Skip to main content

Dam it all

Dams and levees are the usual answer to flooding, but they're not always the right answer. So why do we keep building them?

September 1, 2001


Douglas Clement Senior Writer

The long-established human response to floods has not been to flee them, but to stay put and build walls to keep river water out. The Babylonians leveed the Euphrates. Ancient Rome did it to the Po. In the United States, the Army Corps of Engineers has lined the Mississippi's banks with concrete, with levees longer than the Great Wall of China, and multiple dams and reservoirs. Parts of the Missouri have been channeled so tightly by corps projects that environmentalists consider it a straitjacketed travesty of nature.

The pouring of taxpayer-financed concrete has slowed in recent years as the weaknesses of structural answers have become more evident. Not only can dams and levees fail—1,083 levees failed during the 1993 flood—but they can make floods worse by raising water levels and velocity, increasing pressure on downstream flood control structures. And to the degree that their presumed protection induces people and businesses to move onto the floodplain, structural controls actually create the potential for higher damages when they do fail.

In addition, since a flood control structure in one location will affect river flow elsewhere, communities often react to a neighboring town's new dam or levee by building one of their own, in what becomes a cascading escalation of river armament.

"We are creating more problems when we allow communities to channelize and to avoid their flood problems by putting levees up," observed Rep. DuWayne Johnsrud of Wisconsin's 96th District, which includes many river communities along the Mississippi. "[Because] then [the flood] blows out somewhere else that doesn't have a levee. ... We all live downstream from somebody."

Finally, the costs of maintaining structural controls are substantial, and that maintenance tends to be neglected if floods aren't an imminent threat. "Once you build them you've got to maintain them, and that's the problem," said Gale Selken, a South Dakota Department of Environmental and Natural Resources dam engineer. "Nobody wants to put money into it when it's dry. They only [complain] when the water's coming."

Despite all these problems, structural solutions are still popular in the Ninth District:

  • In July, Grand Forks-East Grand Forks (N.D. - Minn.) began a $350 million construction project to control flooding with earthen levees.

  • In late April, Fargo, N.D., engineers proposed a $130 million flood control proposal, most of it devoted to levees and flood walls.

  • Downstream city dwellers on the Red River hope that upstream rural citizens will endorse a major new dam near Lake Traverse, S.D.

  • The contentious debate over a solution to flooding in Watertown, S.D., involves a choice between one large dam or hundreds of small ones, but the proposed solutions are still structural.

  • Structural flood controls have been used in Rochester, Minn., and they're the leading contenders for solving the rising tide of Devils Lake in North Dakota.

Critics of such solutions call dam advocates "water buffaloes" who throw their weight around to force their projects through, and they say the corps never met a dam it didn't like. Structure supporters counter that environmentalists are so extreme they would just as soon invert the slogan "Damn the torpedoes."

While this debate is inevitably heated, both sides have valid points. Structural controls are effective, up to a point, in preventing rivers from inundating protected lands. Most of the levees built and maintained by the corps did not fail during the 1993 flood, and they did protect millions of people and billions of dollars of property.

By the same token, those same dams and levees also induced floodplain development that otherwise would not have been at risk. And they increased river speed and height, which placed additional force on dams and levees protecting land and property further downstream. According to some analyses, in fact, it was only because of unplanned upstream levee failures during the 1993 flood that St. Louis was not inundated.

Since dams can both save lives and wreak havoc, deciding whether to build one demands careful scrutiny. People who stand in the middle of the dam debate say that there is a definite place for structural flood control, but only if its wisdom can be clearly demonstrated in each given application. Dams are neither a panacea nor an unmitigated evil. To build or not to build is not a simple choice.

The labyrinth of cost-benefit

Cost-benefit analysis is the standard approach used to determine the advisability of a structural project. Analysts compare the costs of a project to the benefits that could be reaped. But for economists, the straightforward idea is far simpler in concept than execution.

Since the benefits are largely hypothetical—what will the dam save if and when there is a flood?—a clear market value isn't readily available. Measuring the benefits of a proposed dam gets complicated when economists debate whether to use property damages avoided (PDA—the usual method), contingent valuation (asking affected people how much the dam is worth to them) or hedonic price (extrapolating from property prices of comparable land already protected from floods). The PDA approach is straightforward, but economists are not of like mind as to its superiority.

Another area of cost-benefit controversy is deciding on the right discount rate—essentially an inverted interest rate—to use in calculating the current value of future benefits and costs. To estimate a project's value, an analyst needs to add up its costs and benefits over a chosen time period, 50 years perhaps. Project costs and benefits are incurred at different points in time—costs are often up-front and benefits accrue later—but decisions to undertake the project must be made in the present. So analysts need to think about the present value of each cost and benefit. Given that a dollar spent on a project now could have been otherwise invested, determining its value in the present requires an inverse interest rate, a discount rate, which says, for example, that $100 worth of benefit one year down the road is worth $90.91 in the present, using a 10 percent discount rate.

