The Financial Engineering of Infrastructure Privatization

Abstract

Problem, research strategy, and findings: Leasing government infrastructure to private investors has been proposed as a practical way to increase both public revenue and investment in aging facilities, yet questions remain regarding lease value. In particular, some recent private auction bids surpassed government's lease estimates for U.S. roads and parking systems by hundreds of millions of dollars. We argue such discrepancies are largely explained by the use of structured finance or financial engineering techniques; these lower capital costs and maximize quick investor payouts, yet are often ignored in lease agreements because governments do not understand them. Our approach models the separate effect of several deal parameters on the investment return of a hypothetical tolled facility. We find even modest financial engineering (such as interest rate derivatives and swaps, deferred payment sweeps, or mark-to-market accounting practices) increases the current value of future facility revenues far more than changes in lease length, tolls, or operating costs. The public sector undercharges for its infrastructure when it ignores how private investors package and assess future revenue.

Takeaway for practice: When leasing public facilities, governments would be smart to better understand potential investors’ capital structure and financial engineering strategies. Doing so avoids leaving money on the table; it also reduces the risks of future underperforming assets.

Research support: None.

Keywords:

• privatization
• infrastructure
• asset lease
• financialization
• toll road

Notes

1. We use the term privatization throughout the article even though the public sector maintains legal ownership of the asset because the leases in question shift the provision of transportation services from the public sector to the private sector through deliberate government action (see Starr, 1988).

2. Existing infrastructure is conventionally referred to as a brownfield asset in contrast to newly built greenfields. To avoid confusion with the other use of the term brownfield in the planning literature (i.e., an environmentally contaminated site), however, we will use the term existing infrastructure throughout the article.

3. Mier and Syal (2010) suggest that we are in the midst of a third wave of privatization deals that has moved away from large capital, large cash-flow enterprises like the Chicago Skyway to smaller parking facilities, surface transportation, port facilities, governmental facilities, and prisons.

4. A full discussion of these instruments is beyond the scope of this article. We provide greater detail in the Methods section (see below); for technical detail, see Yescombe (2002).

5. Scholars refer to the increasing size of financial markets and institutions as well as the increasing role of finance in non-financial sectors as financialization, arguing that it has occurred generally in the global economy over time; see, for example, Krippner (2005) and Orhangazi (2008).

6. The range of bids tendered has also been sizeable. In examples such as the Chicago Skyway and the Indiana Toll Road, Macquarie Investment Group's winning bids were, respectively, 161% and 35% higher than the second-place bids (Bel & Foote, 2009).

7. This earnings multiple is one standard means to compare deals; it measures the ratio of the up-front payment (the asset price) to the asset's annual revenues (earnings before interest, taxes, depreciation, and amortization, or EBITDA; Bel & Foote, 2009).

8. The IRR is a measure of the return on investment (profit) that compares an initial investment with a stream of returns to investors over time. It is reported as a percentage (Yescombe, 2002).

9. Ownership of the concession was often shared with an operating partner through a consortium or partnership legal structure. The firms typically follow the Macquarie model, where “a sponsoring manager—usually but not always an investment bank—establishes a separate publicly traded entity to own infrastructure assets while contracting out management functions to the sponsor” (Lawrence & Stapledon, 2008, p. 2). Most of the infrastructure assets in our sample were parts of larger portfolios stapled together by the sponsoring firm, funded by the sale of stakes to a range of investors. For instance, at the point of its acquisition of the Dulles Greenway concession in mid-2005, Macquarie Infrastructure Group had bundled together stakes in nine toll roads and bridges in the United States, Canada, the United Kingdom, France, Portugal, and Australia (MIG Annual Report, 2006).

10. This figure was the median for a range of projects sampled by Bel and Foote (2009).

11. The deals we examined demonstrated a wide range of debt–equity ratios, ranging from 19% of initial capitalization in the case of the Indiana Toll Road to 47% for the Chicago Skyway (Gray, Cusatis, & Foote, 2008). We assumed August 2006 as the effective date of the concession for the purposes of calculating interest rates and bond spreads (drawn from Federal Reserve Flow of Funds data), corresponding to the height of the market.

12. Nonetheless, there are some shortcomings associated with these models. Discounted cash flow analysis loses some predictive power when risks across the lifespan of the asset are unstable or when demand elasticities (e.g., the relationship between toll rates and traffic volumes) are unknown. Moreover, we are unable to model the effect of capital improvements (e.g., road widening or repavings) on debt, usage, or the timing of returns. This latter issue may be less of a concern given that, ceteris paribus, similar large-scale investments would not change the relative contributions of each parameter to the present value of the asset.

13. The operating expense ratio is the sum of annual operating expenses expressed as a percentage of gross income. As indicated earlier, the median operating expense ratio for a range of deals analyzed by Bel and Foote (2009) was 36%, which we use as the baseline for our analysis.

14. However, some operators won more discretion to raise rates beyond these limits. In the case of the Indiana Toll Road, for example, starting in 2010 annual toll rates are allowed to rise by either 2%, the increase in the consumer price index, or the per capita increase in nominal GDP, whichever is highest. Because the recession years 2008 and 2009 are the first years since World War II in which per capita GDP increased by less than 7% (Bureau of Economic Analysis, 2010; Johnson et al., 2007), the operator will likely increase tolls at a rate higher than the 2.54% inflation rate that has been the average for the past decade.

15. The presence of a revenue-sharing agreement could dampen revenue growth and hence the price proffered by the private concessionaire. Only the Northwest Parkway and the Chicago parking meters include revenue-sharing agreements that allow the public sector to benefit from windfall profits, and these provisions are severely restricted. We thus did not include them in the simulations.

16. Twenty years was a common swap maturity in our sample of projects (see Fitch Ratings, 2006; Macquarie Infrastructure Group, 2008).

17. Note that the Federal Reserve reports rates on 10-year interest swaps at 5.43% in August 2006, lower than the 5.68% rate on unswapped Aaa bonds. As a number of our projects began with ratings in the B range (see Fitch Ratings, 2006), we have taken a conservative stance here and assumed 5.68% as the outcome of a swap scenario.

18. For instance, MIG had “operating cash flow of $306.9 million in the 2006 financial year, but paid distributions totaling $512.9 million in relation to that year” (Lawrence & Stapledon, 2008, pp. 7–8).

19. Several projects (including the Skyway) used short-term bridge financing from bank lending consortia to fund the initial purchase, then refinanced to increase leverage 12 to 24 months later (Macquarie Infrastructure Group, 2008, p. 44).