The economics of climate change: methods and implications for ecological economics. A discussion note The three pillars on which an analysis of society ought to rest are studies of economic, socio-demographic and environmental phenomena. Richard Stone, 1984.

Carlos A. López Morales Rensselaer Polytechnic Institute PhD Student [email protected]

Abstract Using a simple dynamic optimization model, this paper discusses the economics of climate change focusing on some issues highlighted in the aftermath of the Stern Report. The paper advances the notion that many of the conclusions obtained from economic models (i.e., the so-called “policy ramp”) are more a consequence of the particular modeling strategy than a consequence of substantial knowledge on the interactions between the economic and the climatic system. The paper documents some of the comments to the Stern Report coming from leading economists and highlights the unease on the conventional methods used in the economics of climate change. This paper also offers a methodological discussion of the conventional economic approach to climate change, specially regarding to its analytical and explanatory power. The existing deviations from conventional practices have some important implications for the field of ecological economics. Ecological economists are facing the question “How to deal with climate change issues from the field’s methodological perspective?” This paper offers a discussion of methods to contribute to the definition of the approach of ecological economics to climate change analysis.

Introduction Since its publication in the Fall of 2006, the economic literature criticizing the Stern Report began to identify two main methodological problems that reflect deep and non-trivial caveats in the way that the standard economic approach answers the question “what to do?” concerning global climate change. Major attention has being paid to discounting procedures, including both its rationale and its implications to the policy advice that arise from the standard approach. The academic debate promptly pointed out that the main differences between conclusions coming from the Stern Report and those coming from well-known economic models (such as Nordhaus’s DICE model) were due to the use of different discount rates to calculate present values of benefits and costs happening in the distant future. Additionally, academic commentators raised strong criticisms regarding the treatment of deep uncertainty not only in the Report, but also in the standard economic modeling strategy. In a nutshell, decision making and policy advice are not the same when unknown probability functions -concerning disastrous events– are gathered from small samples of observations. Similarly, now the literature also proves that taking into account potential disasters could even preclude the existence of any optimal mitigation policy. Mainstream academic observers consequently suggest that the conventional economic approach is not very useful to deal with the kind of uncertainties involved in the climate change area, thus not very useful in providing a compelling answer to the policy puzzle of determining the timing of optimal mitigation policies.

This paper advances the notion that the main policy conclusions coming from the standard approach (embodied in the so-called “mitigation policy ramp”, widely referred in the relevant literature) have their roots in the particular consumption-smoothing modeling strategy, and not in some specific knowledge of economy-climate systemic interactions. The basic framework presented in this paper is useful to organize the above topics for discussion, and could help to transmit in a relatively easy way some of the main theoretical problems surrounding the conventional economic approach to climate change. The model easily replicates the optimal policy ramp, and illustratively represents the differences in optimal mitigation policies when using different discount rates. This paper also offers a methodological discussion of the conventional economic approach to climate change, specially regarding to its analytical and explanatory power. As can be seen in the current literature, some economists of the main theoretical stream are recently moving away from economic marginal analysis and towards the precautionary principle (as Martin Weitzman), or recognizing the fallacy of misplaced concreteness (as Partha Dasgupta), a concept long ago offered to the ecological economics jargon by Herman Daly. These deviations from conventional practices have some important implications for the field of ecological economics. Ecological economists are facing the question “How to deal with climate change issues from the field’s methodological perspective?” This paper offers a discussion of methods to contribute to the definition of the approach of ecological economics to climate change analysis. Section II presents a simple dynamic optimization model that represents some of the issues of the economics of climate change. Section III presents a brief summary of comments from leading economists that are present in the aftermath of the Stern Report. Finally, section IV concludes with some methodological reflections regarding the notion of misplaced concreteness.

II. The economics of climate change: a very simple model Consider that society needs to consume at each moment of time some amount of a particular nonrenewable resource defined as the capacity of the atmosphere to stock greenhouse gas emissions coming from economic activity. Assume that some amount, which can be measured in metric tons of carbon dioxide or, alternatively, in how many parts per million of it there are, has been defined by climatologists as being relatively safe regarding its climate change implications. Call this safe level S*.1 Thus, S* defines the non-renewable stock that is available to human “consumption”, that is, the amount below which economic activity can stock emissions in the atmosphere without provoking severe disruptions in the climatic system. Consider a general framework in which society plans in an infinite time-horizon. We have it that society gets some satisfaction from emissions. The reason is straightforward. Economic activity, which we consider as being the main source of human-produced GHGs emissions, provides economic welfare and development. Assume that social satisfaction, or “welfare”, or “utility”, is increasing in consumption but that it exhibits a decreasing marginal behavior. 2 Further, assume society discounts future satisfaction at a non-negative rate, defined to be the “discount rate of time preference”, reflecting its degree of temporal impatience. Then, the very basic problem of the economics of climate change is that society seeks the optimal consumption policy that maximizes the present value of the utility stream subject to feasibility constraints, which in this case implies that stocked emissions in the atmosphere does not exceeds the safe-level of atmospheric accumulation. If solved analytically, it is convenient to present the problem in continuous time. Formally, we have 1

2

In what follows we arbitrarily assign S* as equal to 10´000 “units” of emissions. These units could be defined to be parts per million or million tons, and therefore the specific number could be accordingly changed. This is a common assumption in economic theory. Decision makers enjoy consumption in such a way that even when they gather positive amounts of satisfaction with each additional unit of good consumed, they gather it at decreasing rates. In mathematical formalism, this means that the utility function is increasing, but that the second derivative is negative.

∞

max ∫ e − t U C tdt C t  0

subject to

I.

∞

∫ C t dt =S * 0

C t ≥0 Where U(C(t)) represents the utility function, C(t) is the consumption in moment t, ρ≥0 is the pure rate of time preference and S* is the available stock of the non-renewable resource. We have made two implicit assumptions that do not alter the objectives of the model. First, we have assumed that emissions, here represented by C(t), accumulates linearly in the atmosphere. That is to say that if society send, say, 10 units of emissions to the atmosphere, then it consumes the same 10 units of the available amount to be accumulated in the atmosphere. We also neglect the decay rate of accumulated GHGs. This rate, however, is very slow (some researchers suggest that 15% of every emission unit will be stocked in the atmosphere even in 500 years in the future). 3 Here, we assume that once you send an emission unit to the atmosphere, it stays there forever. Before looking to the solution strategy to Problem I, we would like to express the feasibility restriction in a more tractable way. In Problem I, we have it that ∞

∫ C t dt=S * . 0

To make things easier, we define a new variable, S(t), to reflect the amount of resource that is still available at instant t: t

S t =R−∫ C  d .

(1)

0

Now, differentiate the above expression with respect to time to capture the dynamics of the available resource stock: ∂S =−C t. ∂t Thus, Problem I turns into the following problem: ∞

max ∫ e

− t

(2)

U C t dt

C t  0

subject to ∂S =−C t  ∂t C t ≥0

I'.

