Energy Taxes and Subsidies: Worthy Goals, Competing Priorities, and Flawed Institutional Design

By David M. Schizer¹

Draft of April 29, 2015

(Please do not post, cite, quote without author’s permission)

We rely on a host of different sources of energy, including petroleum, natural gas, coal, nuclear, wind, biomass, solar, and geothermal. Each offers a different mix of externalities for the environment, national security, and the economy. As a result, the government has a number of reasons to correct these externalities with Pigouvian taxes and subsidies. Many of these reasons are plausible, and some are even compelling. In response, energy subsidies in the U.S. have increased dramatically since 2009.² Between 2007 and 2010, they grew by 108% (from $17.9 billion to $37.2 billion).³ Renewable energy subsidies increased by 186% during this period (from $5.1 billion to $14.7 billion). In total, the U.S. has spent $150 billion on green energy through 2014.⁴ Yet although these subsidies pursue worthy goals, the results often have been confused, inefficient, and even counterproductive. This Article shows why these subsidies have underperformed, and offers suggestions about how to do better.

Specifically, this Article focuses on three sets of challenges. First, we are pursuing competing goals. Some subsidies are supposed to advance environmental objectives, while others are motivated by national security. Indeed, national security is frequently invoked as a goal of energy policy, but only in general terms. To clarify what is at stake – and also what is not – this Article emphasizes two national security externalities. First, the US could face increased defense costs in securing access to petroleum. Second, some energy producers are geopolitical rivals. Their oil and gas revenue can fund policies that are costly to us (and our allies), including terrorism, invasions of their neighbors, and nuclear weapons programs. While the magnitude of these externalities is uncertain, we can reduce both by cutting our demand for petroleum. This goal is largely consistent with our environmental goals. In contrast, another response – expanding and diversifying our supply of petroleum – usually prompts environmental opposition. Meanwhile, although other policies are routinely justified by national security, this Article takes issue with these claims. First, there is no longer a national security argument for coal, now that hydraulic fracturing has unlocked vast domestic reserves of natural gas. Second, for the same reason, the national security case for renewables is weak, although there obviously is an environmental case for these energy sources.

Second, political constraints limit the type of policy instruments used in the U.S.: subsidizing what we want to encourage is easier than to taxing what we want to discourage.⁵ Unlike many other jurisdictions, then, we have not enacted a carbon tax or increased our gasoline tax since 1993.⁶ Instead, we use a hodgepodge of targeted subsidies that favor particular technologies. In other work, I have considered ways to counter this familiar political dynamic – proposing a type of a gas tax that should be more politically palatable⁷ – so this is not the focus here. But this Article offers some additional analysis of this issue, including another way to improve the political prospects of a Pigouvian tax: making it look more like a subsidy by embedding it inside a lump sum payment. For instance, claimants can receive a tax credit of a fixed dollar amount (e.g., of $300), which is reduced as they use more gasoline (e.g., by 30 cents for each gallon they use).

Since we are likely to keep using subsidies instead of Pigouvian taxes, we have to address a third set of challenges, which is the most important of all: our subsidies under current law are riddled with institutional design flaws. For example, by reducing the price of energy, subsidies have the counterproductive effect of increasing demand. They also create a number of perverse effects on supply. For instance, by nurturing future substitutes for hydrocarbons, green energy subsidies induce hydrocarbon producers to cut their prices today. In addition, some reward production, instead of revenue or profit; as a result, they motivate firms to generate energy that is not needed. Likewise, they bet on specific technologies, even though the government does not have the expertise and incentives to “pick winners” effectively. Most energy subsidies are delivered through the tax system, so they are administered by nonexperts and use a currency that is not equally valuable to everyone. Furthermore, many energy tax expenditures are too generous, subsidizing what claimants would do anyway. Indeed, some subsidize behavior that already is legally required. Some of these flaws are more familiar than others. Most (though not all) arise because we use subsidies instead of Pigouvian taxes, and one contribution of this Article is to be more precise on this score.