But the choice of a discount rate is somewhat arbitrary: Is the proper discount rate high, low or does it vary over time? "We write books on the subject of the proper discount rate," observed University of Minnesota economist Steven Taff. "The state [of Minnesota] uses 8 percent. The Feds have generally used a much lower number, 3 percent, 4 percent." The choice makes a huge difference to the outcome of an analysis, with low rates tending to boost the estimated value of long-run benefits, and critics have accused the corps of using rates that exaggerate project benefits.

A more contentious issue in cost-benefit analysis is "scoping"—that is, who and what is included in the analysis? "Who counts? Where do they count? How long should we consider? What costs and benefits do we consider?" said Taff. "Drawing a circle around what's going to be inside and outside the decision framework is the necessary first step." And as Taff points out, these issues are not strictly economic. "These are intensely political decisions ... but we can deal with that in the cost-benefit framework. We can say, 'if you think that this should count, here's what the costs are'... or, 'these are the people who bear this cost.'"

The trade-offs

The task of an economist, therefore, is to clearly spell out the costs and benefits of different alternatives and clarify the tradeoffs implicit in any decision on flood control measures.

"When you're building these flood control structures, you're basically trading off protection [from] low-level floods against much higher damage when higher level floods come along," said Lester Lave, professor of economics and finance at Carnegie Mellon University in Pittsburgh. The structures allow residents to derive value from the protected land during years of low-level flooding, but when a large flood occurs "it's going to overtop everything." Since people have built in the floodplain in the meantime, "the amount of damage is greater than if you had not built the flood control structure."

From an economist's standpoint, this could be the most efficient use of resources, said Lave. "In present discounted value," he said, it could well be "more valuable to get the high-value use out of the land and then have the [homes and buildings] destroyed than to keep the land in very low-level use."

The same argument applies to rural vs. urban flooding. "If we're narrow-minded economists and put on our efficiency hat," said Lave, "we say, 'Great, that's the efficient solution: Flood all this cropland and lose pennies rather than flooding the urban area and lose dollars.'" But of course, there are other considerations. "From an efficiency standpoint, I think that's right," Lave said. "But when you look at it from an equity standpoint, the farms and these other areas say, 'What the hell are you guys doing? We're not benefiting from that city center. Why are you causing all these problems for us?'"

It is usually in examining such tradeoffs that the seemingly technical cost-benefit analysis breaks down. An economically rational choice will be deemed socially unwise or unpalatable to some decision-makers. What is economically efficient might not be seen as politically expedient.

Indeed, some benefit-cost analyses have been skewed to arrive at the desired outcome, be it support or defeat of any given project. Lave chaired a 2001 National Academy of Sciences review of a corps cost-benefit analysis of its proposed Upper Mississippi lock-and-dam expansion project. Both the NAS assessment and an internal Army review found the corps analysis was highly flawed. The U.S. Office of Special Counsel determined that it had been "manipulated ... in order to justify the expensive extension."

Congressional critics of the corps have called for independent review of all major corps cost-benefit analyses in the future and the exclusion of private benefits from these analyses. They're also pushing for the use of a 1.5 ratio of benefits to costs as the standard for proceeding with a project (rather than 1.0), arguing that taxpayers should expect to receive better than break-even benefits from their dollars spent.

The corps says it is already subject to sufficient scrutiny and not inherently biased toward massive concrete flood control. And local corps staff point to their Red River Basin project, which would combine wetlands water storage with structural control, as an example of their push toward nonstructural solutions.

"Historically, corpswide, there may be some justification" for criticism of the corps for overemphasis on structure, said David Loss, an engineer in the corps' St. Paul office. "But I think the direction the corps is taking right now in fact is very much emphasizing looking at nonstructural alternatives."

Whatever the pros and cons of dams and levees, and regardless of corps biases, its clout has been only slightly blunted by recent attacks on its reputation. When the Bush administration tried to cut the corps budget by 14 percent this year, its congressional supporters promptly replaced the missing funds.

While the distribution of funds to structural flood control will remain a largely political decision, economists point out that shifting the costs and responsibilities for dam-building to local entities would tend to prevent misallocation.

"You have to be very careful with whether the benefit-cost analyses [of structural flood controls] are being done honestly," noted Lave. "And I think that one way of certainly doing that is to have more local share."

That is, if local communities are required to pay a greater portion of flood control costs, they'll consider more carefully what truly is needed, what will actually protect local assets over the long term and whether other alternatives might prove to be a better investment. Academic studies have indeed found that the size of flood control projects chosen by legislators varies inversely with the local cost share. But whether policymakers will be willing to curb federal subsidization of structural flood control remains an open question.


Douglas Clement
Senior Writer

Douglas Clement was a managing editor at the Minneapolis Fed, where he wrote about research conducted by economists and other scholars associated with the Minneapolis Fed and interviewed prominent economists.