Fudging the books

Cost-benefit studies: the record of falsification

Expertise is as saleable a commodity as gold. The old adage that 'he who pays the piper calls the tune' is as apt today as when it was first coined. Indeed, the record of industry generally is littered with examples of cover-ups in order to justify the marketing of products which are unsafe or suspected of causing harm. [1] Always, and everywhere, an 'independent' expert is on hand to tell the public what industry would like it to hear.

Nor should that surprise us. Like everybody else, the scientists and consultants who work in industry are ruled by everyday concerns - the mortgage, the need to provide for a family, the fear of failure and criticism. They know that promotion is not won by rocking the boat and that there is little profit in gaining the reputation of a trouble-maker. Is it any wonder, then, that many are tempted to cut corners and 'believe the best' for the sake of their companies and their careers?

That is not to say that all studies undertaken by a vested interest are necessarily tarred with the brush of deceit. It is to suggest, however, that a little scepticism is a healthy thing. Indeed, it is always wise to ask who funded a study before coming to a conclusion as to its objectivity. In the case of large-scale water projects, the point has been well made by none other than President Jimmy Carter. Thus, when still a candidate for the democratic primaries, he told a rally held to oppose plans to build a dam at New Melones in California:

"In many of the Corps of Engineers' dam projects around the nation, the benefit / cost ratios have been grossly distorted. Data and promises on which project approvals are sought are erroneous and outdated. False justifications of projects are attempted.

Every corps project that was initiated many years ago should be thoroughly evaluated and computations should be confirmed by the General Accounting Office (GAO). This would insure the saving of billions of dollars in taxpayers' money and hundreds of miles of irreplaceable and precious wild rivers.

A recent GAO analysis of the Sprewell Bluff dam project on the Flint River in Georgia indicated vividly the fallacies in existing Corps of Engineers analysis procedures. Construction costs were underestimated, extremely low interest rates were assumed, nearby lakes were ignored, population projections were exaggerated, environmental damage was concealed, power production estimates were based on overloaded generator ratings, no archaeological losses are included, and major recreation benefits were claimed in spite of official opposition from state and federal recreation agencies. Similar distortions exist in the New Melones project." [2]

The GAO report referred to by Carter was the result of an independent audit by Congress's watchdog agency of a cross section of seven dam projects then being undertaken in the US. [3] It contained some stinging criticisms of the major dam building agencies. Thus, as Julian McCaull, former editor of Environment, reports:

"The study shows many specific instances in which sponsoring agencies, intent on having the projects authorised, overstated expected annual monetary gains to be realised from water management activities such as flood control, irrigation, generation of electricity and outdoor recreation. At the same time, estimated costs were sharply underestimated for annual operation and maintenance of the projects, reservoir-flooding of productive land, and loss of out-door recreation sites." [4]

In one case, the US Army Corps of Engineers claimed that Missouri's Pattonsburg Lake Project would bring $1.1 million a year in agricultural benefits. In sharp contrast, the US Department of Agriculture estimated that the project would result in an annual loss of $1 million a year through the destruction of existing agricultural land and the loss of business to local farm industries. Typically, the Corps chose its own figures for the purpose of its cost-benefit analysis. It also claimed $198,700 worth of flood control benefits and $413,000 worth of water supply benefits without any documented evidence to support either claim. [5]

In another instance, the Corps claimed $65,000 worth of benefits a year in irrigation from the Lost Creek Lake Project on Oregon's Rogue River. Yet, as Julian McCaull reports, those irrigation facilities would have had to be provided "by a separate project, the proposal for which was subsequently shelved without ever being presented to Congress." [6]

In a similar vein, the Corps boosted the project's hydroelectric potential by including "not only the economic value of on-site generating capacity of 14,100 kilowatts, but also the value of 10,500 kilowatts which might be added in the future (but which had not been formally proposed to Congress) at some other site in the Rogue River Basin, and which might then be used to supplement the Lost Creek capacity."

