The Civil Service as a Problem-Solving Institution

1 Examples of Problem-Solving

Nevertheless, the civil service can engage in problem-solving. Indeed, it does it all the time. Anonymous civil servants who have done extraordinary things that ensure that our society keeps functioning (Lewis 2018, 2024). In many cases, they solve problems that most of us are unaware even exist.

I will point to two examples of problem-solving I’ve observed firsthand. The first is about performance management systems in the US federal government. Such systems are formal mechanisms intended to identify levels of performance, and motivate greater effort and innovation to improve that performance. Unfortunately, the history of such systems is not very promising. To extend the Scott metaphor, governments tend to impose of high modernist administrative designs when they engage in government reform – they embrace projects that seek to be comprehensive and rational, and which often pay little attention to the context in which civil servants work. Much of the most profound scholarship of public administration, such as that offered by Herbert Simon, Charles Lindblom, and Aaron Wildavsky, all point to the practical cognitive limitations of civil servants to manage information, and to manage conflict. Rational models of reform that do not account for these limitations, collapse under their own unwieldy expectations. Wave after wave of such reforms, such as Program Planning Budgeting Systems, or Zero-Based Budgeting, did not seem to transform government in the way promised, even as they consume a great amount of administrative resources. Such systems were adopted with great fanfare, and then dustbinned by a new Presidency.

In 1993, the federal government passed the Government Performance and Results Act (GPRA). The Act mandated that agencies start to undertake period strategic plans, and measure and report data on specific performance indicators. In 2010, the Act was updated, with the GPRA Modernization Act. The Modernization Act retained key elements of the original Act, and added to them. It incorporated changes from the Bush administration – institutionalizing the position of Performance Improvement Officers, the Performance Improvement Council, and Chief Operating Officers. It required that agency officials meet quarterly to focus on their key goals, as well as cross-agency goals.

With a series of co-authors, I have tracked the relative success of these different performance initiatives, both by talking to officials who worked in the federal government, and relying on surveys collected by the Government Accountability Office (GAO) about performance management practices.^[1] Observing that it is very difficult if not impossible to assess the effects of a broad governmentwide mandate on actual performance, we looked at performance information use as our key dependent variable. Our independent variable was if the employee was engaged in organizational routines related to performance. We found that the first iteration of performance management with GPRA largely failed to change patterns of self-reported performance information use (Moynihan and Lavertu 2012). However, the Modernization Act was associated with greater performance information use by civil servants. The difference reflects the difference in routines: the routines created by GPRA pushed employees to collect and disseminate data. This was not sufficient to encourage the use of that data. The Modernization Act compelled employees to talk with one another about performance metrics, using routines such as quarterly reviews, was associated with purposeful use of performance data (Moynihan and Kroll 2016).

The second example centers on the emergence of the administrative burden framework in the federal government. This framework defines administrative burdens as people’s experience of policy implementation as onerous and directs attention to different types of costs that make up those burdens: learning costs (understanding what a public service is and how to engage with it), compliance costs (time and effort devoted to interacting with the public service) and psychological costs (the emotional responses that people experience) (Herd and Moynihan 2018). During the Biden administration, the federal government adopted key aspects of this framework into policymaking in a number of ways, through and executive order on racial equity, on customer experience, in Circular A-11 (the budget guidance that all agencies must follow), through revised guidance for the Paperwork Reduction Act issued by the Office of Information and Regulatory Affairs (Herd et al 2024). The Government Service Delivery Improvement Act passed at the end of 2024. The Act requires each agency to designate a senior official to oversee service delivery initiatives, while requiring the OMB to identify a Federal Government Service Delivery Lead. The Act includes language requiring that the Service Delivery Lead:

1. understands how individuals, businesses or organizations interact with an agency;

2. review delivery processes, while considering values such as “ease, efficiency, transparency accessibility, fairness, burden, and duration, including wait and processing times”;

3. collecting both qualitative and quantitative data on service delivery;

4. evaluate the quality of service delivery;

5. coordinate with OMB and other stakeholders on service delivery to improve service delivery practices.