This Article also breaks new ground in offering guidance about how to correct, or at least to mitigate, these flaws. For example, we should try to subsidize the behavior we actually want (such as saving gasoline), instead of a proxy that is related but different (such as buying a hybrid). “Proxy” subsidies are more likely to produce imperfect or even counterproductive incentives than “results-based” subsidies. In addition, broader subsidies put less pressure on the government to “pick winners.” We also should try to spend only the minimum needed to incentivize the desired behavior. That level should be zero for conduct that already is legally required. It also can be zero (or at least a very low level) when energy prices are high, since high prices produce their own incentives for energy innovation.⁸

Needless to say, the government does not have to pursue energy policy goals only with Pigouvian taxes or subsidies. It also can use a broad menu of other policy instruments, ranging from command and control regulation, cap and trade, and disclosure (for environmental goals) to foreign aid, surveillance, military deployments, and treaties (for national security), and much more. This Article does not seek to cover the entire waterfront of policy instruments, or to delineate precise boundaries among them. Instead, the premise of this Article is that the government will continue to rely heavily on tax expenditures – and may at least consider Pigouvian taxes – so it is important to understand how structure these instruments more effectively.

The three sets of challenges analyzed in this paper are considered in turn. Part I surveys environmental, national security, and economic externalities associated with energy, and shows how empirical uncertainty and competing goals complicate our efforts to correct these externalities. Part II canvasses the political advantages of subsidies over Pigouvian taxes, considers how to improve the political prospects of Pigouvian taxes, and canvasses the hodgepodge of targeted subsidies under current law. Part III analyzes seven institutional design flaws that arise in subsidies under current law, and offers guidance about how to mitigate them. Part IV is the conclusion.

The first step in evaluating energy subsidies under current law, and in determining how they can be improved, is to understand what they seek to achieve. What goals should the government pursue with energy tax expenditures? There is both good and bad news here. The good news is that there are a number of worthy policy goals here, rooted in the environment, national security, and the economy. But the bad news is that these goals are hard to pursue for two reasons. First, there is significant uncertainty about the relevant effects. Second, our environmental and national security goals are sometimes in tension. This Part canvasses various justifications for the government to intervene, highlighting uncertainty we face in pursuing them, as well as tensions among our goals.

A. Environmental Externalities

Many subsidies – especially more recent ones promoting renewables and energy efficiency – pursue a range of environmental goals. They target climate change, pollution, the environmental damage from economic development, and concerns about running out of finite resources. Familiar market failures can justify government interventions here (at least for most of these goals), but the task is complicated by significant uncertainty about the targeted harms.

1. Greenhouse Gas Emissions and Climate Change

Climate change is perhaps the most common justification for efforts to replace fossil fuels. Combustion of these fuels and industrial processes caused 78% of the increase in greenhouse gas emissions (“GHG’s”) between 1970 and 2010, according to the Intergovernmental Panel on Climate Change.⁹ The familiar concern is that GHG’s raise global temperatures. For instance, the IPCC concludes that the “period from 1983 to 2012 was likely the warmest 30-year period of the last 1400 years in the Northern Hemisphere,” ¹⁰ Likewise, the World Bank cites “growing evidence . . . that . . . warming close to 1.5°C above pre-industrial levels by mid-century is already locked-in to the Earth’s atmospheric system.”¹¹

“[T]he signature effects of human-induced climate change—rising seas, increased damage from storm surge, more frequent bouts of extreme heat—all have specific, measurable impacts on our nation’s current assets and ongoing economic activity,” according to the Risky Business Project.¹² The risks include damage to coastal property, declining agricultural yields in some places (possibly offset by rising yields in others), less productivity from outdoor workers, shortages of water, and outbreaks of disease.

While alleviating these risks is a familiar rationale for taxing hydrocarbons or subsidizing alternative energy, a challenge in doing so is that we face multiple layers of uncertainty. First, there are vigorous debates about the level of emissions from particular activities. For example, when natural gas is extracted and transported, some escapes into the atmosphere. Experts differ about the magnitude of these “fugitive emissions,” and thus about how beneficial it is to replace coal with natural gas.¹³

Second, even if we can ascertain how much emissions are rising, the effect on global temperatures is contested. According to the IPCC, past trends do not supply the answer because they “are very sensitive to the beginning and end dates”; indeed, “the rate of warming over the past 15 years . . . is smaller than the rate calculated since 1951.”

