Projects

Benefits of Black Liquor Gasification

Project Description

Dr. Michael Farmer, Assistant Professor of Public Policy at GIT, is the principal investigator on the project and will perform the three studies with the aid of graduate students in the Public Policy department. Dr. Scott Sinquefield, scientific leader of the gasification research program at IPST, will coordinate the overall project and provide the necessary technical input for the three studies. Dr. Delmar Raymond, Director of New Technology Development at Weyerhaeuser Company, will act as an unfunded advisor and provide industrial perspective.

Environmental Impact Statement
The first task is to complete an environmental impact statement (EIS) on the introduction of BLG in the areas where it is economically feasible to introduce the operation over the next 2 decades.

An EIS is much as its name suggests: it is a catalogue of impacts on the environment from a proposed project or program. The EIS catalogue requires a researcher to mix empirical estimates over quantitative impact projections (such as ambient air quality changes, estimated impacts on wild-life or other projected physical effects) with a qualitative severity index that experts offer: i.e. is the expected change severe, nominal, or unimportant?

The EIS serves the Pulp and Paper Industry in two ways: first it details the positive environmental effects that appear after introducing BLGCC in numerous communities and second, it paves the way to conduct non-market valuations that assign economic, or monetary, values to those positive changes. The BLGCC study provides a unique opportunity to conduct both types of studies in sequence and to integrate them into a complementary exercise.

Data collection on environmental impacts will proceed by a mix of expert interviews and quantitative data nation-wide where BLGCC plants are likely to open. This phase of research will be conducted by three MS level Public Policy students over the fiscal year 8/02 - 8/03 with the close advice of Dr. Sinquefield and under the direction of Dr. Farmer. A survey and one-on-one interviews with technical experts on the expected reactions to environmental impacts will progress under close supervision by Dr. Farmer, who will begin to prepare for the technical demands over summer 02 (one month), through the school year and ending summer 03.

Oddly, EIS and non-market value practitioners view each other's approach as incomplete or biased, one a substitute for the other. We view the processes as complements and seek the joint deployment of both practices. EIS practices benefit Non-market valuation in three ways.

Subjective environmental impact assessments by ecologists, chemists, health officers and biologists jointly treat two dimensions of risk. First experts must assess how convincing they find the actual empirical estimates of environmental impact per se and second, experts must adjudicate how serious they believe a given impact to the environmental system under study to be. It is clearly difficult for evaluators to separate the two influences (how good is the estimate and how serious is the impact estimated), but high quality technical information on BLGCC processes allows relatively more certain impacts. This permits more focus on how uncertainty over possible environmental reactions to those estimated impacts affects subjective assessments of environmental change. Given the ability to focus more on the expected severity of a change rather than on the uncertainty of how much change actually occurs, industry and policy-makers receive a more readable EIS. Also non-market valuation economists, then, gain insight into expert reports over environmental impacts.

EIS reports are more useful to follow-up economic work if experts signal clearly the range of expected environmental impact. Simple signals affect directly economic estimation: are the changes linear with respect to the scale of the activity, does damage or improvement change at an ever increasing rate or at a diminishing rate. We propose to conduct an EIS with subjective conclusions that specify this dimension as the vehicle to assist in functional form estimates of economic benefits from environmental improvements. Strangely, very few EIS statements record this very important property; so our more informative EIS report benefits the public and for industry directly, but also allows for better economic analysis.

Scientific EIS reports often consider the instability of a bio-chemical system. Much as a high school titration experiment reveals a mixture in a vile as stable, resisting reaction to new chemical introductions until the color changes all at once, environmental improvement / damage may alter local stability risks in similar ways. Stability concerns are an often latent, undifferentiated impact in the EIS; and since thresholds are almost never estimated empirically unless stability has been breached, the concern enters through the subjective EIS evaluation without explicit exposition. Yet industry, the public and policy makers want to distinguish stability concerns from other, more marginal expected impacts.

Again the economist benefits from such distinctions as well. Statistical economic estimates may under-estimate both benefits and losses if 'threshold discontinuities' fall within statistical confidence bounds. Helping to elicit these risks more systematically will make for a more powerful and more convincing EIS(1) that the public and industry can use.