The adoption of this framework represents a form of problem-solving, in that it identified administrative burdens as a problem, one that was previously overlooked, and made the case for reducing those burdens. While it remains to be seen how durable the framework is, and the depth of its impact, it offers a counterpart to another example of problem-solving: cost benefit analysis. Cost benefit analysis started as a tool from economics before becoming adopted by the federal government, and maintained by successive administrations to identify and reduce excessive burdens on businesses. The adoption of such frames compels officials to seek information that they may not have previously attended to, analyze that information differently, and take action as a result. In the case of administrative burden, this implies better understanding how citizens experience public services, and redesigning those services to reduce unnecessary frictions.

Administrative burden ideas that came from researchers were embraced by a crucial but unexpected ally, which is the civic tech movement. This movement features technologists interested in improving public services, often with a prosocial motivation. These technologists can be found in federal organizations until relatively recently (such as 18F and the US Digital Service, before they were eliminated by the Trump administration), the nonprofit space (e.g. Civilla, Code for America and US Digital Response), and the for-profit space (Nava and Propel are examples). The administrative burden framework provides a language and logic for this group to demonstrate how their work in improving the design of digital services mattered (Pahlka 2023; Schrock 2024). The engagement of civic tech in government also encouraged the use of new problem-solving tools. An example is human centered design (Herd et al 2024). When done well, human centered design involves ethnographic accounts of user experiences, as well as observations of those experiences. It seeks to directly capture and make sense of how human beings engage with administrative processes, and how those processes could be redesigned to be easier on those humans. We might think of this as an example of a movement and practice that seeks to bring metis into government.

2 Lessons About Problem-Solving

Rather than focus on one big point, I will draw out what I think are several lessons on problem-solving and learning in the civil service system. These lessons are provisional, intended more as provocations than definitive claims.

Are we hiring problem solvers? A classic claim of organizational learning is that learning occurs when organizations encode lessons from individuals. But this requires hiring the right people in the first place. The current version of the federal civil service system does not score well on this count. A requirement to preference veterans, who make up about 3 in 10 civilian federal employees, makes it harder for talented non-veterans to compete. Hiring processes can be painfully slow, especially if security background checks are involved. Screening processes often reward a willingness of the candidate to simply declare their excellence, rather than to show it. Innovations with hiring have offered an effort to solve this problem, by using Subject Matter Expert Qualification Assessments (SME-QA) (Nitze and Sinai 2022). This approach was developed by technologists working in government who were frustrated that the hiring process resulted in them being unable to hire excellent applicants who were not selected as finalists. The SME-QA process involves these experts in reviewing resumes, rather than deferring to HR specialists. It also allows them to ask for demonstration of skills to verify their qualifications. This makes it much more likely that the government ends up hiring those with the best skills rather than those best able to game a brittle hiring process. This represents a model of adapting existing hiring processes to make them more likely to hire talented people. Policymaker frustrations around hiring – and firing – constraints is creating pressure for more radical changes, including the move to at-will processes that eliminate most civil service rules.

Public organizations need to embed institutions of change. Public organizations lack market pressures to innovate. Therefore, they need to find other tools to encourage learning and problem-solving. One approach is to “embed institutions of change” (Moynihan 2022a). This seeks to take advantage of one bureaucratic strength, which is institutionalization: creating offices, routines, and requirements, and modifying these incrementally over time. Institutionalization can be used to embed values of innovation. Both examples identified above – performance management and burden reduction – fit with this model of change. They do not seek to reinvent government but to create or repurpose organizational routines to facilitate change. Performance management reforms most obviously relied on legislation, but was also able to amend the routines it relied on over time, from routines of data collection and dissemination to routines more tied to performance information use. The stability of the legislative base of GPRA meant that the reform was not easily swept away, and created a community who observed how well the original GPRA model was doing, and had ideas to hand when Congress was willing to revisit this model. The administrative burden framework relied on an executive order, but also tied the reform to existing processes: budget reporting and Paperwork Reduction Act requirements. The logic of this approach is that government runs on routines. It stands in contrast to more visible reform efforts, and assumes that change will be gradual, but is more likely to last if it takes hold.

