Bank lending has proven to be quite resilient. The onset of digital financial markets has not upended the fact that most small- and medium-sized companies seek financing from banks, not via the issuance of bonds. And if game-theoretic literature from the 80s and 90s is any guide, this will not change anytime soon. Decentralized bond markets are yet to become an instrument of choice. And even if the blockchain now allows lenders to embed various contingencies, called covenants, into so-called smart contracts, these contracts nonetheless compete in a domain that humans already excel in: flexibility.
The view, prominent in the academic literature since at least the 80s, that we adopt here is that contracts cannot cover all contingencies that might befall a company. Increased competition, innovation breakthroughs, or legal challenges can hardly—or only at a high cost—be codified in a (smart) debt contract. What if new regulations cast doubt on the viability of a once-profitable project, necessitating a reduction in repayment for continued investment? Critics may argue that one might at least try to account for such a contingency. To this, the literature would retort: why bother anticipating these contingencies if efficiency can be (at least partially) restored by renegotiating the loan with one’s bank?
Understanding the contractual terms that will emerge when incentives for renegotiation are mutual requires a model. Following Gorton and Khan (2000), we consider a firm that chooses which project \theta\in\{G, B\}, if any, to pursue. A project leads with probability p_\theta to success and with probability 1-p_\theta to failure. Success results in a project-specific return A+z y_\theta, and failure results in a return A. Here, A is the failed project’s liquidation value (common to both projects), and z\geq 0 is a (random) parameter that scales the project’s return above the liquidation value upon success.
What motivates these payoffs? We seek to capture how non-contractible risk, z in the model, shapes project selection. The firm decides whether to pursue G or B . And while we assume that G is unambiguously more efficient than B , i.e., expected project-specific returns are greater under G so that p_G y_G>p_B y_B , the firm may lack incentives to pursue G . This occurs if the return to being successful under B exceeds the return under G , i.e., if y_B>y_G due to a lower investment cost under B . To see this, consider that the firm’s expected utility when pursuing project \theta is
u_\theta(R,z) = \max \big\{p_\theta \big(A+z y_\theta -R\big); 0 \big\}.
Then, the firm selects G only if the upside risk, as captured via z , is large or the repayment R upon success is low. Formally, for a given repayment level R , we can derive a threshold, \hat{z}(R) , so that the firm pursues G if and only if z\geq\hat{z}(R) . The inefficient project choice B results otherwise. If such a choice is to be reverted, the repayment R must be lower so that the firm gains a more significant upside in the event of success. To achieve this, the firm will renegotiate with the bank and point out the lack of incentives to pursue project G . Such renegotiation may be successful: If the greater probability of repayment under G instead of B outweighs the loss in repayment upon success, the bank will agree to lower the repayment R .
I should mention that the model does not stop here but enters into questions over liquidation rights prior to project selection that further strengthen the case for greater flexibility in corporate finance. Nonetheless, the above already provides a compelling example of how renegotiation can enhance overall efficiency. The initial question remains, however: why is financing exclusive to banks? Could renegotiation be successful in an alternative scenario where the firm had issued bonds? I am doubtful. In the presence of many small lenders, debt forgiveness (i.e., lowering the repayment R ) encounters a collective action problem with free-riding incentives. A situation may arise where lenders collectively—much like the bank—would benefit if each lender lowered the firm’s repayment if needed. The individual lender, however, does not internalize the positive incentive effect on other lenders. And so the debt is less likely to be forgiven.
All told, bond markets may have their virtues, chiefly size transformation and greater liquidity. But for small- and medium-sized companies, the inflexibility of contract characteristics discourages this source of financing. For them, bank lending is the past—and the future.