Th solution routine begins by defining the current value Hamiltonian function: H =U C t −t C t  ,

(3)

where t  is the co-state variable and, in this case, it represents the shadow price (in terms of the 3

See Maier-Reimer and Hasselman (1987) and Azar and Sterner (1996) for a discussion on this.

objective function) of the atmospheric capacity to store GHGs emissions. Denote with UC (C(t)) the partial derivative of U(·) with respect to the control variable, C(t). The first order condition to maximize H with respect to C(t) is U C C t =t . (4) Also, from the Maximum Principle it follows that ∂ − t=0, (5) ∂t which is a differential equation with solution in t= 0 e  t , (6) meaning that the shadow price grows at a constant rate equal to the discount rate used in the discount factor. Differentiating (4) and combining with (5) yields, ∂C 1 − = , (7) ∂t C   C  −U CC C , which is the expression for the consumption-elasticity of marginal utility. UC The parameter   C  is not a constant (unless assumed): it may depend on the level of consumption. It is also a measure of the curvature of the instantaneous utility function, which in turn is interpreted as being a measure for inequality aversion associated with U(·). Expression (7) characterizes the optimal consumption policy: non-negative parameters  and   C  imply that the optimal policy is timedecreasing. Earlier generations enjoy higher consumption of the non-renewable stock than later generations. If sustainability is defined as either generational equal access to the available stock (which would follow Brundtland's definition), or as a situation in which this economy exhibits a constant national product, as measured by the current value Hamiltonian (see Weitzman, 1976), this conventional result led researchers (for example, Farzin, 2002) to conclude that a cake eating economy like the one of problem I can not be sustainable. Further technical assumptions allow to extract more information on the nature of the optimal policy defined in expression (7). If U(·) belongs to the family of isoelastic utility functions, then   C  is a constant, which allows to easily obtain an explicit ∞ function of the optimal consumption policy {C *  t  }t =0 . In such a case, we may have − t  * * (8) C  t = S e  .  Many lessons can be learned from expression (8), and they represent closely some aspects of the debate surrounding the Stern Report, specially those related with its parametrical treatment. Figures 1, 2, 3 and 4 below display a small sensitivity analysis on (8) with different values of parameters  and   C  . To isolate the effect of discounting, we first restrict the example to the special case in which η is a constant. In particular, we have assumed that U  C  t  =ln  C  t   , which is an isoelastic utility function for which   C  =1. As can be seen in the figure below, discounting affects not only the rate at which consumption is declining, but also the level at which the optimal policy starts at the beginning of the planning horizon. Assuming that the externality problem of climate change is observable to the central planner from t=0, this would mean that the degree of social patience affects the optimal amount in which society starts to concentrate emissions in the atmosphere. This is observable also in equation (8), which informs that the higher the value of ρ the larger the initial consumption C(0). where   C  =

Figure 1. Optimal consumption policy for problem I: dependence on time preference

Optimal Policies at Different Time Preference Rates 300 225 150 75 0 0

25

50

75

100

125 Year

150

175

200

225

250

Consumption Function : 3 Percent Consumption Function : 1 Percent Consumption Function : dot 1 Percent

Figure 2 below displays atmospheric concentrations under different discount rates. Note that the cases for ρ=0.01 and ρ=0.03 resembles the shape of the optimal policy in Nordhaus's DICE model (see Figure 6 below): the characteristics of the emission policies come from the modeling strategy -consumption-smoothing over a planning horizon, just as Weitzman (2007) notes. Figure 2 also shows that optimal policies eventually stabilize atmospheric concentrations at the maximum allowed level. The time preference parameter only alter the date in which stabilization is reached. At 3%, stabilization is reached 160 years hence; at 1%, stabilization occurs in 5 centuries hence; while at 0.1% only 75% of atmospheric capacity is consumed 15 centuries hence.

Figure 2. Optimal atmospheric CO2 concentration: dependence on time preference

Accumulated consumption 20,000 15,000 10,000 5,000 0 0

40

80

120

160 200 240 Time (Year)

Accumulated consumption : 3 Percent Accumulated consumption : 1 Percent Accumulated consumption : dot 1 Percent

280

320

360

400

Figures 3 and 4 below show dependence of optimal consumption and concentration policies on the inequality aversion parameter. Three arbitrary values are considered: 0.5, 1, and 2. As can be seen in equation (8), the larger (smaller) the value of η the smaller (larger) the initial consumption at t=0 and the lower (greater) the rate (in absolute value) at which optimal consumption decreases. This implies that if society exhibits a small degree of inequality aversion (say, η=0.5), then earlier generations enjoy higher consumption of the non-renewable stock.

Figure 3. Optimal consumption policy for problem I: dependence on inequality aversion

Optimal Policies for problem I: dependence on inequality aversion 800 600 400 200 0 0

20

40

60

80

100 120 Time (Year)

140

160

180

200

Consumption Function : eta equals one Consumption Function : eta equals a half Consumption Function : eta equals two

Figure 4. Optimal atmospheric CO2 concentration: dependence on inequality aversion

Atmospheric concentration 11,000 8,250 5,500 2,750 0 0

40

80

120

160 200 240 Time (Year)

Accumulated consumption : eta equals one Accumulated consumption : eta equals a half Accumulated consumption : eta equals two

280

320

360

400

If society displays some degree of inequality aversion (say, η=2), then initial consumption is smaller and the rate at which optimal consumption declines is slower. This behavior implies atmospheric concentrations as displayed in Figure 4. The effect of this parameter is as if either mitigates or multiplies the effect of the time preference parameter. In equation (8), for instance, the value of η=0.5 is equivalent to multiplying ρ by a factor of two. If η equals two, then it halves the value of the pure time preference. Where does this discussion leave us? Figure 5 below shows the small exploration of the parametrical space just performed. It implies that any concentration policy lying between the upper and the lower concentration policy could be defined as optimal using the respective combination of parameters. The blue line depicts optimal concentration if a pure rate of time preference of 3% and a value of 0.5 for η are fed into the model. The red one represents optimal concentration if the pure rate of time preference equals 0.1% and if η=2. In the former policy, stabilization is reached after a century, while in the latter approximately 50% of atmosphere capacity is still free to use 15 centuries hence. This is in fact a wide range of optimal policies that could make sense under some combination of parameters. Note, additionally, that we only considered discount rates between 0.1% and 3%, and only constant values for η between 0.5 and 2. Expansion of the parametrical space (to represent a somewhat more impatient society, or to include distinct cases for η) will increase the area in which an optimal policy could possibly lie. Note that this parametrical uncertainty arises even if we assume that climate knowledge is good enough to make sure that S* represents atmosphere's capacity to store greenhouse emissions and that the central planner has perfect knowledge about that value and the way the stock gets affected by economic activity.

