Accounting for State Capacity - American Affairs Journal

The debates over the Department of Government Efficiency have revealed, if nothing else, that the federal budget is obscure even to the political combatants ostensibly responsible for developing and overseeing it. In the executive branch, Elon Musk highlights that billions of dollars of payments are processed by the Treasury without even a memo line. Meanwhile, in Congress, Republican politicians highlight the incompleteness of the bureaucracy’s spending records, while Democrats bemoan the Trump administration’s dissimulation in ceasing to share budgetary guidance documents.¹ The camp followers of these obscure programs are thousands of federal contractors, pursuing vague goals with indefinite timelines. As soon as the ink on a bill is dry, it seems, Congress loses sight of its initiatives until their eventual success or their all-too-frequent failure.

Contrast this with the 1930s, when the Roosevelt administration provided Congress with hundreds of pages of spending reports every ten days, outlining how tax dollars were being put to use in minute detail. The speed and thoroughness with which these reports were produced is hard to fathom, and yet the administration was actually holding its best information back. FDR’s Treasury had itemized information on hundreds of thousands of projects, down to the individual checks that were written. Incredibly, politicians had better dashboards in the era of punch cards than we have in the era of AI. The decline in government competence runs deeper than our inability to match the speed and economy of New Deal construction: even their accounting was better. What happened?

Political scientists discuss the decline in government competence in terms of “state capacity,” which describes a government’s ability to achieve the goals it pursues. Most political scientists agree that the United States not only suffers from degraded state capacity in absolute terms, but has less state capacity today than in the early twentieth century. A popular theory for this decline blames the excessive proceduralism of the U.S. government: the “cascade of rigidity” or the “procedure fetish.”²

But reformers need more than complaints. To rebuild state capacity, reformers need an affirmative vision of what good procedure should look like and, in order to enact it, knowledge of how government procedure is changed. The history of government budgeting and accounting reform illustrates both. There were three major eras of reform to federal accounting in the twentieth century: New Deal reforms of the 1930s, conservative reforms of the 1940s and 1950s, and liberal reforms of the 1960s. This history tells the story of how accounting reforms first built up American state capacity and how later reforms contributed to its gradual decline. These reforms thus offer lessons on rebuilding state capacity today.

First, reform is political. Every accounting reform was pushed by a coalition pursuing goals unrelated to better accounting. Second, successful reforms help the actual decision makers. New Deal reforms built state capacity by giving politicians useful information, while succeeding reformers failed by pursuing reforms that appealed only to technocrats. Finally, simplicity is essential. After pursuing many second-order goals, the United States eventually lost the ability to provide timely information on how money was spent. It will take serious work to recover that basic level of competence.

The 1930s: Accounting Rebuilds America

In the 1930s, the United States faced a once-in-a-lifetime cataclysm. After the stock market collapse in 1929, unemployment rose to 25 percent. Cities and states were functionally bankrupt. FDR’s landslide victory against Herbert Hoover gave his administration a mandate for aggressive reform and experimentation. The administration proposed unprecedented spending for economic recovery, and from 1935 to 1937 was granted the then stunning amount of $8.5 billion for infrastructure investment alone—roughly $200 billion in today’s dollars.³

This New Deal spending built roads and bridges, while the need to account for the spending spree built state capacity. To keep track of these staggering sums, the Treasury Department created an accounting system that provided nearly real-time information on more than two hundred thousand individual projects, down to recording every last check within ten days of it being written. This information was shared with both the White House and Congress, who invested in the bureaucracy due to the trust this information generated. Achieving this unprecedentedly good oversight required teamwork in the form of an alliance between Treasury Department bureaucrats and New Dealer academics. The Treasury bureaucrats were interested in better accounting for its own sake—it was their job, and they wanted to do it well. For FDR’s academic advisers, however, these reforms were a means to an end: increasing executive power.

