Why Asset Specificity Overrides Frequency in the Corporate Make-or-Buy Decision

When a retail consumer decides whether to cook dinner or order delivery, the calculation rests almost entirely on the immediate spot price of ingredients versus the restaurant's markup. But when a multinational corporation decides whether to manufacture its own microchips or outsource them to a foundry, the spot price is largely irrelevant. The corporate make-or-buy decision is governed instead by transaction cost economics, a framework that measures the friction of doing business in the open market rather than just the accounting cost of production.[3]

The theory originated with Ronald Coase in 1937 and was formalized by Oliver Williamson, who won the 2009 Nobel Memorial Prize in Economic Sciences, sharing the 10 million Swedish kronor award for his work. Williamson established that the transaction, rather than the commodity, is the basic unit of economic analysis. He argued that market exchanges are not frictionless; they require searching for partners, negotiating terms, writing contracts, and enforcing agreements. When those frictions become too expensive, a firm will choose to bring the activity inside its own hierarchy.[4][5]

To predict when a firm will abandon the market and build internal capacity, Williamson identified three primary determinants: asset specificity, uncertainty, and transaction frequency. Asset specificity measures how tailored an investment is to a particular relationship. A standard delivery truck has zero asset specificity because it can haul goods for any client. A custom-built stamping die designed exclusively for one specific car door panel has high asset specificity, as it loses its value if the relationship with the automaker ends.[3]

When asset specificity is high, companies face the hold-up problem. If an external supplier builds a highly specialized factory to serve a single buyer, the buyer can later demand price cuts, knowing the supplier has no alternative customers. Conversely, the supplier could threaten to halt production, knowing the buyer cannot easily find a replacement. To avoid this mutual vulnerability, the buyer simply acquires the supplier or builds its own internal division.[5]

The empirical evidence heavily supports this mechanism. A 2006 meta-analysis published in the Academy of Management Journal aggregated 143 primary studies covering over 50,000 firm-level decisions and found a robust correlation coefficient of r = 0.29 between asset specificity and the decision to internalize production. This makes asset specificity the single strongest predictor of vertical integration, consistently outweighing other environmental factors across multiple industries.[2]

Uncertainty acts as an amplifier to asset specificity. In highly unpredictable markets—where technology shifts rapidly or demand fluctuates wildly—writing a comprehensive, long-term contract is impossible. Williamson termed this bounded rationality, noting that human managers simply cannot foresee every future contingency. When uncertainty is high, the cost of constantly renegotiating incomplete contracts pushes firms toward internalizing the transaction, where managers can adapt by fiat rather than by contract.[3][4]

However, uncertainty alone is rarely enough to force a make-or-buy shift. A 2025 study in SciOpen examining platform boundaries found that environmental uncertainty requires a 40 percent higher transaction frequency threshold to trigger integration compared to asset specificity. If the asset is generic, firms prefer to absorb the uncertainty in the open market rather than commit capital to a permanent internal division.[6]

Transaction frequency serves as the economic justification for the overhead of integration. Setting up an internal department requires fixed costs: management salaries, human resources infrastructure, and physical space. If a transaction occurs only once every 24 to 36 months, the firm will endure the high contracting costs of the open market because the fixed costs of internalizing it would be ruinous.[5]

This dynamic is clearly visible in the public sector. A study in Evidence & Policy examining 120 public school districts found that administrators apply transaction cost logic when deciding whether to conduct internal educational research or hire external consultants. Districts only build internal research and evaluation departments when the need for highly specific, localized data occurs with high frequency. For rare, standardized assessments, they default to buying the service from the market.[1]

The rise of digital platforms has introduced new complexities to the make-or-buy equation. Platform companies often rely on third-party developers to build complementary services, effectively outsourcing innovation. The 2025 SciOpen research indicates that platforms only internalize these boundary services when the third-party application becomes so deeply integrated into the core user experience that the platform risks losing control over its own ecosystem.[6]

In these digital environments, data itself has become a highly specific asset. When a vendor requires access to a company's proprietary customer data to function, the risk of opportunism—which Williamson famously defined as "self-interest seeking with guile"—skyrockets. Firms are increasingly pulling data analytics back in-house, not because external vendors are more expensive, but because the contracting costs of securing the data are too high.[4][7]

The make-or-buy threshold rests on a strict balancing act between the bureaucratic inefficiency of internal management and the contracting hazards of the open market. As long as the cost of market friction exceeds the overhead of corporate hierarchy, the boundaries of the firm will continue to expand. The exact placement of that boundary shifts with every new technology, but the underlying calculus of specificity, frequency, and uncertainty remains the definitive rule.[7]