Figure 5. Optimal atmospheric CO2 concentration: dependence on parameters space

Atmospheric Concentration 11,000

8,250

5,500

2,750

0 0

150

rho=0.3; eta=0.5

300

450

600 750 900 Time (Year)

1050 1200 1350 1500

rho=0.001; eta=2

A comment on discounting At least for two decades, discounting in the context of climate change has received much of the attention of the economic debate. But the study of its rationale and of the way in which it affects the results in dynamic optimizing economies is in fact much older. For instance, Ramsey (1928) makes a

now famous critique to the practice of using positive rates of pure time preference;4 Koopmans (1974) makes a case in which discounting can be disastrous in the sense of advancing the end of the planning horizon, when assumed finite. Further, Gale (1967) notes that, notwithstanding the ethical critiques already available by then, an infinite horizon cake-eating model requires positive rates of time preference to actually make formal sense (otherwise, no optimal policy exists). 5 Despite this focus of the literature on the pros and cons of using positive utility discount rates, special attention has being payed to the so-called “social rate of discount”, or “consumption-discount rate”. This rate is different to the “private”, or “utility-discount” rate, ρ. Following Perman, et. al. (2003) and Heal (1998 and 2008), we define the social rate of discount as the negative of the rate at which the value of a small increment of consumption changes as its date is delayed.6 Recall the objective function in problem I: ∞

∫ e− t U  C  t  dt . 0

The value of a small increment in consumption is, simply, e−t U C  C  t   , where U’(C(t)) is the marginal utility from consumption. Then, the negative of the rate of change of this magnitude is − ∂  e−t U C  C  t     C  ∂C  t  ∂t (9) r  t = = =  C  g  t  , − t C t  ∂ t e U C  C t  Note that the term ρ, the “pure rate of time preference”, is the utility-discount rate used in the objective function of problem I. If g(t)>0 and η(C)>0, then the social rate of discount will be larger than the pure rate of time preference, meaning that an additional unit of consumption today is worth more that in the future, when consumption is much more abundant. Under this view, even after assuming an egalitarian treatment of instantaneous utility, the scenario of growing per capita consumption allows for a positive social discount rate. However, in scenarios of constant per capita consumption rates the social rate of discount is equal to the pure rate of time preference. Several approaches to equation (9) are present in the literature. Fundamentally, the main differences come from which variables are assumed exogenous and which endogenous. If a partial equilibrium mood is assumed, g(t) may be assumed exogenous, and then the values of the ethical parameters may be arbitrary or may be determined by calibrations using existing empirical studies. The resulting value of r(t) is therefore determined by the “observed” values in the right hand side of equation (9). Another approach consists in taking r(t) as some approximation of existing interest rates in the market. If knowledge concerning g(t) is available, then the ethical parameters must be chosen so equation (9) holds. One additional approach is assume a general equilibrium setting, just as the one presented before. In this view, r(t) is determined by the values assumed for the ethical parameters and by the optimal trajectory of consumption. In the case of the cake-eating economy considered in paragraphs above, it is easy to see that r(t) is zero no matter the magnitude of the values of η and ρ. This is so because there are no alternative uses of the consumption flow (i.e., no capital). If capital is included, then it is easy to see that consumption and investment flows are chosen so the marginal productivity of capital equals r(t) as defined in (9), which is a familiar 4

5

6

“One point should perhaps be emphasized more particularly; it is assumed that we do not discount later enjoyments in comparison with earlier ones, a practice that is ethically indefensible and arises merely from the weakness of the imagination...” (Ramsey, 1928, pp. 543) Which is a fact that is easy to see. If ρ was zero, equation (7) says that an optimal policy implies having a constant consumption of a finite stock in an infinite horizonm, clearly an infeasible possibility. In Gale's words, this is the “worst posible programme.” (Gale, 1967, pp. 4). Here, since it is preferable to work with positive numbers, we follow the conventional practice of defining the social discount rate as the negative of the rate of change of discounted marginal utility. For more on this, see Heal (1998 and 2008).

relation in the resulting Ramsey model known precisely as the Ramsey equation. Geoffrey Heal recently offered a useful discussion between the use of the pure rate of time preference and the social discount rate (Heal, 2008). From there we can conclude that it does not make sense to treat the consumption rate as exogenous (that is, to work directly with the social discount rate), since that is the way to go if we are running a partial equilibrium analysis -that is, an economic evaluation of a project that has by assumption no major repercussions in the economic system. If the “project” under analysis has the objective to produce systemic implications (e.g., a substantial change in relevant systemic trends), such as in the case of mitigation policies in the context of climate change, then the analysis requires to work in a general equilibrium setting, that is, it requires the consumption trajectory to be endogenously determined, for it is its main purpose to study the outcomes of, say, sensitivity analysis over the optimal consumption policy. This means to work directly with the pure rate of time preference and with the inequality aversion parameter, and then let the model determine the relevant social discount rate.

III. The economists and the Stern Report: a summary of the critics Many reactions within economists have occurred since the document known as “The Stern Review” on the economics of climate change was published in October of 2006. Here, in the sake of being short and concise, we will restrict ourselves to the specific set of critiques coming from certain leading economists. We divide those critiques to attend three main aspects (political, technical and methodological). The two technical issues relate to the Ramsey equation. The first one of these deals with the consistency of its parametrical treatment regarding two main ethical assumptions, while the other technical issue relates to the problem of using that equation in the context of large uncertainties in the very long run. As we will see, these two elements suggest non-trivial methodological issues involving not only the particular setting of the economic modeling used to answer climate changerelated questions, but the overall approach implied in the theoretical framework at economists’ hands when dealing with long run issues, like climate change. The political set of critiques consists on some “sociological” comments on the Review, so to speak, and it is the one with which we start. We found very interesting many of these comments from critics regarding how the Review was put together, and some of the methodological procedures that determined the direction of the Review’s main conclusions.

