Capstone Project!!

CAP AND DIVIDEND:

A STATE-BY-STATE ANALYSIS James K. Boyce & Matthew E. Riddle Political Economy Research Institute University of Massachusetts, Amherst Economics for Equity and the Environment Network August 2009 This report is jointly published by the Political Economy Research Institute and the E3 Network. The Political Economy Rese arch Institute (PERI) promotes human and ecological well-being through original research. Our approach is to translate what we learn into policy proposals that are capable of improving life on our planet today and in the future. In the words of the late Robert Heilbroner, we “strive to make a workable science out of morality.” Established in 1998, PERI is an independent unit of the University of Massachusetts, Amherst, wi th close links to the Department of Economics.

Economics for Equity and th e Environment Network (E3) is a national network of more than 250 economists who are developing new and applied ec onomic arguments for environmental protection with an explicit focus on social justice. E3 economists are commi tted to using their research and expertise to advance the aims of the progressive movement through dialogue and collaboration with NGOs, the public, decision makers, and the media. The views expressed in this report are the author s’ and do not necessarily reflect those of the sponsoring institutions.

Acknowledgements: We are grateful to Jesse Sanes for research assist ance, Mike Sandler for preparation of maps, Jesse Jenkins for calculations of the stat e-level carbon intensity of electricity, and Peter Barnes and Kristen Sheeran for comments on an earlier draft of this paper.

About the authors: James K. Boyce is a professor of economics at the Universi ty of Massachusetts, Amherst, and director of PERI’s program on development, peacebuilding and the environment. He is a member of the E3 steering committee.

Matthew E. Riddle is a doctoral candidate in economics at the University of Massachusetts, Amherst, and a research analyst with the Center for Social Ep idemiology and Population Health at the University of Michigan.

PERI Gordon Hall 418 N. Pleasant St.

Amherst, MA 01002 www.peri.umass.edu Equity and the Environment Network 721 NW Ninth Ave Suite 200 Portland, OR 97209 www.e3network.org CAP AND DIVIDEND:

A STATE-BY-STATE ANALYSIS James K. Boyce & Matthew E. Riddle Political Economy Research Institute University of Massachusetts, Amherst August 2009 ABSTRACT The impacts on consumers of a cap on carbon emissions will vary across income brackets and across the 50 states. This paper provides state- level estimates of these impacts by income decile. We then estimate the net effect of a cap-and-dividend policy in which all carbon permits are auctioned and 80% of the revenue is returned as dividends to the public. We find that inter-state differences are small compared to the differences across income brackets. Within each state, at least 60% of households receive net benefits: the dividends more than offset the impact of higher fossil fuel prices on their real incomes. Differences across states are small in cap-and-dividend compared to inter-state differences in per capita spending for defense and federal farm programs. The high visibility of dividends, coupled with the positive impact on family incomes, could enhance public support for a durable climate policy. s s Key words: Climate change; climate policy; fossil fuels; global warming; cap-and-trade; energy policy.

JEL codes: H22, H23, Q48, Q52, Q54, Q58 EXECUTIVE SUMMARY A cap-and-permit system to curb carbon dioxide emissions from burning fossil fuels will raise prices to consumers. Individual carbon footprints will now carry a price tag. The money that consumers pay in higher prices will not disappear from the nation’s economy, however: it will be transferred to the owners of the carbon permits.

A cap-and-dividend policy would put this ownership in the hands of the people. It would do so by auctioning the permits and returning most or all of the revenue to the public as equal per- person dividends. If 100% of the permits are auctioned, there is no need for permit trading in secondary markets, no siphoning of revenue into trader profits, and no risk that speculators will manipulate the carbon price. The cap-and- dividend policy would provide incentives for businesses and households to curtail their use of fossil fuels, while protecting consumers from the impact of higher prices on their real incomes. In this paper, we examine differences in the impact of a cap on carbon emissions across income brackets and across the 50 states. We then estimate the net effect of a cap-and- dividend policy. We find that in every state the majority of families come out ahead: the dividends they receive more than offset the impact of price increases. Differences across states are shown in Figure A.

They are small compared to differences across income brackets. Because dividends are distributed equally to each person, variations in cap-and-dividend’s net impact arise solely from differences in carbon footprints. Households who consume more carbon, directly via fossil fuels and indirectly via other goods that are produced and distributed using them, will pay more; those who consume less will pay less. The differences between average carbon footprints in the top 10% and bottom 10% of the income distribution are far wider than differences across the states. FIGURE A : IMPACT OF CARBON PRICING ON MEDIAN FAMILY OF FOUR ($/YEAR , AT $25 /TON CO 2) CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / EXECUTIVE SUMMARY / PAGE ii The inter-state differences in net benefits from cap-and-dividend are also small relative to those in many other public policies. Figure B compares them to differences in per capita spending on defense and federal farm programs. The ratio between the top ten and bottom ten states is more than 11:1 in the case of defense spending, and 190:1 in the case of farm programs. In the case of cap-and-dividend, it is only 2½:1.

Cap-and-dividend would return carbon revenue in equal measure to each American. In contrast, the American Clean Energy and Security (ACES) Act, passed by the U.S. House of Repre- sentatives in June 2009, would allocate revenues and free permits in a variety of ways with uneven effects across households. The Congressional Budget Office (2009) estimates that under ACES roughly two-fifths of the carbon revenue (or “allowance value”) would flow to households in the top quintile of the national income distribution. In Figure C this outcome is contrasted with cap-and-dividend, in which each quintile receives the same amount, 20%, equal to its share of the population.

The visibility of the transfers of carbon revenue to the public may be even more important than the net distributional effects of climate policy.

Dividends to the public in the form of checks in the mail or deposits into bank accounts will provide highly tangible benefits to families, against which they can weigh the impacts of higher fossil fuel prices. Transfers to households resulting from ACES – via myriad routes such as capital gains to corporate shareholders and rebates in electricity bills – will be less apparent. For reasons of both economic fairness and transparency, therefore, cap-and-dividend offers a way to secure durable public support for an effective policy to wean the economy from dependence on fossil fuels. A proactive U.S.

policy, in turn, will be a crucial condition for reaching an effective international agreement to confront the global challenge of climate change. s $0 $500 $1,000 $1,500 $2,000 $2,500 Defense expenditure Farm programs Net impacts of cap-and- dividend Mean of top ten recipient states Mean of bottom ten recipient states FIGURE B . TOP TEN AND BOTTOM TEN STATES :

DEFENSE EXPENDITURE , FARM PROGRAMS , AND CAP -AND -DIVIDEND POLICY FIGURE C . DISTRIBUTION OF CARBON REVENUES TO HOUSEHOLDS : ACES V . CAP -AND -DIVIDEND POLICIES 15% 17% 15% 16% 38% 20% 20% 20% 20% 20% 0% 10% 20% 30% 40% Bottom 20% Next 20% Middle 20% Next 20% Top 20% ACES Cap-and-dividend CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 1 I. INTRODUCTION This paper examines inter-state differences in the impact on households of policies that “put a price on carbon,” that is, policies that increase the price of fossil fuels to curtail emissions of carbon dioxide into the atmosphere. In particu- lar, we examine the impact of a “cap-and- dividend” policy that limits the quantity of car- bon entering the U.S. economy, auctions per- mits up to this cap to the firms that supply fossil fuels, and returns all or most of the auction revenue to households in the form of equal per capita dividends.

