White Elephants: Why Some Large Projects Fail

Article excerpt

The much heralded, privately financed Dulles Greenway in Virginia is in trouble. Only three months after opening the $340-million toll-road project, operators have laid off one-third of the toll collectors and are worried about meeting their scheduled loan repayments next fall. What went wrong? The Greenway is a classic case of a bungled large infrastructure project. Hatched during the high-growth decade of the 1980s, the project suffers from excessively optimistic projections of demand, high preconstruction costs, and a failure to take into account changing feasibility conditions..The area's economy had already begun to slow down by 1988 and was in full "meltdown" by 1991.

Since the end of World War II, when large-scale revenue bond financing took off, the United States landscape has been littered with a number of white elephants that share similar traits. These defaults are rare relative to the size of the market, but when problems happen, they happen big. Large, high-profile defaults in the toll-road sector in the 1950s and 1960s have contributed to the common perception that revenue bonds are more risky than general obligation bonds.

Early White Elephants

Early toll-road projects suffered from a number of ills. Coordination of planning was spotty - in some cases state and local governments built competing freeways near the toll roads and bridges. In other cases, the convenience and time savings offered were marginal, and tolls were too high compared with other available routes. Ultimately, some roads suffered from simple lack of demand - they did not take people where they wanted to go.

In 1958, the West Virginia Turnpike commission defaulted on $133 million bonds. In 1963, the Calumet Skyway followed suit, de, faulting on $101 million. (Thirty years later, the skyway has accomplished an investment grade $106 million refinancing.} The Chesapeake Bay Bridge and Tunnel Commission, although able to meet debt service on Series A and B bonds, defaulted on $100 million subordinated Series C in 1970. Numerous smaller-sized defaults occurred during this time, including two bridges in West Virginia, one bridge in Nebraska, and two in Illinois. Bonds sold by East St. Louis for the Martin Luther King Bridge defaulted in 1974 and were finally settled in 1986.

Traits and Trends

Despite the historical record of defaults, project planners persist in repeating similar mistakes. Robert H. Muller, managing director at J.P. Morgan Securities, outlined some of the problems befalling toll-road projects of the 1980s. He analyzed five major projects during the past decade and found the following.

* Two forecasts were accurate. Actual revenues for the Dallas, Texas, North Tollway and the Illinois North/South Tollway moderately exceeded projections five years after project completion.

* Revenue was 20 percent below forecast in one case - the Sam Houston toll road, which is part of the Harris County, Texas, toll-road system.

* In two cases, revenues were more than 50 percent below forecast: the Hardy Toll Road in Harris County, Texas, and the Sawgrass Toll Road in Florida.

* Economic projections were much too optimistic.

* Fundamental demand was lacking.

* Fees were higher than alternative roads.

Unlike earlier toll roads that were built for intercity transport, recent toll roads (like the Dulles Greenway) are being built in the outer urban fringe - to handle intracity traffic. Muller also mentioned a number of other trends. 1) Today's projects are far more expensive than those of a decade ago and commonly include increasing debt service requirements. 2) To support increasing debt, feasibility studies either assume steadily growing demand or build in rate hikes. Earlier projects did not assume rate escalation. 3) Not unlike some of the white elephants of the past, projected revenues far outpace what might be considered a normal growth pattern. …