Political issues As can be seen in the front page of its Website, The Stern Review on the economics of climate change is “a review to the Prime Minister and the Chancellor of the Exchequer on the Economics of Climate Change” (Stern Review Web Site, 2007). This simple fact motivates Nordhaus to say that “the Review should be read primarily as a document that is political in nature and has advocacy as its purpose” (Nordhaus, 2007b, p. 688). Unlike academic papers and documents, Nordhaus says, “[t]he scientific ground rules of government reviews produced by professional scientists and economists are not codified.” (Nordhaus, 2007b, p. 688). This very point is illustrated by other of the Review’s critics, who complains because of the impossibilities to follow analytically the Review’s conclusions from its premises: “The Review uses an IAM [integrated assessment model] called PAGE, on which some numbers have been crunched and some conclusions have been based, but the exact connection between PAGE and Stern’s conclusions is elusive, frustrating, and ultimately unsatisfactory for a professional economist who honestly wants to understand where the strong policy recommendations are coming from.” (Weitzman, 2007, p. 705). This lack of “codification of the scientific grounds”, as Nordhaus called it,

led Richard Tol, in a very early critique of the Review, to hasty say that “[t]he Stern Review can therefore be dismissed as alarmist and incompetent.” (Tol, 2006, p. 35).7 Another point is made pointing to the collective author body and to the time period in which this group of people wrote the report: “A team of 23 people, led by Sir Nicholas Stern and supported by many consultants, worked for a little over a year to produce a report of some 700 pages on the economics of climate change.” (Tol, 2006, p. 32). Nordhaus refers also to this fact: “The Review was prepared in record speed. One of the unfortunate consequences of haste is that the Review is a thicket of vaguely connected analyses and reports on the many facets of the economics and science of global warming. Readers will find it difficult to understand or reproduce the line of reasoning that goes from background trends (such as population and technology) through emissions and impacts, to the finding about the 20 percent cut in consumption, now and forever.” (Nordhaus, 2007b, p. 688). Those are some of the consequences of the Report’s genesis as a political document, and can be well summarized by quoting Nordhaus again: “The central methodology by which science, including economics, operates is peer review and reproducibility. By contrast, the Review was published without an appraisal of methods and assumptions by independent outside experts. Nor can its results be easily reproduced. (…) This deviation from the norm of modern science does not necessarily discredit the Review, but it does mean that fatal flaws in evidence and reasoning, which might have been caught in the early stages under normal ground rules, may emerge after the report has been published” (Nordhaus, 2007b, p. 688). Notwithstanding the problems of the Review regarding its way of doing economic analysis and policy evaluation, economists find some methodological virtues when discussing some analytical issues, specially because the Review makes a case in which the major role of economics in multidisciplinary discussions and policy evaluations is proved, at least in principle. John Quiggin suggests, for instance, that “[t]he Stern Review radically changed the terms of the debate by presenting the issues in economic rather than in scientific terms”. (Quiggin, 2006). Both Dasgupta (2007) and Nordhaus (2007b) emphasize that the conclusions drawn from the Review’s analysis are not dependent on neither new climatic facts nor new economic theory, but dependent on specific assumptions within the standard framework of the dynamic cost-benefit analysis used in economic policy evaluation, assumptions that are expressed in particular choices for crucial parameters values and for a specific utility function. Tol et. al. also show the same pattern, despite Tol’s early dismissal of the Review for being incompetent: “we applaud the Stern Review author team for reconfirming that the climate problem can productively be approached as an economic problem whose solutions can be explored with the tools of decision analysis.” (Tol and Yohe., 2007). While the success of the Stern Review in pointing and spreading out some of the economic issues involved in the climate change problematic can go relatively uncontested, the success of the very set of economic tools being used in “the majority” –as Weiztman calls it– of climate change economic models is under scrutiny, not even by mainstream “outsiders”, skeptical as they may be of such a theoretical framework, but by some of the most visible “insiders”.

Technical issues Every insider of the economics of climate change knows that the main economic modeling strategy consists in using some variation of the Ramsey-Koopmans-Cass model for economic growth, which in principle means to analyze conditions for a so-called optimal consumption-smoothing policy. Nordhaus explains this in several places (1992, 1993, 2007a and 2007b). Basically, this strategy involves a 7

However, as we are about to see, this negative early critique from Richard Tol is tinged positively afterwards, when some of the Review’s implications to economic analysis are taken into account.

dynamic optimization problem that seeks the maximum of some social welfare function subject to dynamic economic restrictions facing “an externality” coming from global warming,8 which in turn is made “endogenous” and “controllable” with some so-called climatic “modules”, as Nordhaus baptized them, containing a set of equations reflecting climatic interactions with the economic system. The optimization process, being based in such an analytical framework, soon needs to deal with a particular object that gathers attention of commentators, the “Ramsey equation”. The Ramsey equation relates four things: the social discount rate, two subjective (or ethical) parameters (the pure rate of time preference and the index of inequality aversion) and one “economic fact” (the rate of growth of consumption, which can be endogenous or exogenous, depending on the particular modeling approach).9 Too much work is being done in the economics literature discussing several issues associated to either the discounting practice or/and the parameter capturing inequality aversion. We do not intend to summarize that discussion here. Two aspects are highlighted in the climate change context, though, and it is worth to pay attention to them. The first one means to look how economists seek for consistency in the assumptions with the “general equilibrium model” they can think of, to use Weitzman’s (2007) expression, consistency that is fruitfully discussed in a deterministic pure-capital model, as in Dasgupta (2007). The second one appears when one crosses the bridge from the deterministic-pure-capital model to either a stochastic-pure-capital model (like in Weitzman, 2007) or to a stochastic-imperfect-economy model (like in Dasgupta 2007). These two aspects yield important conclusions not only withheld to technical details in this “majority” modeling strategy, but conclusions that force researchers to ask questions from a more broader methodological level.10 There are reasons to think that economists participating in the climate change debate are beginning to think that the consumption-smoothing modeling strategy is not a good analytical tool to analyze climate change from an economic point of view, either because the need to treat some non-resolved theoretical puzzles and use them to give policy advice involving the spending of billions of dollars in mitigation-related actions, or because it might be a case for misplaced concreteness, a concept that is familiar enough to ecological economists. Now we turn to briefly look some of the components of those two aspects. 1. Parametrical choices and consistency of modeling. Table 1 below shows the choices for the parameters involved in the Ramsey equation in two of the commented previous economic models and in the Review, along with the reported reasons to make that choice. To be able to compare consequences in the social rate of discount, r, we follow Dasgupta (2007) in looking to the Review’s assumptions of a constant-exogenous rate of growth of consumption equal to 1.3% per year. The difference in the discount rate showed in the fifth column is one of the main critiques, if not the main, to the Review’s economic methodology: to illustrate, today’s value of $1 of damage costs happening in 150 years hence is equal to $0.0552 dollars under Cline’s, to $0.0018 under Nordhaus’, and $0.1242 under Stern’s. This fact, together with the Review’s also criticized treatment of both pessimistic climate scenarios and optimistic mitigation prospects, is the one that leads its main pro-present-mitigation conclusions.11 8

9

10

11

The greatest “externality” the world has ever seen, as Stern (2006) called it. This externality is also thought of by Dasgupta (2007) as being a “massive global commons” problem, or the greatest “market failure”, as Stern (2006) and Heal (2008) call it. If a general equilibrium model, then the rate of growth of consumption is endogenous. If, instead, is a partial equilibrium model, then it is exogenous. Heal (1998) provides a useful discussion on this distinction. Partgha Dasgupta himself has some interesting comments on how economists deal with economics philosophy and methodology while they do economics. (Dasgupta, 2005). If a consensus across critics has to be made, that would be it. Even otherwise extreme economists agree on this: “the key assumption that drives [the Review’s] strong conclusions is the mundane fact that a low interest rate is postulated…” (Weitzman, 2007); “Stern must expect criticism from mainstream economists for raising ethical concerns, specially with

