Review of comments on Restoring the Lower Snake River: Saving Snake River Salmon and Saving Money by the Oregon Natural Resources Council

Executive Summary

The Northwest Power Planning Council asked the Independent Economic Analysis Board (IEAB) to review the economic assumptions and methodologies used in Restoring the Lower Snake River: Saving Snake River Salmon and Saving Money, produced by the Oregon Natural Resources Council with economic analysis principally by Philip S. Lansing. The IEAB questions a number of crucial assumptions in the ONRC study, which leads to questions about the reliability and accuracy of the conclusions. We do feel that, in publishing this evaluation, the ONRC has provided the service of putting many of the basic economic issues relating to breaching the Lower Snake River dams in front of the public for consideration.

The report, by the Oregon Natural Resources Council, estimated that the "net economic benefits" attributable to the hydropower, navigation and irrigation aspects of the current lower Snake dams amounts to a negative $231.9 million. The author estimates that a restored river would yield "net economic benefits" of only negative $145.2 million. Thus, the study concludes that society would be better off by $86.7 million if the dams were breached.

The ONRC assumed that $194.4 million of current fish and wildlife restoration expenditures could be avoided if the dams were breached. This figure includes $164 million of hydropower currently not generated because of fish restrictions. The IEAB agrees that portions of the $194.4 million could probably be avoided. Some fish and wildlife spending would undoubtedly continue. Power would still not be generated at the breached lower Snake dams, and continued fish restrictions would probably continue to affect generation at other dams.

The ONRC assumed that the lost hydropower attributable to a restored river can be properly measured by lost firm power, and that this loss can be replaced from alternative sources at a cost of 1.6 cents per kWh. The ONRC estimate for the total cost of replacement power is $115.6 million. More realistic estimates suggest this figure should be more in the 2-2.5 cents per kWh range or about $160 million. If the value of lost non-firm energy and a more realistic value of firm power are combined, these two adjustments could easily double the ONRC cost of replacement power.

While the IEAB also questioned the magnitude and interpretation of other figures in the ONRC report, the two instances noted above are so crucial to their analysis that changes could easily negate or reverse the ONRC conclusion that restoring the lower Snake River would result in a benefit to society of $86.7 million.

The Lower Snake River Juvenile Salmon Migration Feasibility Study (DREW) is currently addressing a number of alternatives for the lower Snake, including the natural river option addressed by ONRC. The approach being used by DREW will be roughly compatible with most elements of the ONRC benefit-cost methodology, but will be broader and more inclusive in scope. Given the resources that the Corps is devoting to the process, and the open participation of a wide range of agencies and stakeholders in the process, the IEAB anticipates that DREW will provide some definitive answers to the issues raised by ONRC.

Full report

The Northwest Power Planning Council asked the Independent Economic Analysis Board (IEAB) to review the economic assumptions and methodologies used in Restoring the Lower Snake River: Saving Snake River Salmon and Saving Money, produced by the Oregon Natural Resources Council with economic analysis principally by Philip S. Lansing. The Council asked that the assumptions and methodologies of the ONRC report be compared to those proposed for the economic analysis by the Lower Snake River Juvenile Salmon Migration Feasibility Study (a.k.a. the Drawdown Regional Economic Workgroup, or "DREW"). In its review, the IEAB was asked to highlight key assumptions in both reports and key differences between the two. The focus must be on assumptions and methodologies since the latter study is not yet complete.

An Overview of Methods Used in the ONRC Report

The goals of the Oregon Natural Resources Council report are addressed in the report's executive summary:

"This paper addresses this question: Would restoring the Lower Snake River to free-flowing conditions cost or save money?"

"This question is a bit more complex than simply adding up annual spending on keeping the dams in place and comparing it with the costs of restoring the river. What is required is a comparison of the actual net economic benefit provided by the dams at present with the net benefit that would result from river restoration."

"In other words, Which is greater: Net economic benefit of Lower Snake dams and reservoirs or net economic benefit of restored Lower Snake River?"