Third, even if we could forecast temperature changes accurately, there is uncertainty about the welfare effects of these changes. It is possible that some regions will suffer severe or even catastrophic harms (e.g., from droughts or flooding), while others regions will benefit (e.g., from longer growing seasons in northern latitudes).¹⁴ Since these effects are unlikely to emerge for decades, we need to choose a discount rate to value them. A market rate yields a low present value for future harms, counseling against costly responses today.¹⁵ Some defend market rates as the most reliable benchmark for comparing different forward-looking investments.¹⁶ Others advocate a lower rate on normative grounds.¹⁷ Still others believe the real market rate is lower than we think, since climate change is likely to slow economic growth.¹⁸

For all these reasons, there is no consensus about the social cost of carbon. While the IMF values it at $25 per metric ton of CO2, the Obama administration uses $37, and much lower and higher numbers have been asserted as well.¹⁹ Notably, these estimates are based on global welfare costs, rather than national welfare costs, which are considerably lower (e.g., because the U.S. has fewer areas prone to flooding and is less warm than other, more vulnerable parts of the world). There is a vibrant debate about whether the national or global measure is more appropriate.²⁰

To be clear, though, these various sources of uncertainty are not a reason for inaction. The fact that a cost is hard to measure does not mean its level is zero. But it does mean that research is crucial, and that we should approach policymaking with some humility, relying on the best estimates we have and updating policies as we learn more.

Pollution is another well known source of negative externalities from energy. For example, mining for coal can pollute streams, trigger landslides and disfigure landscapes, while burning it causes smog and acid rain.²¹ Nuclear power generates radioactive waste, while accidents can emit radiation into the atmosphere. Oil can pollute land and water when pipelines leak,²² tankers crash, and offshore rigs malfunction. Hydraulic fracturing uses toxic chemicals that can pollute water wells if spilled on the surface or not disposed of properly.²³ In addition to damaging property, pollution sometimes affects human health as well. There also are other dangers in improperly accessing, transporting, and using energy, such as fires, explosions, and seismic activity.

While these harms can also justify government interventions, uncertainty is also a significant challenge here, as it is with climate change. For example, there is a heated debate about whether hydraulic fracturing is likely to contaminate water, and how feasible it is to regulate this risk effectively.²⁴ Likewise, the riskiness of nuclear power, and the health effects of nuclear waste, are also contested. As a result, environmentalists are divided about whether to rely on nuclear power in the effort to combat climate change.

3. Rapid Economic Development and the Consumption of Finite Resources

While energy development has other environmental costs as well, these are less persuasive as justifications for subsidizing alternative energy. For example, energy development can trigger “boom town” conditions, such as traffic jams, housing shortages, and overtaxed public services. But the right way to deal with traffic in North Dakota is to build roads, not to subsidize solar panels. More generally, local governments should respond to these issues with zoning, taxes, and other familiar instruments.

Another concern about fossil fuels is that they eventually will run out.²⁵ Yet this is not a persuasive rationale for subsidizing alternative energy for two reasons. First, although concerns about exhausting the supply of hydrocarbons are longstanding – indeed, a founder of Standard Oil dumped his shares in 1885, believing the world’s supply of oil was nearly gone – new technologies have consistently been developed to access more.²⁶ Second, and more fundamentally, there is no market failure. As Harold Hotelling showed in 1931, the market price of hydrocarbons reflects their inherent scarcity. Producers have to decide how much to sell today and how much to save for the future.²⁷ If they expect to earn more (in present value terms) by selling later, they will do so. As a result, the current price incorporates predictions about the future, including assumptions about anticipated demand, expected extraction costs, and the likely availability of substitutes. A tighter supply causes the price of hydrocarbons to rise, creating stronger incentives to develop alternatives.²⁸

1. 
   National Security
   

1. Energy Shocks and Defense Costs

The most basic connection between energy and national security is that navies cannot sail, fighter jets cannot fly, and tanks cannot run without fuel. In fact, one of the Allies’ advantages in World War II was more reliable access to oil.³⁰ Yet the military’s need for fuel is only part of the story.

The entire U.S. economy depends on energy. Although the “energy intensity” of the U.S. economy has declined – so we use less energy to generate a dollar of GDP³¹ – oil shocks still have significant adverse effects on the economy.³² Consumers and producers have to adjust in ways that can be especially disruptive in the short run. If the oil is imported, there is a further cost as well: higher prices cause more resources to leak out of our GDP.³³ The more oil we use, the more painful these shocks are. Yet “those purchasing oil,” Brown and Huntington have observed, “are unlikely to understand or consider how their own oil consumption increases the economy-wide effects of oil supply shock.”³⁴

Avoiding energy shocks has been a core goal of U.S. foreign policy since the Arab oil embargo of 1973. In pursuing this goal, the U.S. has repelled Saddam Hussein’s invasion of Kuwait, fielded a significant naval presence in the Persian Gulf, sold weapons to friendly governments in the Middle East, and maintained military bases there.³⁵ To the extent that dependence on oil has caused the U.S. to commit more blood and treasure to national defense, these incremental costs are hidden costs of using oil.