Table 1 Author William Cline (1992) William Nordhaus (1994)

ρ 0 3%

η 1.5 1

g 1.3% 1.3%

r 1.95% 4.3%

Stern Review Source: Dasgupta (2007)

0.1%

1

1.3%

1.4%

Reasons Ethical Calibration with market economies Ethical

However, discussion of these Review’s assumptions led Dasgupta (2007) to analyze their consistency with a basic deterministic-pure-capital model, and he did not restrict himself to Stern’s, but included both Nordhaus’s and Cline’s as well. His conclusion is that the overall set of assumptions showed in Table 1 is either “bad economics”, “bad ethics”, or “bad philosophy”. In such a model, the ratio of aggregate saving to aggregate output is dependent on δ, on η, and on the rate of return on investment, r, as Dasgupta (2007) shows and Weitzman (2007) claims. With the parameters values chosen in the Review, 97% of each unit of output is saved in the optimal pure-capital path, situation considered by both authors as a reduction ad absurdum.12 Having preferred a market-based-behavior parametrical calibration instead of the former ethical parametrical prescription may lead the researcher to base herself empirically on other studies to get the set of values used to calculate the “right” social discount rate. Dasgupta suggests that this is the “democratic move” of Nordhaus’ studies, and while that move makes Nordhaus position “consistent” in Dasgupta’s opinion, it is not exempt of analytical problems: “[t]here are two unknowns that Nordhaus must determine (δ and η) but only one equation, ρ=r, that relates them. So he is forced to estimate one of the unknowns from other types of data. There is then a problem of consistency in the ways the parameters have been estimated in the different studies” and there is “a serious possibility that observed behavior offers a wrong basis for calibrating δ and η” in the context of the “massive global commons problem” of climate change (Dasgupta, 2007). 2. Uncertainties, risks and modeling Some of the Review’s critics emphasize the distinction between “known unknows” and “unkown unknowns”, as Nordhaus (2007) calls them. Uncertainty comes basically from the rate of growth of consumption, and there are, at least, two ways in which critics deal with it. The first one is restricted to the conventional dynamic optimization analysis, and involves some of Weitzman’s and Dasgupta’s critiques. The second one puts aside “the marginal analysis”, moves away from standard welfare analysis, based solely as it is on “traditional” consumption, and generalizes towards a comprehensive view of consumption, one that includes, say, existence value of landscapes and many other things largely unrecognized by standard economic welfare analysis. Regarding the first way to deal with uncertainty, recall the Ramsey equation and consider the case in

12

regard to that favourite economists’ red herring, discounting. Indeed, divergence from economic orthodoxy seems to be driving the results with the report claiming novelty in the treatment of future generations, distributional inequity, and uncertainty.” (Spash, 2007) The contradictory situation can be read off as follows: intergenerational and intragenerational equity considerations may lead the researcher to choose parameters close to Cline’s and Stern’s choices. Those choices, however, reveal to optimally prescribe in a deterministic general equilibrium setting (as Weitzman calls it), or in a deterministic pure-capital model (as Dasgupta treats it), a savings policy that takes too much from present generations to give distant-richer generations. This does not reflect much inequality aversion at all, to say it with Dasgupta (2007).

which the constant rate of growth in consumption is a random variable, which probability function is known.13 Then there are as many realizations for the social discount rate as there are possible values for the consumption rate of growth. Analytical treatment of this discount rate uncertainty, however, is not carried on best-guessing over discount rates themselves, but over discount factors, as it is well-known in the hyperbolic discount literature based on Weitzman (1998). Weitzman (1998) showed that the certainty-equivalent discount rate converges in the long run to the minimum possible discount rate. If that is the case, Weitzman (2007) notes, then the Stern assumptions might end up being just about right, despite it seems that no critic gives the Review some credit for its uncertainty treatment. But this is not the only story concerning uncertainty. Making a case in which the Review’s assumptions may be right does not solve further and deeper analytical problems. Carrying out a brief uncertainty analysis based on a macroeconomy represented “by a dynamic stochastic general equilibrium in the Lucas-Mehra-Prescott fruit free model”, Weitzman (2007) suggests that existing unsolved theoretical problems enter into the scene, especially the so-called “risk free rate puzzle” and the “equity premium puzzle”. We do not intend to go further in explaining them, since Weitzman himself does a good job in that. We only want to illustrate that this puzzles add more analytical problems to the economic approach, in which, again, the game between prescription and description is an important issue. For instance, if description is chosen, and if some other assumptions concerning correlation of climate damages to economic activity are met, then the implied social rate of discount is 1.7%, close enough to Stern’s criticized selection; when a prescription mood dominates, however, then the social rate is 6%, even higher to the Nordhaus figure in Table 1 (Weitzman 2007). Going even further, Weitzman (2007) suggests that the values of δ and η consistent with the empirical observation of both the calibrated risk free rate (referring to short-term treasury bills) equaling 1% and the equity risk premium over that rate being around 6% are as disparate as δ=151 and η=150. “One does not know whether to laugh or cry”, Weitzman laments, “at the prospect of what the Stern Review IAM might end up recommending as its preferred policy for climate change in its number-crunching simulations if [those values] were fed into PAGE” (Weitzman, 2007). All this considerations led Weitzman to conclude that “[t]rying to forecast costs and benefits of climate change scenarios a hundred years or so from now is more the art of inspired guesstimating by analogy than a science” (Weitzman, 2007). Dasgupta (2007) also still restricts himself to the standard economic setting, and works uncertainties out using an “uncertain production economy” model. There, he proves that in a stochastic world of large uncertainties –expressed in the form of a high variance in the (still known) probability distribution of consumption growth rates– there are some (relatively) extreme situations in which no optimal policy may exist to solve the optimization problem. The contradiction arises because the optimal consumption-smoothing policy becomes infeasible, since it implies a ratio of aggregate saving to aggregate output larger than 1. After asking “how large is large?”, Dasgupta shows that even a ratio between the distribution’s variance to its mean equal to 7 yields no optimal policy. “[T]he problem of optimal saving when formulated in terms of expected well-being over an infinite horizon,” he says, “is so inadequately posed as to defy an answer. Consumption discount rates cannot be defined and social cost-benefit analysis of projects becomes meaningless” (Dasgupta, 2007). The second way in which uncertainty is analyzed reveals the existence of structural uncertainty that thickens the tails of probability distributions of the rate of growth of consumption. To make his point, Weitzman moves in a rare direction for a mainstream economist, “setting aside” marginal analysis and 13

Below we incluye the case when it is unknown.

