Trade secrets versus Cost Benefit Analysis

Trade secrets versus Cost Benefit Analysis

International Review of Law and Economics 22 (2003) 511–526 Trade secrets versus Cost Benefit Analysis夽 Ekkehard Hofmann a,∗ , Georg von Wangenheim b...

115KB Sizes 1 Downloads 80 Views

International Review of Law and Economics 22 (2003) 511–526

Trade secrets versus Cost Benefit Analysis夽 Ekkehard Hofmann a,∗ , Georg von Wangenheim b a b

Law Faculty, University of Hamburg, Edmund-Siemers-Allee 1, 20146 Hamburg, Germany Max-Planck-Institute for Research into Economic Systems, Evolutionary Economics Unit, Kahlaische Str. 10, 07745 Jena, Germany

Abstract We argue that administrative agencies should base their decisions on Cost Benefit Analysis (CBA) for two reasons: first the well known ex ante effect that CBA make public decisions more rational and second the little discussed ex post effect that the agency’s CBA provides valuable information to judicial or political bodies controlling the agency. The ex post purpose of CBA implies (judicial review of administrative decisions) or requires (political review) the public availability of the CBA. A conflict between this public availability and the protection of trade secrets arises, if the costs or benefits of a decision depend on information which are trade secrets. As a solution to this conflict, we propose a nonpublic in camera court procedure which leaves little discretion to the agency. © 2003 Elsevier Science Inc. All rights reserved. Keywords: Trade secrets; Cost Benefit Analysis; Public availability

1. Introduction Cost Benefit Analysis (CBA) serves two different, though closely related functions. The first function is to make public decisions more rational by properly evaluating and comparing all predictable effects of alternative decisions. We refer to this function as the ex ante function of CBA. Most of the literature1 on CBA investigates questions related to their ex ante function. Debates about the specifics and alternatives of CBA and about their merits with respect to their ex ante function have shown that none of the types is perfect and all 夽 Paper presented at the 18th Annual conference of the European Association for Law and Economics Vienna, September 13–15, 2001. ∗ Corresponding author. E-mail addresses: [email protected] (E. Hofmann), [email protected] (G. von Wangenheim). 1 Some examples are Zerbe and Dively (1994); Boadway and Bruce (1984, pp. 292–328); Pearce (1998); Dasgupta and Pearce (1972); Mishan (1988); McGarity (1998); Revesz (1999); Costanza, Farber, and Maxwell (1989). For a critical discussion of different approaches see Adler and Posner (1999).

0144-8188/03/$ – see front matter © 2003 Elsevier Science Inc. All rights reserved. doi:10.1016/S0144-8188(02)00117-5

512

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

have their drawbacks. In addition, all such techniques require some kind of aggregation of individual preferences to social preferences and thus suffer from the problems developed in Arrows impossibility theorem. The second function of CBA is to facilitate the control of public decisions by the judiciary or some political institutions: a CBA on which a public decision is based and which is available to the controlling body conveys a large amount of information to the controlling body—information which the controlling body would otherwise have to produce on its own, or even might never be able to gather. We label this function the ex post function of CBA. The ex post function of CBA used to be neglected in the literature. Only recently did Cass Sunstein (2001) and Eric Posner (2001) address the CBA’s function to facilitate judicial or political control of public decisions. Of course, the ex post function also serves the ex ante function: the easier it is to control public decisions, the stronger are the incentives of the decision-makers to weigh the pros and cons of a specific decision against each other in a proper way. However, the two purposes of a CBA are clearly distinct and only ex ante function is a widely discussed topic in the literature. Since we mainly deal with the ex post function of CBA, we follow Lesser and Zerbe (1998, p. 221) in using a rather broad definition of CBA.2 We call any decision procedure a CBA if it satisfies the following conditions: it must (1) enumerate the alternative outcomes of the decision, (2) describe the predicted consequences of these alternative outcomes and how these predictions are derived, (3) assign evaluations to these consequences which are numerical and based on the same dimension,3 and (4) choose the outcome for which the sum of the evaluated benefits exceeds the sum of the evaluated costs by the largest difference. We stress that we do not discuss where the evaluations come from—they may be based on the willingness to pay or the willingness to accept but also on personal value judgments of the decision-maker. In this paper, we first (Section 2) briefly review to what extent CBA are used in administrative decision-making in different countries. In Section 3, we discuss why judicial control of administrative decisions, in particular on large scale projects, currently, i.e. without a proper CBA by the administration, cannot comply with the self-imposed standards of courts to control whether “the objective weight of each and every single concern has been properly taken into account by the authority” (German Federal Administrative Court).4 We will argue that political control of administrative decision must suffer from similar pitfalls and that requiring the administrations to perform a CBA in the wide sense described in the previous paragraph solves the problem. We acknowledge that conducting a CBA is costly (Section 4) but we stress that the clear statement of numerical and comparable measures of costs and benefits is the last point to economize on. We also acknowledge and discuss in detail that in addition to these direct costs of a CBA, there may be indirect costs of published reasons of public decisions in the form of an inefficient revelation of trade secrets. In Section 5, we discuss how the courts should deal with trade secrets which are at the base of a public decision. Section 6 concludes. 2 “CBA is simply an accounting framework in which benefits and costs associated with decisions are set out for purposes of information and discussion” (Lesser & Zerbe, 1998, p. 221). 3 Typically dollars or another currency unit, but quality adjusted live years or even saved exemplars of an endangered species would do the job equally well. 4 For details regarding the applicable standard, see Section 3.3.

