(c) Jeff P. Johnson and Harlan J. Onsrud
National Center for Geographic Information Analysis (NCGIA)
University of Maine
Orono, Maine 04469-5711
(207) 581- 2175
Abstract: This article takes a practical view of open access and cost recovery in local and regional government geographic information system (GIS) agencies. Legal issues are discussed in the context of open access and cost recovery. These legal issues present significant disincentives to implementation of cost recovery policy. To address more practical issues, anecdotes and results of recent case study research are presented that support various points of view. This work suggests that partnerships are the primary incentive for cost recovery, and illustrates the potential benefits of cost recovery for negotiating partnerships. However, the article points out case studies of GIS agencies that raise doubts about the likelihood of recovering significant costs in a cost recovery policy. The authors question whether these limited benefits of cost recovery policy are worth the likelihood of significant long-term legal and practical costs.
This article is not concerned with whether cost recovery is `right' or `wrong'. The philosophical arguments on both sides of the GIS information policy debate are well-worn, yet there has been little advancement in the debate from speculative arguments to practical knowledge. Speculative arguments are too easy to justify based on one's philosophy, rather than on substantive observations of the incentives and disincentives for cost recovery in local or county government GIS agencies.
What is needed is a discussion of whether cost recovery policy is worthwhile in a practical sense. Among the relevant issues are both legal and practical issues. What are the legal disincentives to cost recovery, and how significant are those disincentives? What are the legal incentives to cost recovery, and how real are such incentives? Non-legal issues, such as partnerships, data sharing, funding and politics are equally important to a GIS agency evaluating policy alternatives. What are the practical disincentives and incentives to cost recovery? How are these incentives and disincentives functioning in actual cases of information policy implementation? Perhaps most importantly, does cost recovery fulfill the expectations of those who adopt cost recovery policy?
This article addresses these questions by discussing legal issues and their positive and negative implications for cost recovery policy. The article then addresses practical incentives and disincentives for cost recovery, and examines these issues in the context of actual local and county government GIS information policies. This article builds upon master's research that explores relevant legal issues and presents case studies of open access and cost recovery policy in local and regional government GIS agencies. Through that work, the article transcends the traditional arguments over information policy by approaching information policy from a pragmatic yet intellectually informed perspective.
LEGAL ISSUES AND COST RECOVERY
Open Records Law and Exceptions
Geographic information systems are becoming major information resources in many local and regional governments. These systems are used in determining various local government policies or actions in areas as diverse as tax assessment, land use planning, public works, zoning, environmental assessment, and social programs. These applications can determine or influence which landowners are affected by which land use or environmental policies, the amount of taxes assessed to individual properties, where roads and sewers are built, and a host of other public policy decisions. Thus, GIS is an important information resource upon which major local government decisions are based.
Due to the direct influence of local and regional government GIS on local and regional- level decisions, public access to the information contained in a GIS has become a major issue. State public records laws require that such information be open for public inspection and copying. Legislation regarding levels of required access, acceptable fees and definitions of "public records" differ among states, but digital geographic information is clearly covered by most such laws. Under typically strict limits on fees and promotion of access in state public records laws, cost recovery policy is not viable.
However, state public records laws may also provide for exceptions for specific agencies or types of information. Some states have enacted such cost recovery enabling legislation, which exempt GIS data or GIS agencies from public records laws. Passing such legislation can be difficult. In Tennessee, the first effort by a GIS agency to pass an exemption was defeated by media opposition; the second attempt passed after the media was convinced the legislation would not hamper media access to information. In Washington State, two legislative efforts have failed; a third effort is underway this year. Clearly, open records law is a barrier to cost recovery policy. While the law can be overcome through exceptions for GIS data, enacting such exceptions can be difficult and time consuming. If legislation does not already exist in a state or municipality, GIS agencies seeking to implement cost recovery policies will have to go through the process of obtaining this legislation. From a practical perspective, the time and monetary costs of pursuing cost recovery enabling legislation can only be characterized as disincentives to cost recovery.
