Abstract

Benefit cost analysis in a variety of guises has established itself as a useful tool in public policy making. It is an approach widely adopted in appraising a wide range of infrastructure investments and has been regularly used in legal proceedings. In the context of this study, it forms a common procedure for assisting in the assessment of the social benefits and costs of airport investment. It is not, however, a technique without its limitations. Beside a range of technical concerns, conducting a comprehensive benefit cost analysis can be resource intensive and time consuming. Recently, there have been efforts to make its application more efficient by adopting benefit transfer procedures. This involves making use of findings from one study as inputs into other policy making activities. While applying secondary data to a new policy issue has a long pedigree, new areas of application involve taking non-market valuations of externalities from one study and transferring them to a different policy site. This paper looks at some of the limitations of employing benefit transfers and uses noise nuisance aspect of airport investment policy appraisal as an illustrative case. Based upon a meta-regression assessment of hedonic price models, the findings suggest that caution should be exercised in conduction benefit transfers.

Introduction

Benefit cost analysis, in one way or another, now forms the backbone for much of the analysis that underlies public policy making. For example, US federal requirements are that major regulatory programs must "…not be undertaken unless the potential benefits…outweigh the potential costs…" (National Archives and Records Administration, 1985). The benefit cost approach is not, however, without critics, and questioning ranges from such details as the legitimacy of attempting to place monetary values on all items in the benefit cost account to the intellectual validity of using partial equilibrium analysis in public policy making (e.g. Hoehn and Randall, 1989).

The focus here is on the more detailed level of assessing the viability of a particular approach to benefit cost analysis and particular the increasing use made of benefit transfers when conducting such analysis. Benefit transfers involves deploying valuations based on primary data gathering in a specific study to estimate changes in consumer's surplus with regard to another policy⁽¹⁾. The implicit assumption doing this is that parameters derived in one location, at one time and for one type of policy making decision, can legitimately be employed in other decision making exercises. This implies that either is some common parameter applicable to all studies or at least one can explain variations between studies so that parameters taken from one exercise can adequately be modified to be used in a second exercise. If neither of these conditions hold, original independent analyses may be required for each case study.

The issue is one of degree. No modelling exercise produces a completely accurate picture, and benefit transfers are no exception. The practical point is whether benefit transfers provide sufficiently acceptable accuracy for the task at hand. Acceptance criteria is likely to differ quite considerable across a range of applications (Smith, 1992)⁽²⁾. The focus here is on one particular type of transfer relating to the use of aircraft noise nuisance valuations, although the findings may have wider implications.

This study is concerned with examining which benefit transfers may be legitimate in the context of applying aircraft noise nuisance valuations derived at one airport location to decision making at other sites. Airports are considered environmentally intrusive, but as air transport is growing rapidly, there are pressures for both the expansion of air transport infrastructure and a more extensive use of existing facilities. Benefit cost procedures now normally form an integral part of assessing any major developments of large airports and increasingly being used in appraisals of smaller, local facilities. High levels of aircraft noise are usually the main concerns expressed by objectors to expanding airports infrastructure or when there are efforts to push more traffic through existing terminals.

There is a well developed methodology for placing monetary values on aviation related noise nuisance, but the approach to date has generally been to rely on individual case studies for each assessment. For reasons of economy and to speed up decision-making benefit, transfer may be seen to offer a tractable method of improving current practices. This study initially elaborates on both the benefit transfer concept and the airport noise nuisance issue prior to conducting an analysis of variations in noise nuisance values that have been found in a range of individual cases. The underlying hypothesis being that if there are wide and difficult to explain variations in the values found, then it is difficult to legitimize the use of benefit transfers in this context.

Benefit Transfers

The underlying idea of benefit transfers is that one can take findings or parameters from one case study and adopt them to assist in policy making elsewhere⁽³⁾. Conceptually, this is not a particularly new idea, and economists have a long tradition of applying parameters such as demand elasticities in consumer analysis or input-output coefficients in macroeconomic policy assessments in work other than that from which they were initially derived. Even in public appraisal procedures where non-market factors are of importance, benefit transfer has a long pedigree; in the United Kingdom's, Department of Transport's computerized COBA framework of road investment appraisal, standardized values and parameters synthesized from previous studies are often included in the benefit cost calculus. This has included values for reduced risks of accidents, for travel time savings and for changes in vehicle operating costs. In the United States, unit-day values were used as early as 1962 to evaluate recreational resources.