"Net economic benefit is a technical term meaning economic return to society after all costs are accounted. Net benefit is typically a positive amount, but it can be negative when hidden costs are included in the reckoning. For the Snake River dams, the benefit is negative."

"Net economic benefit is very different from economic impact. An impact study might focus, for example, on the impact of a proposed course of action on a local community. There will be many different impacts in different areas if the Lower Snake River is restored, some positive and some negative. Our benefits analysis does not address local impacts. Instead, it takes a broader view and focuses on changes in overall economic wealth." (ONRC, pages 2 and 3)

The IEAB agrees that benefits and costs to society are the appropriate focus of economic evaluation. We also agree that an overemphasis on short-term and regional "impacts", such as the directly affected jobs or incomes of stakeholder groups, can distort the true benefits and costs of the drawdown decision or any other strategy for salmon recovery. The DREW analysis will also address the benefits and costs of drawdown to a natural river as one of several alternative recovery actions, using the benefits to society as a whole as the decision focus.

The ONRC report used an unconventional approach of estimating the "net economic benefits" attributable to three uses (hydropower, barge transport and irrigation) of the lower Snake system with the dams, and then comparing them to the "net economic benefits" in the system without the dams. The difference between these amounts is one indicator of the benefits to society from dam removal or breaching.

The ONRC's approach differs in economic terminology from what will be used by DREW. DREW's more conventional economic approach will focus on particular proposed actions (in this case, restoring the lower Snake to free-flowing condition) and then look only at the marginal changes in benefits and costs attributable to this proposed action. Thus DREW will focus on the changes in benefits and costs between a base case (the present situation) and several alternative recovery actions, one of which is the dam breaching scenario looked at by ONRC.

Conceptually, the marginal change in benefits and costs attributable to a policy action as estimated by DREW can be directly comparable to the difference between the "net economic benefits" with the policy and without, as estimated by ONRC.

The ONRC report and the DREW process also have in common the recognition of a wide range of kinds of costs. They both recognize that some costs are direct, such as the costs of moving earth to breach a dam. Some costs could be avoided if the dams are breached, such as the costs of operating and maintaining the navigation system. Some costs are opportunity costs, the cost of something one gives up, for example the lost hydropower generation if the dams are breached. However, as discussed below, the IEAB questions the way some costs were classified in the ONRC report, and, more specifically, the magnitude of some of the ONRC cost estimates.

The ONRC presentation is also very complicated -- with numbers, assumptions, and calculations spread across text, appendices, sidebars and footnotes. This makes the analysis not only initially difficult to evaluate, but also difficult to follow. As complicated as it is, the ONRC approach still gives an incomplete picture of all of the benefits and costs associated with the system - as the author notes, recreation effects and the value of the anadromous fish themselves are not counted. In contrast, DREW will attempt to measure and include the impacts of the alternatives on recreation and fish value in their analysis.

In publishing this evaluation, the ONRC has provided the service of putting many of the basic economic issues relating to breaching the Lower Snake River dams in front of the public for consideration. The IEAB welcomes the ONRC's approach of using documented assumptions, rather than unquantified and unqualified assertions. The IEAB agrees that the process of trying to agree on the magnitude of economic effects and their proper interpretation should help us better evaluate the range of fish recovery alternatives. Untangling the web of assumptions and estimates has also given the IEAB a headstart in its evaluation of the DREW analysis. An Evaluation of the ONRC Assumptions and Findings

The ONRC report summarizes its findings as follows:

"The Lower Snake dams and reservoirs require the Bonneville Power Administration to spend $194.4 million every year on salmon restoration."

"Taxpayers and electric ratepayers subsidize electric power production, river transportation and irrigation from the Lower Snake dams and reservoirs. With all costs accounted, these three Lower Snake dam "benefits" actually produce a net benefit loss to the economy of $114 million every year."

"Electric power from the Lower Snake dams is not competitive. It costs 2.44 cents per kilowatt-hour. If we restore the Lower Snake River and purchase power elsewhere, we could provide energy for 1.87 cents per kilowatt-hour."

