When dealing with complex human interactions, people have often found it useful to construct simplified models which, while they do not capture the full intricacies of a given interaction, are close enough so that one can abstract away from the human and draw conclusions based on mathematics or logic. In international relations one find the Realist model, which posits that all nations are unitary actors that have the same basic motivation, to become more powerful. This allows an analyst to construct scenarios in advance, which can be used to establish baselines for states' behavior when real crises erupt. (That this model becomes disastrous when academics confuse Realism with reality, unfortunately, is another matter.)
The trick when building such models is that they have to be simple enough to use, yet must include enough moving parts to do a good enough approximation of the real world. When your simplification is too simple, it will inevitably lead you down the wrong path.
In market economics, the standard model consists of two groups of actors: buyers and sellers. They engage in positive-sum interactions to their mutual benefit, efficiently allocating resources and creating a veritable paradise of industry and affluence. No formal place in this model is given to governments or analogous actors, despite their omnipresence. Governments are seen as intruders into the market to the extent that they are involved at all, spoiling with their touch the pristine operation of pure economics. Governments are expected to keep the peace and leave it at that.
As pleasant as such a vision might be for some, it bears no relation to reality, anywhere. The point need not be belabored; governments everywhere intervene in the market routinely. And not just governments, either; increasingly, elements of civil society act when they perceive a threat from seemingly unrelated transactions, for example when farmers will protest the construction of factories that could pollute their water.
The truth is, economic transactions do not occur in isolation. Third parties have an inherent and inalienable interest in the outcome of market interactions. First, such transactions could have incidental effects (such as pollution, or the disruption of established industries, or consuming common resources such as roads) that directly impact third parties. Second (especially with regard to governments), economic transactions cause power shifts. For a government (or an established company or community) to allow the free market to increase the power of an adversary, incidental to its normal operation, would be foolhardy; this is why we have blockades and trade embargoes.
Interested third parties, far from being intrusions onto the free market, are an integral part of that market and always will be. Market transactions will only occur once their expected utility outweighs any opposition put up by third parties. Sufficiently motivated third parties can disrupt markets entirely. That being the case, the typical market model is flawed for neglecting interested third parties.
An improved model would ideally take into account two additional factors: direct incidental impacts on surrounding parties, and changes in relative power. Can this be done without being bogged down in excessive complexity? I don't know. But thinkers who venerate the austere genius of the free market should acknowledge that their model is critically deficient in the same way that political Realism is, and it is equally naíve to rely exclusively on either.