Yet the magnitude of these externalities is hard to estimate. After all, having a powerful military advances a number of important goals unrelated to energy, including promoting our values, deterring conventional attacks, countering terrorism, discouraging the proliferation of weapons of mass destruction, protecting our allies, and facilitating global commerce. It may be, moreover, that only a large change in oil supply or demand would affect U.S. military policy, while a modest change would not. As a result, it is hard to know how much dependence on oil actually adds to U.S. defense burdens and commitments.³⁶

Given this uncertainty, economists generally have not tried to quantify these costs, and many have concluded that they are too speculative to be considered by policymakers.³⁷ Others go further and conclude, as the National Research Council does, that “the marginal cost is essentially zero.”³⁸ But as with climate change, the fact that a cost is conceptually and empirically difficult to pinpoint does not mean its level is zero. While it would be misguided to attribute the entire cost of the U.S. military presence in the Middle East to oil, it also is a stretch, in my view, to conclude that none is attributable to oil. In any event, the National Research Council acknowledges that substantial changes in oil markets could affect the analysis. The shale oil boom was just beginning in 2010, when NRC published their study. The dramatic price decline of 2014, which was driven primarily by growth in U.S. production and took many by surprise,³⁹ was still years away. In light of these tectonic shifts in global oil markets, this issue warrants further study.

In any event, if policing access to oil does add to U.S. defense burdens, more domestically-produced oil is often invoked as a solution. Yet this is not quite right. Even if the U.S. uses only domestic oil, U.S. prices will still spike if a major source of global supply suddenly becomes unavailable. For example, if the Persian Gulf is sealed off, European and Asian buyers will bid up the price of U.S. oil.⁴⁰ Admittedly, price spikes are especially harmful when oil is imported, since they drain more resources out of our economy. But shocks are still disruptive even when oil is domestically produced, creating uncertainty and causing producers and consumers to make painful short-term adjustments.⁴¹

To mitigate these risks, U.S. production is helpful. But new discoveries in Brazil, Canada, and the U.K are helpful as well. When there is more slack capacity in the global oil market, the loss of one source is less likely to cause a price shock.⁴² In addition, when new production comes from stable parts of the world, unstable sources represent a shrinking percentage of global oil production, which further reduces the likelihood and magnitude of shocks.⁴³

Notably, slack capacity can be attained not just with supply enhancement, but also with demand reduction. Since petroleum is mainly used in transportation, greater fuel efficiency is helpful, as is the development of alternative energy sources for transportation.⁴⁴ Fortunately, U.S. petroleum consumption has declined significantly since 2006, in part due to greater fuel efficiency, and in part due to slower economic growth.⁴⁵

2. A Source of Strength for Geopolitical Rivals

Dependence on oil can undermine national security not only in imposing burdens on us, but also in strengthening geopolitical rivals. For example, Russia earns 75% of its export revenue and 50% of its tax receipts from hydrocarbons. Oil revenue can strengthen rivals in three ways. First, they can strengthen their hold on power by buying influence with key domestic constituencies. Second, this money supports their militaries and, in some cases, funds terrorist organizations.⁴⁶ Third, the ability to stop selling energy – and thus to undermine their customers’ economies – is itself a source of power. For example, Russian threats to shut off natural gas sales give it significant sway over its neighbors.⁴⁷ If our reliance on particular types of energy strengthens adversaries in these ways, the energy generates additional negative externalities.

Again, we can mitigate these costs with both demand reduction and supply enhancement. Both strategies put downward pressure on prices, reducing the revenue and leverage of hostile energy producers. For example, a commonly observed advantage of exporting U.S. natural gas is countering Russia’s energy-based influence in Europe. Similarly, the recent surge in U.S. oil production helped build support for sanctions on Iran, since new U.S. production could substitute for Iranian oil.⁴⁸ Increased U.S. oil production has also added to U.S. “soft” power by arguably substituting the U.S. for Saudi Arabia as the world’s “swing” oil producer.⁴⁹

So far, the analysis assumes that cutting the oil revenue of geopolitical rivals affects their capacity to threaten the U.S., but not their motivation to do so. Yet this may not always be true. Some regimes could respond by seeking a better relationship with the U.S. (e.g., to secure aid or an end to sanctions), while others could ratchet up their rhetoric or even precipitate a crisis (e.g., to distract and rally domestic constituencies). There could even be a change in the regime. Thus, although the U.S. is likely to benefit from weakening hostile energy producers, it is possible to imagine other scenarios as well.