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

513

2. The application of CBA in different countries Even with our broad definition of CBA, its use differs largely between countries. In the US, CBA often is applied to (regulatory) rule-making but also to large scale single projects.5 This inclination to rely on CBA has its historic roots already in the 1930s. As Dorfman (1997a) points out, the Flood Control Act of 19366 for the first time in US legislation required administrative agencies to rely on CBA in their decision-making. The Act stated that projects should be approved “if the benefits to whomsoever they may accrue are in excess of the estimated costs.” In more recent history, the widespread use of CBA may partly7 be attributed to efforts of the US federal administration to require CBA since President Reagan’s Executive Order No. 12291. The cornerstones of this order are the requirement for Federal agencies to use a CBA whenever “significant regulatory action” is at stake.8 Pursuant to President Clinton’s Executive Order 12866, any regulatory action that is likely to result in a rule that may have an annual effect on the economy of $100 million or more or adversely affect in a material way the economy is subject to formal evaluation of costs and benefits.9 CBA and regulatory impact analysis have been employed by US EPA in a broad array of areas including agricultural pesticides, lead in drinking water,10 waste disposal (full cost accounting and centralized waste treatment11 ), ground water and drinking water regulation,12 and air pollution.13 Various state governments in the US are also required to employ CBA under state law.14 For instance, Washington’s agencies are obliged to “determine that the probable benefits of the rule are greater than its probable costs, taking into account both the qualitative and quantitative benefits and costs and the specific directives of the statute being implemented.”15 Even at local level, CBA is not unknown. A survey carried out by Zerbe and Dively (1991) covering 72 US cities showed that a substantial number of larger cities in the US make use of formal policy evaluations. Of course, European administrative and political decisions cannot be characterized by a complete absence of structured comparisons of costs and benefits of the alternative 5 Although our argument concentrates on large scale single decisions, we hold that its core is transferable to rule making. Nevertheless we do not deny the considerable differences between them (for details cf. Ogus, 1998, pp. 57–64). 6 Flood Control Act ch. 688, § 1, 49 Stat. 1570, 1570 codified as amended 33 U.S.C. § 701a (1994). 7 For a brief history of benefit-cost analysis see Chakravarty (1987) and Dorfman (1997a, 1997b). For further reference see Morrison (1998), and Porter (1995, pp. 148–189). 8 Executive Order No. 12866, 3 C.F.R. 638 (1994), reprinted in 5 U.S.C. s 601 (1994); Executive Order No. 12498, 3 C.F.R 323 (1986), reprinted in 5 U.S.C. s 601 (1988); Executive Order No. 12291, 3 C.F.R. 127 (1982), reprinted in 5 U.S.C. s 601 (1988), revoked in 1993. 9 Executive Order No. 12866, 3 C.F.R. 638 (1994), section 3f. A similar requirement is set up by the Safe Drinking Water Act (section 1412 (b)(3)(C), 42 U.S.C. § 300g-1). For application to EPA regulation on arsenic, see 40 CFR Parts 9, 141, and 142, pp. 7010–7023. 10 40 CFR Parts 9, 141, and 142, p. 1956. 11 http://www.epa.gov/ostwater/guide/cwt/cwteac.html. 12 http://www.epa.gov/safewater/economic.html. 13 http://www.epa.gov/ttn/ecas/reg doc.html. 14 Arizona, California, Colorado, Florida, Illinois, Oregon, Virginia. See Zerbe (2000). 15 Chapter 403, Washington Law, 1995 (Engrossed Substitute House Bill 1010, partially vetoed, effective July 23, 1995).

514

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

decisions. The European Commission, Directorate General for the Environment, uses formal approaches—mainly called “economic evaluation”—for policy proposal assessment.16 In Germany, section 7 subsection 2 of the Statute on General Rules for the Federal Budget (Bundeshaushaltsordnung) requires the federal17 administration to perform a CBA18 for any action which is relevant for the budget. One can find similar provisions in a number of more specific statutes. In practice, however, CBA are conducted only at a very superficial level, if at all. In Section 3, we give a typical example for how costs and benefits are dealt with in preparing state decisions on large scale projects.

3. Weighting interests with or without numbers and the scope of control 3.1. The Airbus example In this section, we first report on the case of the enlargement of the Airbus production site at Hamburg as an example for how administrative decisions are made without a properly conducted CBA. We then argue that this lack of a CBA, which is a general feature of German administrative decisions on large scale infrastructure projects, not only makes administrative decisions unpredictable but also conceals information needed for controlling the administration according to the rules stipulated by German courts. The plan of the Airbus consortium (now EADS: European Aeronautic Defense and Space Company) to produce a large aircraft19 requires the enlargement of an aircraft construction plant situated on the bank of the river Elbe. The site’s enlargement includes an extension of the factory’s runway by which a quarter of the river’s width will be cut off. Furthermore, the project includes filling about one-fifth of the Mühlenberger Loch, a freshwater tidal mudflat which is a unique resting site for migratory birds and spawning ground for fish for the lower Elbe River. The mudflat is 1668 acres (675 hectares) in size, and it provides critical habitat for 70 species of migratory birds. The Mühlenberger Loch received protection by the Ramsar Convention on Wetlands. It is an area which has not been classified as a “special 16 For an overview of project assessment, see http://www.europa.eu.int/comm/environment/pubs/studies.htm. Examples: European Commission, Auto-Oil II Cost-effectiveness Study, Draft Final report, August 1999 (available at http://www.europa.eu.int/comm/environment/enveco/auto-oil/aopces part1.pdf); research is done on strategic decision-making its formal evaluation (http://www.europa.eu.int/comm/environment/eia/sea-studiesand-reports/sea integration main.pdf), and for diverse environmental policies, see for instance European Commission, DG Environment, Economic Evaluation of Air Quality Targets, November 1998; for Tropospheric Ozone http://www.europa.eu.int/comm/environment/enveco/air/tropozone-a.pdf. 17 For the Länder, section 6 of the Statute on General Rules for Budgets (Haushaltsgrundsätzegesetz) stipulates the same requirement. 18 The term “cost benefit analysis” (“Nutzen-Kosten-Untersuchung”) was replaced by the term “efficiency analysis” (“Wirtschaftlichkeitsuntersuchung”) in 1997 by a “Statute on the further development of the Budgetary Law” (“Haushaltsrechts-Fortentwicklungsgesetz”). According to the reasons of the statute, the new term “efficiency analysis” is broader than the term “cost benefit analysis”: in the taxonomy of the legislator, the latter term applies only to actions which affect the economy as a whole (“Gesamtwirtschaft”) while the new term also includes actions which have no or only negligible effects on the economy as a whole but affect the federal budget or a state budget. In our broad definition of CBA, the new term is equivalent to a CBA. 19 Airbus A380, 260 ft long, 79 ft high, with a wingspan of almost 262 ft.