The embedding of new routines must ultimately align with the organizational culture. Routines that provide permission structures to identify and solve existing problems tilt the culture toward innovation. Such changes can also encourage the adoption of new tactics. For example, Veterans Affairs had embraced greater attention to improving service quality, creating a customer experience office, embracing tools of human centered design, and consistently using metrics to drive up trust in the services (Herd et al 2024).

Blame avoidance, reputation, and negativity bias affect public solving. In the absence of a bottom line, the employees of public organization place more emphasis on building and maintaining other forms of capital. This includes the reputation of both the individual and the organization. A variety of work has examined this, most strikingly Weaver (1986) on blame avoidance, and Carpenter (2014) on reputation. The implication then is that problem-solving efforts are observed through the lens of credit and blame. In a politicized environment, civil servants tread cautiously. Organizations have incentives to develop and showcase expertise in specific reputational domains that allow them to protect their autonomy (Carpenter 2014). When obvious organizational failures occur, actors scramble to cast blame on one another (Hood 2010).

The emphasis on blame and reputation may not be problematic if actual failures correlate with reputational failures. Indeed, accountability for failures provides a compelling political incentive to identity and solve problems in the future. But this correlation does not always hold. Research on negativity bias in public settings points to an asymmetry in how success and failure are rewarded (James et al. 2020). Negativity bias draws from the psychological concept of loss aversion, which suggests that humans are more emotionally engaged by bad events than good. This simple insight offers a significant number of implications. While the shift to performance measurement was intended to improve public sector outcomes, people tend to focus on negative outliers more than high performers. As a result, civil servants do not have strong incentives to create excellent public services; they instead are incentivized to avoid visible and embarrassing failures. It is possible to build a problem-solving strategy around this approach using performance metrics, which is that civil servants should use data to proactively identify poor performers and invest resources in addressing those outliers (Holm 2018). But this also means that civil servants are more broadly concerned about minimizing scandals that are not materially related to performance. For example, when the General Service Administration held an employee retreat in Las Vegas in 2012, it was deemed unseemly enough to become front page news. It led to the federal government significantly constraining travel for years, reducing investments in professional development for its employees.

Proceduralism as a barrier to problem-solving. Many formalized practices are bad ideas, or constraints that prevent innovation. The legal scholar Nicholas Bagley refers as “the full panoply of formal legal obstacles that an agency must negotiate in order to complete a particular action” (Bagley 2019, 351). While the idea of unaccountable bureaucrats has become a fixture in political discourse, the opposite may be true: civil servants are accountable to too many processes and demands. It is harder to fire them relative to their private sector peers, but in other respects they must act (or often, fail to act) in a context of multiple rules and constraints that discourage bold action. In most cases, these accountability mechanisms are hard to disagree with in isolation. Who could oppose the transparency offered by FOIA, the participation opportunities offered by public comment requirements, or protections against corruption in the procurement and grant-making processes? But cumulatively, they slow decision processes or encourage caution.

Folk proceduralism: beliefs about the nature of constraints can become more powerful than the constraints themselves. Agencies will often add rules and SOPs that over-comply with formal direction. Over time, it becomes assumed that these rules are required by statute, rather than reflect an interpretative guide to that statute, one that can be changed and reinterpreted. These beliefs become a form of folk wisdom embedded into organizational culture. This tendency toward folk proceduralism is encouraged by overly cautious General Counsels, but it can also be addressed by more action-oriented general counsels. For example, two organizations with the same set of constraints can act differently depending on whether the organizational leader, and culture, has a pro-action perspective or a cautious approach (Pandey et al 2007).

