| 摘要 | My friend and former colleague Irwin M. Stelzer has written a short but interesting essay on climate policy, arguing that by asking the right question we will be oriented toward a useful set of policy prescriptions given the substantial uncertainties about the underlying facts and projections of anthropogenic warming. His arguments, which reflect closely those he articulated as a panelist at an AEI event in November 2018, are thoughtful and represent a vast improvement over the overconfident assertions, hysteria — yes, that is the proper term — and political machinations that are typical in this debate.
But his position remains unpersuasive, for the reasons discussed below. For now it is important to lay out his argument, which can be summarized as follows.
The debate over climate policy is afflicted with uncertainty analogous to that confronting a homeowner considering whether to purchase fire insurance. Because the homeowner cannot know whether a fire will consume his house in the next year, purchasing an insurance policy is prudent; and a similar rationale counsels adopting a climate policy that would provide some insurance against the uncertain possibility of nontrivial damage from anthropogenic warming.
A regulatory approach — “Californiacat[ing] the entire country” — would yield mounting inefficiencies, imposing costs “that often are commensurate with the benefits only in the eyes of the burgeoning [regulatory] agencies.”
A carbon tax would “correct a massive flaw in the way that the price system is operating” and would “do something about the outbreak of fiscal imprudence that is careening us toward an unpleasant reckoning with unsustainable trillion-dollar deficits.”
It is true no more that a carbon tax is a political loser, and it would drive “consumers to make choices based on properly computed prices . . . reflect[ing] all of the costs of producing and distributing those products.”
“Because quantification [of climate phenomena and/or the proper tax] is difficult . . . a modest tax that will reduce emissions and deficits, while also reducing the mountain of debt” to be left to future generations would be wise.
My observations on those arguments are as follows.
Climate Policy as Insurance
Stelzer fails to ask how much actual “insurance” his modest carbon tax would provide. The homeowner can tailor a policy in the context of the value of his house and the premiums needed to protect that value, and Stelzer implicitly is correct that it is usually the case that the premiums are sufficiently affordable and the insurance benefits sufficiently high to make an insurance policy worthwhile. Such is not the case with climate policy, whether in the form of a tax, regulations, or, a fortiori, something as radical as the Green New Deal. The most prominent recent proposals for a carbon tax are those promoted by the Alliance for Market Solutions and the Climate Leadership Council; if we apply the climate model used by the Environmental Protection Agency, the temperature reduction in 2100 yielded by each proposal would be 0.015°C, under the highest assumption made by the Intergovernmental Panel on Climate Change (IPCC) about the effect of reductions in greenhouse gas (GHG) emissions. (The standard deviation of the surface temperature record is about 0.11°C.)
That same modeling result obtains for the domestic regulatory approach promoted by the Obama administration; and at an international level, the entire Paris agreement, if we assume the emissions cuts (the nationally determined contributions) to be serious (they are not), would have reduced temperatures in 2100 by 0.17°C. Add an additional 0.01°C if we assume that the contemporaneous pseudo-agreement between the Obama administration and China was meaningful. (It was not.) Even the electricity components of the Green New Deal, perhaps surprisingly, would have a temperature effect of less than 0.2°C.
To take this one step further, assume a truly serious international effect to reduce GHG emissions: a 20 percent reduction by China, a 30 percent reduction by the rest of the industrialized world, and 20 percent by the rest of the developing world, all in addition to the Obama reductions and all by 2030. The temperature effect by 2100: a bit more than 0.5°C. In short, Stelzer’s insurance analogy is not very useful in the context of climate policy because the value of the insurance yielded by any plausible policy would be effectively equal to zero.
