Self-regulation and precaution

1. Self-regulation and precaution Mariam Thalos

2. Plan of the talk • The logic of CBA/RA and why it might be unsatisfactory as a means of precaution • Other “logics” of precaution? • 1. Risk aversion • 2. Bayesian decision • 3. “buying an option” • Proposal: Adaptation points

3. Wingspread statement “We believe existing environmental regulations and other decisions, particularly those based on risk assessment, have failed to adequately protect human health and the environment, as well as the larger system of which humans are but a part…. “While we realize that human activities may involve hazards, people must proceed more carefully than has been the case in recent history. Corporations, government entities, organizations, communities, scientists and other individuals must adopt a precautionary approach to all human endeavors.Therefore it is necessary to implement the Precautionary Principle: Where an activity raises threats of harm to the environment or human health, precautionary measures should be taken even if some cause and effect relationships are not fully established scientifically.” --Wingspread Statement on Precautionary Principle, 1998

4. 2 take-home points from Wingspread • Risk assessment has failed to protect. • More precaution!

5. What is precaution? Apparently, not risk assessment. But why not?…

6. Risk=Probability x Impact • Risk = weighted utility • Utility = relative value (on a ratio scale)

7. A utility matrix

8. RA/CAB involve a conversion of 2 ideas: Utility (often monetized) Expectation

9. The Expectation Equation • Example: EMV(my going to work today)=P(getting paid)xPay + P(not getting paid)x$0

10. Expected utility calculation

11. Example EU(my going to work today)= P(getting paid)xUtility(Pay) + P(not getting paid)xU($0)

12. RA/CBA is CONTRASTIVE.

13. Example EU(current environmental practices)= P(climate disaster) x Utility(climatedisaster but we got to do whatever we wanted) + P(small climate change) x Utility(small change but we got to do whatever we wanted) + P(no climate change) x Utility(no climate change and we got to do whatever we wanted)

14. Example (cont’d) EU(change environmental practices)= P(climate disaster) x Utility(climatedisaster even though we curtailed some of our previous practices) + P(small climate change) x Utility(small change and we curtailed some practices) + P(no climate change) x Utility(no climate change and we curtailed some of our previous practices)

15. So where has RA gone wrong?

16. in RA, all risk is equal: this is why we can sum increments

17. How to be cautionary (take #1): be AVOIDANT/AVERSE • NEGATIVE reference (e.g. climate disaster) • Regulatory goal: AVOID

18. What we know (?)

19. What we know(!)

20. A (VERY) different circumstance

21. A different type of situation deserves a different decision rule: MAXIMIN

22. MAXIMIN is an AVOIDANT imperative: ELIMINATE options with unacceptable outcomes!

23. a very different choice!

24. MAXIMIN is compatible with CBA • Why? Because CBA does not apply when MAXIMIN applies.

25. Varieties of MAXIMIN • (Partial): ELIMINATE options with unacceptable outcomes • (True): ELIMINATE all but the option/options with the highest Min;

26. Option elimination • is an enormously risk-averse strategy; Maximin advises it ONLY under conditions of extreme uncertainty (no knowledge of probabilities) • It removes your shot at good outcomes as much as it does unacceptable ones

27. Risk aversion is “folk” Normative theories advise against risk aversion, simply because it removes good outcomes from the table in what seems an arbitrary way

28. When to be averse:CPP---a “Core Precautionary Principle” (Gardiner 2006) • (1) the “absence of reliable probabilities”; • (2) there are “unacceptable outcomes”; and • (3) the “care little for gains” condition

29. ‘care little for gains’ “The person choosing has a conception of the good such that he cares very little, if anything, for what he might gain above the minimum stipend that he can, in fact, be sure of by following the maximin rule. It is not worthwhile for him to take a chance for the sake of a further advantage, especially when it may turn out that he loses much that is important to him.” --John Rawls

30. NOTES • Any form of Maximin will (in general) violate EU when applied in conditions where probabilities can be assigned; • “care little for gains” suggests a rationale (a rational source) of such violation

31. there is a logic to risk aversion: it begins with a REFERENCE

32. Prospects (a theory of risk aversion)(Kahneman & Tversky 1979) VALUE losses gains referencepoint

33. Reference points and RA/CBA • CBA is neutral on regulatory reference: its commendations are conditioned by ALL the utilities in your matrix, and all comparisons are global • No approaching, and no avoiding • But for circumstances that fall outside CBA’s regime…

