Why should government spatial data be free?

Executive summary

(links are within this page)

Digital geographic data is used to the point of being essential by most governments in Canada for mapping and geographic analysis. Decision makers use government geographic information to help decide where to create a new park, which hillside is to be logged, or where a pipeline or highway should go. This information represents a valuable public asset, just like roads, schools, hospitals, or ferries. Since the 1980's the general trend has been towards cost recovery of government information. At first glance, this policy appears to be a good way of sharing the costs of a government service among those who benefit from it, like a toll booth on a bridge or a user fee on an ambulance. However, there are significant problems with this approach. Chief among those are:

So we hope we have given you some food for thought and perhaps spurred you to action. Feedback is welcome.

Why should government spatial data be free?

Digital Geographic Data

Most governments in Canada use computers for mapping and geographic analysis. They have large amounts of digital (computerised) geographic data stored in their databases, which have become essential to government business.

Decision makers use government geographic information to help decide where to create a new park, which hillside is to be logged, or where a pipeline or highway should go. Cities are planned and managed using geographic information. Search and rescue groups need geographic information to effectively search for missing persons, academics need it for research, schools need it for learning, and companies need it to make business decisions and develop value added products and services. As well, governments need each others' data.

This information represents a valuable public asset, just like roads, schools, hospitals, or ferries. However, digital information does not ?wear out? with use like these physical objects. On the contrary, widespread use of digital geographic information increases its value to the government and society by encouraging a consistent framework for mapping locations of objects in the world.

Cost Recovery Policy

The Mulroney government started a trend in Canada towards ?cost recovery? for government information in the 1980?s and 1990?s. (http://www.usask.ca/library/gic/18/prophet2.html). Digital data necessary for the conduct of government has been marketed and sold like a free market good, in an attempt to recover the costs of its creation and maintenance.

At first glance, this policy appears to be a good way of sharing the costs of a government service among those who benefit from it, like a toll booth on a bridge or a user fee on an ambulance. However, there are significant problems with this approach.

Public policy problems with cost recovery for digital spatial data

Public access is denied

The greatest benefits to society from geographic data may come from a large number of marginal users ? those who derive an incidental benefit from it. For example, a backcountry guide could use digital data to create a detailed map on a web site to advertise her tours. An elementary school can create a detailed map of fish counts in nearby streams. An environmental group can accurately map locations of rare oak trees for a regional district planning process. Search and rescue volunteers can take a laptop computer with detailed contour maps to plan a helicopter evacuation. These are all users who could not afford to purchase digital topographic data at its current price to the public, and might not be able to purchase it at any price.

We believe that geographic data should be compared with electronic delivery of other government information. For example, weather forecasts could be sold to each individual user who required them. It costs money to pay meteorologists, so perhaps their salaries could be funded by whoever needs to predict the weather. The result would be that businesses such as airlines, farmers, and house painters would pay, and normal citizens would take their chances with their picnics. Perhaps there would be more hikers lost in unexpected snowstorms. The point is that free distribution of information about our environment can be a huge benefit to society, even though it is difficult to measure economically.

Government is balkanized

When data is exchanged freely between different levels of government, each can use the information it needs however it sees fit, and there is an incentive to reduce costs by avoiding duplicate data collection. This encourages consistency and cooperation between governments.

, outlines one way this situation can deteriorate. It begins when one individual government may see an advantage in selling its data to the others, increasing its own revenues. Other governments will then begin charging as well, to at least provide bargaining power when dealing with the first. Soon, every government is charging, with no net benefit except to the accountants recording the transactions.

A side effect of these charges is that the public is cut out of the deal. The data charges are a barrier to local groups and individuals that need access to spatial data to provide input to local planning or help with local government, but don?t have taxpayer funded data to trade.

This situation is happening now in Canada.. Access to topographic and cadastral mapping is not free to local governments, who have become reluctant to share their own private cadastral mapping with senior governments. Sometimes a local government will share its data with one branch of a provincial Ministry, but not others.

Charging is expensive

For example this quote from a letter by Andrew Hubbertz in June, 1996 about the Canadian Atmospheric Environment Service:

How successful has AES been in their efforts "to recover some portion of the costs of collecting, quality assuring and archiving the data"?

According to information obtained under the Access to Information Act, the cost of developing this CD-ROM product since 1991 was $292,000, not counting the cost of collecting the data itself. As of May 1995, sales of the product came to $16,291.

A similar CD-ROM product compiling streamflow data cost $247,400 and returned $47,102.

Despite disappointing sales, AES remains committed to revenue generation from these products. This is all the more baffling if one considers that the federal government, through the Green Plan, is spending $3 billion over five years on environmental research and development. It is unclear how research will be advanced through restricting access to basic data.

The KPMG study cited below found that is rare for a government agency to make a significant profit from its data sales, if revenues from other levels of government are not included.

