Remote Sensing of Earth from Space: Economic and Policy Issues

It began by accident in 1963. An astronaut smuggled a simple snapshot camera aboard a Mercury spacecraft and brought back the first civilian photos from space. Tremendous scientific acclaim was accorded to these synoptic views of Earth, leading eventually to a civilian remote-sensing satellite program in the United States.

A host of countries—including the United States, Japan, Canada, Europe, Brazil, India, and the former Soviet Union—has invested nearly $3 billion to date in remote sensing of Earth by civilian satellites. Sensors on the satellites use energy waves to detect numerous features of Earth and its atmosphere. The data collected by the sensors are transmitted to stations on the ground and either translated into photographs or, increasingly, into digitized tapes for computerized analysis.

Remote-sensing systems are used to map Earth's topography, identify mineral deposits, delineate land uses, monitor soil and climate conditions, assist in flood control, identify archaeological sites, and monitor utility rights-of-way to help ensure safe operation of gas pipelines. In addition, these systems are a powerful tool for learning more about the complex processes governing Earth's oceans and atmosphere, thus improving man's ability to understand and respond to potential climate change and other human impacts on natural systems.

Because remote-sensing data provide information about both public goods, such as environmental quality, and private goods, such as oil deposits, substantial controversy has always surrounded public policy about who should pay for remote sensing. This controversy was evident in congressional debate during 1991 about shortcomings of the Land Remote Sensing Commercialization Act (LRSCA), which had been passed by Congress in 1984. LRSCA sought to transfer operation of Landsat, the civilian remote-sensing system owned, operated, and funded by the government, to a private contractor. Under the act the government would retain ownership of Landsat and, for a few years, would subsidize the private firm taking over its operation. Congress also faced a controversial proposal to spend $53 billion on the Earth Observing System (EOS), a set of large satellites that would monitor climate change.

The policy challenges

Remote-sensing data, like other types of information, are an unusual commodity. Information has value only when it influences a decision. In the case of Earth observations from space, one public policy challenge arises from the difficulty of determining how much and what kind of remote-sensing information would be most valuable when information-collecting sensors and the spacecraft that carry them are still in the design stage. The fact that remote sensing provides information about both private goods and public goods complicates this determination. Establishing the value of information about public goods is problematic because the valuation of these goods is itself uncertain.

Another policy challenge concerns determination of the appropriate level and composition of investment in the supply of remote-sensing information. Particularly challenging is the question of what would be the most cost-effective scale of system in which to invest. The traditional notion that space technologies, including remote sensing, must be large-scale and serve multiple purposes is being increasingly questioned. Smaller satellites, each specializing in the collection of different types of data; photographs taken from aircraft rather than from satellites; and conventional ground-based data collection may provide close substitutes for large space-based systems. Analogous technological substitution is found in the electric utility industry, where large-scale systems for electricity generation are giving way to cheaper and more flexible smaller systems.

Yet another policy challenge concerns how remote-sensing information should be priced. Once there is an investment in infrastructure to obtain remote-sensing data, copies of the data typically can be made at little expense. Since it costs so little to satisfy the demand of additional users for the data once the data have been collected, the total benefit to users is greatest if the information is sold at prices near zero. But at such prices the market would not cover infrastructure costs, provide incentives for technological innovation, and protect intellectual property. The problem also arises in the manufacture and distribution of software, in publishing, in the supply of video movie rentals, in the provision of stock market quotations, and in public utility services. In these activities, the pricing problem is addressed principally by the use of customer-differentiated prices, though debate continues about the efficacy of such pricing policies.

In the United States, broad concerns about issues such as climate change partly motivate a desire to expand remote sensing by preserving Landsat and implementing the planned Earth Observing System. There is no doubt that remote-sensing data could provide some of the information needed to formulate policies concerning public goods such as climate. Given the host of uses of remote-sensing data by the private sector, however, a focus on public goods alone would illuminate neither the relative sizes of private versus public demands for information nor the relative urgency of different demands and the nature of potential substitutes for data collected via space-based remote sensing.

Conventional wisdom holds that the private market for remote-sensing data is quite small. In fact, the size of demand depends on the quality, quantity, and availability of data as well as on ancillary infrastructure like the availability and price of computer hardware, software, and human training. (Even today, space remote-sensing courses are not routinely part of curricula in relevant disciplines such as geology.) As hardware, software, and training become readily available, however, private sector demands for remote-sensing data are likely to increase markedly. This factor is especially important in the planning of remote-sensing systems to be launched far in the future (in the case of the Earth Observing System, nearly a decade from now).

The supply side

Small scale, redundancy and modularity of capacity, diversity of production methods, and capability for quick turnaround are attributes of a growing number of industries, including computers, telecommunications, and publishing. They are also becoming attributes of traditionally large-scale industries like steel and auto production and electricity generation. In addition, growth in employment and sales revenue is increasingly the province of small- and medium-size businesses and new entrepreneurial companies, especially in high technology, software, and information services industries.

