with the current STI systems, the results of a search are usually no more than a few pages of information. Large volumes can be directed to the high speed printer at the STI facility; where the cost is less than online printing. If full text retrieval were available under present conditions, the costs of high speed communications and printing at the user's location would require a careful evaluation as to whether the text was time critical. This situation, of course, could change if the economics of STI system user were reduced. At present, it appears that 30 characters per second communications speeds are sufficient for most STI system users.

APPENDIX B

THE ROLE OF TRANSACTION COSTS IN THE

DESIGN OF ROYALTY PRICING SCHEMES FOR STI

by

Y. M. BRAUNSTEIN* and J. A. ORDOVER*

NEW YORK UNIVERSITY

*The views presented in this paper are solely those of the authors and do not necessarily represent those of New York University.

In this paper, we shall discuss various schemes for collection of royalty payments for reproduction of scientific and technical information (STI). We shall concentrate on the costs that accompany the enforcement of royalty price schedules. These costs are referred to in the economic literature as transaction costs. These costs have to be included in the design of actual pricing schedules. Indeed, a major argument for excluding certain users from payment is that the transaction costs associated with the collection of payment from these uses exceed the benefits of doing so. We would expect that such arrangements can be worked out between the users and sellers without a necessary intervention of the legislature or the courts.

The blanket license system involves a set payment to the owner of a copyright. Once payment is made, an unlimited amount of photocopying can be done.

There are two species of blanket licensing:

(i) direct licensing;

(ii) clearinghouse licensing.

Under (i) the owner of the copyright negotiates directly with the user of a journal/library for a fee. Under (ii) the copyright owner negotiates indirectly through the clearinghouse which pools various copyrights. System (ii) is analogous to the one employed in the music

area.

The

Comparative economics would seem to favor the second variant. major saving is in transactions costs: both in the case of a publisher negotiating agreements with a multitude of users and in the case of the user (library) negotiating with a multitude of copyright holders. Another major saving for the user results from a reduction in the number of payments that will have to be made. A similar reduction exists on the publisher's side. There are, however, important additional costs that appear if (ii) is used, rather than system (i). The major new cost is associated with the necessity of monitoring the photocopying in individual libraries in order to determine an equitable distribution of the proceeds. Those costs may be significant. According to their own estimates, ASCAP's transaction costs amount to approximately 20 percent of their gross revenue. Such costs would not be expanded under the direct licensing scheme. This is not to say that the direct licensing scheme does not require some monitoring of use, since under this scheme the extent of use will be important in setting the fee.

Overall, however, we suggest that the monitoring costs ought to be significantly lower under (ii) than (1).

are,

What is the economic impact of a blanket fee? In the limit it may not affect the amount of photocopying. This perhaps paradoxical result will be obtained if the library finances the cost of the photocopying permission fee by means of a lump sum (i.e., poll) tax which is levied uniformly on both users and non-users of the photocopying privilege. The poll tax places, however, an undesirable burden, on non-users who in effect, called upon to subsidize the users. On equity grounds the poll tax is clearly undesirable. Whether it should be implemented depends on how much the society would suffer from a reduction in socially desirable photocopying, which indubitably would occur if user fees were employed. Since unquestionably, a good deal of photocopying does not have any benefits over and above those that accrue to the researcher himself, arguments from both efficiency and equity standpoints would support our preference towards user fees. should be noted that if a library utilizes the user fee to collect the revenue, it commits resources to generating the same information that is necessary under the per-use license. If it is believed that the collection costs associated with the user fee are excessive, then at the risk of some unfairness a lump-sum tax ought to be imposed. The lump-sum tax is in essence in use now; all faculty members, students, and others contribute at least part of the library budget either in the form of lower salaries or higher tuition fees. Such payments are clearly independent of the use a particular individual makes of the library resources.

B.2.2 PER-USE LICENSE

It

The efficiency of per-use licensing depends on the expense associated with monitoring the use. Herein lies the main disadvantage of the peruse license over the blanket license. The costs of monitoring are technologically determined. At present these costs are probably high in the area of journal use, but relatively low in the area of bibliographic and data base use. Furthermore, the costs will depend on how coarsely use is defined. For example, different user fees may be imposed on recent journal copies as opposed to older copies. Medical journals may have different user fees than physical science journals, etc. The finer the partitioning of users, uses, and used objects, the better will the pricing system function as a signal towards efficient allocation of resources. Those gains in allocative efficiency must be weighted against the attendant information costs.

B.2.3 TWO-PART TARIFFS

The third system is a combination of the two preceeding ones. The two-part tariff pricing scheme involves a fixed entry fee, independent of use, and the per-use price.2 Such a system is currently employed

by the telephone company, for example, which charges a connection fee as well as the per-call charge. Such pricing systems have been recommended for industries in which production costs involve a substantial fixed cost element and in which, as a consequence, socially desirable pricing at marginal cost is not feasible in that it does not cover the total cost of output. A form of the two-part tariff would be a system whereby a library would purchase the license to photocopy by purchasing the hard copy of a journal and also pay a fee for each photocopy of an article from a journal in its collection. This would suggest that a pure per-use license is difficult to conceive of because the hard copy price of a journal is in fact an entry fee. (And we note that often libraries pay higher subscription prices than do individuals.) This may be so, but we prefer for reasons of taxonomy to think of the entryfee component as being an explicit payment for the right to photocopy.

It is clear that the current system does not fit neatly into either of these three categories of exclusion/collection mechanisms. There is in the library price an implicit component of a license to photocopy. But the extent of photocopying which such a license allows is not clear since the meaning of "fair use" is not apparent to either the publishers or to the librarians. Publishers expect some recompense for photocopying of their journals when such photocopying violates the existing statute. This brings on the element of the per-use license discussed above with an additional complication that some forms of use are exempt from that license, the "educational exemption" for example.

The first step in thinking about the appropriate form of a new copyright law should involve a clear understanding of the kinds of pricing mechanism that ought to be employed. This outline provides a basic classification scheme. In the next section, we shall begin to assess more precisely the various transactions costs associated with the three fundamental pricing mechanisms. [Note that for ease of exposition we have not followed here a suggestion often found in the literature that per-use and blanket mechanisms are but degenerate forms of the two-part (or multi-part) tariff system.]

B.3

ESTIMATING TRANSACTIONS COSTS

Although economic efficiency can be improved by the institution of per-use charges, it is obvious that some resources must be used to collect these charges. These "transactions costs" that are associated with an "exclusion mechanism" may be a negligible or significant sum relative to the charges that are imposed. In this section, we shall develop alternative estimates of their magnitude.

Exclusion mechanisms are the procedures by which one can determine who is using a good or service and then bill them for that usage. The difficulties of establishing such mechanisms have been cited as part of the rationale for the collective provision of public goods. At the present time there is a large but unknown amount of photocopying of