FDR’s academic advisers were reformers centered around budget pioneer Frederick Cleveland, working mainly in FDR’s home state of New York. Cleveland built this network through running the New York Bureau of Municipal Research, a good governance institute, and through directing research on budgeting at the Institute for Government Research, a think tank now called the Brookings Institution. These reformers had a villain: legislatures, and their penchant for micromanaging the way the bureaucracy hired workers and spent money. This legislative nitpicking reached amusing lengths, with one notorious provision requiring the federal government to hire somebody who was between forty-five and fifty years old and had at least ten years’ experience on the Supreme Court of the Philippines.⁴ Of course, this described precisely one person, who the government now had to hire regardless of his competence. Some of Cleveland’s associates, however, sought a diagnosis as well as a culprit. Cleveland’s researcher W. F. Willoughby proposed that the issue was distrust: Congress didn’t know how the bureaucracy ultimately spent money, so it attempted to micromanage the decision-making process. Accordingly, he proposed that with better accounting, the executive branch could give better information to Congress, thus encouraging it to relax these procedural constraints.⁵

The reformers pursued a vigorous executive branch that could resist legislative micromanagement, starting in their home state. Cleveland was the main author of a report for New York’s 1915 constitutional convention, which argued for sweeping expansion of gubernatorial power.⁶ The state eventually revised its constitution in line with Cleveland’s views, leading to the expansive powers that Governor Roosevelt later enjoyed, such as a line-item veto. When FDR became president, he tapped this network, commissioning them to write a series of reports on improving the federal government. Many reports were on technical issues, such as accounting—not coincidentally, written by Cleveland’s protégé A. E. Buck. The main report, called the Brownlow Report, argued for a more powerful president with expanded staff assistance. These reformers had strengthened the New York governor, and FDR hoped they could do the same for the presidency.

The reformers were joined by a different group: Treasury bureaucrats who were disenchanted with the existing system. Up until that point, Congress had never received adequate reporting on how money was spent, although it received a confusing array of useless reports. The Treasury, for instance, reported on how money was spent, but until 1927, it had only reported how money was intended to be spent, as opposed to how it was actually spent.⁷ Even after 1927, its reports were inadequate. The General Accounting Office, the government’s auditor, was supposed to release annual reports, but these were typically months or years late. In fact, it didn’t submit even a single report from 1933 to 1936, exactly when (given that this was the height of the Great Depression) such reports would have been invaluable.⁸ The mess of reports reflected the tangled way the government made payments.

Making payments required two things: first, government agencies had to pay their bills; and second, the Treasury had to have money in the bank when the agencies paid these bills. Reconciling these two requirements was surprisingly difficult.

First, each government agency tasked its own employees with making payments. The Department of Agriculture had employees who paid farmers, the Navy Department’s employees paid sailors, and so forth. These employees kept records on the payments they made, which were sent to the Treasury and constituted one set of the Department’s books.

Second, the Treasury also had to have money in banks throughout the United States. For example, there were many government agencies in Chicago, which required the Treasury to place money in a Chicago bank in order to cash all these agencies’ checks. This was especially inconvenient outside of cities, where the government maintained relationships with hundreds of small rural banks to make payments. Banks sent the records they kept to the Treasury, which then constituted another set of books.

The Treasury thus had two poorly integrated sets of books: one kept track of officials who wrote checks, while the other kept track of banks that handled money. Reconciling the two was a nightmare. The Department could eventually figure out how money had been spent, but by the time they detected malfeasance the money would be long gone.

Both academics and bureaucrats were, therefore, interested in improving government accounting, and the New Deal infrastructure programs gave them their chance. The project was set up by three administration officials: Secretary of the Treasury Henry Morgenthau; Daniel Bell, a career Treasury official serving as the director of the Bureau of the Budget; and E. F. Bartelt, a career Treasury official who ran the accounting division. They aimed to acquire the best technology of the day but felt that they first had to reform the confused procedure for making payments.

President Roosevelt promulgated executive orders to streamline the payment process, whereby the Treasury had to track how hundreds of assorted government employees made payments from hundreds of banks. These orders instructed the Treasury to place money in only a single bank per state, and to only allow the Treasury’s own employees to make payments (thus taking this responsibility away from individual agencies).⁹ After the reforms, the Treasury only kept books on one set of employees who transacted through roughly fifty banks. Having improved the process for all government spending, the administration invested in technology for the infrastructure program’s accounting.

