In Support of Cost Recovery for Publicly Held Geographic Information
Harlan J. Onsrud
Department of Spatial Information Science and Engineering
University of Maine
Let us assume that a local or state government jurisdiction has justified its GIS acquisition and operating expenses solely through cost savings and increased efficiencies in the operation of its governmental functions. After being in operation for some time, the jurisdiction begins to receive requests from private citizens and businesses for GIS data, products, and services that the government has available or is capable of producing. What approach should it take in handling the financial burden of responding to such requests?
Within the current GIS community, there appear to be three dominant approaches: (1) distribute GIS data, products, and services free to requesters, (2) recover the marginal costs required to respond to citizen requests, and (3) recover more than the marginal costs, such as (a) the full cost of producing the agency's GIS data, products, and services; (b) some portion of that total cost; or (c) greater than the total cost.
Free Distribution of GIS Products and Services
Under the "free distribution" approach, the position is taken that records contained in a government geographic information system have been acquired at taxpayer expense and should be free to anyone who has an interest in them and requests them. Those advocating this approach argue that the additional time and expense incurred by government in responding to private citizen and private business requests for such information should be borne as a public service and paid for through the government's general budget. Thus, the costs of responding to requests would be paid for in the same way local governments often pay for operating schools and maintaining streets.
It is assumed under this approach that the well-being of the community as a whole is promoted through the practice of "free distribution" or that many in the general population are interested in acquiring digital data and other GIS products from the government. However, this assumption seems to be inappropriate when applied to most local government jurisdictions in the U.S. because only a very small percentage of private citizens or businesses currently have any desire to acquire, for instance, digital spatial products from their local government GIS. The argument that providing "free access" to a few businesses or people results in substantial benefits to the overall community usually appears to be tenuous. Only when the costs are negligible in providing "free access" to all requesters or when the cost of administering the fees is larger than the fees collected does the free distribution approach appear to be logical. Even then, distribution at taxpayer expense may promote such increased numbers of requests over time that the costs become significant.
The establishment and maintenance of GIS operations in local government require substantial expenditures. The ability to develop GIS capabilities may be considerably lessened if government officials lack the ability to cover at least the costs of responding to private citizen requests. In addition, the greater the revenue the public GIS operation generates from outside sources, the greater the likelihood that government officials will value and maintain the operation over time. We may reasonably argue also that if public employees know that the costs of responding to private requests are being recovered and that the time and effort they expend in responding will not be a drain on other programs and responsibilities, they are likely to respond better to citizen requests for GIS services and products. This in turn enhances citizen access to the information housed in our government's records.
Recovering the Marginal Costs Necessary to Respond to Citizen Requests
In many jurisdictions, local government traditionally has had the ability or right to recover costs in excess of what it needs to meet its own needs. Marginal costs for responding to citizen and business requests for GIS data, records and services might typically include the costs of "staff time for selecting, down loading and handling data, computer time, consumable supplies (such as disks, tapes, and printouts), pro-rated hardware and software maintenance, and (staff time to assist requesters in using the data)" (Taupier 1992). These costs might also include clerical time to process requests and staff time to produce a specifically requested product or time to develop a computer program necessary to respond to a request.
Most people would agree that allowing the collection of marginal costs for providing citizen access to government-housed GIS records or for providing citizens with discretionary services is reasonable. Those items that may actually be included in the computation of marginal costs are often controlled by the language of open records laws in individual states. Rather than computing the actual expenses for responding to individual requests, the current laws of many states allow the establishment of fee schedules that are based on estimates of the average marginal costs for the requested products and services.
More problematic is the situation in which a government agency attempts to restrict subsequent use of records or information during the act of distributing records at marginal cost. If the agency has recovered the marginal cost of providing a data set request, it can be argued that the agency no longer has an interest in whether the data set is reproduced a hundred times for "free-riders." The reasonable and actual costs are all the agency would have acquired from those free riders. By not having to respond to the potentially large numbers of additional requests by "free-riders" who have satisfied their information needs elsewhere, the GIS operation can be run with minimal cost and the agency can retain its focus on those governmental goals for which the agency was created.
In addition, requiring private citizens or businesses to sign a statement limiting their subsequent use of the publicly-housed GIS data might be viewed by the courts as either an attempt to force a unilateral contract on the requesting party or an attempt by a party with unequal bargaining power to force a contract of adhesion on another. It may be argued that by signing the agreement with the government agency, private citizens or businesses gain nothing of value that they do not already have by right under the open records laws. If they gain nothing of significant value by giving up their right to reproduce and disseminate the GIS data, the argument that the contract is void and unenforceable may be fairly credible.