“Californicat[ing] the Entire Country” Through Regulation
Stelzer certainly is correct that regulators can be driven by perverse incentives, among them the imperative of ever-larger budgets, and it is not difficult to envision a legislature/bureaucracy negotiating outcome in which budgets rise with the increasing stringency of the regulatory regime. At the same time, much (not all) of the California climate regulatory structure is facilitated by federal policies that allow California to shift some substantial part of the costs of its regulations onto taxpayers and consumers in other states. The federal subsidies for wind and solar power are a good example; another is the federal light-vehicle fleet mileage rules, which engender pricing cross-subsidies from purchasers of trucks and SUVs to the producers or purchasers of electric vehicles. An attempt to shift the costs of federal regulatory policies onto overseas economies would be much more difficult, although it is true that federal policies raising the prices of conventional energy would create a wealth transfer, as a crude generalization, from red states to blue.
I am sure that Stelzer counts himself among the many economists — perhaps most economists — who argue that a tax on GHG emissions would reduce the overall cost of achieving GHG emissions cuts below the cost that would obtain under command-and-control regulations, because emitters would be able to choose the least costly means of achieving emissions reductions in the face of the tax. That outcome is unlikely to be the case with a regulatory approach because regulators cannot know the specifics of every industrial process and the like and have only weak incentives to learn or to apply them.
The central problem with that argument is straightforward: The emissions goal is not fixed or exogenous. Instead, it must be chosen. “Efficiency” requires both an efficient emissions goal that equates the marginal benefits and costs of emissions reductions and tools to achieve that amount of reductions that minimize the cost of doing so.
Once government derives revenues from a system of carbon taxes, with ensuing political competition for those revenues, it is not difficult to predict that under a broad range of conditions the chosen emissions reduction goal will be more stringent than that emerging from a regulatory process. This is particularly the case if policymakers have time horizons shortened by the imperatives of the next election; the implicit use of a discount rate too high will yield a preference for increased revenues in the short run at the expense of lower revenues over the longer run.
Because the marginal members of the congressional majority are likely to be the incumbents in greatest danger of defeat in the next election, it is not difficult to predict that the political equilibrium for a carbon tax will be a rate maximizing revenues over a period shorter rather than longer, precisely because for those marginal members of the majority, the time horizon is the next election and the political benefits of greater spending are more or less immediate. This is particularly the case given that the burdens of the carbon tax would be hidden in the prices of myriad goods and services.
It might be the case that regulators too have incentives to choose emissions goals that are too stringent, if doing so is consistent with the larger goal of maximizing their budgets (or discretionary budgets), and because overly stringent regulations may serve an ideological agenda. But in the case in which Congress must approve or has the power to repeal given regulations, there are strong reasons to believe that a tax approach would prove less efficient overall than the regulatory approach.
A Carbon Tax as a Pricing Corrective
Stelzer argues that a carbon tax would “correct a massive flaw in the way that the price system is operating” and would drive “consumers to make choices based on properly computed prices . . . reflect[ing] all of the costs of producing and distributing those products.” Stelzer here gives the politicians far too much credit: Do they know the marginal social cost of GHG emissions? And precisely how does Stelzer know that the difference between the market price of energy and its social cost is “massive”? Stelzer’s description of that difference as “massive” is not consistent with his earlier argument that uncertainty is the central characteristic of our understanding of the climate effects of increasing GHG concentrations and that it is that uncertainty that leads us to opt for policies yielding some insurance.
Notice that in the integrated assessment models commonly used for climate policy analysis, the social cost of GHG emissions changes by one or two orders of magnitude merely by changing the assumed discount rate. At a discount rate of 7 percent — the rate specified in the Office of Management and Budget regulation on benefit/cost analytic methodologies — the social cost of GHG emissions in 2050 falls to $14.27 per ton in the DICE model and to $0.73 per ton (!) in the FUND model. In the two prominent carbon tax proposals noted above the proposed taxes are around $40 and $55, respectively. The Obama administration calculated the social cost of GHG emissions by including only the domestic cost of regulations, combined with the global benefits, which is a blatant exercise in stacking the deck and inconsistent with elementary principles of benefit/cost analysis. The Trump administration has recalculated those numbers using only domestic benefits; the social cost of GHG emissions in 2050 is $2.20 at a 7 percent discount rate and only $10.10 at a 3 percent discount rate.