Each Treasury-affiliated bank kept track of its state’s infrastructure projects and their expenses. Every payment and contract would be recorded on a punch card, which encoded information such as the cost, the project number (perhaps for a bridge), and the type of payment (perhaps an unpaid invoice). Every ten days, the state banks closed their books and airmailed punch card records to Washington, where the Treasury Department had built a state-of-the-art IT system to process these cards. These machines could sort and read hundreds of cards every minute—at the time, a quite impressive performance—and add up the information that the cards contained. The Treasury would calculate, for instance, a per-state breakdown of each program’s spending.¹⁰

The system recorded an astounding level of detail. The full reports that were automatically generated every ten days were almost 2,500 pages long, breaking down the data on approximately 220,000 projects according to thirty-two different categorizations of spending.¹¹ This level of detail was reserved for the White House, which could examine individual projects in order to reward economy and punish profligacy and tardiness. Congress got somewhat less information—but it was still the best information it had ever received.¹²

Every ten days, congressional leaders were presented with an approx­imately two-hundred-page report, whose comprehensive information included tabular breakdowns of spending and time-series graphs.¹³ They could examine everything from how an agency had spent money over the past few years to its performance over the past ten days. If any agency had failed to make sufficient progress or had misappropriated funds, Congress was likely to find out almost immediately. Congress rewarded that trust by giving the administration carte blanche to choose its own infrastructure projects without congressional micromanagement.

The accounting system was such a success that the administration even made a public interest film, The President Accounts, that ascribed the good management of the New Deal to these reforms.¹⁴ The accounting project had provided the information that the president needed for management and that Congress needed for oversight. The bureaucracy’s ability to move quickly was thanks in large part to the trust this information garnered, and the administration publicly said so. Better accounting had indeed strengthened the executive branch.

Before its ultimate demise, the experiment was put to its greatest use for oversight of World War II military spending.¹⁵ Although the wartime information was less detailed (owing to the military’s incompatible accounting equipment), the Treasury was providing information even faster than before: E. F. Bartelt could, according to his obituary, tell a congressman the status of a military appropriation within half an hour.¹⁶ Better accounting not only rebuilt America, but also helped win a world war.

The system didn’t last, however. The automated accounting system was only ever a pilot project reliant upon emergency infrastructure funds (and later, upon emergency wartime funds). As this funding ran out after World War II, the pilot experiment automatically expired, and Congress appears to have forgotten that it had once received such detailed information. And not only did Congress forget, but academia did as well—there does not appear to be any scholarly discussion of this experience.

The Treasury’s experiment—by almost any standard, an outstanding success—offers three lessons for rebuilders of state capacity. First, the need to build a coalition: although academics had proposed accounting improvements to fix the government’s management woes, these ideas would have gone nowhere if not for cooperation with the bureaucracy. Second, the need to focus on what decision-makers actually want: the accounting system gave the White House the information it needed to manage projects efficiently, and it gave Congress the information it wanted to perform effective oversight. Finally, the virtue of keeping things simple: the system only provided information on contracts and spending, and although it did nothing else, it achieved its goals thoroughly and expediently.

After the Treasury’s experiment had faded from memory, government accounting couldn’t measure up to the heroic achievements of the 1930s and 1940s, but it was by no means as bad as the pre-reformed system. The streamlined procedure for payments and accounting, enabled by FDR’s executive orders, still remained. But as conservatives reclaimed their influence in the postwar era, even these improvements fell under attack.

Revenge of the Gray Flannel Suits

To 1950s businessmen, the dream of rolling back the size of government seemed like a lost cause. The federal government undertook a range of responsibilities that would have been unthinkable before the Great Depression and which had become unchallengeable after war’s end. Democrats had won nearly uninterrupted control of Congress since 1932, and the Republicans only won the Presidency by nominating the war hero Dwight D. Eisenhower, who professed his acceptance of the New Deal.

Conservatives had lost the hearts of the voters but hoped they might yet win the hearts of government accountants. Their suggested reform was program budgeting, which aimed to measure the total cost of government programs more comprehensively than before. This ap­proach won the endorsement of the landmark Hoover Commissions and was ultimately enacted in law. These reforms, however, inadvertently decreased state capacity. To measure costs in a broader way, agencies had to expand the role for budget offices at the expense of technical experts. This cost data was not useful or even desirable to the politicians in Congress who oversaw agencies. It did appeal to its backers, however: an alliance of accountants and businessmen.