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

515

protection area” pursuant to Art. 4 sec. 1 of the directive on the conservation of wild birds (79/409/EEC),20 but presumably falls under its regime.21 The directive on the conservation of natural habitats and of wild fauna and flora (92/43/EEC) is also applicable.22 On one hand, the project generates benefits in terms of an expected increase in employment, in income and in tax revenue for the city and state of Hamburg, while on the other hand, it brings about environmental damages in the form of noise, air pollution, destruction of (parts of) biotopes and impairments of waterways with respect to their various functions as navigable waters, drinking water supply, wetland habitat for animals and plants. The enlargement of the production site is highly subsidized (in total about 1500 million euro). 3.2. The structure of agencies’ own proffered rationales Although the city of Hamburg was required by law to provide for a written explanation that states the essential reasons for permitting the project including all the pros and cons,23 the published reasons of the permit fail to include any statement of evaluations of the factual effects of the project. They rather proceed in the following steps.24 They first categorize the goods which are negatively affected according to how severely they are affected (“very high,” “high,” “intermediate,” “low,” or “no impairment”).25 In a next step, the authorities compare each impairment of a good (or group of goods) separately to the entirety of the benefits. Step by step, the published reasons justify the destruction of parts of the Mühlenberger Loch, the additional noise and air pollutant emissions resulting from additional air traffic, and the reduction of the value of real estate around the site by the total of the positive effects of the project. Without ever attempting to aggregate the costs of the project, the authorities then simply state that the positive effects on the labor market and the regional economy are of “exceptionally large relevance” for the metropolitan region of Hamburg and that the competent authorities, therefore, incur the considerable costs of the project. This assertion does not go beyond a mere claim that the Hamburg authorities complied with the German legal doctrine of fair commensuration26 and weighed the total of the pros against the total of the cons against each other. Such a segmented comparison of costs and benefits of a project is typical for reasons in German administrative rulings. Authorities segment either the costs or the benefits of a project, depending on whether they grant or deny permission, respectively. One may see the legal background for this segmented comparison of costs and benefits in a fundamental doctrine of German administrative law: parties affected by a public permit may challenge the violation of their individual rights in the courts, but not the violation of “objective” law as 20

OJ 1979 L 103, p. 1. See the “Basses Corbi`eres” judgment of the European Court of Justice C-374/98 (December 7, 2000), para. 47 (http://curia.eu.int/en/index.htm). 22 OJ 1992 L 206, p. 7. 23 For details regarding the applicable standard, see Section 3.3. 24 While this is not the actual sequence of arguments as presented in the decision, it is their logical structure. 25 One should note that these levels of how severely the goods are affected have no relationship to the importance of the goods for social welfare. 26 For details of this standard, see Section 3.3. 21

516

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

such (“Schutznormtheorie”).27 As a consequence, a plaintiff may argue that his or her own interest had been undervalued, but he or she will not prevail with the claim that other costs, e.g. destruction of values protected by the Federal Law on Nature Protection or the losses imposed on other individuals, had been undervalued so that the costs of the entire project are larger than its benefits and therefore the violation of his or her interests is illegal.28 The effect of the mentioned doctrine and the consequential reluctance of administrations to aggregate both the costs of a project and its benefits and then compare the total of the costs to the benefits is threefold. First, administrative decisions become less predictable relative to a comparison of the proper sums of costs and benefits.29 On the basis of the same factual predicted effects of a project and the same value judgements, administrations are able to justify both positive and negative decisions in a court proof way. For a positive decision, they compare each impairment of goods separately to the total of the benefits resulting from the project, and for a negative decision, they do the reverse: compare each benefit of the project to the total of the costs. Whoever wants to file a law suit against such a decision must show that the impairment of his own rights and not that of others or of merely “objective” law outweigh the entirety of the benefits or of the costs, respectively. It is obvious, second, that with such a weak standard of judicial review, the aggregation of the benefits or the costs of the project (whichever supports the administrative decision) need not be done in a sophisticated way, even a very crude aggregation will be sufficient from the point of view of an administration which eschews repeals of its decisions. Consequently, the reasons of public decisions in Germany rarely go beyond a poorly structured30 collection of benefits and costs of the project which are never aggregated in a proper way. Third and most important, without an explicit aggregation of costs and benefits on a common numéraire the administration avoids articulating its value judgments in a way which is open to judicial review or political control. There is no way courts could review the evaluations of (single) goods as long as public decision-makers do not aggregate the costs and benefits on a common numéraire, i.e. use numerical decision procedures to present the pros and cons. This is partly due to structural limitations of ordinary language (what do you get if you multiply “massive” by “intense” or sum them up?). As we will argue in Section 3.3, this does not accord with the language of the relevant rulings of the German Federal Administrative Court. In Section 3.4, we will extend the argument to, and concentrate on, political control of administrative decisions. 27 Federal Administrative Court (Bundesverwaltungsgericht), Official Collection of Decisions Vol. 1, p. 83; Vol. 2, pp. 203, 205–206; Vol. 16, pp. 187–188; Vol. 41, pp. 51, 63; Vol. 55, pp. 280, 285–286; Vol. 60, pp. 154, 156; Vol. 61, pp. 256, 262; Vol. 81, pp. 329, 334. 28 Obviously, this doctrine misses the truism that all evaluation is relative. 29 By this argument we do not imply that a state-of-the-art CBA necessarily induces perfectly predictable administrative decisions. Rather we acknowledge that estimation of costs and benefits for a CBA leaves ample space for different decisions, for instance due to different evaluation techniques (cf. e.g. Costanza et al., 1989) or different discount rates (e.g. Revesz, 1999). How severe this unpredictability of CBA-based decisions is, depends on whether one requires consistency in evaluations over time. Although we tend to favor rather strict consistency requirements, we leave this question for further research. 30 The table of contents of the reasons of the Airbus permit may serve as an impressive example, but any other German permission of a large scale project would also do the job.

