Incentives, precision technology and environmental protection

Ecological Economics 23 (1997) 25 – 43 ANALYSIS Incentives, precision technology and environmental protection Madhu Khanna a,*, David Zilberman 1,b a Department of Agricultural and Consumer Economics, Uni6ersity of Illinois, Urbana-Champaign, 431, Mumford Hall 1301, W. Gregory Dri6e, Urbana, IL 61801, USA b Department of Agricultural and Resource Economics, Uni6ersity of California, Berkeley, 207, Giannini Hall, Berkeley, CA 94720, USA Received 18 May 1996; accepted 16 October 1996 Abstract This paper calls for a reevaluation of the pollution problem in order to devise efficient strategies for its control. It contends that a fundamental cause of pollution is inefficiency in input-use during the production process. This inefficiency manifests itself as input-waste and polluting residues. Precision technologies can increase the effectiveness of inputs used in production and reduce pollution generation. Despite their potential economic and ecological benefits, adoption of precision technologies has been slow. This paper presents a conceptual framework to identify and analyze the specific factors that affect their diffusion among microunits. These include various heterogeneous characteristics of microunits, prices which do not reflect the relevant scarcities due to distortionary regulatory policies and lack of institutional mechanisms for efficient allocation of inputs and outputs. It rationalizes the observed variations in the appropriate choice of technology across microunits and argues for the need to broaden the range of incentive-based instruments for environmental protection beyond pollution taxes. © 1997 Elsevier Science B.V. Keywords: Environmental policy; Heterogeneity; Policy distortions; Precision; Technology adoption 1. Introduction * Corresponding author. Tel.: + 1 217 3335176; fax: +1 217 3335502; e-mail: khanna1@uiuc.edu 1 Tel.: + 1 510 6426570; fax: + 1 510 6438911; e-mail: zilberman@are.berkeley.edu. There is a widespread belief that modern technologies of production with their heavy reliance on fossil fuels and chemical inputs are inherently polluting. Strategies suggested for environmental protection therefore include reduction in the use 0921-8009/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 9 2 1 - 8 0 0 9 ( 9 6 ) 0 0 5 5 3 - 8 26 M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 of these inputs and in production levels, a return to natural or biological inputs, use of ‘alternative (low-input) agricultural practices’, and a shift away from corporate farms to small scale family farms (Mishan, 1993; National Research Council, 1989; Batie, 1988; Buttel et al., 1981). This paper argues that these strategies do not fully address the root cause of the problem. The inherent cause of pollution is not modern technology, levels of input-use or production per se. Instead, a fundamental cause of pollution is inefficiency in the utilization of inputs in the production process. Input-waste manifests itself as residuals due to the law of materials balance. This law states that the mass of all material inputs from the environment (energy and raw materials) to the economy must equal the mass of final products plus the mass of residuals discharged to the environment minus the mass of materials recycled (Ayres and Kneese, 1969). Residuals from inputs such as irrigation water, pesticides, fertilizers and fossil fuels are a major cause of environmental problems like mineralization and salinization of soil, chemical contamination of ground and surface water, and accumulation of greenhouse gases. Agronomic and engineering research shows that there exists a tremendous range of proven alternative technologies in agricultural, industrial, and energy production processes that have the potential to reduce pollution at source by increasing the efficiency of input-use. Efficiency of input-use is defined as the ratio of the input effectively utilized in production (or converted into output) to the input that is applied to the production process. We refer to these technologies as ‘precision technologies’. In agriculture, these technologies include low drift spraying equipment for pesticides and herbicides, monitored and timed applications of fertilizers together with soil nitrate tests, and drip/sprinkler irrigation. In the other sectors these technologies include fluidized bed processors for electricity generation, high efficiency air classifiers for cement manufacture, targeted paint spraying techniques for metals and wood and fuel-efficient appliances and automobiles. All these technologies apply inputs more precisely to the production process, thereby increasing the efficiency of input- use and reducing input-waste and pollution per unit input at source. These technologies also have the potential to alter the link between pollution and output. Controlling pollution at source through their adoption is likely to be more costeffective than abating it after it is discharged, using equipment that only adds to costs and does not increase input-productivity (Guinn, 1994; Porter and van der Linde, 1995a,b). Although these technologies have the potential to provide both economic and ecological pay-offs, their adoption has been paradoxically slow. Older technologies with lower efficiency, which we now refer to as traditional technologies, continue to be used. Does this imply that microunits (microunit is defined as the smallest producing unit that can be managed independently and has homogeneous production characteristics) that continue to use traditional technologies are irrational? Are there $10 bills waiting to be picked up? Porter and van der Linde (1995a,b) and Ayres and Walters (1991) suggest that lack of awareness among microunits about the costs of inefficiency and about these technologies results in non-adoption of profitable and environmentally friendly technologies. This paper argues that even when profit-maximizing microunits have complete information about alternative technologies, it may be rational for some of them not to adopt these technologies. This could occur for two reasons: (1) microunits are heterogeneous and the gains from adoption as well as the costs of adoption, in the form of investment in human and physical capital, vary across them; and (2) existing market imperfections reduce the incremental benefit of adoption of precision technologies. These imperfections include distortions in related input and output markets, institutional barriers to efficient allocation of inputs, input prices that do not reflect their true scarcity values, and lack of penalties for residual generation. In order to provide insight into the technical and behavioral factors that influence the adoption of precision technologies, this paper presents a conceptual framework that juxtaposes discrete technology choice and common attributes of a broad range of precision technologies with a behavioral economic model. This generic framework M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 recognizes both the heterogeneity among microunits in an industry and the inherent bio-physical linkages between production techniques, and input-use and pollution. It generalizes the models developed to analyze policies for controlling pollution from two specific but diverse production processes, namely irrigated crop production and electricity generation (Dinar and Zilberman (1991), Caswell et al. (1990), and Dinar et al. (1992), analyze irrigation technology choice and its implications for drainage problems in California. Khanna (1995) analyzes the impact of adoption of clean coal on greenhouse gas emissions from electricity generation in India). This framework is valid for analyzing choices for precision technologies in numerous other production processes where inefficiency in input-use is the underlying cause of pollution. It abstracts from the details of specific production and pollution processes and focuses on the structure of the production process and on the linkages between production and waste generation based on scientific knowledge. This framework draws on the work on residual management by Ayres and Kneese (1969), but focuses on source reduction of residuals through technology adoption. It also extends the threshold model of adoption (David, 1975) and its application to irrigation technology choice (Caswell and Zilberman, 1986) to examine adoption of a broad range of precision technologies and its implications for environmental policy. This framework rationalizes the observed variations in choice of appropriate technology across microunits by bringing forth the various sources of heterogeneity among them and ways in which they affect the incentives to adopt precision technologies. Although adoption invariably reduces the input-output and pollution-output ratios, this framework shows that it may not lead to reductions in input-use or pollution in all cases. As a result, imposition of a pollution tax may not invariably increase the diffusion of these technologies in an industry. However, incentive-based environmental policies that allow flexibility in technology choice among heterogeneous microunits are likely to be more efficient for controlling pollution than mandated adoption of precision technologies, particularly when the costs of adop- 27 tion are high. Further, our analysis brings forth the need to broaden the range of incentive-based policy instruments beyond those aimed directly at pollution. It argues that these instruments should also include elimination of other market imperfections which prevent economically sound resource management. By pricing inputs, outputs, and pollution to reflect the relevant scarcities and costs, such policies can harness technological progress to protect the environment. 