The paper is organized as follows. Section 2 re- views the basic features of the cap-and-dividend policy, including the rationale for carbon pricing, differences between a cap-and-permit policy and a carbon tax, and how to return auction revenue to the public as dividends.

Section 3 provides a brief overview of the distri- butional impact of cap-and-dividend at the na- tional level. We examine both the gross impact of higher fossil fuel prices and the net impact when revenues are returned to the public. For the latter calculation, we assume that 80% of the revenues are returned to the public as divi- dends – a percentage roughly the same as what President Barack Obama proposed in his Feb- ruary 2009 budget. An attractive feature of cap- and-dividend is that the policy delivers positive monetary benefits to low-income and middle- income households, even without counting the environmental benefits of mitigating climate change. At the same time, it rewards house- holds at any income level who reduce their car- bon footprints.

Section 4 examines inter-state variations in the impact of higher fossil fuel prices. We analyze three sources of variations: (i) differences in in- come; (ii) differences in consumption patterns; and (iii) differences in the carbon intensity of electricity consumed. Because the impact varies across the income distribution, we present these results by income decile (tenths of the population ranked by per capita income) as well as for the median household in each state. We then provide a state-by-state analysis of the net impact of the cap-and-dividend policy on a dec- ile-by-decile basis. We show that inter-state variations are minor relative to variations based on income.

Section 5 discusses other, non-dividend uses of carbon revenues. Specifically, we discuss (i) transitional adjustment assistance, the main aim of which is to create jobs in communities adversely impacted by reduced production and use of fossil fuels; and (ii) the mix of uses pro- posed in the American Clean Energy and Secu- rity (ACES) Act of 2009, also known as the Waxman-Markey bill.

Section 6 summarizes our main findings and offers some concluding remarks. II. CAP-AND-DIVIDEND: THE BASICS Any policy that limits the supply of fossil fuels will raise their price. The economic logic binding price to scarcity holds true, regardless of the cause of scarcity. When OPEC wants to increase the price of oil, it cuts production. If lawmakers place a cap on carbon emissions from burning fossil fuels, this too will increase their price. 1 There is a crucial difference, however, between higher prices caused by a carbon cap and higher prices due to other forces. The higher prices from a carbon cap will be a cost to con- sumers, but not to the economy as a whole. In- stead they are a transfer. Every dollar paid by consumers in higher fuel prices will go to the holders of carbon permits. Unlike price rises due to market forces or OPEC supply restrictions, the price rise due to a carbon cap simply recycles dollars within the United States.

A key question is: who will get these dollars?

There are three possible answers:

Profits to corporations: If permits are given free- of-charge to corporations, they will reap windfall profits. Consumers will pay higher prices, and CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 2 the firms and their shareholders will get the money. This is a “cap-and-giveaway” policy.

Revenues to government: If permits are auc- tioned rather than given away, the permit value (the counterpart to the higher prices paid by consumers) will be captured by the government.

If this money is used to fund public expendi- tures or cut taxes, the distribution of benefits to the public will depend on the specifics of these uses. This is a “cap-and-spend” (or “cap-and- invest”) policy.

Dividends to the people: If the revenue from permit auctions is returned to the public as equal per capita dividends, consumers will be partially or fully insulated from the impact of higher prices. Households with small carbon footprints will come out ahead, receiving more in dividends than they pay in higher prices. This is a “cap-and-dividend” policy.

The stakes are high. A carbon cap will bring the greatest allocation of new property rights in the United States since the Homestead Act of 1862.

The value of permits under a cap that cuts emis- sions 80% by 2050 – the goal endorsed by cli- mate scientists and embodied in legislation now before Congress – will amount to trillions of dol- lars over the next forty years. The mechanics of cap-and-dividend A carbon cap will be most efficiently adminis- tered “upstream,” by requiring permits (some- times called “allowances”) to be purchased by the first sellers of fossil fuels into the economy.

The cap will reduce supply and raise fuel prices; in this respect it is akin to a carbon tax (for dif- ferences between permits and taxes, see the sidebar on page 3). The resulting market signals will spur businesses and households alike to invest in energy efficiency and clean energy. In a cap-and-dividend policy, the permits are auctioned by the government and all or most of the auction revenue is returned to the public as equal payments per person. This is what economists call a “feebate” arrangement: indi- viduals pay fees based on their use of a scarce resource that they own in common, and the fees are then rebated in equal measure to all co- owners. In this case, the scarce resource is the U.S. share of the carbon storage capacity of the atmosphere; the fee is set by the carbon foot- print of the household; and the co-owners are the American people.

One way to disburse dividends is via ATM cards, similar to those used today by many Americans to access Social Security payments. At the ATM, individuals can check on the auction revenue deposited into their accounts and withdraw funds at their convenience.

With auctions, no need for permit trading In his budget proposal submitted to Congress in February 2009, President Barack Obama af- firmed the principle that 100% of carbon per- mits should be auctioned.

With 100% auction, there is no need for permit trading. Auctions can be held monthly or quar- terly, with the number of permits on offer being reduced gradually as the carbon cap tightens over time. The permit allows its holder to bring a fixed quantity of fossil carbon into the economy in a certain time frame, say over a two-year pe- riod from the date of purchase. Firms simply buy the number of permits they want at the auction.

Most permits in our society are not tradable.

Driving permits, gun permits, parking permits, landfill disposal permits, and building permits cannot be traded in markets. There is no reason why carbon permits should be different.

