Positive v. Normative Justifications for Benefit-Cost Analysis

Transcription

1 Positive v. Normative Justifications for Benefit-Cost Analysis James K. Hammitt Harvard University (Center for Risk Analysis) Toulouse School of Economics (LERNA-INRA)

2 Positive (descriptive) v. Normative (prescriptive) Conventional BCA includes mix of individual/public and expert/scientific inputs Values of benefits & costs based on individual preferences "Objective" assessment of risks and other consequences Individual behavior and perceptions sometimes inconsistent with economic model Cognitive errors? Oversimplified model? How should BCA incorporate departures from model? Populism v. paternalism? 2

3 Trouble in Happyville (Portney, 1992) You are Director of Environmental Protection All 1,000 residents of Happyville Believe (naturally occurring) contaminant in drinking water is carcinogenic Are willing to pay $1,000 per capita cost of treatment 10 leading risk analysts all agree While one can never prove contaminant is harmless All would stake their reputations on it You have communicated skillfully, but residents distrust government & want treatment What should you do? 3

4 Behavior often differs from standard economic model Behavior is wrong? Cognitive error (susceptibility to framing, nonlinear use of probabilities) Self-control problems (procrastination) Model is wrong? Omits important attributes (type of mortality risk) Idealized assumptions (perfect information & processing) 4

5 Costly information explains some deviations Simplest standard model assumes perfect information and zero processing cost Health & safety regulations of private goods are harmful if binding (limit consumer choice) Automobile airbags, food-safety standards With costly information, optimal for consumers to delegate information collection & processing Consumers will not (& should not) have good estimates of benefits, risks, costs Information disclosure may be harmful, if consumers will misuse it (over-react to selected attributes) 5

6 BCA includes two steps Predicting consequences of alternative policies Positive question Predict as accurately as possible, use descriptively accurate models Departures from standard economic model may be appropriate Evaluating consequences of alternative policies Normative question Use consumers (reflective, informed) preferences With costly information: Consumers may be efficiently ill-informed Departures from revealed preferences may be appropriate May be optimal to delegate evaluation (e.g., QALYs) 6

7 Alternative justifications for BCA Positive Kaldor-Hicks compensation test If value of benefits exceeds value of harms, winners could compensate losers leaving everyone better off So what? Normative Utilitarian Maximize social welfare (sum or other function of individual well-being) Weight monetary values inversely by income? Pragmatic BCA approximates utilitarian calculation and is practical On average, everyone gains Compared with what alternative? BCA protects against cognitive error Excessive salience of few attributes Rationale & assumptions open to public challenge 7

8 Fundamental questions for social Individual policy What is "Good" for an individual? By individual, do we mean family, household, community? Distributional When does more Good for some justify less Good (or forgoing more Good) for others? 8

9 Good for individual Consumer sovereignty = preference satisfaction Neoclassical economics Other concepts of well-being Capabilities (Sen), opportunities (Sugden) primary goods (Rawls), moral life (religions), QALYs (health economics) 9

10 Interpersonal tradeoffs & distribution Economics assumes there is no objective method to compare effects on individual utility Who suffers more from the "same" level of pain? Utilitarianism Assumes one can define utility units that are interpersonally comparable Practical (common) measures Money benefit-cost analysis (BCA) QALYs cost-effectiveness analysis (CEA) 10

11 Positive v. normative justifications for BCA Positive: If policy is adopted, could transfer payments be arranged such that policy with compensation yields Pareto improvement? Kaldor-Hicks compensation test Individual: Consumer sovereignty Distributional: Money metric Normative: Is affected population better off Even if they do not perceive themselves to be? Individual: Well-being may differ from preference satisfaction Distributional: Utility metric Social welfare may incorporate distributional considerations 11

12 Divergence between positive Individual & normative models Framing effects Omission v. commission WTP WTA disparity Discounting hyperbolic v. exponential Discount rate Qualitative attributes of mortality risks Non-proportionality of WTP to risk reduction Ambiguity aversion Disclosure and value of information Distributional Income, age & other adjustments (e.g., to VSL) 12

13 Positive & normative discount rates Ramsey model r = r + h g r = consumption discount rate r = utility discount rate h = elasticity of intertemporal substitution g = economic growth rate Author r h g r C tax (2015) Stern Nordhaus DICE DICE

14 Qualitative risk attributes Perception and tolerance of risk depends on Dread Uncontrollable, involuntary, catastrophic, inequitable distribution of benefits, affects future generations Uncertainty Unobservable, not understood scientifically, delayed consequences, newly recognized Limited evidence that these affect WTP to reduce risk Most good studies suggest modest effect (less than fold) Other qualitative attributes Natural v. artificial/synthetic chemicals Genetic modification v. conventional breeding 14