2. Literature review Much of the existing research in the environmental policy literature ignores the physical and biological linkages between production and pollution and the potential for altering these linkages through discrete technological alternatives (Helfand and House, 1995; Kohn, 1993; Kohn (1988); Helfand, 1992; Baumol and Oates, 1988; Dewees, 1983 (Kohn (1988) shows that abatement may increase the scale of production of a firm in long run competitive equilibrium. However, because abatement raises average and marginal costs of production and thus the equilibrium output price, it must reduce aggregate output). Unlike this paper, these studies assume identical firms and focus on abatement of pollution either after it is generated or through input substitutions that involve movements along the existing marginal productivity curve of a polluting input and not an outward shift in it. As a result, they invariably show that incentive-based environmental policies lead to a reduction in production levels. In contrast to Milliman and Prince (1989) and Downing and White (1986) (these studies compare the incentives provided by alternative environmental policies for innovation/adoption of technologies that reduce marginal cost of abatement among identical firms), this paper shows that the availability of these technologies could induce their adoption and reduce emissions even in the absence of any environmental policy. Unlike their analysis, this paper also shows that an emissions tax may not always provide incentives for adoption. The impact of such a tax on adoption rates in an industry depends upon the distribution of 28 M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 heterogeneous characteristics of microunits in the industry and on the technical attributes of the precision technology. Several studies have emphasized asymmetric information between regulators and polluters and uncertainty about discharges, particularly in the case of nonpoint pollution sources. Hansen (1995), Xepapadeas (1991), and Segerson (1988) design policies to control pollution based on measurable ambient pollutant levels while Smith and Tsur (1995), Laffont (1994) and Dasgupta et al. (1980) design mechanisms to induce firms to reveal their true type (pollution) to the regulator. This literature treats pollution primarily as a behavioral problem and under-emphasizes heterogeneity and the possibility of discrete technological alternatives that affect the linkages between production and pollution. Information about the characteristics of heterogeneous microunits and their efficiency of input-use is increasingly being made available through mechanisms such as geographical information systems (GIS), global positioning system (GPS) and computerized systems that track performance of equipment and flow of inputs in a manufacturing facility. This paper, therefore, suggests that this information about microunits together with observations on their technology choice and inputuse can be used to infer the extent and sources of pollution and to design efficient environmental policies. 3. Examples of proven precision technologies This section discusses examples of precision technologies (summarized in Table 1) from a variety of production processes that appear to be very diverse. It brings forth the attributes that are common to all of them and that infuse these technologies with the potential for reducing pollution at source. These common attributes include an ability to increase efficiency of input-use by improving application of inputs in the production process while reducing residuals per unit input. These technologies also have the potential to increase input-productivity levels above and beyond those achieved through increased input-use effec- tiveness in both agricultural and industrial production processes. In agriculture, these technologies avoid deficiencies and excesses in input application and reduce biological stress on crops, resulting in a synergistic impact on their yields larger than simply due to more efficient input-use (Wallace and Wallace, 1993). Their adoption also results in improvements in production conditions and in maintenance of industrial equipment that reduces wear and tear and shutdown time of equipment (Porter and van der Linde, 1995a). The adoption of precision technologies does not necessarily require radical changes in production processes. In many cases these technologies can be adopted in the short or medium run with small investments in capital. A study on US chemical plants found that 80% of the projects implemented by them to reduce pollution at source involved simple technological changes and 25% required no capital investment (INFORM, 1992). However, the requirement for higher human capital skills, improved monitoring and maintenance of the production process, higher quality inputs and capital equipment that accompanies the adoption of many precision technologies is likely to raise the fixed costs of production and application costs per unit input. Sustained improvements in efficiency are likely to require further investments in R and D and in advanced equipment in the long run. Modern irrigation methods, such as drip irrigation, are an example of precision technologies. They apply water more accurately and at appropriate times, thereby increasing the percentage of applied water which is effectively used by crops from 60 to 95%. This reduces run-off from 23 to 1% and deep percolation from 18 to 4% (Hanemann et al., 1987). In contrast, gravity-based irrigation can cause severe problems of waterlogging and salinity of groundwater due to run-off and deep percolation of excess water (Dinar et al., 1992). It is also responsible for more than 40% of nitrogen runoff and 50% of nitrogen leaching in high plain regions (Wu et al., 1994). Laser-controlled leveling of fields also improves water distribution and water intake, and increases crop yields by 8 – 22%. Table 1 Examples of precision technologies for source reduction of pollution Input/source Industry Source of heterogeneity Improved technology Salts/minerals Residue of irrigation Agriculture Water holding capacity of soil Nitrates Residue of chemical fertilizers Agriculture Plant health, soil quality Chemicals of pesticides Residue of pesticide Agriculture Carbon-dioxide Coal combustion Electricity generation; cement; steel Weather, wind, and land quality Vintage and age of capital; quality of management Carbon-dioxide Electricity/heat from fossil fuel Households combustion Sulfur-dioxide Coal combustion Acetylene Calcium carbide Volatile organic compounds Solvent-based paints Hydrocarbon halogens, acetic acid, phosphates Dyes Textile Operating and maintenance practices Chromium III Chromium sulfate Leather tanning Operating and maintenance practices Carbon monoxide, VOC Gasoline Automobiles Age, maintenance practices Drip or sprinkler irrigation instead of furrow irrigation More accurate timing and application; pre-plant and spring nitrate tests IPM; ground application instead of aerial spray Washed coal; natural gas or coal gas. Dry process and high efficiency air classifiers for cement; high quality coal or electric furnaces for steel. Energy-efficient appliances, improved ceiling and wall insulation. Washed coal; low sulfur coal; natural gas. Calcium carbide with finer particles Use of high-volume low pressure spraying of paint; powder-coating and electrodeposition, water-based solvent. Alternative chemicals (Remazol or Procion); alternative dyeing processes (spraying, pad-batch dyeing, application through foam media) Synektan Tal, a non-chrome tanning agent;Baychrome high exhaust/intake process. Ultra light fuel efficient vehicles, electric vehicles Education, income, age of buildings. Electricity generation Vintage of capital; quality of management Ductile iron Design of equipment and operating practices Wood furniture fabricated Operating and maintenance metal products practices M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 Pollutant 29 30 M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 In the US about two-thirds of all insecticides and fungicides are applied aerially and only 25% of them reach the target area. The amount of insecticide reaching the target area can be increased to 50% through use of ultra low volume equipment on aircraft and to 75% through the use of ground application equipment. Improved techniques for rotation and tillage further reduce insecticide use by 6% (Pimentel and Lehman, 1993; National Research Council, 1989). Improved lowdrift nozzles and ‘air assisted’ application technologies can significantly reduce pesticide drift and achieve greater surface area coverage per unit of applied pesticide. Similarly, by using field scouts to measure insect pressure, integrated pest management improves the accuracy and timing with which insecticides are applied. This increases yield and reduces pesticide use (Ferguson et al., 1993; Thomas et al., 1990). Spatial variability in soil characteristics can be addressed by using soil testing to guide precision nutrient management. Pre-plant soil nitrate tests (Schmitt and Randall, 1994), late-spring and other nitrate tests for soil (Babcock and Blackmer, 1992; Musser et al., 1995) can significantly reduce applied nitrogen while having a positive impact on yields and profits. Nitrogen testing of soils in Nebraska can reduce nitrogen use by 18– 27% without reducing yields (Fuglie and Bosch, 1995). Variable rate technologies that adjust fertilizer applications in response to spatial variations in a field and allow more timely, precise and judicious use of fertilizers, increase the efficiency of fertilizer use and reduce nitrate losses (Sawyer, 1994; Blackmore, 1994). Similar benefits of precision technologies have been documented in many other production processes. Conversion of energy from fossil fuels into electricity generates residuals such as carbon and sulfur dioxides (the combustion of fossil fuels combines oxygen from the atmosphere with hydrogen and carbon in the fossil fuel to form residuals such as water, carbon and sulfur dioxides). The conventional coal technology for