Finally, low oil prices obviously create challenges for energy producers that are not hostile to the U.S. Some are close allies and trading partners, such as Canada. Others are at risk of being replaced by hostile insurgencies. For example, Nigeria is battling the Boko Haram insurgency.⁵⁰ Likewise, Egypt and Jordan could be destabilized if Saudi Arabia cuts its support for them in response to declining oil prices. In theory, we can increase our foreign aid budget to offset these adverse effects, but we need to have the institutional capacity and political will to do so effectively. In any event, to the extent that these increases are needed, they offset some of the national security advantages from reduced demand and enhanced supply.⁵¹

2. 
   Economy
   

1. Property Rights, Innovation, and Basic Research

First, it is well understood that property rights are not always broad enough to afford innovators all the benefits of new ideas. “If many people are able to exploit, or otherwise benefit from, research done by others, then the total or social return to research may be higher on average than the private return to those who bear the costs and risks of innovation,” Ben Bernancke has written. “As a result, market forces will lead to underinvestment in R&D from society's perspective, providing a rationale for government intervention.”⁵² This argument is not unique to energy. It is especially persuasive for basic research, since “[t]he most applied and commercially relevant research is likely to be done in any case by the private sector.”⁵³

3. Network Effects, Transaction Costs, and Infrastructure

A second set of market failures, grounded in network effects and transaction costs, impede the development of energy infrastructure. For example, a key advantage of petroleum-powered cars is the prevalence of gasoline filling stations. Since there is no comparable network of natural gas filling stations or electric charging stations, natural-gas- and electric-vehicles are less appealing. Yet if consumers do not buy them, filling and charging stations are unlikely to be built. This network effects creates a “chicken and egg” problem that could justify government intervention.

Similarly, the safest way to transport oil and natural gas is via pipeline, instead of rail or truck. But it is extremely costly to secure rights of way from thousands of landowners. Indeed, a great deal of natural gas in North Dakota is burned at the well, instead of shipped to consumers, because of a lack of pipelines. A government role (e.g., through eminent domain) can be justified to address these transaction costs.

3. Investment Inefficiency

Third, consumers might underinvest in energy efficient cars, appliances, heating systems, and other technologies – a market failure known as “investment inefficiency.” Cognitive biases could keep some from focusing on this opportunity. Others might not take the time to figure out how much energy (and money) they could save, although mandatory disclosure undercuts this explanation. Credit constraints might prevent some upfront investments. Agency costs could also play a role, since those who often are best positioned to make the energy-saving investment (builders and landlords) may not benefit (directly) from reduced energy bills.

Yet this market failure is a slim reed for justifying energy subsidies. For one thing, the empirical evidence is mixed about whether consumers actually do underinvest in energy savings technologies.⁵⁴

In addition, the tax system already offers an (unintentional) subsidy for buying these technologies (though not for renting them), which has been overlooked in the literature: the economic return on these investments – reduced energy bills – is untaxed “imputed” income. For example, assume a taxpayer has two options: first, she can pay $3,000 for an energy efficient refrigerator and save $50 per year in electric bills; second, she can pay $2,000 for a less efficient refrigerator, invest the extra $1,000, and earn $50 per year in interest to fund her higher electric bill. This $50 return is taxable in the second case, but not the first. This favorable treatment of imputed income derives from administrability concerns, rather than a policy choice.

4. Economic Advantages That Do Not Justify a Subsidy

Finally, there are other economic advantages when oil and gas development and “green tech” succeed in the U.S., including economic growth, more jobs, and enhanced purchasing power. These advantages can be substantial.⁵⁵ But although these benefits are sometimes invoked to justify a subsidy,⁵⁶ there is no market failure. As a result, a subsidy actually could induce too much investment, draining resources from other valuable activities. Indeed, there is evidence that green tech subsidies are a costly (and even counterproductive) way to create jobs.⁵⁷

Similarly, although reducing oil imports has economic advantages,⁵⁸ the case for a government intervention is unclear. Presumably, the goal would be for U.S. consumers to keep more of their wealth, instead of transferring it to foreign energy producers. Some argue that the U.S. should pursue this goal by using its market power as a huge consumer of energy (so-called “monopsony” power). If cutting our consumption reduces prices, we would save money on the oil we still import. Yet using monopsony power in this way does not repair a market failure; it actually creates one.⁵⁹ In response, some try to justify this step as a response to another market failure: the market power exercised by OPEC. In any event, the premises of these “market power” arguments are somewhat dated. The price decline in 2014 raises questions about OPEC’s market power.⁶⁰ In addition, the U.S. carries less weight as a consumer, given our successful efforts at energy efficiency and the increasing energy appetites of China and others.⁶¹ Finally, the domestic oil boom in the U.S. means the U.S. is importing much less; we impored 20% in 2014, compared with 60% only a few years ago, and are importing 14 million barrels per day less than the U.S. government forecasted less than a decade ago.⁶² As a result, cutting prices would trigger a transfer to US consumers – not only from foreign producers – but also from U.S. producers.

