(The following additional remarks were submitted for the record:) EXTENSION OF REMARKS BY DR. CECIL B. HAVER RELATING TO ADDITIONAL AUTHORIZATIONS FOR THE ARKANSAS AND THE ALABAMA RIVERS FOR NAVIGATION AND OTHER PURPOSES

In extending these remarks, I would like to present a systematic summary of our research on (1) the Arkansas River navigation proposal and (2) the maltiple-purpose development proposals for the Alabama-Coosa River. Summary of "An Economic Analysis of the Navigation Proposal for the Arkansas River and Its Tributaries," by Dr. Cecil B. Haver, Dr. W. B. Back, and Mr. L. A. Sjaastad

In this economic analysis we review the navigation development of the Ar kansas and Verdigris Rivers to Catoosa, Okla., as proposed by the Corps d Engineers. The corps proposal, however, goes beyond the navigation aspect analyzed herein. We leave for later consideration such questions as the economic feasibility of developing the Arkansas River and tributaries for hydroelectric power, water supply, recreation, and flood control. We shall address ourselves to such questions as the following: Is the navigation proposal as developed by the Corps of Engineers (published in H. Doc. 758, 79th Cong., 2d sess., and revised in 1954), economically feasible? Is the project or any separable segment o increment thereof able to accomplish a given purpose as economically as any alternative means, public or private? Are the assumptions used by Corps Engineers analysis sound and reasonable?

The authors recognize that public policies governing the development of this Nation's water resources are not determined solely on the basis of economic considerations. But even in these cases economic analyses can and do serve a valuable purpose by ascertaining the extent to which costs must be incurred to accomplish tangible as well as intangible results. Although agencies concerned with river basin planning may view economic horizons not extending beyond their own operations, one hopes that our elected Representatives in Congress may transcend this obstacle and seek to achieve decisions consistent with maximizing the general welfare. Our function as economists in analyzing this public investment proposal is to aid the decisionmakers of this Nation in promoting the maximization of the general welfare.

The methodology for this analysis cannot be stated in simple terms. Each commodity expected to contribute a substantial tonnage on the Arkansas is investigated in terms of production or consumption levels and trends within and outside the area. In addition, information about the behavior of traffic in commodities on similar waterways was used in this study. It was necessary to adjust the analysis to each commodity to take into account circumstances unique to the individual commodities and to the area in question.

The entire study is carried out within the context of the benefit-cost cri terion; that is, benefits, to whomever they may accrue, exceed the costs. Th criterion, commonly used in appraising public works projects, is a necessary but insufficient condition for project approval. It is indeed unfortunate that such a criterion should be applied as it is to these projects. It is easily demonstrated that the benefit-cost criterion cannot correctly choose between competing projects or uses of resources, if we allow the proposition that returns on available capital should be maximized.

The manner in which benefit-cost analysis has been applied permits the cre ation of vested interests and the use of pressure on the Congress by that interest. Since the costs of a Federal project to any area are proportional to that area's share of total Federal taxation, it is clearly to the area's advantage to favor any project in which its benefits exceed its share of the total costs. For example, in the case of Arkansas River project, about 2 percent of the total cost will be borne by Oklahoma and Arkansas taxpayers, while the major portion of the benefits accrue to those States. Thus, Oklahoma and Arkansas would be net gainers from the Arkansas project even if the benefit-cost ratio were far below unity. Thus, the benefit-cost criterion, as applied, can create local interest in a project which, from the standpoint of the Nation as a whole, should not be undertaken. This defect in the application of the benefit-cost ratio, of course, could be remedied by an application of “user” charges to the services of projects. However, this does not necessarily imply the impossibility of worthwhile public works projects without the application of "user" costs.

