College and Research Libraries JOSEPH J. KOHUT and JOHN F. WALKER Allocating the Book Budget: Equity and Economic Efficiency Gold's cost-benefit 1nodel for allocating the book budget is critiqued from the point of view of practicability, economic theory, and equi- ty. It is concluded that allocative formulas are preferred alternatives for distributing the book budget among departmental funds. Eco- nomic efficiency via cost-effectiveness analysis is suggested as potential- ly useful for within fund acquisition of library-resource units. THIS PAPER WAS ORIGINALLY CONCEIVED as a brief discussion of Gold's marginal cost- benefit model for allocating the book budget among academic depart- ments.1 However, it soon became ap- parent that a more lengthy paper was necessary. Several features of cost-bene- fit analysis must be explored in order to put the marginal cost-benefit model in perspective. Furthermore, Gold advo- cates economic efficiency versus equity, an issue that challenges a fundamental premise by which librarians have tradi- tionally allocated the book budget. This problem brings up the role of allocative formulas and their validity. Finally, there was the need to comment on cost- effectiveness analysis and its potential impact on economically efficient acquisi- tion of library-resource units. DIFFICULTIES OF CosT-BENEFIT ANALYSIS In their oft-cited review article, Prest and Turvey list several limitations in principle to cost-benefit analysis. 2 Among these is the problem of intangible bene- fits ; that is, those benefits that do not I oseph I. Kohut is science librarian and John F. Walker is associate professor of economics, Portland State University, Port- land, Oregon. lend themselves to quantification. The scope of this problem is magnified for complex social institutions I programs wherein there are many different bene- ficiaries and many kinds of benefits, both personal and social. Within this category are academic libraries with multiple objectives (except in the most general sense) and diverse readership with diverse needs. The problem of benefit enumeration is, therefore, ex- ceedingly difficult, if not impossible. Nonetheless, the analysis demands that all benefits, many of which are subtle and indirect, be listed without double counting. The evaluation of intangible benefits must also be considered, for it is insuf- ficient that they be merely noted. But how does one put a numeric value on such personal reading experiences as esthetic appreciation, joy of discovery, insight, boredom? Or on social benefits such as the educative value of reading, information as a community resource? Even if some benefits are quantifiable, they must be valued in dollars. Most of these are not money-valued benefits; that is, their worth has not been estab- lished in a competitive marketplace. It may, with considerable effort, be possi- ble to assign shadow prices or arbitrary I 403 404 I College & Research Libraries • September 1975 weights, but it is doubtful whether these would add much to decisions made on the basis of intuition or experience. 3 In any case, incommensurables cannot be ignored, a danger not uncommon in cost-benefit analyses. On the other hand, bias must not be introduced by under or over-evaluation. To simply list incom- mensurables separately is inadequate if the bulk of the benefits cannot be as- signed reasonable dollar values. In his summation of the first Brookings con- ference on government expenditures, Dorfman concludes: The practitioners were very skeptical and inclined to doubt whether the most important social effects of gov- ernment investments could ever be ap- praised quantitatively by cost-benefit analysis or any other formalized meth- od. One of them likened the problem to appraising the quality of a horse- and rabbit-stew, the rabbit being cast as the consequences that can be mea- sured and evaluated numerically, and the horse as the amalgam of external effects, social, emotional, and psycho- logical impacts, and historical and aes- thetic considerations that can be ad- judged only roughly and subjectively. Since the horse was bound to dominate the flavor of the stew, meticulous eval- uation of the rabbit would hardly seem worthwhile. 