It should be clear, then, that a number of different market failures – rooted in the environment, national security, and the economy – can justify government efforts to favor or discourage the use of particular types of energy. Yet empirical uncertainty is an important challenge in doing so. In addition, there is an even more fundamental challenge: some of our goals are in conflict. Specifically, the same source of energy can have positive externalities for national security and negative externalities for the environment, and vice versa. Yet it is important to be precise about where these tensions arise, and where they do not. Too many commentators either ignore or overstate these tensions. The main conflict arises with petroleum, so the analysis is different for natural gas, coal, and renewables. Each of these energy sources is considered in turn.

Starting with petroleum, the U.S. is especially vulnerable to oil shocks – and thus is particularly motivated to police access to oil – when supply is constrained relative to demand. Obviously, there are two ways to recalibrate this balance between supply and demand: reducing demand and increasing supply.

The first of these strategies – demand reduction – does not create a tension with environmental goals. By promoting fuel efficiency, car pooling, mass transit, and the like, we become less dependent on petroleum. This eases the need to police access, and also reduces the oil export revenue of geopolitical rivals. In addition to these national security benefits, reducing the demand for oil has familiar environmental advantages. Burning less petroleum means lower GHG emissions and less air pollution. Since there is less need to extract and transport oil, the environmental risks from these activities are reduced as well.

In contrast, the second strategy – expanding and diversifying supply – has national security advantages, as well as environmental costs. Diverse and ample supply reduces the need to ensure that any specific source is available. It also puts downward pressure on price, weakening geopolitical rivals who are funded by oil exports. Yet more supply means more consumption, and thus more GHG’s and pollutants. Offshore oil exploration has become more controversial since the Deep Water Horizon accident in the Gulf of Mexico. In the U.S., onshore development has relied increasingly on hydraulic fracturing, and many environmentalists worry that this drilling technique contaminates drinking water and causes earthquakes. Likewise, national security and environmental advocates have squared off about whether to allow drilling and build pipelines in particular places. The debate about the north slope of Alaska is a longstanding example, and the Keystone pipeline is a more recent one. No doubt, there will be others over the coming years, as the government mediates between these competing sets of externalities.

While petroleum is essentially the only fuel used for transportation in the U.S., it is used only rarely to generate electricity, power industry, and heat homes. For these energy needs, the U.S. relies on natural gas, coal, nuclear, and renewables.

Ten years ago, domestic natural gas reserves were dwindling, and the U.S. was preparing to import significant amounts of natural gas. At the time, the world’s leading natural gas exporters were Russia, Iran, and Qatar. Importing natural gas from these sources would pose many of the same national security issues that arise with petroleum.

As with petroleum, then, there are national security advantages in enhancing and diversifying supply. In the past ten years, hydraulic fracturing has unlocked vast new domestic reserves of natural gas, offering over one hundred years of supply. Because of this “shale revolution,” the price of natural gas has fallen from over $12.00 per mbtu in June of 2008 to less than $3.00 per mbtu in February 2015.⁶³ By assuring a domestic supply of this energy source, the shale revolution has had significant national security advantages. There are no security challenges in ensuring access, and dollars spent on this energy do not strengthen foreign producers that are hostile to the U.S.

To the extent we can replace petroleum with natural gas, the national security advantages multiply. In a limited way, this is happening in two contexts. The first is heating commercial and residential buildings, where cheaper natural gas is replacing more expensive heating oil.⁶⁴ The second is transportation, albeit on a modest scale. Instead of relying on petroleum, electric cars obviously rely on electricity, which can be generated with natural gas. In addition, about 150,000 vehicles in the U.S. have engines powered by natural gas. These vehicles are gaining a modest foothold, since natural gas is significantly cheaper than petroleum.⁶⁵ An important challenge, though, is a lack of natural gas filling stations. This issue is less serious for municipal buses, delivery trucks, and other vehicles that refuel in a depot.⁶⁶ In addition, natural gas filling stations are cropping up near interstate highways to accommodate natural-gas-powered long-haul trucks.⁶⁷