electricity generation converts a maximum of 35% of the heat energy of coal into electrical energy. Generally, this conversion efficiency is much lower due to use of low quality (high ash) coal, aging plants, and poor operating practices. Use of a pre-combustion technology to wash coal removes extraneous materials such as ash and improves the precision with which carbon is applied for electricity generation. Washing reduces the ash content of coal from 23.5 to 3.5% and its sulfur content by 13– 35%. It also increases the heating value per ton of washed coal, improves the efficiency of electricity generation, and reduces carbon and sulfur emissions per kW h (EPRI, 1988). It also reduces wear and tear of equipment and decreases shutdown periods of the plant for maintenance, thus increasing the productive capacity effectively available for electricity generation. In the long run, conversion efficiency can be further increased by using fluidized-bed processors and advanced coal gasification technologies. Precision technologies also increase efficiency of input-use in cement and steel manufacture. In the case of steel, high quality coal reduces the coke rate (the quantity of coke needed per ton of hot metal output) by 23%, increases productivity of the furnace by 28% and reduces carbon emissions (Mongia et al., 1994). Production of cement with a dry instead of a wet process, together with use of high quality coal, reduces coal/ton of cement by 19.6% and generates 7% less CO2/ton of cement. Further, a high efficiency air classifier reduces electricity consumption in cement production (kWh/ton of cement) by 24– 30%. It increases productivity of electricity by increasing the production rate of cement (tons/hour) by 80– 100% and decreases CO2/ton of cement (Cienski and Doyle, 1992). Residue of calcium carbide used for the production of ductile iron generates a pollutant, acetylene. Using calcium-carbide with finer particle size, instead of the regular carbide, increases carbide-use efficiency from 28 to 39% and reduces carbide residue from 10.4 to 0.65% (Barker, 1992). In the fabricated metals product and the wood furniture industries, use of improved spraying processes and alternate materials instead of conventional methods of applying solvent-based paints using hand-held spray guns or electrostatic spray methods increases transfer efficiency of paint from 30– 80% to 95– 100%. This reduces use of materials and wastewater and eliminates emis- M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 sions of volatile organic compounds (VOC) (Denny et al., 1995; Heltzer, 1995). In the textile industry, use of alternative chemicals, such as Remazol and Procion, and cleaner technologies, such as pad patch dyeing, increase fixation efficiencies of the dye to the fabric to 95– 98% and reduce use of auxiliary chemicals, water, and energy per unit output and reduce generation of toxic waste-water (Snowden-Swan, 1995). In the leather tanning industry, chromium III effluents can be reduced by up to 95% through the Baychrome process and supplemental use of Synektan Tal, a non-chrome tanning agent. They reduce chrome requirement and increase chrome intake efficiency (Alexander and Donahue, 1990). Several technologies that increase efficiency of consumption of energy in buildings, appliances, and transportation are also available. Improving insulation to reduce heat loss in buildings, using task lighting and high efficiency fluorescent lamps, using ultra-light hybrid automobiles (with fuel efficiency greater than 80 miles per gallon) can aid the conservation of resources and protect the environment (Lovins and Lovins, 1995; NorbergBohm, 1990). These technologies, though seemingly unrelated, share common attributes that increase efficiency of input application and reduce pollution per unit input. We now present a model for analyzing the choice for these technologies that incorporates these attributes in a stylized manner. 4. The micro-level decision model Consider a production process that requires a variable input and a fixed asset (for example, land or machinery). For simplicity, we assume that microunits have a choice of two technologies, i =1, 2 where i= 1 is the traditional and i= 2 is the precision technology. A distinction is made between the amount of applied input, and the amount of effective input only a part of the applied input with technology i, ai, is used effectively in the production process, ei, and a multiplicative relationship is assumed between the two (Caswell and Zilberman, 1986): ei =hi (a)ai ; h %i \0 and h ¦i 31 (1) where hi (a) is a measure of efficiency of input-use and depends upon technology i and on a, an index of the physical conditions under which production occurs. The index a includes variables such as quality of the fixed asset, weather, managerial ability, and structure of ownership of the fixed asset. Index a is scaled to range between 0 and 1 and varies across microunits. It is also reasonable to assume that the efficiency of inputuse increases as the physical conditions of production, a, improve, but at a decreasing rate. We assume a constant returns to scale production technology with output per unit asset, yi, being a function of effective input per unit asset and technology choice: yi =f(e,i ); fe \0, fee B 0 (2) More specifically, the production function is specified as: yi =f(bi hi (a)ai ) (3) with fa \ 0, faa B0 and bi representing an index of the productivity of the effective input. This production function can be adapted to represent alternative production processes. In representing the crop-water production technology, Caswell et al. (1990) assume bi = 1 for all i, while Dinar et al. (1992) postulate: yi = f(h(a, i, a, c), i, c), where a denotes land quality or soil/water salinity, and c denotes weather/temperature. It can also be interpreted as a household production function and used to model output produced by automobiles or appliances with varying energy efficiencies. This function can be further modified to represent production processes with putty-clay technologies (Johansen, 1972) as for cement and electricity in the short run (Sterner, 1990; Khanna, 1995) and the linear-plateau model for crop production using irrigation water (Letey et al., 1985) or nitrogen (Berck and Helfand, 1990). In both these cases, the production function can be specified as: yi =bi hi (a)ai for yi 5ȳi. Following the law of material balances, the amount of residuals generated is directly related to the amount of input-waste produced and thus inversely related to the effectiveness with which an M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 32 input is used. The link between residuals and input-use varies not only with technology, but also with the heterogeneous conditions under which production takes place and which are captured by a. We hypothesize the pollution coefficient of an input with technology i, gi, to be an inverse function of the prouction condition, a. Pollution per unit asset with technology i is expressed as a linear function of applied input use: zi =gi (a)ai ; g %(a)B 0 i (4) In the limiting case, all of the input wasted may become a polluting residual and gi =[1− hi ]. 4.1. Technical features of alternati6e technologies Based on common characteristics of a broad range of precision technologies, drawn from the examples discussed above, we assume that a precision technology has three technical effects. First, it increases efficiency of input-use applying inputs more precisely, in terms of quantity, quality, and timing to the production process. This implies: h2(a) \h1(a) for 0B a B1 (5) This is the precision-effect of the precision technology which reduces the input-output ratio. Secondly, a precision technology has a direct effect on output that is over and above the increase caused by the precision effect. By improving the production conditions under which a given amount of effective input is used, it further raises the marginal productivity of the effective input. This effect of precision depends only on technology choice and is independent of the existing efficiency of a unit (or proportion of applied input that is effectively used). Thus: b2 \ b1 = 1 for all a (6) We refer to this as the productivity-effect of the precision technology. Thirdly, the adoption of a precision technology reduces the generation of pollution (waste) per unit of applied input. Thus, for any given a, g2 Bg1 and we refer to this as the pollution-effect of a precision technology. 4.2. Costs of alternati6e technologies Microunits are assumed to be price-taking units both in the input and the output markets. Because the precision technology is embodied in a higher quality variable input (e.g. in higher quality of coal or irrigation water with low salinity) or because its application is more skill or time intensive (e.g. pest scouting) or requires professionals, the price per unit of input applied with a precision technology is assumed to be higher than the price per unit input applied with a traditional technology. Thus we assume that w2 \w1. The annualized fixed cost per unit asset is denoted by ki. Adoption of a precision technology may require fixed investment in the form of modifications to the input application equipment and, thus, k2 \ k1 ]0. Both fixed and variable input prices may vary across microunits due to differences in location of the microunit relative to source of the input (location of a power plant relative to the sources of coal, depth of the well from which groundwater is drawn for irrigation) and differences in the discount rates used for annualizing the fixed costs. 