In practice, benefit transfer applications can be divided into three broad types: estimates based upon expert opinion; estimates based upon observed, revealed behavior; and estimates based upon stated, preference elicitation mechanisms (Brookshire and Neill, 1992). The distinctions, however, are somewhat blurred. Expert opinion, for example, is seldom formed in a vacuum but normally relies on judgments arrived at after assessing either revealed or stated preferences studies. The basis of these assessments are subjective and usually opaque. <--!Notes incomplete-->

The recent interest in benefit transfer is mainly associated with more fully incorporating environmental externalities into the benefit cost framework⁽⁴⁾. The issue is whether one can legitimately transfer non-market valuations of these externalities, particularly those valuations deploying stated preference (contingent) valuation techniques that have become a major focus of the literature (O'Doherty, 1995). This may be seen as a belated switch in emphasis from methodological concerns about intellectual legitimacy of alternative non-market evaluation techniques to questions of application and policy relevance of empirical findings.

While there have been important developments in evaluation methods in terms of revealed and stated preference methodologies, external environmental costs and benefits are often given only sparse and partial coverage in many benefit cost analyses. Increasing public concern about environmental implications of various policy options is leading to a broader approach to benefit cost analysis being sought. This is also taking place at a time when many inter-governmental (ranging from the European Union to World Bank), national and local agencies are being committed to the adoption of comprehensive appraisals of policy options and project proposals.

From a pragmatic perspective, benefit transfer has a number of attractions. In terms of financial expediency, taking parameters from one study or a synthesis of a set of previous studies and employing them more widely is much less costly than conducting separate evaluations for each individual decision made. Linked is a growing body of studies providing estimates of case specific parameters, and it appears sensible to see if they can usefully be mined for additional and useful insights (Bergh et al, 1997). Benefit transfer can also help streamline decision making. Deploying previously derived monetary valuations estimates of environmental externalities can significantly speed up what is often considered a lengthy process of information collection, collation, and analysis.

From a methodological viewpoint, benefit transfer may be seen to introduce a degree of consistency into decision making through the use of common parameters across studies. This may particularly relevant when the degree of accuracy in parameters does not have to be very high; as in the initial screening of projects. Luken et al (1992) for example, discuss benefit transfers in terms of establishing limits within which parameters may lie. It is also relevant when a large number of relatively standard but linked policy issues are being addressed.

Additional to public policy making, legal requirements to provide forms of compensation to those adversely affected by environmental degradation and legal processes often seek out evidence from earlier cases as precedents.

Benefit transfers, however, are not without their limitations. A central issue is the decision regarding which values can legitimately be transferred and which are study specific. In the latter case, benefit transfers may remain legitimate if appropriate adjustments can be made to allow for specificity in individual case studies. One criteria for deciding on the potential transferability of results is to examine the variability between previous case studies and to explore the extent to which that can be explained and allowed in subsequent transfers.

Airport Noise Studies

As a result of rising incomes, increased leisure time, the growing importance of service sector activities and improvements in air transport technology, air passenger traffic since 1960 has grown world wide at an average rate of 9% a year and freight and mail traffic by some 11% and 7% respectively. Further, all the indications are that as a sector, it will continue to expand into the foreseeable future, albeit at differential rates in various geographical sub-markets. It seems likely that passenger traffic will grow at a rate of between 5% and 7% into the foreseeable future (Boeing Commercial Airplane Group, 1996).

This growth has taken place when there has been comparatively limited expansion in the physical capacity of airport infrastructure. Initial excess capacity and better utilization of existing space have been the main facilitators of growth. The ability of segments of airport infrastructure to handle the forecast growth in traffic is now in doubt. Capacity has already been reached at 10 of the largest 46 airports in Europe and is being approached at another 16. As a consequence, there are pressures for additional physical capacity to be built (Comité de Sages for Air Transport, 1994). Similar problems exist in several parts of North America.

