Politically misaligned bureaucrats incur in greater cost overruns, are generally less motivated, & are perceived as less loyal to their missions

Ideology and Performance in Public Organizations. Jorg L. Spenkuch, Edoardo Teso & Guo Xu. NBER Working Paper 28673, April 2021. DOI 10.3386/w28673

Abstract: We combine personnel records of the United States federal bureaucracy from 1997-2019 with administrative voter registration data to study how ideological alignment between politicians and bureaucrats affects the personnel policies and performance of public organizations. We present four results. (i) Consistent with the use of the spoils system to align ideology at the highest levels of government, we document significant partisan cycles and substantial turnover among political appointees. (ii) By contrast, we find virtually no political cycles in the civil service. The lower levels of the federal government resemble a "Weberian" bureaucracy that appears to be largely protected from political interference. (iii) Democrats make up the plurality of civil servants. Overrepresentation of Democrats increases with seniority, with the difference in career progression being largely explained by positive selection on observables. (iv) Political misalignment carries a sizeable performance penalty. Exploiting presidential transitions as a source of "within-bureaucrat" variation in the political alignment of procurement officers over time, we find that contracts overseen by a misaligned officer exhibit cost overruns that are, on average, 8% higher than the mean overrun. We provide evidence that is consistent with a general "morale effect," whereby misaligned bureaucrats are less motivated.

---

Panel A of Figure 7 shows that the overrepresentation of Democrats increases as we move up the hierarchy. Among employees in grades 1-12 of the GS, we find about 50% of Democrats (30% of Republicans and 20% of independents), which rises to approximately 56% at the top of the GS (grades 13-15), and to 63% among career SES.34 Intriguingly, this finding appears to be driven in large part by selection on observables.  First, Democrats have, on average, higher levels of human capital than Republicans. In Table 4, we report estimates from regressing indicators for educational attainment on a bureaucrat's political affliation. In order to measure education at entry, we restrict the sample to the first quarter in which the employee is observed.35 According to our results, Democrats are 6.6 p.p. more likely than Republicans to hold a college degree (column 1), and 8.3 p.p. more likely to have some form of post-graduate education (column 4). We continue to observe differences in human capital after controlling for bureau (columns 2 and 5) and pay-level fixed effects (columns 3 and 6)|although the gap between Democrats and Republicans does narrow. The pattern of coefficients in Table 4, therefore, suggests that higher human capital allows Democrats to be hired in bureaus and occupations that require more advanced skills as well as at higher steps of the hierarchy (see also Appendix Table A3).  Moreover, the fact that there do remain residual differences after accounting for bureau and pay grade implies that, even within comparable jobs, Democrat civil servants tend to be positively selected.

In addition to being positively selected at the time of hire, Democrats are more likely to be promoted after they enter the bureaucracy. In Table 5, we present estimates of partisan differences in promotions from grades 13{15 of the GS to the career SES (columns 1{3), as well as promotions from grades 1{12 of the GS to grades 13-15 (columns 4{6).36 Given that promotions are rare events at the quarterly level, all estimates in Table 5 are multiplied by 1,000. The results show that Democrats are more likely than Republicans to be promoted to higher levels of the hierarchy (columns 1 and 4), with a sizeable share of the gap being attributable to differences in educational attainment and the bureaus in which they serve (columns 2{3 and 5{6).37 The second factor that helps to explain greater overrepresentation of Democrats at higher levels of the bureaucracy is their lower propensity to exit. To illustrate this, Panel A of Figure 8 plots survival curves by partisan affiliation. While about 5% of civil servants of either party exit after the first quarter, the share of those who remain within the federal government as time progresses is significantly higher for Democrats. In Panel B of Figure 8, we repeat the exercise in regression form, controlling for bureau  quarter-of-entry fixed effects. After 10 years, Democrats are about 4.5% more likely than Republicans and independents to be still employed in the civil service.

5 Conclusion

A central question in the governance of any organization is how to align the objectives of leaders with those of their subordinates. In this paper, we turn to the U.S. federal bureaucracy to study the role of mission alignment in organizations.

To this end, we combine administrative data on the near universe of federal government workers with data on all registered voters in the U.S. The resulting dataset allows us to shed some of the first light on the ideological leanings of a large number of individual civil servants, and thereby peek into the black box of \bureaucratic politics." We establish three stylized facts. First, politicians do use the limited power they have over personnel policies in order to achieve greater ideological alignment between themselves and the upper echelon of the bureaucracy. The political cycles in our data are consistent with the use of the spoils system to better align the highest layers of the bureaucracy with the goals of the president. Second, we find a remarkable degree of political insulation among career civil servants. In contrast to political appointees, we see virtually no political cycles in the civil service. Our findings, therefore, suggest that, at its lower levels, the federal government resembles a \Weberian" bureaucracy, which is largely protected from political interference. Third, Democrats make up the plurality of civil servants. In addition, we show that Democratic civil servants are especially overrepresented in higher layers of the bureaucracy. Any observed difference in career progression, however, is in large part due to selection on observables. Democratic-leaning bureaucrats have on average higher levels of educational attainment, and they are less likely to exit the civil service, which results in a greater accumulation of experience. Both of these two facts are consistent with the idea that Democrats have a higher proclivity for public service.

The existence of an impartial and politically insulated career civil service is often seen as the hallmark of good governance and a \Weberian state." While the insulation of the career civil service prevents political interference, civil servants may have their own preferences and ideological leanings, which can conflict with those of the president. As a consequence, to implement an administration's agenda, politicians and department heads often need to work with bureaucrats whose personal values are not aligned with the present mission of the organization. To shed light on the costs of such misalignment, we focus on a subset of civil servants who work across all departments of the government and for whom we can measure performance: procurement ofcers. Linking procurement contracts to the matched personnel and voter registration data allows us to study the mission-alignment of procurement ofcers across nearly all departments of the federal bureaucracy. Strikingly, we find that political misalignment increases cost overruns by 8%. We provide evidence that suggests that a general \morale effect" is an important mechanism behind this finding, whereby bureaucrats who are ideologically misaligned with the organizational mission have lower motivation. As political turnover leads to sizable mission-misalignment between politicians and civil servants, our findings provide direct evidence on the costs of political insulation of the bureaucracy, which should be traded off against the benefits of avoiding political interference. As more and more organizations embrace a mission-driven focus, our findings may have implications beyond the public sector.