There is a theory that miners have no power. This is distinct from the closely related Proof of Work Fallacy. The theory rests on the assumption that miners are subject to economic pressures that preclude sustained effective attacks. This theory leads people to believe that mining can be strongly pooled as long as merchants are not centralized, as the economy can control the behavior of mining, rendering the system secure. The consequence of this invalid theory is complacency regarding the insecurity caused by pooling.
The theory holds that if majority hash power double-spends then merchants will necessarily increase confirmation depth requirements, increasing the cost of subsequent attacks. At some point an equilibrium is reached where greater depths are considered sufficient for exchange. Given that this would preclude double-spending altogether, there would be no advantage to sustaining the attack. The theory accepts that attacks can happen, but not frequently enough to materially reduce utility.
The theory also holds that a miner cannot avoid selecting the highest fee transactions as this reduces relative reward, enriching other miners. This is presumed to lead to a loss of majority power and therefore an inability to continue. This aspect of the theory implies that miners cannot effectively censor.
The theory also considers that selfish mining by majority hash power is feasible, but in the absence of double spending and censorship, there is no adverse consequence to the economy. In this case the majority simply becomes the one miner as all others are unable to retain rewards. Despite lack of competition, hash rate and fee levels are maintained by the ever-looming possibility of competition.
Yet miners and merchants are trading partners, engaged voluntarily in mutually-beneficial activity. As explored in the Balance of Power Fallacy, neither can control the other and price is the resolution of all preferences. This would seem to support the theory, however the theory does not address the threat, and is actually a red herring. Bitcoin is designed to defend against non-market forces, specifically the state. Market forces are never a threat to the market itself.
The pooling of hash power eviscerates security, as states can simply co-opt it. But states can also build their own mines to the same effect. Bitcoin therefore requires both significant hash power and distribution of that power among people who are willing and able to risk state controls.
The state is an economically rational actor. Inflation is profitable for the issuer. Bitcoin's widespread use would prevent states from effectively levying the inflation tax. State attacks are therefore expected, and analogous attacks are commonplace. It is practically inevitable that states will subsidize attacks, but even the possibility invalidates the theory.