The need for tradable permits (“cap-and-trade”) is premised on the assumption that some or all of the permits are given away free-of-charge rather than sold by auction. Such giveaways must A CARBON CAP WILL BRING THE GREATEST ALLOCATION OF NEW PROPERTY RIGHTS IN THE UNITED STATES SINCE THE HOMESTEAD ACT OF 1862. CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 3 be based on some formula (like historic emis- sions). Some firms will get more permits than they need, while others will get fewer than they want; trading is necessary to redistribute them from the former to the latter. If 100% of the car- bon permits are auctioned, however, permit trading becomes unnecessary. 23 With non-tradable permits, trader profits do not drive a wedge between the amount paid by con- sumers in higher prices and the amount of available revenue from permit sales. None of the carbon revenue is siphoned off by specula- tors or trading firms. Non-tradable permits also safeguard the policy from the perception or real- ity of market manipulation by players seeking to game the system. 4 Dividends versus other uses of carbon revenue Rather than returning 100% of carbon revenues to the public, policymakers could dedicate a por- tion of the revenues to other uses. In his Febru- ary 2009 budget, for example, President Obama proposed using 81.4% of projected carbon reve- nues for the years 2010-2019 for lump-sum tax credits (extending the “Making Work Pay” credits that were initiated in the economic stimulus pro- gram) and devoting the remainder to investment in clean energy technologies. 5 Apart from clean energy investments, other po- tential uses for carbon revenues include offset- ting the impact of higher fossil fuel prices on the purchasing power of federal, state, and local governments; transitional adjustment assis- tance to workers, communities, or firms ad- versely affected by the transition away from fossil fuels; and other government expenditures, tax cuts, or deficit reduction.

Following the contours of President Obama’s budget proposal, in the following analysis we assume that 80% of carbon permit revenues are returned to the public as dividends, and that the remaining 20% are allocated to other uses.

In section 5 we further discuss some of these potential uses. PERMITS VERSUS TAXES An alternative way to put a price on carbon is by means of a tax. A carbon tax is simply a permit with a fixed price. A cap-and-permit policy sets the quantity of permits (and hence emissions), and lets demand determine the permit price; a carbon tax sets the price, and lets demand determine the quantity of emissions. In both cases, higher prices provide a market signal to encourage energy efficiency and investments in alternative energy.

If policymakers could have perfect foresight as to future demand for fossil fuels – knowing what new technologies will become available, when the economy will boom and slump, and so on – then setting either the quantity of permits or the carbon price could achieve exactly the same result. In reality, there is much uncertainty about future demand, so the relationship be- tween quantity and price cannot be predicted with much precision.

The fundamental aim of climate policy is to re- duce emissions to reach the 2050 target. There- fore, a compelling case can be made for “getting the quantity right” by setting the number of per- mits and letting their price vary with demand, rather than vice versa. Moreover, in the face of uncertainties as to the relation between quantity and price, there may be political pressures to set the carbon tax too low, based on optimistic pro- jections of the resulting emission reductions.

On the other hand, political pressures may also undermine the efficacy of a cap-and-permit policy. This can happen in two ways: first, by set- ting the cap at a level that is inadequate to achieve the necessary emission reductions; and second, by allowing “offsets,” whereby instead of curtailing fossil fuels, firms can get credits for other actions such as planting trees, slowing deforestation, or reducing carbon emissions in other countries. 2 The case for permits rather than taxes is prem- ised, therefore, on a “tight” cap: one that reduces emissions to meet the 2050 target, without offsets that transform the cap into a porous sieve. If policymakers instead opt for a carbon tax, the question of how to distribute the revenue will remain. The analysis presented in this paper would apply equally to a “tax-and- dividend” policy. 3 CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 4 III. DISTRIBUTIONAL IMPACTS OF CAP-AND-DIVIDEND AT THE NATIONAL LEVEL The cap-and-dividend policy will have a progres- sive impact on income distribution nationwide.

Households with smaller-than-average carbon footprints pay less in higher fuel costs than they receive as dividends; households with larger-than- average carbon footprints pay more than they re- ceive. In general, lower and middle-income house- holds will come out ahead, for the simple reason that they consume much less carbon than upper- income households. Overall, roughly three- quarters of American families will obtain positive net benefits in purely monetary terms, not count- ing the environmental benefits that are the main rationale for a carbon-pricing policy.

To calculate the net impact across income brackets, we first estimate the carbon footprints of households: the carbon dioxide emissions re- sulting from not only their direct fuel consump- tion but also the production and distribution of other goods and services that they consume. 6 Data on expenditure patterns are drawn from the Consumer Expenditure Survey conducted by the U.S. Bureau of Labor Statistics. Lower-income households generally devote a larger fraction of their expenditure to direct fuel consumption than upper-income households (in economic parlance, fuels are “necessities” not “luxuries”).

Carbon emissions per dollar expenditure for dif- ferent items are based on input-output data. As one might expect, this ratio varies greatly across expenditure categories. In the case of electricity and household fuels, one dollar of spending generates about 7 kg of carbon dioxide emis- sions. In the case of services, the corresponding amount is about 0.3 kg. The distribution of carbon emissions across ex- penditure categories is shown in Figure 1. Gaso- line and electricity consumption each account for about one-quarter of the average household’s carbon footprint. Natural gas and heating oil con- tribute a further 12%. Indirect uses – including consumption of food, industrial goods, services, and other transportation – account for the rest. Because low-income households consume less than high-income households, they generally have smaller carbon footprints. Differences across the income spectrum are shown in Fig- ure 2a. In the highest income decile, carbon emissions per capita are more than six times greater than in the lowest decile.

As a share of their income, however, the poor consume more carbon than the rich – that is, 0 5 10 15 20 25 30 12345678910 decile heating oil 3% natural gas 9% indirect 37% electricity 25% gasoline 26% FIGURE 1 : HOUSEHOLD CARBON FOOTPRINT BY EXPENDITURE CATEGORY (NATIONAL AVERAGE ) FIGURE 2A : CARBON FOOTPRINT BY INCOME DECILE ( METRIC TONS CO 2 PER CAPITA ) more carbon per dollar of their income. This is primarily because, as noted above, direct fuel consumption accounts for a bigger fraction of their household budgets: they spend more on necessities and less on luxuries. Carbon per dol- lar of expenditure is more than twice as high in the poorest decile as in the richest, as shown in Figure 2b. Hence, a price on carbon is regres- sive in and of itself, hitting the poor harder as a fraction of their incomes than the rich. FIGURE 2B : HOUSEHOLD CARBON FOOTPRINT BY INCOME DECILE (KILOGRAMS CO 2 PER $ INCOME PER CAPITA ) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 12345678910 decile The net impact of the policy depends, however, on who receives the money generated by the carbon price. If this money is captured by auc- tioning the carbon permits – rather than giving them away free-of-charge – and if most of the resulting revenue is returned to the public in dividends, the net impact turns progressive. To illustrate, we assume that the permit price is $25 per ton of carbon dioxide, all permits are auctioned, and 80% of the revenue is returned to the people as dividends. This price is within the range of projections based on current legis- lative proposals; for example, the Congressional Budget Office (2009) estimates that the Wax- man-Markey bill would result in a permit price of $28/tCO 2 in the year 2020. A more aggressive policy, with a more ambitious schedule for emission reductions and/or fewer “offsets,” would generate a higher price. This would in- crease the magnitude of the impacts of the cap- and-dividend policy, but it would not alter their distributional incidence.

