Long-range forecasting of the global warming alarm phenomenon: Predictions from structured analogies Kesten C. Green Graduate School of Business, University of South Australia J. Scott Armstrong The Wharton School, University of Pennsylvania International Symposium on Forecasting 2011, Prague 9:30 − 10:30 AM, 27 June, University of Economics, RB213 21 June, 2011, AGW analogies v16 1
Global warming alarm is anti-scientific Alarming forecasts of warming for public policy decision makers come from • processes that violate the scientific method, and are • promoted by a political movement*, that tries to • suppress scientific and other opposition. We use a method developed for forecasting outcomes of conflicts among interest groups to make forecasts about the movement. *See next slide 2
“Climate science mired in politics” Declaring that “science is politics in climate change; climate science is politics” , Union Environment Minister Jairam Ramesh on Wednesday urged Indian scientists to undertake more studies and publish them vigorously to prevent India and other developing countries from being “led by our noses by Western (climate) scientists who have less of a scientific agenda and more of a political agenda” . The minister’s remarks evoked a loud applause from over 100 Indian scientists who had gathered on the Indian Space Research Organisation (ISRO) premises in Ahmedabad… Indian Express, 9 June 2011 http://www.indianexpress.com/news/Climate-science-mired-in-politics--Jairam/801199/ 3
Analogies in forecasting Analogies are commonly used • in an unstructured manner , and • after making a forecast in order to support the forecast. Analogies do contain useful information and • can aid forecasts • if identified and analyzed by experts… • in a structured and unbiased manner . 4
Structured analogies proposed at the ISF in June 2003 and it was published in mid-2007 in the IJF Google search for “structured analogies” 2005, February 120 2008, February 500 2010, March 3,000 2011, June 4,040 Google Scholar cites IJF paper: June 2011= 25 “Structured analogies” & forecasting = 70 5
Validation of structured analogies method (SA) Findings from forecasts for 8 diverse conflicts: Method % accurate Guessing 28 Experts unaided judgment 32 (106) SA (2 or more analogies) 56 (97) Source: Green & Armstrong (2007). Structured analogies for forecasting. International Journal of Forecasting , 23 , 365-376. 6
Structured analogies applications • Recommended as a tool for improving thinking quality (Crittenden & Woodside 2007) • Design of sales promotion campaigns Providing decision makers with analogies improves creativity (Althuizen & Wierenga 2008) • Forecasting the economic impact of new policies (Savio & Nikolopoulos 2009) • Forecasting effectiveness of policy implementation strategies: working with semi-experts SA with semi-experts not better than UJ (Savio & Nikolopoulos 2009) • Time series forecasts of a chemical process SA better than Neural Nets (Xue & Zhu 2010) • Forecasting the effectiveness of policy implementation strategies SA better than UJ (Savio & Nikolopoulos 2010) 7
Structured analogies procedure for global warming alarmism 1. Identify possible analogies : Suggestions from experts with different views (invited alarmists). Literature review (e.g. Kabat 2009; Booker & North 2007). 2. Screen for similarity : Meet criteria? Outcomes known? 3. Code outcomes of analogous situations 4. Forecast target outcomes : Obtain forecasts by using predetermined rule to select outcomes (e.g., use the mode of the analogy outcomes). 8 8
Criteria for selecting analogies to the dangerous manmade global warming alarm 1. Based on forecasts of material catastrophe arising from effects of human activity on the physical environment 2. Endorsed by scientists, politicians, and media 3. Accompanied by calls for strong action 9
Our statement of the problem “The Intergovernmental Panel on Climate Change and other organizations and individuals have warned that unless manmade emissions of carbon dioxide are reduced substantially, temperatures will increase and people and the natural world will suffer serious harm. Some people believe it is already too late to avoid some of that harm. Have there been other situations that involved widespread alarm over predictions of serious harm that could only be averted at considerable cost? We are particularly interested in alarms endorsed by experts and accepted as serious by relevant authorities. ” 10