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

517

The only way to avoid these problems is to require administrative agencies to explicate not only the factual background of their decisions but also their evaluations.31 Hence, in addition to what administrations already do, i.e. discuss the alternative outcomes of the decisions and describe the predicted consequences of these alternative outcomes and how these predictions are derived, administrations should be required to spell out on the basis of which evaluations of the concerned goods they aggregate the total of the costs and the total of the benefits of the projected decision.32 In other words, what we petition for is that administrative decisions be based on CBA in the sense defined in Section 1 and that the complete CBA be included in the reasons of the decisions. 3.3. Legal standards of review Our petition presupposes a specific understanding as to whether courts should be able to control the administrative decisions including the underlying value judgments. We argue that according to the standards laid out by the German Federal Administrative Court they should, and we concur to the extent sketched out in this section. Provided that at least one plaintiff has standing, the German administrative courts are required to review administrative decisions in the field of large infrastructure projects on the basis of the principle of proportionality and the “doctrine of fair commensuration” (“Abwägungsgebot”).33 What the courts claim to control in such reviews is stated in almost every relevant mandate of the German Federal Administrative Court since the 1970s. The agency is required to regard every relevant matter, and its predictions of effects must be reasonable. Moreover, the courts purport to investigate whether “the importance of private interests has not been misjudged and the objective weight of each and every single concern has been properly taken into account by the authority”34 as the crucial last step of the doctrine of fair commensuration in order to strike a balance of all pros and cons.35 31 Patrick McNutt, our esteemed discussant at the Vienna EALE-Conference, drew our attention to an alternative approach: a court-like procedure for making large scale administrative decisions. Such a procedure might produce all information needed to derive good decisions and would be more flexible than a legal requirement to perform a proper CBA. However, we conjecture that at the end of the day, the effects of such procedure would be the same as our proposed legal requirement to perform a CBA: since the party negatively affected by a projected decision could request a CBA at very low costs but could gain additional chances to challenge the decision in court, such request would be brought forward in every (court-like) proceeding to prepare a public decision on large scale projects. 32 We stress that value judgments do not only concern the evaluation of predicted outcomes of alternative public actions, but also the determination of these predicted outcomes: the selection of the predictive methods may substantially depend on value judgments like the relevance of low-probability outcomes or the concentration on a small number of alternative scenarios. 33 Federal Administrative Court (Bundesverwaltungsgericht), Official Collection of Decisions (Entscheidungen des Bundesverwaltungsgerichts, amtliche Sammlung) Vol. 34, pp. 301, 309; Vol. 48, pp. 56, 63; Vol. 52, pp. 237, 243; Vol. 55, pp. 220, 225–226; Vol. 56, pp. 110, 116; Vol. 59, pp. 253, 256; Vol. 61, pp. 295, 301; Vol. 64, pp. 270–272. 34 Expressed in this way for the first time in: Federal Administrative Court (Bundesverwaltungsgericht), Official Collection of Decisions (Entscheidungen des Bundesverwaltungsgerichts, amtliche Sammlung) Vol. 34, pp. 301, 309. The other citations given in footnote 33 repeat the same wording. 35 Federal Administrative Court (Bundesverwaltungsgericht), Official Collection of Decisions (Entscheidungen des Bundesverwaltungsgerichts, amtliche Sammlung) Vol. 75, pp. 214, 254; Vol. 104, pp. 236, 251.