4 The rabbit is neither . sufficiently large nor the horse sufficiently small to make a significant quantitative impression on the book-budget stew. There is yet another impediment to describing the marginal-benefit curves demanded by Gold. Benefits must be measured annually for each curve, one for each academic department. There · must then be a single, objective value- measure of benefit applicable to all de- partments. This point introduces the problem of incomparabilities. The uni- versity is a pluralistic society with a com- plex mix of program objectives (e.g., teaching; research; providing a liberal education; training businessmen, teach- ers, etc.). To take only major reference tools, for example, would all patrons agree as to the absolute value of the Art Index, Business Periodicals Index, Chemical Abstracts, Bibliography of the Modern Language Association? Because the value of intangible benefits is pri- marily dependent on one's point of view, there is no single measure that could be agreed upon. Willis H. Shape- ley has commented: In theoretical discussions, it is nice to visualize a curve, with "decisions" being made by picking the maximum or something that is lower or higher than something else. But for utility in the real world of decision-making, agreement is needed, not only on the many inputs to the measure, but also on the validity of the measure itself, including the underlying theoretical concepts and me~od of calculation. The possibilities of getting a measure with enough objectivity so that people whose interests are going to be ad- versely affected will accept it as a judge or a major tool in the decision- making seem to me very small in- deed.5 It is, therefore, naive to assume that fundamental differences in outlook among the diverse groups compnsmg a university can be resolved and an ob- jective value-measure of benefit devised. How then can the library, which is ob- ligated to support this heterogeneous group of users, all of which are com ... peting for library resources, presume to make the value judgments necessary to produce a cost- beneficial solution? In addition to these problems relating to the vertical axis of Gold's marginal cost-benefit model is a major difficulty with the horizontal axis. There must be a uniform measure of output; that is, a unit of consistent size and quality. Since there is no good correlation be- tween quantity and quality of litera- ture output, such a measure cannot be derived. It is unreasonable, for exam- ple, to expect that a long novel (or part thereof) is always "better" than a short poem. Moreover, there is the difficulty expressed above in that not everyone would always agree on what is "better" because "better~' is a subjective concept. Even if these obstacles could somehow be overcome, such a hypothetical unit would be unworkable in practice. Li- brarians must deal with library-resource units. The unit of output of the book budget is a monograph, a serial, etc. Though discrete items, they are non- equivalent, nor were they intended to be. To put it another way: generation of the marginal-benefit curves requires that library-resource units be homogene- ous. Neither are they, nor can they be homogeneous because virtually all are copyrighted. In this sense homogeneity is illegal. The notion of copyright raises the question of monopoly. Monopoly prices disturb the conditions of Pareto opti- mality and create inefficiencies in Gold's sense of the word. The rules of ''the general theory of the second best" ap- ply when any element assumed by the model is not fulfilling the rules of gen- eral economic equilibrium. Though tb.e rules of second best are not well de- veloped, it is a fair interpretation that if all conditions for equilibrium cannot be met, then none of them should be enforced. 6 Thus, Gold's "rules" for al- locating the book budget, which are imi- tations of the conditions for general equilibrium in a competitive economy, may produce less efficiency, not more, since elements of monopoly are ines- capable. In effect, Gold is advocating a plan- ning programming and budgeting sys- tem ( PPBS) for every subject depart- ment of the library. A standard objec- tion to such an approach is that a full cost- benefit analysis of any public ex- penditure is meaningful only when Book Budget: Efficiency I 405 compared to similar studies of all other expenditures. That is, a comparative study of each book in every library is theoretically required. The time and cost of such a study would be over- whelming, particularly when spillovers, uncertainties, inadeql)ate information, time-stream cost- benefit discounting, and other difficulties create distortions un- less compensated. 