The impact of the cap-and-dividend policy is shown in Table 1. The annual carbon charge – TABLE 1 : DISTRIBUTIONAL IMPACT OF CAP -AND -DIVIDEND AT THE NATIONAL LEVEL ( $25 /T CO2 , WITH 80% OF REVENUE DISTRIBUTED AS DIVIDENDS ) $ per capita % of income Per capita income decile Per capita income (in 2003 dollars) Average household size Carbon charge Dividend Net impact Carbon charge Dividend Net impact 1 3844 4.5 135 386 251 3.5% 10.0% 6.5% 2 6538 3.6 177 386 209 2.7% 5.9% 3.2% 3 8968 3.2 209 386 177 2.3% 4.3% 2.0% 4 11544 2.9 238 386 148 2.1% 3.3% 1.3% 5 14481 2.7 267 386 119 1.8% 2.7% 0.8% 6 18034 2.4 299 386 87 1.7% 2.1% 0.5% 7 22623 2.3 337 386 49 1.5% 1.7% 0.2% 8 29120 2.1 385 386 1 1.3% 1.3% 0.0% 9 39942 2.0 457 386 -71 1.1% 1.0% -0.2% 10 67940 1.7 618 386 -232 0.9% 0.6% -0.3% Mean 23657 2.5 317 386 69 1.3% 1.6% 0.3% Median 16160 2.0 283 386 103 1.7% 2.4% 0.6% CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 6 the cost to consumers from higher prices for fos- sil fuels, and for other goods and services that use them in their production and distribution, ranges from $135 per person in the lowest- income decile to $618 per person in the high- est. 7 Each household receives the same per cap- ita dividend, $386. The bottom seven deciles come out ahead, receiving more in dividends than they pay as a result of higher fuel prices; the eighth decile breaks even; and the top two deciles experience a net cost. As a percentage of income, the lowest decile sees a 6.5% gain, while the top decile sees a 0.3% loss.

The monetary winners outnumber the losers for two reasons. The first is that the U.S. income distribution is strongly skewed to high-income people. As shown in Appendix Table A.1, the na- tional mean (average) per capita income in 2003 was $23,657, whereas the median in- come – that of the “middle American,” 50% of the population having higher incomes and 50% lower – was $16,160. Just as mean income is pulled above the median by the high incomes at the top, per capita dividends are pulled up by the outsized carbon footprints of high-income households.

The second reason is that our calculations are based on the assumption that 80% of total car- bon revenue is returned to households. House- hold consumption accounts for only 66% of total carbon emissions in the United States, however, and hence for roughly the same share of total carbon revenues. The remaining emissions come from local, state and federal government (14%), non-profit institutions (8%), and produc- tion of exports (12%). 8 While the results in Table 1 show the broad pat- tern of distributional impacts from the cap-and- dividend policy, the impact on individual house- holds will depend on their consumption choices.

Upper-income households who reduce their car- bon footprints well below the norm for their in- come bracket can derive positive net benefits, too; conversely, lower and middle-income households with disproportionately large carbon footprints can come out behind. Regardless of income level, higher fuel prices provide incen- tives for energy efficiency and alternative fuels.

Those who respond strongly to these market signals fare better than those who do not curtail their consumption of fossil fuels. 9 In sum, the progressive impact of per capita dividends more than offsets the regressive im- pact of higher fossil fuel prices. The majority of American families are “held harmless” by the policy: their real incomes are protected, and in many cases increased. This, in turn, protects the nation’s climate policy from the political backlash that higher fuel prices could other- wise trigger.

IV. STATE-BY-STATE IMPACTS OF CAP-AND-DIVIDEND One issue that has received attention in Con- gress is the differential effects that carbon pric- ing may have across the states. In a June 2009 interview with The New York Times, President Obama alluded to this issue when he described the compromises in the Waxman-Markey bill as having been “necessary to moderate the differ- ent effects of greenhouse-gas controls on dif- ferent parts of the country” (Broder 2009).

Two broad sorts of inter-state differences can be distinguished. The first is on the consump- tion side of the economy, arising from differ- ences in household use of fossil fuels (both direct and indirect) and hence in the impact of higher prices on consumers. The second is on the production side, arising from differences in how firms and workers are affected by the tran- sition away from burning fossil fuels. In this sec- tion our focus is the consumption side.

REGARDLESS OF INCOME LEVEL, HIGHER FUEL PRICES PROVIDE INCENTIVES FOR ENERGY EFFICIENCY AND ALTERNATIVE FUELS. CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 7 Impact of higher fossil fuel prices on households The higher fossil fuel prices that result from any policy that puts a price on carbon will have dif- ferent impacts on consumers in different states for three reasons:

Income differences: States vary in both average income and income distribution. Just as people in upper-income households tend to have larger carbon footprints than lower-income households (see Figure 2a), people in higher-income states tend to have bigger carbon footprints, all else equal, than their counterparts in lower-income states.

Differences in consumption patterns: Energy use is affected, among other things, by public policies and the weather. In California, for ex- ample, policies to promote energy efficiency have paid off by reducing the state’s per capita electricity use considerably below the national average. Gasoline consumption varies due to differences in commuting distances, public transportation, and the level of state gasoline taxes. In the northern states, households spend more to heat their homes; in the southern states, they spend more to cool them.

Differences in the carbon intensity of electricity:

Some states rely mostly on coal-fired power plants, which generate higher carbon emissions per kilowatt-hour than other electricity sources.

Some states rely more on hydroelectric power, nuclear power, or other low-carbon technolo- gies. Electricity accounts for roughly one-quarter of the typical household’s carbon usage (see Figure 1); differences in the carbon intensity of electricity affect this component of the impact of carbon pricing on consumers.

Table 2 presents data on the extent of inter- state differences in these respects. Per capita income varies from about $11,500 in Missis- sippi to about $21,000 in Connecticut.

