, Senator John McCain, most of whose views otherwise differed from Proxmire’s, also came to be known for spotlighting pork-barrel spending. Informal groupings have mattered too. For a time in the 1980s, the congressional Military Reform Caucus numbered well over 100 members. They held widely divergent positions on national security, precluding any sort of substantive agenda. Even so, the reform caucus focused attention on DoD management and choice of weapons in ways that set off alarms at high levels in the Pentagon (Chiarelli & Gagnon, 1985).³ In its diversity the reform caucus was atypical. The F-35 caucus illustrates the more usual pattern: As Congress prepared to consider the 2020 budget, members lined up more than 100 flag-level retired military officers to protest plans backed by other lawmakers to buy upgraded versions of the older F-15, fearing this might somehow jeopardize the F-35 (Trimble, 2019).

Perspectives on the PMIC

Images and Analyses

From the close of World War II into the 1960s, popular accounts of military programs might as often laud engineers designing X-planes and test pilots flying them as deplore goldplated weapon systems. For many Americans, communism was an unholy menace and the Soviet Union, in Ronald Reagan’s later words, an evil empire. Then the fighting in Vietnam brought reportage of napalm and B-52s carpet-bombing jungles—what the U.S. military knew to do—even as soldiers on the ground struggled with faulty small arms such as M16 assault rifles that jammed and sometimes blew up in their faces (McNaugher, 1984). Popular views began to shift, slowly at first before gathering such momentum that Lyndon Johnson left the White House without seeking reelection.

By the 1970s, a growing slice of the public agreed with critics who portrayed ranking military officers as collaborators with industry executives and legislators in spending billions of dollars on ever more weapons capable of ever greater overkill. In such depictions, well-paid lobbyists, including retired flag officers and one-time congressmembers, aid and abet the efforts of defense firms to sell their wares. (In 2019, lobbying expenditures by defense and aerospace firms came to some $111 million, according to OpenSecrets.org, 2020.) If something of an oversimplification—the military, not industry, has always been the dominant force in the PMIC—an academic literature grew up around it. Bernstein and Wilson (2011) note much of this, little of which needs further mention. More popular accounts from this period include Fitzgerald (1972), Fallows (1981), and Stubbing with Mendel (1986).

The earlier literature gave little attention to how the PMIC fit into the larger structure of business-government relations. More recent work, though rarely defense-specific, reveals intricate webs of cross-industry ties among companies, whether small businesses or multinationals, that support the PMIC directly or indirectly (Soffer, 2001; more generally, Phillips-Fein, 2009). These ties extend far beyond groups such as the Aerospace Industries Association and Business Executives for National Security to include local officials and civic boosters, state-level officials, and corporate leaders across many different industries. Collectively such “economic elites and organized groups representing business interests” (Gilens & Page, 2014, p. 565) exert disproportionate influence over public policy generally. In issuing or reinforcing warnings of threats to national security they bolster the PMIC even if they have no connection with the defense industry. Shermer (2008), for instance, showed how such interests came together in the 1950s Sunbelt to attract inward investments by defense contractors, which went on to alter the political complexion of cities and states, in part as a result of the voting preferences of their employees. Still, so long as the Cold War continued, broad business support for the PMIC hardly needed to be constructed: It came near to self-assembly. The voices were far from uniform, ranging as they still do from the Red-baiting that has been a constant theme in conservative circles for more than a century to rhetorical support for “stronger” defense, but they have been loud and persistent.

Narrowing the view to defense, studies based on “issue networks” (Heclo, 1978) or “garbage can” models (Cohen et al., 2008) should help reveal how particular national security policies and programs gain or lose momentum. Such models envision fluid structures that actors enter and exit opportunistically, depending on the course of events and their particular interests. Applications to the PMIC do not seem to have been appeared as yet, but if attempted could yield further insight into the ways in which costly new weapons programs or policy shifts such as that resulting in the Space Force come together, with more of an analytical thrust than typical journalistic and historical accounts.

Militarism

The militarization of U.S. foreign policy, with influence and power in the councils of government accruing to the Pentagon at the expense of the Department of State (DOS), began well before the Cold War. Into the 1930s, DOS overshadowed the War and Navy Departments even on matters of direct concern to the armed forces (see, e.g., documents on the 1930 London Naval Conference in FRUS, 1945, p. 1ff.). That changed as the decade continued and the United States entered World War II. More than most presidents, Franklin D. Roosevelt sidestepped cabinet officers and their departments. He distrusted the judgment of his secretary of state and “lacked confidence in the State Department bureaucracy. Thus the State Department was excluded from most of the significant wartime diplomacy.” (Yergin, 1977, p. 57; also see May, 1955). Overseas, field commanders relied on organic sources of intelligence, and their concerns often extended beyond purely military considerations, to include, for instance, the capacities of the German and Japanese economies as wartime damage accumulated. For these and other reasons, “the State Department and its overseas personnel were rather generally pushed aside” (Briggs, 1965, p. 133). After the war, it was military officers who governed occupied territories. The State Department regained prominence with the success of the Marshall Plan. Yet George Marshall himself had been army chief of staff before becoming secretary of state, and he would later return to the Pentagon as secretary of defense.