Investments in new airports and expansions together with the up-dating of existing ones are financially costly. Airports also generally impose a variety of serious adverse environmental effects on those living nearby. Despite much improved technology residents and businesses experience considerable aircraft noise. For these reasons, airport investments are generally the subject of benefit cost analysis. Putting a monetary value on the aircraft noise nuisances stemming from airport development as part of benefit cost calculations has a long pedigree dating back at least to the seminal work of the Commission on the Third London Airport (1971). The vast majority of evaluations has been case specific, and the methodologically generally deploys some form of revealed preference, hedonic price approach. These studies sought to indirectly place a non-market monetary value on the noise nuisance by examining the impact on property values in residential areas adjacent to an airport. Traditionally, they have looked at marginal changes in real estate values associated with additional units of airport related noise nuisance.

Recently, there have been a number of innovations in the evaluation methods employed, and several stated preference models deploying contingent valuation techniques have sought to directly elicit from individuals their willingness-to-pay for aircraft noise abatement (e.g. Feitelson et al, 1996). This broad dichotomy of approaches poses some problems for benefit transfers. Revealed preference and stated preference frameworks are based upon differing sets of underlying assumptions. The appropriate use of one or the other requires that the entire benefit cost framework into which they are fitted conform to their underlying theoretical basis. This is not a problem in itself, but it does mean there are inherent limitations to its general application.

More problematic is the considerable diversity of values generated both within and between the two groups of evaluation methodologies. With the hedonic pricing method, conventional surveys have produced a range of estimates of the impact of airport noise on local property values. Nelson (1980), for example, looked at 13 studies and produced a property value discount range of 0.4% to 1.1% for each additional decibel of aircraft noise nuisance. These values compare to the 2.4% to 4.1% for home owners in Feitelson et al (1996) stated preference work.

Meta-analysis

Meta-analysis was initially developed in the physical sciences and involves the statistical synthesis of existing case studies to extract additional information concerning, for example, representative parameters or factors (moderator variables) that result in study specific results. Glass (1976) provides a widely accepted formal definition as, "… the analysis of analysis...the statistical analysis of a large collection of analysis results from individual studies for the purpose of integrating the findings. It connotes a rigorous alternative to the casual, narrative discussions of research studies which typify our attempts to make sense of the rapidly expanding research literature."

Most of the statistical synthesis work concerned with environmental evaluation has employed some form of meta-regression procedure. In particular, it has been concerned with moderator variables that account for difference in the values of environmental externalities found in various case studies. Meta-regression analysis takes the general form (Stanley and Jarrell, 1989):

  bj = b + SakZjk + uj (j = 1,2,…L) (k = 1,2,…M) (1)

where bj is the reported estimate of the relationship of interest in the jth study from a total of L studies; b is the summary value of b, Zjk are variable that reflect the relevant characteristics of an empirical study that could explain variations amongst studies; ak are the coefficients of the M different study characteristics that are controlled for and uj is the error term

Transport induced noise nuisance values can be explored in this framework. The general functional form that seems appropriate when seeking meta-analysis moderator variables in this context can be summarized along the lines of equation 1. The right hand side of the equation effectively extends the argument of Hunter et al, (1982) that much of the observed variance in correlations across studies can be accounted for by three statistical artifacts: unreliable data due to small sample size, inter study differences in the reliability with which dependent and independent variables are measured, and inter study differences in restrictions of range.

Y = f (P, X, R, T, L) + Error (2)

• Y represents an outcome of interested. This may be a single measure, such dbA in the case of noise, or it may     reflect a variety of differing effects such as level, duration and pitch;

• P can be treated as the specific cause of the problem (such as air traffic levels and proximity to source);

• X represents features of those affected by the nuisance (such as individuals' age and income);

• R, since the analysis is not based upon primary data but rather the combination of other studies, represents the characteristics of the research methods used in each study (e.g. such as econometric or survey) and the data used (e.g. time series or cross-sectional);

• T indicates the period covered by each study to allow for any underlying dynamic effects (such as systematic changes in social preferences); and

• L is the location of the study (which could be spatial, such as urban and rural, or may relate to the country forming the basis for each data set).