alysis involved a critique of the proposal made by the Corps of Engihose evaluation we judge to have underestimated the possible tonnage is instances and overestimated them in others. We chose to make the d prospective tonnage of individual commodities as liberal as the analy1 permit. Our estimate of total prospective barge traffic on a navigable s river is 1,936,924 tons per annum during the period of amortization. le.) As reported in House Document 758 (1947) the U.S. Army Corps eers estimated 9,015,000 tons as prospective barge traffic. The corps his estimate upward in 1954 to 13,221,171 tons. Our estimated total ve tonnage in barge traffic is about 22 percent of the original total estithe corps, and it is about 15 percent of their revised estimate. ble following summarizes the results of our commodity analysis and · barge tonnage projections with comparisons to those made by the Corps

eers.

nparisons of projected tonnage of barge traffic on the Arkansas River

alyzed in this study; estimates by corps in 1954 reevaluation extended.

ltural products and supplies comprised 16 percent of the tonnage 2 tons) and 17 percent of the savings projected by the corps; in our they accounted for 31 percent of the tonnage (609,154 tons). However, ysis suggests that the corps report greatly overestimates the prospecage of sugar, coffee, canned meat, canned goods other than meat, corn, , wheat, flour, and feed, and underestimates soybeans and phosphate barge-lot commodities account for 60 percent of the corps tonnage 2 tons) and 65 percent of our tonnage projection (1,235,770 tons). mary table indicates that the corps estimates are considerably larger most optimistic estimates that our analysis yields.

d a rather extensive analysis of possible petroleum movements on the ive waterway. The corps has estimated 3.9 million tons, which reprepercent of the projected traffic for the waterway. Our most optimistic n would be 356,000 tons, which is less than one-tenth of that estimated

265-60-pt. 5- -29

by the corps. Iron and steel products are another major item that was raised substantially in the Corps of Engineers 1954 restudy. These items accounted for 3.2 million tons, or 23.9 percent of the aggregate tonnage proposed by the corps. This projection is quite extraordinary when it is considered that in 1954 the total consumption of iron and steel in these two States was less than 800,000 tons, and in order for the corps projection to be a physical possibility, the consumption of these commodities in the 1975-2024 period must increase by 400 percent and all of that consumption must be carried into these two States by this river. This would be extraordinary in light of the fact that the terl movement of steel products on all inland waterways is only 6.5 percent of attional output. Our final projection was 323,000 tons, about 10 percent of that made by the corps.

A coal estimate for the Arkansas is a difficult one. The corps set its esti- la mate at 1.27 million tons in 1954, which represented a 10-percent decrease from its original estimate. Our analysis of trends and costs in this region relative to others, and movement of coal by various forms of transportation, leats as to estimate not more than a token movement of 100,000 tons, a mere 8 percent of the corps projection.

J

The corps projects 594,000 tons of scrap iron and steel. This is a surprisizzy high estimate when waterway statistics indicate that in 1957 all waterway (4 only carried 1,400,000 tons and that was only 1.9 percent of the total movemes of that product. The corps also estimated a large less-than-barge-lot and less than-carload-lot traffic for the Arkansas Waterway. Its 2,586,000 tons p jection represents 24 percent of the sum of all other traffic estimated for waterway. Waterborne statistics suggest that this is a gross overestim for instance, the traffic summary for the Ohio suggests that the 10 tep modities, none of which can be said to be package freight, take 96.5 percent | the tonnage; package freight then must be less than 3.5 percent of the to handled annually.

Our analysis does not attempt to estimate actual traffic savings expected t accrue to the proposed Arkansas River development, rather we inquired i the range of savings that would be required to justify by the benefit-to-cost crite rion the navigation portion of the project. The Chief of Engineers set t benefit-to-cost ratio for the entire project as 1.08 in 1949 and then increased the to 1.19 in the 1954 revision. The benefit-to-cost ratio for the navigation phas according to the Corps of Engineers estimates, is 1.15.

Using our own projection of 1.94 million tons of traffic and the ers estimate of $3.07 per ton savings, we obtain an estimate of the annual adjusted benefits to navigation of $5.96 million. The corps estimate of anne charges attributable to navigation is $34,782.000. Thus, the benefit to-cost r3 would be 0.17 to 1. A more likely estimate of the benefit-to-cost ratio obtained by using the $1.63 per ton, the average savings experienced on t 10 of the Nation's most active inland waterways; thus, the estimated an gross benefits of $3,157,000 are obtained, yielding benefit to-cost ratio of to 1. Both of these benefit-to-cost ratios are much less than unity and the are both still on the basis of unadjusted benefits. Net savings to navigate development would reduce the unadjusted benefits by one-half (if we ass a straight-line demand curve) when evaluated on the basis of consumer surplus in transportation development.