7 Finally, economists cannot agree on the criteria to judge a professionally acceptable cost-benefit analysis. Har- burger observed "the need for an ac- cepted set of professional standards for this type of study should be obvious."8 The purpose of Harburger~ s article was to suggest three criteria. One of these criteria has been challenged by Boad- way, who concluded that it improperly ignored equity relations which must be weighted on "non-economic grounds."9 Since economists can~t agree on the cri- teria to judge a good cost-benefit analy- sis, how can ·they expect librarians to do so? EcoNOMIC EFFICIENCY AND EQUITY There appears to be a failure by Gold to recognize that allocation of li- brary-resource units is not so much a capital investment project as it is one of general welfare economics. In business it is obvious that profits (benefits) should be maximized for the concern. In the matter of "public" fund expend- itures, such as a library, where there are many beneficiaries, the issue is not so clear cut. Moreover, the libr.ary~s goal is to produce mainly intangible benefits, not to increase future income. When benefits are diffuse, mainly noneconom- ic, and nontransferable into standard units of output, the cost-benefit ap- proach is inapplicable. In general welfare economics there are really two questions that must be answered: ( 1) Is the program econom- ically efficient? and ( 2) Is the distribu- 406 I College & Research Libraries • September 1975 tion of resources "desirable"? According to Weisbrod: ... advice by economists to decision makers tends to be restricted to ques- tions of the first type. At the same time actual decisions do, and should, also reflect questions of the second type. As a result economists often are dis- appointed that their advice carries lit- tle weight, and decision makers are disappointed that economists do not provide more complete advice.1° Whether economic efficiency is sufficient in itself is a moot point. The fact is that library administrators are con- cerned about equity and the direction of collection development. These con- cerns have been evident at least since allocation by academic department be- came common at the tum of the cen- tury. Part of the problem here is a confu- sion of roles between the .analyst and the decision maker. According to Bon- nen, for example, the question of who should benefit is a normative matter in- volving value judgments that an econo- mist as a scientist is not in a position to make.n Similarly, the need for explicit decisions in allocation of the book bud- get is not obviated by an assignment of costs and benefits and simply allowing the resources to fall as they may. In truth, cost- benefit analysis is not even a politically neutral tool, for like it or not there will likely be pressures de- pending on whose ox is being gored. The danger that the analysts will dic- tate decisions is very real for: We live in a culture that worships quantities. A computer printout on tis- sue paper can in a political environ- ment thicken into an arras behind which no one can see. It insulates against common sense. A man who says, "My experience and judgment tells me this should be done," is hard put to defend his opinion against a man who says, «The computer tells me that should be done." In which of these men it is wise to put one's faith depends on how much of the hard-bit- ten reality has been conveyed to the computer. In other words, the oppor- tunity cost of benefit-cost analysis, nar- rowly defined can be the debilitation of the power of judgment.12 We do not mean to deny the utility of mathematical models, or even cost-bene- fit analysis where applicable. But their limitations must be realized, and the analysis alone must never be allowed to preempt good judgment in matters that are largely ((noneconomic" and "non- scientific." ALLOCATING BY FoRMULA Gold seems to think that inefficiencies will result by any allocative model not in accord with the rules governing the general economic equilibrium. In fact, it has been argued that quick rules of thumb about pricing and purchasing that keep costs of decisions down actual- ly increase the probability that the sys- tem acts the way the theory predicts. 