"River transportation on the Lower Snake is expensive and heavily subsidized. Although river shippers pay only $1.23 per ton to go from Lewiston, Idaho to Kennewick, Washington, taxpayers and electric ratepayers pay an additional $12.66. The total cost to ship one ton of goods on the Lower Snake is $13.89. In comparison, rail costs only $1.26."

"Thirteen agribusinesses pump water from the Ice Harbor reservoir. Together, these farms earn a net $1.9 million per year. But taxpayers and electric ratepayers subsidize these farms with $11.2 million. If the farms paid their full costs, they would lose $9.3 million every year. It would be cheaper to buy these farms outright and end their production altogether."

The conclusion of the ONRC report is that the region is faced with negative "net economic benefits" from the river system irrespective of whether it is restored or stays in its current state. They estimate that annual "net economic benefit" with the dams is a negative $236.2 million, while annual "net economic benefit" with a restored river is only negative $149.5 million. Thus they conclude that the difference of $86.7 million is a measure of how much society would be better off if the dams are breached and the river restored.

The ONRC conclusion cannot be compared with results from DREW, since the latter study is not yet complete. Draft pieces of the DREW study are just now being made available for review, and the complete study will not appear until 1999. Nevertheless, as requested by the Power Council, in the discussion below the IEAB offers some evaluation of the ONRC assumptions and results. (While we use the DREW procedure for comparison, our review should also not be interpreted as a premature endorsement of the approach or conclusions of the forthcoming DREW report.) In trying to understand and evaluate the results arrived at by the ONRC the IEAB found it useful to begin by placing the ONRC's own numbers in the familiar benefit-cost analytical framework that will be followed in the DREW analysis. The results of this exercise are presented as Table 1. The table is intended to focus directly on the basic question posed by ONRC - whether breaching dams would save money. Again, Table 1 does not evaluate the ONRC estimates, but simply restates them in a form more familiar to economists working on these issues.

Table 1: ONRC Evaluation of the Economic Impacts of Restoring the Lower Snake River as Rearranged by IEAB Using Benefit-Cost Categories.

  Direct Costs
(Million $)
Avoided Costs
(Million $)
BENEFITS (No Change from Restoring River) -- --
Implementation Costs    
Dam Removal 25.6  
Current Dam O&M   33.6
Current F&W Restoration Costs   194.4 [1]
Opportunity Costs    
Replacement Power (Value of Lost Power) 115.6 [2]  
Increased Private Cost 0.1  
Lost Power from Locks   0.8 [4]
Port Property Tax Subsidy   1.3 [5]
Net Crop Value 1.9 [3]  
Purchase of Irrigated Farms 2.0 [3]  
Lost Power Generation from Water Use   0.5 [6]
Pumping Subsidy from Electric Ratepayers   0.5 [7]
Crop Subsidies   0.8 [8]
TOTALS 145.2 231.9
NET COST (Direct Cost Minus Avoided Cost)   -86.7

Because Table 1 uses the same numbers presented in the ONRC report, it also adds up to the ONRC bottom line, a benefit (a reduction in negative benefits) of $86.7 million from dam breaching. The numbers in brackets refer to our comments, provided below, on the validity of many of the entries in the table.

The first line of the table reflects the ONRC assumption that there would be no change in gross benefits from restoring the river. That is, the ONRC assumes that the same amount of electricity would still be available for consumption (although now from alternative sources), that irrigated farmers would be made whole (by being bought out), and that grain would still be transported to market (although by alternative modes). As noted above recreational benefits and cost have not been included in the analysis. Neither has the value of increased salmon been evaluated as a benefit, although that increased value is clearly the intended policy objective. Such an exclusion has the result of underestimating the potential benefits of breaching.

The two cost columns distinguish between "direct costs" and "avoided costs." "Direct costs" are payments necessary to accomplish the restoration alternative (such as the cost of breaching the dams or purchasing irrigated farms). "Avoided costs" represent savings that result from restoring the river. Such savings result because some payments (such as current dam operation and maintenance (O&M) , or if one accepts the ONRC view, the majority of salmon recovery expenses) that would otherwise be necessary to maintain the current state of the river, are no longer necessary if the dams are breached.