By then Senator Joseph McCarthy was flinging accusations of communist influence at the State Department, inflicting lasting damage by driving out experienced personnel and discouraging others from seeking a career in the foreign service. The department itself stumbled for internal reasons on issues such as arms control, failing to develop staff capabilities or call on outside experts to counter the claims of the armed forces during interagency deliberations on, for example, the likelihood of detecting clandestine Soviet nuclear tests. With the Pentagon opposed to almost any sort of restriction on nuclear weapons, and the AEC, congenitally wary of displeasing the military, mostly going along, DOS, the natural home for more nuanced viewpoints, did little to provide them. The State Department compounded its failings in Washington by sending poorly prepared delegations to international negotiations; U.S. representatives sometimes lacked insight even into the substance and rationale of policies they were ostensibly backing (Kistiakowsky, 1976, p. 268; more generally, see Marcella, 2008). Once Robert McNamara took over at DoD, he quite overtly pushed the Pentagon deeper into foreign policy (DoD, 2013). In the decade following, Richard Nixon and Henry Kissinger bypassed or ignored the State Department, NSC, and the usual apparatus of diplomacy as they pleased.

If anything, militarism gained momentum after the Cold War. The George W. Bush administration invaded Iraq despite lack of support in other capitals and withdrew from the Anti-Ballistic Missile Treaty. Donald Trump went much further, rejecting diplomacy and international cooperation outright. Any reversal will be a long-term and politically fraught undertaking. The first part of this article has surveyed the PMIC. The second and shorter part will examine major acquisition programs and explore the government’s long-running and largely ineffectual efforts to limit cost and schedule overruns and otherwise manage MDAPs to desired outcomes. Much of the difficulty stems from military control over the “requirements” that specify what each system is to do and the acquiescence of those in industry, Congress, and the executive branch to the wishes of the senior officers who write and approve these contractual provisions.

Major Acquisition Programs

MDAPs constitute the core of the PMIC. They go on for years and cost billions, as illustrated by two aircraft programs, the F-35 fighter and V-22 Osprey. Lockheed Martin is building the F-35 in three versions to requirements set down by the Air Force, Navy, and Marine Corps. (Information on the F-35 following comes from DoD, 2019g, and Gertler, 2020; cost figures include the engines, supplied by Pratt & Whitney under a subprogram with its own budget lines.) RDT&E began with design studies in the early 1990s, flight tests got underway in 2006, low rate production followed, and in 2015 the Marine Corps accepted their version as “operationally capable,” the first of the three to be so designated. As of 2020 about 550 F-35s had been delivered. Current plans call for production to continue until the middle of the century, perhaps beyond, depending on foreign sales.⁴ RDT&E expenditures have been put at more than $70 billion and procurement at some $400 billion (both figures in inflation-adjusted 2012 dollars), with lifetime operating and maintenance costs, which DoD has begun to call sustainment, projected to reach $1.2 trillion (then-year dollars). Cost overruns, on what is so far the most expensive acquisition program ever, are expected to approach $150 billion (GAO, 2020, p. 203). Like other recent MDAPs, computer software accounts for much of the cost of both RDT&E and anticipated sustainment expenditures (Box E).

Box E. “Software Is Never Done”

F-35s fly with the aid of computer code that will total some 24 million lines when finished. Modifications will then continue more-or-less indefinitely. Hence the saying “software is never done,” the title of a report from the Defense Innovation Board (DoD, 2019c). Variants include “software is never finished … just released,” notwithstanding bugs unknown and perhaps known but not yet fixed. For software, sustainment accordingly consists not of fixing breakage and fighting corrosion and wear but of error-correction and upgrading, familiar as downloads to everyday smartphones—sometimes introducing new bugs.

For the F-35 and many other MDAPs software has become a substantial source of budget and schedule overruns. As long ago as 2003, software may have accounted for something like two-fifths of all DoD RDT&E (GAO, 2004, p. 1). While more recent estimates are evidently lacking, the figure can only have risen. And RDT&E is only part of the story, since dollars from nonacquisition accounts pay for sustainment, sometimes including major revisions.

As commercial tablet and smartphone sales took off and apps seemingly multiplied day by day, firms selling into these markets had to find new ways of developing software. Migration of storage and processing to the cloud, dissociated from user hardware at the “edge,” added to pressures for rapid development. Traditional methods originating in the mainframe era, with design, development, and testing proceeding step by step until the entire package was deemed ready for release, could not keep pace. Called waterfall processes because in theory the flow is unidirectional from start to finish, they were replaced by continuous and iterative code development, termed agile processes, in which engineers work, not with the entire package, but with discrete modules small enough to be readily manageable. The objective: to complete each module quickly, in weeks or months rather than years, test and modify based on user feedback, and piece these blocks together in ever-larger assemblages, these too tested along the way and modified iteratively. The process viewed from a distance might accordingly be visualized more like a series of eddying whirlpools than a steady downstream flow.