During the war, accounting had become important as never before, as prominent accountants such as Percival Brundage noted at the time.¹⁷ Accountants did brisk business helping to navigate the heavy income tax, the wage and price controls, and, above all, the wartime production contracts, which entailed intricate recordkeeping. The contracts for airplane production, for instance, typically allowed a certain profit plus reimbursement for the cost of producing the airplane. Accordingly, businesses had to keep detailed records on the cost of their products, both to ward off government investigators and to guard against claims of profiteering.

Accountants, in turn, began to examine the government, with Brundage making postwar speeches urging accountants to take a greater interest in government.¹⁸ He took his own medicine: Eisenhower appointed Brundage as his budget director, where he now turned a critical eye to the government’s books. He and his ilk balked at the government’s accounting practices, which recorded only the things that the government bought (that is, its contracts and payments). The accountants argued that, instead, the government should concentrate less on things it buys and more on things it does. In particular, they pushed to measure these activities’ non-monetary costs. Military aid illustrates their concern: the government might donate planes or munitions from its warehouses without spending any money. Measuring the true cost of military aid, in their view, required accounting for these materials’ costs.

Joining the accountants were business executives, who wanted the government to use accounting for managerial purposes, as the private sector did when deciding to open factories or enter new markets. Managerial accounting was old hat by the 1950s, but executives, too, had taken a recent interest in the government—particularly its debt. Businessmen hoped that measuring expenses would encourage better control over what they saw as the “gigantic public debt and high costs of current government activities.”¹⁹ Using their influence in local organizations such as New York City’s Citizens’ Budget Commission²⁰ and the Taxpayers’ Federation of Illinois,²¹ they pushed local government managers to use accounting as a tool for cutting spending.

Both accountants and executives wanted a new approach to accounting for the cost of government programs. The accountants wanted better financial data, whereas the executives wanted better management. That is, accountants thought that improved accounting would be more accurate, while executives thought it would be more useful. Both soon got their chance to realize their vision during the era of postwar reform.

Following the end of the war, the U.S. government was a sprawling mess of duplicative agencies. In 1947, Congress appointed a bipartisan commission run by former President Herbert Hoover to clean things up. The first Hoover Commission, as it came to be known, reformed nearly all aspects of the government, from the military to records management. Its success inspired Republicans to later convene a second, much less successful Hoover Commission in 1953.

Both Hoover Commissions treated accounting as a priority, and both won significant legislative victories that infused government with business ideas.²² Previously, budgeting had focused on purchasing things and showed an agency’s salaries, travel expenses, and so forth. The new approach focused on activities and their costs. For example, the Coast Guard reworked its budget to show the cost of “Airlift rescues” and “Sealift rescues,” these being two of its major operations. Moreover, the Commissions proposed cost-based budgets that would show the cost of all materials used in programs: “Airlift rescues” would include, for instance, the cost of spare planes that the Air Force donated.²³ The Commissions’ accounting recommendations were rapidly enshrined in law.

This application of program budgeting was officially encouraged by the Budget and Accounting Act of 1950. Further, amendments to the National Security Act in 1949 required that the Department of Defense adopt both program budgeting and cost accounting. The Department was required to create powerful financial officers, called comptrollers, who were given expansive responsibilities such as collecting statistics and studying management issues.²⁴ Other agencies similarly increased the power of their budget offices.

Congress went still further, and through the Act to Improve Government Budgeting of 1956, encouraged the executive branch to submit cost-based budgets: these were budgets that would estimate the total cost of materials that a program would use, rather than the number of dollars to be spent. The executive branch complied: in 1957, four agencies submitted cost-based budgets, and by 1959, almost one hundred did.²⁵ Agencies similarly began applying cost accounting to their property. This move required agencies’ budget offices to begin recording a wide variety of new data: organizations such as the Army had always kept track of the number of munitions, for instance, but now they kept track of the value of munitions, too. By the mid-1950s, at least eleven agencies had applied cost accounting to their $12 billion in property.²⁶

Agencies didn’t just record property values—they started charging other agencies to use their property. For instance, printing offices’ accountants would calculate the cost of paper and, based on this, would charge other agencies for the printing services they requested. The printing office would use a device called a working capital fund, essentially a bank account that stored the revenue from printing services, which could then be used to buy more paper. This device had always existed, but it was taken to extremes in the pursuit of discovering the true cost of programs.