And because no account was taken of the economic loss that would be incurred by local fishermen when Lost Creek was flooded, the 'fishing use' benefits for the project were over-estimated by some $22,500 a year. In total, the GAO estimated that the value of expected annual benefits for the seven projects it considered had been inflated by $12,607,080. [7]

It is worth putting those figures in perspective. In the case of the Pattonsburg Lake Project, "the claimed annual benefits" which could not be justified were valued at $873,000: 'unsubstantiated' claims - those benefits which had been overestimated, but which were in part justified - at $413,000 a year: while costs were found to have been under-estimated by $446,000 a year. As Julian McCaull points out, on the basis of the first item alone, "Over the 100-year period used to predict the economic future for the project, this yearly overstatement would total $87,300,000 or 52 percent of the entire estimated cost of the project." [8]

Elsewhere we find other examples of water projects where the cost-benefit figures are now open to doubt. Thus:

Reviewing the Final Environmental Statement (FES) submitted by the US Bureau of Reclamation in support of its Garrison Diversion Project, the Washington-based Institute of Ecology concluded that the scheme had "no economic justification". [9] Under the scheme, a 77-mile open channel and a string of dams are being built in order to channel the waters of the Missouri River into a massive reservoir for the purpose of irrigating some 250,000 acres.

Yet, as Onno Kremer, vice-chairman of the Manitoba Environmental Council, reports in the Canadian journal Alternatives, those 250,000 acres already "support a prosperous agriculture" and to irrigate them it will be necessary to "convert 220,000 acres of farmland and wetlands to drains, ditches, service roads and other facilities." [10] Indeed, the Institute of Ecology estimated that the project would in fact "reduce cropland by 8,148 acres, grassland by 39,172 acres and woods by 6,276 acres." Moreover, the irrigation costs were likely to be so high that they would "amount to a subsidy of $122,000 per person or $469,771 per farm."

The Institute of Ecology also pointed out that: "Energy consumption per acre for the Garrison Diversion Unit is nearly nine times the national average for irrigation agriculture"; the "degradation of water quality by the project will be far greater than indicated in the FES"; that "the destruction of family farms, the uprooting of entire families and the concentration of land ownership due to the project are inadequately discussed"; and that, "a disregard for potential effects on complex ecosystems is a basic fault of the FES.". [11]

The Institute felt obliged to conclude:

"The FES prepared by the Bureau of Reclamation for their Garrison Diversion Unit in North Dakota does not represent a comprehensive and objective examination of the complete impacts of the Garrison Diversion Unit. The FES is final for the principal supply works only and not for the total authorized project. Information included in the FES is grossly inadequate and defies objective evaluation. Further, the Bureau of Land Reclamation's analysis of its own data is often inaccurate, insufficient, or misleading. The project is described in very general and frequently tentative terms and the existing environment is considered in a cursory manner. Major adverse impacts are ignored and alternatives to the project are not considered. It appears ... that the FES was prepared as a justification for the Garrison Diversion Project rather than a detailed analysis of its environmental impacts." [11]

Authorised in 1946, the Tenn-Tom Canal was intended to link the Tennessee and Tombigbee Rivers at a cost of $316 million. George Laycock reports:

"When the calculations were completed, the cost-benefit ratio came out at 1.24 to 1. To reach that profit making conclu\ sion, the engineers had tossed in all the benefit ingredients they could justify from among those allowed by Congress, including several million dollars for recreation, fish and wildlife 'enhancement' and wage payments to those employed to work on the canal." [12]

By 1976, however, the benefit-cost ratio had fallen to 1.08 to 1 -and even that ratio, claim critics, was only achieved by underestimating costs and over-estimating benefits. Thus, in 1981, R. Jeffery Smith reported in Science:

"In 1976 ... the Corps and its economic consultant, A. T. Kearney of Chicago, found several dozen firms in the region of the waterway who said they planned to use it after it was completed. Early in 1981, the General Accounting Office contacted 17 of these firms - representing the bulk of the predicted shipments -and discovered that only about half were still interested. The GAO found that some of the predictions 'were not based on 'definite' company plans.' In other instances, the GAO said, Kearney's estimating practices 'may have been too liberal'." [13]

By 1981, the cost of the project had leapt to 1 billion dollars - and a further $960 million dollars was estimated to be necessary to straighten and widen the third leg of the canal in order to avoid bottlenecks.

"Other hidden federal project costs detected by the GAO could include as much as $31.5 million to soften the waterways impact on fish and wildlife: $360 million to deepen and widen the port of Mobile where barges will enter the Gulf: and $48 million to construct waterway-related recreational facilities. Mississippi and Alabama, which are obligated under federal water project rules to spend $170 million for highway and bridge relocations, are actually receiving $90 million of this amount from the Federal Department of Transport ... None of these costs are included in the Corp's cost-benefit calculations." [14]

Despite such criticisms, however, Congress voted in 1981 to continue with the project - at the cost of $10 million a month - on the grounds that construction work on the canal had proceeded too far to put an end to it. Stopping the project, commented one of the scheme's backers, would leave behind "the largest swamp in America".