518

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

We refer to this concept of intensive judicial review as the “justification standard”: the courts shall examine whether the agency’s own proffered rationale—consisting of predictions of the effects of a decision and the evaluation of these effects—justifies the decision. The courts shall deem the decision justified if and only if the (aggregated) pros outweigh the (aggregated) cons on the basis of the evaluations of the administration, provided that all these evaluations remain within the limits of the constitution. Administrative courts thus follow the doctrine of deference when it comes to value judgments. They will not set aside the authority’s decision unless one or more relevant administrative value judgments are deemed “extremely unacceptable.”36 Nevertheless, the application of the justification inquiry calls for sufficiently precise communication of the “importance of interests” and the “weight” of concerns. Specifically, the selected method of expressing importance or weight must allow for a transparent aggregation of all the effects brought about by different alternatives. In principle, words can do the job, although the use of numbers is by far more feasible to articulate comparative value judgments. By calling a matter, take for example air pollution, “important,” or a certain loss in biodiversity “massive,” the agencies remain silent on the weight of the respective concern relative to other goods at stake. Hence, on the basis of the information typically included in the reasons of an administrative decision on large scale projects,37 the courts are limited to testing the basic tenor of these administrative adjudications at a minimum standard: they can only set aside those decisions which cannot be justified on the basis of any constitutional set of value judgments in combination with the effects of the decision as they are predicted in the reasons (“justifiability standard”). In other words, the way public decisions are commonly presented deprives the courts of the option to perform the self-imposed check whether private interests have been “misjudged” or the explanation by the agency failed to comply with the “objective weight” of each and every matter concerned. This conflicts with the above outlined German Federal Administrative Courts’ opinions about the required scope of judicial review. Therefore, we admonish that the German Federal Administrative Court did never set aside an agency decision on the grounds that it was not based on a numerical presentation of the pros and cons during more than 30 years of applying the doctrine of fair commensuration. 3.4. Political control of administrative decisions As we showed in Section 3.3, presenting explanations of the pros and cons without employing numerical decision procedures puts a severe constraint to the scope of judicial review and apparently does not comply with the language of the German “doctrine of fair commensuration.” We now turn to the alternative to judicial review—political control of administrative decisions. Following the insights of Public Choice theory, economists today do not assume that public decision-makers—be they politicians or be they administrators—are benevolent. 36 Federal Administrative Court (Bundesverwaltungsgericht), Official Collection of Decisions (Entscheidungen des Bundesverwaltungsgerichts, amtliche Sammlung) Vol. 56, pp. 110, 122. The wording differs in the relevant rulings of the last decades, but the essence is always the same. 37 The situation for small scale projects is even worse, but this may be justified by the relatively small benefits from improving these decisions.

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

519

They rather assume that public decision-makers have their own preferences, possibly due to lobbying activities of interested parties, and thus must be subjected to some kind of control. In terms of Principal Agent Theory, public decision-makers are agents of a principal or of a multitude of principals: the people of a jurisdiction. We thus follow the Positive Political Theory (for an overview on this branch of literature cf. Epstein & O’Hallaran, 1999, and parts of von Wangenheim, 2000) in assuming that all institutions involved in public decision-making have their own preferences (value judgments) and private information on some aspects of their decisions. Different from that approach, we do not concentrate on the specifics of the interaction, but rather on how asymmetric information may be reduced in the problem of controlling administrative (or political) institutions making public decisions. Obviously, asymmetric information in the control of a decision-maker may substantially be reduced if the decision-maker is forced to make all reasons which were relevant for making his or her decision available to the controller. This information must cover both the factual background of the decision including the predictions of the effects of the decision which has eventually been taken and of its alternatives and the value judgment on which it is based. The information required for effective control is thus identical to the information provided by a CBA in the broad sense as we use the term in this paper. Stating what information is needed for an effective control still leaves open who would be able to process such information and would thus reasonably take the role of the controller. In a political system which is based on the separation of powers, administrative (as well as political) decisions may either be controlled by the judiciary or by (other) political institutions including the voters (the latter perhaps supported by mass media). Judicial control puts the courts in a role of super administrators: the courts duplicate the administrators’ decision based on the courts’ values judgments on how social preferences are, or—more profane—based on the courts’ preferences which may be influenced by interest groups as well.38 While value judgements of courts may be closer to efficiency related reasoning, and thus a control of public decisions exclusively by the judiciary may bring about more efficient results of the administrative cum judicial process, such an approach would violate the constitutional foundations of modern democracies.39 Political control in modern democratic systems, however, does not assign all aspects of control to the political institutions but also relies on the judiciary. In particular the control of procedural questions and inner consistency of the reasons of administrative decisions lies in the domain of the judiciary. In this realm, courts can substantially increase the costs of faking CBA and thus ensure that political control refers to the true reasons of the decision and not only to pretended reasons.40 Value judgments, though, are mainly left to the political sphere, especially when they refer to the aggregation of individual to social preferences. The judiciary controls only 38

For a discussion of this assumption, see von Wangenheim (1993). Some economists tend to give up these foundations in order to promote efficiency or liberty (e.g. von Hayek) as the ultimate value in society, but this is of course, based on a value judgement. We do not want to discuss this issue in any detail, but just take the structure of the existing modern democratic constitutions as given. 40 Posner (2001, p. 30) also sees the role of the courts as curb to fake CBA. Given his narrower definition of CBA (aiming at efficiency), he argues that this curb is not necessarily in the interest of the controller who need not be efficiency oriented. 39

520

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

the crudest violations (“extremely unacceptable value judgments” in the words of German Federal Administrative Court41 ) of generally accepted and constitutionally determined value judgements and leaves the control of less extreme value judgments to the political institutions. What “the crudest” means, i.e. how these control competencies are allocated among institutions, differs substantially between nations.42 These variations in degrees leave the substance of our argument unaffected as long as there is some strictly positive control of administrative decisions by political institutions. Admittedly, our petition to include CBA in the reasons of administrative decisions in order to ease control of such decisions including their underlying value judgments seems to become weaker the less courts control value judgments. However, deference of courts does not necessarily mean complete absence of control. Political institutions may step in. Whether and to what degree they do, depends on the costs of doing so. The information provided by CBA in the reasons of administrative decisions reduces these control costs. The ex post function of CBA thus provides justification for incurring the costs of performing CBA in addition to its ex ante function. However, democracies allocate the control function among courts and political institutions, controllers have to become informed about all essential background of the decision as argued above: unless both the competent court and the political entity who controls (be it the people or a representative body) are well informed about the decision and its background including all interests involved, judicial and political control, respectively, will fail and administrative decisions will deviate from the controllers preferences—however these preferences are determined. 3.5. Public availability of CBA Obviously, the information required for control must be known to all members of the controlling institution. If this institution is the people, the information must be public by definition, if the controlling political institution is a representative body, the information must be public to all its members. Hardly ever will it be possible to restrict the information to these individuals (unless they are very few) since they interact in the political sphere and at least some of them will make use of the information in the public. Thus political control requires or implies information of the general public. The effect is hardly different for judicial control for two reasons: first, judicial control must be triggered substantiated by some party affected by the public decision. While one could imagine a right to trigger judicial control without the necessity to substantiate the claim that the public decision violates some legal restrictions, such system would quickly overburden the judiciary with the necessity to duplicate all public decisions. If a party affected by the public decision must substantiate the claim that some legal restriction has been violated it needs to know, of course, the content and the reasons of the decision. 41 Federal Administrative Court (Bundesverwaltungsgericht), Official Collection of Decisions (Entscheidungen des Bundesverwaltungsgerichts, amtliche Sammlung) Vol. 56, pp. 110, 122. 42 For example, the US-American judiciary seems to be even more hesitant to decide on value judgments than the German judiciary. This is also reflected in the importance of publicity in administrative decision-making in the US and in Germany (cf. Rose-Ackerman, 1994, 1995).