The table also distinguishes between "implementation costs," which are actual payments to achieve restoration of the river (e.g. the cost of earthmoving to breach dams) versus "opportunity costs," which are benefits given up as the result of restoration (e.g. less hydropower, lower harvest of some crops.)

Comments on Items in Table 1

The IEAB found Table 1 useful as an aid in understanding the relationship among the various numbers presented in the ONRC report. With this as a start, concerns were identified about the way the particular numbers were used in the ONRC analysis, and sometimes about the magnitude of the number used. Our comments in the section that follows are keyed to the numbers in brackets in Table 1. These comments explain in more detail our substantial concerns about particular assumptions used by the authors of the ONRC study.

[1] Current Expenditures for Fish & Wildlife Recovery

This number is the single most important number in the ONRC analysis. The rest of our comments are far less significant in magnitude of effect compared to this. The authors themselves note (on page 11, in a footnote on page 13, and in the appendices on page 28) that their results depend crucially on the correct magnitude and interpretation of this number. They use this $194.4 million number, together with $33.6 million of current dam O&M, as the measure of current annual cost of the dams, which they then allocate to the various uses (hydropower, navigation and irrigation). The ONRC study assumes that if the dams are removed, then these costs will go away.

However, it does not appear reasonable to assume that all of the $194.4 million fish and wildlife restoration costs will be avoided with removal of the four lower Snake dams. Removing these four dams will not really "restore the river to free-flowing condition", because Idaho Power Company dams, Dworshak Dam, and a large number of irrigation water users will continue to store and divert water. The ONRC report quotes a study that concludes that if the Lower Snake River is restored, there is an 80 to 100 percent probability that spring and summer Chinook salmon populations can recover to the levels of the 1960's within 24 years. However, current indications from the PATH (Plan for Analyzing and Testing Hypotheses) process are that even under the best of conditions, recovery will be neither fast nor assured. If this is the case, It is likely that there will be continued spending on fish recovery, and it is highly unlikely that spending on fish monitoring, research and habitat restoration will stop even if fish numbers recover.

Of the $194.4 million cost attributed to current fish recovery efforts in the ONRC study, $164 million consists of the value of lost hydropower generation due to restrictions on water flows, spill requirements, temperature controls, and timing of releases. As the fish-related flow regime reduces the amount and value of hydropower generation, BPA or regional utilities purchase power from other, more expensive sources. The ONRC uses BPA's estimated cost of $164 million for hydropower costs associated with "fish restrictions" on the system.

ONRC assumes that this hydropower cost will disappear when the dams are breached. While ONRC does recognize that its results crucially depend on this assumption, that dependence is not noted among the results highlighted in the report. We think that this treatment of hydropower cost is a major flaw in the ONRC report. The hydropower costs of removing the four lower Snake River dams depend upon: (1) the degree to which that removal would alleviate needs for water releases to assist fish in the Snake and Columbia rivers, thus allowing more hydropower generation to occur on Columbia River and upper Snake River dams; and (2) the amount of power generation sacrificed on the lower Snake. ONRC must be assuming that the loss of hydropower at the breached Snake River dams would be fully compensated by an increase in power production elsewhere, and that hydropower production on the Columbia and upper Snake would increase markedly as spring seasonal flow requirements are diminished. But ONRC provides absolutely no information on the relative magnitudes of these two factors to support the assumption that $164 million in hydropower costs associated with "fish constraints" will be saved (i.e., disappear) when the four dams are removed. To the extent that breaching does not recover all $164 million of these costs, because generation is impossible at breached dams and some "fish constraints" continue on the remaining dams, this difference would come directly off of ONRC's $86.7 million bottom line.