Central offices were given monopolies, with other agencies being required to purchase services from them. For instance, the General Services Administration was created in 1949 to be the government’s property manager; all other agencies were required to rent property from it. Similarly, the Department of Defense created working capital funds for its inventories of uniforms, munitions, and so forth. In order to obtain equipment, the military’s teams were required to pay into these various funds, such as the stock-fund and the industrial fund, as if they were buying from a private company. In theory, this pseudo-market discipline would force agencies to economize on their use of services and supplies.

The government now measured costs more comprehensively in every respect, just as the accountants and businessmen wanted. But although it gave federal accounting some abstract virtues, it came with three significant downsides that each decreased state capacity.

First, the new process of cost measurement added red tape, empowering budgeteers at the expense of scientists and engineers. Budget offices were tasked with managerial studies, collecting statistics, and other responsibilities previously left to technical experts. More than that, financial officers were placed in charge of technical programs. Research was increasingly seen as a financial task for which cost accountants were the proper overseers. In the Department of Defense, for example, program budgeting led to the growth of financial agencies at the expense of technical agencies.²⁷ The Department of Agriculture placed all scientific research under the finance office, when previously the career scientists had called the shots.²⁸ Accountants acquired an impressive and unprecedented level of influence over various sectors of government activity, ranging from ballistic missile development to cattle research.

Second, while these cost measurements appealed to accountants, they were rarely useful for decision-makers. In the executive branch, ac­counting simply could not lead to spending cuts in the way proponents claimed—the issue was fundamentally political. Even the Department of Agriculture’s chief finance officer attacked the exaggerated claims about what his office could accomplish.²⁹

If program budgeting was of dubious value for the executive branch, it was clearly unhelpful to the legislature. Congress had been skeptical of program budgeting from the get-go: when the Department of Defense adopted it in 1953, members bitterly complained about the lack of clear oversight information.³⁰ These concerns were never allayed. Fundamentally, Congress appropriates money rather than total costs and was never able to think in any other terms.³¹ Cost-based budgets were counterproductive for the legislators who actually passed the budgets.

Third, it was not only unhelpful, but cost accounting actively under­cut Congress’s ability to perform traditional oversight. Cost accounting precluded immediately sharing spending information, due to cost calculations requiring accountants to laboriously review the books. More than that, congressional control was undercut by the financial machinery of cost accounting, namely by cost-based budgets and work­ing capital funds.

Cost accounting’s largest problem is that it is fundamentally arbi­trary: there is no unique cost of a single good. For instance, equipment might be valued at its replacement cost or at its original cost, but whichever cost is chosen, there are many ways to allocate it to the different goods the equipment produced. There is no principled reason to prefer any one approach over another. This issue is seen today with military aid to Ukraine. Although Congress attempted to control the total amount of military aid, the Pentagon repeatedly evaded this cap by arbitrarily revaluing the munitions it had donated. It first decided that its accounting tricks allowed it to donate $3 billion dollars more, which then became $6 billion, and ultimately $8 billion more than Congress had thought it had authorized. Congress was displeased, with one mem­ber calling it an “attempt at deception.”³² The most powerful case for cost-based budgets is to increase control over military aid, but even here, the control is illusory.

Working capital funds were still more destructive of congressional control. Traditionally, Congress gives agencies money and then ques­tions them about how they spend it. Working capital funds, by contrast, are outside that system: the agencies that sell supplies or services are funded by other agencies, not Congress. Moreover, these agencies have monopolies on their products and are subject neither to actual market discipline nor to rigorous congressional oversight. The negative consequences are seen in the obscurity of the Defense Department’s working capital funds.³³ They were also seen in the GSA corruption scandals of the 1970s, which revealed graft and sleaze on an industrial scale.³⁴ The theory was that if agencies sold services, it would encourage the economical use of these services. But the agencies that sell services were now outside of congressional control, with nobody to impose thrift on them.