A recent survey shows that approximately 21% of local and county government GIS agencies assert copyright (Johnson, work in progress). Implementing a cost recovery policy suggests the need to impose control over or protect ownership rights in a GIS through copyright. The greatest value of a geographic information system lies in the information contained in databases and digital maps. If a GIS is to be protected, these two assets are clearly the most critical elements of the system. Copyright is often seen as an appropriate vehicle for protection of databases and maps; however, the extent of copyright protection of such resources is limited.
Under the Copyright Act of 1976, a database is defined as `a work formed by the collection and assembling of preexisting materials or of data that are selected, coordinated or arranged in such a way that the resulting work as a whole constitutes an original work of authorship'(17 U.S.C. Sec. 101, emph. added). In Feist v. Rural Telephone, the Supreme Court gave constitutional stature to the requirement of originality in selection or arrangement of facts. While originality can be found in GIS datasets, the greatest value of the information resource is in factual material.
The most direct problem with copyright of dataset, from a dataset producer's point of view, is that the information (the facts) in the dataset are not copyrightable. Under copyright law, those facts are part of the public domain. Equally troubling, the `selection and arrangement' of an electronic dataset are inherently pliable. Secondary users, anyone with a knowledge of computers and databases, could easily compile their own copyrighted dataset by rearranging the data in another dataset (Hicks 1990, 117), selecting only a portion of the data, or both. `Much of the fruit of a compiler's labor may be used by others without compensation' (Feist 1991, 1289). The malleable character of an electronic database, then, can render database copyright virtually useless.
Similar to digital datasets, protection of maps through copyright is weak. Maps appear to be copyrightable, provided they contain the required level of creativity or originality. Despite this protection of maps, the extent of this copyright must be questioned. The protection most likely extends to wholesale copying, but copying of specific attributes or features from a map might not be protected in certain circumstances. For example, a downstream user could digitize features from several different maps and create a new copyrightable map. Digital files could be manipulated to develop a substantially different map, copyrightable in its own right.
Originality is a requirement for copyright of maps, just as for databases. While maps as a whole can be original, each specific feature on a map may not. Locations of natural and cultural features could be seen as factual. The greater the accuracy of a map, the more likely such features could be seen as factual. Under such an interpretation, map copyright could extend to maps as a whole, but not to specific features such as roads, waterbodies, and vegetation. Digital maps could easily be manipulated, similar to databases, to produce new, copyrightable maps. Local governments may seek to avoid such uses through copyright or other legal mechanisms similar to the contract and licensing provisions commonly used for databases. Such protections are likely to encounter many of the same difficulties confronted in protection of databases. However, local governments must be aware that copyright alone is unlikely to protect their spatial data from subsequent use by requesters of either hardcopy or electronic maps. Just as copyright is a weak protection for databases, maps are probably only protected as a whole, leaving the individual features open to the public domain.
In sum, copyright is insufficient for protection of ownership and control to enable cost recovery policy. While one may consider this weakness to be a disincentive, it is not a barrier to protecting a GIS agency's ownership of information. Contracts or licenses may be used to achieve similar objectives.
Contracts and Licenses
A survey of local and regional government GIS agencies indicates that approximately 32.4% are implementing contracts and 26.8% licenses when providing copies of GIS data (Johnson, work in progress). Indeed, more respondents to the survey were using contracts than copyright to protect their ownership of data. Use of contracts instead of or in supplement to copyright is a reflection of the difficulty of applying copyright law to compilations of information.
Ginsburg discusses the use of `contracts as copyrights' (Ginsburg 1994, 325). `By contract, the provider may ensure a broader scope of protection than copyright would afford...Moreover, by contract, the information provider may secure protection for material that may not be copyrightable...From the provider's point of view, contract may therefore prove a more attractive means of obtaining the same, or more, protection than that available under copyright' (Ginsburg 1994, p.325). Indeed, the benefits of using contracts as a substitute for, or supplement to, copyright are already pursued by many local and regional government GIS agencies. Another benefit of contracts and licenses is that a GIS agency may include disclaimers of data quality in a contract, to warn users of shortcomings in data. Yet, local or county governments may not be able to enforce contracts.