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

521

Again, it will hardly be possible to restrict this knowledge only to the parties affected by the decision, even if it is substantially smaller than the entirety of the voters. The second reason why judicial control implies publicity of the reasons of a public decision, is the control of the judiciary. Judicial decisions are controlled in two ways: by an appeals process and, perhaps more important for the decisions of higher courts, by peer review in the scholarly legal discussion. The latter requires that judges not concerned with the case and legal scholars must be informed about the public decision, its reasons and the court’s appeals decision. Again, one can hardly imagine a restriction of information to this large group. We thus conclude that, independently of who controls, an effective control of administrative or political public decisions not only requires a CBA in the broad sense we use in this paper, but also requires or implies the information of the general public about the content and the reasons of the decision. 4. The costs of published CBA The question remains whether and to what degree an effective control makes sense. Costs and benefits of a published CBA thus have to be compared. The benefits are hard to determine, but some general statements are possible. The benefits, of course, depend on the interests which are at stake—small scale decisions like a single building permit for a small house for a single family may justify only a very superficial CBA, while for large scale projects like the enlargement or construction of an international airport, a profound CBA will avoid errors of substantial costs ranging in the millions or even billions of Euro. Also, for projects which are under heavy attack in the public (the media) a profound CBA may save vast costs in the form of “endless” political and judicial quarrels while the CBA of projects which trigger very little discussion before the public decision will save only very little litigation and public distress. On the costs side of CBA, there are two aspects, the first of which is very simple: the costs of performing a CBA, e.g. the costs of gathering information, the costs of developing reasonable forecasts of the effects of alternative decision, and the costs of preparing a written and publishable version of the analysis.43 These costs mainly depend on the size of the analyzed project and on the depth to which the analysis goes. The second type of costs is more subtle and will be at the base of this and Section 5. These costs accrue from the publication of the reasons of the public decision which may include trade secrets which will loose their value in consequence. If the public decision is based (or should be based) on information which the holder of the information does not want to be published because it would then lose its value, a trilemma results: • either the control of the public decision is effective, but some value of the project at discussion becomes destroyed,44 43 The cost of developing a system of value judgments for the affected goods may also be substantial if a multi-person body has to make the decision. However, with respect to value judgments, economies of scale will be substantial: once such value judgments have been made, they may (and should) be used for further decisions again and again. 44 Note that complete destruction of the value of the secret information would make it unnecessary to include it in the reasons as a consequence of which it would gain its value again.

522

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

• or the value of the private information is saved, but control becomes incomplete, • or the information is withheld from the public authorities and thus the decision is more likely to deviate from the preferences of the controller as the principal of the decision process. An example for the first case may be45 the enlargement of the airport of Stuttgart.46 The enlargement seemed to produce so large benefits to the owners and the customers of the airport, that the costs in the form of reduced land values around the airport were outweighed. However, some owners of real estate close to the airport claimed that their losses were broadly underestimated. Unfortunately, they felt unable to prove their claim because this would have required publication of trade secrets which were at the very basis of their profits endangered by the airport enlargement. Their claim was that putting the burden of proof on them would essentially reduce their choice to one between Scylla and Charybdis: either they lose their profits due to the enlargement of the airport or they lose their profits due to publication of their trade secrets. Requiring a publicly available record of the administrative decision on the permission for the enlargement of the airport in order to allow political control would thus destroy valuable resources.47 In other cases, the benefits from using land in a way which requires an administrative permission may depend on keeping some aspects of the projected land use secret. Revealing the secret in the public records of the administrative decision would destroy the benefits from the projected land use. Then the prerequisites of effective control may be welfare reducing: a socially valuable project may be given up because it cannot be performed without the administrative permission but becomes privately worthless if the permission is sought for and the information necessary to substantiate that the project is welfare enhancing is published in the procedure required for the administrative permission. Our initial illustrating case may serve as a (preliminary) example: EADS was not willing to provide information on how they came to the prediction that the plant at Hamburg would offer employment for 8000 people in the first step of the permission procedure. EADS denied this information based on the argument that this would weaken its competitive position in the market for aircrafts.48 Statutes of many countries acknowledge the basic conflict underlying the problem discussed here: the value of information as a first step of safeguards against activities (not only public activities) which have too severe external effects versus the value of trade secrets 45