4.3. The adoption decision Microunits choose the technology and level of input-use that maximizes their quasi-rents per unit asset by undertaking a two-stage decision-making process. In the first stage, a microunit chooses the optimal level of input-use per unit asset with both technologies by: Max Pi (a) =Pf(bi hi (a)ai ) −wi ai −ki ; i= 1, 2 (7) The optimal a *i with each of the technologies is determined such that its value of marginal product is equal to its application cost: Pf %bi hi −wi =0 (8) In the second stage, a microunit determines the maximum profits, Pi*, per unit asset with each technology and chooses technology 2 if: P*(a) P*(a)] 2 1 (9) M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 The quasi-rent differential between the two technologies can be represented as: P*− P*= DP* =PDy −w2Da −a1(Dw) −Dk 2 1 (10) where Dx= x2 −x1 for x =y, a, w, k. A precision technology always decreases applied input-use per unit output, but its adoption affects applied input-use per unit asset in two opposing ways —it increases efficiency of the applied input and its per unit price. When the existing efficiency of applied input-use is high, the gain in efficiency with adoption is lower than otherwise. The benefits of increasing applied input-use with adoption can then be more than offset by the higher cost of applied input and thus the net effect is a reduction in applied input-use. However, due to the positive precision and productivity effects, it can be shown that adoption always leads to an increase in effective input-use per unit asset and thus it increases output per unit asset. The effect of adoption on pollution generated per unit acre is: Dz =a1(g2 − g1)+g1 (a2 −a1). Adoption is pollution-saving if it lowers applied input-use or when the reduction in pollution due to the pollution-effect, a1(g2 −g1)B 0, is larger than the increase in pollution due to increased input-use (g1 (a2 − a1)). Even when adoption increases applied input-input-use, it may reduce pollution per unit acre if the pollution-effect is sufficiently large. Although adoption may increase pollution per unit asset, the percentage increase in pollution is always less than the percentage increase in output and thus adoption reduces the pollution-output ratio (proofs of all these statements are in Khanna (1995) and can be obtained from the authors). The technical characteristics of a precision technology (precision, productivity, and pollution-effects) interact with the existing a of a microunit to influence changes in its profit maximizing levels of input-use, output, and pollution per acre with adoption. The sign and magnitude of the profit differential in Eq. (10) depends upon the signs and magnitudes of the following effects of technology adoption: (a) the output-increasing effect, Dy(\ 0); (b) the input- 33 saving effect, Da ( -0); (c) the input application cost effect Dw( \0); and (d) the fixed cost effect, Dk(\ 0). The magnitude of the first two effects is determined by the characteristics of the technology such as the magnitudes of the precision and the productivity-effects as well as the input and output prices and the production conditions (a) within which it is to be adopted. The positive differential in input application costs per unit input and in the fixed costs per unit asset reduce incentives to adopt precision technologies. The choice of technology, thus, involves a trade-off between the benefits of adoption in the form of input savings and higher revenues, and higher application and fixed costs per unit asset. This trade-off varies across microunits due to differences in the input prices they may face and because the impact of adoption on input-use and output depends upon the existing a of a microunit. It can be shown that when the precision-effect of these technologies is large, microunits facing the same input prices but with relatively low a (adverse physical conditions) are likely to adopt precision technologies. When the precision-effect is low, but the productivity-effect is strong, adoption is more likely among microunits with high a. This analysis brings forth two key points. One, the introduction of a precision technology in an industry, where it was previously not available, could create voluntary incentives for its adoption, even in the absence of any environmental policy, if the profit differential in Eq. (10) is positive. When a precision technology is pollution reducing, its availability could simultaneously lead to increased profits and lower pollution for the adopting microunits. Secondly, even when microunits have information about existing precision technologies, adoption may not be economically viable for all of them. As argued above, the net gains from adoption vary across microunits depending upon their heterogeneous characteristics and the input and output prices faced by them. We now discuss the impact of specific sources of heterogeneity on incentives for adoption. 34 M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 4.4. Heterogeneous factors that influence adoption of precision technologies Empirical studies that use field data to examine the pattern of adoption of precision technologies demonstrate that a significant portion of observed variability in technology choice can be explained by the observed heterogeneous characteristics of microunits. These characteristics include their physical environment, ownership of assets, human capital skills and the local infrastructural conditions and institutions. 4.4.1. Physical en6ironment The physical environment influences the technology adoption decision of a microunit. The likelihood of adoption of precision technologies is positively related to the age and vintage of microunits in the electricity supply industry (Khanna, 1995) and in the steel industry (Oster, 1982). Soil quality, topology, and weather also influence the choice of input application techniques. Empirical analysis for the San Joaquin Valley shows that modern irrigation technologies were more heavily adopted on steeply sloped fields with lower water retaining soils and with salinity problems (Dinar et al., 1992). Nitrogen testing in Nebraska was adopted more frequently on irrigated fields with higher organic content and pH (Bosch et al., 1995). Another source of heterogeneity among microunits is their location relative to the source of alternative inputs. The latter affects the conveyance cost of inputs and thus their price. In the case of groundwater used for irrigation, farms with deeper wells were more likely to switch to drip irrigation because it reduced the costs of pressurization of water (Caswell and Zilberman, 1986). Power plants located closer to the source of low quality coal are less likely to switch to higher quality coal which comes from a distant location (Khanna, 1995). It is generally believed that large scale, industrialized farms overexploit the environment (Buttel et al., 1981). However, numerous studies find that farm size is positively related to the adoption of soil conservation practices (Nowak, 1987; Heffernan and Green, 1986), IPM (Thomas et al., 1990) and modern irrigation (Dinar et al., 1992). This could be due to the greater ability of large farms to hire professional applicators and crop consultants, easier access to credit and information and more contacts with extension agents and representatives of agribusinesses. 4.4.2. Ownership of the fixed asset Ownership of land or capital equipment impacts management quality and the operating and maintenance practices of a microunit, and can influence incentives to adopt productivity-enhancing technologies (Feder et al., 1988). Unified organizations where labor and management are incorporated in one person, market-oriented farm organizations, and tenured/owner farmers are more likely to adopt irrigation technologies (Dinar et al., 1992), soil conservation techniques (Nowak, 1987) and IPM (Yee and Ferguson, 1994). Commercially-operated microunits have higher efficiencies than those under state ownership or cooperatives (Khanna, 1995; Sterner, 1990). The nature of occupancy of houses — owner, tenant, single or multiple —affects energyefficiency of homes. Owner occupation and single ownership of houses results in greater incentives for energy conservation and for installation of energy-efficient appliances than tenancy in rented homes (Fisher and Rothkopf, 1989; Jaffe and Stavins, 1994). 4.4.3. Human capital, learning costs, and discount rates The age of farmers, number of years and quality of education, as well as their personal learning ability and motivation affect their human capital skills, managerial capabilities and the technical efficiency of production on farms (Kalaitzandonakes and Dunn, 1995). Differences in these attributes together with differences in experience and exposure to extension services lead to heterogeneity in the learning costs and other transition costs of adoption of precision technologies. Higher levels of education and younger age of farmers is significant in inducing adoption of computer technologies for farm management, modern technologies for irrigation (Dinar and Yaron, 1990), soil tests for nitrogen (Fuglie and M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 Bosch, 1995) and IPM (Yee and Ferguson, 1994; Thomas et al., 1990). High levels of experience with traditional technologies may, however, discourage adoption of precision technologies. Studies show that implicit rates of discount used by producers and consumers are typically much higher than real rates of interest, and vary widely with income and age of consumers and the financial and credit standing of producers (Ayres and Walters, 1991). Effective discount rates are estimated to be ten times as high as actual market discount rates (Ruderman et al., 1984). This raises the annualized fixed costs of precision technologies and lowers incentives for adoption. 