defining consumption “comprehensively”, which means to include “the consumption of natural environments, ecosystems, species, and the like” (Weitzman, 2007). “With an evolutionary stochastic process like global climate change”, he writes, “the world is not standing still long enough for us to accumulate the relevant information to accurately assess tail probabilities. The net result is thicker left tails for the distribution of g under dynamically evolving global climate change than we are accustomed to dealing with in our much more familiar dynamic stochastic general equilibrium macro models, which in practice are based upon the stationary thin-tailed stochastic processes that we use to model a rational expectations equilibrium whose structure is (supposedly) fully known and understood.” (Weitzman, 2007).14 This assessment on the uncertainties of very distant future events is not new.15 Clive Spash made precisely that point in his comments to the Review: “Stern do note the Keynsessian differentiation between weak and strong uncertainty. One approach to complexity and strong uncertainty is scenario analysis. Thus, the forty scenarios which informed the IPCC work under the third assessment report were explicitly stated to be ‘equally valid with no assigned probabilities of occurrence’ (…). Stern then employ expected utility modelling which is known to be an inadequate representations of human behaviour e.g., assuming away loss aversion (Perrings, 2003). Subjective probability density functions then give precise computer generated outcomes. (…) Comparative statistics, shifting from one equilibrium to another, conceal complex processes of change. Thus Stern manage to convert unknown and unkownable futures into events with known probabilities, and miraculously strong uncertainty becomes weak uncertainty” (Spash, 2007). While paragraphing the language of the Review, Weitzman says that “the economics of climate change is the greatest application of subjective uncertainty theory the world has even seen” (Weitzman, 2007). Weitzman conclusions have the merit to be quoted: “The take-away message here is that the burden of proof in the economics of climate change is presumptively upon whomever wants to model optimal expected-utility growth under endogenous greenhouse warming without having structural uncertainty tending to matter much more than risk. Such a middle-of-the-distribution modeler needs to explain why the inescapably thickened tails of the posterior predictive distribution, for which the thick left tail of g represents rare disasters under uncertain structure, is not the primary focus of attention and does not play the decisive role in the analysis.” (Weitzman, 2007). To paraphrase Weitzman’s language, we can say that, indeed, this changes the rules of the research game. Dasgupta’s final paragraph is self explanatory as well, and also deserves to be quoted extensively: “[e]conomics helps us to realise what we are able to say about matters that will reveal themselves only in the distant future. Simultaneously, it helps us to realise the limits of what we are able to say. And that too is worth knowing, for limits on what we are able to say are not a reason for inaction. Climate change and biodiversity losses are two phenomena that are probably not amenable to formal, quantitative economic analysis. We economists should have not pressed for what I believe is misplaced concreteness. Certainly, we should not do so now.” (Dasgupta, 2007). We shall come back to the fact that the concept of misplaced concreteness, a not unfamiliar one to ecological economists, arises in the discussion.

14 15

This point is made in literature previous to the Stern Review. Since we restrict to some comments on the Review, we can only refer readers to the next set of four papers for an interesting discussion over uncertainty and cost-benefit analysis in the context of climate change: Tol (2003), Yohe (2003), Azar and Lindgren (2003) and Howarth (2003).

Methodological issues: it might be model! The last set of issues we want to discuss includes the relationship between the nature of the policy prescriptions coming from climate change economics and its very analytical framework. The underlying idea is that policy implications coming from the economics of climate change can be understood more as a consequence of the specific family of (dynamic optimization) models used than being a result of the actual theoretical content of those models. In other words, economists are finding answers (for instance, spend 1% of World's GDP in emissions mitigation, as in the Stern Report) to questions (for instance, what is the optimal economic policy regarding atmospheric carbon concentration?) whose main characteristics are determined by the analytical framework used in their scientific inquiry. Figure 6 below illustrates some runs of the DICE model for different policy strategies, as presented in Nordhaus (2007a). The strategy labeled as “optimal”, for instance, implies that carbon concentration starting at 2008 follows closely for more than a decade the baseline scenario, while substantial differences between these two profiles are observable only by year, say, 2045. Two centuries hence, the optimal policy stabilizes carbon concentration at 700+ ppm, while the baseline scenario jumps out of the chart at a figure higher than 1100 ppm.

Figure 6: Atmospheric CO2 concentration for different strategies

Source: Nordhaus (2007a). Relating this result with the modeling strategy followed, Martin Weitzman writes: “The upward sloping 'climate policy ramp' of ever tighter emissions reductions in the majority of other models (but not beginning just yet, please) is a familiar consumption-smoothing consequence of discounting (...) An efficient trajectory has a cost minimizing substructure similar to a Hotelling extraction problem: consumption flows are smoothed over time by maximizing present discounted utility subject to a stock

constraint on accumulated CO2e, which result in an 'as if' CO2e shadow tax grows over time at (approximately) the rate of interest.” (Weitzman, 2007, pp. 705). The Stern Report yields both a “flatter policy ramp” and higher emissions reductions “in its Hotelling-analogous consumption smoothing time profile by imposing discounting at a bare-minimum rate of interest” (Weitzman, 2007, pp. 705), as we noted in the above paragraphs. This seems to suggest that if we are using standard settings of Hotellingtype models for consumption smoothing (which can be understood as a general category including the Ramsey-Koopmans-Cass family of models for economic growth, which is the base for the DICE model, as Nordhaus (2007b, pp 690) makes clear), then we will obtain optimal policies that can be seen as a ramp, like in the case of climate change. Further, many of the problems that feed theoretical controversies (i.e., the “right” discount rate, the “right” inequality aversion parameter) will be present again. That is, different optimal consumption “policy ramps” coming from “explorations of the parametrical space”, which constitute the main suspects for creating unease and debate among economists in this matter, are implications of the specific modeling strategy, and not conclusions obtained from specific knowledge about interactions between the economic and climatic systems. Those conclusions can be easily obtained from a simple Hotelling-type model, like the one presented in a previous section. The model is not by any means new, and it is commonly known as the “cake-eating model”.16

IV. Conclusions: misplaced concreteness and research attitudes In 1989 Herman Daly and John Cobb published a book that constitutes now part of the classic literature in ecological economics. For the common good, is the title. They dedicate a substantial part of the book to alert ecological economists about some limitations and dangers of the modern organization of knowledge in academic disciplines, being the fallacy of misplaced concreteness, as was called by Alfred North Whitehead, the one that receives the attention. Economics as an academic discipline is a successful one, Daly and Cobb say, perhaps the most successful among social disciplines. Success comes along with a high level of abstraction, though, “yet the whole ethos of the university in general, and of the department of economics in particular, discourages the full realization of the extent of the abstracting that has gone on.” (Daly and Cobb, 1989, pp. 35). But the awareness do not take the form of an attack to abstractions, but of a call to realize their limitations. After all, as Whitehead wrote and the authors report, “[t]he methodology of reasoning requires the limitations involved in the abstract.” (Whitehead, 1925, p 200; quoted in Daly and Cobb, 1989, p 36). The fallacy of misplaces concreteness is “involved whenever thinkers forget the degree of abstraction involved in thought and draw unwarranted conclusions about concrete actuality.” (Daly and Cobb, 1989, p.36). In other words, it can be said that this type of “fallacy” might appear when researchers cross the bridge between abstract models and concrete economic reality. Daly and Cobb provide several examples for what they believe are cases of misplaced concreteness in economics -in fact, their book is organized in part having those examples as a structure. Four main topics receive attention: the market, the measures of economic success, the notion of rationality embedded in the so-called homoeconomics, and the economic concept of land. However, the word “fallacy” is not used in a precise way, since misplaced concreteness does not refer to some sort of erratic logic, as the authors promptly recognize: “we simply cannot think without abstraction. 'To abstract' means literally 'to draw away from' (...) It seems we must always commit this fallacy to some degree, and we must think of 16