Despite early successful experiences with alternatives on defense and space systems, some as far back the 1960s, defense and intelligence agencies mostly continue with waterfall methods (Larman & Basili, 2003; Robson et al., 2020, pp. 87–117, compare waterfall and agile methods in DoD context). Only recently has DoD begun to adopt the newer practices, in part because waterfall methods align with normal acquisition procedures (GAO, 2020, pp. 33–40 and 66–69). As for hardware, requirements and specifications come first, spelled out in detail for contractors to meet (DoD, 2018a). Yet experience with the World Wide Military Command and Control System (WWMCCS) long ago demonstrated that acquisition practices evolved to buy Army tanks and Navy ships do not suit large-scale information systems. WMMCCS began as an effort to combine and consolidate the independent communications systems of the Army, Navy, and Air Force. Each had its own global network, together incorporating something over 150 computer systems developed for particular applications, whether messaging or inventory management or operational planning. With each user group intent on having things its own way, “from the very beginning WWMCCS [was] dominated by subunit concerns, emphasizing services’ needs and requirements and not infrequently working to the detriment of the larger national interest” (Pearson, 2000, p. 343). This was a recipe for disaster. Trying to accommodate a nearly endless list of specialized requirements saddled WMMCCS with painfully slow and costly development of both hardware and software. When finally delivered, much of this was obsolescent or obsolete. Software-intensive programs like the F-35 grapple with similar issues today.

Agile methods might seem analogous to acquisition approaches tried previously under rubrics such as concurrency and spiral development, the basic idea being to begin production, for instance of a new aircraft, before completion of engineering development and testing, with subsequent modifications incorporated as they come along. This was common practice during World War II when output was the priority and the costs of post-production changes, which often included scrapping tooling and tearing apart just-finished equipment for rebuilds, could be justified. In peacetime the cost penalties proved unacceptable. Brown (1992) includes numerous examples from earlier decades, and the F-35 seems another instance of the perils of hardware concurrency. For software, of course, there is no tooling, and installation does not entail reworking a physical system.

DoD has recently begun to talk of an all-inclusive “Information Enterprise,” a term that covers “communications; spectrum management; network policy and standards; information systems; cybersecurity; positioning, navigation and timing policy; and the DoD information enterprise that supports DoD command and control” (DoD, 2019d, p. 7, parentheticals omitted). To take full advantage of all this, as well as software embedded in weapon systems, defense agencies have little choice but to move toward agile software practices; failure to do so promises to leave fielded systems, like WMMCCS, lagging perpetually behind the technological state of the software arts.

The impetus for the F-35 stemmed from cost overruns leading to drastic production cuts on an earlier program, the F-22 air superiority fighter. Conceived in the early 1980s to stay ahead of anticipated Soviet advances and therefore pushing hard against multiple technological frontiers, the Air Force had wanted 762 F-22s. (Note the appearance of analytical precision, implying that 761 would be one too few and 763 an unneeded extra.) It was forced to settle for fewer than 200. At this point a new platform, less ambitious technically and therefore expected to be less expensive, became the priority. Once again spending rose, leading to reductions in planned production, although in this case only by about 500 F-35s, to some 2,500 (for all three service-specific versions, but excluding foreign sales). In this case the Air Force established a requirement for its version of 1,763, a figure set in 1991 and unchanged since.

With the F-35 seemingly secure, the Air Force began seeking a new stealth bomber, the B-21, for which it has already received more than $10 billion in RDT&E funding, with another $2.8 billion requested for 2021 (DoD, 2020d, pp. 1–19). At the same time, the Air Force, Navy, and Army have all been pursuing, for largely self-defined missions, means of executing “prompt global strike,” the ability to deliver warheads from within U.S. borders on targets that might be anywhere on Earth in as little as an hour (Sayler, 2020; Woolf, 2020). In the past this could only be accomplished with ICBMs. Since these follow predetermined paths, at least in principle they can be tracked from takeoff, their trajectories predicted, and warheads intercepted and destroyed. Given maneuverable platforms traveling at five or more times the speed of sound, targets would remain unknown or at least highly uncertain until just before impact, making interception far more difficult. Extraordinary technical difficulties long blocked development, with many billions of dollars spent since the 1950s and feasibility still not fully assured (Hallion, 1998). The Pentagon contends that Russia and China are well along in development, with early versions perhaps fielded already, leaving the United States no choice but to proceed—and with defenses against such weapons too (DoD, 2019a; Stone, 2020). From threat to threat, the PMIC rolls along.