4.4.4. A6ailability of precision technologies Precision technologies and the infrastructure (e.g. skilled manpower) needed to support their operation may not be available locally or at the appropriate scale to allow widespread adoption. The availability of crop consultants and agricultural input-supply dealers is distributed unevenly in the US and is likely to have a strong influence on the rates and geographic patterns of adoption of precision technologies (Wolf, 1995). In the case of developing countries, lack of foreign exchange to buy energy efficient technologies or critical spare parts or lack of capital to manufacture these technologies domestically and an inability to create the required infrastructural support has inhibited their adoption (Lashof and Tirpak, 1990). Government policies to meet objectives other than those of environmental protection and economic efficiency, such as equity, may restrict availability of precision technologies. For example, import of high quality coal for power generation is restricted in India to protect employment in the domestic coal sector. 5. Impact of alternative policies on adoption The profit differential between the two technologies in Eq. (10) is affected not only by the heterogeneous characteristics of microunits, but also by policy changes that influence the output price, the application costs per unit input, and the differences in fixed costs per unit asset with both 35 the technologies. Additionally, penalties on pollution are also likely to influence the net gains from adoption because of the differential effect of the two technologies on pollution generated. 5.1. En6ironmental policies With the imposition of a pollution tax, t, the optimization problem of a microunit can be represented as Max Pti (a) =Pf(bi hi (a)ai ) −wi ai − ki −tgi ai (11) Optimal applied input-use, a ti, is determined such that the value of marginal product of the input is equated to its social cost per unit, defined as 6i =wi +tgi (a). The social cost of an input is the sum of its market price per unit, wi, and its pollution cost, tgi (a). The latter depends upon the tax rate and the pollution coefficient, gi, of technology i. It can also be regarded as a microunitspecific input tax. Adoption of the precision technology now occurs if Pt2 − Pt1 =DPt =PDy −w2Da −a1(Dw) −Dk + t(g1 −g2)a1 −tg2Da \0 (12) The term t(g1 −g2)a1 reflects the additional incentives provided by the pollution-effect of a precision technology for adopting it after the imposition of a tax, t \0. However, the last term in Eq. (12) shows that if adoption is input-use increasing (Da \0) then the imposition of a tax reduces incentives for adoption. The imposition of a tax is likely to induce a microunit to adopt a precision technology when adoption is input-saving or when it has a sufficiently strong pollutioneffect that offsets its input-increasing effect. The tax could also cause microunits with Dz\ 0 that had earlier adopted the precision technology to switch back to the traditional technology to reduce their tax burden. Although their pre-tax profits are higher with the precision technology, their post-tax profits may be higher with the traditional technology. Hence, the impact of a pollution tax on adoption depends upon the technical characteristics of technologies and the existing a of a microunit. 36 M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 The implementation of a pollution tax or a microunit-specific input tax policy, however, requires the ability either to monitor pollution or to make inferences about its magnitude based on knowledge about the efficiency of input-use (or pollution coefficients), technology choice and variable input-use by microunits. When information is available about the distribution of efficiencies at the regional level but not about site-specific efficiencies, then pollution taxes or site-specific input-taxes are no longer a feasible policy. Second-best policies based on more easily observable decisions on input-use and/or technology choice need to be designed. The choice of policy now depends upon the extent of information available. When input-use and technology choice are both observable, input-taxes differentiated by technology can be designed. These taxes reduce the differential in application costs per unit input and achieve pollution reduction both by raising the social costs of input-use (more so with the traditional technology) and by creating incentives to switch to precision technologies. Alternately uniform input taxes can be used even though they are less efficient than technology-differentiated taxes. In many cases, the feasibility of input-based taxes is limited, because very high tax rates have to be imposed to reduce input-use and pollution by the desired amount or to induce a switch to alternative inputs or input application techniques. This is the case for inputs such as fertilizers and pesticides because fertilizers have a low price elasticity of demand, and share of pesticides in gross expenditures is low while the revenue to cost of pesticide ratio is high. It is estimated that a doubling of nitrogen prices will reduce nitrogen use by 20% (Nutzinger, 1994). Returns on pesticide use are high and estimated to vary between $4 and $13 per $1 spent on pesticides (Pimentel and Lehman, 1993). In such cases, as well as in cases where only technology choice is observable easily, technology-tax-subsidy schemes can be used. Taxing the fixed costs of a traditional technology at higher rates than those of a precision technology or subsidizing the fixed costs of a precision technology creates incentives to adopt it. This policy, however, does not affect variable input-use with a given technology. There are several programs in the US for cost-sharing and providing technical assistance for improving fertilizer, pesticide and irrigation management. All of these incentive-based policies allow flexibility in methods of controlling pollution to varying degrees. As the information becomes more limited, policies tend to become less flexible and less cost-effective. Given the variability in appropriate choice of technology among heterogeneous microunits, flexible policies are preferable to mandatory adoption of technologies by all microunits. Mandated technologies include certain forms of crop rotation and tillage practices, pesticide spraying equipment, bans on certain chemicals, scrubbers to abate sulfur-dioxide from new coal-fired utilities, and automobiles with minimum average fuel economy standards. Uniform technology standards burden heterogeneous microunits with excessive costs relative to flexible strategies, particularly if the fixed costs of adoption are high. Mandating installation of scrubbers by electric utilities or higher fuel efficiency standards for automobiles also discourages adoption of pollution-preventing strategies such as use of washed coal or coal with low sulfur content (Portney, 1990) and reduction in gasoline use in automobiles (Nivola and Crandall, 1995). 5.2. Other policies Apart from non-existence of markets for pollution there are many other market imperfections that contribute to environmental degradation. These include cheap and abundant resources whose prices do not reflect appropriate scarcities, subsidies on inputs and outputs, barriers to trade in environmentally beneficial inputs, and institutional impediments to conservation such as a lack of markets for resources like surface and groundwater. Policy reforms that eliminate these distortions are likely to create incentives for adoption of appropriate technologies and reduce pollution, even in the absence of environmental regulations. 5.2.1. Output price regulations Agricultural markets are typically protected from operation of free market price mechanisms M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 through various devices such as state-guaranteed crop prices, tariffs, import quotas, and export subsidies. Simultaneously, in order to limit agricultural surpluses, governments use production quotas and financial incentives for acreage reductions and set asides extensively. Kalaitzandonakes and Taylor (1990) and Kalaitzandonakes (1994) provide empirical evidence that high protection of the beef/sheep industry in New Zealand and of the vegetable industry in Florida reduced incentives to improve efficiency of input-use and productivity. From Eq. (10) above we see that price supports will encourage adoption of precision technologies that are output-increasing, but they do not provide incentives for adoption of technologies that have large precision or pollution-effects. Moreover, they reinforce acreage reduction policies in inducing excessive use of inputs such as nitrogen and pesticides by providing incentives for intensifying production practices and by making the costs of these inputs cheap relative to the artificially high output prices. Price and income support programs encourage crop specialization and act as barriers to the adoption of diversified farming systems with reduced fertilization and pesticide requirements (Nutzinger, 1994; Reichelderfer, 1990). Replacing existing price support programs with a ‘free-land market’ discourages intensive use of inputs; for example in the case of cotton production it is estimated that this would reduce insecticide use by 10% (Pimentel and Lehman, 1993). Changes in existing agricultural policies can lower pollution by controlling the extensive and intensive margins of cultivation (Just and Antle, 1990). Replacement of deficiency payments by green payments that reward farmers for adoption of precision technologies would lower incentives for increasing production in an environmentally harmful manner (Braden et al., 1994). Targeting acreage reduction programs to land on which production poses an environmental hazard could also integrate agricultural and environmental policies in a mutually beneficial way. Price regulation in the electric utility industry in the US has encouraged investments in additional supply as opposed to improvements in demandside efficiency, even when the latter is less expen- 37 sive, by linking profits to sales (Wiel, 1989). Allowing utilities to profit from conservation by permitting equivalent returns on such investments creates perverse incentives to do the most expensive conservation investments first (to earn the greatest return) with the least impact on demand (to maintain sales). Giving bonuses to utilities for achieving demand side goals, by allowing them to keep a portion of the customer savings and by relating their return on equity to the decrease in the average bill of customers, can effectively provide utilities incentives for undertaking conservation investments. A study for nine developing countries shows that electricity prices in most of them were below long run marginal costs, and in the case of India and Argentina they even fail to cover production costs. Thermal plant efficiencies were also at their lowest in India followed by Argentina (Gutierrez, 1993). Pricing reform in these countries would significantly increase incentives for investment in efficiency improving technologies both for the generation of electricity as well as for the end-use of electricity and promote energy conservation. 