We would need more concrete information to fully assess the plausibility of ONRC's assumption that the $164 million cost of "fish constraints" will be saved when the Snake River dams are removed. One source of such information is the 1994 comprehensive analysis performed by the Bonneville Power Administration, US Army Corps of Engineers, and US Bureau of Reclamation in their Columbia River System Operation Review (SOR). In addition, the hydropower modeling work currently being done for the Lower Snake River Juvenile Salmon Migration Feasibility Study will provide updated estimates of the effect of dam breaching on hydropower generation, and will attempt to track which current fish and wildlife expenditures are avoidable. Thus, the ONRC assumptions should, hopefully, be subject to testing through the DREW process.

It should also be noted that not all of the estimated $33.6 million of O&M expenditures on dams would be eliminated. Substantial maintenance expenditures for mothballing of the dams in a breaching scenario would be necessary, although unquantified at this time.

[2] Replacement Value of Lost Energy

 The assumptions on which this number is based comprise the second largest contributor to the ONRC's bottom line. The study asserts that there is currently a surplus of power in the Pacific Northwest (page 18), and that the power currently available from the Snake River dams (7,227,000 MWH) could be replaced from alternative generation sources at a cost of 1.6 cents per kWh. Since the ONRC presents no alternative rate scenarios for the future, it presumably is projecting that the 1.6 cent replacement cost rate is applicable in perpetuity.

There are two problems with these assertions. First, the most recent projections by the Bonneville Power Administration indicate that the Northwest currently relies on imports of power under so-called "critical water" conditions (1997 Pacific Northwest Loads and Resources Study, 1998). Second, the 1.6 cent projected cost is certainly at the very low end of conventional wisdom. Power Council modeling currently projects rates from 1.5 to 3.5 cents per kWh, depending on a range of possible future conditions, with the low- to mid-2 cents range considered most likely (Northwest Power Planning Council. "Analysis of the Bonneville Power Administration's Potential Future Cost and Revenues." Staff Analysis Report 98-11. June 1998, pages 23, 24). The DREW study will apparently use Power Council projections to quantify alternative generation costs. An increase in the cost of replacement power from $0.16/kWh to $0.25/kWh would increase the cost of replacement power from $115 million to over $180 million, erasing $65 million of the ONRC's calculated $86.7 million increase in economic welfare. This suggests that the ONRC conclusions are not robust in the face of uncertainty regarding future power costs.

If the dams are breached, a likely alternative generation source in the region is fossil fuel (e.g., natural gas) burning, which creates environmental concerns of its own. To make a balanced with-and-without comparison, one should add to the opportunity costs an estimate of the social costs of the increased carbon dioxide and particulates introduced into the atmosphere by increased combustion of natural gas. The Corps EIS is expected to note these carbon dioxide and particulate effects, but apparently will not attach an economic value to these effects.

Today's generation facilities also have economic and social value arising from their contribution to secondary energy, capacity, and power system reliability. (Information from NWPPC staff suggests that nonfirm energy generation may add almost 50% to the lost firm generation figures used by ONRC to get the $115.6 million figure.) If the actual value of lost power exceeds the $115.6 million replacement cost estimated by ONRC, this difference would reduce, and possibly reverse, ONRC's estimated net societal benefits of dam breaching.

[3] Lost Net Crop Value and the Purchase Price of the Farms

Including both the lost net crop value and the purchase price of the farms double counts this loss. Either measure taken alone could be an estimate of what it would take to compensate the farmers for lost access to irrigation water. If we as a society decided to set up a program to purchase land currently in irrigated agriculture, the price we would be willing to pay should be no higher than the net value of the crops that could be grown on that land. The ONRC study used average crop values for counties adjacent to the Ice Harbor pool to estimate the lost net crop value, and estimated that the farmland could be purchased for $1500 per acre.