Our analysis points out that Arkansas barge traffic would on an average ba to produce savings of $21.91 per ton, or 121 mills per ton-mile, for an estimate 1.94 million tons per year if the navigation project is to cover all costs associate with its development and operation. Not even this Nation's most navigab waterways ever approach such savings. Further, if the United States by this navigation project, the annual operation and maintenance costs wol be greater than the expected traffic savings; therefore, it would be in the bes interests of the Nation to leave the project unused.

It is obvious from our projected tonnages that the savings fall far short! those necessary to render the project economically feasible. On the basis of o analysis, one can recommend that the Arkansas project should receive a fu scale reanalysis and reevaluation. Our own study suggests that it is not fear ible and is not justified on the basis of the benefit-cost criterion. If you are willing to accept our analysis as definitive, then the navigation aspects of the project should be deauthorized, as there is no evidence of justified navigaties development for this stream.

ummary of “An Economic Analysis of the Proposed Public Development of the Alabama and Coosa Rivers," by Dr. Cecil B. Haver, Mr. L. A. Sjaastad, and Dr. Irving Dubov

Our research summary on the navigation aspects of this project are covered the direct testimony. Therefore, I would like to raise only one point with spect to navigation and then elaborate on our hydoelectric power benefits timate.

I am concerned with the question of improving navigation benefits estimates. ader present practice there are two problems: (1) To obtain reliable tonge projections and (2) to ascertain economically defensible per ton traffic vings. The method used in our analysis, namely, a mathematical projection based on economic variables such as trends in economic activity, production, and nsumption levels and trends, would appear to have considerable merit over an most wholly subjective approach to estimating potential traffic. The subctive approach for estimating tonnage has some merit, if unbiased sampling ethods are used. Certainly, such subjective methods would aid the corps in terpreting estimates that it obtains by objective methods.

There are a number of questions concerning the estimate of traffic savings. he basic problem is the method of calculating savings is inconsistent with onomic theory and our notions of a downward-sloping demand curve. It is vious that the corps has a downward-sloping demand function in mind when it nsiders sources of traffic. For instance, in the Alabama-Coosa report the rps indicates three sources of potential traffic: (1) Existing traffic now using her transportation routes, (2) undeveloped traffic which could not previously velop because of existing freight rates or lack of outlets, and (3) new traffic nerated by the waterway investment.

The implication of this statement is that the traffic projected for the AlabamaDosa is made possible because of the postulated lower barge rates. This incates that the demand curve for transportation in general is downward sloping. the demand curve were not downward sloping, the “undeveloped" traffic would already moving by some alternative to waterway transport, as would the Dotential" traffic predicted to result from industrial growth made possible by e proposed improvement.

It is fairly clear from this that we are faced with a declining demand curve r waterway transport, and that this curve can be expected to lie below that the other, more speedy, modes of transport. Under these circumstances, it is ot correct to assume that savings to this Nation's economy of a cheaper, but no means perfect, substitute for existing transportation is simply e difrence between rates multiplied by tonnage. The savings that are relevant e related to the area under the respective demand curve for waterway and ternative transportation, up to the quantity of transportation demanded of ch; i.e., the "consumer's surplus" is the relevant measure of savings to the ipper. In order, then, to make an accurate calculation of traffic savings, one ould need to know the respective demand curves (and there will be a different e for each assumption about the price of alternatives, ceteris paribus). Howver, we can make a first approximation only if we are willing to assume that e typical shipper is rather indifferent between waterway and nonwaterway ansportation (at the same price and quantity), and also that the resulting mand curve can be satisfactorily approximated with a linear function. This st approximation tells us that the real savings to the shipper involved in supying water transportation would be of the order of one-half the difference rates times the quantity of water transportation. This is so because we are king only the area under the assumed linear demand curve rather than the ctangle suggested by the price differential and the quantity of traffic on the aterway.