While the F-35 originated in the Air Force push for a next-generation multirole fighter, some major programs begin tentatively. What became the V-22 Osprey, a tiltrotor that takes off and lands vertically like a helicopter while swinging its rotors forward in flight for speeds greater than helicopters can reach, got its formal start in 1983 with a $69 million contract award to a joint Bell-Boeing proposal, the only submission despite years of earlier exploratory work, much of it by NASA (Gertler, 2012; more generally, see Whittle, 2010, the source of much of the information following). Development posed daunting technical obstacles, including instability during transitions between vertical and horizontal flight. V-22s experienced multiple fatal crashes and Defense Secretary Richard Cheney tried to halt the program. He could not. The Marine Corps resisted fiercely and the service enjoys widespread congressional favor going back to its amphibious assaults on Japanese-held islands during World War II. Fearful of absorption into the Army during the 1930s, the Corps had worked out methods for this mission, in which the Army had no interest and would not have willingly cooperated with the Navy, just as the Navy did not want Army troops on its vessels (Millet, 1996). Having found a set of tasks more compelling than guarding embassies and protecting warships, when guided missiles forced troop transports farther offshore the Corps at first substituted helicopters for landing craft and then latched onto tiltrotors as promising still greater standoff distances without extended flight times (Sedivy, 1992). In 2007 ten V-22s went to Iraq, by 2020 deliveries had reached 375, and the 2021 budget seeks $1.8 billion for another batch (DoD, 2020d, p. 15). Well before, the Marines, with other services, had begun exploring next-generation vertical lift platforms.

Unlike the F-35 and V-22, with their checkered histories, some few MDAPs run to completion smoothly enough to reap widespread acclaim. The Navy’s Polaris submarine and missile is perhaps the best-known example (Sapolsky, 1972). Among aircraft programs, the F-16 receives generally high marks (Aronstein & Piccirillo, 1997), as have designs from Lockheed’s Skunk Works, a very different model (Rich & Janos, 1994). At the other extreme, perhaps as many as one-fifth of major programs run into so much trouble—some combination of cost overruns, schedule delays, and performance shortfalls—that they are canceled before completion (McNicol, 2018).

Efforts to cancel a program most often originate in the executive branch. The common response in Congress has been to curtail production quantities to limit cost overruns, as for the F-22. High-level officials in OSD at times succeed—Cheney failed with the V-22 but managed to halt the Navy’s A-12 at about the same time. On rare occasions the services themselves change course, as the Army did in 2004 when, after much internal debate, it ended the Comanche helicopter program after two decades in hopes of spending the funds elsewhere (GAO, 2009, pp. 17–18).

In a few cases Pentagon civilians have pushed initiatives over military resistance. McNamara did so with the F-111, to much criticism; in that case, the Air Force and Navy argued that OSD’s push for commonality would seriously compromise their respective missions (Richey, 2005). Ronald Reagan’s Strategic Defense Initiative (SDI) posed a different sort of threat: To the leaders of the armed forces it was a chimera poised to draw billions from programs they had been pursuing for years. They also feared that even the faint prospect of some sort of shield against enemy missiles would devalue the cherished deterrent triad of winged aircraft, land-based ICBMs, and missile-firing submarines. Facing “substantial opposition from both the Air Force and Navy” (Werrell, 2000, p. 22), the administration established an independent organization within DoD, which lives on as the Missile Defense Agency. In yet other cases, Congress votes to keep buying equipment the military no longer wants, such as the C-130 transport, designed in the early 1950s and still in production today (GAO, 1998).

Major program outcomes depend on a host of technical decisions. These affect everything from the degree of commonality among the three F-35 versions to the risks of fielding systems that after years in development turn out to be poorly suited to whatever threats the U.S. military then finds itself facing. All this might suggest that DoD’s choices reflect some sort of rigorously structured analysis. A veneer of reasoned judgment, and sometimes the genuine article, does feature in published documents on national security flowing from the Pentagon, White House, other parts of government, and official or self-appointed advisory bodies and think tanks. Yet as many examples in this article show, humans and their governments do not necessarily display much capability for sensible choices and effective decision making. For these and other reasons, major weapons programs sometimes end up meriting the epitaph “designed by committee,” a not unfair characterization of the F-35, the more so as the three versions ended with much less in common than initially advertised.

There are three proximate causes of acquisition decisions that at least in hindsight seem questionable. First, great uncertainty surrounds geopolitical settings, as demonstrated by the unexpected end to the Cold War, and evolving technologies too, innovation being unpredictable by definition. Second, the armed forces largely control decisions on what to buy and the senior officers who have the final say are experts in warfighting, not technology, while the practical utility of weapon systems hinges on tradeoffs and compromises involving both technical and military matters. Third and related, few in Washington outside the Pentagon have sufficient understanding of technology and sufficient political clout to oppose whatever the service chiefs, Secretary of Defense, NSC, and intelligence agencies say is needed to ensure national security, the more so as national security itself is a political construct, more than a little malleable.

Regulating Acquisition

A forbidding array of laws, rules, and regulations governs DoD contracting. In 2016 Congress created the so-called Section 809 panel, shorthand referring to the portion of that year’s defense authorization act establishing the Advisory Panel on Streamlining and Codifying Acquisition Regulations. Another in a lengthy series of reviews intended to map out directions for reform, the panel’s findings, released in three volumes in 2018–2019, run to more than 2,300 pages—as good an indication as any of the labyrinth of extant rules and regulations (DTIC, n.d.).

