Onsrud, H.J., J.P. Johnson and J. Winnecki, GIS Dissemination Policy: Two Surveys and a Suggested Approach, Journal of Urban and Regional Information Systems, 1996, 8(2): 8-23

GIS Dissemination Policy:  Two Surveys and a Suggested Approach

Harlan J. Onsrud, Jeffrey Johnson, Judy Winnecki
Department of Spatial Information Science and Engineering

University of Maine

Selection of appropriate policies for dissemination of GIS data, products, and services continues to be an important issue for many local government agencies. GIS agencies are currently making dissemination policy decisions with little data or information on the economic, legal and political ramifications of choosing one policy or practice over another. While some progress has bee n made in exploring the practical benefits and drawbacks of different policy alternatives, there is still little information available about the current state of dissemination policy among local government GIS agencies. This article presents empirical evidence of the current state of dissemination policy in US local government agencies, analyzes potential implications of the information compiled from the completed surveys, and suggests an approach for GIS dissemination policy in US local governments that accommodates both open access and revenue-generation objectives.

I. Introduction

A. The GIS Dissemination Policy Problem

Information about the character and location of natural and cultural resources and their relationship to human and economic activities is essential to making decisions about the future. In response to this need, geographic information systems have proliferated rapidly in recent years among all levels of government . With dwindling tax revenues and pressing needs to build and maintain digital data bases, local governments are being encouraged to turn geographic data, products and services into profits. These factors are leading GIS agencies to confront the conflicts and mismatches between revenue-generation policies and open access policies in the dissemination of digital data, products, and services.

Several legal and practical issues are inherent aspects of the dissemination policy debate. State open records laws define the types of information that must be made available and the extent of allowable charges (see Archer and Croswell 1989; Onsrud 1992a ; Onsrud 1992b; Dando 1992; Dando 1993). Liability is a legal concern due to the potential for inaccurate government data or misuse of spatial data by the private sector or citizens (see Anderson and Stewart 1995; Johnson and Dansby 1995, Perritt 1995a). To protect an agency investment in GIS or to minimize liability exposure, some GIS agencies are implementing restrictions on secondary uses of the spatial data supplied by their agencies. Copyright law is one potential restriction on secondary uses of spatial data, albeit a weak restriction (Johnson 1995a, Johnson & Onsrud 1995). In light of the inability of copyright to adequately protect spatial data compilations, contracts and licenses are additional important legal mechanisms for controlling secondary uses of spatial data. These and other legal issues are addressed in many recent publications in the GIS community.

Dissemination policy affects the extent to which data sharing is promoted among GIS agencies and the availability of spatial data for the private sector and citizens. Dissemination policy may influence GIS implementation in local governments. Dissemination policy determines the extent of benefits that the public gains from a government housed GIS and may stifle or promote economic activity in the community. Problems with dissemination policy may adversely affect the operations of a GIS agency by requiring administrative, political or legal activities that reduce the focus on the primary agency objectives. Because of these important effects, dissemination policy is an important issue worthy of further research.

B. Need for a Survey

Despite the importance of the dissemination issue, no empirical assessment ha s yet been available of the state of dissemination policies in local government GIS agencies. Many relevant articles have been published in journals and conference proceedings on why one policy is more suitable than another (Johnson & Onsrud 1995; King 19 95; Lopez 1995; Perritt 1995; Archer 1994; Mechling 1994; Dando 1993; Lawrence 1993; McDermott 1993; Stipek & Toccaca 1993; Blakemore & Singh 1992; Brown 1992; Castle 1992; Dando 1992; Lerner 1992; Onsrud 1992a; Onsrud 1992b; Rhind 1992; Dando 1991; Epstei n 1991; Friedley & Colbert 1991; Tosta 1991; Epstein & McLaughlin 1990; Lawrence 1990; Maffini 1990; Archer and Croswell 1989). In addition to being informed on the issues involved with implementation of alternative information dissemination policies, it seems important that the GIS community know what dissemination policies are actually being implemented. Without an awareness of what policies are actually being implemented, the debate over open access and revenue generation is not sufficiently grounded i n reality. Thus, surveys of local government GIS dissemination policies are needed contributions to the ongoing debate over dissemination policy in GIS.

The first survey reported in this article was conducted in January through February of 1995. A questionnaire was mailed to local and county government GIS agencies throughout the U.S. A second survey conducted in July through August of 1995, took a look at the dissemination policy in a single state, the State of Minnesota. Minnesota was chosen as the single state for comparison with the national survey because of the close familiarity of one of the authors with the government sector GIS community in that state and because numerous Minnesota agencies and local governments have been active in GIS for m any years. This paper presents the results of the national and Minnesota surveys, discusses the implications of the results, and discusses the future of dissemination policy in local government GIS by considering the potential impact of information techno logy, particularly the Internet, on GIS dissemination policy.

II. The Surveys

A. Targeting the National Survey

The national survey was directed at local and county government GIS administrators who are involved with dissemination policy adoption or implementation. We could not locate a comprehensive list of local and county government GIS agencies or GIS administrators. However, the Urban and Regional Information Systems Association (URISA) publishes an annual list of member names, affiliations and addresses. Several hundred of these individuals are involved with local or county government, many of whom use geographic information systems. Most addresses in the 1994 URISA membership directory provide little evidence of whether GIS is being used or not, so the authors refined the membership list to individuals most likely to be dealing with GIS in local and county government. Members involved in URISA s Integrated Systems , Land Records , or Public Information special interest groups (SIG s) were selected from the list, resulting in 184 different local government agencies that could be using GIS. To supplement that list, another 27 local government agencies were located, primarily through personal contacts and published articles referring to operational GIS, for a total of 211 mailed questionnaires.

B. The National Survey Questionnaire

The questionnaire was designed as a simple, brief list of questions that would produce information about the current state of dissemination policy in US local government GIS agencies. Because of the desire for simplicity and to limit the questionnaire to a single page, it contained only ten questions (See Figure 1). We mailed the survey in early 1995, and did not do follow up mailings.