13 Gordon concluded that to the extent "rules of thumb" used by decision mak- ers .are inconsistent with theory, it is be- cause the theory is weak.14 Quick rules of thumb or formulas for allocating the book budget are not, therefore, without economic efficiency. In fact, these formulas meet another frequent- ly mentioned justification for cost- bene- fit analysis. That is, they make explicit factors formerly unconsidered or only intuitively sensed in the decision-mak- ing process. Ever since the Clapp-Jordan formula for estimating liminal adequacy of aca- demic libraries, value judgments ex- pressed by weighing factors have been an important aspect of formula budget- ing. In general, four major factors have been considered in the allocative formu- la: subjective judgments based on col- lection evaluations and historical in- equities, size of academic departments, level of program and usage, and litera- ture size.' All emphasize the importance of equity. Collection evaluation and standards were common long before allocation by formula. The literature on this topic is extensive and varied. 15 The process is obviously subjective and dependent on knowledgeable analysts. An element of "objectivity" may be introduced by using numeric guidelines for allocating book funds according to preestablished, arbitrary standards of adequacy. It is generally assumed that there is some cor- relation between the size of a collection and its value. Goyal stressed the importance of de- partment size.16 He initially recognized the importance that society and the uni- versity attach to the work of a given de- partment. The problem is greatly sim- plified by assuming that society and the university attach equal importance to the work of each department and that bias in favor of one or another would be expressed in the size and growth of such departments. Thus, allocation by departmental size allows librarians to avoid independent value judgments be- tween incomparabilities (e.g., whether physics is more important than philoso- phy). The judgment is, in fact, in large part made for society and the li- brary by the university, including the students. The organizational factor in alloca- tion has been further refined by de- scribing departments according to level of program. Burton, for example, used weighing factors based on library usage (as measured by circulation data) for lower and upper division undergradu- ates, graduate students, and faculty. 17 Thus, usage is considered in conjunction with department size and composition. McGrath described departments by their curriculum which he related to annual literature output for each sub- ject.18 Literature output and other vari- Book Budget: Efficiency I 407 abies were later submitted to factor analysis as a basis for an allocative for- mula.19 Inherent in the literature-size factor are several assumptions. For ex- ample, the potential utility of a given subject literature is proportional to the size of the literature of that field re- gardless of the subject field. · This as- sumption avoids such irreconcilable ar- guments as: "One book in physics is worth three books in education." (In fact, no book in education may be con- sidered of any value to a physicist, and vice versa.) The assumption also avoids such unprovable-given the present state of the art-statements as: "Quality control is higher in physics than in edu- cation, therefore, physics books are po- tentially more valuable to a physicist than education books to an educator." Another obvious assumption is: If books I serials are not particularly impor- tant to a "reader" in a given subject, then fewer books I serials will be pub- lished in that subject than in more lit- erature intensive fields. Or, to put it more simply: Those disinclined to read are disinclined to write. Given the above assumptions, it seems reasonable, from the point of view of equitable access to the literature, that annual literature output for each subject be considered in allocating the book budget. All other factors being equal, each department would receive an equal percentage of the total annual literature output in the corresponding discipline. Dillehay used the book review litera- ture to compile quantity and cost figures for each subject. 20 These data were then used in an "augmentation" formu- la for allocating the book budget. An additional advantage of this approach is a book-review file which is an aide in evaluating proposed purchases. Implicit in Dillehay's model is an understanding that allocation has two aspects: ( 1) an equitable distribution of books among funding units; and ( 2) an efficient allo- 408 I College & Research Libraries • September 1975 cation of funds within each funding unit. This distribution is in accord with the heterogeneity of users/needs on one level and a more or less homogeneous group of users/needs on another. Once an equitable distribution of re- sources is achieved, Kohut is concerned with maintaining the desired balance of library-resource units among the various funding units.21 