The Section 809 study followed others going back decades; Fox et al. (2011) provide a concise survey, while Schwartz and Peters (2018) summarize more recent legislation. Few reforms have come close to their stated objectives. The daunting thicket of multiple approval stages and reviews remains because it was put in place piece by piece over many years to build in reliable oversight mechanisms and insulate against corruption and is still needed. “Waste, fraud, and abuse” became the charge du jour in the 1980s, with congressmembers pointing to high-priced military items such as the infamous $2,000 coffeemaker; one senator displayed a “spare parts Christmas tree,” each decoration with its price tag (Chiarelli & Gagnon, 1985, pp. 12–13). Long before, in the 1930s, Senator Gerald Nye had led a well-publicized series of hearings in search of collusion, scandal, and profiteering on military contracts, without uncovering much. Still, bribery, bid-rigging, and self-dealing do surface from time to time, just as in purely private transactions and state and local government contracting. Thus in 2004 Air Force procurement official Darleen Druyun pled guilty to criminal charges for steering sweetheart deals to Boeing, while Randy “Duke” Cunningham was later sentenced to eight-plus years in prison for taking $2.4 million in bribes as a member of the House Intelligence and Appropriations committees.

No way around the basic dilemma has been found. Repeated efforts to reform acquisition incrementally—the thrust of most recommendations, including those of the Section 809 panel—get at marginal, not systemic, issues. The 809 panel adopted “Think Bold” as something of a motto, and urged creation of a “Dynamic Marketplace Framework,” but said nothing concrete about how to accomplish any of this. Indeed, past attempts to break sharply from existing practices, such as Total Package Procurement—“a totally new procurement concept specifically designed to control costs” (Knaack, 1998, p. iii)—have sometimes made matters worse. First employed for the C-5A, a mammoth airlifter built by Lockheed during McNamara’s tenure at OSD, the idea was simple enough: “development, production, and support requirements for a system would be bought under a single overarching contract” with “price and performance commitments … finalized during the contract’s definition phase” (Knaack, 1998, p. 31). After negotiating the contract, Lockheed could not come close to fulfilling its terms. These called for an airframe with low empty weight, high load-carrying capacity, and the ability to take off and land on short and rough airfields. The combination proved technically impossible to meet. This had been foreseeable and as it became obvious the officers overseeing the program, along with high-level civilians in Washington, repeatedly dismissed or concealed the accumulating deficiencies. Lockheed, its commercial business also stumbling, nearly went bankrupt, and the Air Force ended up paying far more than planned for a system that could not do what it was supposed to. As poor experiences on Navy programs also showed, total package procurements, intended to end low-ball cost estimates and overpromising by bidders, accomplished nothing of the sort (Alic, 2014).

Requirements: The Nexus

General Bernard Schriever once said of Curtis LeMay, Air Force chief during the first half of the 1960s, “He was of the school that ‘Goddamn it, I know what I need. This is a requirement, goddamn it, now go out and do it’” (Sheehan, 2009, p. 167). Fixated on nuclear war, as head of the Strategic Air Command LeMay first turned SAC into “an air force within the Air Force” (Greenwood, 1979, p. 236) and then came near to remaking the entire service in SAC’s image, as the Air Force purchased thousands of supersonic fighter-bombers designed for all-out war with the Soviet Union—aircraft that proved ill-suited to the war that did come, in Vietnam (Box F).

Box F. Requirements: LeMay and the Bomber Generals

Within the Air Force, the dominating presence of General Curtis LeMay and his allies, many of whom had served under his command, shaped the Air Force for a generation and more. From the 1930s on, Army aviators dismissed tactical missions such as ground attack near the fighting front. They believed strategic bombing, intended to destroy the enemy’s homeland-based warmaking potential, to be the primary instrument of air power and key to victory, and the Army Air Forces went on to fight World War II on this basis. (A great deal has been written on strategic bombing; see, e.g., Alic, 2007, pp. 18–31 and 89–102, and citations therein.) LeMay, a hard-driving combat leader during that war, took over SAC in 1948. He became Air Force Vice Chief of Staff in 1957 and four years later rose to the topmost position.

In the 1950s, LeMay’s vision of nuclear war shaped the designs of supersonic fighter-bomber-interceptors of the Century Series, F-100 through F-106. “Strategic fighters” were intended to accompany SAC’s heavy bombers to target areas and once inside enemy territory conduct opportunistic nuclear strikes on air defenses, airfields, and any aircraft they could catch on the ground (Boyd, 1988). The Air Force purchased nearly 6,000 in six different types from five different firms (Knaack, 1978; there was no F-103).

Fighter-bombers configured to penetrate Soviet airspace at high speeds and altitudes performed poorly in Vietnam. Mission profiles were much different than those for which aircraft such as the mainstay F-105, nearly the size and weight of a World War II “heavy” bomber, had been conceived. Pilots flying them in Southeast Asia struggled to locate targets on the ground while dodging antiaircraft defenses and hostile interceptors; many were lost (Werrell, 1998; one cannot do better than Broughton, 1985, for what it was like to fly F-105s over Vietnam). As the lessons became plain, OSD compelled a reluctant Air Force to buy F-4s, a Navy aircraft that, since it had to take off and land from carrier decks, was more maneuverable than F-105s and others of the Century series. Eventually, fighter pilots who had risen through the attenuated tactical aviation ranks displaced bomber generals in the topmost echelons of the Air Force command structure (Worden, 1988).

