You can't trust people to be free - at least not in all ways, or all of them, or perhaps even all the time. I am not aware of any comprehensive study of the nature of corruption. Therefore, I'm working on the intrinsic nature of corruption myself.

In the market process, corruption is renamed collusion. It is the Achilles heel of capitalism and it arises in corporations at the same rate that corruption arises in governments. Some other forms it takes are planned obsolescence, artificial scarcity, and even artificial desire by rebinding instincts of people to create markets. That is all to say it is a systemic property and it appears with absolute certainty given enough time.

The AnCap style solution of disbanding higher organization to eliminate corruption at higher levels does impact the problem. It does eliminate corruption from those levels. It doesn't eliminate corruption altogether but it does pragmatically eliminate Pyramids, space travel, high rises, terraforming, etc. It eliminates any endeavor which implicitly requires the organization of people to act cohesively to accomplish. It is whack-a-mole. The problem re-expresses itself in a different form and place. There is value in whack-a-mole solutions - they are like experiments that give you another instance observation and more information about the rules that define the phenomena.

Corruption exists at the individual level as well, in fact it winds into the individual. The dopamine reward system which has evolved to shape behavior and ensure the survival of species is hijacked by addictions and other behavioral viruses. Viruses are a concise instance of the 'corruption' phenomena. Cryptography is another concise framing with its trust system/code breaker adversarial model. Godel's incompleteness theorem ensures that the competition will be a perpetual part of systems. Informatics supplies the proof of universal computability that enables it to spread to unrelated systems.

I find that corruption is a fundamental feature of information. I think that looking at its anthropic forms can yield useful insights, but the real key to understanding it will come from observing it in its most irreducible forms in physics, thermodynamics, informatics, game theory, etc.

At this point, most will sort of throw their hands up and conclude you should just accept it and focus on something actually solvable. Except, there are some existential implications for mankind that will arrive in the near term future. Without a deep understanding of this to develop defenses, global nuclear war is going to seem like a quaint alternative.