This experience offers several warnings. First, coalitions of reformers can just as easily make things worse as better. Well-intentioned accountants and businessmen pushed for reforms that ultimately swelled the ranks of the bureaucracy they opposed. Second, reforms will fail if the actual decision-makers do not consider the reforms useful. Few politicians used the cost accounting data, but the cost accounting system rapidly took on a life of its own, with budget offices annexing roles previously reserved for subject-matter experts. Finally, complex solutions create lasting problems. The reforms have created a set of issues that still plague the U.S. government today, such as unaccountable working capital funds and arcane cost accounting standards. All in all, the 1950s show the problems of glibly assuming that what works in business must work in government.

With the dawn of the 1960s, the United States returned to liberalism—but did not return to the practical liberalism of the Roosevelt era. Instead, the ascendant technocratic liberalism embraced the theory of the program budget and decided that it should go even further. As long as we’re budgeting according to total costs, the liberals thought, the government should measure benefits as well.

The 1960s: The Slide Rule Reformers

Liberals of the 1960s were ready to break with the stodgy conservatism of the Eisenhower era. From closing the missile gap, to improving national education, to dealing with southern racism, the U.S. political system they took control of appeared unable to confront issues of dire importance. In one significant area, however, liberals doubled down on conservative reforms: the accounting and budgeting procedures with which they intended to fund a more vigorous government.

The government was confronting a daunting set of crises, and in their view, it was time to get down to brass tacks. The bureaucracy had to escape the squabbling of politics and figure out what to tackle head-on. It had to set its long-run goals and ruthlessly prioritize its programs to achieve them, an approach they called Planning, Programming, Budgeting (PPB). Robert McNamara, the secretary of defense in the Kennedy and Johnson administrations, pioneered this approach in the Department of Defense, eventually persuading Johnson to mandate it governmentwide. Unfortunately, far from achieving reformers’ ambi­tious goals, the approach dealt a lasting blow to government competence: PPB effectively centralized power with the planners and froze out technical experts, eventually leading the government to farm out policy to contractors and NGOs. The ultimate results were the last thing in the world its supporters—businessmen and the military’s social scientists—would have wanted.

The Air Force’s research arm, the RAND Corporation, began thinking about the benefits of big government long before the 1960s student radicals. They had good reason to. Consider, for instance, a proposal to make a missile 10 percent more accurate but five times more expensive. This might seem wasteful—but in a nuclear war, taking out extra Soviet missiles before they retaliate would make it the bargain of the century. The cost was meaningless without knowing the benefits being purchased: in this case, a better chance to survive an atomic war. To consider such issues, in 1950, RAND’s cost guru Charles Hitch ordered that the organization establish a cost analysis department.³⁵ Its maximally comprehensive approach to cost accounting included future costs, such as the projected cost of refurbishing equipment. These comprehensive costs of government activity could then be compared to their benefits.

From here, RAND scholars began to apply cost-benefit analysis—or systems analysis, as they called it—across the entire government. One of their researchers, Roland McKean, studied water resource policy, eventually publishing a study in 1958. His stated goal was to make the government as efficient as private businesses. Businessmen’s options, he said, can “be translated into effects on cost or, better yet, on profits. Government may be able to do something along the same lines.”³⁶ The “same lines” were to measure the benefits of government programs, which he analogized to profit. At the end of his study, he suggested that the program budget could be used as a guide for bringing cost-benefit analysis to the rest of the government, and sketched out what this might look like. In 1959, he was consulting for the government’s Office of Education.³⁷ After that, it was off to the races for cost-benefit analysis.