5.2.2. Subsidies on inputs and protectionism Many agricultural and industrial inputs such as water and fossil fuels are generally sold at prices that are well below the real resource cost of their use which consists not only of costs of production but also includes scarcity value and costs of pollution. Surface water prices in California have generally followed decreasing block pricing or a charge per-acre instead of a charge per unit used. The high elasticity of demand for water in urban and agricultural use shows that a significant reduction in water use can be expected to accompany real water price increases (Wiley, 1985). This would also induce farmers to switch to advanced irrigation technologies and reduce drainage (Caswell et al., 1990). The rapid increase in fossil fuel consumption worldwide has primarily stemmed from their artificially low prices (Bates, 1993). Oil-producing and exporting countries such as Mexico, Egypt, and Libya regularly subsidize petroleum-based fuels. Coal is subsidized in China and India to encourage domestic production, employment and 38 M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 self-reliance in energy. Cross country studies show that higher energy prices significantly increase efficiency of energy-use and can reduce carbon emissions (Larsen and Shah, 1992; Kosmo, 1987). On the other hand, subsidies on low quality domestic coal and prohibitive tariffs on high quality imported coal have led to a substantial decline in thermal efficiency of the electricity sector in India. Trade liberalization in coal could reduce carbon emissions by 6.5% relative to baseline emissions even in the absence of an environmental policy such as a carbon tax (Khanna, 1995). Similarly, fuel price deregulation could increase efficiency and reduce costs of carbon abatement in Indonesia (Sasmojo and Tashrif, 1991). Rising gasoline prices during the 1970s spurred automobile producers to increase fuel efficiency of automobiles even more than mandated by the corporate average fuel efficiency standard in the US. As fuel prices have declined, since 1982, improvements in average fuel economy have slowed dramatically. Raising fuel prices to reflect resource scarcity and environmental costs is likely to be more effective in improving efficiency of automobiles and conserving fuel than use of mandated standards while keeping fuel prices low (Nivola and Crandall, 1995). In short, the efficiency of resource-use can be increased by relying more on markets for their pricing and allocation and limiting government interference to incentivebased policies for correcting externalites from their use. 5.2.3. Lack of markets In the case of certain inputs such as surface water, markets have not been allowed to develop leading to inefficiencies in water allocation and reduced incentives to adopt modern irrigation methods. In various states in the US, non-market (rationing-oriented) mechanisms are currently employed to allocate surface water among individual users in the agricultural sector. In the absence of water markets, the opportunity cost of water is effectively zero for the prior appropriators. A water market would result in higher prices for water, and could significantly reduce water use while inducing farmers to switch to advanced irrigation technologies. This would also reduce drainage, even in the absence of drainage disposal costs. Imposing drainage disposal costs achieves still greater conservation (Shah et al., 1993). 6. Costs of environmental protection The steeply rising costs of environmental protection are a major source of concern in the US and elsewhere. One reason for these high costs is a focus of environmental policy on abatement of pollution at end of the pipe by diverting resources away from productive activities using mandated technologies. By ignoring the heterogeneity among microunits and the potential for reduction of pollution at source, by increasing efficiency of input-use and minimizing waste, such policies do not achieve pollution reduction in a cost-effective manner. The 1990 Acid Rain Program of the USEPA allowed flexibility of technology choice for reducing emissions at source by relying on tradable permits. It led most of the electric utilities to choose the lower cost option of preventing emissions by using low sulfur coal instead of using scrubbers to recapture sulfur after it was discharged. The general accounting office has projected that this approach saved $3 billion per year or cost half as much as a command-and-control approach (GAO, 1994). The other reason for the high cost of environmental protection is a disregard for policy distortions in related input and output markets that reduce incentives for cost-effective methods of waste-minimization and create price imbalances that favor resource-intensive and pollution-intensive production practices. By eliminating these policy distortions and by inducing the adoption of precision technologies it is possible to not only achieve environmental protection, but also to increase social welfare, net of environmental benefits. The gain in welfare due to removing distortions in related markets can more than offset costs of pollution reduction and achieve environmental protection at negative costs to the economy. Combining environmental policy with reforms that remove existing distortions can thus be more effective than environmental policy alone. For example, imposition of drainage costs M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 in the presence of water markets is shown to result in higher net returns for farmers than in the absence of markets because farmers can now earn revenue by selling excess water (Dinar and Letey, 1991). The removal of tariffs on clean technologies and products, removal of domestic regulations on fuel prices and barriers to the availability of precision technologies can lead to win-win links between the goals of environmental protection and the goals of increased production and social welfare (Khanna, 1995; Beghin et al., 1995; Lovins and Lovins, 1992; Sasmojo and Tashrif, 1991). 7. Distributional implications Adoption of precision technologies has varying distributional implications for producers and input suppliers depending upon the pattern of adoption they induce. When the availability of precision technologies induces voluntary adoption among microunits with low a, it leads to greater equity in the distribution of profits because adoption increases their profits per unit asset relative to those with the traditional technologies and brings them closer to the profits of the microunits with high a. On the other hand, if fixed costs of the precision technology are high and the precision effect is low relative to the productivity effect, adoption is more likely among microunits with high a. This increases the divergence in profits among heterogeneous microunits. These distributional effects get further compounded by existing policy distortions in related markets. For example, when adoption of precision technology has a large productivity effect and increases output, commodity price support programs result in larger benefits for microunits which produce a larger volume of output. Adoption of precision technologies by high a units in the presence of these programs can exacerbate the inequitable distribution of profits in a region. The distributional implications of imposition of pollution taxes or input taxes vary with the tax rates imposed and the pattern of technology adoption they induce. The impact of tradable permits varies with the initial distribution of these 39 permits among microunits. These permits are likely to lead to transfers from pollution intensive microunits with low a and those relying on traditional technologies to the more profitable microunits with a high a and those using precision technologies. The burden of pollution taxes and input taxes is also likely to fall as efficiency of a microunit increases since pollution-levels and input-use are generally inversely related to a. Tax payments are thus relatively higher for microunits with low a and to the extent that these are also less profitable units, these taxes can be regressive. Further, for any given a, these taxes are likely to penalize microunits using traditional technologies more than those using precision technologies. By reducing input-use per unit output, precision technologies are likely to reduce the burden of input taxes. Moreover, taxation of agricultural inputs such as fertilizers and pesticides imposes upward pressures on commodity prices. This adversely affects farmers who rely heavily on taxed inputs, while leading to net gains in revenue for those who have low rates of input-use or use other substitute inputs. When the price elasticity of input-use is low and the input-costs are a small part of the total production expenditures, as in the case of fertilizers and pesticides, very high taxes are required to reduce their use adequately. Given the infeasibility of high taxes, Dubgaard (1990) proposes a hybrid policy for controlling pollution. He suggests combining a tax-free quota of nitrogen with taxes