34. Framing (how to locate the reference point) • Mr. Paul owns shares in Company A. During the past year, he considered switching to stock in Company B, but decided against it. He now finds out that he would have been better off by $1200 if he had switched to the stock of Company B. Mr. George owned shares in Company B. During the past year he switched to stock in Company A. He now finds that he would have been better off by $1200 if he had kept his stock in Company B. Who feels the greater regret? • Mr. Paul 8% Mr. George 92% (N=138)

36. adaptation • A reference point is an adaptation point (in the language of perception theory): stimuli above this point produce one sort of response, whereas stimuli below produce a response of a different sort • Having an adaptation point in the operating characteristics of one’s theory of VALUE does suggest a sort of caution: the sort we refer to as risk aversion. But is it a desirable sort of caution?

37. A victim compensation experiment • Miller & McFarland (1986) conducted a series of studies vis-à-vis abnormality and a victim’s fate. Subjects were presented with a brief description of an incident and then asked to indicate how much compensation the victim deserved. In one experiment, the victim was a man who’d been hurt in a robbery. • In one condition, the robbery took place in the store at which the victim shopped most frequently. And in a second condition, it occurred at a branch of the same store where the victim proceeded after finding his regular store closed for renovations. • The probability of being injured (shot) in either condition was rated very low. But subjects recommended significantly more compensation (more than $100,000 more) for the same injury in the exceptional circumstance (second condition).

38. Responsibility and precaution • Subjects thus seemed to think that the victim in the second condition was LESS responsible for his fate than in the first condition. Why? (“because through no fault of his own he went farther afield than he would have gone”?) • Norm theory (Kahneman & Miller, 1986) says it’s because we can think of a “highly available alternative” in which he would not have sustained the harm, so we attribute to it greater causal effectiveness than in a contrasting condition--in spite of judging both outcomes as equally highly unlikely • This suggests a certain way of thinking about caution: “don’t go outside the norm if you’re not sure”

39. Self-regulation physiology • Sensitivity to environment variables • Points of adaptation • Feedback loops (interlocking positive and negative)

40. Understanding path dependence • History (or at any rate history of what is “normal”) matters, or so suggests the compensation experiments • initial conditions matter • NOTE: history represented--as such--is nowhere represented in a utility matrix

41. Away from the open future • we need a map: Actual

42. Regulatory reference • back to the map:   Actual

43. Potential asymmetries:how approach might be different from avoidance • Approach gradient v. avoid gradient: one might be steeper than the other, depending on whether you’re focused on avoiding a negative outcome or approaching a positive one • One involves inhibition, the other production of behavior • Might an approach orientation be more inherently stable? Might an avoid orientation, because it does not involve an express end point, be inherently less stable? (Carver and Scheier believe this to be the case) • All overstated

44. distinction will help • goal-level strategy v. implementation-level strategy study Failing grade  distractions  Grade of A

45. Regulatory Focus:a “population-level” analysis of behaviors • approach orientation: positive-valence reference; aim: produce HITS (and avoid errors of omission); becomes more susceptible to errors of commission, overproduction of Type I errors (in Neyman-Pearson terms) • avoidant strategy: negative-valence reference; aim:produce MISSES (and avoid errors of commission; becomes more susceptible to error of omission, overproduction of Type II errors. • Asymmetry actually documented

46. focus • Go in • Keep out • We keep track of your record  

47. CBA is ergodic • future is not constricted by past • shakes off the effects of history • matrix of utilities is a depiction of a very open future (no regulatory focus)

48. Market economies are not ergodic • Examples: QWERTY keyboards, 4’8.5” railway track gauge, PC (not Mac) • We preserve history • This can lead to inefficiencies (so-called “market failures”)

49. WE are not ergodic • emotions and moral judgments keep track of history and “path” • we keep track of path, and in particular we track the ease of undoing outcomes • Is this what Wingspread is commending? Would we want to endorse it, if so? YES

50. In favor of focus • Reference points are useful as anchors • Regulation via reference points is a physiological universal vis-à-vis irreversible outcomes (ex: thermal regulation) • Regulation vis-à-vis a reference point as anchor is a way of exercising genuinely laudable (rational) precaution in relation to irreversible disasters when CBA treats increments of hangnail risk on a par with same-size increments of (irreversible) catastrophic risk.