This paper quantitatively examines the marginal value of corporate cash holdings (MVC)—that is, the change in shareholder value resulting from an additional unit of cash. In a frictionless market, the marginal value of one unit of cash equals one. However, in reality, various frictions and constraints can cause MVC to deviate from this benchmark. For instance, financing constraints may restrict investment in profitable projects, making internal cash more valuable (MVC > 1), while agency conflicts may lead to inefficient use of cash, reducing its value (MVC < 1). Clearly, MVC plays a critical role in corporate financing and investment decisions, and is also vital for understanding corporate governance, which explains its long-standing importance in both academic and practical research.

Currently, the mainstream method for estimating MVC is the reduced-form approach developed by Faulkender and Wang (Journal of Finance, 2006, hereafter FW), which regresses stock returns on changes in cash holdings (scaled by beginning-of-period market value). The coefficient on the cash change is interpreted as the MVC. However, this paper identifies critical problems in the FW method that prevent it from accurately capturing the true marginal value of cash. The key issue lies in the requirement of the cash flow identity inherent in corporate financial data. The FW method omits relevant variables (such as cash flows), and under the constraints of the identity, adding omitted variables results in perfect multicollinearity, rendering the regression invalid. Using both simulated data and real financial data, this study demonstrates that the FW coefficient is not a valid estimator of MVC and is more consistent with the marginal value of cash flow—an omitted variable. As a result of this issue, many studies using the FW method report implausible MVC estimates. For example, some studies find an MVC of 1.7, which implies that the average cost of external financing would need to be at least 0.7 per dollar raised—an assumption rarely justified.

To better estimate MVC, this paper proposes a structural approach. A dynamic model of shareholder value is constructed, incorporating key factors such as financing and agency frictions that influence corporate cash holdings. Based on this model, the MVC functions for shareholders and managers are derived. The model parameters are then estimated using the Simulated Method of Moments (SMM) with firm-level financial data. Finally, combining the model's parameter estimates, the derived MVC functions, and the observed financial data, more accurate MVC estimates are obtained.

Empirical findings show that MVC estimated via the structural model deviates from one in economically reasonable ways. Firms with higher capital stock and cash reserves exhibit lower MVC, as they rely less on external finance. MVC follows an inverted-U pattern with respect to leverage—moderate debt increases MVC, while excessive leverage leads to debt overhang and lowers MVC. Smaller firms and those facing greater financing constraints have higher MVC, since additional cash helps ease investment frictions. The model’s validity is further supported by two quasi-natural experiments. Following industry tariff reductions, corporate governance improves, managerial incentives increase, and MVC changes accordingly. Similarly, the 2004 American Jobs Creation Act (AJCA), which triggered profit repatriation, shows that MVC significantly predicts a firm’s likelihood of repatriating oversea profits.

This study contributes to the literature by identifying and explaining the limitations of the commonly used FW method in MVC estimation, and by offering a more robust structural estimation framework. It provides new tools and perspectives for analyzing corporate investment and financing decisions.

About Di Li

Li Di is a tenured associate professor at the HSBC Business School of Peking University and the deputy director of the Peking University HSBC Financial Research Institute. His main research areas include corporate finance, corporate governance, mergers and acquisitions, and structural estimation. He has published papers in renowned academic journals such as The Journal of Financial Economics, Management Science, Journal of Financial and Quantitative Analysis, and China Economic Quarterly.