on additional nitrogen purchase. The effective use of the tax-free quota of nitrogen, and hence reduction in nitrogen purchases, can be greatly aided by the adoption of GPS, GIS, and soil nitrate tests that allow precise applications of fertilizer. Hybrid policies that combine input-taxes with compensation schemes to encourage adoption of precision technologies can also temper the adverse distributional implications of these high tax rates. These schemes could take the form of cost-sharing for adoption of alternative pest-management strategies and fertilizer application techniques. By reducing the fixed costs of adoption of precision technologies, these policies can induce adoption while serving as a transfer mechanism. These policies, however, can result in windfall gains to participants, attract 40 M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 new entrants with lower efficiency of input-use, and increase financial well-being of microunits in the polluting industry. Unlike pollution or input taxes that generate revenues, subsidies and costsharing arrangements impose a heavy financial burden on the government. The advent of precision technologies also has implications for the incomes of suppliers of inputs such as high ash coal, fertilizers and pesticides and for crop consultants whose incomes are related to input sales. These agents are likely to witness declining incomes as a result of precision technology adoption, if these technologies reduce use of their products. In the case of consumption technologies, such as automobiles, a tax on inputs such as gasoline can be regressive depending upon the tax rate and the share of incomes spent on gasoline. Chernick and Reschovsky (1992), however, show that the regressivity of a gasoline tax is likely to be modest because gasoline expenditures are a relatively small share of total consumer expenditures even in the lowest income brackets. Moreover such a tax would offset some of the decline in the real price of oil that has occurred since the mid 1980s. 8. Implications for the long run A question that arises inevitably is whether the gains in efficiency of input-use through adoption of precision technologies are a one-time gain only and whether further reductions in pollution will inevitably require reductions in output. As argued above, increasing efficiency in application of inputs requires information and knowledge about the techniques and timing of input use as well as careful monitoring and control of input requirement and application. Thus investments in human capital and stock of knowledge are likely to increase input-use efficiency beyond levels currently achieved, subject to the laws of thermodynamics. A simple extension of the static framework presented in this paper could model input-use efficiency as a function of the stock of knowledge in the economy which in turn depends upon expenditures on education, and research and development. These expenditures can be met by foregoing present consumption (Romer, 1986). Hence, the level of efficiency and technology at any instant in time is endogenously determined. Growth and pollution reduction in the long run are both to some extent driven by increasing efficiency of input-use, although growth can also be achieved through increasing the use of capital intensive techniques that have no beneficial effect on the environment. Further research is needed to identify conditions and policies under which the objectives of growth and environmental protection can be made complementary by inducing investment in the development of environmentally friendly innovations. 9. Conclusions This paper argues that it is misleading to link pollution directly to the use of modern technologies or to production levels. Instead the pollution intensity of inputs depends upon the efficiency of their use and can be lowered through adoption of precision technologies that increase input-productivity and reduce residuals. To explain the slow rate of adoption of precision technologies, it is important to recognize the technical characteristics of these technologies and the sources of variability in the gains from adoption among heterogeneous microunits as well existing policy distortions in related input and output markets. In the past, technological advancement has occurred without much consideration for resource use and implications for the environment, primarily because of lack of incentives to do otherwise. Lack of penalties for residual generation, market failures in input and output markets, and institutional mechanisms that prevent the price system from rationing scarce resources have hampered the development and adoption of environmentally friendly technologies. The analysis above does not examine policy choice and its implications when there is asymmetric information between regulators and polluters. It, instead, explores the implications of policy changes that can be based on observable information about microunits. This paper also focuses on the implications of alternative policies for the adoption of currently avail- M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 able precision technologies and ignores their impact on incentives for research and development of such technologies. Further research is needed to explore the extent to which such policies, by reducing existing market imperfections, can induce innovation of environmentally benign precision technologies and of technologies that reduce asymmetric information by providing detailed information about the characteristics of microunits, and can improve monitoring of pollution. References Alexander, K. and Donahue, V., 1990. Cleaner technologies in the tanning industry: the environmental challenge of the 1990’s, Proceedings of international conference of pollution prevention: clean technologies and clean products. Office of Research and Development, USEPA, Washington, DC: 19 – 31. Ayres, R.U. and Kneese, A.V., 1969. Production, consumption and externalities. Am. Econ. Rev., 59 (3): 282 – 297. Ayres, R.U. and Walters, J., 1991. The greenhouse effect: damages, costs and abatement. Environ. Res. Econ., 1(3): 23 – 26. Babcock, B.A. and Blackmer, A.M., 1992. The value of reducing temporal input nonuniformities. J. Agric. Res. Econ., 17 (2): 335 – 47. Barker, B.J., 1992. An engineered calcium carbide desulphurizer for lowering slag reactivity. CIM Bull., 85 (961): 119 – 121. Bates, R.W., 1993. The impact of economic policy on energy and the environment in developing countries. Ann. Rev. Energy Environ., 18: 479 – 506. Batie, S.S., 1988. Agriculture as the problem: new agendas and new opportunities. South. J. Agric. Econ., July: 1 – 11. Baumol, W.J. and Oates. W.E., 1988. The Theory of Environmental Policy, second edition. Cambridge University, NY. Beghin, J., Roland-Holst, D. and van der Mensbrugghe, D., 1995. Trade liberalization and the environment in the Pacific basin: coordinated approaches to Mexican trade and environment policy. Am. J. Agric. Econ., 77(3): 778 – 785. Berck, P. and Helfand, G.E., 1990. Reconciling the von liebeg and differentiable crop production functions. Am. J. Agric. Econ., 72(3): 985 – 96. Blackmore, S., 1994. Precision farming: an introduction. Outlook Agric., 23(4): 275 – 281. Bosch, D.J., Cook, Z.L. and Fuglie, K.O., 1995. Voluntary versus mandatory agricultural policies to protect water quality: adoption of nitrogen testing in Nebraska. Rev. Agric. Econ., 17: 13 – 24. 41 Braden, J.B., Netusil, N.R. and Kosobud, R.F., 1994. Incentive-based nonpoint source pollution abatement in a reauthorized Clean Water Act. Water Resour. Bull., Am. Water Resour. Assoc., 30 (5): 781 – 791. Buttel, F., Gillespie, G., Larson, O. and Harris, C., 1981. The social basis of agrarian environmentalism: a comparative study of Michigan and New York farm operators. Rural Socio. 46 (3): 391 – 410. Caswell, M.F., Lichtenberg, E. and Zilberman, D., 1990. The effects of pricing policies on water conservation and drainage. Am. J. Agric. Econ., 883 – 890. Caswell, M.F. and Zilberman, D., 1986. The effects of well depth and land quality on the choice of irrigation technology. Am. J. Agric. Econ., 798 – 811. Chernick, H. and Reschovsky, A., 1992. Is the gasoline tax regressive? University of Wisconsin-Madison, Institute for Research on Poverty, Discussion Paper 980 – 92. Cienski, T. and Doyle, D., 1992. Energy conservation in the comminution of industrial minerals. CIM Bull., 85 (958): 101 – 109. Dasgupta, P., Hammond, P. and Maskin, E., 1980. On imperfect information and optimal pollution control. Rev. Econ. Studies, 47 (October): 857 – 860. David, P.A., 1975. Technical choice, innovation and economic growth. Cambridge University Press, Cambridge. Denny, D., Freward, B., Pava, T.F. and Appley, E., 1995. Pollution prevention in the fabricated metals products industry. In: H.M. Freeman (Editor). Industrial Pollution Prevention Handbook. McGraw-Hill, NY. Dewees, D.N., 1983. Instrument choice in environmental policy. Econ. Inq., 21(1): 53 – 71. Dinar, A., Campbell, M.B. and Zilberman, D., 1992. Adoption of improved irrigation and drainage reduction technologies under limiting environmental conditions. Environ. Res. Econ., 2(4): 360 – 373. Dinar, A. and Letey, J., 1991. Agricultural water marketing, allocative efficiency, and drainage reduction. J. Environ. Econ. Manage., 210 – 223. Dinar, A. and Yaron, D., 1990. Influences of quality and scarcity of inputs on the adoption of modern irrigation technologies. West. J. Ag. Econ., 15(2): 224 – 233. Dinar, A. and Zilberman, D., 1991. Effects of input quality and environmental conditions on selection of irrigation technologies. In: A. Dinar and D. Zilberman (Editors). The Economics and Management of Water and Drainage in Agriculture. Kluwer, Boston. Downing, P. and White. L., 1986. Innovation and pollution control. J. Environ. Econ. Manage., 5: 1 – 25. Dubgaard, A., 1990. The need for a common nitrogen policy in the EC. In: R. Calvet (Editor). Nitrates-Agriculture-Eau (International Symposium, Nov. 7 – 8, 1990). Institut National de la Recherche Agronomique, Paris: pp. 131 – 136. Electric Power Research Institute, 1988. (EPRI) Proc.: Reducing electricity generation costs by improving coal quality, CS-5713, pp. 10 – 9 and pp. 9 – 33. Feder, G., Onchan, T., Chalamwong, Y. and Hongladarom, C., 1988. Land Policies and Farm Productivity in Thailand. Hopkins University for The World Bank, MD. 42 M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 Ferguson, W., Yee, J. and Parvin, D., 1993. Cotton yield and yield risk of Mississippi farmers using professional scouting services. Mississippi Agric. For. Exp. Station Bull. 1002. Fisher, A.C. and Rothkopf, M.H., 1989. Market failure and energy policy. Energy Policy, August: 397 – 406. Fuglie, K.O. and Bosch, D.J., 1995. Implications of soil Nitrogen testing. Am. J. Agric. Econ., 77(4): 891 – 900. GAO, 1994. Air pollution allowance trading offers an opportunity to reduce emissions at less cost. Report to the Chairman, Environment, Energy, and Natural Resources Subcommittee on Government Operations. House of Representatives. Washington DC. Guinn, L., 1994. Pollution prevention and waste minimization. Nat. Res. Environ., 9 (2): 10 – 12. Gutierrez, L.E., 1993. International power sector experience: a comparison with the Indian power sector. In: Conference of Power Sector Reforms in India. Report No. 166/94, ESMAP, Ind. Energy Dep., The World Bank, Washington DC. Hanemann, W.M., Lichtenberg, D., Zilberman, D. et al. Economic implications of regulating agricultural drainage to the San Joaquin river. West. Consortium Health Prof., San Francisco. Hansen, L.G., 1995. A variable tax-rate mechanism for regulation of non-point emissions-complex regulation problems. Presented 6th Annu. Conf. Eur. Assoc. Environ. Resour. Econ. Umea, June. Heffernan, W. and Green, G., 1986. Farm size and soil loss: prospects for a sustainable agriculture. Rural Sociol., 51 (1): 31 – 42. Helfand, G.E., 1992. The simple economics of pollution prevention. Toxic Sub. J., 12:1 – 11. Helfand, G.E. and House, B.W., 1995. Nonpoint source pollution. Am. J. Agric. Econ., 77 (4): 1024 – 1032. Heltzer, J.M., 1995. Wood furniture finishing. In: H.M. Freeman (Editor), Industrial Pollution Prevention Handbook. McGraw-Hill, NY. INFORM, 1992. Environmental Dividends: Cutting More Chemical Wastes. INFORM, NY. Jaffe, A.B. and Stavins, R.N., 1994. The energy paradox and the diffusion of conservation technology. Res. Energy Econ., 16 (2): 91 – 122. Johansen, L., 1972. Production Functions: An Integration of Micro and Macro, Short Run and Long Run Aspects. North-Holland, Amsterdam. Just, R. and Antle, J.M.. 1990. Interactions between environmental and agricultural policies: opportunities for coordination and limitations for evaluation. Am. Econ. Rev., Papers Proc. 80 (2), May: 197 – 202. Kalaitzandonakes, N.G., 1994. Price protection and productivity growth. Am. J. Agric. Econ.. 76 (4): 722 – 732. Kalaitzandonakes, N.G. and Dunn, E.G., 1995. Technical efficiency, managerial ability and farmer education in Guatemalan corn production: a latent variable analysis. Agric. Res. Econ. Rev., April: 36 – 46. Kalaitzandonakes, N.G. and Taylor, T.G., 1990. Competitive pressure and productivity growth: the case of the Florida vegetable industry. South. J. Agric. Econ., 22: 13 – 21. Khanna, M., 1995. Technology adoption and abatement of greenhouse gases: the thermal power sector in India. Ph.D. dissertation, University of California, Berkeley. Kohn, R.E., 1988. Efficient scale of the pollution-abating firm. Land Econ., 64 (1): 53 – 61. Kohn, R.E., 1993. An economic model of pollution prevention. Math. Comput. Modeling, 17(12): 21 – 28. Kosmo, M., 1987. Commercial energy subsidies in developing countries: opportunity for reform. Energy Policy. 17 (3): 244 – 53. Laffont, J., 1994. Regulation of pollution with asymmetric information. In: C. Dosi and T. Tomasi (Editors). Nonpoint Source Pollution Regulation: Issues and Analysis. Fondazione Eni Enrico Mattei, Kluwer Dordrecht: 39 – 66. Larsen, B. and Shah, A., 1992. World fossil fuel subsidies and global carbon emissions. Background Paper for World Development Report 1992, Policy Research Working Paper Series, WPS 1002, The World Bank, Washington, DC. Lashof, D.A. and Tirpak, D.A., 1990. Policy options for stabilizing global climate. Off. Policy Planning Evaluation, US EPA, Washington, DC. Letey, J., Dinar, A. and Knapp, K.C., 1985. Crop-water production function model for saline irrigation water. Soil Sci. Am. J., 49: 1005 – 1009. Lovins A.B. and Lovins, L.H., 1995. Profitably stabilizing global climate, Climate Change, 22(2): 90 – 97. Lovins A.B. and Lovins, L.H., 1995. Reinventing the wheels. Atlantic Monthly, January: 75 – 93. Milliman, S. and Prince, R., 1989. Firm incentives to promote technological change in pollution control, J. Environ. Econ. Manage., 17: 247 – 265. Mishan, E.J., 1993. The Costs of Economic Growth. Praeger, Westport, CT. Mongia, N. et al., 1994. Energy use and carbon implications in India: focus on industry. Energy Policy, Special Issue, September. Musser, W.N., Shortle, J.S., Kreahling, K. et al., 1995. An economic analysis of pre-sidedress Nitrogen test for Pennsylvania corn production. Rev. Agric. Econ., 17 (1): 25 – 35. National Research Council, 1989. Alternative Agriculture. National Academy, Washington, DC. Nivola P.S. and Crandall, R., 1995. The Extra Mile: Rethinking Energy Policy for Automotive Transportation. The Brookings Institution, Washington, DC. Norberg-Bohm, V., 1990. Potential for carbon-dioxide emissions reductions in buildings. Global Environmental Policy Project Discussion Paper, Energy and Environmental Policy Center, J.F. Kennedy School of Government, Harvard University, Cambridge, MA. Nowak, P.J., 1987. The adoption of agricultural conservation technologies: economic and diffusion explanations. Rural Sociol., 52 (2): 208 – 220. Nutzinger, H., 1994. Economic instruments for environmental protection in agriculture: some basic problems of imple- M. Khanna, D. Zilberman / Ecological Economics 23 (1997) 25–43 mentation. In: H. Opschoor and K. Turner (Editors). Economic Incentives and Environmental Policies. Kluwer, Dordrecht. Oster, S., 1982. The diffusion of innovation among steel firms: the basic oxygen furnace. Bell J. Econ., 13 (1): 45 – 56. Pimentel, D., and H. Lehman (Editors), 1993. The Pesticide Question: Environment, Economics and Ethics. Chapman and Hall, London, pp 231, 293, 317. Porter, M.E. and van der Linde, C., 1995a. Green and competitive. Harv. Bus. Rev., September – October: 120 – 138. Porter, M.E. and van der Linde, C., 1995b. Towards a new conception of the environment competitiveness relationship. J. Econ. Pers., 9 (4): 97 – 118. Portney, P., (Editor) 1990. Public Policies for Environmental Protection. Resources for the Future. Washington, DC. Reichelderfer, K., 1990. Environmental protection and agricultural support: are trade-offs necessary? In: K. Allen (Editor). Agricultural Policies in a New Decade. Resources for the Future and National Planning Association. Washington, DC. Romer, P., 1986. Increasing returns and long-run growth. J. Pol. Econ., 94(5): 1002 – 1037. Ruderman, H. et al., 1984. Energy efficiency choice in the purchase of residential appliances. In: Doing Better: Setting an Agenda for the Second Decade. American Council for an Energy-Efficient Economy, Washington, D.C. Sasmojo, S. and Tashrif, M., 1991. Measures for CO2 emissions reduction through energy price deregulation and fossil fuel taxation: a case study for Indonesia, Energy Policy, Special Issue, 19 (10), December. Sawyer, J.E., 1994. Concepts of variable rate technology with considerations for fertilizer application, J. Prod. Agric., 7 (2): 195 – 201. Schmitt, M.A. and Randall, G.W., 1994. Developing a soil nitrogen test for improved recommendations for corn, J. Prod. Agric., 7 (3): 328 – 333. Segerson, K., 1988. Uncertainty and incentives for nonpoint pollution control. J. Environ. Econ. Manage., 15: 87 – 98. . . 43 Shah, F., Zilberman, D. and Chakravaorty, U., 1993. Water Rights Doctrines and Technology Adoption. In: K. Hoff, A. Braverman and J.E. Stiglitz (Editors). The Economics of Rural Organization, Theory, Practice and Policy, published for the World Bank. Oxford University, NY. Smith, R.B.W. and Tsur, Y., 1995. Asymmetric information and the pricing of natural resources: understanding the case of unmetered water. University of Minnesota Working Paper. Snowden-Swan, L.J., 1995. Pollution prevention in the textile industries. In: H.M. Freeman (Editor), Industrial Pollution Prevention Handbook. McGraw-Hill, NY. Sterner, T., 1990. Energy efficiency and capital embodied technical change: the case of Mexican cement manufacturing. Energy J., 11 (2): 155 – 167. Thomas, J.K., Ladewig, H. and McIntosh, W.A., 1990. The adoption of integrated pest management practices among Texas cotton growers. Rural Sociol., 55 (3): 395 – 410. Wallace, A. and Wallace, G., 1993. Limiting factors, high yields and law of the maximum. Hort. Rev., 15: 409 – 448. Wiel, S., 1989. Making electric efficiency profitable. Public Utilities Fortnightly, 6: 9 – 16. Wiley, Z., 1985. Economic Development and Environmental Quality in California’s Water System. Institute of Governmental Studies, University of California, Berkeley. Wolf, S., 1995. Cropping systems and conservation policy: the roles of agrichemical dealers and independent crop consultants. J. Soil Water Conserv., 50(3): 263 – 269. Wu, J. et al., 1994. An evaluation of Nitrogen runoff and leaching potential in the High Plains. Draft manuscript, Dep. Agric. Econ., Oklahoma State University. Xepapadeas, A.P., 1991. Environmental policy under imperfect information: incentives and moral hazard. J. Environ. Econ. Manage., 20: 113 – 126. Yee, J. and Ferguson, W.L., 1994. Scouting and pesticide use in cotton production: a sample selection model. Presented at the Annu. Meeting Am. Agric. Econ. Ass., San Diego.