The survey was designed in three basic sections. Questions 1 through 4 are introductory questions that establish the level of government, the storage requirements of the database, and the means by which agencies provide access to their spatial data. Questions 5 through 7 attempt to establish the type of dissemination policy (open access or revenue-generation) by exploring what costs are recovered by an agency in three different ways: the specific charges assessed (Question 5), the agency s definition of those charges (Question 6), and the actual amount charged for a copy of the entire database (Question 7). Through collection of this redundant information, we ensure that the survey provides a clear view of the charges levied by GIS agencies. The third section, Questions 8 through 10, explores liability waivers and restrictions on secondary uses of information. By assessing combinations of responses to questions 5 through 10 we may determine whether each respondent agency is primarily adhering to open access policies or re venue generation policies.

C. The Minnesota Survey: Targeting and Modifications

The Minnesota survey was mailed in July of 1995 to any city, county or state agency that had a contact person on the Minnesota GIS/LIS Consortium mailing list. Ninety-one ( 91) questionnaires were sent to Minnesota government agencies: twenty-eight (28) cities, fifty-five (55) counties and eight (8) state agencies. The Minnesota Survey eliminated one question from the national survey (Question 2) because the answer was already known for this sample. In addition we added three questions. These are shown as items 7, 8, and 9 in Figure 2. Otherwise the questionnaires used the identical wording for each item.

III. Survey Results

This section presents the responses from all usable questionnaires. The implications of the responses are discussed in Part IV.

A. Respondent Characteristics

National Survey: Ninety-six (96) local and county GIS agency administrators responded to the national survey, a response rate of 45.5%. Considering the difficulty in refining the list of potential GIS administrators and the use of a single mailing, we consider this response to be extremely successful. Seventy-nine (n=79) of the respondents reported that they used a geographic information system in their local government agency (Table 1). Of the returned questionnaires, thirty-one were from city or town agencies, forty-three were from county agencies and five were from multi-district partnerships (Table 2). Usable responses were receive d from nearly every state in the United States.

Minnesota Survey: For the Minnesota survey the response rate was 57% with fifty two (52) of the ninety-one (91) surveys returned. Of the fifty-two (52) questionnaires returned, twenty-eight (n=28) reported using a GIS (See Table 1). For these twenty-eight, six were from city or town agencies, sixteen were from county agencies, and six were from state agencies (Table 2).

National and State Surveys: In both surveys we sought and received only one response per government jurisdiction or agency. Thus sixteen county responses in the Minnesota survey represents responses from sixteen different counties. The returns by non-users of GIS were not include d in the subsequent compilations about local government dissemination policy. Thus, the numbers presented in Table 2 and subsequent tables are based on only those agencies that reported having a geographic information system.

B. Storage Requirements

GIS database storage requirements of respondent GIS agencies were widely dispersed in both surveys. Most respondents fell in the one to nine gigabyte and ten to twenty-nine gigabyte ranges. (Table 3)

C. Means of Access

The bulk of the agencies provide access primarily through written and verbal requests. In the national study, fifty-four of the seventy-nine agencies or 68% provide access through verbal requests, five agencies or 6% report using public terminals and six agencies or 8% use dial-up or on-line access mechanisms. Ten agencies or 13% specifically wrote in that they responded to written requests (National Question 4). Similarly in Minnesota, most access was through written or verbal requests. Only one agency (4%) reported using a public terminal, and three, (11%) had some type of dial-up access (Minnesota Question 3).

D. Restrictions on Secondary Uses

The questionnaires ask two questions about restricting secondary uses of spatial data. Table 4 shows the responses to the first question.

In the national survey, sixty percent of respondent GIS agencies (47 of 79) report that requesters are not allowed to distribute the digital data to others and twenty-five percent (20 of 79) do not impose restrictions on further distribution. Fifteen percent (12 of 79) did not respond to the question (National Question 9).

The Minnesota responses were similar but with a larger proportion stating that they do not restrict secondary distribution. In the Minnesota survey, fifty-four percent (15 of 28) report that requesters are not allowed to distribute the digital data to others and forty-three percent (12 of 28) do not impose restrictions on further distribution. Four percent (1 of 28) did not respond to the question (Minnesota Question 5).

A second question asked how secondary use is restricted (Nation al Question 10 and Minnesota Question 6). The responses are shown in Table 5. In both the National and State surveys the means of restricting secondary uses of electronic GIS data are primarily by contract, license, or copyright with approximately 20 to 30% of the sample using one of these methods. Of course, some GIS agencies are using combinations of restrictions to more strongly secure the digital data resources. Those responding in the other category typically indicated that they restrict secondary use not with a signed contract or license, but enclose a letter to the recipient of digital data that states that the data may not be transferred to others. Some of these respondents indicated that they rely on a state statute or local ordinance which bars the further transfer of the data without permission.

It is interesting to note on the national survey that 42% of respondents do not designate a means of restricting secondary use under this question. This is expected since 60 % of respondents said they are imposing restrictions in the previous question and only they would be expected to respond to the last question. The same compatibility between the two questions is not present in the Minnesota survey results. Fifteen respondents say they are imposing restrictions on further transfer in question 5, yet only twelve of those (28-16) are able to indicate explicitly how the restrictions are imposed. At any rate, restrictions on secondary uses of electronic GIS data are being imposed by approximately 50 to 60% of respondents in the two samples tested here. Only 25% of the national sample and 43% of the Minnesota sample are able to respond that they are imposing no restrictions on further distribution of local government GIS data.

E. Liability Waivers

Concern about liability is highly evident, with 61% of the national respondents and 39% of the Minnesota respondents requiring data recipients to sign liability waiver statements. Whether or not liability exposure is a real threat, largely a percept ion, or primarily an unknown, the results of these surveys show that concern with liability by local government agencies is real.

F. Cost of Access

The survey includes three questions about the price of access (National Question 5, 6 and 7 or Minnesota Question 10, 11, and 12). Because the terms used in the questions are subject to interpretation, we phrased each question as a request for a digital copy of the entire GIS database in order to add clarity to the nature of the request. We felt that if a GIS agency indicates adherence to open access principles in response to a request for their entire database it is highly likely that they would respond similarly to responses for smaller portions of their databases. It is interesting to look at the three cost questions together in the context of open access vs. revenue generation policies and within the limits imposed by most state open records laws. Most state open records laws were originally patterned after the federal Freedom of Information Act (FOI A) which generally requires dissemination of government records at the cost of dissemination with no questioning of requesters regarding their intended use of the public records.