The above attributes and resulting trends in growth of employment and sales revenue have not generally been characteristic of space activities and enterprises. Why is this the case? With respect to space-based remote sensing, a partial answer is that the provisions of the Land Remote Sensing Commercialization Act restrict flexibility in setting market-driven prices for data. They also create uncertainty about how government will regulate system capabilities such as licensing, data access (for national security or the conduct of basic science research, for example), and operating parameters like spatial resolution (how much detail the spacecraft's sensors are permitted to "see"). In many business plans for new private sector ventures in Earth observation, legal costs budgeted to comply with or challenge LRSCA regulations alone add some 40 to 80 percent to the expected costs of establishing the venture.

These regulatory impediments and costs could be accepted without too much thought if there was a clear-cut case for government monopoly in the provision of Earth-observation information. But such a case has not been made, and there is in fact growing evidence that points to the disadvantages of government monopoly. For instance, larger-scale, more comprehensive remote-sensing systems may well be disproportionately costly to operate relative to smaller-scale, less comprehensive systems that could be operated privately for commercial purposes or under contract to the government to satisfy governmental data needs.

The difficulty of setting a price for an information commodity like remote sensing has also been at the heart of debate about Earth observations from space. Proposed congressional legislation (H.R. 3614) would amend LRSCA to establish two tiers of pricing, one tier for educational and other nonprofit users set at marginal cost, and one tier for for-profit users set at a higher level to provide opportunities for recovering infrastructure costs. It is not clear how the groups would be distinguished or how marginal cost would be defined under this policy, but if past experience with the pricing of services of the nation's space shuttle is any guide, charges to favored users might be set equal to "additive cost"—the direct out-of-pocket expenses for extending data availability to another user, without any provision for recovering capital depreciation and replacement charges. A price equal to additive cost likely would be very low.

Marginal-cost pricing, particularly when based on additive costs, is likely to have several undesirable side effects. Presumably the capital costs would be recovered from other users, especially commercial users. However, it is not clear that this approach would provide fiscal solvency for either a private or a public remote-sensing program. If the price of remote-sensing data is set too high, commercial users could turn to foreign vendors of space-based information or to alternative modes of information acquisition, such as data gathering on the ground or by airplane. Thus, while marginal-cost pricing would not thwart access to data for important public purposes (science and education), it could have the ironic side effect of perpetuating one of the very problems driving current policy change: the inability to make remote sensing financially sustainable, whether or not it is privatized.

Loading capital costs on nonfavored customers also creates difficulties in enforcing limitations on the acquisition of remote-sensing data by favored users—difficulties acknowledged by H.R. 3614. For-profit firms in the Landsat industry typically use academic or other nonprofit organizations as consultants, as do most for-profit firms in the environmental and resource management industries. A natural result of the proposed two-tier pricing system, then, would be "free" data acquisition by for-profit firms via nonprofit associates. If prices to for-profit users were set high enough to cover Landsat's capital costs, the attractiveness of data acquisition through a nonprofit associate would be especially strong and, like the effect of marginal-cost pricing, would threaten the financial sustainability of the system.

These disadvantages of marginal-cost pricing for favored customers are even more serious for potential entrants into the Landsat industry. They might be less of a problem if the best design of a remote-sensing system was self-evident. However, with uncertainty about the best design, barriers to entry may prove insurmountable. Data prices set at additive cost also offer little clue as to the value of different types of information, thus clouding investment decisions for entrepreneurs if the remote-sensing technologies are supplied by the private sector, and for taxpayers if these technologies remain in the public sector.

Because privatization of the provision of remote-sensing data under LRSCA has thus far been ineffective in resolving demand side, supply side, and pricing issues, it might be tempting to conclude that ownership and operation of Earth-observation systems must be largely if not wholly the province of government. As noted, however, this belief is not well substantiated and may indeed be false. Participation by the private sector in the provision of remote-sensing data offers several possible advantages and may well be feasible if economically sound government policies are pursued.

At least two strategies might offer significant opportunities for private sector involvement without risk to the financial inability of space-based remote sensing. One strategy is to remove barriers to the use of small-scale remote-sensing systems. If the viability of such small-scale systems is in doubt because ventures involving these systems are considered too risky, capital costs still too large, or markets still too small, an alternative strategy would be to allow consortia of public and private information providers to jointly undertake investments in new remote-sensing capacity. These providers would still remain competitors in the downstream provision of different information services to users. Such competitive joint ventures (CJVs) have been sanctioned by the U.S. Department of Justice in cases involving other industries.

The design of a CJV would be based on the number and diversity of products to be provided by a consortium. A neutral party (perhaps a contractor hired by the consortium) would operate the venture to avoid favoritism. No party would be barred from the consortium as long as it paid a share of the fixed costs of general facilities and overhead as well as the incremental cost that the consortium would bear in providing an additional product (for example, the cost of adding and operating another sensor). Nor would joining a consortium be obligatory. (A potential supplier would not want to become a member if it thought it could supply a product at lower cost on its own.) Pricing of services by consortium members would be unregulated, assuming there would be sufficient competition in downstream marketing.