We believe the quality of data, access to technical expertise and the design of an EIS as both a stand alone document and a platform for future applied analysis leads to an accessible and informative EIS that benefits industry, the public and analysts who follow-up.

Given the size of the potential energy gains, one Impact of BLGCC is the reduction in use of other dirtier fuels that may contribute to atmospheric deterioration. The process still releases CO2 and steam. This benefit will be listed in physical quantities in the EIS and listed as an un-estimated non-market benefit in the second phase below. Range estimates from existing studies will be cited; but the state of the art in global climate change estimation suggests such an estimate is a study unto itself given existing research.

Economic Valuation of Environmental Change
With relatively good applied economic feasibility data, a Production Functions Approach can be used to assess the likely changes within the firm. Feasibility studies are more accounting based approaches to the capital demands. This leaves a more complete, traditional neoclassical production estimation required for our study to be accomplished.

Much of the primary data, in raw form, has been collected in order to complete profitability studies. The BLGCC process is technically separate from the primary functions of making paper other pulp products - they are parallel plant operations. So the capital costs can be measured against increased energy production directly for this segment of the total joint production process (energy output and paper).

What is unsure is if the delivery of more energy from the plant operations dedicated to treat black liquor affects the mix of variable input choices in the paper producing sector proper. It may be that energy services are weakly separable from other input demands wherein the delivery of cheaper energy affects only a small number of fixed asset decisions with very little impact on other inputs such as labor. Yet if reliable and perhaps cheaper energy supply are more substitutable (or relatively complementary) to other inputs, then the reaction of other production decisions in the paper output sector of the firm may be important to understand fully environmental and health affects.

A standard weak separability test among inputs will explore if a more complete production analysis of the firm is necessary. We do not expect this to be the case and a Non-market valuation study can proceed on BLGCC alone. To conduct the test, existing data used in other research [1-8] on input use and price response will allow a 'within group test' (weak separability test) to establish which inputs can reasonably be aggregated into a larger input factor group along with BLG source energy in a production function.

If other input components loosely 'grouped' with energy, then the presence of many noticeable inter-active effects requires the research team to estimate a system of inputs demands derived from an aggregate cost function. This allows for more stable parameters and unbiased confidence intervals on the individual input demand equations. We expect the number of observations over the industry to be small, so a full translog function (a second order Taylor series Cost Function) is unlikely to be feasible. We are hesitant to assume a Constant elasticity formula without more production review. If our weak separability tests allow inputs to be aggregated into a few input systems (3-5), the flexible functional forms including the Weibull, or a translog function reduced by F-tests to a manageable number of parameters. If the input demand system performs better, then we will use these input demands and test the expenditure constrained production system that this implies.

Existing studies seem to attribute the main contribution to energy supply in the paper production sector of the firm to reduced volatility in the input price of energy. In an expenditure constrained production function (production subject to a budget constraint is common in strained industries), the risk factor suggests that energy and, say, labor are separable but that the overall cost function is constrained to a budget region (capital markets may be reluctant to support full scale in a risky environment). In this case, a production estimate will generate a dual cost function different from a standard form. The combination of the input demand system and a cost function estimate (reconciled together), should provide an efficient estimate of any response from other inputs to the provision of greater and more reliable energy to the paper making operations. If any of these inputs affect the environment of community development goals, then this additional BLGCC economic impact can be established. Our expectation is that these will not be required, but they need to be checked.

The shift in production process requires a new set of input demand equations. Fortunately, most of the pulp paper production process remains unaffected; yet the changes in capital structure of BLGCC, already available as accounting data from feasibility studies, will require that we estimate input demand curve shifts (i.e. new projected regression coefficients). (2)

Non-Market Approaches
Most of the local effects include choices by laborers to engage in a particular employment activity, local demand for energy and its transportation costs, reduction of steam into the air and water, possible ambient particulates from existing production into the air and water that can be produced.

We propose a wage hedonic to estimate characteristics that draw or repel workers and the price, in wages, that induce them to enter or exit that job. Since some laborers also appear to make a workforce participation decision jointly, we will adapt the wage hedonic to a tobit analysis. To my knowledge this has not been done inside the structure of the wage hedonic which compels us to code the estimator.