For the first question about costs (Table 6), 43% of the national survey respondents and 32% of the Minnesota survey respondents answered affirmatively to one or more of a, b, or c without also responding to d through h. The inference is that these agencies meet the traditional open access test for distribution of digital data at the marginal cost of dissemination or lower.

Anyone checking items d, e, f, g, or h is attempting to generate revenues beyond the cost of dissemination by the definition for "marginal cost of dissemination" as defined by the Federal Freedom of Information Act and as mirrored in many state open record s laws. Thirty-four percent (34%) of national respondents and 50% of the Minnesota respondents implied through this question that they are revenue generation agencies.

It is interesting to note that although 60% of respondents in the national survey are trying to impose restrictions on the further transfer of the digital data (Table 4), only 34% of respondents indicate that they charge beyond marginal cost for the entire database. Thus it appears that a substantial number of those restricting subsequent transfer of data are doing so for other than revenue generation or investment protection reasons.

In the second question about costs, we asked respondents to self define the terminology for the costs the y would charge for the entire digital GIS database. The results are shown in Table 7. As seen from a comparison of Tables 6 and 7, the "self definitions" of the terms for the costs being collected vary from the traditional legal definitions distinguishing "open access" from "cost recovery." It appears that some sites self-identify as partial cost recovery sites when Table 6 shows that they meet the traditional criteria as an "open access" site (43% in Table 6 versus 37% in Table 7 for the national survey and 32% in Table 6 versus 14% in Table 7 for the Minnesota survey).

Finally, to provide greater evidence of the extent of charges, the third question about costs asks for the price a GIS agency would charge for a copy of the entire GIS database. The breakdown of charges in Table 8 vary widely among GIS agencies, from $0 t o $4 million.

These values may be compared to the levels of charges identified previously. Even though in the national survey, 43% of respondents in Question 5 (Table 6) and 37% of respondents in Question 6 (Table 7) indicate that they do not charge beyond marginal costs of dissemination, Table 8 shows that only 12% of respondents indicate that they would charge less than $2,000 for the entire GIS database. Surely for an open access site, the costs of media and other expenses in conjunction with duplicating and disseminating the entire GIS database should not exceed $2,000. Thus, it appears that most of those implying or claiming an open access policy did not respond to this quest ion. Indeed, it is interesting to note that the majority of respondents in the national survey (65%) and many in the Minnesota survey (46.5%) did not answer this question. This seems to indicate that many GIS agencies have not yet confronted or thought through this issue. In another instance of conflict, one agency that indicates charging only "costs of disseminating" under Question 6 would charge $25,000 for a copy of the entire GIS database. Thus, confusion and uncertainty seem to be evident in regard to what agency policies are or should be in response to such a request.

G. Status of GIS Dissemination Policy

1. Clearly Open Access Sites
Assuming that one wants to determine the number of respondent GIS agencies that clearly would be held by a court of law as adhering to open access principles, one should probably select those agencies that charge only for dissemination and duplication cost s (above the double line in Table 6), charge less than $1,000 for a copy of the entire GIS dataset (see nat ional question 7 and Minnesota question 12 - beyond $1,000 clearly would appear to contradict federal Freedom of Information Act (FOIA) provisions and many of the state open records laws patterned after FOIA), and would apply no restrictions on secondary u se of the government GIS data (see national question 9 and Minnesota question 5). Based on meeting all three of these criteria, 24% of respondents in the national sample (19 out of 79) are clearly adhering to open access policies (See Table 9). Twelve of these nineteen sites meeting the open access conditions left blank question 7 regarding the amount they would charge for a copy of the entire GIS database. Although they simply might not know the amount, leaving the response blank might also express misg ivings about distributing the entire database at merely the cost of dissemination. Only two of the nineteen sites stated explicitly in supplementary comments that they are currently reconsidering their open access policy. Although only 5 multi-district p artnerships are included in the survey, none use a clearly open access approach as shown in Table 9. This may indicate that GIS partnership arrangements are likely to promote more restrictive dissemination policies.

Applying the same open access criteria to the Minnesota data, 28% of respondent agencies (8 out of 28) are clearly adhering to open access principles (see the lower half of Table 9). Of this number one of six cities, three of sixteen counties, and four of six state agencies fully meet the open access criteria.

2. Clearly Revenue Generation Sites
In a similar manner, one may determine the number of respondent GIS agencies that clearly adhere to revenue generation principles. For instance, those that charge for more than dissemination and duplication costs (responses d through h in Table 6), would charge in excess of $1,000 for their entire GIS database (see national question 7 and Minnesota question 12), and impose restrictions on secondary use of digital data (see national question 9 and Minnesota question 5) would surely be held to violate open access principles as set forth in the federal Freedom of Information Act (FOIA) and as mirrored in many traditional and typical state open access laws. Based on meeting all three of these criteria , 20% of respondent GIS agencies in the national sample are clearly adhering to revenue generation policies (see the second column of Table 9). This includes 5 of the 31 cities, 8 of the 43 counties, and 3 of the 5 multi-district GIS agencies.

Applying the same revenue generation site criteria to the Minnesota data, 36% of respondent agencies (10 out of 28) are clearly adhering to revenue generation principles. Of this number two of six cities, eight of sixteen counties, and none of the six state agen cies fully meet the "clearly revenue generation" criteria.

3. Mixed or Conflicting Dissemination Policy Sites
The last column of Table 9 shows those GIS agencies that meet one or two of the revenue-generation criteria but not all of them. Thus, those GI S agencies listed in this column have various combinations of both open access and revenue generation characteristics.