Meanwhile, businessmen were adding long-run planning to the social science toolbox. The management fad at the time was strategic planning, which held that a company should systematically identify its top goals and formulate a rigid, multiyear plan for achieving them. Starting in the early 1960s, businesses embraced the planning process, which was the work of a specialized planning staff in the executive’s office. Just as business homed in on its top priorities, so too could the government select a narrow set of goals.³⁸

These influences came together in McNamara, who translated his early work in operational research into a meteoric rise in business. His early career saw him use statistical analysis in the Air Force during World War II, after which he was hired at Ford to inject quantitative skill into the flailing company. He was named the comptroller and introduced methods for quantitative analysis and long-run planning. Shortly after becoming CEO, he was tapped to head the Pentagon by President Kennedy. He brought several RAND scholars with him, including Hitch as the Department of Defense’s comptroller. It was this team that developed the aforementioned Planning, Programming, Budgeting system.

The three-step approach of PPB combined the cost-benefit analysis of social science with the long-range planning of business. With planning, the government agency identifies its top long-run goals. Social scientists then analyze the agency’s programming options using cost-benefit analysis and select the top choice. Budgeting is the step where the selected choice is given a dollars-and-cents budget. At every point, generalist policy analysts were in the driver’s seat.

In 1961, McNamara imposed PPB at the Pentagon, which (with modifications) is still the Department’s budgeting procedure today. For a brief and heady time, it was the entire government’s budgeting procedure. After being convinced of the approach’s merit, in 1964, President Johnson announced that all government agencies would adopt PPB. Not a single agency had been told in advance.

The immediate result was embarrassing. Agencies were caught off guard and admitted that they were unfamiliar with the system or how to adopt it. Some agencies requested guidance from the Bureau of the Budget, which had imposed the PPB mandate and the deadlines for its adoption. They heard in response that the Bureau itself disagreed with the mandate and thought it was unworkable, adding that it was therefore unable to give them guidance. Most agencies made desultory efforts to comply, with every agency except the Department of Defense eventually abandoning the system.³⁹

The ultimate result of this short-lived experiment was a heavy blow to state capacity. Although PPB failed, the attempt to impose it was influential. Attempting to measure the benefits of government programs empowered those who measured the benefits, namely the government’s social scientists. There were three main consequences of the experience.

First, it led to the divestiture of in-house expertise to the advantage of contractors. This was seen most dramatically in the Department of Defense itself. Until the 1960s, the Pentagon had maintained significant in-house expertise through the operation of its own shipyards and arsenals. The Department used the technical experts from these facilities to negotiate as equals with contractors, as they could distinguish accurate technical claims from misleading ones. McNamara privatized many of these facilities and urged the military to hire contractors instead, arguing that clear definitions of military need would allow policy analysts to solicit bids from the entire private sector instead of the few military facilities.⁴⁰ The military accordingly invested in policy analysts at the expense of engineers. But in the end, losing these technical experts prevented the government from negotiating as equals, giving contractors the upper hand in negotiations.

Second, PPB led to the farming out of policymaking to NGOs. The civilian adoption of PPB was due, in fact, to the War on Poverty. After Johnson established the Office of Economic Opportunity and tasked it with the antipoverty effort, its director brought on McNamara’s assistant, Adam Yarmolinski. In true PPB fashion, the agency’s leadership first identified their major goals and then solicited the most cost-effective options to pursue them. The contractors they hired were frequently former RAND and aerospace employees who were doubtfully qualified to, say, reform urban governance. More than that, the office encouraged formal cooperation with community activists, in accordance with their goal of building morale and cohesion (as they attempted in Vietnam) and due to activists being more cost-effective than government employees. Career social workers and poverty experts were left out of the conversation.⁴¹ The trend of policy outsourcing only expanded over the next few years. The first major concerns about the growth of contractors may have been voiced by none other than the succeeding Richard Nixon administration, which pushed back against some of the Great Society’s programs.⁴² The PPB experiment sputtered out in the 1970s, but NGO-captured community involvement remains.

Third, PPB failed at its purpose, which was to be a better system for making and explaining decisions. Few government agencies ever understood it, and it was so obscure to members of Congress that the Committee on Government Operations formed a PPB Inquiry that attempted to learn what the system was and why the bureaucracy had adopted it. Although the civilian bureaucracy abandoned PPB, the Department of Defense still uses this approach, adapted from postwar corporations that collapsed as soon as Japanese companies gave them their first real competition. Today, commissions sponsored by government and by outside reformers study the possibility of reforming PPB to reinject the dynamism that the military and its engineers once had.⁴³ In the meantime, the U.S. military has given up its in-house expertise in return for the worst of midcentury corporate practices—complete with a five-year plan!