Like Burton,22 he rec- ognizes the need to differentiate acquisi- tions by form of publication (mono- graphs, serials) and that the cost and relative importance of monographs and serials vary by subject. Both employ li- brary-resource units as a basis for allo- cation. Burton, however, is interested in deriving book-budget requirements to support academic programs. Kohut is concerned with the opposite problem wherein a book budget which is given must be appropriately distributed by form of publication (monograph, seri- al) within each funding unit. Kohut's correction is intended to adjust for an- nual changes in the total book budget and/ or fluctuations in average mono- graph and serial costs by supject. ( Ko- hut's correction has been replaced with a simple algebraic solution devised by Gary Sampson, Library, Portland State University. It has been successfully ap- plied in a pilot study using internal cost data generated in the last two years.) It is assumed that there are no inten- tional changes in the relative distribu- tion of library-resource units among funding units (e.g., to reflect changes in institutional programs). By tying budget aiiocation directly to library-re- source units, equity is maintained, and an explicit view of collection develop- ment and its cost by subject and form of publication is presented. BooK SELECTION AND EcoNOMIC EFFICIENCY A strong case for economic efficiency may be made on the level of individual departmental funds. A promising tool is cost-effectiveness analysis, a study de- signed for single objective problems and more or less homogeneous groups of users and values. Moreover, it avoids the obstacles of disparate benefits and their dollar values. The concern is not with collection worth, but with collection ef- fectiveness (measured by usage) in re- lation to costs. Cost-effectiveness analysis may be used as an adjunct to more subjective, tradi- tional means of book selection. Lancas- ter, for example, considers cost-effective- ness in relation to coverage (collection content) of an information retrieval system. 23 The calculation of pay-offs is becoming more promising with advances in bibliometrics (e.g., Bradfordian- Zip£ distributions, literature obsoles- cence rates, acquisition characteristics curves, etc. ) . Such bibliographic studies are becoming a prominent feature of library research.24 However, as the pos- sibilities of cost-effectiveness analysis be- come greater, we must be ever alert to its limitations: It is important to remember that all analysis of choice falls short of scien- tific research. No matter how we strive to maintain standards of scientillc in- quiry or how closely we attempt to follow scientific methods, we cannot turn cost-effectiveness analysis into sci- ence. Its objective, in contrast to that of science, is primarily to recommend -or at least to suggest-policy, rather than merely to understand and pre- dict. ... Human judgment is used in designing the analysis, in deciding what alternatives to consider, what factors are relevant, what the interre- lations between these factors are, and what numerical values to choose, and in interpreting the results of the analy- sis. This fact-that judgment and in- tuition permeate all analysis-should be remembered when we examine the results that come, with apparent high precision, from analysis.25 Together with a recognition of user idiosyncracies and the experience, com- mon sense, and intuition of knowledge- able bibliographers, cost-effectiveness analysis may eventually help to improve the economic efficiency of book selec- tion. SuMMARY There are a multitude of difficulties in principle and practice that militate against a cost-benefit resolution to allo- cating the book budget. In fact, there is disagreement among economists that cost-benefit analysis can ever be done well for complex social institutions. It is, therefore, unreasonable to expect that librarians do the kind of detailed economic analysis that economists them- selves do badly. In a more general sense, our quarrel is not with economic effi- ciency per se, but with those who would make economic efficiency the only rele- Book Budget: Efficiency I 409 vant value. It is prudent to keep in mind the admonition of Wildavsky: It seems unfair to blame the evangel- ical economizer for spreading the gos- pel of efficiency. If economic efficiency turns out to be the one true religion, maybe it is because its prophets could so easily conquer. 