We propose using averting expenditures to identify marginal market demand for special clothing of workers, purchases of pollution mitigating gadgets by households to reduce the impact of any exposure they may fear. While averting expenditures under-estimate the total value of pollution exposure as an existing entity, changes in expenditure at the margin are less biased in their effects on welfare estimates.

For externalities that transcend local impact, such as the marginal value of green-house gasses on the environment, the quantity change emerges from the production function (cost function and derived demand) estimates above. The value will be drawn from existing studies and sensitivity analyses will be conducted to express the range of impacts.

We remain alert to additional approaches and impacts as the EIS takes form that will dictate most of the local impacts. One of the challenges will be to estimate direct impacts site by site; but many smaller impacts will be location specific. Since we are evaluating an entire program, not a site, we hope to establish some standards for adopting direct estimation techniques fused with the judicious use of so-called benefits transfer, the practice of using existing site benefits estimates from another place for the existing site.

Finally, for local biotic water systems and ground water systems subject to possible thresholds effects, the benefits of lowering these potential environmental instabilities requires that we endogenize possible discontinuities into the production estimation and averting expenditure functions (which are non-linear) or recreation functions if biota or recreation sites are now at lower critical risk of a threshold effect change. This would be the first applied Safe Minimum Standard analysis conducted in this way, integrating an EIS warning into a full non-market valuation study.

There are several competing approaches for related statistical questions beyond the maximum likelihood structure that we suspect will be numerically intractable and not guarantee a globally most likely answer. There are Bayesian approaches, modified ridge regression models, mixed logit models with nested alternatives that include, as yet, wholly unobserved possibilities, et cetera. (3)

Finally an ex post compare and contrast between the complementarities and distinctions between EIS and economic welfare estimation makes for an easy summary that promises to assist the applied economist, but also to unify expert information into a more coherent and accessible form for policy-makers, industry groups and the local public.

Community Development Impact
These flow largely the derived labor demand schedules estimated above, buttressed by identification of candidates for economic survival under the "with' versus "without" conditions of a BLGCC operation replacing the existing boiler recovery operations.

A standard employment appraisal for the area (or county) will discern the penetration of the labor force potentially affected by the introduction of a BLGCC plant. If by chance the state extension service has conducted a regional input-output or even a simple industry survey, a signal regarding the local multiplier effects of a possible plant closing (or failing to expand) can be assessed. This analysis will be formal but largely not technical.

Also from an inventory of communities likely to switch in the next 12 years, we will explore their sources of energy and their current means through which they satisfy peak energy demands on a regular basis. Many rural areas where pulp and paper plants exist rely on hydro-power. With modest local growth, these turbines no longer satisfy peak energy demand. If the released energy from a local BLGCC operation can be shown to make up the difference, this perhaps small marginal energy contribution can be disproportionately valuable.

Policy implications abound. In general, if BLGCC reduces local energy dependence, lowers some caustic chemical risks, and shores up employment prospects, then the technical costs and risks of conversion constitute a compelling public interest. Shared technology adaptation, supplemental research and innovation grants and technical assistance through university extension are all viable alternatives with established precedent in rural community and manufacturing development programs nation-wide.

(1) It has long been established that persons value a small loss with much greater intensity than an equally sized gain, much more than many economists might expect. Experts clearly fall victim to the asymmetry and are charged as tending toward alarm, exaggerating loss risks. Yet if instability fears drive the asymmetry of harm over gain, then some of the distance is explainable and more directly policy relevant. Teasing out the degree to which instability affects an impact evaluation by better elicitation may make the gains herein more valuable to society and will improve the confidence in the information quality of the EIS report.

(2) It is this shift in capital on one dimension of production that makes the stable, joint input demand system method up front useful as it reduces the confidence interval around estimated demand curve shifts - more efficient cross effect variable estimation. Technically, we are moving from single production to a joint production process, but the effects appear separable (i.e. the energy production process in the ex post position can be run independently and some business economists and industry analysts wonder if this side operation can be sold to operate independently under contract.

(3) Depending on the structure, we hope to treat such an effect by refining or adapting programming methods of Gausian quadrature to project these instability prospects into a fully endogenous statistical estimation model. Forms of these solutions exist in studies of inventory adjustment by producers when future states of the world enter the possibility set that are not far enough beyond past experience for the producer that the solution is not tractable from a stable parameters alone.