Contracts and licenses must provide each party in an agreement with some `consideration' (i.e. compensation) for entering the agreement. Since access to and copies of GIS data is an unabridged right in some states, the provision of data from a GIS agency to a citizen might not be construed as legally valid consideration if challenged. Further, the one-sided nature of the contract (usually drafted by the GIS agency) could be another hindrance to enforcement of the agreement. The dataset recipient receives little or nothing through the contract or license if the transfer is already mandated by an open records law. Because the recipient receives little or nothing of value that was not already a right of that requester, a finding of `no consideration' could invalidate contracts between GIS agencies and the public.
Thus, the use of contracts or licenses is a potential mechanism for creating the control over GIS data needed for many cost recovery programs. At the same time, the enforceability of such contracts is in question. Maintaining physical control over GIS data is not easy- and may be impossible (Cleveland 1985). That difficulty, and the legal maneuvering that must accompany any attempt to control GIS data, is another disincentive to cost recovery policy.
Perhaps the most serious legal disincentive to a cost recovery policy is liability. Potential liability is important in many GIS agencies, which often are concerned with how requesters will use GIS data. The survey of local government GIS agencies found that 63.4% are using disclaimers when releasing data, demonstrating a common concern with liability (Johnson, work in progress). At first glance, it might appear that a cost recovery policy (or a restricted access policy) might reduce liability exposure by reducing the number of people who use the GIS data. Some GIS agencies consider using restrictions on secondary uses of data to limit liability, rather than to raise funds. As an example, one GIS agency in a Washington State county is considering adopting both copyright and contract restrictions, despite the agency's plan to disseminate data at marginal costs of dissemination. Liability concern can influence information policy.
However, liability concern should motivate GIS agencies to freely disseminate GIS datasets, rather than charging or restricting secondary uses. Charging for information or restricting secondary uses may actually increase liability exposure for GIS agencies through reduced opportunities to claim sovereign immunity protection and greater exposure to specific theories of liability.
Cost recovery policy agencies will often face greater liability exposure based on a public domain v. `limited-distribution' distinction.
`The law recognizes relatively limited property rights in generally disseminated information products...., but it also implies only narrow assurances about their quality, thus arguably contributing to ready access and broad use. By way of contrast, for limited-distribution information products, the greater recognition of property rights is directly related to the effort to protect the private data. The law also creates a more robust structure of qualitative assurances for these products (Nimmer & Krauthaus 1992, p.104).
Limited distribution, accomplished through restrictions on secondary uses, could increase a GIS agency's exposure to liability compared to a policy which puts that information in the public domain. Further, one my distinguish between proprietary and non-proprietary government activities, while assessing liability exposure. In performing a proprietary function, such as charging greater than marginal costs for copies of datasets, a GIS agency may more likely to encounter liability than in an open access policy. Indeed, `...[t]he intention of the parties could be important. For example, if a government set out to market data, one could ascribe to the government an intention to supply information fit for the purpose for which it was intended' (Dansby 1992, p.11). The proprietary nature of cost recovery could reduce the likelihood of a successful governmental immunity defense by a GIS agency (Onsrud 1992a, 5). In sum, restricting secondary uses of data and charging for data can each increase a GIS agency's exposure to liability.
Specific sections of liability theory support this notion, including negligent misrepresentation, fraudulent misrepresentation, professional malpractice, and strict liability (see Anderson & Stewart 1995). The two types of liability that appear most relevant are negligent misrepresentation and strict liability. If a local government realizes some direct benefits of releasing data and the recipient of the data suffers a loss due to reliance on that data, a local government could face liability under a negligent misrepresentation theory (Anderson & Stewart 1995, p.4-5). Because this theory of liability depends upon the exchange relationship in the data transaction, negligent misrepresentation `...should apply to situations in which local governments seek to raise revenues...through the sale of data' (Ibid., p.5). Emphasis on the business nature of the transaction suggests that GIS agencies that disseminate data for marginal costs should not be subject to this type of liability.