In a further instance of his forceful role within the Pentagon, LeMay sought initially to ensure that ICBMs would not be allowed to eclipse SAC’s big bombers. He insisted, for example, that “the prime characteristics of the ICBM can be exploited best by using it initially to damage and disrupt Soviet air offense and defense systems on the ground, holding them down until they can be destroyed by the manned bomber” (LeMay, 1957, p. 10; emphasis in original). After a time, LeMay and other bomber generals were forced to switch gears; they then argued for building vastly more ICBMs than the numbers arrived at by the rational calculators in McNamara’s OSD; after all, the Minuteman was an Air Force program (Ball, 1980).

LeMay’s imprint on the Air Force contrasts with that of Navy Admiral Hyman Rickover, an equally imperious but otherwise very different politico-bureaucratic operator. For decades, Rickover and his hand-picked subordinates kept tight control over nuclear submarine design (Duncan, 1989). Yet while LeMay had broad and deep support within the Air Force, many senior naval officers resented Rickover and opposed his efforts, and more than a few submariners believed his technology choices too conservative, claiming that Soviet design concepts promised superior performance. Rickover did not need their support; he got his way through assiduous cultivation of Congress and the parallel position he held as a high-level AEC manager even while remaining on active duty in the Navy.

Now as then “the warfighter writes the requirement” (Army Strong, 2011, p. 32). This sets governing parameters for system design, yet few of the high-level officers who generate and endorse requirements can claim much beyond a rudimentary grasp of science, engineering, and production. Rather, they “tend to envision technology as a servant that must meet their needs as they define them” (O’Neil & Porter, 2011, p. 126). The Army’s Future Combat System (FCS) illustrates. This was to be an “ensemble of manned and unmanned ground and air platforms” (DoD, 2004b, p. 49) with electronic networks linking everything from individual soldiers to robotic combat vehicles. From 2003 until it vanished from budget documents at the end of that decade, the FCS consumed some $20 billion (Alic, 2013). The verdict of Frank Kendall (2017), former Under Secretary of Defense for Acquisition, Technology, and Logistics: “driven by a ‘vision’ that was divorced from reality,” the program “delivered basically nothing to the Army” (p. 6).

Engineering design and development—the central activity in any large-scale technical effort, military or commercial—is heavily path-dependent, in the sense that early decisions constrain those that follow (National Research Council, 1991). Setting tightly binding requirements at the beginning restricts flexibility downstream and can in fact make it impossible to meet, or even come near to meeting, those requirements. For the FCS they drove the entire program down a dead-end path. Well-conceived requirements, conversely, are a necessary if not sufficient condition for satisfactory outcomes. As explained by aircraft designer George Spangenberg, who worked on over 100 programs during a lengthy career as a Navy civilian, the F-8 fighter “went from the day we let the contract to fleet introduction in three years … One of the best [ever]. It just shows what can be done … in this case we had a good requirement; it was not overstated. We weren’t stretching for the stars” (Spangenberg, 1990, p. 183).

To someone like Spangenberg, sensible requirements are little more than Engineering 101, and for successful technology-intensive businesses they become part of organizational DNA. Unlike their counterparts in commercial firms, however, managers in the defense industry cannot operate on this basis. They must heed another, often contrary, dictum: “you do what your customer wants, you don’t try to tell your customer what he wants” (Whittle, 2010, p. 115, quoting from an interview with a Bell Helicopter vice president).⁵ Contractors who bridle, decline to exaggerate what they can do for military customers—less likely to be “smart buyers” than, say, airlines negotiating with Boeing or Airbus over what next-generation jetliners should look like—risk losing out to more pliable rivals.

Planning and Management

Each service choses its own weapons, or tries to. DoD never has enough money for everything, and advocates within services and service branches discuss and debate what should be next. Armored personnel carriers or scout helicopters? Electronic warfare suites or air-to-ground missiles? A new generation of submarines or surface combatants? By law, the chief of the acquiring service must sign off on major programs. Unlike in the world of business there are few avenues of appeal once top-level decisions have been reached. At that point, service commitments become all but irrevocable and the agenda shifts to persuading Congress, the media, and the public of the need for the new program, while fending off possible threats from Pentagon civilians and the financial watchdogs at the White House Office of Management and Budget.

Whittle (2010) explores these dynamics for the V-22. In an older but especially revealing case, Bergerson (1980) shows how a shifting set of officers, mostly in the middle ranks, managed to get armed and armored helicopters accepted by a reluctant Army high command. This was a task of some delicacy: “Policy entrepreneurs” had to build a coalition through persuasion without stirring up too much opposition too early. In other settings, as illustrated by the Air Force’s LeMay and the Navy’s Rickover, powerful bureaucratic actors assemble and steer the necessary coalitions from the top down. This was also how the Future Combat System originated: General Eric Shinseki, the Army’s chief of staff, bypassed the usual evaluation and approval stages to push the program directly into the acquisition pipeline (Porter et al., 2009).