To what extent do these national security advantages come at an environmental cost? The picture is somewhat mixed. On the positive side of the ledger, natural gas pollutes the air much less than coal. In addition, burning natural gas emits about half as many GHGs as burning coal. Indeed, since 2005, GHG emissions have declined more in the U.S. than anywhere else, and the substitution of natural gas for coal is one of the reasons.⁶⁸ U.S. GHG’s declined by about 9.7% from 2005 to 2013 (from 5753.5 million metric tons (“mmt”) to 5195.5 mmt), while GHG’s from power generation declined by 15% (from 2,402.1 mmt to 2,040.5 mmt).⁶⁹ Even so, natural gas is itself a potent GHG, as noted above. The volume of these “fugitive” emissions – and thus the extent of natural gas’s advantage over coal – is debated. In addition, even though natural gas probably is less damaging than coal, it obviously emits more GHG’s than wind or solar. Cheap natural gas makes it more difficult for renewable energy to become economically competitive. Thus, although cheap and plentiful natural gas has clear advantages for national security, its implications for climate change are mixed. In addition, the natural gas boom in the U.S., which relies on hydraulic fracturing, has prompted environmental concerns about water contamination and earthquakes, as noted above. Environmentalists also warn that drilling damages habitats and contributes to congestion and traffic. In addition, it is common for environmentalists to oppose the construction of pipelines, which are needed to bring natural gas to market. With natural gas, then, there also is some tension between environmental and national security goals.

In contrast, the analysis of coal is more straightforward. It is a familiar point that coal has especially large environmental externalities. But are there offsetting national security reasons to use coal?

Because the U.S. has ample domestic reserves, the national security advantages of coal are commonly invoked. To reduce dependence on Middle Eastern oil, the government has periodically sought to replace petroleum with coal in transportation. The “synfuels” program pursued this national security objective unsuccessfully the 1970’s and 1980’s. Electric cars are a more recent mechanism for replacing petroleum with coal, since coal-fired power plants can generate electricity for these cars.

Yet the boom in natural gas production has an important implication, which has not been adequately recognized: As long as we produce domestic natural gas, there is no national security case for coal. After all, natural gas offers the same national security advantages – since we also have sizable domestic reserves – but at lower environmental cost. Given the ready availability of a superior substitute, the best national security argument for coal is diversification. If we find ourselves in a situation where we no longer can use domestic natural gas, we could fall back on coal. But this is a contingent case at best. Although the government continues to spend billions of dollars on “clean coal” technology, the case for these expenditures is significantly weaker than it used to be.

Similarly, although advocates of renewable energy tout its national security advantages, the real case for this technology is grounded in the environment, not national security. The heat and electricity from renewables can be generated with domestically produced natural gas or coal instead. Likewise, renewables can replace petroleum by powering electric cars, but so can natural gas and coal.⁷⁰

In the near term, the only national security argument for renewables is a contingent one: if we stop using domestically-produced natural gas and coal for environmental reasons – for instance, if there is a national ban on hydraulic fracturing and the use of coal – we would need another domestic source of energy. Yet this is not so much a national security defense of renewables, as it is a further affirmation of the potential tension between national security and environmental goals.

Over the (very) long term, if natural gas reserves run out in one hundred years, we might need renewables then. Likewise, the environmental benefits of renewables could translate into national security benefits, avoiding famines, droughts, and other sources of instability from climate change.⁷¹ Yet this payoff is uncertain and is fifty or one hundred years off.

***

To sum up, then, energy taxes and subsidies can be justified by a range of externalities, grounded in the environment, national security, and the economy. Some are more persuasive than others, but overall the case for government intervention is strong. Yet an important challenge is that a number of the relevant externalities are difficult to measure or predict. In addition, some types of energy generate both positive and negative externalities, requiring policymakers to mediate among competing priorities. Even in the absence of political and administrative constraints, then, these goals are not easy to pursue. Once these factors are added to the mix – as the next two Parts show – the challenges multiply.

The last Part surveyed a range of negative externalities associated with various types of energy. Since consumers do not bear these costs, they do not consider them in deciding how much energy to use. In theory, we could solve this problem by adding a tax equal to these hidden costs, so consumers take them into account. This Part offers a simple example of how this might be done. Yet current law takes a very different approach, using a hodgepodge of targeted subsidies that favor particular technologies. Why do we use targeted subsidies instead of Pigouvian taxes? Politically, it is easier to subsidize what we want to encourage than to tax what we want to discourage. After briefly surveying this familiar political dynamic, this Part consider how to structure Pigouvian taxes so they are more politically plausible, and then canvasses the scattered array of energy tax expenditures under current law.