Justification for this type of framework above is necessary. First, much of the work using meta-analysis in the physical sciences tends to focus on P and X in equation 1. This is mainly because of the ability of those working in these fields to compare strict experiments where methodologies are identical and results tend to be reported in a more standardized way. The importance of so-called "artifact effects", that are normally embraced in R in the social sciences, would seem to be much greater where the need for quasi-experimentation has often resulted in a considerable range of diverse model specifications, information gathering procedures and econometric estimation methods being adopted. This obviously poses practical problems when attempting to express various studies' approaches consistent with the inherently numerical variables that are the basis of most scientific meta-analysis.

The question of estimation methodology may also be particularly important in many sub-areas of economics now expanding. Such is the case with several aspects of the environmental economics because different procedures can embrace differing components of cost (Pearce and Markandya, 1989). In part, this is due to the elements that make up the external cost of many environmental considerations. Boulding and Lundstedt (1988), for example, argue that individuals' values influence behavior in at least three ways; observed choice (i.e. revealed preference), peoples' conversations (i.e. stated preference), and adaptations as part of the learning process. Smith (1989) suggests that all three provide insights into measuring the economic value placed on environmental amenities, although there is no reason to expect the values to be the same. A simpler breakdown argues that individuals' preference for environmental protection consists of three components:

Total Economic Value = User Value + Option Value + Existence Value (3)

User value is that which may be observed from behavior, option value is the result of individuals desiring to keep an environmental asset in case they wish to use it in the future, and existence value is the importance individuals attach to the very existence of an asset even if they never have personal contact with it. These categories can be further refined, but the point is that estimation techniques that rely only on revealed preference tend to understate option and existence values. Stated preference techniques can embrace them more fully, thus yielding a higher valuation. Similar types of issues can emerge in other areas of economics.

The same sort of argument apply to T and L, especially when time series and cross-sectional work is combined in the meta-analysis or there is considerable geographical variation in the sources of the data used. The importance of when and where the studies were conducted stem from the underlying economic preference model. At one level, preferences are context specific, and differing contexts may result in varied valuations. One of these element is information. Before the 1960s lead additive was added to fuel and was felt to have negligible environmental effects. But medical work brought this into question and thus, preferences were changed. We are now moving back in the other direction and concerns are being expressed about the health implications of the additives being used in place of lead. Equally, local conditions can be important because they can influence the choice set. A wealthy country or region could provide a richer choice set than a poorer one and preferences may be affected. Added to this are the constraints imposed by political and technical environments that are often space specific.

Applying meta-analysis to environmental evaluations has been done before, generally in the context of revealed preference valuations of externalities. Smith (1989) considered the empirical work of 35 hedonic price studies and found consistency in the results that emerged once allowance was made for local conditions and assumptions. Smith and Huang (1995) examined some 167 hedonic models of the marginal willingness to pay for reducing particulate matter in the air, although their meta-analysis forced them to rely on only 86. Their results show that market conditions and the procedures used to implement the hedonic models were important in explaining variations in the values individual studies produced.

Work by Smith and Kaoru (1990), looked at some 200 studies of recreational demand employing travel cost evaluation methods. They used regression based meta-analysis to examine the results of 77 of these studies. One of the main conclusions that can be drawn from this body of analysis is that variations across studies can often be explained in terms of the specific nature of the recreational resources and the underlying assumptions made in the estimation models employed.

While variations exist in the parameters examined, the evidence from this body of environmental literature is that they can, in part, be explained in a way that potentially allows for adjustments if they are adopted in benefit transfers. Findings from other related areas are less strong. Waters' (1996) meta-analytical work on travel time savings values and Button and Kerr's (1996) on the effectiveness of urban traffic restraint policies do not offer any statistical confidence that benefit transfers can readily be recommended in these fields. Similar conclusions can be drawn from the meta-analysis of the output elasticities of infrastructure investment (Button and Rietveld, 1997)

In terms of putting monetary values on airport noise nuisance, empirical findings of 18 economic studies deploying hedonic price techniques are examined using a simple meta-regression analysis. Looking at results produced through the application of a single revealed preference evaluation methodology is likely to produce conservative estimates of any limitations in adopting benefit transfers since the technique has been more uniformly applied than contingent valuation methods. If there is little consistency across hedonic price studies then there is likely to be even less across contingent valuation estimates.