A very similar setting was in fact presented in Hotelling's original paper (Hotelling, 1931), and analytically explored in Gale (1967) and in Dasgupta and Heal (1974). Further, Farzin (2002) provides a discussion on sustainable development. Gowdy and Julia (2007)recently presented an application of this model in the context of climate change. The nonrenewable resource in their paper is, however, the overall stock of carbon available to exploitation.

minimizing it rather than eliminating it entirely. For this reason it is a very subtle fallacy -more a general limitation of conceptual thought than an error in logic.” (Daly and Cobb, 1989, p. 41). Two ways to avoid the fallacy are discussed in Daly and Cobb's book. One is tracked back to Whitehead himself: “recurrence to the concrete in search of inspiration”. The other one requires to avoid excessive professional specialization. Despite there are some interesting issues concerning these two proposed ways of avoiding the fallacy (e.g., the technical repercussions of having not very specialized professionals) we do not discuss them further. For our purposes it is enough to remember the notion of the fallacy, and the potential dangers it represents to academic inquiry in economics. In the context of the economics of climate change, as represented in the basic model presented in a previous section of this paper, we can say that there is not enough focus on the things that often go unobserved in academic discussions. One in particular: What is the main question we are trying to answer with the economic models used so far in the context of climate change? There seems to be particular attention to a specific number: the percentage of World's GDP that need to be invested in climate change mitigation. Is one number all what we might be looking for? Assume for a moment that we have a reasonable method to come up with a number that causes consensus between experts. Even the knowledge of the “right” number presumably is not going to give us all the information policymakers need to generate sound policy related with climate change. How does the investment will be distributed among regions? Does the world need a multilateral organization to organize it? Or can we assume that all nations will agree with the optimal plan and act correspondingly? What about all the strategic behavior that might appear? In other words, what we have been calling “the economics of climate change” is by far not all the economics we might want to be doing when facing climate change. The economics of climate change is not resolved with the finding of an aggregate number (which, it is clear now, has been problematic enough). There are so many things we do not know, and so many things we know we can not know, to believe that an integrated assessment model will provide all the information we require to design climatic policy. To complicate further, any policy associated with climate change needs to be associated with the goals of development and poverty alleviation in a by no means negligible number of countries. We do have to pay attention to integrated assessments models. We need that information. But we do not have to conclude that such a model finds the solution to the problems associated with climate change. A small look to some research attitudes will help to show the point this discussion note tries to make. We take four economists, which differ not only in the time in which they worked, but also in the field and methodological perspective they choose to develop their research. Ariel Rubinstein, associated with both the School of Economics at Tel Aviv University and the Department of Economics at New York University, is a main author in the field of game theory. He published in 2006 an introspective paper that tries to ask the question “What are we trying to accomplish as economic theorists?” (Rubinstein, 2006). Rubinstein talks about the perils of using economic models: “As economic theorists, we organize our thoughts using what we call models. The word 'model' sounds more scientific than 'fable' or 'fairy tale' although I do not see much difference between them (...) A good model in economic theory, like a good fable, identifies a number of themes and elucidates them. We perform thought exercises that are only loosely connected to reality and that have been stripped of most of their real life characteristics. However, in a good model, as in a good fable, something significant remains.” (Rubinstein, 2006, p. 881). A very similar account is offered by Robert J. Lucas, who won the Nobel prize in Economics in 1995 and works in the Department of Economics at the University of Chicago. He has an informal manuscript entitled “What economists do”, in which he is trying to answer pretty much the same question. “I'm not sure whether you will take this as a confession or as a boast,” he writes, “but we are

basically story-tellers, creators of make-believe economic systems.” Economists gain understanding about economic phenomena using their “toy models” that represent make-believe worlds, like in his example of monetary-managed depressions in the imaginary Kennywood amusement park. To apply the knowledge to reality, “we must be willing to argue by analogy from what we know about one situation to what we would like to know about another, quite different situation. And, as we all know, the analogy that one person finds persuasive, his neighbor may well find ridiculous. (...) In any case, that is what economists do. We are storytellers, operating much of the time in worlds of make believe. We do not find that the realm of imagination and ideas as an alternative to, or a retreat from, practical reality. On the contrary, it is the only way we have found to think seriously about reality.” (Lucas, 1988). Analogy, make-believe scenarios and toy-models are all inside the toolbox of economists, as we can see. Some people may think that that way has been successful, some people may not. In any case, there seems to be a strong feeling that in the context of climate change success is not guaranteed. A very different research conviction is shown in two other Nobel prizes in economics, Richard Stone and Wassily Leontief, both known by their effort to provide economists a good input of information from economic reality to perform empirical analysis. Stone received the Nobel prize in 1984 by his work in developing the system of national accounts. Leontief received it in 1973 for the development of the input-output method and its application in economic problems. We first cite some general ideas Stone advanced when presenting the so-called “Programme for Growth” for Great Britain in the1960’s: “We should approach the economic system as an engineer approaches a complicated piece of machinery or as a doctor approaches his patient. Any adjustment or treatment depends on a sound diagnosis. (…) So let us follow the normal order of action: analysis, diagnosis, prescription, treatment. We shall continue to get nowhere if we continue to shortcircuit the first two stages.” (Stone, 1965; quoted in Johansen, 1985; p. 24) This very conviction favoring a somewhat exhaustive description of current economic situations are very related to some remarks that Wassily Leontief addressed to the US Mathematical Society in1954. Again, we cite extensively: “It is as if we were asked to reproduce the blueprint of a complicated motor on the basis of our knowledge of the general principles of operation of internal combustion engines and no other specific information but that conveyed by the few dials located on the dashboard and possibly the noise coming from under the closed hood. And as if that were not difficult enough, the structural characteristics of the engine the economist is studying are known to change under the impact of its continual operation. (…) It certainly becomes much easier if we are allowed to look under the hood. (…) But look under the hood [the economist] can, although in economics as in the garage it is an inconvenient and often a dirty operation. Admittedly, had we been able to reproduce the blueprint of the engine by indirect inference from the behavior of the gauges, such intellectual accomplishment would earn a much higher rating. Nevertheless, some economists rolled up their sleeves and looked under the hood”. (Leontief, 1954). The problem of climate change is perhaps the major social problem scientists are facing in present time. However, scientist do not need to forget about other major social problems, such as development, income inequality and poverty. With these four perspectives coming from equal number of leading economists we have two major approaches to economic research, and each one would have different implications when coming to the associated degree of misplaced concreteness involved. Geoffrey Heal, in his meta-review of the economics of climate change, concluded that “it is very clear that most of the models analyzed to date are so aggregated as to miss many important issues.” (Heal, 2008, p. 22). Academic community has the challenge to go more specific in climate change issues, and economists