The last Part showed that enhancing and diversifying the supply of petroleum (and to an extent natural gas) generates positive externalities for national security, but negative externalities for the environment. Declining to tap new reserves can create national security costs, while tapping them can cause environmental harm. Each of these goals is important, so how do we reconcile them?

Pigouvian taxes offer a straightforward solution: to account for these competing effects, we should impose a separate Pigouvian taxes on each harm. In deciding how much energy to use, then, consumers will pay a price that equals the sum of the private cost and these various social costs. The analysis above suggests we would need at least four different Pigouvian taxes, whose levels would vary for each energy source.

First, a carbon tax, C, could be imposed on hydrocarbons, based on the GHGs they emit when burned, extracted, and transported. Thus, C would be higher for coal than for oil and natural gas. A range of familiar implementation challenges should be considered.⁷² For example, even if the tax seeks to maximize national (as opposed to global) welfare, we would want to impose it on foreign consumers as well, since their emissions affect us.⁷³

Second, another Pigouvian tax, P, could internalize pollution risks from using, extracting, and transporting different sources of energy. For instance, the P for offshore oil drilling would reflect the risks of oil spills. The P for hydraulic fracturing would internalize risks of water contamination. The P for nuclear energy would reflect the risk of accidents and the cost of disposing of nuclear waste, and so forth. A range of implementation issues obviously would have to be considered. In some cases, we will want effluent taxes (which measure and tax pollution), while in others it may be easier to have “product” taxes (which tax the use of a product that involves pollution, instead of the pollution itself). Likewise, if the tax is supposed to maximize national welfare, P is likely to be lower for imported fuels, since pollution in extracting these fuels should have less effect on the U.S.⁷⁴

Third, additional Pigouvian taxes could internalize national security risks. For example, a tax, NR, could apply to fuels imported from geopolitical rivals (the “R” in “NR”), reflecting the risks in providing them revenue that can fund terrorism, nuclear programs, and the like. A fourth tax, ND, could internalize defense costs (the “D” in “ND”) associated with policing access to the fuel.

Admittedly, implementing a national security tax would be difficult, and I am not aware of any work analyzing how to do so. For example, classifying nations as hostile or friendly – or as hard or easy to defend – would give the State Department another diplomatic lever, but one that could be challenging to use.

Quantifying ND would also be quite difficult. Since the oil market is global, even users of domestic oil are affected by oil shocks, as noted above, so they have a stake in securing access to Middle Eastern oil. This means ND probably is positive even for crude from North Dakota and Canada. But ND should be higher for Middle Eastern crude, if only to incentivize the development of more stable sources.

In varying taxes by country, we face not only these empirical and conceptual challenges, but also the constraints imposed by trade treaties. National boundaries pose a further challenge as well. Even if the tax is supposed to advance only national welfare, we still would want to impose it on consumers outside the U.S., since their consumption choices – for instance, in supporting hostile regimes – create negative externalities for us as well. The diplomatic challenges in arranging this would be immense. Yet fortunately, most of our trading partners already have high taxes on petroleum.

In addition, we can also add taxes to account for externalities from traffic and congestion, as well as the cost of reads. We also can add instruments to internalize externalities from research, network effects, and so forth.

Under this segmented approach, then, the total Pigouvian tax would depend on the type of hydrocarbon, as well as its source. For example, oil from North Dakota or Canada would not be subject to NR, but would be subject to C and P, as well as a low ND. Oil from Saudi Arabia would be subject C, a higher ND, and a lower P. Nuclear energy would not be subject to C or NR, but it would be subject to P and perhaps also a modest ND (reflecting the cost of securing the plant and its nuclear fuel from terrorist threats). Domestic coal would be subject to C and P, but not NR or ND. The same is true of domestic natural gas, although C would be lower.

Needless to say, the administrability (and political) challenges here are significant. Even if we had perfect information about the magnitude of various externalities, there are incremental administrative costs in tailoring individualized taxes in this way. Pervasive uncertainty about the magnitude of these various externalities – and how they vary in different contexts – makes this an even taller order. Given these challenges, the most we can hope for is a rough estimate, which accounts for relative, as well as absolute, levels. For example, even if we cannot be precise about the size of the externalities from natural gas and coal, we know the environmental harm from coal is likely to be worse, so taxes and subsidies should reflect this difference. Likewise, we would impose an extra national security tax on oil, which would not be needed for coal or natural gas. We could try also to vary the tax, at least to an extent, depending on the source of the oil.