TABLE 2 : INTER -STATE DIFFERENCES IN INCOME AND ENERGY USE Expenditure per capita of median household ($) State Median income (annual per capita) Electricity Gasoline Natural gas Fuel oil Carbon intensity of electricity supply (kg CO 2/MWh) Alabama 13,308 416 446 100 23 669 Alaska 18,806 345 481 136 22 546 Arizona 15,544 314 412 126 9 558 Arkansas 12,772 411 437 98 23 630 California 16,616 195 525 106 12 454 Colorado 18,829 332 450 134 9 913 Connecticut 20,964 219 481 107 187 412 Delaware 18,527 330 431 157 69 933 District of Columbia 17,795 453 513 110 25 734 Florida 15,925 384 441 8 6 672 Georgia 15,895 438 487 106 24 708 Hawaii 16,969 392 454 8 7 857 Idaho 14,231 317 422 124 20 459 Illinois 17,521 348 484 212 23 556 Indiana 16,350 341 468 208 23 1,041 Iowa 15,925 304 471 212 14 933 Kansas 16,138 305 473 213 14 918 Kentucky 13,417 321 425 194 21 1,002 Louisiana 12,179 405 426 97 23 745 Maine 15,398 200 418 97 172 455 CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 8 Per capita expenditure on electricity by the me- dian household in each state ranges from $195/year in California to $458/year in Vir- ginia. Variations in per capita gasoline expendi- ture are less pronounced, ranging from $377/year in New Mexico to $537/year in Texas. Natural gas use is highest in the upper Midwest, and heating oil use is concentrated in the northeastern states. The carbon intensity of electricity varies widely across the states. North Dakota, a state that is heavily reliant on coal-fired power plants, emits 1134 kg of carbon dioxide per megawatt hour TABLE 2 : INTER -STATE DIFFERENCES IN INCOME AND ENERGY USE , CONTINUED Expenditure per capita of median household ($) State Median income (annual per capita) Electricity Gasoline Natural gas Fuel oil Carbon intensity of electricity supply (kg CO 2/MWh) Maryland 20,192 339 448 161 71 681 Massachusetts 19,428 214 465 105 184 648 Michigan 17,297 347 481 211 23 666 Minnesota 18,534 318 505 223 14 780 Mississippi 11,531 398 414 95 22 631 Missouri 15,311 334 454 203 22 899 Montana 13,475 312 410 122 20 765 Nebraska 15,722 302 468 212 14 780 Nevada 17,276 324 433 131 9 702 New Hampshire 19,423 214 465 105 184 387 New Jersey 20,330 339 449 162 71 474 New Mexico 12,994 297 377 119 9 935 New York 16,298 212 391 114 112 442 North Carolina 15,512 435 481 105 24 618 North Dakota 14,126 293 444 204 13 1,134 Ohio 16,360 341 469 208 23 852 Oklahoma 13,407 288 432 201 13 790 Oregon 16,395 331 451 130 21 227 Pennsylvania 15,950 316 403 150 66 613 Rhode Island 16,417 203 431 99 175 550 South Carolina 14,305 425 463 102 24 442 South Dakota 13,845 291 440 203 13 631 Tennessee 14,463 426 465 103 24 645 Texas 14,492 388 537 78 8 729 Utah 14,907 322 431 126 20 1,028 Vermont 16,560 204 432 100 176 73 Virginia 18,413 458 521 111 25 645 Washington 18,049 341 472 134 21 160 West Virginia 12,219 312 405 188 21 948 Wisconsin 17,355 347 482 212 23 840 Wyoming 15,237 324 436 127 20 1,099 U.S. average 16,160 312 448 119 38 667 Note: For data sources, see Appendix.

CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 9 (MWh). Vermont, where the main power sources are nuclear and hydro, emits only 73 kg CO 2/MWh. Taking these differences into account, Figure 3 depicts the impact of higher fossil fuel prices on the median-income household in each state, with a carbon price of $25/tCO 2. The results show that inter-state differences are not terribly large, ranging from $239 in Oregon to $349 in Indiana. The map in Figure 4 (page 11) depicts these impacts on a median-income family of four. Table 3 shows the impact on consumers by income decile across the states, with the results expressed as a percentage of income. The dollar $0 $50 $100 $150 $200 $250 $300 $350 $400 Indiana Delaware D.C Wisconsin Minnesota Virginia Ohio Colorado Maryland North Dakota Kansas Iowa Wyoming Missouri Michigan Georgi a Kentucky Utah Massach u setts Nebraska Illinois New Jersey Connecticut Haw aii North Carolina Texas Alaska West Virginia Tennessee Nevada New Hampshire Oklahoma Al ab ama Pennsylvania Louisiana Rhode Island South Dakota New Mexico Arkansas Montana Florida South Carolina Maine Arizona Mississi p pi California New York Idaho Washington Vermont Oregon Indirect Costs Heating Oil Natural Gas Electricity Gasoli ne ` FIGURE 3 : PER CAPITA CARBON EXPENDITURE OF MEDIAN HOUSEHOLD BY COMMODITY GROUP (PRICED AT $25 /TCO 2) TABLE 3 : CARBON PRICE IMPACT BY STATE AND INCOME DECILE (PERCENTAGE OF MEDIAN INCOME ) Decile State Median 1 2 3 4 5 6 7 8 9 10 Alabama 2.1% 4.5% 3.4% 2.9% 2.5% 2.2% 2.0% 1.8% 1.6% 1.3% 1.0% Alaska 1.6% 2.8% 2.3% 2.0% 1.8% 1.6% 1.5% 1.4% 1.2% 1.1% 0.9% Arizona 1.7% 3.2% 2.5% 2.2% 1.9% 1.7% 1.6% 1.4% 1.3% 1.1% 0.9% Arkansas 2.1% 4.3% 3.3% 2.8% 2.5% 2.2% 2.0% 1.8% 1.6% 1.4% 1.1% California 1.5% 2.9% 2.3% 2.0% 1.8% 1.6% 1.4% 1.3% 1.2% 1.0% 0.8% Colorado 1.7% 3.4% 2.6% 2.3% 2.0% 1.8% 1.6% 1.5% 1.3% 1.2% 0.9% Connecticut 1.4% 2.9% 2.2% 1.9% 1.7% 1.5% 1.4% 1.2% 1.1% 0.9% 0.8% Delaware 1.8% 3.6% 2.8% 2.4% 2.1% 1.9% 1.7% 1.6% 1.4% 1.2% 1.0% D.C 1.9% 4.6% 3.3% 2.7% 2.3% 2.0% 1.8% 1.5% 1.3% 1.1% 0.8% Florida 1.7% 3.4% 2.6% 2.2% 2.0% 1.8% 1.6% 1.4% 1.3% 1.1% 0.9% Georgia 2.0% 4.1% 3.1% 2.7% 2.3% 2.1% 1.9% 1.7% 1.5% 1.3% 1.0% CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 10 Decile State Median 1 2 3 4 5 6 7 8 9 10 Hawaii 1.8% 3.5% 2.7% 2.3% 2.1% 1.9% 1.7% 1.5% 1.4% 1.2% 1.0% Idaho 1.7% 3.2% 2.5% 2.2% 2.0% 1.8% 1.6% 1.5% 1.3% 1.2% 1.0% Illinois 1.7% 3.5% 2.7% 2.3% 2.0% 1.8% 1.7% 1.5% 1.3% 1.1% 0.9% Indiana 2.1% 4.2% 3.3% 2.8% 2.5% 2.2% 2.0% 1.8% 1.6% 1.4% 1.1% Iowa 2.0% 3.9% 3.1% 2.7% 2.4% 2.1% 1.9% 1.7% 1.6% 1.4% 1.1% Kansas 2.0% 4.0% 3.1% 2.7% 2.4% 2.1% 1.9% 1.7% 1.5% 1.3% 1.0% Kentucky 2.3% 5.0% 3.8% 3.2% 2.8% 2.5% 2.2% 2.0% 1.7% 1.5% 1.1% Louisiana 2.3% 5.0% 3.8% 3.2% 2.8% 2.4% 2.2% 1.9% 1.7% 1.4% 1.1% Maine 1.7% 3.2% 2.5% 2.2% 1.9% 1.8% 1.6% 1.5% 1.3% 1.1% 0.9% Maryland 1.6% 3.1% 2.4% 2.1% 1.9% 1.7% 1.5% 1.4% 1.2% 1.1% 0.9% Massachusetts 1.6% 3.2% 2.4% 2.1% 1.9% 1.7% 1.5% 1.3% 1.2% 1.0% 0.8% Michigan 1.8% 3.6% 2.8% 2.4% 2.1% 1.9% 1.7% 1.6% 1.4% 1.2% 1.0% Minnesota 1.8% 3.4% 2.7% 2.3% 2.1% 1.9% 1.7% 1.5% 1.4% 1.2% 1.0% Mississippi 2.2% 4.7% 3.6% 3.0% 2.7% 2.4% 2.1% 1.9% 1.7% 1.4% 1.1% Missouri 2.1% 4.3% 3.3% 2.8% 2.5% 2.2% 2.0% 1.8% 1.6% 1.4% 1.1% Montana 2.0% 3.9% 3.0% 2.6% 2.3% 2.1% 1.9% 1.7% 1.5% 1.3% 1.1% Nebraska 1.9% 3.7% 2.9% 2.5% 2.3% 2.0% 1.9% 1.7% 1.5% 1.3% 1.0% Nevada 1.7% 3.2% 2.5% 2.2% 1.9% 1.8% 1.6% 1.4% 1.3% 1.1% 0.9% New Hampshire 1.5% 2.7% 2.2% 1.9% 1.7% 1.5% 1.4% 1.3% 1.2% 1.0% 0.9% New Jersey 1.5% 2.9% 2.3% 2.0% 1.7% 1.6% 1.4% 1.3% 1.1% 1.0% 0.8% New Mexico 2.1% 4.3% 3.3% 2.8% 2.5% 2.2% 2.0% 1.8% 1.6% 1.3% 1.1% New York 1.5% 3.2% 2.4% 2.1% 1.8% 1.6% 1.5% 1.3% 1.2% 1.0% 0.8% North Carolina 1.9% 3.9% 3.0% 2.6% 2.3% 2.0% 1.8% 1.6% 1.5% 1.3% 1.0% North Dakota 2.3% 4.6% 3.5% 3.0% 2.7% 2.4% 2.2% 2.0% 1.7% 1.5% 1.2% Ohio 2.0% 4.0% 3.1% 2.7% 2.4% 2.1% 1.9% 1.7% 1.5% 1.3% 1.0% Oklahoma 2.1% 4.3% 3.3% 2.8% 2.5% 2.2% 2.0% 1.8% 1.6% 1.4% 1.1% Oregon 1.5% 2.6% 2.1% 1.9% 1.7% 1.5% 1.4% 1.3% 1.1% 1.0% 0.8% Pennsylvania 1.8% 3.5% 2.7% 2.3% 2.1% 1.9% 1.7% 1.5% 1.3% 1.2% 0.9% Rhode Island 1.7% 3.3% 2.6% 2.2% 2.0% 1.8% 1.6% 1.4% 1.3% 1.1% 0.9% South Carolina 1.8% 3.6% 2.8% 2.4% 2.1% 1.9% 1.7% 1.6% 1.4% 1.2% 1.0% South Dakota 2.0% 3.8% 3.0% 2.6% 2.3% 2.1% 1.9% 1.7% 1.5% 1.3% 1.0% Tennessee 2.0% 4.2% 3.2% 2.7% 2.4% 2.1% 1.9% 1.7% 1.5% 1.3% 1.0% Texas 2.1% 4.2% 3.2% 2.8% 2.4% 2.2% 1.9% 1.7% 1.5% 1.3% 1.0% Utah 2.1% 4.0% 3.1% 2.7% 2.4% 2.2% 2.0% 1.8% 1.6% 1.4% 1.1% Vermont 1.5% 2.6% 2.1% 1.9% 1.7% 1.5% 1.4% 1.3% 1.2% 1.0% 0.8% Virginia 1.8% 3.6% 2.8% 2.4% 2.1% 1.9% 1.7% 1.5% 1.4% 1.2% 0.9% Washington 1.4% 2.4% 1.9% 1.7% 1.6% 1.4% 1.3% 1.2% 1.1% 1.0% 0.8% West Virginia 2.4% 5.1% 3.9% 3.3% 2.9% 2.6% 2.3% 2.0% 1.8% 1.5% 1.2% Wisconsin 1.9% 3.7% 2.9% 2.5% 2.3% 2.0% 1.8% 1.7% 1.5% 1.3% 1.0% Wyoming 2.1% 4.2% 3.3% 2.8% 2.5% 2.2% 2.0% 1.8% 1.6% 1.4% 1.1% TABLE 3 : CARBON PRICE IMPACT BY STATE AND INCOME DECILE (PERCENTAGE OF MEDIAN INCOME ), CONTINUED CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 11 amounts from which the percentages are de- rived are reported in Appendix Tables A.1 and A.2. The impact on the median household is shown in the first column. The biggest impact is in West Virginia, where the costs from higher fossil fuel prices are equivalent to 2.4% of me- dian income. This is mainly due to the state’s relatively low incomes: West Virginia’s median carbon charge is only 4% above the national median (see Appendix Table A.2), but its median income is almost 25% below the national level (see Appendix Table A.1). The smallest impact is felt in Connecticut, the state with the highest median income, despite the fact that the me- dian Connecticut resident pays a little more in dollar terms than the median West Virginian.