Since the information was withheld, the truth will always remain with the victims. Bundesverwaltungsgericht (Federal Administrative Court, decision of October 5, 1990, reprinted in Neue Zeitschrift für Verwaltungsrecht–Rechtsprechungs-Report 1991, pp. 118–129), Bundesverfassungsgericht (Federal Constitutional Court, 1988, Official Collection of Decisions, Vol. 77, pp. 121–125). 47 On first sight, one might argue that this is only a loss to a private party, since either another private party is able to use the trade secret in a more profitable and thus socially preferable way or publication of the trade secret does not matter. However, if an owner of a trade secret must fear that his secret becomes published when he applies for a permission necessary to exploit the secret, he will have far less incentives to produce the secret, i.e. to invent, in the first place. By these disincentives to invent, the purely private loss resulting from the publication of the trade secret in the single case transforms into a social loss when one looks at the general rule. 48 Note, however, the example has two drawbacks: (1) EADS provided an expertise by PROGNOS on this matter to the Hamburg authorities after the public hearings and this expertise might enter the court procedures in a public way; and (2) EADS may have had merely a guess on the labor market effects when applying and the competition argument may have been a pretext to conceal the true lack of profound information. 46

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

523

and privacy as well as state secrets. The US Freedom of Information Act49 as well as the European Directive on Environmental Information and its pursuing transformation acts50 try to tackle the dilemma by enlisting types of information which need not or must not be revealed.51 As a general rule, the German Administrative Procedures Act grants the right to see the files to those persons directly involved in or affected by an (intended) administrative act and protects trade secrets against revelation in administrative procedures.52 The ostensible solution to the trilemma, namely to reveal the information only to the competent public authorities is at odds with the ex post function of CBA: without further communication of the information, control becomes impossible. In Section 5, we outline a general rule on how one could deal with this problem in administrative court procedures. 5. Proposed solution to the trade secrets trilemma To solve the trilemma described in the previous section, we propose that relevant trade secrets and other relevant information subject to privacy needs must be revealed only to the public authorities who make a public decision. In the reasons, of the eventual decision, the authorities then only publish the predicted influence of the trade secret on the benefits and costs of the decision taken and its alternative and mentions explicitly which part of the CBA is incompletely published due to the protection of trade secrets.53 The control of the public decision would then have to be left to the courts in an in camera procedure, i.e. a closed procedure in which only a very small number of people, possibly not even the parties or the attorneys, participate. Such procedure is unusual, but has been established for German administrative courts by the Federal Constitutional Court (1999: 1 BvR 385/90) in a similar field: the right of public employees to inspect their employment files. The case concerned a former high-ranking public employee who was to be concerned with classified matters but was to be fired because the German internal secret service (Landesamt für Verfassungsschutz) had doubts on his reliability as he seemed to “have a characteristic which might make him become a victim of a secret service conspiracy.” The employee quit the job in order to avoid being fired, but nevertheless suffered a severe breakdown of his career. The Constitutional Court ruled that the Administrative Court must get the complete employment files in an in camera procedure54 to decide whether the information of the internal secret service was rightfully classified. In a first step of the in camera procedure, the administrative court would have to decide whether the concealed information deserves protection as a trade secret or for privacy 49

5 U.S.C. 552 sec. (b). For example, sections 7 and 8 German Environmental Information Act. 51 Further examples are the newly enacted Freedom of information acts of the German Länder: section 7 Informationsfreiheitsgesetz Berlin (1999), section 11 Informationsfreiheitsgesetz Schleswig-Holstein (2000), and section 5 Brandenburger Akteneinsichts-und Informationsgesetz (1998). 52 Sections 29 and 30 Verwaltungsverfahrensgesetz. 53 If a larger political body (e.g. a state parliament) makes the decision, the information might be restricted to a small committee thereof which only reveals to the entire political body the data which would appear in the published version of the CBA of an administrative unit. 54 For the time being, the court restricted the in camera procedure to the presiding judge, but left the specifics of a new law on the in camera procedure to the legislator. 50

524

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

reasons. Once the court has so found, the court must replace the political control of the public decision, i.e. the court must control the underlying predictions and value judgments which are usually in the political realm. Four alternative rules could guide the court. In any of these rules, the court rejects the administrative decision, if there is no set of predictions of the effects of the decisions and value judgments which (a) allows the decision to persist when the concealed information is considered, (b) is not a “crudest violations of generally accepted and constitutionally determined value judgements,” and (c) is consistent with the other predictions and value judgments in the reasons of the decision. The four alternative rules differ in additional requirements for approval of the administrative decision: 1. Under the “permissive rule,” the court approves if and only if it can find at least one set of predictions of the effects of the decisions and value judgments which satisfy conditions (a) through (c). 2. Under the “restrictive rule,” which is at the other extreme of the four possible rules, the court approves if and only if allsets of predictions of the effects of the decisions and value judgments satisfy conditions (a) through (c). 3. Under the “extrapolating rule” which is between the two extremes, the court approves if and only if the set of predictions of the effects of the decisions and value judgments which the court thinks to be closest to the other predictions and value judgments in the reasons of the decision satisfies conditions (a) through (c). 4. Under the “replacing rule” which again is between the two extremes, the court approves if and only if the set of predictions of the effects of the decisions and value judgments which the court likes most satisfies conditions (a) through (c). As one can easily verify, the restrictive rule allows only very few public decisions based on concealed information to pass the in camera procedure, while the permissive rule is hardly a filter at all. The extrapolating and the replacing rules are somewhere in between. Comparing the rules, we think the permissive rule to be too permissive: it gives strong incentives to abuse the in camera procedure to avoid political control of public decisions. Suppose this rule governs the procedure. Then all affected parties in an administrative procedure will try to declare as much information they provide to the public authorities as confidential as possible. The reason is that it is much easier to capture the deciding body when relevant information is concealed. If the request is denied by the deciding body or by the courts, the requesting party may, in the worst case, be as well off as without the request but may also be better off. Similarly, the deciding body may, in the worst case, become subjected to public control by the courts, but may also avoid public control of its value judgments and predictions of effects. The permissive rule thus weakens the control function of CBA too much. On the other extreme, the restrictive rule will fail to alleviate the indirect costs of published CBA in most cases, since only very few cases will pass the test which the court applies. As a consequence, applicants for public law permissions will refrain from filing an application if they expect their project to be valuable only if the trade secret remains concealed. In addition, the rule cannot be applied if both the proponents and the opponents of the public decision want to rely on concealed information. However, the number of cases in which trade secrets or similar private information tips the administration’s CBA seems to be small. Then the costs accruing from this rule are

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

525

small compared to the increased risk of reduced political control of public decisions. As a consequence, we conjecture that the restrictive rule is optimal: courts should uphold public decisions which are based on concealed information only if all sets of relevant value judgments make the decision socially worthwhile.