Opportunity is lost

New techniques are putting digital spatial data to many uses beyond environmental mapping. It is possible to look for correlations between fields and trends as varied as health problems, urban pollution and agricultural practices. It can be difficult to make a business case for spatial data purchase for this type of research when its outcome was very uncertain.

The cost of computer software and hardware for spatial analysis is falling rapidly, but the cost of the data that feeds them is not. This discourages innovative, experimental research. If it can?t promise a benefit to justify the data cost, it won?t be funded.

Liability may be increased

Currently, most government data has a standard disclaimer, that the user must not use the information for any business purposes, etc. However, if there were serious errors in government data, which did cause personal or economic hardship, the case against the supplier would be much stronger if the data had been sold to them in a commercial transaction. It is very difficult to sue a Canadian government, but this has been cited as a consideration in the United States.

Policy is inconsistent

Some governments (e.g. the Canadian federal government) have broad policies on information pricing, they are not followed consistently. It seems that senior managers get to choose whether ?their? data is to be subject to cost recovery or not.

To quote from the KPMG Geospatial Policy study referred to below:

Many pricing policies are inconsistent in their implementation across agencies in the same level of government, or even within the same departments. Many policies have competing priorities.

Technical problems with cost recovery for digital spatial data

Fair pricing is difficult

The costs of collecting geographic data can be very large. For example, the TRIM program to collect 1:20,000 scale topographic mapping for all of British Columbia cost about tens of millions of dollars.

From the point of view of the single agency responsible for these costs, it is reasonable to charge high prices for copies of the data. them, as has been done with 1:20,000 topographic mapping in British Columbia. The digital data sells for $400.00 per mapsheet (approx. 10 km x 15 km). 7,000 mapsheets cover the area of B.C. Even with a good discount, these high prices restrict access to only ?big players? who need it for justifiable business reasons, such as senior governments or large corporations. On the other hand, low prices don?t bring enough revenue to cover accounting costs or get a ?fair share? of proceeds from big players.

There is no price that is fair to both an elementary school student and a large agribusiness company.

The root of this problem is that the product is intangible - once it has been created, extra users cost no more. Digital data does not ?wear out?. There is no incremental cost of more users.

Distribution is cheap

In the Internet age, digital data can be distributed extremely cheaply. It can be put on a network server computer where anyone can access it or download it with virtually no transaction cost. There is no point in charging for the distribution process, because managing the charges would cost more than the transfer itself.

Metering is impossible

Unlike electricity or vehicles, it is virtually impossible to measure how often or for what a copy of digital data is used. Once access is granted, there is no reasonable method of charging by ?amount of use?.

These are new to the digital age

Up until the 1970?s, spatial data was all on paper maps. Producing high quality maps was an expensive process involving hand drawing and complex multi-colour printing. It was very practical to sell paper maps, because the purchase price was significantly lower than the cost for most purchasers to make an individual copy.

It is only in the past fifteen years or so that computer power and storage capacities have increased enough to make it cheaper to manage large amounts of spatial data digitally. Digital data can be copied at no cost without any loss of quality, changing the fundamental economics.

It is only in the last decade that computer networks have become fast enough to transmit large amounts of digital spatial data electronically. This eliminates the requirement for any physical media (diskettes, tapes) so that digital data can be distributed instantly at a cost which is so small as to be immeasurable.

The result is that topographic data for all of BC, which cost $70 million dollars to compile over a decade, can be stored on a $200.00 disk drive, and transferred in a few hours on a high-speed line. This was not imagined even ten years ago.

Spatial data collection is also becoming much cheaper. In the 1970?s, human technicians would laboriously interpret information from stereo air photographs. Now, satellite imagery, Global Positioning Satellites, radio collars, and other electronic technologies capture data far more cheaply and directly.

How it costs government

Setting up the store

Governments are not often good entrepreneurs. When a manager or executive proposes to begin selling spatial data, it is tempting to spend a lot of money 'up front' to create an elaborate sales mechanism based on promises of future revenue.

Unlike real companies, there is no real risk or accountability for the costs. If sales don't materialise as expected, the government will not go out of business. Ministers change, and these costs get buried in historic records.

For example, the B.C. government spent approximately $11 million over 9 years on software development for the LandData BC distribution mechanism. The ongoing sales were not sufficient to amortise this cost.

Data sharing agreements

Governments in Canada are writing data sharing agreements with other levels of government (federal, provincial, American state, regional district, and municipal) on an ongoing basis. These require management time to negotiate, draft, execute, and administer. They are required to ensure that each government does not ?give away? its data without getting something else in return. None of them contribute new funds or services to the taxpayer.

How it costs other organizations

Cost to high technology industry

According to many people in the Geographic Information Systems field, Canadian data pricing policies are an impediment to the growth and development of the geographic software industry. The difficulty of obtaining spatial data is a barrier for smaller companies, agencies, and governments which prevents them from using modern computer technology for geographic decision making. This retards the development of the consulting community and software companies that could provide these services. At a time when the governments are working hard to stimulate the growth of high technology employment, it is ironic to be pursuing a data policy that resists better use of information technology.