26 In problems of welfare economics, equity must be considered in addition to economic efficiency. Thus, the con- cept of equity is justly entrenched in the library profession. Though equity is dominant in formula budgeting for books, there are also economic efficien- cies realized by such rules of thumb. Furthermore, economic efficiency via cost-effectiveness analysis may have a role in book ·selection as bibliographic investigations of subject literatures pro- gress. REFERENCES 1. Steven D. Gold, "Allocating the Book Bud- get: An Economic Model," College & Re- search Libraries 36:397-402 (Sept. 1975). 2. A. R. Prest and R. Turvey, "Cost-Benefit Analysis: A Survey," Economic Journal 75: 683-735 (Dec. 1975). 3. Roland N. McKean, "The Use of Shadow Prices," in Samuel B. Chase, Jr., ed., Prob- lems in Public Expenditure Analysis (Washington, D.C.: The Brookings Insti- tution, 1968), p.33-77; Julius Margolis, ''Shadow Prices for Incorrect or Nonexis- . tant Market Values," in Robert H. Have- man and Julius Margolis, eds., Public Ex- penditures and Policy Analysis (Chicago: Markham, 1970), p.246-90. 4. Robert Dorfman, "Introduction," in Robert Dorfman, ed., Measuring Benefits of Gov- ernment Investments (Washington, D.C.: The Brookings Institution, 1965), p.2. 5. Willis H. Shapeley, "Comments" on Fred- eric M. Scherer, "Government Research and Development Programs," in Robert Dorfman, ed., Measuring Benefits, p.61- 62. 6. R. G. Lipsey and Kelvin Lancaster, "The General Theory of the Second Best," Re- view of Economic Studies 24: 11-32 (1956). 7. Kenneth S. Allen, Current and Emerging Budgeting Techniques in Academic Librar- ies, Including a Critique of the Model Bud- get Analysis Program of the State of Wash- ington (Seattle?, 1972). The arguments pre- sented by Allen against PPBS for libraries are valid in the present context. 8. Arnold Harburger, "Three Basic Postulates for Applied Welfare Economics: An Inter- pretive Essay," Journal of Economic Litera- ture 9:785 (Sept. 1971). 9. R. W. Boadway, "The Welfare Founda- tions of Cost Benefit Analysis," Economic Journal84:939 (Dec. 1974). 10. Burton A. Weisbrod, "Income Redistribu- tion Effects and Benefit-Cost Analysis," in Chase, ed., Problems in Public Expenditure Analysis, p.177. 11. James T. Bonnen, "The Absence of Knowl- edge of Distributional Impacts: An Ob- stacle to Effective Policy Analysis and De- cisions," in Haveman and Margolis, eds., Public Expenditures and Public Policy, p.246-90. 12. Ruth P. Mack, "Comments" on Burton A. Weisbrod in Chase, ed., Problems in- Pub- lic Expenditure Analysis, p.220--21. 13. Richard V. Clemence, Readings in Eco- , nomic Analysis, vol. 2, Prices and Produc- tion (Cambridge, Mass.: Addison-Wesley, 1950). See the articles by F. Machlup and R. A. Lester. 14. R. A. Gordon, "Short Period Price Determi- 410 I College & Research Libraries • September 1975 nation in Theory and Practice," American Economic Review 38: 265-89 (June 1948). 15. George S. Bonn, "Evaluation of the Collec- tion," Library Trends 22:265-304 (Jan. 1974 ). 16. S. K. Goyal, ''Allocation of Library Funds to Different Departments of a University- An Operational Research Approach," Col- lege & Research Libraries 34:219-22 (May 1973). 17. Robert E. Burton, "Formula Budgeting: An Example," Special Libraries 66:61-67 (Feb. 1975). 18. William E. McGrath, "Determining and Al- locating Book Funds for Current Domestic Buying," College & Research Libraries 28: 269-73 (July 1967). 19. William E. McGrath, Ralph C. Huntsinger, and Gary R. Barber, "An Allocation For- mula Derived from a Factor Analysis of Academic Departments," College & Re- search Libraries 30:51-Q2 (Jan. 1969). 20. Bette Dillehay, "Book Budget Allocation: Subjective or Objective Approach," Special Libraries 62:509-14 (Dec. 1971). 21. Joseph J. Kohut, "Allocating the Book Bud- get: A Model," College & Research Librar- ies 35:192-99 (May 1974). 22. Burton, "Formula Budgeting." 23. F. W. Lancaster, "The Cost-Effectiveness Analysis of Information Retrieval and Dis- semination Systems," Journal of the Amer- ican Society for Information Science 22: 12-27 (Jan.-Feb. 1971 ). 24. Allan D. Pratt, "Libraries, Economics, and Information: Recent Trends in Information Science Literature," College & Research Libraries 36:33-38 (Jan. 197 5). 25. Edward S. Quade, "Introduction and Over- view" in Thomas A. Goldman, ed., Cost- Effectiveness Analysis; New Approaches to Decision-Making (New York: Praeger, 1967), p.8. 26. Aaron Wildavsky, "The Political Economy of Efficiency: Cost-Benefit Analysis, Sys- tems Analysis, and Program Budgeting," Public Administration Review 26: 308 (Dec. 1966).