It may be argued that as technology advances and efficient networking becomes a reality, the value of much government data and information will lie in its timeliness. If this is true, failure to impose restrictions on the subsequent use or distribution of that data may have little practical effect on continued demands for information services from the government. For instance, a person attempting to determine who owns a particular parcel of land, determine where a power line is buried, or a person carrying out a study which involves data on a range of economic and spatial factors will almost always want to use the latest data, not something down-loaded from the government GIS six months, six weeks, or six days ago. If a local government provides on-line electronic access to its GIS and allows citizens and businesses to access such information and distribute it at will, few would object based on open records grounds to the government's generation of a substantial profit from such an information service.
In favor of imposing post-dissemination restrictions on data use, one might argue that comprehensive general-purpose GIS serving the needs of many government agencies are less likely to be built without the restrictions. If private citizens can obtain access to GIS data from a specific government agency at its marginal cost, then other government agencies can as well. Since the marginal cost of GIS data is typically far less than the total cost of originally collecting the data, government agencies will wait for other government agencies to collect the data for them and then acquire it for essentially the duplication cost. As a result, all agencies wait for other agencies to initiate data collection. Thus, GIS data sets, which would be of significant value to all government agencies in the jurisdiction as well as to the general public and the business community, are never compiled. Under this scenario, few incentives exist for individual agencies to collect any data beyond that which the agency minimally needs for accomplishing its own internal purposes, to inform each other of the data they have collected, or to share that data. An operational system of great value is never created due to disincentives and lack of incentives to share and cooperate.
On the other hand, if one restricts use of GIS data after its dissemination by an agency, it is better possible to construct incentives to encourage governmental agencies to produce data of benefit to other agencies and to share that data with them. Taupier (1992) has evaluated government agency incentives to share geographic data by considering three groups; those agencies that both produce and use geographic data, those that primarily use, and those that primarily produce. Using this approach, let us assume that all agencies using or producing geographic information in a jurisdiction join together in creating a "GIS data pool" and that the data contributed to the pool is required to meet pre-specified standards. The government-housed data pool might be housed in a centralized facility, exist in a decentralized networked environment connecting all cooperating agencies, or might exist only for administrative purposes with the pool existing in fact only as separate files within individual agencies (i.e. sharing capabilities are not necessarily directly connected electronically).
Let us further assume that an agency that both produces and uses GIS data contributes data to the pool and thereby receives the right to draw from it one hundred data elements for every element contributed. Because of economies of scale and cooperation, the agency gains those hundred elements at less than the marginal cost of gathering the data from the several separate agencies from which the data elements might have originated. In addition, the agency has paid for the cost of the data acquired through the use of data the agency itself has originally collected. Thus the agency benefits first by paying less and second by having the option of paying with data rather than with cash.
An agency that is primarily a user rather than a producer of geographic information will tend to buy data from the pool at marginal cost rather than bear the full cost of producing the data itself. By paying marginal costs to the cooperative pool, it also will pay less for the information it needs and receive a more comprehensive and standardized product than if it was forced to go to several non-cooperating agencies to acquire the information.
An agency which is primarily a producer rather than a user of geographic information will tend to want to continue to collect marginal cost payments from others and avoid joining in the common pool. The reason for this is that they have little need for information from others to carry out their own agency's traditional missions. As a producer of geographic information, the agency would contribute far more data to the pool than it would ever draw out. However, assuming that the agency's productivity in producing geographic information is measured by the extent to which other agencies and the private sector use the information the agency produces, the agency's productivity is likely to expand greatly by having its data more readily accessible to others through a cooperative arrangement. Thus, such an agency also would benefit by cooperating.
Unless restrictions are placed on the use of data after it has been drawn from the pool, this scenario tends to be far less workable. Agencies have less incentive to participate. For instance, a private sector firm might acquire the government's entire GIS database at a fee set at the government's marginal cost of duplicating it. The private firm then resells the database to fifty other individuals at less than what the government can reproduce it for, perhaps with or without some significant value-added enhancements. A government agency user of geographic information looking for a "free ride" might be able to purchase its needed GIS data more cheaply from the private sector supplier and not be bothered with data sharing arrangements with the pool. If one agency can ride for free then others can too and we return to the condition in which comprehensive GIS data sets never get built because each agency would rather choose to ride for free than spend money from its budget.