The regressive impact of carbon pricing is evi- dent in these inter-state comparisons. Within states, the regressive impact of higher fuel prices is even clearer. In every state, the biggest impact as a percentage of income is in the low- est-income decile, and the least impact is in the highest-income decile. The carbon charge as a fraction of income steadily declines from the bottom to the top of the income profile. Impact of recycling revenue as dividends The net impact of cap-and-dividend differs markedly from the impact of higher fossil fuel prices alone. The dividends (here assumed to be 80% of carbon revenues) have a strong pro- gressive impact on family incomes, as they rep- resent a larger fraction of income for the low- income households than for high-income house- holds. This outweighs the regressive impact of higher fossil fuel prices. Table 4 shows the net dollar impact by state and income decile. In every state, the median household (shown in the first column) sees a positive net impact: the amount it receives as dividends exceeds what it pays as a result of higher fossil fuel prices. Fig- ure 5 (page 13) depicts these effects for a family of four at the median income level in each state.

The largest positive effects, as can be seen in Table 4, are consistently in the lowest-income IN EVERY STATE, THE BOTTOM SIX DECILES EXPERIENCE POSITIVE NET BENEFITS. FIGURE 4 : IMPACT OF CARBON PRICING ON MEDIAN FAMILY OF FOUR ($/YEAR , AT $25 /TON CO 2) CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 12 Decile medians State Median 1 2 3 4 5 6 7 8 9 10 Alabama 103 250 208 177 148 119 87 50 2 -67 -222 Alaska 91 231 190 160 133 106 76 43 0 -62 -195 Arizona 127 262 223 195 168 142 112 78 35 -30 -175 Arkansas 114 252 212 183 156 129 99 65 21 -42 -181 California 135 278 237 207 179 150 118 81 32 -40 -207 Colorado 61 220 173 139 108 77 43 4 -47 -121 -287 Connecticut 84 249 203 168 135 102 66 22 -34 -120 -321 Delaware 48 207 160 126 95 64 30 -9 -58 -131 -289 D.C 49 243 190 150 111 71 26 -27 -98 -208 -475 Florida 118 260 220 190 162 133 101 64 16 -56 -221 Georgia 71 230 184 150 119 88 54 14 -36 -111 -275 Hawaii 85 231 188 157 129 100 68 32 -15 -83 -234 Idaho 141 265 228 202 178 154 127 97 59 2 -119 Illinois 81 238 193 159 128 97 63 24 -26 -100 -266 Indiana 37 198 150 115 84 53 20 -18 -66 -135 -283 Iowa 63 214 169 136 107 78 47 11 -33 -98 -235 Kansas 62 218 172 139 108 78 45 8 -40 -109 -258 Kentucky 72 229 184 150 120 89 55 17 -32 -104 -261 Louisiana 105 251 209 178 150 121 89 52 5 -65 -218 Maine 127 258 220 192 167 141 113 81 40 -20 -152 Maryland 61 220 174 140 109 77 43 4 -47 -122 -288 Massachusetts 80 240 194 160 129 97 62 21 -32 -112 -293 Michigan 70 227 181 147 117 86 53 15 -33 -104 -259 Minnesota 54 214 166 132 101 70 37 -2 -51 -122 -276 Mississippi 128 263 224 196 169 142 112 78 35 -30 -171 Missouri 63 221 175 141 111 80 46 8 -41 -112 -266 Montana 116 249 210 182 156 130 102 70 29 -31 -160 Nebraska 80 227 183 152 123 95 64 29 -15 -78 -215 Nevada 97 241 199 169 140 112 81 45 -1 -68 -217 New Hampshire 98 239 198 168 140 113 83 48 5 -60 -200 New Jersey 83 242 197 163 132 100 65 24 -29 -110 -293 New Mexico 114 251 211 182 155 128 99 64 21 -44 -186 New York 135 278 238 208 180 150 118 79 29 -49 -231 North Carolina 88 238 195 163 133 104 72 34 -13 -83 -235 North Dakota 62 213 168 136 106 77 46 10 -35 -100 -237 Ohio 58 217 170 136 105 74 41 2 -46 -118 -272 Oklahoma 103 246 204 174 146 117 87 52 7 -58 -200 Oregon 147 275 238 211 186 160 133 100 59 -2 -138 TABLE 4 : NET IMPACT OF CAP -AND -DIVIDEND BY STATE AND INCOME DECILE ($ PER CAPITA ) CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 13 decile. In every state, the bottom six deciles ex- perience positive net benefits; in 45 states, the bottom seven deciles do so.

Table 5 shows these net impacts as a percent- age of income. For the median household (first column), the range is from a net benefit of 0.2% of income in Indiana to 1.1% in Mississippi. In the lowest-income decile, where net benefits are greatest, the range is from 4.1% in Maryland to 10.1% in Mississippi. In the top decile, the net cost ranges from 0.2 to 0.5% of income.