Strict liability, which originated in cases involving the sale of goods (Nycum & Lowell 1981, p.461), is another theory of liability potentially more relevant to the sale of GIS data than the dissemination of datasets. Because strict liability is imposed without consideration of fault, a strict liability claim is more difficult to defend against. Even disclaimers are not a defense. A GIS agency could conceivably confront such liability for providing information that caused a requester economic or physical harm. One case, Brocklesby v. United States and Jeppesen, involves an information provider held liable for accidentally providing faulty topographic information which caused several deaths during aircraft landing attempts (767 F.2d 1288; see Johnson 1995). A GIS agency might be held liable under similar circumstances, though most GIS data would probably not be used for such dangerous purposes. A more reasonable example is found in McCain v. Florida Power Corporation (593 So.2d 500). In this case, a mechanical trench digger operator was shocked when he struck an underground power line in an area marked `safe to dig' by Florida Power. Florida Power had obviously incorrectly located the position of underground facilities. He won a suit for $175,000. If Florida Power had obtained incorrect geographic information from a GIS agency, and that hypothetical GIS agency knew or should have known that the data might be used for inherently dangerous activities, that GIS agency might be held liable if the inaccuracies were related to the error in locating the underground facilities, even if they had a contract disclaiming liability (see Johnson 1995).
While strict liability is a relevant type of liability to any GIS, strict liability seems more applicable to proprietary data policy. The Restatement of Torts section relevant to strict liability states that no fault liability may apply to one acting in a `sale, rental or exchange transaction with another' ([[section]]552C). Whether the payment of marginal costs by a requester would constitute a sale or exchange relationship is unclear. Yet, cost recovery would clearly constitute a sale or exchange transaction. Cost recovery policy appears to increase the likelihood of a successful liability claim under strict liability or negligent misrepresentation.
On a different note, statutory liability is another relevant liability consideration. It is conceivable, though not likely, that local government GIS agencies could face legal challenges under antitrust or privacy statutes. Far more likely, some GIS agencies may encounter liability for violating or stretching the permissible interpretations of public records laws. In municipalities that engage in cost recovery programs, challenges could be likely. Regardless of the monetary costs of losing statutory challenges, the court costs of defending a lawsuit would be expensive. Since open dissemination policy would not likely face statutory liability, statutory liability is an important concern unique to cost recovery policy.
Charging for data or restricting secondary uses of GIS datasets invites greater exposure to liability than an open access policy.
Summary of Cost Recovery and Legal Issues
No legal disincentives are insurmountable obstacles to implementation of cost recovery. Exceptions to open records laws can be obtained for a GIS agency. Yet, obtaining such legislation could be expensive in terms of time and costs. Contracts may be useful for disclaiming liability and maintaining control over GIS information, though those contracts may be unenforceable or difficult to defend if challenged. Copyright is largely a useless tool in attempting to protect an agency's investment in a GIS. While these legal issues complicate the adoption and implementation of cost recovery, liability is a potentially greater disincentive. Cost recovery policy could potentially create liability exposure where an open access policy would be less likely to invite liability. In sum, the complexity of the legal issues associated with cost recovery demonstrates that the adoption and implementation of a cost recovery policy could have significant costs to a GIS agency.
However, other more immediately practical issues are confronting many local and regional government GIS agencies. In spite of the legal disincentives to cost recovery, pressing practical incentives for adopting cost recovery policy may outweigh those legal disincentives. The remainder of this article explores the practical incentives and disincentives to cost recovery, and compares the potential benefits of cost recovery to the legal and practical disincentives to such a policy.