Early estimates of the cost of the power generated by Kentucky's Devil's Jump Dam made it clear that the price per unit of electricity would be 8 to 10 times as high as that produced by the Tennessee Valley Authority, the dam's major anticipated customer. So, too, with other dams already built upstream at Lake Cumberland, flood-control benefits were seen to be negligible. Unable to offer cheap power or flood control benefits, the dam seemed likely to be slated. The Corps of Engineers, however, was not to be deterred. As George Laycock reports:

"Its solution was simply not to drain Lake Cumberland as low as usual for the winter. Automatically, this would mean that Lake Cumberland no longer had storage capacity to hold back as much water during periods of high run-off. Then these waters, for which there had been no room left in Lake Cumberland, could be accommodated in the proposed Devil's Jump Reservoir. This, at least in theory, would enable the Corps to claim that Devil's Jump would hold back 256,000 acre-feet of water for flood control in spite of the fact that it offered no such benefits in any real sense." [16]

The dam was nevertheless approved by Congress, but subsequently halted as a result of public protest.

Such blatant falsification is by no means unique to America's federal dam-building agencies. Indeed, it would appear to be typical of the industry worldwide. Thus, when the Indian economist Vijay Paranjpye reviewed the cost-benefit analyses submitted by India's hydro-electric authorities, he found them so consistently wrong that he concluded:

"The estimates had very little to do with the realistic assessment of costs ... (their) sole purpose was, and still is, to get the approval of the Planning Commission and obtain the necessary financial sanction." [17]

To support that allegation, Paranjpye cited the findings of an Indian government committee which had analysed 64 dam projects. The Committee, notes Paranjpye, found that "the percentage rise in revised costs over initially estimated costs turned out to be an average of 108 percent."

More specifically, Paranjpye took a detailed look at the figures submitted by the Karnataka Power Corporation in support of its Bedti Dam. On the basis of his own calculations, Paranjpye argued:

"Even if the most obvious and direct costs of the project are taken into account, the project is economically non-viable. The government of Karnataka will be experiencing an annual loss of over Rs 3 Crores (30 million rupees or £2 million) due to this project throughout the lifetime of the dam, unless the Karnataka Power Corporation raises the average price of power to well over 80 paise per unit, so as to derive a reasonable rate of return which could cover the social and economic cost." [18]

At that price, the power from the dam would be selling at almost five times the amount that Karnataka Power had forecast.

Paranjpye calculates that the benefit-cost ratio for the project is 0.8 to 1. He concludes: "It is reasonable to presume that if the true costs had been revealed, the Planning Commission would never have sanctioned the project." Most obviously, Karnataka Power had juggled the figures:

• by underestimating the costs of resettlement and compensation;
• by ignoring totally the cost of transmission lines;
• by omitting to take account of the revenue from the forests which would be lost to the dam;
• and by allocating derisory sums for such vital measures as soil conservation.

In the event, Karnataka Power was forced at least to recalculate the cost of resettlement. Even then, however, the figures were still underestimated. Thus, where Karnataka power upped the number of people whose land would be flooded from 1,500 to 3,706 (a percentage difference of 147 percent), Paranjpye put the true figure at 5,193. So, too, he estimated:

• the true figure for paddy land submerged was 1,429 hectares as against Karnataka Power's raised figure of 459 hectares;
• for dry land lost, 384.2 hectares as against 201.6 hectares;
• for gardens, 190 hectares as against 36.4 hectares;
• and, for the number of houses and hamlets to be submerged, 741 as against Karnataka power's figure of 50.

Indeed, Paranjpye agreed with the authorities on only one set of figures: the amount of forests which would be lost - some 10,000 hectares.

Over-estimating benefits, under-estimating costs

The above examples highlight just a few of the means used to manipulate cost-benefit ratios in favour of projects which, under any realistic appraisal, would never have a chance of even getting off the drawing board. Other techniques used to 'massage' the figures in order to give a favourable economic assessment of a project include the following:

1. The use of unrealistically low discount rates

Until 1971, the discount rate* used to evaluate US water projects was often as low as 3.125 percent. [19] That level was set in 1962 under Senate Document 97 and, even at the time, bore no relation to then interest rates, let alone their likely future value. When, in 1971, the rate was revised, the US Water Resources Council - "recognising both the objectives of subsidising water resource projects and the objectives of an efficient combination among and between federal and non-federal investment activities" - set the new rate at 7 percent rather than the 10 percent it acknowledged to be the rate of return of non-federal investments. [20] In 1973, however, the Council reduced the rate to six and seven-eighths percent.