15 Non-proportionality of WTP to risk reduction WTP for (small) reductions in mortality risk should Increase with magnitude of risk reduction Be nearly proportional to magnitude of risk reduction (yield same VSL) For other goods (e.g., birds saved, lakes preserved) WTP should increase with magnitude (nonsatiation) No clear answer to how much increase is enough 15

16 Continuity of marginal rate of substitution implies near proportionality for small risk changes Wealth Indifference curve 0 Survival (= 1 risk) 1 16

17 Empirical results often show sensitivity, not proportionality WTP to reduce mortality risk (on foreign coach trip) from 8/100,000 (Jones-Lee, Hammerton & Philips, 1985) Risk reduction 4/100,000 7/100,000 Mean WTP VSL 3.4 million 2.2 million Median WTP = 50 for each risk reduction 42% of respondents would pay same amount for each risk reduction (8% would pay more for smaller risk reduction) 17

18 Non-proportionality of WTP to risk reduction Initial risk reduction is valuable, additional risk reduction is worth much less? Two actions that each reduce risk by Dr are valued more than one action that reduces risk by 2Dr + e Value action or consequence? 18

19 Ambiguity aversion Humans dislike risk, uncertainty, ambiguity, ignorance Risk: "objective" probabilities Uncertainty: subjective probabilities Ambiguity: unknown probabilities Ignorance: unknown possible outcomes Should we take greater precaution when risks are more uncertain? 19

20 Precautionary principle Activities which are likely to pose a significant risk to nature shall be preceded by an exhaustive examination; their proponents shall demonstrate that expected benefits outweigh potential damage to nature, and where potential adverse effects are not fully understood, the activities should not proceed 1982 UN World Charter for Nature 20

21 Perils of prudence (Nichols & Zeckhauser 1986) Conservative assumptions, worst-case analysis, ambiguity aversion can increase risk Technology Deaths Probability Expected deaths Ambiguous , Certain Using upper-bound risk estimates, Certain would be preferred to Ambiguous 21

22 Perils of Prudence If decision is repeated for 10 pairs of technologies (and risks are independent) Technology Deaths Probability Ambiguous < 1, Certain 1, Policy of choosing Certain (with smaller upperbound risk) is almost sure to kill more people 22

23 Value(s) of information Increase chance of choosing decision that is best (or not bad) for actual conditions "Expected value of information" in decision theory Overcome burden of proof needed to depart from status quo policy or default assumption Compensate for decision rule that does not maximize expected value of outcome Reassure decision makers and affected public that decision is appropriate Enhance compliance, minimize opposition & legal challenges Incorporate compliance and challenges as factors in analysis? 23

24 Information disclosure Provision of accurate information generally viewed as Not harmful Possibly beneficial Individuals may be misled Over-emphasize salient attributes (e.g., possibility of carcinogenesis, neglect probability) Aversion to irrelevant(?) attributes (e.g., synthetic v. natural chemicals, GMOs) Prohibiting (accurate) information disclosure may be appropriate Probative v. prejudicial value of evidence 24

25 Distributional income, age & other adjustments (to VSL) Intuitively and empirically, VSL varies across people and over time Wealth and income (increases) Age and life expectancy (theory and empirical evidence less clear) Applied BCA ignores differences Violates positive justification 25

26 Implications for benefit-cost analysis Positive Predict whether everyone in affected population would perceive self to be better off with policy (and appropriate compensation) Measure individual preferences and perceptions as accurately as possible Recognize differences by age, income, etc. Scientific enterprise, subject to empirical testing Normative Evaluate whether (everyone in?) affected population would be better off, even if they do not perceive themselves to be How are individual good & social welfare determined? Advocate policies knowing that affected population would reject Analysts are just another interest group? 26

27 Role of government Populist Direct democracy Use polls, focus groups to provide electorate what they want now Paternalistic Exercise leadership Lead electorate in direction they will ultimately recognize they should have preferred 27

28 Role of government Thomas Jefferson (1820) I know of no safe repository of the ultimate powers of society but the people themselves; and if we think them not enlightened enough to exercise their control with a wholesome discretion, the remedy is not to take it from them, but to inform their discretion by education. Edmund Burke (1774) Your representative owes you, not only his industry, but his judgment; and he betrays, instead of serving you, if he sacrifices it to your opinion. 28