This is by far the most interesting column in the national survey. The table shows that 15 out of 31 cities (48%), 27 out of 43 counties (63%), and 2 of the 5 multi-district agencies (40%) have both open access and revenue-generation characteristics. Twenty of the 44 sites are included in this column solely because they require that requesters not distribute digital data to others ( national question 9). These twenty sites are charging only marginal costs or less for the dissemination of data or they left the cost question blank. Except for their response on restricting subsequent use of data, these twenty sites might readily be labeled as "clearly open access sites."

However, sixteen of the 44 sites in this column state explicitly or strongly imply through written comments in response to questions 5, 6, 10 or in the margin of the form that their dissemination policy is still und er development or under active review. The implication of most of these sixteen is that they may be moving further away from open access principles toward revenue generation principles. In addition, 33 of the 44 respondents in this column left the answer to question 7 on how much they would charge for their entire GIS database blank. Again, this may indicate an inability to estimate the dissemination cost but it may also indicate uneasiness in distributing the entire database at the cost of dissemination . Thus, many of the 44 mixed policy sites may be moving toward the "clearly revenue generation" approach.

In summary, the national survey clearly shows that over half of the respondent agencies (56%) are in some manner split between following a traditional open access approach versus a more restrictive government revenue generation approach. The long hand comments on the forms indicate that many agencies are actively considering their information policies.

In the Minnesota survey, the third column of Table 9 shows that 3 out of 6 cities (50%), 5 out of 16 counties (31%), and 2 of the 6 state agencies (33%) exhibit both open access and revenue-generation characteristics. It is particularly noteworthy that two of the five counties in this column indicate that their policies are under active consideration and that four out of the five counties did not respond to the cost question for the entire database. With eight of the 16 counties surveyed already evidenced as clearly revenue generation site s and with a further five in probable transition, there appears to be a strong movement toward revenue generation approaches in county government GIS agencies in Minnesota.

H. Pricing Approaches

It is interesting to look at the relationships among the responses to questions 5, 6, and 7 for those agencies in the national survey following revenue generation approaches. The responses to these questions as well as written comments on the forms indicate a wide range of pricing methods for those jurisdictions choosing revenue generation approaches. Some revenue generation agencies are distributing data at quite low costs but restricting secondary uses. Others would request as much as $4 million for their entire database. In one case, an agency would charge $2.2 million plus costs of media . In another case, an agency is charging $250 per Mb for copies of data. In yet another, the respondent indicates that the volume of GIS data requested is converted to its paper copy equivalent and the cost is then computed at 15c per page plus media cost and postage. In oral remarks, one GIS administrator indicates that he determines the price that private commercial firms charge for similar geographic data and then undercuts that price.

These survey results evidence a lack of consistency in the pricing approaches being used by those agencies pursuing revenue generation approaches and perhaps evidence lack of sound economic theory upon which to base a price under government data sales circumstances. Several of the pricing approaches seem irrational as applied to digital data and others would appear to be harmful to the economic well being of the commercial sector in the community.

I. Market Demand for GIS Agency Data

If GIS agencies are implementing revenue generation policies for the express purpose of fully or partially covering their agency expenses (e.g. data maintenance costs, hardware maintenance costs, database development costs, and capital investments), it is appropriate to investigate the demand for spatial data. The Minnesota Survey asked two questions about demand (Question 7 & Question 8)

The findings show that demand for digital data to date has been quite low from those Minnesota GIS agencies responding to the survey (Table 10). More than half of the respondents report receiving less than one request per month over the previous year with a fourth of those receiving no requests at all. Approximately 9 out of 10 respondent agencies receive less than one request per week for digital GIS data (24 of 27) and only one agency reports receiving more than two requests per week. The highest demand is in metropolitan areas. One reason for the overall low demand may be that digital GIS data in its raw form is not very usable and currently requires manipulation by a value added intermediary to allow the data to be fully utilized by the typical user in the private sector.

Regardless of the pricing structure, low demand results in low revenues (Table 11). Twenty-four out of twenty-eight respondent agencies in Minnesota reported revenues of less than $10,000 in the past fiscal year. Only two counties, both in the metropolitan area with multi-million dollar systems reported revenues over $50,000. One of these reported over $100,000 in revenues in the past fiscal year but also reported this to be an unusual year, with revenues historically in the $20,000 range. Considering the costs incurred by the GIS agencies in instituting and servicing revenue generation approaches, the argument that charging for data allows a GIS agency to reduce the burden on the local taxpayer by charging users is unsupported by these survey findings. The costs recovered in proportion to the overall GIS agency annual budgets typically appear to be negligible or offset by the costs to service the revenue generation policy.

IV. Reflections on the Survey Results

The authors are surprised and a bit dismayed by the results of the surveys. We have tracked the "open access" versus "cost recovery" debate in the GIS community for many year s and have helped articulate the policy arguments on both sides of the debate (e.g. Onsrud 1992a and Onsrud 1992b). After evaluating the competing arguments on most of the issues, we have arrived at information policy recommendations for GIS agencies that are at odds with the possible trend evidenced by this survey work toward more restrictive government information policies in local and county GIS agencies. We believe that a middle ground approach is possible that largely achieves the goals of both open access and revenue generation approaches. Yet we have seen few jurisdictions using such an approach.

A. Economic Effects

Long-term economic vitality for a community suggests the avoidance of government monopolies over GIS data (Perritt 1995a, Lopez 1995) and yet the present surveys evidence widespread attempts by local and county governments to control subsequent uses of geo graphic data and transfers to third parties. This restrictive information practice eliminates much of the potential for experimentation and innovation by private individuals in competing with each other in offering value-added products and services to the community at large. The practice also portends an adversarial relationship between government GIS agencies and local businesses, similar to the evolving situation in some of the European nations, in which lawsuits by government against businesses over information products become more commonplace and overall economic vitality is harmed (CEC 1993, CEC 1995a, CEC 1995b, CEC 1995c , IMO 1995).

For an information economy to thrive, government information policy must encourage a diversity of information producers and distributors and permit the disaggregation of production so that a variety of value-added products can be introduced into the market (Perritt 1994, Perritt 1996). Restricting government data by imposing contractual and licensing conditions on private parties prior to its use in the market stymies diversity and disaggregation of production. In such instances, the economy suffers as a result of placing the power to add value to government information in a fewer number of privileged hands. In contrast, government policies that promote reuse of public information resources without imposing government intellectual property rights in the public information lead to greater innovation in the private and public sectors.