Cost to higher education

University geography students have no budget to pay the costs of spatial data for their projects. This frequently forces them to use data from the USA instead of their own country. One student pointed out that it is more difficult to verify the results of a student project when it cannot be done with an area of the earth familiar to the student. Someone who has never been to Colorado or Los Angeles has difficulty verifying that they are not using false assumptions that would be more evident in their own neighourhood.

Cost to search and rescue

There are Search and Rescue agencies in Canada which have to use paper maps because they do not have access to digital data. This means that they cannot use laptop computers for coordinating searches, but need to carry bulky rolls of paper maps around instead.

Cost to democracy

Freedom of Information legislation is founded on the principle that for citizens to have effective input into governmental decisions, they must have access to the information that the government uses. Canada?s geospatial data pricing policies keep data out of the hands of organizations that could use it for community purposes. For example, local groups wishing to monitor stream pollution or community watershed degradation do not have access to the best digital topographic maps. Only the government itself does, putting citizens at a disadvantage.

Lack of enforcement encourages cheating

Prosecution for violation of copyright on data is rare. Offences would be very difficult to detect and prove without searching premises and computer systems, so there is little fear of prosecution. When goods with a large nominal value are exchanged in an honour system without legal consequences, distortions and cheating are almost inevitable

Free data distribution can be very inexpensive

Placing data on a web site or FTP site for free distribution is very inexpensive. The costs for network bandwidth are miniscule, compared to the value of the information. For example, the  USGS Geographic Data Download is a website with topographic information for all of the United States. Another is the  Yukon government Department of Renewable Resources GIS.

This technology makes information available for download without any human intervention or accounting required.

Recent Developments

KPMG Study

The Geoconnections branch of the Canadian Government released the GeoConnections Canadian Geospatial Data Policy Study in June, 2001. It was produced by KPMG Canada under contract.

The recommendations from the Executive Summary (PDF):

  1. Recommendation -- Digital geospatial data that are collected or created by any level of government should be made as readily available electronically to the public as possible by improving access mechanisms and processes, unless there are privacy, security or competitive reasons not to do so. Specifically, in implementation, the following points should be taken into consideration:
This study has encouraged debate on the topic in Canada, but as of spring, 2002 has not been adopted by the Federal government.

Treasury Board order

On April 1, 2000 the Government of British Columbia Ministry of Finance approved a new policy on data distribution within the provincial government. In summary, information and data held by B.C. Crown Ministries must to be made available to all other B.C. Crown Ministries in its original format, at the cost of distribution. This policy does not apply to Crown corporations, or other levels of government.

This limited decision has already begun to opened up access within the provincial government. The KPMG study advocates applying this logic to a wider range of public agencies. Ultimately, there is only one taxpaying public in Canada, and cost recovery pricing for digital information only creates artificial barriers between its representative governments.

Other jurisdictions

Things are very different in the United States. In 1994, President Clinton ordered that all spatial data collected with federal funds must be made available to the public at the cost of distribution, without copyright restrictions. This has led to huge volumes of government data downloadable at no charge on the Internet.

The New Zealand government has decided to make their topographic database available free of copyright.

On 25 September 2001, the Australian Government announced a new Commonwealth Policy for Spatial Data Access and Pricing,

Recommendations

By dismembering the bureaucracy that charges for geospatial data, Canadian governments can :

Why haven't things changed?

There aren't bad or stupid people to blame for this problem. People are simply doing the best they can to follow their mandates, and we expect they are probably doing a good job in the context of their narrow mandates. GDBC for example, is following the orders of Treasury Board. Treasury Board is trying to recover the costs associated with developing these datasets in the first place. And to the politicians that influence Treasury Board, who have no knowledge of this issue, present policy probably seems to make sense. Smart people are making the best decisions they can.
Granted, it certainly seems absurd that it could be more expensive to charge for information than give it away at the cost of distribution. Except that it isn't absurd. It is absolutely the case.

How to fix the problem

The first step in resolving this problem is to make government geospatial information free from copyright and available at the cost of distribution immediately. Look at the academic literature; revenue in the form of taxes will be generated by industry doing innovative things with the data.

We could try a petition. Someone else has tried that. A petition is one document with a bunch of signatures or names. What is needed is dozens, or even hundreds, of voices and documents imploring politicians and Treasury Board to change the system.

Many people, especially in government, support better access to government geospatial information. The problem is that many people have been unable or afraid to speak out given their positions. There is tremendous support for this cause, but in order to make things right, we all need to speak out.

Another step is to seed the pot. Wherever possible make the geospatial data you create and work with free for the public good. It doesn't matter how small a piece it is. Every little bit helps to build network effects and sooner or later critical mass will be reached.