Algorithms Don't Take Bribes: Empirical Evidence That Automated Governance Reduces Corruption and Abuse of Power in Public Procurement

People screw over people. Algorithms do not care about taking advantage of a position of power. Algos speed up everything and keep it more transparent. In recent years, public discourse has been dominated by the narrative that algorithmic systems in governance are inherently biased, opaque, and prone to exacerbating inequality and abuse. Critics argue that automation removes human accountability while embedding new forms of technological discrimination. This paper presents a contrarian empirical perspective: when properly designed and implemented, algorithmic governance can dramatically reduce corruption and abuse of office precisely because algorithms are impartial—they have no friends, no bribes, and no career incentives. Drawing on lessons from natural science, the paper shows that the most powerful governance mechanism in biology is itself an algorithmic process. Natural selection, formalized by Dennett (1995) as a substrate-neutral algorithm, operates without regard for status or alliances. This principle is mathematically captured by the replicator equation from evolutionary game theory (Taylor & Jonker, 1978; Hofbauer & Sigmund, 1998): \[ \dot{x}_i = x_i (f_i(\mathbf{x}) - \bar{f}(\mathbf{x})) \] where \( x_i \) denotes the frequency of strategy \( i \), \( f_i(\mathbf{x}) \) its fitness, and \( \bar{f}(\mathbf{x}) \) the average population fitness. Under these rules, corrupt or self-serving strategies are driven toward extinction once an evolutionarily stable strategy (ESS) of fairness emerges (Maynard Smith & Price, 1973). The analogy is direct: discretionary human governance tends toward stable “corruption equilibria,” while algorithmic systems shift populations toward honest, transparent equilibria—exactly as nature has done for billions of years. Empirical ResultsThe study examines large-scale public procurement reforms in Estonia’s fully digitized e-procurement system and Ukraine’s open-source ProZorro platform using Open Contracting Data Standard (OCDS) datasets (2015–2025). Results show striking reductions in corruption-risk signals (single-bidder contracts, price anomalies, supplier concentration, etc.). In Ukraine alone, ProZorro generated more than US$ 6–8.7 billion in documented savings. ReplicationThe full replication package (code + data) is included in the attached ZIP file `gandolfi-2026-algorithms-dont-take-bribes-replication-1.0.0.zip`. All analyses are fully reproducible.