Shinseki could have sought and heeded technical advice on the FCS from experts inside the Pentagon or on the outside. DoD employs around 100,000 engineers and scientists, about half of them in R&D laboratories and many more in acquisition. Leaving aside the occasional maverick, only rarely do they serve as effective counterparts to senior officers. This is because, quite simply, all except for a small number on the civilian side of the Pentagon report directly or indirectly to military officers who are themselves almost universally reluctant to deliver bad news upward.

The R&D laboratories themselves are “managed and operated within the military service chain of command” (GAO, 2018, p. 3). Of the more than 80,000 DoD employees with technical skills of one sort or another working in acquisition—not all in jobs classed as engineering or science (Table 2)—civilians outnumber military by about 9:1; however, some four-fifths of military personnel are commissioned officers and most exercise some level of managerial authority (DoD, 2020f). Like their counterparts in private industry DoD’s civilian employees mostly do as their superiors expect or demand.

Table 2. Acquisition Workforce and Budget by Service, 2020

	Acquisition Personnel^a	Technical Personnel^b	Acquisition Budget (billions of dollars)^c
Army	43,100	20,400	$38.8
Navy (including Marine Corps)	68,300	39,500	$82.8
Air Force	41,100	18,500	$97.2
Other	28,600	3,820	$33.7
All DoD	181,000	82,200	$252.5 billion

a Note: Second quarter, fiscal year 2020. Totals may not add because of rounding. Does not include nongovernment contract personnel, for which no figures are available. Other includes Defense Contract Management Agency, Defense Logistics Agency, Defense Contract Audit Agency, Missile Defense Agency, and several smaller entities;

b Approximated as total of acquisition personnel in engineering, facilities engineering, information technology, science and technology management, and test and evaluation positions;

c Total obligational authority, fiscal year 2020.

Source: For Acquisition Personnel, Technical Personnel, see DoD (2020f), and for Acquisition Budget, see DoD (2020h), p. 2.

Program offices staffed by several hundred to over 1,000 people (plus contractors, uncounted but often in considerable numbers) oversee each MDAP. Mid- to high-ranking officers from the responsible service head nearly four-fifths of these offices (Hunter & Tate, 2019, p. 60). As the day-to-day interface with Washington, they should be the first to learn when contractors struggle to satisfy requirements, overspend, or fall behind schedule. Those in charge may know but not tell, since it is DoD practice to rotate officers in and out of assignments every few years and program managers may prefer to sit on bad news, leaving their successors to worry over what to do. Beyond this, the Navy’s inquiry into the troubled A-12, later canceled by Cheney, concluded that “the fundamental problem … is to create appropriate incentives to enable senior leaders to rely upon responsible, accountable line managers for realistic perspectives on the cost, schedule and technical status of their programs” (Beach, 1990, p. 35). Yet even if incentives were realigned, this would not accomplish much so long as the root cause of acquisition maladies remains the requirements laid down by the services. Personnel below the topmost levels can do little or nothing about these.

Table 2 also points to another organizational contrast among the services. The Army has fewer than half as many dollars to spend as the Air Force, yet employs about the same number of acquisition personnel. Throwing people at process does not seem to have led to better outcomes; available comparisons, while limited, point to cost growth greater than for Air Force and Navy programs (Arena et al., 2006, pp. 11–16). Indeed, the Army has never seemed that comfortable with technology, notwithstanding its Corps of Engineers and the Ordnance Department’s 19th-century contributions to manufacturing. As noted earlier, the Army’s aviation arm depended on private firms from the beginning, and the ground Army stumbled repeatedly in buying trucks and tanks, even though drawing on technical knowledge that advanced slowly compared with aviation (Beaver, 1993; Johnson, 1998, pp. 78–80). The Navy, on the other hand, had no choice but to develop competence in mechanical systems once steam power entered the fleet, since ships must be self-reliant at sea (although it too turned to private industry for aircraft).

Shinseki could have looked beyond DoD civilians for technical advice on the FCS. Even so, he might not have received unfiltered views. Although the Army, like the other services, has a long-established board of advisers on science and technology, something over one-quarter of the current members are retired officers and the others include many consultants and academics who have received DoD contracts and grants (Army Science Board, n.d.). At a further distance, the National Academies of Science, Engineering, and Medicine has a standing Board on Army Research and Development, and the academies might alternatively have organized an ad hoc group of experts. Still, the academies undertake nearly all their studies under contractual “terms of reference” that the funding agency can write to constrain the inquiry as it wishes. In fact, DoD and the services have quite effectively insulated themselves since the first half of the 1960s from sources of advice they cannot control.

Conclusion

Over the years, the Pentagon, defense industry, and Congress have changed internally and in their relationships with one another. In an overall sense, most of these changes have had marginal to minor impacts on the PMIC as a whole. The few that have been substantial followed from uncontrollable geopolitical events, notably wars (World War II and Korea) and threats deemed existential (the Soviet Union after 1945). Otherwise, U.S. military institutions exhibit considerable stability. Given the popular and political appeal of “national security” and a “strong military,” whatever these may be taken to mean, the armed forces can usually find ways to buy whatever weapons they want with the funds Congress makes available. Given the workings of American politics, who is to say no?