Even so, in its 1973 cost-benefit analysis for California's New Melones Dam, the Corps of Engineers still insisted on using a discount rate of three and one-eighths percent. Dr. Thomas Parry and Professor Richard Norgaard argue.

"The reason is clear when benefits and costs are examined in the light of the effect of a six and seven-eighths or 7 percent discount rate. Using the six and seven-eighths percent figure, the interest foregone would be approximately $13,500,000 annually (as opposed to $5,700,000 at three and one-eighths percent). Adding figures for amortization, maintenance and operation, and taxes foregone, the total annual costs become about $15,400,000, or significantly more than the $13,500,000 which the corps estimated to be the annual benefit. This disparity, of course, would make the benefit-cost ratio less than one-to-one and, consequently, would make the entire project economically unfeasible." [21]

The discount rate is described by Parry and Norgaard:

"The discount rate is the rate of interest used to convert future dollar values to present dollar values, and depending on the discount rate assigned to a benefit-cost analysis, it can have a significant influence on the outcome of the analysis. To illustrate what we mean by discount rate, a capital investment which will yield you $10 one year from now really is not worth $10 to you today because you have to wait a year to get the money. If the interest rate in society is 5 percent, that $10 is only worth about $9.50 to you today since you could be receiving regular interest money at the 5 percent rate on the $10, and thus have 10.50 rather than just $10.00 at the end of the year. [B. T Parry and R. B. Norgaard, 'Wasting a River', Environment Jan / Feb 1978, Vol. 17 No. 1, p.25].

In other cases, not even a low discount rate can save a project. Idaho's Teton Dam, which impounded 17 miles of river in a massive irrigation and flood control project, is a case in point. Even using the Bureau of Reclamation's own figures and its chosen discount rate of 3.5 percent, a group of environmentalists found that the claimed benefit-cost ratio of 1.2 to 1 was in reality only 0.73 to 1 - that is just 73 cents worth of benefits for every dollar spent. Using an interest rate of 6 percent, it was found that the ratio tumbled to 0.4 to 1. [22]

It is also important to recognise that whilst unrealistically low discount rates are used for the water projects themselves, unrealistically high discount rates are employed to evaluate the ecological benefits that eventually will be destroyed. This is of particular relevance in the Third World. Indeed, as the Nairobi-based Environment Liaison Centre points out:

"The use of a discount rate over time seems unrealistic for such rehabilitative projects as timber and food production or the conservation of wildlife and scenic areas. These resources carry only a slight risk of becoming worthless or obsolescent through technological advancement, changes in the weather, or fashion in contrast to machinery, chemical factories, or dam projects. With a growing world population, the risk that soil, water and air will be valued any less in 50 years is almost zero. Their value can only go up, unless we colonise other planets. Yet, if a figure of 10 percent is adopted as a discount rate, which is a much lower figure than in most cost-benefit analyses, it would appear 'uneconomic' to plant trees for fuelwood production. If it takes £100 and 50 years to produce a crop worth £2,000 (at 1975 prices), it would not appear 'economic to do so because interest charges would amount to $10,000. With this analysis, our future generations may have to do without basic requirements of survival, such as forests, fertile soils, clean water and scenic areas. Clearly a new method of evaluating protective measures is needed." [23]

2. Over-estimating job-creation potential

One of the major stated reasons for building the James Bay Project was that it would create 125,000 jobs - at an admitted cost, incidentally of $80,000 a job. In the event, only 22,000 jobs were created during the busiest period of construction in 1977: and by 1978 there were only 12,000 - 13,000 on-site jobs available, most of which were unskilled. The Committee for the Defence of James Bay comments,

"Indirect jobs will be created as well, of course, but as the project moves towards completion workers will be laid off, boom towns will become ghost towns, and the economic benefits will evaporate. A completed dam requires very few maintenance personnel. This fact is especially important to the native population, which might find themselves permanently unemployed after working for a few years - having lost their land and their livelihood in the process." [24]

Elsewhere, we find the job-creation benefits of large scale dams similarly overestimated. The Tenn-Tom Canal is a case in point. Fred Powledge points out:

"Ordinarily, redevelopment benefits are calculated on the basis of hiring local workers who are otherwise unemployed: the assumption is that hiring these people does not reduce productivity anywhere else in the nation, and so wages paid to such workers are not included as project costs. The Corps avoided even these questionable procedures and made three insupportable assumptions: that a 'local' worker was one who had been in the area for more than one day; that 80 percent of the work force was 'local'; and that 100 percent of the local workers hired by the project were unemployed." [25]

3. Failure to account for the energy costs of building the dam

Cost-benefit analysis rarely takes account of the actual energy required to build and operate a dam. Where such 'energy accounting' has been carried out, however, the results are often surprising. Thus, Linney and Harrison point out:

"Blocking a river's course requires that an enormous amount of earth must be moved and a structure raised, either out of landfill and rock or concrete. Access roads must be built, often over long distances and steep terrain, as dam sites are usually isolated. If the dam is constructed of concrete, more petrol will be used to power the mixing plant on the site. Heavy, earth-moving equipment will be run during the 2-6 years of construction. With the exception of the use of dynamite in blasting, all these construction activities will require fossil fuel. Given the increasing cost of fossil fuel and its scarcity in developing countries due to foreign exchange constraints, the influence of fuel consumption should be evaluated. Some dams built in the past create a questionable net gain in energy (energy used in construction minus power generated) due to construction in isolated energy intensive sites, and a short life from sedimentation." [26]

More specifically, when Dr. Philip B. Williams studied the energy accounts of the New Melones Dam in California, he found that the dam would result in a net loss of energy. As Tim Palmer, author of Stanislaus: The Struggle for a River, notes:

"From the projected average year's 430 million kilowatt hours of electricity, Williams subtracted the energy costs of construction and maintenance: the loss of an existing power plant; energy for irrigation and pumping; and the energy costs of reservoir recreation. Williams' bottom-line? A net loss of 39 million kWh per year." [27]

4. Over-estimating the benefits of flood control

As we have seen in Chapter 10, the flood control benefits of large dams are frequently over-stated. No dam has yet been built to take account of the worst possible flood - the cost of doing so would be simply prohibitive. The problem, as Grant Ash, the Corps of Engineers' own officer, points out, is that dams give people a false sense of security, tempting them to build on flood plains where common sense might otherwise tell them to avoid.

But although the Corps of Engineers is clearly aware of the dangers of building on flood plains, it nonetheless chooses to include the value of the property built there among the benefits it claims for a dam. Such a statistical sleight of hand, argues Brent Blackwelder of the Washington-based Environmental Policy Institute, allows the Corps to justify almost any project:

"Many of the structures now protected by upstream dams and levees were not in place and would not have been built were it not for the dam or levee. Using the Corps' argument, almost any dam could be justified provided enough expensive developments were located in the flood-plain downstream from the dam." [28]

In many cases, the claimed flood-control benefits of a dam often only benefit a tiny minority. That point was clearly documented in A Cost Share of Water Resource Projects when Project Benefits are not Widespread. [29] Among the examples given in the report, the following stand out as particularly profligate:

A planned $17.7 million dollar flood control scheme in Hendry County, Florida, will benefit just 21 farmers. Of those 21, 13 will receive most of the benefits - and four of those are large corporations, owning 61 percent of the 34 square miles to be protected. Moreover, two landowners have expressly stated that "they are against the current project because it would overdrain their land and they would rather have their land in its present state." Ironically, the project is considered necessary in order to mitigate the flood damage caused by an earlier Corps' scheme. As the GAO explain:

"Four levees were constructed in the mid-1950s to prevent floodwaters originating on the then sparsely developed land west of the levees from flooding the agricultural lands to the east. This construction and the subsequent increased development have aggravated flood problems on the lands west of the levees. The Hendry County project, authorised in 1965, provides additional flood protection west of the levees." [30]

The South Sumter flood prevention scheme, costing some $2,742,610 is intended to reduce the damage of local floods through a 47-mile network of drains and channels. Although the original plans estimated that 300 people would benefit, it is clear that the major beneficiary has in fact been T.G. Lee Dairies, the largest landowner in the area. Indeed, according to the district conservationist, T.G. Lee purchased the land because of the benefits that would accrue from the project.