The argument is sometimes made that government data must be licensed or contractual conditions must be imposed to ensure that only the most up-to-date and accurate government data sets are used by the public and private sectors. However, less economically harmful methods exist for accomplishing these goals. For instance, causing a ban in the use of a government trademark to toll upon some condition, such as availability from the government of a replacement data set and publication that the government's requirements for timeliness and accuracy are no longer met in the prior dataset. Those desiring use of the government's trademark in conjunction with data would need to return to the government for the updated data. This and similar alternative arrangements often exist for accomplishing government's goals without stifling the potential for increased diversity of value-added offerings in the market.

B. Liability and Additional Legal Issues

Although the surveys evidenced strong concern about liability by local and county GIS agencies, the realistic extent of liability exposure in a GIS and spatial data context is unclear. The authors found no case law that directly addresses government liability in the direct context of local or state government geographic information systems (Johnson & Onsrud 1995, Johnson & Dansby 1995). However, it is probably safe to conclude that government GIS offices assume at least some liability exposure in collecting and disseminating land-related data, whether or not they compete with the private sector in the sale of GIS data (Dansby 1992, Anderson & Stewart 1995). Liability exposure may not be eliminated; it may only be reduced. I n addition, whether the signing by a recipient of a liability waiver statement prior to receiving government GIS data has any effect in law is not yet clear. If the data recipient had a prior right through an open access law to access and copy the data, subsequent conditions imposed by an agency through contract, license, or administrative action may not be enforceable.

It is probably safe to say also that those GIS agencies or political subdivisions choosing to enter or compete with the commercial market in selling geographic information data, services or products incur greater liability exposure than those government jurisdictions that do not, regardless of the business practices and contract language used by the GIS agency to minimize exposure (Johnson and Dansby 1995, Perritt 1995b). As previously stated, whether or not liability exposure is a real threat, largely a perception, or primarily an unknown, the concern with liability by local and county government GIS agencies as evidenced by the instant surveys is real. Because the concern is real, the liability ramifications in moving from an open access approach to a revenue generation approach for the dissemination of GIS data should be thoroughly considered

The discussion of open access versus revenue generation approaches is incomplete without also discussing the possible and probable effects of open records laws and their exceptions, copyright, contracts, and licenses. Detailed legal issue discussions regarding sales of government data are be yond the scope of this article but are adequately addressed elsewhere. A recent analysis indicates that the legal issues related to GIS dissemination policy strongly mitigate against revenue generation practices (Johnson & Onsrud 1995).

It has also been argued by some that government should restrict access to government's spatial data in order to protect personal information privacy. The logic of the argument purports to be that government should restrict access to government geographic data in order to hamper the widespread use of detailed geodemographics for customer profiling by the private commercial sector. But, of course, selling government data increases privacy problems and thus has the opposite of the desired effect. Information insiders such as government agency personnel and those large businesses that can afford to buy data gain access to the information that government collects on individuals, households, and property. At the same time, selling of government data decreases the typical citizen's ability to determine the extent of personal information being collected since full access to the data and experimentation with the data using modern computational and processing techniques may only be acquired through purchase. If protecting information privacy is truly a concern, the issue should be dealt with directly and not through a means that increases the gap between the "information have's" and "information have not's." (Onsrud, Johnson & Lopez 1994)

C. Problems Among Government Agencies Due to Dissemination Policy Conflicts

Results of the survey show that there is no current consensus about dissemination policy among local government GIS agencies. Some local government GIS agencies are implementing revenue generation policies while others are continuing to support open access principles. Related work indicates that substantial differences in dissemination policy among agencies creates barriers to spatial data sharing, potentially inhibiting efficient implementation of GIS in some agencies (see Johnson 1995b). As an example, recent case study research identifies one situation where a county GIS agency with an open access policy is having difficulty sharing data with a city GIS agency imposing revenue generation policies within that county (Johnson 1995b; Johnson & Onsrud 1995). Because of differences in philosophies and policies between the agencies, the city indicates that it may take the county's data at will since the county follows an open access philosophy and yet the city refuses to freely share data with the county. The County is more than willing to share its data and work cooperatively with the city yet the city GIS agency demands that the county pay $80,000 per year or a one-time payment of $700,000 for its data. For about eighteen months, the county refused to pay for what it considers to be public information, but has recently decided to pay a negotiated fee for the data. In this case, differences in dissemination policy in the same jurisdiction (with many of the same taxpayers) are inhibiting implementation of GIS at the county level. The county is giving serious consideration to implementing a revenue generation and data contract approach as a means to prevent the city from "free riding" on its data and to coerce cooperation even though the county believes such an approach to be unwise for the overall interests of the community. Such a situation demonstrates the problems that can occur due to conflicts in dissemination policy within a single jurisdiction. Similar problems can be expected in city to city, county to county, and region to region data sharing.

Recent research (Johnson 1995b) as well as the surveys reported in this article indicate that the expected fiscal benefits of revenue generation are unlikely to be achieved by most local or regional government GIS agencies. Due to low revenue returns, one of the strongest practical incentives for following a revenue generation approach, as discussed above, is the improved negotiating position for data or cost sharing partnerships with other government agencies that can result from imposing a revenue generation policy (Johnson 1995a & b). However, imposing broad restrictions on information in order to address a specific problem may be short sighted, particularly if alternatives exist for pressing an agency's negotiating position. "Free riding" by other government agencies should be dealt with in the political arena or with very narrow legislation focused explicitly on that problem.

As local and regional government GIS agencies proliferate and accumulate greater amounts of spatial data, much of our nation s spatial data will be affected by local level GIS dissemination policy. An individual agency may perceive that restrictive information practices and revenue generation are in that agency s best interests. However, if each agency constrains access to their spatial data in expectation of revenue generation or in expectation of increased data or cost sharing among agencies, the amount of information available in t he public domain will be greatly reduced (Lopez 1995). It is in the best interests of the public domain of spatial data, upon which the public and private sector GIS communities depend, if local government dissemination policies promote free and open use of spatial data by public agencies, citizens and the private sector.