Expertise needs protection from politicization. The patterns of problem-solving described here would be improved if they were supported by political actors. For example, Congressional indifference to performance data has dramatically weakened the incentives for civil servants to invest marginal effort into performance improvement. Similarly, the Trump administration has shown much lower interest in using burden reduction frameworks than their predecessor. Even under normal governance conditions, patterns of political inconsistency and neglect undermine any framework for problem-solving.

Under conditions of outright political hostility to civil servants, the prospect for problem-solving become even bleaker. One of the most compelling theories from political science about the civil service is the expertise model that Gailmard and Patty (2012) have advanced. The theory uses both formal modeling and historical accounts to understand why civil service systems exist. A key assumption is that many problem-solving skills that civil servants develop have limited value in the private sector. Therefore, they need incentives to invest in developing expertise. Employees who work in unstable environments, where they can easily be fired, will be less likely to invest in expertise.

Employees who cannot put their expertise to use in policy making will likewise be discouraged from cultivating their expertise. Politicians have an interest in a high-functioning public sector, and therefore should be willing to offer that stability alongside with some promise of autonomy to public employees. The most credible way in which they have found to do so is through formal civil service systems that provide job security and at least some measure of influence.

A good historical demonstration of this point comes from Aneja and Xu (2024), who show that the creation of the civil service system improved the performance of a core administrative task: the postal system. Mail was more likely to be delivered more reliably and with fewer errors after civil service reforms. The key mechanism was a reduction in turnover, allowing mail workers to accumulate experience. The centrality of stability to the accumulation to expertise remains clear more than a century later. In studies that cast light on why political appointees scored lower on program performance assessment, Lewis (2007) and Gallo and Lewis (2012) show that less experience in task-specific roles, and greater experience working on campaign positions, made political appointees less effective in managing program performance.

The traditional equilibrium that gave rise to civil service systems breaks down if the politician values loyalty over competence, leading them to revoke civil service systems, and exclude civil servants from meaningful opportunities to solve problems. Evidence from the first Trump administration has found that as employees perceived their work environment becoming more politicized, they became less likely to invest in expertise, and more likely to consider exiting (Richardson 2019).

At the time of writing, we stand on the edge of a wave of politicization that the federal government has not seen since the creation of the civil service system. President Trump has claimed that executive authority allows him to reclassify federal employees engaged in policy discussions to political appointee status (Moynihan 2022b). In his second term, he has gone even further, firing federal employees for reasons unrelated to their performance, and targeting agencies perceived as liberal for larger cuts (Bonica 2025).

The threats of reclassification and firing are sometimes accompanied with patterns of intimidation of individual employees (Moynihan 2024a). For example, the prior and incoming head of the Office of Management and Budget, Russ Vought, has said:

We want the bureaucrats to be traumatically affected. When they wake up in the morning, we want them to not want to go to work because they are increasingly viewed as the villains. We want their funding to be shut down so that the EPA can’t do all of the rules against our energy industry because they have no bandwidth financially to do so. We want to put them in trauma.

The Heritage Foundation, a conservative think tank that helps Republicans plan their presidency, has submitted tens of thousands of FOIA requests, seeking the emails of individual employees in the hopes of finding damning information. The American Accountability Foundation, a group funded by Heritage, created a website listing the name, picture and salaries of civil servants that they classified as “targets” for the incoming Trump administration. Elon Musk posted the identities of individual employees he regards as performing “fake jobs”, under the guise of efficiency (Moynihan 2024b). The result of individualized enemies list was that many such employees were fired or resigned, and faced threats to their personal safety (So et al 2025). Hostility toward bureaucrats is nothing new, but conditions of personal attacks and overt intimidation, led by the governing party, is different from what we have seen before in the United States. Such patterns are more apparent in authoritarian regimes elsewhere, such as Brazil under Bolsonaro (Story et al 2023).

What might conditions of overt hostility and intimidation, alongside the threat of loss of their jobs, do for bureaucrats? Under these conditions, more will leave, and those who stay are unlikely to be willing to challenging assumptions or claims of the governing party, even when they are wrong. It is hard for civil servants to solve problems if their political masters treat them as the problem to be solved.