As a result of the project, the company has been able to upgrade 2,390 acres from unimproved pasture to 1,405 acres of cropland and 985 acres of pastureland, bringing annual benefits worth $25,095. Moreover, the scheme has also raised the value of the company's land by $1,195,000. Nonetheless, neither T.G. Lee or any other of the beneficiaries of the scheme have had to make any contribution to the costs of the project. [31]

5. Ignoring the costs of decommissioning

The costs of decommissioning dams are rarely taken into account when the costs and benefits for a project are totted up. Where the dam has served flood control purposes, that is particularly relevant. Indeed, as Dr. Philip Williams points out: "Decommissioning costs may be extremely large if development has occurred in the floodplain downstream and is then exposed to flood damage." [32]

6. Over-estimating the life of dams

Sedimentation is a major factor in reducing the projected, useful life of a dam, (see Chapter 16). In India, for example, dams are known to silt up between three and - in the case of the Nizamgara Dam - 17 times faster than expected. Nonetheless, it is clearly in the interest of a dam builder to estimate the longest life possible for a dam - thus spreading the initial high construction costs over a longer period, and hence reducing the proportion of those costs which are charged annually. Indeed, by extending the claimed useful life of a dam, a poor cost-benefit ratio can be transformed into a more favourable ratio.

Thus, in its 1961 analysis of the benefit-costs for the New Melones dam in California, the US Corps of Engineers estimated a 50-year lifespan. On the basis of that figure, it arrived at a benefit-cost ratio of 1.6 to 1. But Parry and Norgaard report,

"The very next year, the Corps decided to increase the period of analysis to 100 years. With other factors, such as increasing prices, the Corps derived a new and much more favourable benefit-cost ratio of 2.5 to 1. The change to the 100 year period obviously was of great advantage in promoting the New Melones Project." [33]

7. Under-estimating construction costs

The problems encountered at Sri Lanka's Kotmale Dam - part of the Mahaweli Scheme - which had to be moved from its original site owing to geological problems emphasises the dangers of under-estimating construction difficulties. Yet all too often the estimates of construction are based on what can only be described as 'wishful thinking'. Thus, in the case of Papua New Guinea's Purari project, the government claimed that a feasibility study, undertaken jointly by the Snowy Mountains Engineering Corporation and Nippon Koei, gave the project an all-clear at least as far as its economic viability was concerned.

Nothing, in fact, could have been further from the truth. Indeed, as the Purari Action Group point out, some of the assumptions used to estimate construction costs were 'highly tentative'. Thus:

"A feature of the feasibility study is its vagueness in costing a number of key elements in the dam construction. This vagueness is mainly to do with assumptions about material availability and suitability. As the area is so remote and there is so little background data on soils, geology and hydrology and negligible construction experience, more than the normal amount of care in evaluation should be expected. However, the study reveals that only a relatively small amount of testing of materials has been carried out. In some important areas it recommends that further testing and analysis should be done before the study is formally adopted - an indication that the authors themselves feel that the amount of materials testing had been inadequate." [34]

In particular, the Purari Action Group pointed out that the costs of stripping away landslide debris in order to construct a proposed saddle dam were uncertain; and that it was unknown whether suitable earthfill was available locally.

8. Failing to count land flooded as a cost

Writing in the Annals of the Oklahoma Academy of Science, biologists Drs. R. John Taylor and Constance Taylor, note:

"Paradoxically, there are about as many acres inundated in a reservoir (in Oklahoma water projects) as are protected downstream. In most cases the potential agribusiness yield of an entire floodplain, even receiving periodic flooding, and its adjacent uplands, exceeds the yield of the land protected on a flood-plain below the dam." [35]

It is rare, however, for such agricultural losses to be accounted for in cost-benefit analysis. Thus, the 1980 House Committee on Government Operations for the US Senate learnt that 27,500 acres of farmland would be lost in flooding the Tennessee Valley Authority's Columbia Dam Reservoir; in exchange, flood control would be provided for a mere 9,000 acres downstream. Yet, the project's original planning documents "made no mention of the multi-million dollar loss of farm production and farm-related business." [36]

Powledge also points out: "The builders like to forget that private land that is inundated by a federal project is, by definition, taken off the local tax rolls." [37] That cost, too, is frequently ignored. In New Zealand, a local electricity authority tried to procure planning permission for a dam omitting discussion of the land which would be flooded. The land in question was, in fact, a substantial proportion of the district's fertile farmland. In other cases - particularly in the tropics - there has been little attempt to put a value on the thousands of acres of natural forest which are frequently destroyed in order to create the reservoir for a dam.