Results of these studies, all based on hedonic price index methods and offering a common measure of noise value (the change in property value associated with a change in noise level) provide a range of estimated airport noise nuisance values (see Table 1). The set of estimated values has a standard deviation of 1.28. The studies do differ, however, in terms of when they were conducted, location of the airport considered and the level of aggregation at which the hedonic estimation was completed. To assess the importance of these case specific features and reflect the arguments of some advocates of benefit transfer, there is a need to transfer the entire demand function. The meta-regression embraces such effects in the form of dichotomous variables (Boyle and Bergstrom, 1992).

Table 1. Studies used in meta-analysis

Table 2 shows the parameters from the meta-regression exercise, together with indicators of their statistical significance. The results provide little by way of overall explanation for the variability in results or of any important influence that may be relevant. The low overall degree of explanation for variations in the noise nuisance values provided by the meta-regression, even allowing for the essentially cross-sectional nature of the analysis, offers little objective support for adopting any of the reported values, or average of them, for benefit transfer purposes.

This does not mean benefit transfers are invalid; any one of the values found in the case study may be appropriate for transfer in specific situations. The problem is rather instead one of selection. The mean value of case study parameters would certainly not seen appropriate. Any selection of a noise nuisance value from these previous studies would inevitably be subjective. Although the importance of expert opinion should not automatically be discounted (Button, 1997), this must inevitably raise questions about the widespread usefulness of benefit transfers if they amount to little more than legitimizing established judgments.

The argument about judgments in benefit transfers can actually be extended into the evaluation process itself. While all benefit cost exercises require making a large number of assumptions, the real issue is making these as transparent as possible. Therefore, if it does ultimately come down to the analyst selecting a particular case study's parameter to transfer, justification for the selection should be transparent.

  TABLE 2. Meta-regression parameters.

Conclusions

Benefit transfers are increasingly seen as cost effective and expeditious in incorporating environmental effects into benefit cost studies. In some areas, with due care and caution, there is evidence that one can adopt parameters derived in similar contexts for some aspects of benefit cost analysis. A major issue, however, is the generality with which this can be done across various types of parameters, especially when dealing with non-market valuations of external benefits and costs.

Examination of a range of values obtained for noise nuisances associated with airports indicates that there is little justification for applying benefit transfer procedures at this time. The relatively wide range of airport noise nuisance values previously obtained and the difficulty of explaining it also raises further questions involving the overall legitimacy of the techniques that have been employed. The range of parameters cannot be explained by factors such as the country concerned or the data based employed and this raises concern about the overall usefulness of the hedonic method or, at the very least, the way it is applied.

Before benefit transfers are too widely adopted in this age of comparative research austerity, there are grounds for advocating a more systematic and rigorous analysis of individual legitimacy. This testing should cover a range of approaches, not just the meta-regression or even the broader meta-analytical approaches, which is favored here. Taking values from current studies and back-casting them into earlier benefit cost studies could prove insightful, as might splitting samples used in externality evaluations and exploring whether the parameters estimated differ significantly between them. To test for temporal stability of parameters, follow-ups should be fostered with new case studies replicating old ones that use the same samples and methodology.

Our findings may also have implications on the ways primary studies are conducted. In the past, they have been case specific and little, if any, thought given to their potential deployment in benefit transfers. Recognition of this may lead to reconsidering how case studies are conducted and their results reported. There are arguments, for example, that on-going case studies be more systematic so that the results presented may be assessed more completely by anyone wanting to consider them for a benefit transfer exercise.