PRACTICAL ISSUES AND COST RECOVERY
Politics may at first seem to be a speculative consideration in cost recovery policy. However, recent case study research indicates that politics may be the most important incentive or disincentive to cost recovery in some GIS agencies (Johnson 1995b, master's thesis). Some local or regional governments have individuals who oppose cost recovery. For example, one County GIS in Washington State has been influenced by the strong open access perspective of a upper level individual in the government. Within that County, that political perspective is one disincentive to adoption of a cost recovery policy. A political figure with a strong cost recovery perspective can similarly create strong incentives for cost recovery in a GIS agency. In a County GIS in Tennessee, one administrative-level individual championed the potential to recover costs from the County's investment in GIS; cost recovery was implemented. Though this article seeks to focus on substantive legal and practical issues, political influence or policy champions may be defining incentives or disincentives on their own.
Thus, political influence and politics could be a significant incentive or disincentive to cost recovery, depending upon the situation in a particular local or regional government.
Partnerships between GIS agencies and other public or private entities can be beneficial to all participants. Data can be shared to avoid duplication of effort and expense. Costs can be shared to develop a more comprehensive GIS that suits all participants needs, or to simply ease funding constraints in a GIS agency.
One of the strongest arguments for cost recovery is that cost recovery promotes or allows partnerships between GIS agencies and other public agencies or the private sector. Different types of information policy may complicate efforts to form partnerships between GIS agencies. Open access policy may make partnerships more difficult to negotiate for a GIS agency. Because of a GIS agency's open access policy, few proprietary agencies or firms are likely to be willing to share their data or costs. Those proprietary entities could obtain the open access agency's data, without joining a partnership agreement (note the `free rider' problem discussed in King 1993). In one case, a County GIS with an open access policy in Washington State, the County attempted to obtain data from the local utility, but was unable to negotiate a data sharing agreement because any County data would be available to the utility without an agreement. The County was forced to pay $100,000 for the utility's dataset; yet the County is obligated to provide their own data to any utility at the cost of dissemination. Also, because open access GIS agencies disseminate GIS data at marginal costs with no restrictions, proprietary agencies are typically unwilling to share data that might be accidentally released into the public domain by the open access GIS agency. Thus, open access policy can be a disincentive to partnerships.
Conversely, cost recovery may facilitate negotiation of partnerships. With a restrictive cost recovery policy, a GIS agency may better be able to barter data or negotiate cost sharing with potential partners. Indeed, some argue that this strength of cost recovery policy is the most compelling reason for adoption and implementation of cost recovery policy. GIS agencies that are likely to seek partnerships could see cost recovery as a necessary mechanism for successfully negotiating data or cost sharing partnerships. The potential of cost recovery to facilitate partnerships, then, is a significant incentive to some GIS agencies to implement cost recovery.
In spite of the difficulties of developing partnerships for an open access agency, partnerships based on exchange of services and expertise, or coordination of data quality and types could succeed in an open access environment. One open access County GIS agency in North Carolina is engaged in loose inter-local agreements with municipal government GIS agencies for sharing data, applications and expertise. Thus, partnerships can be developed in an open access environment, though the weaknesses of open access policy for serious negotiation and deal-making are obvious.
Another aspect of partnerships is that the line between private data and public records may become blurred. This may be seen as a strength by GIS agencies that seek to implement cost recovery despite open records laws. However, a local government could conceivably lose ownership of its digital GIS data through some partnership agreements. This could result in political costs if private companies are able to use and market local government data, but public interest groups and private citizens are unable to gain similar access. One County in Florida refused to join a GIS data and cost sharing partnership for this very reason: the long-term costs of losing control or ownership of local government GIS data were considered to be greater than the short term benefits of sharing costs of development. While some GIS agencies are able to negotiate agreements that favor the government, many local and regional government GIS agencies may agree to terms that provide significant short-term benefits to the agency and tremendous long-term costs to the agency and the public.