A bit like in natural ecosystems, feedback mechanisms built into the PMIC tend toward gradual rather than rapid change. In the biosphere, big disturbances—asteroid impact, ice age, anthropogenic greenhouse gases—disrupt whole systems and set them on a new path. The PMIC has not been so disrupted since the Korean War. Vietnam did not bring transformation, even if the Air Force began to train its pilots more realistically and bought aircraft better suited to tactical warfare; the Army, for its part, vowed not to repeat the experience, and other lessons that might have been learned were not. In 1986, following the invasion of Grenada, which had taken place with a good deal of confusion among participating services and service branches, Congress passed the Goldwater–Nichols Act. The intent was to push the armed forces toward more effectively integrated “joint” operations and a more nearly corporate military, as opposed to a loose collection of quasi-independent units. Requirements and the acquisition process were to shift accordingly, leading to weaponry conceived and designed for joint operations. As the 1991 Gulf War showed, Goldwater–Nichols helped with the planning and conduct of warfighting. But by the evidence reviewed in this article on cost, schedule, and performance outcomes, it has had little effect on acquisition. As in the past, the “Services dominate the current requirements process” and, as a consequence, “‘Jointness’ is forced into the program late in the process during an adversarial and time-consuming program review” (DoD, 2004a, p. iii).

Built for big wars, the U.S. military has never shown much appetite for the smaller wars the nation has asked it to fight. Since the collapse of the Soviet Union, warnings of new threats have come from DoD, intelligence agencies, various commissions, members of Congress, the White House, and multiple outside groups. Perceived threats have included an expansionist China, portrayed as a near-peer semi-superpower—“a powerful and prosperous China that is equipped with a ‘world-class’ military” (DoD, 2019a, p. i), a resurgent Russian Federation, along with North Korea and Iran. Uncertainty itself has been depicted as a threat, as if something other than an ever-present variable: “Force posture and employment must be adaptable to account for the uncertainty that exists in the changing global strategic environment. Much of our force employment models and posture date to the immediate post–Cold War era, when our military advantage was unchallenged and the primary threats were rogue regimes” (DoD, 2018c, p. 7). Should big wars return, the military means to be prepared: “DoD’s FY 2021 research and development budget is the largest ever requested and laser focused on the development of crucial emerging technologies to expand our warfighting advantages over strategic competitors” (DoD, 2020a, pp. 1–9).

The obvious danger is that of the “big war” challenge becoming a self-fulfilling prophecy. The Cold War created powerful incentives for cooperation among the United States and its allies. These have diminished. Many in Congress, self-identified conservatives especially, show little interest in alliance politics or rebuilding U.S. soft power and diplomacy. So far, the contradictions between U.S. actions in pulling back globally and its still accreting military power have barely been addressed. (What would be the reaction in the United States should China subdivide the globe absent only demilitarized Antarctica, plus exoatmospheric space, into “combatant commands”?) The contradictions probably will not be addressed, much less resolved, because both the public and much of the policy community believe in giving the armed forces what they say they need to fight big wars. If this remains the ongoing pattern, the PMIC will probably continue to change in small ways only.

Department of Defense	$716 billion
National Nuclear Security Administration^a	$28 billion
National Intelligence Program^b	$62 billion
International Affairs^c	$51 billion
Other^d	$10 billion
Veterans Benefits and Services^e	$239 billion
Total	$1.106 trillion

Browse All

Share Link

Article contents

The U.S. Politico–Military–Industrial Complex

The U.S. Politico–Military–Industrial Complex

Summary

Keywords

Subjects

Bureaucracies and Politics

Spending

Eisenhower and the Military–Industrial Complex

Box A. Military Technologies: Episodic Evolution

Box B. Eisenhower and Military Affairs

A Sketch of the PMIC

Box C. Toward the PMIC: Slow Start, Fast Finish

Acquisition

The Defense Industry

Box D. Rapid Innovation

Congress

Perspectives on the PMIC

Images and Analyses

Militarism

Major Acquisition Programs

Box E. “Software Is Never Done”

Regulating Acquisition

Requirements: The Nexus

Box F. Requirements: LeMay and the Bomber Generals

Planning and Management

Conclusion

Further Reading

References

Notes

Related Articles

Browse All

Share Link

Article contents

The U.S. Politico–Military–Industrial Complex

The U.S. Politico–Military–Industrial Complex

Summary

Keywords

Subjects

Bureaucracies and Politics

Spending

Table 1. Fiscal Year 2021 National Security-Related Budget

Eisenhower and the Military–Industrial Complex

Box A. Military Technologies: Episodic Evolution

Box B. Eisenhower and Military Affairs

A Sketch of the PMIC

Box C. Toward the PMIC: Slow Start, Fast Finish

Acquisition

The Defense Industry

Box D. Rapid Innovation

Congress

Perspectives on the PMIC

Images and Analyses

Militarism

Major Acquisition Programs

Box E. “Software Is Never Done”

Regulating Acquisition

Requirements: The Nexus

Box F. Requirements: LeMay and the Bomber Generals

Planning and Management

Table 2. Acquisition Workforce and Budget by Service, 2020

Conclusion

Further Reading

References

Notes

Related Articles