In conducting the case studies, an analyst may also take the view that the explanatory variables employed should be selected and specified in such a way that parallels information available for other potential case policy study sites. At the more fundamental level, the arguments of Loomis (1992) that more headway would be made by seeking an acceptable benefit function transfer methodology rather than simply looking to transfer parameters has intellectual merit and would make transparent the underlying assumptions transferred. None of this rules out the need for continued basic research into improving fundamental parameter estimation procedures.

References

Abelson, P. (1979), "Property prices and the value of amenities", Journal of Environmental economics  and Management, 6: 11-28. Bergh, J. van den, Button, K.J., Nijkamp, P. and Pepping, G.C. (1997)Meta-analysis in Environmental Economics, Kluwer, Amsterdam.

Boeing Commercial Airplane Group (1996) Current Market Outlook, 1996, Boeing, Seattle.

Borins, S. (1981), "Meiszkowski and Saper's estimate of the effects of airport noise on property values: a comment", Journal of Urban Economics, 9: 125-128.

Boulding, K.E. and Lundstedt, S.B. (1988) "Value concepts and justifications, in Peterson, G.L., Driver, B.L. and Gregory, R. (eds.) Amenity Resource Valuation: Integrating Economics with Other Disciplines, Venture Publishing: State College.

Boyle, K.J. and Bergstrom, J.C. (1992) "Benefit transfer studies: myths, pragmatism and idealism", Water Resources Research, 28: 657-663.

Brookshire, D.S. and Neill, H.R. (1992) "Benefit transfers: conceptual and empirical issues", Water Resources Research, 28: 651-5.

Brookshire, D.S.(1992) "Issues regarding benefits transfer", paper to the Association of Environmentaland Resource Economists Workshop, Snowbird.

Button, K.J. (1997) "The three faces of synthesis", to the European Science Foundation TERM Conference, Amsterdam.

Button, K.J. and Kerr, J. (1996) "Effectiveness of traffic restraint policies: a simple meta-regression analysis", International Journal of Transport Economics, 23: 214-225.

Button, K.J. and Nijkamp, P. (1997) "Environmental policy assessment and the usefulness of meta-analysis", Socio-economic Planning Sciences (forthcoming).

Button, K.J. and Rietveld, P. (1997) "Infrastructure policy and pan-European regional development", in Kohno, H., Nijkamp, P. and Poot, J. (eds.) Regional Cohesion and Competition in the Process of Globalisation, Spinger-Verlag, Berlin (forthcoming)

Collins, A and Evans, A. (1994), "Aircraft noise and residential property values: an artificial neural network approach", Journal of Transport Economics and Policy, 28: 175-197.

Comité de Sages for Air Transport (1994) Expanding Horizons, European Commission, Brussels.

Commission on the Third London Airport (1971) Report, HMSO, London.

De Vany, A. (1976), "An economic model of airport noise pollution in an urban environment", in Lin, S. (ed.) Theory and Measurement of Economic Externalities Academic Press, New York.

Dygert, P. (1973), Estimation of the Cost of Aircraft Noise to Residential Activities, Ph.D. Dissertation, University of Michigan.

Emerson, F. (1969), The Determinants of Residential Value with Special Reference to the Effects of Aircraft Nuisance and Other Environmental Features, Ph.D. Dissertation, University of Minnesota

Emerson, F. (1972), "Valuation of residential amenities: an econometric approach", Appraisal Journal, 40: 268-278.

Feitelson, E.I., Hurd, R.E. and Mudge, R.R. (1996) "The impact of airport noise on willingness to pay for residences", Transportation Research D: Transport and Environment, 1: 1-14.

Freeman, A.M. III (1984) "On the tactics of benefit estimation under Executive Order 12291", in Smith, V.K. (ed.) Environmental Policy under Reagan's Executive Order: the Role of Benefit Cost Analysis, University of North Carolina Press, Chapel Hill.

Gautrin, J. (1973), The Economics of Aircraft Noise: An Essay with British Data, Ph.D. thesis, University of London.

Gautrin, J. (1975), "An evaluation of the impact of aircraft noise on property values with a simple model on urban land rent", Land Economics, 51: 80-86.