V. Accommodating Both Open Access and Revenue Generation Concerns

We recommend adherence to the following principles in establishing information dissemination practices for a public agency. We believe these principles allow government GIS agencies to be responsive to private citizen and business requests for GIS data and processing results while avoiding a financial burden on the public agency in meeting those requests. The approach avoids most of the policy and legal criticisms voiced against GIS agencies currently following revenue generation policies and yet allows GIS agencies to charge at substantially higher rates for value-added services than many GIS agencies are currently ch arging. The approach creates a level playing field in that it eliminates the possibility of government monopolies in geographic information and allows private firms to compete on an equal footing with government's GIS agency for-profit operations. We believe the approach fully meets the goals and intents of most state open records laws and encourages a competitive environment that should enhance long term economic vitality in the community and stimulate increased use of governments' geographic data. The approach is summarized by Figure 3.

Principle 1. Government GIS Data
GIS digital data files accumulated for government purposes (illustrated by the lower base layer in Figure 3) should be treated as public records and made available to any requester at the cost of dissemination. These dissemination charges would typically include a reasonable approximation for the costs of distributed disks, tapes, or other media, pro-rated hardware and software expenses based on the proportion of public versus government use of the GIS, mailing expenses, and the staff time necessary to copy the material and ship it. No restrictions should be imposed on the use or further dissemination of this digital data.

Principle 2. Government GIS Data Products
GIS data products created by and for government (illustrated by the upper base layer in Figure 3) should be treated as public records and made available to any requester at the cost of dissemination. No restrictions should be imposed on the use or further dissemination of the data products.

Principle 3. Public and Private On-line Services
If a GIS agency chooses to offer access through a networked server, the GIS agency should make available to the public and maintain in electronic form the data and data products illustrated in the two base layers of Figure 3 as well as any indices and metadata files used by government in accessing those digital files. This information should be made available to the public through the largest nonproprietary cooperative public computer network available (e.g. the Internet). Because maintaining a GIS data server would substantially benefit other government agencies in increasing their access to the GIS data and because the marginal cost to additionally allow private users to download data from such a server would be negligible, we recommend that no fee should be charged to the public as a condition of accessing this information through the network.

The GIS agency may offer a fee-based electronic information service that provides value-a dded services such as on-line GIS software processing capabilities, automated notice that indicates files have been updated, various means of searching full text documents, metadata, and spatial data, and similar value-added services. As illustrated by the right side of the center column in Figure 3, the agency may charge at any rate it desires for such value-added services and impose any contractual conditions it chooses because the value-added services are for private needs and not for government purpose s.

The establishment of on-line services by a government GIS agency should in no way prohibit private companies or individuals from accessing the two base layers of information shown in Figure 3, either electronically or otherwise, to provide, either commercially or on a voluntary basis, services similar to the public GIS agency on-line services.

Principle 4 - Public and Private GIS Service Centers
The vast majority of citizen and private business requests are not for the raw GIS data or digital data products already existing in government files. Instead, most current requests are for specific data processing services to extract information for a private purpose. User fees set at a profit rate for the GIS agency may and probably should be charged for these data processing services.

For digital data processing services taking only a few minutes to respond to or those that require only the copying or printing of an existing file, cost of dissemination fees such as those itemized earlier should be charged. However, services that involve more than a few minutes to respond to and that involve the processing of data to create new files should be billed at profit generation rates (see the right side of Figure 3). Reasonable profits could probably be generated by charging at a minimum rate of two or three times the total costs of labor and materials required in responding to the request. Because the resultant value-added products are for private rather than public purposes, the GIS agency should also fee l free to individually negotiate or impose contractual conditions and intellectual property rights in the resultant value-added products.

The establishment of a Public Agency GIS Service Center (profit center) should in no way prohibit private companies or individuals from accessing the two base layers of information shown in Figure 3, either electronically or otherwise, in order to provide, either commercially or on a voluntary basis, services similar to the Public Agency GIS Service Center.

We believe t he above approach meets most of the open access and revenue generation objectives currently being pursued by local and county government GIS agencies. The potential for profit generation lies primarily in the provision of value-added services and products . The outlined approach allows and encourages government to generate revenue and profits by providing such services and products if it so chooses, particularly where demand for the products and services is not being met by the private sector information industry. Yet the approach also creates a level playing field. Because the approach comports with the open records laws of most states, there is probably no need to gain legislative authority in most instances in order to implement the approach. The approach also appears to meet the "draft principles for public information policy" advocated by Perritt (Perritt, 1996, pp. 498-499).

If a Public GIS Service Center using public employees and resources is operating effectively and efficiently in providing value-added GIS services and products, it is unlikely that private firms will choose to compete with the government-housed GIS for-profit operation. However, if the operation operates inefficiently or offers very limited value-added services, competing private firms have equal access to the underlying digital public records and are free to compete with the government-housed profit center.

Note that the above approach avoids partnership arrangements between a government GIS agency and one or two private firms. Such arrangements typically result in restrictions on third party use and dissemination of the two base layers of information shown in Figure 1. The principle of no restrictions on subsequent use and dissemination of the two base layers of Figure 3 is key to our suggested dissemination approach.

In discussing dissemination policy with GIS administrators, one of the primary concerns appears to be efficiency of their operations. The time, effort, and costs to respond to citizen and business request s is a substantial burden on government GIS agencies. Using on-line access mechanisms in the manner indicated above should greatly reduce this burden. The agency simply places GIS digital data files and indices on a networked server and lets the public and businesses access and download the files for themselves either at a public terminal, or if sufficient bandwidth is available, through the network. As use of networks becomes more pervasive over time and bandwidths improve, daily burdens on public GIS agency staffs in responding to requests should decrease.

