Climate Change: Discounting and Damages PDF Fall 2024

Climate Change: Discounting and Damages EC 434/534 University of Oregon Emmett Reynier Fall 2024 Housekeeping... 1. Problem set 3 due Friday 11/15. 2. Problem set 4 is posted, is is due Tuesday 11/26 (before Thanksgiving), this is the last problem set. 3. Meet with me about your project if som...

Climate Change: Discounting and Damages EC 434/534 University of Oregon Emmett Reynier Fall 2024 Housekeeping... 1. Problem set 3 due Friday 11/15. 2. Problem set 4 is posted, is is due Tuesday 11/26 (before Thanksgiving), this is the last problem set. 3. Meet with me about your project if someone in your group has not yet! Remember that the written project and presentations are 12/3. 2 / 59 Any news you'd like to discuss? 3 / 59 Introduction 4 / 59 The greenhouse effect Greenhouse gases absorb infrared radiation from the surface of the earth. Doesn't make more heat reach the earth, but rather stops heat from leaving the earth → more gas means less heat escapes into space. CO , CH , N O , CFCs, etc. 2 4 X The greenhouse effect is a very good thing in the "right" amounts. It keeps the planet hospitable, unlike the moon. Venus has a much higher concentration of GHGs, resulting in a surface temp > 450 C.∘ 5 / 59 The science of climate change The greenhouse effect is high school-level physics that has been well understood since the 1800s. Much more complicated to measure and predict planetary level effects. We don't have an "earth 2.0" to run experiments on. Scientists are nearly unanimous in their belief that (a) humans are causing climate change, and (b) the impacts will be bad. Consensus is based on decades of work and multiple lines of evidence (GCMs, paleoclimate, instrumental temperature record, observed sea level rise, etc.) Skeptic arguments are pretty weak and appear to be based largely on poor analysis and political tribalism. 6 / 59 Heterogeneity in climate change Change in the number of very hot days Differences taken between 5 year moving averages in 1990 and 2019 A very hot day is when avg temp is above 90th percentile (~80F) 7 / 59 Heterogeneity in climate change Change in the number of very cold days Differences taken between 5 year moving averages in 1990 and 2019 A very cold day is when avg temp is below 10th percentile (~32F) 8 / 59 What to do about this? While climate science can tell us about physical effects and trends... It can't tell us how much it will cost and what our response should be. This is much more the province of economics (and political science, philosophy, etc.) E.g. How do we deal with uncertainty? Measuring costs and benefits across generations? Etc. This is about all we'll say about the physical basis of climate change. If you are interested in learning more, some helpful references include: Real Climate, Skeptical Science, Gavin Schmidt's (NASA GISS) excellent TED talk, and, of course, the IPCC. 9 / 59 The economics of climate change Climate change is the ultimate public bad The usual theories and policy responses apply, but the sheer scope and scale of the problem (incl. political) make it very difficult to implement them in the real world. We'll focus on two key topics: 1. Discounting (today) 2. Damage estimates (next lecture) Some things we're skipping: game theory (climate treaties and coalition formation), innovation (R&D and technological diffusion), etc. We also won't much into actual policy, although we've covered some of that in previous lectures (e.g. EU ETS). 10 / 59 Social cost of carbon Social cost of carbon (SCC) ∼ Present value cost of one extra ton of carbon. Basically, what is the marginal cost of emitting one extra unit of CO ? 2 When economists talk about putting a price on carbon, they are generally referring to the SCC. Aside: To calculate an SCC, you typically need an Integrated Assessment Model (IAM). A simulation model that integrates simplified versions of the economy and climate system. IAMs play a major role in climate change economics, but are another aspect that we'll only discuss in passing. 11 / 59 Discount rate basics 12 / 59 The "miracle" of compound interest Imagine a policy that will yield costs of $1,000 a century from now. Question: What is the PV from today's perspective? Answer: It depends on the discount rate, r. Seemingly small differences in r can make a huge difference over time... 100 r = 1% ⟶ PV = 1, 000/(1 + 0.01) ≈ 370 100 r = 3% ⟶ PV = 1, 000/(1 + 0.03) ≈ 52 100 r = 5% ⟶ PV = 1, 000/(1 + 0.05) ≈ 8 13 / 59 How to think about the discount rate Intuitively there are two ways to think about the discount rate 1. Opportunity cost consumption Abatement costs today requires ↓ consumption. Similarly, costs in the future require ↓ consumption then. Compare consumption across periods. 