The Carbon Tax as a Revenue/Deficit Reduction Tool
Stelzer certainly is correct that the current federal deficit/debt path is not sustainable. A basic condition for the stability of government borrowing is that the real rate of return to government debt must exceed the growth rate of real government debt in the long run. If the market expects the debt to grow over the long term more quickly than the real return to the debt, then interest rates sufficient to induce the market to offer real resources to the government must increase, perhaps massively. Moreover, the credibility of the government promise to repay its debts in real terms cannot be assumed to be independent of the magnitude of the debt itself, in that incentives to renege on promised debt service must rise as the debt rises. And potential lenders to the government would assume that dollar-denominated debts promising extremely high interest returns (say, 50 percent) in fact would not be repaid in real terms, as the incentive to renege would become overwhelming.
But do we really believe that Congress would not spend the revenues anyway? One would have to make a heroic assumption about the time horizons of politicians to assume that they would be willing to take the political heat for imposing a carbon tax while rejecting the political benefits of the spending that it might finance. Can anyone believe that fierce competition for the revenues would fail to emerge in Congress and that innumerable interest groups would remain modest in their demands? The losers from the carbon tax would insist on compensation in some form, and it is difficult to believe that their demands would be ignored, particularly in the complex bargaining process yielding budget/political outcomes in the Senate.
The assumption that the revenues from a carbon tax would be used to reduce the debt burden inflicted on future generations is not to be taken seriously: After all, why has the equilibrium outcome characterizing recent budget decisions shifted massive amounts of debt onto future generations in the first place? Stelzer certainly is correct that massive deficits and the growing debt represent enormous problems, but he assumes far too much about the virtues of any new tax as a (partial) solution. I think we would be wise to keep in mind the observation made by the late Herbert Stein: When the problem becomes truly unsustainable, it will be addressed one way or another. Stelzer’s implicit argument that it would be far better to deal with it now may be true, but it is irrelevant.
Stelzer’s argument that a carbon tax no longer is a political loser is far from obviously correct. Two initiatives proposing carbon taxes in Washington state — hardly a Republican stronghold — in 2016 and 2018 were defeated decisively, by 59–41 and 57–43, respectively. The Waxman-Markey proposal for a cap-and-trade system (similar in many ways to a carbon tax) passed the House of Representatives in 2009 by 219–212; the party breakdown was 257 Democrats to 178 Republicans. Rep. Bob Inglis (R-SC) and Rep. Carlos Curbelo (R-FL) proposed carbon taxes in 2009 and 2018, respectively, and were defeated decisively in the following elections. It is clear that congressional Republicans would vote almost unanimously against a carbon tax and that significant numbers of congressional Democrats would do so also.
Some Additional Observations
Stelzer’s uncertainty principle can be restated as follows: There is some nontrivial likelihood that anthropogenic climate change will prove serious or even catastrophic. That is the so-called “fat tail” argument: The right-hand tail of the statistical distribution of future climate damage might be “fat,” so the likelihood of such serious damage is not trivial. The problem is that there are at least three potential fat tails. A second is the potential benefits from anthropogenic warming. Merely examine the NASA “greening” analysis of the earth: The peer-reviewed literature estimates that 70 percent of that effect is from carbon dioxide fertilization. A well-known Lancet study reports that far fewer people die from heat than from cold. Perhaps more speculatively, the likelihood of a future glaciation, however distant in time, approaches certainty, and anthropogenic warming under such conditions might prove a significant benefit. And the third fat tail is the unanticipated adverse effects of government policies: Can anyone argue plausibly that such potential effects are unimportant?
Once conservatives (or Republicans) endorse a carbon tax or any policy to address the purported crisis, they will leave themselves with no principled approach for opposition to even the most extreme proposals. After all, if anthropogenic climate change is an “existential threat,” no cost is too high and no benefit is too small. Stelzer implicitly is assuming that the political left will be satisfied if a (modest) carbon tax is implemented; he gives them far too little credit for the extremism of their ideological opposition to fossil fuels narrowly and with respect to the fundamental antihuman core of their environmental goals. Instead, a modest carbon tax will be only the beginning, and, again, once conservatives endorse a carbon tax they will have endorsed all the environmental left’s assumptions.