If those using the GIS data pool are restricted from reselling or redistributing the data to others, free riding is less problematic. We may not care under a marginal cost recovery approach whether the private sector reproduces and widely distributes valuable data sets produced by government. As argued in a previous article, we may even want to encourage this (Onsrud 1992). However, we don't want individual agencies riding for free since this condition results in sacrificing the cooperative GIS data sharing arrangement which could otherwise be created.
Rather than imposing restrictions on all users, an alternative would be to impose the restrictions only on other governmental agencies. In establishing a GIS data sharing arrangement within government, it may be possible to convince high level administrators within the governmental jurisdiction that those agencies who choose to enjoy the benefits of the GIS data pool without cooperating in its support and maintenance should be forced to pay to the pool the costs they would have borne if they had played fair in their use of the data. This approach is advantageous in that most agencies are likely to participate in the GIS data sharing arrangement because of the advantages it will bring and because those agencies who initially choose to refrain from participation will not cause problems as free riders. In addition, all data user agencies will have an incentive to continually question whether their internal data needs are met best through the sharing arrangement or through their own agency's data gathering efforts. This tension should force sharing arrangements to continually adapt to new conditions and thus produce an efficient data pool operation .
Recovering Greater than Marginal Costs
The third approach in implementing GIS capabilities in government is to view GIS operations as revenue generators. This approach to protecting and recovering government's investment in commercially valuable products has often been justified using custodianship and public trust arguments. These concepts argue that government held databases are not owned by the individual government agencies where they might be housed or controlled. Instead, it is argued that individual agencies hold such databases as custodians for the state or perhaps for the general public, such databases must be held in trust for the benefit of the entire public, and the trust corpus may not be given away or sold. (For more in depth discussions relevant to this position see, for instance, ALIC Numbers 1, 3, and 4 (1990) and Archer and Croswell (1989). For counter arguments, see Onsrud 1992).
Based on custodial principles, some jurisdictions take the approach that the price of GIS services should be set, if at all possible, to pay for the governments entire cost in implementing and maintaining its GIS operations. For instance, the average cost of providing a particular GIS service would be set at the total cost of providing that service divided by the number of individuals likely to purchase the service at that price. If the demand proves to be as predicted, user fees would pay for both the costs of meeting the governments internal GIS needs and the costs of responding to private citizen and business requests. Thus, even though the conversion to computerized GIS operations might be justified solely on savings from increased efficiencies and cost reductions for government, revenues generated from the new GIS capabilities would be used to offset the costs of the government's entire GIS operations. Revenue for meeting government's internal information needs traditionally has been raised through local property taxes, sales taxes, income taxes, or other general taxes. This approach has the potential for relieving pressure on those other revenue sources.
This approach has its difficulties. The number of private purchasers of a GIS product set at a particular price is difficult to predict. If the actual number of purchasers is less than predicted, full costs will not be recovered and it may be that full cost recovery is unlikely at any set price. The higher the price, the more likely it is that inequities in ability to access the information will occur. If the actual number of purchasers is higher than predicted, greater than full cost will be recovered and GIS revenues will be subsidizing other operations of government.
Based on custodial principles, some have argued that government agencies should extract the maximum value possible through sales of data sets to the private sector . Osborne and Gaebler (1992) argue convincingly that revenue generation by local governments through other than traditional taxing arrangements can be highly successful and that local governments should take advantage of profit generating opportunities whenever they can. None of the examples they describe involve the sale of information products or services nor do any of the examples appear to give rise to open access questions. Regardless, the successful examples they cite in which local government, private firms, and local citizens all appear to be satisfied with non-traditional entrepreneurial activities by government suggest there is room for profit seeking or at least cost recovery in the delivery of geographic information services to the private sector.
Osborne and Gaebler go on to suggest that user fees for public services will work only when three conditions have been met: "when the service is primarily a "private good" benefiting the individuals who use it; when those who don't pay for it can be excluded from enjoying its benefits; and when fees can be collected efficiently" (Osborne and Gaebler 1992, p. 204). Again, it can be argued that these criteria are either already met or that appropriate methods can be developed in government GIS operational environments to meet these criteria.
It is possible to argue that government-housed geographic information, in theory, is primarily a public good. Upon reaching this conclusion, the inevitable recommendation is that free enterprise should be allowed to reign by letting private citizens and businesses freely copy publicly-housed GIS data sets. From a practical perspective, however, few private businesses or citizens have the capital resources to take economic advantage of free government geographic information opportunities. Large information brokerage firms and perhaps a small number of for-profit local businesses gain an economic windfall. These businesses have the ability to pay for the data, products, and services they desire. Thus, unrestricted ability to copy government-housed data sets primarily results in private gain rather than in general gain for the public.
