Institutional Economics and the Cost of Capital for Infrastructure Projects

ABSTRACT

The growing use of public-private partnerships (PPPs) in infrastructure delivery has rekindled the public versus private cost-of-capital debate. Early debate concluded that the social cost of public-sector capital is below that of private capital because risk can be spread across numerous taxpayers. Intervening research focused on the agency costs that arise due to increased separation between equity holders and managers and the various mechanisms to control those agency costs. We analyze differences in public versus private residual claims and their associated agency costs. We conclude that earlier discussions that omitted agency-cost analysis should be revisited.

KEYWORDS:

• Cost of capital
• public-private partnerships
• infrastructure
• Arrow-Lind Theorem
• agency problems
• risk sharing
• risk spreading
• corporate governance

JEL Classifications:

• H11
• H21

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Global private investment in infrastructure is increasing. Private equity funds raised more than $300 billion in capital between 2013 and 2018 for investments in energy, transportation, telecom and other infrastructure assets. The highest fundraising year on record was 2018 (Infrastructure Investor 2020; Gottfried 2018).

2. Investors today have access to a wide range of financial products, including direct or indirect investments in private equity funds that offer exposure to infrastructure projects. Depending on the jurisdiction of a public-sector project (i.e., in the United States, the federal government can achieve broader diversification than state and local governments) and the identity of equity providers in a particular project (which often include private equity funds with large pension funds as a substantial portion of their LP base), private sector investment vehicles may spread risk more effectively than traditional public financing. Instead of empirically analyzing this question, we focus on the tradeoffs between greater risk-spreading and agency problems that arise with ownership dispersion. We study the different governance mechanisms available to taxpayers and private equity investors as residual claimants in infrastructure projects.

3. Revenue bonds are backed only by project revenue, as opposed to general obligation bonds. We assume that taxpayers do not either explicitly or implicitly guarantee the bridge’s debt. Taxpayer guarantees of project bonds represent an implicit transfer of risk from bondholders to taxpayers. Our goal is to abstract away from such risk transfers.

4. Since public pensions often represent a significant portion of the limited partner (LP) base of infrastructure-focused private equity funds, the distinction between public and private residual claimants may be less pronounced than is often realized. Also, under both financing scenarios, the public retains ownership over the bridge (PPP contracts often include a long-term lease to the operating private consortium). Private financing through a PPP, however, at least partially separates performance risk from project ownership.

5. We also assume that the β in the capital asset pricing model discussed below does not vary with the nature of the risk bearers. A more complete model would allow the project’s revenue and costs to depend on the identity of the residual claimants. Although managerial performance is likely to affect the amount of cash flow generated by the project, we assume for simplicity that the choice of financing (i.e., PPP vs traditional public financing) does not affect project quality or the cash flows.

6. Most observers view this as occurring through the tax and transfer system. See e.g. Lucas (2012).

7. Concerns have been expressed regarding reliance on the risk-free rate as the cost of capital, since taxpayers ultimately bear risk of cost over-runs and asset performance. See e.g., Lucas (2012).

8. The Modigliani-Miller Theorem holds that, under certain assumptions, the value of the firm is independent of its capital structure. We recognize that our analysis suggests that the optimal financing structure $\left(\frac{E^{P\ast }}{V},\frac{E^{T\ast }}{V}\right)$likely depends on the identity of project risk bearers, and that differentiating between the two equity types implies that V should also vary based on the risk-bearing group. If, for example, taxpayer-provided equity capital were less costly than private-investor capital, then the optimal capital structure may include a greater proportion of equity in the taxpayer case. We leave those issues for future work.

9. We suppress the subscript e in all subsequent equations for brevity.

10. See e.g., “Federal Reserve’s Key Policies for the Provision of Financial Services,” Board of Governors of the Federal Reserve, 27 November 2019, available at: https://www.federalreserve.gov/paymentsystems/pfs_policies.htm (accessed 24 January 2020).

11. Courts occasionally “pierce the corporate veil” and see through legal entities to pursue the assets of a firm’s residual claimants in circumstances where they find shareholders are exploiting the corporate form to commit fraud or violate laws. However, that happens relatively infrequently as the bar for veil piercing is usually quite high.

12. Limited liability may affect expected returns by providing prospective investors with greater certainty as to future net cash flows, rather than directly affecting the cost of capital. Through either mechanism, however, it reduces the risk-bearing compensation that equity investors require for a given company or project.

13. This is also true of early literature in finance. The Modigliani-Miller Theorem, for example, assumes zero agency costs.

14. Severability of risk-bearing obligations from tax residence requires tradable residual claims. Tiebout (1956) argues that taxpayers “vote with their feet” by changing jurisdictions in response to varying baskets of government services at a variety of prices (i.e. tax rates). The transaction costs of this option are so high relative to selling tradable ownership shares in firms that we do not view the two as substitutes. It is unlikely that taxpayers would change jurisdictions solely in response to the costs and benefits of a particular infrastructure project. We thus take the fraction of taxpayer ownership to be fixed over time.

15. Today, private capital for infrastructure is largely delivered by special-purpose vehicles controlled by private equity firms specializing in infrastructure and real assets. Although limited partner (LP) interests in these firms are typically not tradable, the funds compete for LP capital, and the management fees, carried interest, and other terms they negotiate with their investors can be seen as analogous to market prices. LPs are often restricted from withdrawing committed funds, or face penalties for doing so. Some infrastructure investment vehicles (e.g., energy yieldcos) are publicly traded. For simplicity, we assume that residual claims in private infrastructure investments are tradable.

16. Numerous empirical studies support the proposition that stock markets help mitigate agency problems. Managerial removals are correlated with negative abnormal stock returns. Also, managerial pay packages are often structured to align managerial and shareholder incentives. See for example Coughlan and Schmidt (1985), Weisbach (1988), and Warner, Watts, and Wruck (1988).

17. As discussed above, limited liability facilitates takeover transactions by allowing shareholders to invest without considering the identity and wealth of other shareholders.

18. Notably, some literature shows that the takeover threat can induce officers and directors to take protective measures that destroy shareholder value. See e.g. Bebchuk and Cohen (2005).