An early forecast*: The global warming movement will wither “Given the history of such environmental scares— over all of human history—my guess is that global warming is likely to be simply another transient concern, barely worthy of consideration ten years from now… ” (Simon 1996, p.266). *Based on unaided judgment, but supported by analogies. 11
Progress on indentifying analogies Began search on 25 January 2009,… is still in process. T o date, we have a list of 71 proposed analogies… of these, we judged 26 to be relevant. Information on the analogies is available at publicpolicyforecasting.com. 12
Assess similarity • Initial assessments by Green and Armstrong • Independent assessments will be sought from experts • Contact Kesten Green at kesten@me.com to participate 13
Relevant analogies T o date, we have tentatively identified 26 relevant analogies. Six examples: • DDT and cancer ( Silent Spring) 1962 • Electromagnetic fields and Childhood Cancer 1979 • Acid Rain in the U.S. 1974 • Uncontrolled reproduction and degeneration (Eugenics) 1883 • Population growth and famine (Club of Rome) 1968 • Natural resource shortages and economic collapse (Ehrlich & Ehrlich 1974) 14
List of the 26 relevant analogies Population growth and famine (Malthus) 1798 Organophosphate pesticide poisoning 1976 Timber famine economic threat 1865 Electrical wiring and cancer , etc. 1979 Uncontrolled reproduction and degeneration (Eugenics) CFCs, the ozone hole, and skin cancer , etc. 1985 1883 Listeria in cheese 1985 Lead in petrol and brain and organ damage 1928 Radon in homes and lung cancer 1985 Soil erosion agricultural production threat 1934 Salmonella in eggs 1988 Asbestos and lung disease 1939 Environmental toxins and breast cancer 1990 Fluoride in drinking water health effects 1945 Mad cow disease (BSE) 1996 DDT and cancer 1962 Dioxin in Belgian poultry 1999 Population growth and famine (Ehrlich) 1968 Mercury in fish effect on nervous system development Global cooling; through to 1975 1970 2004 Supersonic airliners, the ozone hole, and skin cancer , etc. Mercury in childhood inoculations and autism 2005 1970 Cell phone towers and cancer , etc. 2008 Environmental tobacco smoke health effects 1971 Population growth and famine (Meadows) 1972 Industrial production and acid rain 1974 15
Thumbnail sketches • Prepare descriptions of about one page for each analogous situation. • Obtain peer review by experts to improve objectivity, accuracy, and clarity. • This is process is still underway • Descriptions of 9 of the analogies are posted at publicpolicyforecasting.com 16
The sketching process: Mercury in fish analogy Date: 2004/2005 (Environmental Protection Agency) Forecast of impending disaster: Babies born with brain and nervous system damage Forecasting method: Extrapolation to a near-zero dose of a genuine effect from a large dose Actions called for: Pregnant women and others to avoid fish Regulations to reduce mercury emissions Forecast endorsements and challenges: Government agencies: “agency warns about danger of…” 10 million hits Scientists: “scientist warns about danger of…” 2 million hits Media: mercury in fish in Google News 618 hits Blog search 4.7 million; Google search 66 million Outcomes of the conflict $11bn p.a. cost of EPA regulation of electricity generators; consumer fish avoidance 17
Alarming forecasts unscientific So far , none of the analogous alarms were based on scientific forecasts; they were based on experts’ unaided judgments in the form of: n 1. Unrealistic mathematical models: 3 2. Extrapolating an effect from a large dose to an alarmingly widespread effect at near-zero dose: 7 3. Extrapolating that a weak effect might become important over time or for a large population: 16 26 18
How accurate were the alarming forecasts? Preliminary coding revealed of the forecasts made in the 26 analogous situations: categorically wrong 19 wrong in degree 7 accurate 0
Government role Among the 26 analogous situations, government intervention …was called for in 25 (96%). 20
Government policies adopted Among the 25 analogous situations that involved calls for intervention, government policies were implemented in 23. Government actions involved: Increased government taxes Increased government spending Restricting individual liberties 21
Did government intervention help? Among the 23 analogous situations in which policies were implemented: n Harm was caused 20 Policies were ineffective/uncertain 3 Policies were effective 0 22
Forecasts from the use of structured analogies Likelihood of: • alarming forecasts coming true: ~0% • government action being harmful: 87% 23
Recommend
More recommend