9. Benefits of irrigation over-estimated

Irrigation is one of the major benefits claimed for large-scale water projects. In some American irrigation schemes, however, as few as 100 farms have benefitted from projects costing as much as $100,000,000. [38] That is, 1 million dollars has been spent per farm irrigated. Indeed, a 1981 GAO audit of six federal irrigation schemes found that the crops to be grown could never recoup the costs of providing irrigation at between $54 and $130 an acre foot of water. [39]

How then are such projects justified by the Corps of Engineers and other large dam-building agencies? The answer lies in part in overestimating the value of the crops that will be grown on the irrigated land. Once again, the New Melones dam provides an example. The Bureau of Reclamation estimated net irrigation benefits for the dam at $3,610,000 annually, with costs at $1,987,000. When Parry and Norgaard re-analysed the figures, however, they pointed out:

"The Bureau of Reclamation predicted irrigation benefits for speciality crops, such as grapes, deciduous fruits, oranges and vegetables - cotton - and feed grains. The Bureau estimated that the price of speciality crops would be the same during the lifetime of a 100 year project as they were in 1961. However, a 1970 study by Gerald Dean and Gordon King of the Department of Agricultural Economics, University of California, Davis, indicates that if present trends and plans continue, the speciality crop supply will increase, and, as a result of the inevitable effect of supply and demand, real grape prices will correspondingly decline 17 percent; orange prices will decline 4 \ percent; deciduous fruit prices will decline 7 percent; and potato prices will decline 43 percent between 1970 and 1980. " [40]

As a result, argue Parry and Norgaard, the irrigation benefits from the complex of dams and levees of which New Melones is a part, had been over-estimated by $12 million to $24 million (at 1974 prices). Annual irrigation benefits from New Melones alone, they claim, should be "adjusted downward by at least $2,200,000 and possibly by as much as $4,403,000". When over-estimates for likely cotton production and sales had been taken into account ($2,103,950 for the New Melones project), Parry and Norgaard found that there were "no or even negative" benefits for irrigation from the dam.

10. Over-estimating the benefits of recreation

In calculating the recreational benefits of a dam, the number of possible 'visits' to the dam and its reservoir by holiday-makers, fishermen, schoolchildren and the like are estimated and then multiplied by a dollar value per head. However, as Powledge points out:

"The arithmetic generally overlooks the facts that reservoirs tend to get built in places (such as the Tennessee Valley) where there are many other reservoirs, that the other reservoirs were justified in part on the number of projected recreational visits, and that a recreational visit to a New Reservoir means one less recreational visit to Old Reservoir B or C." [41]

Indeed, in the case of Tennessee's Columbia Dam - whose major stated benefit is recreation - there are another nine reservoirs within 50 miles of the dam site.

Moreover, it is difficult to square the claimed benefits from recreation with the more utilitarian demands that a dam itself makes upon its reservoir in order to generate hydro-electricity and fulfill its role in controlling floods. Again, Powledge makes the point:

"A controlled reservoir is not like a natural lake, with a fairly predictable waterline - its water level fluctuates, sometimes wildly, and when it is down, the recreation seeking public is treated to the spectacle of boat docks and launching ramps separated from the water by a hundred yards or so of impassable red mud."

In its report on the Columbia River Dam,

"the Government Operations Committee ... found that in the summer the reservoir would be held at 630 feet above sea level, creating an artificial lake 12,600 acres in size. After October 15th, the level would drop to 603 feet above sea level to leave more room for storage in the event of winter flooding. The Winter Lake would only be 4,300 acres in size. Between it and the normal shoreline would be more than 8,000 acres of mud-flats - hardly a modern version of Walden Pond. And, said the Congressional report, it was estimated that during one summer out of every four, the lake would not be able to make it back to its desired level because of lack of rain." [42]

11. Over-estimating the economic benefits

Clearly, if it can be argued that a dam will bring industrial growth into a region, the benefits claimed rise correspondingly. Frequently, however, those industrial spin-offs are chimeras. Thus, in the case of Ghana's Volta Dam, it was argued that the new lake would augment dramatically the business of the Volta Lake Transport Company (VLTC). Indeed, in 1961, it was estimated that the company would be carrying 150,000 tons of cargo a year by 1970: in 1964, more optimistically still, a figure of 453,000 tons was predicted.

In fact, the company's fortunes have proved disastrous: as the Volta River Authority reported in 1976, "the VLTC continued to be faced with an acute shortage of cargo capacity, lack of spare parts and maintenance facilities and inadequate cargo-handling equipment." [43]

Elsewhere, industrial benefits have been claimed which are purely speculative. Thus, the US General Accounting Office points out - in respect of the Blue River flood control project in Kansas - that many of the industries which it was claimed would be set up as a direct result of the scheme would in fact, have "been realised elsewhere, even without the project." [44] It is doubtful, then, that they should have been included in the cost-benefit analysis.