The term `cost recovery' inherently implies that GIS agencies can recover some useful amount of funds by charging for copies of digital GIS data. However, GIS agencies that have recovered a significant amount of money are difficult to find. One cost recovery GIS agency in Tennessee expects to receive $16,000 from digital data requests in fiscal year 1995. In light of their consultant's initial projections of more than $700,000 annually and a target of $5 million of recovery, that amount is unimpressive. Another cost recovery GIS agency in Florida, which charges slightly more than $6000 for their entire digital base map, has recovered only $2800. In contrast, one open access GIS agency in North Carolina recovered $11,000 in 18 months of charging less than $30 per hour for disseminating data that itself was free. Thus, the open access jurisdiction is providing open access to data while generating more revenue through hourly rates for staff time than some similarly sized cost recovery policies. Clearly, some cost recovery policies are not seeing large monetary returns.
However, a County GIS base map agency in California, included in recent master's thesis research, has recovered a significant amount of funds through a partnership with a private vendor. The costs of converting formerly hardcopy maps to updated digital form was about $700,000. The first year of the cost recovery policy yielded more than $100,000, and the second year is expected to yield $1.2 million in returns for the County and the partner. Thus, significant cost recovery is certainly possible, but such a fiscally successful site was extremely difficult to find. Further, this county relies on an interpretation of state public records law that is questionable, and could be eliminated if current state legislative amendments are passed. Even with this amount of recovery, the policy is likely to face difficulties with state law in the future. Despite recovery in one exemplary GIS agency, there are many more others that have not successfully recovered funds. The inherent incentive of cost recovery, to recover funds to supplement funding of a GIS agency, is not an easy achievement. In some cases, recovery will be far below expectations.
Summary of Cost Recovery and Practical Issues
There are significant practical incentives for cost recovery. Just as the specific legal influences on GIS agencies will differ between states, practical influences vary between localities. Politics may be an incentive or disincentive depending on the context of a particular GIS agency. If strong political influence is exerted for any type of policy, politics quickly becomes an important disincentive to adoption of a different policy. Partnerships are clearly a strong incentive for cost recovery policy, in GIS agencies that need to join partnerships. One could argue that the better public policy decision is for all agencies to implement open access policy and share data freely. However, in this real world, some GIS agencies will have proprietary data policies and private interests will naturally protect their own property. Since cost recovery and data restrictions can strengthen a GIS agency's negotiation position with potential data or cost sharing partners, partnerships are a strong incentive for cost recovery policy. Finally, visions of recovering substantial funds from cost recovery can be another strong incentive to adopt cost recovery policy. However, in some cases, cost recovery is minimal or disappointing. The recovery of substantial funds is not a certainty in any GIS agency. Thus, as a practical incentive for adoption of cost recovery policy, profits may often be illusory. In summary, there are practical incentives for cost recovery policy. The weight placed on any particular incentive, or the array of disincentives, will vary between different GIS agencies.
CONCLUSIONS: IS COST RECOVERY WORTHWHILE ?
It depends, but cost recovery does not hold the benefits that many might expect. There are a broad number of disincentives that make implementation of cost recovery problematic. Even if a cost recovery policy is implemented with these legal issues in mind, some of the legal difficulties could arise at any point during implementation of the policy or well into the future. Citizens could challenge contracts or the policies themselves. GIS agencies could have difficulty preventing the spread of their GIS data, regardless of copyright or contract protections. GIS agencies could be subject to liability that could rapidly erase any recovered funds. These possibilities are real costs of cost recovery in local and regional government GIS agencies.
Yet, there can be political and fiscal incentives for cost recovery. If political influence is an incentive for cost recovery, that incentive alone could overcome all of the legal disincentives. Similarly, the expectation of reduced budgetary constraints through partnerships or recovered funds could outweigh the potential legal obstacles. Just as legal disincentives, such as liability, are not certain to occur, neither is a cost recovery policy destined to succeed. Even if some legal problems do not ensue, the practical incentives may not yield practical results. A GIS a