Consider the recent IPCC report on limiting temperature increases to 1.5°C by 2100. Put aside the utter silliness and dishonesty of that report; buried in its supporting documentation is a discussion (page 2–79) of the taxes (in 2010 dollars) on GHG emissions that would be needed to limit warming in this century to 1.5°C.
The midpoint of the range for 2030 is $3,156 per ton in 2017 dollars. That works out to a tax per gallon of gasoline of over $29. (Combustion of a gallon of gasoline/ethanol 10 percent blend emits about 18.9 pounds of CO2.) Put aside the higher figures among the ranges and the upward shifts as the end of the century approaches: Can anyone believe that such taxes on conventional energy are feasible politically anywhere in the world? Stelzer obviously is not proposing anything even remotely approximating such a preposterous tax; but IPCC now is doing so, and the American environmental left has endorsed such policies, as the “zero-emission” transportation system advocated in the Green New Deal is a call for extreme implicit taxes on conventional transportation fuels. Again, endorsing a modest tax now leaves no grounds for opposition later.
Stelzer’s uncertainty argument is a vast improvement over the loud assertions now characterizing the climate debate, but it remains a bit too facile. Yes, there is much that we do not know. But we do know some things, foremost among them the actual data on climate phenomena, which are rather inconsistent with the “crisis” view. It is absolutely correct that some predicted effects of increasing GHG concentrations are observable in the data. But there is little evidence — thus far — of serious climate impacts attendant on increasing GHG concentrations. Temperatures are rising, but as the Little Ice Age ended around 1850, it is not easy to separate natural from anthropogenic effects on temperatures. The latest research in the peer–reviewed literature suggests that mankind is responsible for about half a degree of the global temperature increase of about 1.5°C since 1850.
There is little trend in the number of “hot” days for 1895–2017; 11 of the 12 years with the highest number of such days occurred before 1960. Global mean sea level has been increasing for thousands of years; it may or may not be accelerating. The Northern and Southern Hemisphere sea ice changes tell different stories. US tornado activity shows either no trend or a downward trend since 1954. Tropical storms, hurricanes, and accumulated cyclone energy show little trend since satellite measurements began in the early 1970s. The number of US wildfires shows no trend since 1985. (Wildfire acreage is far more fundamentally driven by federal forest-management practices.) The Palmer Drought Severity index shows no trend since 1895. US flooding over the past century is uncorrelated with increasing GHG concentrations. The available data do not support the ubiquitous assertions about the dire impacts of declining pH levels in the oceans.
The IPCC in the “Fifth Assessment Report” is deeply dubious (Table 12.4) about the various severe effects often hypothesized or asserted as future impacts of increasing GHG concentrations. The one exception is the disappearance of the summer Arctic sea ice, which IPCC views as “likely,” with “medium confidence,” but only under the most extreme GHG concentration path (RCP8.5). Under RCP8.5, GHG concentrations through 2100 rise at 11.9 parts per million (ppm), over six times faster than the average (1.9 ppm) for 1985–2017, in a world in which natural gas use is growing relative to coal use internationally, however slowly and unevenly.
Stelzer at the end of his essay mentions watchful waiting and adaptation (“steps to ameliorate the impact of [climate] change”); given the enormous uncertainties that we agree are present, such adaptation over time is a serious policy. It would be far from doing nothing.
It does not take much time to think of several low-probability catastrophes. Asteroid impacts. Mass volcanic eruptions. Powerful earthquakes. Tsunamis. Mass contagion. Terrorist use of bioweaponry. Nuclear war. Gamma ray storms. Massive crop failures. Does Stelzer believe that we should spend 1–2 percent of gross domestic product as insurance against each of them?
Let me conclude with an observation about the forest as distinct from the trees. Whatever our disagreements, Stelzer’s small essay serves to focus the debate away from the histrionics and silliness that characterize much of the public “discussion” of climate issues in a way that points toward policy questions vastly more relevant and useful. For that alone he has my sincere thanks; and he deserves those of everyone pursuing rigor rather than rhetoric. |