But perhaps the most spectacular example of over-optimistic claims for industrial spin-offs comes in the case of the Tennessee Valley's Tellico Dam. A major justification for building, the dam was that it would power a new industrial city to be built as a company town for Boeing Aerospace. In the event - when the dam was three-quarters complete - Boeing withdrew from the scheme. By then the dam had already cost over $100,000,000. Indeed, so small were the other benefits claimed for the dam that when a government committee reviewed the economic justification, it concluded:

"The interesting phenomenon is that here is a project that is 95 percent complete, and if one takes just the cost of finishing it against the total benefits ... it doesn't pay, which says something about the original design." [45]

Elsewhere we find that dams have been built even though it was clear that the revenue they produced could not possibly recoup their costs. For example, the modern era of environmental awareness in New Zealand is usually seen as having begun with the controversy over the Manipouri Power Scheme. After years of official secrecy, the price paid by the aluminium smelting consortium for the electricity from this wholly publicly-funded project was revealed to be 0.2 c/kWh, and a later government raised the price about sixfold. Despite continuing secrecy on the costs of the project, it is doubtful whether even the new price recovers costs. No cost-benefit analysis was ever published. [46]

So, too, we find that the Indonesian government agreed to supply Japan's Asahan Aluminium Company with 80 percent of the hydro-electricity for the Asahan Dam. The electricity was to be supplied at cost. Moreover, as the Purari Action Group reports:

"Even after the power plants are transferred to Indonesia in 30 years, the electricity is still to be provided for the company at cost. The Japanese company will determine for itself the cost of the power." [47]

12. Under-estimating the environment costs of irrigation

More generally, the disastrous environmental effects of perennial irrigation are rarely taken into account in cost-benefit analyses. On the contrary, where dams provide irrigation, that irrigation is almost always counted as a 'benefit'. Yet, it is rare indeed for perennial irrigation schemes to avoid the twin evils of salinisation and waterlogging - the more so when there is such clear reluctance on the part of the authorities to earmark any funds for drainage and other remedial measures.

Inevitably, the ecological costs come home to roost. In the US, the cost of building a master drain to the sea in order to remove the accumulated salts from the San Joaquin Valley has been estimated at a minimum of $1.2 billion. [48] In order to remove the salt from the Colorado River before it enters Mexico, the US has been forced to commission a desalinisation plant at Yuma. The expected cost is over $1 billion.

Yet, despite such expenditure, the problem of salinisation is far from solved. Indeed, according to Arthur Pillsbury, only the construction of the NAWAPA scheme can prevent South California from becoming a salt-encrusted desert. That scheme is estimated to cost $2 trillion - at 1973 prices. [49] Had those costs of establishing perennial irrigation been taken into account, would the dams that provided the irrigation water have seemed so attractive? And could they ever have been justified on economic, let alone ecological, grounds? In almost every case, the answer is undoubtedly 'No'.

Conclusion

It thus seems that those who stand to gain politically and financially from the building of a large dam are willing to go to inordinate lengths to ensure that it will be built. Among other things, they are willing purposefully to mislead those who must be persuaded of the dam's desirability and viability before the go-ahead to build it will actually be given. This they do by grossly exaggerating the dam's likely benefits and seriously underestimating its probable costs - in particular its social and ecological costs which, as we have seen, are often totally ignored.

The power, prestige and financial resources of the politicians, bureaucrats and industrialists involved in dam projects greatly facilitates that deceit - as does the credulity and apathy of the public. Moreover, unlike the authorities, those who oppose dams - often local tribal or peasant leaders, obscure academics or youthful environmentalists - have meagre financial resources and little credibility.

To add to their difficulties, they must also confront the entrenched belief that large-scale water development schemes are an essential part of the process of economic development - a process which we have been taught to see as the only means of combating poverty and malnutrition, and of assuring health, longevity and prosperity for all. To challenge dams is thus to challenge a fundamental credo of our civilisation.

But enough of criticism. As we have already mentioned, the problems that dog modern irrigation schemes have not, historically, affected the irrigation systems of traditional societies. It is clear, therefore, that we have much to learn from the traditional irrigation societies of both the past and the present. In the next section we shall see how, in five such societies, all the problems associated with modern irrigation agriculture have been avoided and why.