Likewise, to argue that the private sector should be charged only the "marginal costs" of GIS data sets is very close to arguing that the private sector should be able to acquire the data sets for free (Archer and Croswell 1989, King 1992). Data duplication costs, even if they include liberal estimates for all conceivable costs that might be reasonably charged for providing the service, are a very small fraction of the total expenditures by the government for the hardware, software, personnel, and data collection necessary to create and maintain an operational GIS.
Perhaps the greatest difficulty in imposing fees for public GIS services is in excluding those from enjoying the benefits who fail to pay the requisite fees. Although statutes and regulations may be passed and contracts may be signed, the inherent characteristics of electronic information and the inability to control those characteristics using the assumptions which underlay many of our work product protection laws make it increasingly difficult to enforce social arrangements which rely on the non-dissemination of information for any continued period of time. (e.g. For an extensive discussion, see Cleveland 1985).
One argument used in urging "open access" records environments is that citizen access to commercially valuable government GIS records for free or at marginal cost creates a social system environment which is more successful in promoting innovation and economic efficiency in the long run than one in which the government charges fees for its information and restricts the subsequent use of that information (Onsrud 1992). In reviewing options and in contrasting the federal "open access" approach in the U.S. with the national "cost recovery" approaches prevalent in Europe, Rhind states that "...the international differences in government ethos and legal constraints may - notwithstanding the widespread encouragement of private enterprise - ensure that the variety of uses and the level of experimentation and hence innovation will be greater in the U.S.A. than in Europe" (Rhind 1991, p.944). Similarly, Maffini asks "...(C)ould we imagine a substantial increase in GIS development in the UK if that country adopted the distribution policies of the U.S. (federal government)? I believe that ... a significant difference would be evident." (Maffini 1990, p.52). However, little proof has been offered to substantiate these and similar claims. The legislatures of numerous industrialized nations have chosen to promote and pursue user fees for government information services as a matter of course (i.e. greater than marginal cost recovery) under the belief that such is by far the better long term economic policy. They believe such an approach reduces the demand for government information services generally and equitably passes the burden of the services that are provided primarily to those who use them.
Evidence which supports either of the competing economic efficiency stances tends to be primarily anecdotal. Perhaps an in-depth comparative study should be made of the micro and macro economic benefits that have resulted from the various approaches taken by state and local government GIS operations. For instance, comparisons of economic effects might be made in the U.S. at the local government level among free access approaches, marginal cost recovery approaches with no restrictions on the subsequent use of data, marginal cost recovery approaches with subsequent restrictions, and approaches pursuing greater than marginal cost recovery. Similarly at the national level, an economic study and comparison might be made between the effect of U.S. freedom of information laws on the uses of federally-produced geographic information and the effect of, for instance, the cost recovery approaches of the Ordnance Survey of Great Britain on the uses of similar geographic information. However, until controlled studies are made, we have little reliable evidence to suggest that cost recovery approaches which recover more than marginal costs are any less economically efficient than unrestricted subsequent use approaches.
Another argument against allowing government administrators to have control over the copying and dissemination of some or all of their GIS records is that such restrictions are dangerous to the effective operation of a democracy (Onsrud 1992, Florida 1992.) Underlying the status quo of the U.S. Freedom of Information Act and the open records laws of the individual states is the belief that a presumption of open citizen access keeps government accountable. However, all such laws have exceptions. For instance, one of the narrowly drawn exceptions to the U.S. Freedom of Information Act is that information held by the government which is commercially valuable to a business from which the government obtained the information may be withheld from open disclosure (5 USCS [[section]] 552 (b) (4)). It can be argued that a similar narrowly drafted exception to state open records laws might be constructed to protect digital geographic information which is commercially valuable to state and local agencies. An alternate approach would be to pass state legislation that protects specifically named government-housed GIS data sets. Exception provisions such as these have been upheld as valid in the past and it seems reasonable that a narrow exception could be drafted to protect the commercial interests of state and local governments in their GIS data, products and services while not significantly infringing on the rights of citizens to know what the government has in its GIS records.
The extent to which properly and narrowly drafted exceptions to open records laws would provide a shield behind which potential government abuses might hide appears to be slight. Nations such as Canada, England, and Australia have experience with national and local government policies which place burdens on post-dissemination use of GIS products of commercial value to the government (Rhind 1991, ALIC No. 4 1990.) To date, little substantive evidence has been offered to show that economic and social equity principles or democratic principles have suffered due to governmental policies for charging fees for GIS data, products, and services. If socially valuable GIS operations could be developed more rapidly by placing only slight economic restrictions on the ability of individual citizens to access public geographic information data sets, then the cost recovery approaches being pursued by these other nations should be seriously considered.