6. Conclusions In this paper we have argued that a CBA should be at the base of every public decision. We defined CBA in a broad sense and allowed CBA of small scale projects to be superficial, but not disguise the predicted effects nor the underlying value judgments. Our main argument in favor of such requirement was not the ex ante function of CBA (the guidance of the deciding body towards a good decision) but rather its ex post function (allowing effective control of the decision). In discussing the costs and benefits of performing a CBA, we concentrated on the cost side. We showed that in addition to the direct costs of performing a CBA, the publication of the CBA which is required for political control and results from judicial control may conflict with society’s desire to protect trade secrets. While publication of trade secrets may sometimes also occur under the current weak standards for the reasons of a public decisions, the problem becomes eminent only when one accepts a CBA with its far more profound informational content as the standard of review. As a solution to the problem, we propose an in camera, i.e. a nonpublic, court procedure in which those parts of the reasons of a public decision which concern a trade secret are dealt with. Since an in camera procedure reduces the potential for political control of public decisions, we propose a relatively strict standard of review in the in camera procedure. The ex post function of CBA is under-estimated and under-discussed in the literature. Our paper only discusses some aspects and thus further research should dwell on questions related to the control function of CBA.

Acknowledgements We gratefully acknowledge the helpful comments of Patrick McNutt, our discussant at the 2001 EALE conference, and two anonymous referees. All remaining errors are ours.

References Adler, M. D., & Posner, E. A. (1999). Rethinking cost-benefit analysis. The Yale Law Journal, 109, 165–247. Boadway, R., & Bruce, N. (1984). Welfare economics. Oxford: Blackwell. Chakravarty, A. K. (1987). Cost-benefit analysis. In J. Eatwell (Ed.), The new Palgrave—a dictionary of economics (Vol. I, pp. 687–690). Costanza, R., Farber, S., & Maxwell, J. (1989). Valuation and management of wetland ecosystems. Ecological Economics, 1, 335–361. Dasgupta, A. K., & Pearce D. W. (1972). Cost-benefit analysis: Theory and practice. London: Macmillan.

526

E. Hofmann, G. von Wangenheim / International Review of Law and Economics 22 (2003) 511–526

Dorfman, R. (1997a). Forty years of cost-benefit analysis. In R. Dorfman (Ed.), Economic theory public decisions, selected essays of Robert Dorfman (pp. 323–339). Dorfman, R. (1997b). Why benefit-cost analysis is widely disregarded and what to do about it. In R. Dorfman (Ed.), Economic theory public decisions, selected essays of Robert Dorfman (pp. 372–376). Epstein, D., & O’Hallaran, S. (1999). Delegating powers: A transaction cost politics approach to policy making under separate powers. Cambridge, UK: Cambridge University Press. Lesser, J., & Zerbe, R. O., Jr. (1998). A practitioner’s guide to benefit-cost analysis. In F. Thompson & M. F. Green (Eds.), Handbook of public finance (pp. 221–268). McGarity, T. O. (1998). A cost-benefit state. Administrative Law Review, 50, 7–79. Mishan, E. J. (1988). Cost-benefit analysis: An informal introduction (4th ed.). London: Unwin Hyman. Morrison, E. R. (1998). Judicial review of discount rates used in regulatory cost-benefit analysis. University of Chicago Law Review, 65, 1333–1366. Pearce, D. W. (1998). Cost benefit analysis and environmental policy. Oxford Review of Economic Policy, 14, 84–100. Ogus, A. I. (1998). Regulatory appraisal: A neglected opportunity for law and economics. European Journal of Law and Economics, 6, 53–68. Porter, T. M. (1995). Trust in numbers: The pursuit of objectivity in science and public life. Princeton, NJ: Princeton University Press. Posner, E. (2001). Controlling agencies with cost-benefit analysis: A positive political theory perspective. University of Chicago Law Review, 68, 1137–1199. Revesz, R. L. (1999). Environmental regulation, cost-benefit analysis, and the discounting of human lives. Columbia Law Review, 99, 941–1017. Rose-Ackerman, S. (1994). The economic analysis of public law. European Journal of Law and Economics, 1, 53–70. Rose-Ackerman, S. (1995). Controlling environmental policy—the limits of public law in Germany and the United States. New Haven: Yale University Press. Sunstein, C. (2001). Cost benefit default principles. Michigan Law Review, 99, 1651–1723. von Wangenheim, G. (1993). The evolution of judge-made law. International Review of Law and Economics, 13, 381–411. von Wangenheim, G. (2000). Production of legal rules by agencies and bureaucracies. In B. Bouckaert & G. DeGeest (Eds.), Encyclopedia of law and economics (Vol. V, pp. 559–586). Zerbe, R. O., Jr., & Dively, D. D. (1991). The discount rate policy for municipalities (Working paper). Zerbe, R. O., Jr., & Dively, D. D. (1994). Benefit cost analysis in theory and practice. New York, NY: Harper Collins College Publishers. Zerbe, R. O., Jr. (2000). The role of benefit cost analysis in legal and policy decisions (Mimeo). Invited lecture at the 17th Annual EALE Conference, Ghent.