DREW is following two lines in its analysis of the impacts of drawdown on irrigation from the lower Snake. Initially DREW agreed to just look at the cost of mitigating the effects of drawdown by providing alternative pumping facilities, but has now expanded its analysis to include a brief look at net farm income as a way of estimating what it would cost to buy out the farms. Preliminary indications are that both net incomes and land values for some of the irrigated farmland supplied by pumps from the Ice Harbor pool may be quite high. Nearly a fifth of the land is in orchards or vineyards, where stand establishment can represent a $10,000 per acre investment. Nearly a quarter of the land is in hybrid poplar plantations. Preliminary indications are that DREW estimates of the cost of alternative pumping, and their estimates of lost farm income, will both greatly exceed the ONRC figures. Thus, although the ONRC report uses an approach that double-counts the value of land in irrigated agriculture, it may nonetheless understate the direct economic cost of removing the land from its current uses.

[4] Lost Potential Hydropower Generation from Water Used for Lockage

When a barge is passed through a lock, this uses water -- water that does not go through the turbines. The ONRC's treatment of this issue implicitly assumes the lost power potential will be recovered if the lower Snake dams are breached. However, while decreased barge traffic caused by the closure of the lower Snake navigation could possibly decrease barge traffic and lockage losses at downstream dams, the lost power potential at the breached dams themselves cannot be recovered. This logical flaw in the ONRC report is similar to that associated with lost hydropower, discussed above.

[5] Port Costs and Taxes

There are two issues with the port costs. First, to the extent these represent taxes used for recovery of investment costs in port facilities, they are "sunk" and cannot be recovered or avoided, even if the dams are breached. Second, not all "port costs" are related to river transportation and those related to rail and trucking may even increase with the loss of river navigation. The ONRC analysis is incomplete without this second effect.

[6] Lost Potential Hydropower Generation from Water Used for Irrigation

It is certainly true that diverting water for irrigation has resulted in less water through today's turbines, and less hydropower generation. However, removing the dams and turbines on the lower Snake will not recover the lost power generation there. Some power generation now lost at downriver projects may be recovered, unless there is increased irrigation on other Snake and Columbia River farms to make up the lost crop production. If some lost direct river water pumping is replaced by increased ground water pumping from aquifers linked to the river, this would also affect downstream flows and generation.

[7] Reduced Irrigation Electricity Price Subsidy

This small item also depends on whether the irrigation using subsidized electricity rates is replaced elsewhere in the region using similarly subsidized electric pumps. If it is, then breaching the lower Snake dams will not result in an avoided cost for society, and this component of the ONRC estimate should be eliminated.

[8] Reduced Crop Subsidies

There are two issues here. First, many of the crops grown at the farms in question are not subsidized crops. Second, assuming crop prices and therefore demand are unaffected, lost production will shift to other locations, probably within the U.S., and there will be no savings to national taxpayers at all. Thus, the ONRC study probably overstates the benefits of breaching dams in this area as well.

In measuring the net economic costs or benefits of the proposed natural river drawdown, crop program subsidies are mostly irrelevant. They are a transfer from taxpayers to farmers, and cancel out in national economic accounting. Admittedly, subsidies can lead to costs if they promote inefficient resource use. In terms of regional economic impacts, subsidies can contribute to regional income, at the expense of taxpayers elsewhere. However, the ONRC study did not examine these efficiency and equity effects.

DREW will apparently not include crop program subsidies in its analysis.

Conclusions about the ONRC Report

The IEAB concludes that the ONRC report suffers from a number of serious estimating problems. Two assumptions seem especially indefensible: (1) that all of the system costs and salmon recovery expenditures (including costs of fish related constraints on generation) will be ended when the dams are breached, and (2) that the power lost if the dams are breached can be measured by the lost firm energy only and can be replaced at only 1.6 cents per kWh. Both of these estimated savings are so important to the calculation, that even small changes could reduce or even reverse the estimated benefits of dam breaching.

The Lower Snake River Juvenile Salmon Migration Feasibility Study (DREW) is addressing most of the issues raised by the ONRC study, including the salmon recovery spending, cost of power from alternative sources, and costs avoided by breaching, which are so critical to the ONRC bottom line. Because the DREW process has been an open one, with opportunity for interest group involvement and peer review, we are optimistic that its methodology will be sound. We expect this to be the definitive analysis of the lower Snake drawdown alternative.