Agencies will inevitably have the technical ability to charge for data over the Internet, through the use of digital cash or credit cards. Thus, charging for government s value-added spatial data services and products will become easier as well. However, GIS agencies should be aware that proprietary value-added data products are likely to become much more difficult to control in the Internet environment, where data can be easily and instantaneously transferred to innumerable individuals, agencies and firms. Even with widespread use of encryption on the Internet, maintaining proprietary interests in factual data that changes little over time, such as spatial data, will be exceedingly difficult due to marketplace, legal and practical difficulties (Onsrud 1995; Johnson & Onsrud 1995). In sum, enforcement of claims of proprietary interests in value-added GIS products will continue to be a major challenge, whether the product is produced by a Public GIS Agency Service Center or a private firm.

VI. Summary

Local and regional GIS agencies throughout the US are confronting a myriad of legal and practical issues associated with dissemination policy. At the same time, state and national efforts are pursuing greater sharing of information through standards efforts, clearinghouses and digital libraries. A lack of information about the current state of GIS dissemination policies among local and county agencies caused the authors to conduct a survey of dissemination policy issues in local government GIS agencies. With nearly an even split among the survey respondents between "clearly open access" and "clearly revenue generation" approaches and with as many or more respondent agencies exhibiting characteristics of both approaches, the authors observe that information policy consensus is lacking among local and county government GIS agencies.

Analysis of the numeric responses to the questionnaire (Table 9) and the numerous supplementary comments on the survey forms indicate that many local and county government GIS agencies have yet to make steadfast decisions regarding the directions their long term information policies will take. The benefits and drawbacks of open access versus revenue generation approaches are still being actively debated and experimentation with variations of the approaches is evident. While some states have enacted legislation that enables local and county government GIS agencies to use revenue generation approaches, if they so choose, other states are exhibiting strong legislative support for strengthening open access principles and increasing public and business access to government's data. If there is any trend evident in the survey findings reported here, the trend is more toward revenue generation approaches than open access approaches at the current time. However, the revenue generation approaches themselves are not in conformance with each other and exhibit a mix of secondary restrictions, pricing methodologies , actual charges and legal justifications. Observation in real cases of the benefits and drawbacks of following differing information policies would be valuable in providing a firmer grasp of the practical and legal ramifications of following one approach over another.

In summary, this article provides empirical evidence of the current state of dissemination policy in US local and regional government GIS agencies. As more counties and cities implement comprehensive geographic information systems, more spatial information will be impacted by GIS dissemination policy. The authors argue that conflicting GIS dissemination policies across jurisdictions and levels of government will hinder the beneficial uses of GIS by constraining the availability of spatial data in the nation s public domain. We also argue that compromise positions supporting many of the underlying goals of both revenue generation and open access approaches may be achieved. The current deadlock between the two policies is likely to be broke n over time as local and regional government GIS agencies gain more experience with the legal and practical disincentives to maintaining some of the more restrictive aspects of revenue generation approaches. The Internet (or some future incarnation thereof) will have a dramatic effect in significantly easing the burden of providing public access. Agencies will be able to allow access to and copying of data with little or no direct staff involvement. With such benefits in efficiency, and with far fewer le gal problems, open access is likely to gain favor as the optimum policy for local government GIS agencies in the long term. We argue that support of open access principles is extremely important to economic vitality of the community. Numerous studies in the last fifty years, and additional ongoing studies, have shown that increased diffusion of scientific and technical information results in greater innovation in a wide variety of applications ranging from basic science to the provision of commercial services (Lopez, 1995). One need only look at the TIGER files from the Census Bureau, DLG data from the US Geological Survey and GPS from the Defense Department as examples of how an open access policy can lead to greater efficiency for secondary users and can result in a wealth of new business applications.

However, many of the goals of a government agency in pursuing revenue generation approaches, such as providing value-added services to meet private needs and generating profits from these value-added services, may also be achieved if a GIS agency so desires without violating open access principles or stifling economic activity in the private sector information industry. We offer one such approach that we believe meets these objectives.

VI. Acknowledgments

This article is based on survey work at the University of Maine accomplished by Jeffrey Johnson as part of a master's degree program and Judy Winnecki as part of her work supported by a Bush Fellowship. Harlan Onsrud advised the work and is Chair of the Department of Spatial Information Science and Engineering at the University of Maine. Jeffrey Johnson is currently a student at Stanford Law School and Judy Winnecki has returned to her position as the Principal Land Surveyor for the Minnesota Department of
Natural Resources.

This work is based upon work partially supported by the National Center for Geographic Information and Analysis (NCGIA) under NSF grant No. SBR 88-10917. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

VII. References

Anderson, Jerome R. and Alan R. Stewart, 1995. "Local Government Liability for Erroneous Data: Law and Policy in a Changing Environment," in Proceedings of the Conference on Law and Information Policy for Spatial Databases , Tempe, AZ: NCGIA, University of Maine, pp. 267-279.

Archer, Hugh (1994). Expanding Affordable Public Access to Information: Where do we go from here? URISA Journal : 5- 12.

Archer, Hugh and Croswell, Peter (1989). Public Access to Geographic Information Systems: An Emerging Legal Issue. Photogrammetric Engineering and Remote Sensing. November 1989: p. 1575 - 1581.

Blakemore, Michael, and Gurmukh Singh (1992). Cost Recovery Charging for Government Information. A False Economy? Gurmukh Singh and Associates Ltd.

Brown, Kay (1992). In My Opinion: A Response to Earl Epstein. URISA Journal 4.1: 6-8.

Castle, Gilbert (1992). A Conceptual Framework For Setting Prices For Local Government Information Related Products and Services. Annual Conference of the URISA . Washington, DC.: URISA. 4: 26-32.

CEC, 1993, A Report to the Commission of the European Communities on an Evaluation of the Implementation of the Commission's Guidelines for Improving the Synergy between the Public and Private Sectors in the Information Market. PUBLAW 2 Europe Final Report . CEC Directorate General XIII. Luxembourg: CEC .

CEC, 1995a, INFO2000 Draft, Council Directive. COM(95) Brussels: Directorate General XIII

CEC, 1995b, GI2000: Towards a European Geographic Information Infrastructure (EGII). Discussion Document for Consultation with the European GI Community. Brussels: Directorate General XIII.

CEC, 1995c, Workshop on Commercial and Citizens' Access to Government Information. Luxembourg, 26-27 June, 1995. Luxembourg: European Commission Directorate General XIII.