While this calculation is admittedly arbitrary, it represents a decided imovement over the corps' method. Also, it will be more accurate for determing savings on potential traffic not moving by rail or truck because of prohibitive tes than for traffic currently moving which would be diverted to the waterway that waterway was developed. This is so because in the former case we can reasonably certain that the demand curve for barge transport intersects e price axis at the current rail rate, while we can make no such assumption -the latter traffic. In the latter case, the intersection could be either above below the current rate (since barge transport is currently not available);

under these circumstances knowledge of the actual demand functions would be necessary to determine savings by the consumer's surplus approach.

Another important consideration in determining traffic is the necessity of prorating those savings among the waterways involved. For example, all of the savings of freight moving on the intercoastal waterway and then up the Alabama-Coosa should not be accredited to the Alabama-Coosa; rather these savings should be prorated by some predetermined criterion, so that only benefits directly attributable to the waterway in question are, in fact, accredited to that waterway. In order to avoid, wherever possible, the problems posed by this considerstion in the following analysis, comparative rates, which differ from those used by the corps, will be chosen. These refer only to the Alabama-Coosa.

It should be noted, however, that the navigation aspects of the Alabama-Coosa Rivers are not economically feasible on the basis of expected tonnage alone. Even if we took the corps' estimate, as we do, of gross traffic savings, each and every segment of this navigation proposal is not feasible.

Benefits and costs of hydroelectric power on the Alabama-Coosa

I feel I should extend my remarks concerning the benefits and costs of hydroelectric power because this appears to be a crucial point in the development of this system at the present time.

The benefits of hydropower produced by any water resource project may be ascertained by measuring the amount that users would be willing to pay f such power. Thus the value of power to meet such demands could be obtained from alternative sources in most parts of this Nation. Therefore, the cost £ power from the most likely alternative source may be taken as a measure of the value of the potential power output of a project. The so-called green book points out the following, "From a public viewpoint, a power project will be ece nomically justified if it provides power needed to meet expected power require ments at a cost not greater than that of the alternative source of power most likely to be used in the absence of the project."

As a practical procedure, the Federal Power Commission, in accordance with its interpretation of the recommendations implied in the green book, has meas ured the expected benefits of project power by obtaining power values on the cost of equivalent power, including interest and taxes payable on alternative sources of power. These sources may be public or private, but should represent the most likely alternative source to be utilized to serve the same market with out the project.

We are not particularly interested in the issue of public versus private power; the arguments on such are quite often irrational and irrelevant. However, we are interested in certain basic things that this controversy seems to rest upon, namely, differences in the way inputs and outputs are priced and the wat taxes are treated. In the main, in project reports on Government projects, the Federal power investment is charged an interest cost comparable to the GeF ernment borrowing rate. Private power developers must, of course, pay the prevailing market rates of interest which are higher. Similarly, private power producers pay taxes which, in the main, many Federal power projects do pay; or, charges of payments in lieu of taxes are not considered as a cost. À great deal of confusion with respect to water resource problems can be attrib uted to the failure to distinguish between the following three objectives whic are not unrelated and bear on the discussion above: (1) marginal cost pricing. (2) income redistribution, and (3) optimal resource allocation and investment efficiency. I refer you to our analysis for a discussion of these questions.

The Corps of Engineers' interim report on the Alabama-Coosa is very brief concerning the determination of power benefits from the project's two multiple purpose dams at Millers Ferry and Jones Bluff.

The corps' estimates were based upon data obtained from the Federal Power Commission. The value of hydroelectric power benefits are estimated to be the value of alternative private steam-electric power. Thus in the case of the Alsbama-Coosa, the Federal Power Commission could, as we did, draw on the very recent experience of the Alabama Power Co. in building plants in this area. In our analysis, in table form, we present the costs and annual production expenses of the Alabama Power Co.'s Gorgas No. 3 plant and compare this to the power value estimate used by the corps in appraising the Alabama River Basin power values.