The history of McNamara’s reforms offers several cautionary tales. First, the results of reform can be totally contrary to what reformers intend. The coalition pushing PPB hoped it would usher in a new era of government competence, but the opposite occurred. Second, once interest groups are looped into policymaking, they can be impossible to get rid of. The policy analysts were only all-powerful for a few years, but the cottage industry of consultants, contractors, and NGOs they enabled has had a chokehold on government policy for decades. Third, bad goals can have lasting costs. Outside of the Department of Defense, PPB was a flash in a pan. Nonetheless, the PPB-era outsourcing of government competence harms the state more than most reforms that still exist today. The PPB experiment, on the whole, shows the problems of assuming that what works in theory must work in practice.

Government Accounting: Back to the Future?

To rebuild state capacity, reformers need to know what good government processes look like, and they need a vision of how to convince the government to adopt them. The history of federal accounting, from FDR to the Hoover Commissions to Secretary McNamara, offers three broad lessons.

First, all reform is coalitional. In all three eras, groups surveyed the dry subject of federal accounting and judged that they could profit from reform. In the New Deal, bureaucrats who wanted better payments and academics who wanted a stronger executive collaborated to build an unprecedentedly good system of oversight. In the 1950s, executives who wanted stronger management and accountants who wanted more comprehensive records led to the rise of cost accounting. The brief but far-reaching reforms of the 1960s were set in motion by an improbable coalition of businessmen and social scientists who together hoped that long-range planning and cost-benefit analysis could be used to tackle the Cold War’s challenges.

Second, successful reforms help the actual decision-makers. Accounting reforms are not neutral—they affect how institutions make decisions through empowering some groups at the expense of others. The 1930s reforms gave the White House the information it needed to ensure that bridges and roads were built quickly and under budget. More than that, it gave Congress the information it needed to feel secure in giving the bureaucracy flexibility during the Depression and World War II, thereby empowering engineers and other experts. By contrast, the 1950s and 1960s reforms both created a layer of process that empowered proceduralists. Budget officers and policy analysts did not, in their new roles, give Congress or government executives any information they wanted to have. Accordingly, these procedures merely added red tape. The newly empowered proceduralists subsequently undercut subject-matter experts such as scientists and engineers, thereby decreasing government competence. To avoid pushing lousy ideas, reformers should embrace politics. Government is run by politicians, not technocrats, and reforms can only succeed through helping politicians make better decisions.

Third, simplicity is a virtue, as ancillary goals can eventually overshadow a system’s main purpose. In the New Deal, the Treasury’s system only recorded spending, but did so with astounding speed and comprehensiveness. By contrast, the expansive goals of 1950s cost accounting were not inherently bad, but they came with excessively costly secondary goals. Carefully measuring cost was, in practice, incompatible with quickly reporting spending. Still worse, PPB in the 1960s not only added a new layer of process in pursuit of secondary goals, but these processes stunted U.S. state capacity through sidelining technical experts. Complex goals are often mistakes that can be difficult to correct, even decades later.

If reformers such as the leaders in DOGE are to succeed, they need to imitate FDR and shun the postwar technocrats. Their success in crafting a smaller, more competent government will depend on learning all three lessons. First, build a coalition. DOGE can’t go at it alone—while excoriating the bureaucracy is entertaining, lasting change requires collaborating with reformist bureaucrats and legislators. Second, help the actual decision makers. It isn’t enough to parachute into agencies and slash at things. If the government is to banish paper-pushers and hire engineers, DOGE must wrest power away from proceduralists and give politicians effective control over the bureaucracy. Third, embrace simplicity. Flashy software engineering is all well and good, but truly ambitious goals, such as improving government software—which is certainly necessary—depend upon first streamlining and simplifying dry government procedure.

For motivation, perhaps DOGE should plan a movie night at the National Archives, where they could watch the Roosevelt administration’s film The President Accounts. As its final lines say, “good accounting means good government.”

This article originally appeared in American Affairs Volume IX, Number 2 (Summer 2025): 137–55.

---

Notes