Dando, Lori Peterson (1991). Open Records Law, GIS, and Copyright Protection; Life After FEIST. Annual Conference of the URISA . San Francisco, CA: URISA. 4: 1-17.

Dando, Lori P. (1992). A Survey of State Open Records Laws in Relation to Recovery of Data Base Development Costs: An Ends in Search of a Means. Annual Conference of the URISA: Washington, DC. Vol. 4. p. 85-110.

Dando, Lori P. (1993). Balancing Cost Recovery, State Open Records Laws URISA News , No. 133, p.1-4.

Dansby, H. Bishop (1992). Public Records and Governmental Liability. GIS Law . Vol. 1, No. 2. p. 8-13.

Drucker, Peter F. (1968). The Age of Discontinuity. New York: Harper & Row

Epstein, E. F. (1991). In My Opinion. URISA Journal 3.1: 2-4.

Epstein, Earl F. , and John D. McLaughlin. (1990). A Discussion on Public Information: Who owns it? Who uses it? Should we limit access? ACSM Bulletin . 128: 33-38.

Freedom of Information Act, Pub. L. No. 89-487, 80 Stat. 250 (1966), Pub. L. No. 90-23, 81 Stat. 54 (1967), Pub. L. No. 93-502, 88 Stat. 1561 (1974), Pub. L. No. 99-570, 100 Stat. 3204-48 (1986)

Friedley, Dale, and Larry Colbert (1991). Reaching for Consensus in Public Information Access Policies. Annual Conference of the URISA . San Francisco, CA: URISA. 4: 50-64.

IMO, 1995. Geographic Information Systems in Europe: Problems and Potential, Information Market Observatory, Working Report 95/2. CEC DG-XIII. Luxembourg.

Johnson, Jeff P. (1995a) Information Policy for Local Government Geographic Information Systems: Adoption, Implementation and Results . M.S. Thesis. University of Maine. Spatial Information Science & Engineering. pp. 263

Johnson, Jeff P. (1995b). Case Studies of Dissemination Policy in Local Government GIS Agencies . Computers, Environment and Urban Systems (behind in publication)

Johnson, Jeff P. and H. Bishop Dansby (1995). Liability in Private Sector GIS . Proceedings of the Conference on Law and Information Policy for Spatial Databases : Tempe, AZ. NCGIA, University of Maine, Orono. pp. 285-292.

Johnson, Jeff P. and Onsrud, Harlan J. (1995) Is Cost Recovery Worthwhile? Proceedings of the Annual Conference of the URISA. San Antonio, TX: URISA 1: 126-136.

King, John L. (1995) Problems in Public Access Policy for GIS Databases: An Economic Perspective In Sharing Geographic Information, Ed s Harlan J. Onsrud and Gerard Rushton. CUPR: New Brunswick, NJ. pp. 255- 276.

Lawrence, Jeff (1990). The Economics of Pricing GIS Products. Annual Conference of URISA . Edmonton, Alberta: 2 : 81-87.

Lawrence, Jeff (1993). Pricing GIS Participation and Products. Annual Conference of the URISA . Atlanta, GA: URISA. 2: 165-172.

Lerner, Nancy B. (1992). Developing a Successful Information Access Policy. Annual Conference of the URISA . Washington, DC.: URISA. 4: 79-85.

Lopez, Xavier R. (1995). From Gravel to Diamonds: The National Spatial Data Infrastructure at a Crossroads . Proceedings of the Annual Conference of the URISA. San Antonio, TX: URISA 1: 611-625

Maffini, G. (1990). The Role of Public Domain Databases in the Growth and Development of GIS. Mapping Awareness 49-54.

McDermott, Twyla (1993). Charging System for GIS Products and Services. Annual Conference of the URISA . Atlanta, GA: URISA. 2: 182-192.

Mechling, Jerry (1994). The Hidden Data Pricing War. Governing October: 82.

Onsrud, Harlan J. (1995). The Role of Law in Impeding and Facilitating the Sharing of Geographic Information. in Sharing Geographic Information, Harlan J. Onsrud and Gerard Rushton, eds. New Brunswick: New Jersey, CUPR Press. p.292- 306.

Onsrud, Harlan J. (1992a). In Support of Open Access for Publicly Held Geographic Information GIS Law . Vol. 1, No. 1., p. 3-6.

Onsrud, Harlan J. (1992b). In Support of Cost Recovery for Publicly Held Geographic Information GIS Law . Vol. 1, No. 2., p. 1-7.

Onsrud, H.J., J. Johnson, and X. Lopez (1994). "Protecting Personal Privacy in Using Geographic Information Systems" Photogrammetric Engineering and Remote Sensing , Vol. LX(9), pp. 1083-1095.

Perritt, Henry H. (1994). Commercialization of Government Information: Comparisons between the European Union and United States Internet Research. Vol. 4, No. 2., pp. 7-23.

Perritt, Henry H. (1995a). Should Local Governments Sell Local Spatial Databases Through State Monopolies? Proceedings of the Conference on Law and Information Policy for Spatial Databases. Tempe, AZ: NCGIA, University of Maine. pp. 52-72.

Perritt, Henry H. (1995b). Tort and Criminal Liability as a Function of the Right to Control Content Proceedings of the Conference on Law and Information Policy for Spatial Databases. Tem pe, AZ: NCGIA, University of Maine. pp. 313-315.

Perritt, Henry H. (1996). Law and the Information Superhighway (Somerset, NJ: John Wiley & Sons, Inc.) pp. 81- 82, 498-499.

Rhind, D. (1992). Data Access, Charging and Copyright and their Implications for geographical information systems. International Journal of Geographical Information Systems 6.1: 13- 30.

Stipek, Lawrence, and David Toccaca (1993). Implementing and Assessing the Value of Direct Public Access to the Loudon County, Virginia Geographic Information System. Annual Conference of the URISA . Atlanta, GA: URISA 2: 202-208.

Tosta, Nancy (1991). Public Access: Right or Privilege. Geo Info Systems Nov/ Dec: pp. 20.