Editors’ note: This column is based on CEPR Discussion Paper No. 21671 “Do Not Drain the Swamp! Populism, Bureaucracy and Economic Performance”. 

“Because we are draining the swamp, it’s very simple, and the days of rule by unelected bureaucrats are over”, declared President Donald Trump in his address to Congress on 4 March 2025. In Rome, Italian Prime Minister Giorgia Meloni has similarly warned that bureaucracy, the judiciary, oversight bodies, agencies, and other unelected institutions can operate like a “shadow government”, describing the “interference of the Deep State” as an obstacle to her government’s agenda. Similar rhetoric has also appeared in the Netherlands, and many other countries. Across the US and Europe, then, high-profile populist leaders increasingly cast public-sector experts, oversight bodies, and career civil servants as illegitimate elites standing in the way of the popular will.

Yet, this anti‐bureaucracy rhetoric obscures a real paradox: strong bureaucratic institutions may actually protect the economy from the very disruptions that populists invite. Recent studies confirm that populist regimes typically underperform economically, both at national and local levels. For example, Funke et al. (2023) find that long stretches of populist rule leave a deep mark: after 15 years with populist leaders, GDP per capita is on average about 10% lower than under a non‐populist counterfactual (see also Schularick et al. 2021). Bellodi et al. (2024) show that, in Italy, electing a populist mayor leads to worse fiscal performance and a larger share of procurement contracts with cost overruns, among other things (see also Morelli et al. 2021).

Our new analysis builds on this and asks what role the bureaucracy plays, if any, in mediating these negative effects of populist government (Morelli et al 2026). In so doing, we combine two approaches. First, we use cross‐country synthetic control comparisons around populist takeovers, replicating the analysis of Funke et al. (2023) and adding data on the independence of bureaucracy. Second, we exploit rich US state‐level data going back to 1929. We measure populism in governors’ rhetoric: we assemble a corpus of 3,355 gubernatorial speeches and measure populism with LLMs and the holistic-grading prompt of Tamaki et al. (2025). Moreover, we take advantage of the staggered introduction of civil-service reforms to test for the mediating role of an independent bureaucracy.

The results suggest that bureaucracy matters. In countries and states with career civil‐service protections, the usual drag of populist governance largely vanishes. But without those institutional cushions, the economy takes a hit. What matters is whether the bureaucracy is professional and insulated from political control. In the US state analysis, civil-service reform captures precisely this: meritocratic recruitment, tenure protections, and limits on patronage. These rules affect how easily a populist executive can turn potentially harmful policies into reality. 

We illustrate this with two cases. In Maine, Governor Paul LePage’s 2018 moratorium on new wind-energy permits threatened a sector that depended on predictable permitting rules, long investment horizons, and stable expectations about state energy policy. Yet not all parts of the state’s energy apparatus were equally exposed to direct gubernatorial control. Efficiency Maine Trust was not simply a line agency of the governor; it was a statutory public instrumentality governed by an independent board. This meant that, even while the governor sought to freeze wind permitting, other energy-efficiency programmes could continue under a separate institutional mandate. Illinois shows the same logic from a different angle. Rod Blagojevich’s administration was associated with attempts to politicise appointments and administrative decisions, but Illinois also had a long-standing civil service and anti-patronage framework. These rules did not prevent abuse, especially due to exempt appointments, boards, and politically controlled channels. But they made some personnel decisions more visible, legally contestable, and partly reversible.

These examples clarify why bureaucratic independence matters. Populist governments may create uncertainty by attacking agencies, politicising appointments, or making sudden commitments that are electorally attractive but costly for firms, investors, and households. An independent civil service cannot eliminate those risks, nor should it replace democratic choice. But it can slow down the most damaging moves, force them through legal and technical scrutiny, and make wholesale politicisation harder. This is why our results point to bureaucratic independence rather than administrative size: a large bureaucracy under direct political control may become an instrument of populist rule, whereas an independent civil service can act as a buffer against it.

More specifically, in our preferred specification, which compares each state with its own long-run trajectory while absorbing shocks common to all states, a populist governor (as identified from their speeches using large language models) is associated with 3.8% lower per-capita income when civil-service reform is not in force. In states that enacted civil‐service reforms and protected career bureaucrats from political hiring and firing, no such income penalty appears (and in fact the trend is slightly positive, offsetting the effect).

Figure 1 provides complementary evidence. In the 2010–2023 event study, income declines relative to control states at all four post-takeover horizons among states without reform. In states with civil-service protections, the estimates remain close to zero. Put differently, populism is associated with sharply different economic outcomes depending on whether governors inherit a professional civil service or a patronage-based bureaucracy.

Figure 1 Populist governors’ effect on state-level incomes with professional civil service versus patronage-based bureaucracy, 2010–2023

Notes: Event-study estimates for 2010–2023. Vertical bars show 90% and 95% confidence intervals. The sample contains eight switchers in states without civil-service reform and thirteen in states with reform.

Populist leaders often campaign on eliminating ‘unelected’ experts, but our evidence suggests that at least some expertise in government is precisely what economies need to stay stable. This mechanism is also consistent with Woo-Mora (2025), who studies the cancellation of Mexico City’s New International Airport and shows how a single populist policy shock can generate uncertainty, reduce investment, and impose sizeable short-run output costs. In a related column, Woo-Mora (2024) draws the broader institutional lesson that checks and the rule of law matter even when populist actions are presented as serving the ‘common good’. 

Our results speak directly to this institutional interpretation: without a professional civil service to temper erratic policy swings, populist governments end up doing more harm. In practical terms, preserving bureaucratic insulation and merit-based public administration is therefore critical for economic resilience. If reformers allow populists to dismantle these institutions, the short-term appeal of ‘draining the swamp’ could translate into larger long-term economic swings and weaker growth.

In conclusion, our paper is clear: do not drain the swamp! The very institutions that populists revile can be the economy’s lifeline under their rule. Weakening civil‐service protections will likely amplify populism’s already large costs for citizens’ prosperity. In short, a healthy democracy and a healthy economy both need strong, impartial bureaucracies, even (and especially) when populist leaders tell us otherwise.

References

Bellodi, L, M Morelli, and M Vannoni (2024), “A costly commitment: Populism, economic performance, and the quality of bureaucracy”, American Journal of Political Science 68(1): 193–209.

Funke, M, M Schularick, and C Trebesch (2023), “Populist leaders and the economy”, American Economic Review 113(12): 3249–88.

Morelli, M, M Vannoni, and L Bellodi (2021), “Expelling the experts: The cost of populism for bureaucratic expertise and government performance”, VoxEU.org, 7 April.

Morelli, M, D Petrukhin, and M Vannoni (2026), “Do Not Drain the Swamp! Populism, Bureaucracy and Economic Performance,” CEPR Discussion Paper No. 21671. 

Schularick, M, C Trebesch, and M Funke (2021), “The cost of populism: Evidence from history”, VoxEU.org, 16 February.

Tamaki, E R, Y J Jung, J Chatterley, G Mitchell, S Dzebo, C Sandoval, L Littvay, and K A Hawkins (2025), “Populism meets AI: Advancing populism research with LLMs”, arXiv preprint arXiv:2510.07458.

Woo-Mora, G (2024), “Populism’s original sin: Short-term economic consequences of populist policy”, VoxEU.org, 27 October.

Woo-Mora, L G (2025), “Populism’s original sin: Short-term populist penalties and uncertainty traps”, European Economic Review 172: 104917.