2. Opportunity cost of capital Focus on investments: Every dollar spent on abatement could have been invested elsewhere. The L-R return on the stock market is 7-8%. Implies that every $100 of environmental policy must return $108 of benefits in one year. 14 / 59 How to think about the discount rate Problems: If the environmental problem has a long time frame, are current measure of return on investment relevant? What if the decision to abate today affects the opportunity cost of capital? 15 / 59 Discount rate disagreement The discount rate is arguably the defining issue in climate change economics. It may seem a strange thing to fixate on everything else at stake Unfortunately, there is substantial disagreement about what the correct discount rate is As we have already seen, small changes end up making a huge difference over long time spans This controversy is perhaps best illustrated by the debate that followed the release of The Stern Review... 16 / 59 The Stern Review 17 / 59 The Stern Review Full Title: The Stern Review on the Economics of Climate Change (2006). Commissioned by the UK govt & led by (Lord) Nicholas Stern. LSE prof, former chief economist at the World Bank. While not peer-reviewed, it was a major, comprehensive analysis of climate change, taken very seriously. Published to much fanfare: Stern was flanked by successive UK prime ministers, Tony Blair and Gordon Brown, during the report's launch The main question: Taking the climate scientists' predictions at face value, what should we do? 18 / 59 The Stern Review Take-home message was for swift, decisive action. Business-as-usual (BAU) would be very bad. Irreversible changes Uneven distribution of benefits/damages. Some countries would likely benefit during transition (Russia, Canada, possibly the USA). However, even those countries would be harmed in the very long run. Poor countries would suffer most → Heat waves, dwindling water resources, disruption to agricultural systems, etc. 19 / 59 The Stern Review Stern's recommendations: We probably can't achieve CO stabilization at 450 ppmv (currently ± 2 415 ppmv and rising sharply). If we aim for 500--550 ppmv, we can achieve stabilization at a cost of 1% of GDP per year. This would slow/stop "catastrophic" climate change. Estimated carbon price of $85 per ton. Environmentalists, many politicians, and many economists loved this message However, many others were not impressed. The chief point of contention = Stern's low discount rate. Let's take a look at his framework... 20 / 59 The Ramsey Rule 21 / 59 The Ramsey Rule Due to Ramsey (1928): r = δ + η ⋅ g δ ∼ pure rate of time preference. η ∼ elasticity of the marginal utility of consumption. g ∼ per capita growth rate. In words: Discount rate, r, reflects our innate impatience, δ and the fact that future generations will probably be richer than us, η ⋅ g. Using the Ramsey Rule to derive a discount rate is not controversial. (Everyone uses it.) However, the choice of inputs can be. Stern used δ = 0.001, η = 1, and g = 0.013 Implies super low discount rate of r = 0.014 = 1.4%. 22 / 59 δ : Pure rate of time preference The tendency to discount future enjoyment because it is in the future. Impatience, mortality, morbidity, etc. Stern chooses δ = 0.001 using ethical arguments. Only reason it isn't zero is b/c he assumes there is a small risk of humanity going extinct (nuclear war, asteroid, etc.) Stern is in pretty good company here. Ramsey himself: We do not discount later enjoyments in comparison with earlier ones, a practice which is ethically indefensible and arises merely from the weakness of the imagination. (p. 543) However, some notable economists were not happy. E.g. Dasgupta (2006) insisted δ should reflect observed market savings and interest rates. (Although: Is from ought?) 23 / 59 η : Elasticity of MU of consumption How much the value of gaining (or losing) one unit of consumption changes as we become wealthier. Enters the utility of consumption function as u =. 1−η c 1−η Stern chooses a "unitary" elasticity η = 1. Gains to poor people are more valuable than they are to rich people. Hard to argue with the sentiment, but many felt Stern should have been even more progressive. E.g. Arguments that empirical literature shows η = 2 Assume Al is 100× richer than Bob. Now take $1 from Al and give it to Bob. How big is Bob's utility gain compared to Al's loss? If η = 1 ⟹ 100× bigger If η = 2 ⟹ 10, 000× bigger 24 / 59 g : Economy's per capita growth rate How much richer are we going to be in the future? Stern chooses a growth rate of g = 0.013 i.e. The economy will grow at 1.3% per year. Thinking back to the Ramsey Rule, the value of g is important because it acts as modifying variable on η. How we value tradeoffs in consumption between future generations and ourselves depend on how much richer they are than us. The least controversial of the Review's parameters, still can have massive impacts over time 100 Y (1.03) ≈ 20Y 100 Y (1.015) ≈ 4Y 25 / 59 Other views 26 / 59 Nordhaus William Nordhaus (Yale) was among the most prominent and luminary critics of The Stern Review. Particularly harsh on Stern's choice of a 1.4% discount rate → argued instead for a more conventional rate of 5%. Low discount rate overweights PV of (very) distant costs. Stern Review: "The cost of climate change is the equivalent of a 20% cut in per capita consumption, now and forever." Nordhaus: "In one respect, the Review inflicts cruel and unusual punishment on the English language." (Zing!) Nordhaus SCC closer to $10. (Recall: Stern $85.) Policy could be summarised as "wait and see" Lower carbon price to start, but ramp up R&D to absorb high costs in the future. 