Glass, G.V. (1976) "Primary, secondary, and meta-analysis of research", Educational Researcher, 5: 3-8.

 Hoehn, J.P. and Randall, A. (1989) "Too many proposals pass the benefit cost test", American Review of Economics, 79: 544-51.

Hunter, J.E.; Schmidt, F.L. and Jackson, G. (1982) Advanced Meta-analysis: Quantitative Methods for Cumulating Research Findings across Studies, Sage, Beverly Hills.

Levesque, T. (1994), "Modeling the effects of airport noise on residential housing markets: a case study of Winnipeg International Airport", Journal of Transport Economics and Policy, 28: 199-210.

Loomis, J.B. (1992) "The evolution of a more rigorous approach to benefit transfer: benefit function transfer", Water Resources Research, 28: 701-5.

Luken, R.A., Johnson, F.R. and Kibier, V. (1992) "Benefits and costs of pulp and paper effluence controls under the Clean Water Act", Water Resources Research, 28: 667-74.

McConnell, K.E. (1992) "Model building and judgment: implications for benefit transfers with travel cost models", Water Resources Research, 28: 695-700.

Mieszkowski, P. and Saper, A. (1978), "An estimate of the effects of airport noise on property values", Journal of Urban Economics, 5: 425-440.

 National Archives and Records Administration (1985) Presidential Proclamations and Executive Orders, USGPO: Washington.

Nelson, J. (1979), "Airport noise, location rent, and the market for residential amenities", Journal of Environmental Economics and Management, 6: 320-331.

Nelson, J. (1980), "Airports and property values: a survey of recent evidence", Journal of Transport Economics and Policy, 14: 37-52.

O'Byrne, P., Nelson, J. and Seneca, J. (1985), "Housing values, census estimates, disequilibrium, and the environmental cost of airport noise: a case study of Atlanta", Journal of Environmental Economics and Management, 12: 169-178

O'Doherty, R.K. (1995) "A review of benefit transfer: why and how", British Review of Economic Issues, 17: 1-15.

Opaluch, J.J. and Mazzota, M. (1992) "Fundamental issues in benefit transfer and national resource damage assessment", paper to the Association of Environmental and Resource Economists Workshop, Snowbird.

Paik, I. (1972), Measurement of Environmental Externality in Particular Reference to Noise, Ph.D. thesis, Georgetown University.

Pearce, D.W. and Markandya, A. (1989) Environmental Policy Benefits: Monetary Valuation, OECD, Paris.

Pennington, G., Topham, N. and Ward, R. (1990), "Aircraft noise and residential property values adjacent to Manchester International Airport", Journal of Transport Economics and Policy, 24: 49-59.

Price, (1974), The Social Cost of Airport Noise as Measured by Rental Changes: The Case of Logan Airport, Ph.D. Thesis, Boston University.

Smith, V.K. (1989) "Can we measure the economic value of environmental amenities?", Southern Economic Journal, 56: 865–878.

Smith, V.K. (1992) "On separating defensible benefit transfers from 'smoke and mirrors'", Water Resources Research, 28: 685-94.

Smith, V.K. and Huang, J-C. (1995) "Can markets value air quality? A meta-analysis of hedonic property value models", Journal of Political Economy, 103: 209-227.

Smith, V.K. and Kaoru, Y. (1990) "Signals or noise - explaining the variation in recreation benefit estimates", American Journal of Agricultural Economics, 72: 419-433.

Stanley, T.D. and Jarrell, S.B. (1989) "Meta-regression analysis: a quantitative method of literature surveys", Journal of Economic Surveys, 2: 161–70.

UK Department of Transport (1992) Assessing the Environmental Impact of Road Schemes, Report of the Standing Committee on trunk Road Assessment, HMSO, London.

Uyeno, D., Hamilton, S. and Biggs, A. (1993), "Density of residential land use and the impact of airport noise", Journal of Transport Economics and Policy, 27: 3-18.

Waters, W.G. (1996) "Values of travel time savings in road transport project evaluation', in Hensher, D., King, J. and Oum, T. (eds.) World Transport Research: Volume 2, Transport Policy, Pergamon, Oxford.