may have to equally explore the two perspectives above to provide policy makers with sound knowledge at a nation-specific level. For ecological economists, we advance the notion that perhaps it is time to roll the sleeves and to look under the hood of our aggregated economic models. References Azar, C. and T. Sterner (1996), “Discounting and distributional considerations in the context of global warming,” Ecological Economics 19, p. 169-184. Azar, C. and K. Lindgren (2003), “Catastrophic events and stochastic cost-benefit analysis of climate change,” Climatic Change 56, pp. 245-255. Cline, W. (1992), The Economics of Global Warming, Institute for International Economics, Washington, D.C. Daly, H. and John Cobb (1989), For the Common Good. Redirecting the Economy toward Community, the Environment, and a Sustainable Future, Beacon Press, Boston. Dasgupta, P. and G. Heal (1974), “The Optimal Depletion of Exhaustible Resources,” The Review of Economic Studies 41. Symposium on the Economics of Exhaustible Resources, pp. 3-28. Dasgupta, P. (2005), “What do economists analyze and why: values of facts?” Economics and Philosophy 21, pp. 221-278. Dasgupta, P. (2007), “Discounting Climate Change,” mimeo. http://www.econ.cam.ac.uk/faculty/dasgupta/08/VISCUSI3_july.pdf Dasgupta, P. (2007), “Comments on the Stern Review's Economics of Climate Change”, mimeo. http://www.econ.cam.ac.uk/faculty/dasgupta/stern07.pdf Farzin, Y. (2002), “Can an Exhaustible Resource Economy be Sustainable?” Nota di Lavoro 47.2002, Fundazione Enio Enrico Mattei, Italy. Gale, D. (1967), “On Optimal Development in a Multi-Sector Economy,” The Review of Economic Studies 34 (1), pp. 1-18. Gowdy, J. and R. Juliá (2007), “The economics of the Mega-greenhouse effect: a conceptual framework,” Rensselaer Working Papers in Economics 0711, Troy. http://www.economics.rpi.edu/workingpapers/rpi0711.pdf Howarth, R. (2003), “Catastrophic Outcomes in the Economics of Climate Change,” Climatic Change 56, pp. 257-263. Heal, G. (1998), Valuing the Future: Economic Theory and Sustainability, Columbia University Press, New York. Heal, G. (2008), “Climate Economics: A Meta-Review and Some Suggestions,” NBER Working Paper 13927, NBER, Cambridge. Hotelling, H. (1931), “The Economics of Exhaustible Resources,” The Journal of Political Economy 39 (2), pp. 137-175. Johansen, L. (1985), “Richard Stone's Contributions to Economics,” Scandinavian Journal of Economics 87 (1), pp. 4-32.

Koopmans, T. (1974), “Proof for a Case where Discounting advances the Doomsday,” The Review of Economic Studies 41, Symposium on the Economics of Exhaustible Resources, pp. 117-120. Leontief, Wassily (1954), “Mathematics in Economics,” Bulletin of the American Mathematical Society 60 (3). 215-233. Lucas, R.E. (1988), “What Economists Do,” mimeo. http://www.economics.ox.ac.uk/members/catia.batista/lucas.htm Maier-Reimer, E. and K. Hasselman (1987), “Transport and Storage of CO2 in the Ocean -An Inorganic OceanCirculation Carbon Cycle Model,” Climate Dynamics 2, p. 63-90. Nordhaus, W. (1992), “An Optimal Transition Path for Controlling Greenhouse Gases,” Science, New Series 258 (5086), pp. 1315-1319. Nordhaus, W. (1993), “Optimal Greenhouse-Gas Reductions and Tax Policy in the “DICE” Model,” The American Economic Review 83 (2), pp. 313-317. Nordhaus, W. (2007a), “The Challenge of Global Warming: Economic Models and Environmental Policy,” Yale University, mimeo. (April 4 version). Nordhaus, W (2007b), “A Review of The Stern Review on the Economics of Climate Change,” Journal of Economic Literature 45, pp. 686-702. Perman, R., Y. Ma, J. McGilvray and M. Common (2003), Natural Resources and Environmental Economics, Pearson Higher Education, United States. Perrings, C. (2003), “The economics of abrupt climate change,” Phylosophical Transactions of the Royal Society of London, Series A: Mathematical and Physical Sciences 361 (1810), pp. 2043-2057. Ramsey, F. (1928), “A Mathematical Theory of Saving,” The Economic Journal 38 (152), pp. 543-559. Rubinstein, A. (2005), “Dilemmas of an economic theorist,” Econometrica 74 (4), pp. 865-883. Stern, N. (2006), The Economics of Climate Change: The Stern Review, Cambridge University Press, Cambridge and New York. Stern Review Website (2007), http://www.hmtreasury.gov.uk/independent_reviews/stern_review_economics_climate_change/sternreview_index.cfm Spash, C. (2007), “The economics of climate change impacts à la Stern: Novel and nuanced or rhetorically restricted?” Ecological Economics 63, pp. 706-713. Tol, R. (2003), “Is the uncertainty about climate change too large for expected cost-benefit analysis?” Climatic Change 56, pp. 265-289. Tol, R. (2006), “The Stern Review on the Economics of Climate Change: A Comment,” The Stern Report. Some Early Criticisms, The Center for Science and Public Policy, Washington. Tol, R. and G. Yohe (2007), “The Stern Review: A Deconstruction,” mimeo. http://www.fnu.zmaw.de/fileadmin/fnu-files/publication/working-papers/sterngecwp.pdf

Weitzman, M. (1976), “On the Welfare Significance of National Product in a Dynamic Economy,” The Quaterly Journal of Economics 90 (1), pp. 156-162. Weitzman, M. (1998), “Why the Far-Distant Future Should Be Discounted at Its lowest Possible Rate,” Journal of Environmental Economics and Management 36, pp. 201-208. Weitzman, M. (2007), “A Review of The Stern Review on the Economics of Climate Change,” Journal of Economic Literature 45, pp. 703-724. Quiggin, J. (2006), “Stern and the critics on discounting,” mimeo. Yohe, G. (2003), “More trouble for cost-benefit analsys,” Climatic Change 56, 235-244.