27 / 59 Nordhaus Nordhaus has since become much more pessimistic Nordhaus (2017) argues for a SCC of $31 per ton. Hasn't changed his mind about the discount rate, but rather our estimates of the damages and the underlying uncertainties → downside risk has increased. The subject of our next lecture... Asides: Nordhaus built DICE (Dynamic Integrated model of Climate and the Economy). The original and, probably, still most prominent IAM. Nordhaus was awarded the Nobel Prize in Economics in 2018. 28 / 59 Nordhaus 29 / 59 Nordhaus 30 / 59 Nordhaus 31 / 59 Weitzman Marty Weitzman of Harvard effectively argues that Stern was right, but for the wrong reasons. Stern's policy prescription of "act fast and act decisively" can be supported without his imposed choice of a low discount rate. Two key ideas: 1. Gamma Discounting AER, 2001. Uncertainty about the discount rate means it shouldn't be constant → Start at 4% and then decline to zero 2. Dismal theorem ReStat, 2009. Non-trivial risk of extreme events causes standard CBA to break down → Prudence as the optimal strategy. For less technical expositions of these two ideas, try here and here. 32 / 59 Weitzman Source: Arrow et. al. 2017. 33 / 59 Weitzman The effective discount rate declines with the passage of time because increasingly greater relative present-value weight is being placed on the states with the lower rates (e.g. 1%). In comparison, the higher-discount-rate states (e.g. 7%) are relatively less important over time, because their present value has been more drastically shrunk by the power of compound discounting at the higher rates. Weitzman, 2001 Translation: A high discount rate also "discounts" its own future weighting. 34 / 59 Sterner Thomas Sterner (Gothenburg): Another in the "Stern was right, but for the wrong reasons" brigade. Together with U. Martin Persson in "An Even Sterner Review (2008)" argues for the need to account for relative prices and imperfect substitutability between man-made and natural capital. Builds on previous work: Hoel and Sterner (2008) E.g. Agriculture "only" accounts for 25% of global GDP. However, wiping out global agriculture would cost much more than 25% of the world's wealth. Key idea: Relative price changes that arise from increasing environmental losses mean that aggressive climate policy can be sustained even in the presence of a high(er) discount rate. 35 / 59 Sterner Source: Sterner and Persson (2008) 36 / 59 Damages 37 / 59 From discounting to damages In the wake of the Stern report, much of the discussion in climate change economics centered on the discount rate. In recent years, the focus has moved back to trying to understand the likely impacts of climate change. What do we know about existing damages? What can we say about future damages? Who gains, who loses? Adaptation vs. mitigation? Good news: Major improvements in our understanding of climate impacts over the last 10+ years. Bad news: We've become a lot more pessimistic. 38 / 59 Uncertainty in IAMs From last time, recall that we need Integrated Assessment Models (IAMs) to obtain a social cost of carbon (SCC). Two main difficulties 1. Where do we get our inputs from (i.e. damage estimates)? 2. How do we summarise the damages (i.e. functional form)? 39 / 59 IAMs Source: Interagency Working Group on Social Cost of Carbon 40 / 59 Dated inputs? Source: Climate Impact Lab 41 / 59 Dated inputs? Despite substantial advances in climate change impact research in recent years, the scientific basis for damage functions in economic models used to calculate the SCC is either undocumented, difficult to trace, or based on a small number of dated studies. Here we present new damage functions based on the current scientific literature and introduce these into an IAM in order to estimate a new SCC. We focus on the agricultural sector (...) These new damage functions reveal far more adverse agricultural impacts than currently represented in IAMs. Impacts increase from net benefits of 2.7 $/ton CO2 to net costs of 8.5 $/ton leading the total SCC to more than double." Source: Moore et al, 2018 42 / 59 IAM functional form Quadratic damage functions Major IAMs basically assume some kind of quadratic damage functions E.g. The DICE Model from Nordhaus assumes that global GDP damages D at time t are function of global average temperature T AT 2 D(t) = −0.236[TAT (t)] Where does this functional form come from? Intuition: The effect of climate change will become increasingly worse as the world gets hotter. Empirics: Try to fit a regression curve through existing damage estimates/projections. The first of these makes perfect sense. The second sounds relatively easy, but has proved to be remarkably controversial... 