Conclusion
This is the second in a series of two articles. The first article presented advocacy arguments in favor of minimal restrictions on the access and use of government GIS databases. This article has raised additional issues and has presented at least a sampling of counter arguments. The objective of the two articles has been to expose issues and explore options rather than to promote a particular stance.
In this author's opinion, whether to use a cost recovery approach and in what form is largely a political decision. Studies by the academic community and experiences by the practitioner community may raise issues and provide guidance on the likely ramifications of following certain policies. However, it may matter little from a practical perspective whether a group of experts show through authoritative studies that following one particular cost recovery approach provides greater economic and social equity benefits than other approaches. Democracies allow citizens to select government officials who may chose to ignore the advice of experts. Citizens also have the right at the ballot box to make mistakes. Therefore, the initial critical issue in determining which cost recovery alternatives are practically feasible in a specific jurisdiction may be to answer the question of who has the power in that jurisdiction to make decisions - whether those decisions are considered by the experts to be rational or irrational.
If through the political process, citizens have been convinced that leaders advocating "for-profit government information operations" are appropriate, such practices are likely to be implemented. However, political realities do not negate the responsibility of citizens, practitioners, and researchers to continually question and investigate whether specific approaches provide greater or lesser economic and social equity benefits than others. In democracies, irrational governmental policies are inevitably exposed over time with the result that the system corrects itself. Whether the policies actually implemented by elected officials and government bureaucrats are successful or unsuccessful is a determination which again is ultimately made at the ballot box.
Counter arguments exist to any cost recovery stance government officials may take in implementing GIS in their locality. Local jurisdictions need to fully investigate options and then select cost recovery policies and practices appropriate to their circumstances.
References and Further Selected Readings
5 USCS [[section]] 552. Freedom of Information Act (FOIA)
Archer, Hugh and Peter L. Croswell. "Public Access to Geographic Information Systems: An Emerging Legal Issue," Photogrammetric Engineering and Remote Sensing, Vol. 55, No. 11, November, 1989, pp.1575-1581.
Australian Land Information Council, Data Custodianship/Trusteeship, Issues in Land Information Management-Paper Number 1 (Belconnen ACT, Australia: Australian Land Information Council) 1990.
Australian Land Information Council, Charging for Land Information, Issues in Land Information Management-Paper Number 3 (Belconnen ACT, Australia: Australian Land Information Council) 1990.
Australian Land Information Council, Access to Government Land Information: Commercialization or Public Benefit, Issues in Land Information Management-Paper Number 4 (Belconnen ACT, Australia: Australian Land Information Council) 1990.
Blakemore, M. "Access and security issues in the provision of geographic information" in Medyckyj-Scott, D. et. al., eds., Metadata in the Geosciences (Loughborough, U.K.: Group D Publications Ltd.) 1991.
Cleveland, H. "The Twilight of Hierarchy: Speculations on the Global Information Society," Public Administration Review, Jan/Feb 1985, pp. 185-195.
Florida Legislature Joint Committee on Information Technology Resources, A Critical Analysis of Proposed Amendments to Florida's Public Records Law, Feb. 1992.
Gandhi, S. "Public and Private Access to Geographic Information Systems - Economic and Legal Policy Implications," GIS/LIS '91, Vol. 2, 1991, pp.945-953.
King, John L. Oral comments at NCGIA Initiative #9 Specialist Meeting on Institutions Sharing Geographic Information, Feb.28, 1992.
Maffini, G. "The Role of Public Domain Databases in the Growth and Development of GIS," Mapping Awareness, Vol. 4, No.1, 1990, pp. 49-54.
Onsrud, H.J. "In Support of Open Access for Publicly Held Geographic Information," GIS Law, Vol. 1, No. 1, Jan/March 1992, pp. 3-6.
Osborne D. and T. Gaebler. Reinventing Government: How the Entrepreneurial Spirit is Transforming the Public Sector (Reading, MA: Addison-Wesley Publishing) 1992.
Rhind, D. Data Access, Charging and Copyright and their Implications for GIS," EGIS91, 1991, pp. 929-945. Reprinted in International Journal of Geographical Information systems, Vol.6, No. 1, 1992, pp. 13-30.
Taupier, R. Draft comments at NCGIA Initiative #9 Specialist Meeting on Institutions Sharing Geographic Information, Feb.28, 1992.