For Oliver E. Williamson, existence of firms derives from 'asset specificity' in production, where assets are specific to each other such that their value is much less in a second-best use. This results in problems if the assets are owned by different firms (like supplier and purchaser). The reason behind it is that it will results in protracted bargaining regarding the gains from trade since both agents are likely to become locked in a position where they are no longer competing with a (possibly large) number of agents in the whole market and incentives are no longer there to represent their positions honestly: large-number bargaining is transformed into small-number bargaining.

If transaction is a recurring or lengthy one, a continual power struggle occur regarding the gains from trade, further increasing transaction costs. So re-negotiation may be essential. Furthermore there are liable to be situations where a purchaser may necessitate a particular, firm-specific investment of a supplier that would be profitable for both. Though after the investment has been made it becomes a sunk cost and the Purchaser can endeavour to re-negotiate the contract such that supplier may make a loss on the investment (this is hold-up problem that takes place when either party asymmetrically incurs substantial costs or benefits before being paid for or paying for them). In this type of a circumstances, the most efficient approach to overcome continual conflict of interest between two agents (or coalitions of agents) may be the removal of one of them from the equation by takeover or merger. Asset specificity can also apply to some extent to both human and physical capital so that hold-up problem can also take place with labour (for example labour can threaten a strike due to the lack of good alternative human capital though equally the firm can threaten to fire).

Perhaps, the best constraint on such opportunism is reputation (instead of the law, due to the difficulty ofwriting, negotiating and enforcement of contracts). If a reputation for opportunism significantly damages an agent's dealings in the future, this changes the incentives to be opportunistic. 

Williamson opines that limit on the size of the firm is partly an outcome of costs of delegation (as a firm's size increases its hierarchical bureaucracy does too) and partly the result of large firm's increasing inability to replicate high-powered incentives of residual income of an owner-entrepreneur. To a certain degree this can be attributed to the nature of large firms to make sure that its existence is more secure and less dependent on the actions of any one individual (increasing the incentives to shirk) and since intervention rights from the centre characteristic of a firm be likely to be accompanied by some form of income insurance to compensate for the lesser responsibility, thus diluting incentives. Milgrom and Roberts (1990) show that increased cost of management is a result of employee's tendency to provide false information beneficial to themselves that increases the cost offiltering information. This grows worse with firm size and more layers in the hierarchy.

Empirical analyses of transaction costs have hardly ever attempted to measure and operationalize transaction costs. Research that efforts to measure transaction costs is the most critical limit to efforts to validation and potential falsification of transaction cost economics.