43 / 59 Is climate change good for us? Source: The Spectator 44 / 59 Is climate change good for us? Source: Tol (2009) 45 / 59 Gremlins Not so fast: Tol (2009) dropped minus signs, introduced typos and wrong numbers, and used questionable modeling strategies. First Correction: "Gremlins intervened in the preparation of my paper The Economic Effects of Climate Change..." Second Correction (a correction to the correction!): "This version also contained errors that were soon pointed out by various researchers." Bottom line: Climate change will not be good for us. (More here and here.) 46 / 59 New damage estimates 47 / 59 New damage estimates Source: Burke et al, 2018 48 / 59 How are we improving? Improving our damage estimation Need a framework that is both plausibly causal and reflects adaptation 1. What is the causal effect of a change in climate on health, energy expenditure, labor, etc. 2. How will this effect change as the world gets richer Climate Adaptive Response Estimation (CARE) Two step approach in Carleton et al 2022 and Auffhammer 2022 1. Estimate many different response functions using weather variation 2. Explain variation in these response functions using observables (e.g. income or current climate) Then we can use that heterogeneity to to predict future adaptation 49 / 59 How are we improving? Source: Climate Impact Lab 50 / 59 CARE Assumption: Response function from hot places today tell us how a location that is currently warm/moderate will respond to the climate in the future if the baseline climate gets hotter. 51 / 59 Climate change in your city The cutting edge research is trying to provide extremely detailed projections of climate impacts, with lots of geographic variation. We now have a pretty good idea of how climate is going to affect different parts of the US and western Europe. E.g. Hsaing et al 2017 gives an interactive map of the US. Projections for the developing world still need some work. 52 / 59 Adaptation 53 / 59 The problem Source: Fuss (2014) Climate change and emissions scenarios 54 / 59 Adaptation The "Paris Agreement" notwithstanding, achieving a 2 ∘ C target requires fairly heroic assumptions Negative emissions! We are going to have to live with at least some climate change. Adaptation is about changing our behavior or systems to better respond to this new state of the world. The objectives of adaptation vary with individual and social preferences (e.g. risk). Maintain status quo: eliminate all impacts, good and bad. Reduce negative impacts to minimize welfare loss. Capture positive impacts to maximize welfare gain. 55 / 59 Adaptation versus mitigation Adaptation is typically private and local in scale We can rely on individuals, cities and countries to adapt optimally in accordance with their own self interest. Can estimate adaptation costs using revealed preferences, assume that households adapt to the point where the marginal benefit of adaptation equals the marginal cost Mitigation is public and global in scale Inherent difficulties of international climate treaties: free-riding, prisoner's dilemma, etc. Adaptation and mitigation are best viewed as substitutes. For many aspects of climate change, adaptation is "simpler" Exception is long-lived infrastructure, particularly w/uncertainty. Implies it own public goods aspect (e.g. levees against flooding). 56 / 59 Conclusions 57 / 59 Conclusions Climate change is the ultimate public good (bad). Global in scope and centuries in duration Social Cost of Carbon (SCC) is the PV cost of extra ton of CO emissions. 2 Discount rate is paramount, lots of disagreement over correct rate Worse: Small disagreements lead to huge divergence in optimal policy. Controversy in the wake of The Stern Review's low discount rate. However, even critics (e.g. Nordhaus) have become increasingly aggressive in their policy prescriptions Others (e.g. Weitzman, Sterner) argue that Stern was right, but for the wrong reasons ⟶ Uncertainty and relative prices both call for decisive action even in the presence of a high discount rate. 58 / 59 Conclusions Alongside the discount rate, damage estimates are the most influential and (albeit to a lesser extent) controversial aspect of climate economics. The "good" news is that we have made significant progress in our understanding of likely climate impacts. Unfortunately, we have generally become more pessimistic about these impacts. E.g. Early studies suggested net benefits in the agricultural sector from moderate warming. Newer research suggests net costs are much more likely. Increasing resolution at the local scale allows us to better pinpoint likely "winners" from climate change. Oregon does pretty well! (If the Cascadia Subduction Zone doesn't kill us first...) 59 / 59