Scanning the variations in Table 5 horizontally across columns (by income decile) and vertically across rows (by state), it is clear that the former exceed the latter by a wide margin. Inter-state differences are modest relative to differences across income groups. 10 Decile medians State Median 1 2 3 4 5 6 7 8 9 10 TABLE 4 : NET IMPACT OF CAP -AND -DIVIDEND BY STATE AND INCOME DECILE ($ PER CAPITA ), CONTINUED Pennsylvania 105 248 207 176 148 120 89 53 7 -61 -212 Rhode Island 112 255 214 184 156 127 96 60 14 -55 -210 South Carolina 125 261 222 193 166 139 109 75 32 -32 -172 South Dakota 113 249 209 180 154 127 99 66 25 -35 -164 Tennessee 95 244 201 169 140 110 78 41 -7 -77 -233 Texas 88 243 199 166 135 104 71 32 -18 -91 -252 Utah 74 216 173 143 116 89 59 26 -16 -76 -204 Vermont 143 268 231 205 180 156 129 98 59 1 -125 Virginia 57 221 174 139 107 74 39 -2 -54 -131 -302 Washington 141 273 235 207 181 155 126 93 50 -14 -156 West Virginia 92 240 197 165 136 107 76 39 -7 -74 -221 Wisconsin 50 205 158 125 95 65 33 -3 -49 -116 -257 Wyoming 63 214 169 137 108 79 47 11 -35 -101 -243 U.S. average 103 251 209 177 148 119 87 49 1 -71 -232 FIGURE 5 : CAP -AND -DIVIDEND : NET BENEFIT FOR MEDIAN FAMILY OF FOUR ($/YEAR , AT $25 /TON CO 2) CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 14 TABLE 5 : NET IMPACT OF CAP -AND -DIVIDEND BY STATE AND INCOME DECILE (PERCENTAGE OF MEDIAN INCOME ) Decile medians State Median 1 2 3 4 5 6 7 8 9 10 Alabama 0.8% 8.2% 4.0% 2.4% 1.6% 1.0% 0.6% 0.3% 0.0% -0.2% -0.4% Alaska 0.5% 4.2% 2.2% 1.4% 0.9% 0.6% 0.4% 0.2% 0.0% -0.2% -0.3% Arizona 0.8% 6.8% 3.5% 2.2% 1.5% 1.0% 0.7% 0.4% 0.1% -0.1% -0.3% Arkansas 0.9% 8.2% 4.1% 2.6% 1.7% 1.1% 0.7% 0.4% 0.1% -0.1% -0.3% California 0.8% 7.3% 3.6% 2.3% 1.5% 1.0% 0.6% 0.3% 0.1% -0.1% -0.3% Colorado 0.3% 4.5% 2.2% 1.3% 0.8% 0.5% 0.2% 0.0% -0.1% -0.3% -0.4% Connecticut 0.4% 5.3% 2.5% 1.5% 0.9% 0.5% 0.3% 0.1% -0.1% -0.2% -0.3% Delaware 0.3% 4.2% 2.0% 1.2% 0.7% 0.4% 0.1% 0.0% -0.2% -0.3% -0.4% D.C 0.3% 7.9% 3.2% 1.7% 0.9% 0.5% 0.1% -0.1% -0.3% -0.4% -0.5% Florida 0.7% 7.0% 3.5% 2.2% 1.4% 0.9% 0.6% 0.3% 0.1% -0.1% -0.3% Georgia 0.4% 6.0% 2.9% 1.7% 1.0% 0.6% 0.3% 0.1% -0.1% -0.3% -0.4% Hawaii 0.5% 5.2% 2.6% 1.6% 1.0% 0.7% 0.4% 0.1% 0.0% -0.2% -0.4% Idaho 1.0% 6.9% 3.7% 2.4% 1.7% 1.2% 0.8% 0.5% 0.2% 0.0% -0.2% Illinois 0.5% 5.6% 2.7% 1.6% 1.0% 0.6% 0.3% 0.1% -0.1% -0.2% -0.4% Indiana 0.2% 4.4% 2.1% 1.2% 0.7% 0.4% 0.1% -0.1% -0.2% -0.4% -0.5% Iowa 0.4% 4.8% 2.4% 1.4% 0.9% 0.5% 0.3% 0.1% -0.1% -0.3% -0.4% Kansas 0.4% 5.2% 2.5% 1.5% 0.9% 0.5% 0.3% 0.0% -0.1% -0.3% -0.4% Kentucky 0.5% 7.3% 3.4% 2.0% 1.3% 0.7% 0.4% 0.1% -0.1% -0.3% -0.5% Louisiana 0.9% 9.3% 4.4% 2.7% 1.7% 1.1% 0.7% 0.3% 0.0% -0.2% -0.4% Maine 0.8% 6.4% 3.3% 2.2% 1.5% 1.0% 0.7% 0.4% 0.1% -0.1% -0.3% Maryland 0.3% 4.1% 2.0% 1.2% 0.7% 0.4% 0.2% 0.0% -0.1% -0.3% -0.4% Massachusetts 0.4% 5.2% 2.5% 1.5% 0.9% 0.6% 0.3% 0.1% -0.1% -0.2% -0.4% Michigan 0.4% 5.1% 2.5% 1.5% 0.9% 0.6% 0.3% 0.1% -0.1% -0.3% -0.4% Minnesota 0.3% 4.3% 2.0% 1.2% 0.7% 0.4% 0.2% 0.0% -0.2% -0.3% -0.4% Mississippi 1.1% 10.1% 5.0% 3.1% 2.1% 1.4% 0.9% 0.5% 0.2% -0.1% -0.3% Missouri 0.4% 5.8% 2.7% 1.6% 1.0% 0.6% 0.3% 0.0% -0.2% -0.3% -0.4% Montana 0.9% 7.1% 3.6% 2.3% 1.6% 1.1% 0.7% 0.4% 0.1% -0.1% -0.3% Nebraska 0.5% 5.3% 2.6% 1.6% 1.1% 0.7% 0.4% 0.1% -0.1% -0.2% -0.4% Nevada 0.6% 5.3% 2.7% 1.7% 1.1% 0.7% 0.4% 0.2% 0.0% -0.2% -0.3% New Hampshire 0.5% 4.4% 2.3% 1.5% 1.0% 0.6% 0.4% 0.2% 0.0% -0.1% -0.3% New Jersey 0.4% 5.0% 2.4% 1.4% 0.9% 0.5% 0.3% 0.1% -0.1% -0.2% -0.3% New Mexico 0.9% 8.0% 4.0% 2.5% 1.7% 1.1% 0.7% 0.4% 0.1% -0.1% -0.3% New York 0.8% 8.2% 3.9% 2.4% 1.6% 1.0% 0.6% 0.3% 0.1% -0.1% -0.3% North Carolina 0.6% 6.2% 3.0% 1.9% 1.2% 0.7% 0.4% 0.2% 0.0% -0.2% -0.4% North Dakota 0.4% 5.6% 2.7% 1.6% 1.0% 0.6% 0.3% 0.1% -0.1% -0.3% -0.4% Ohio 0.4% 5.2% 2.5% 1.5% 0.9% 0.5% 0.2% 0.0% -0.2% -0.3% -0.4% Oklahoma 0.8% 7.5% 3.7% 2.3% 1.5% 1.0% 0.6% 0.3% 0.0% -0.2% -0.4% Oregon 0.9% 6.5% 3.4% 2.2% 1.6% 1.1% 0.7% 0.4% 0.2% 0.0% -0.2% CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 15 To put these inter-state differences in perspec- tive, we can compare the impact of the cap-and- dividend policy to that of two major items in the federal budget: defense spending and farm pro- grams. Figure 6 depicts per capita spending un- der these two programs in the top ten states and the bottom ten states, and compares this to the net impact of the cap-and-dividend policy on median households in the top ten and bottom ten states. 11 In the case of defense spending, the ratio between the top ten and bottom ten states is more than 11:1. In the case of farm programs, it is 190:1. In the case of cap-and- dividend, it is 2½ :1.

V. NON-DIVIDEND USES OF CARBON REVENUES A climate policy that incorporates cap-and- dividend is likely to dedicate some fraction of carbon revenues (20% in the preceding analy- sis) to other uses, while returning the rest to the people as equal dividends. If all the carbon permits are auctioned, then these non-dividend uses are funded by a fraction of the revenue.

Alternatively (as in the Waxman-Markey bill), some fraction of the permits may be given away instead of being auctioned; this has an equiva- lent effect, transferring “allowance value” rather than cash to the recipients. 12 In this section, we briefly discuss potential non- dividend uses of carbon revenues or allowance value. First, we discuss transitional adjustment assistance to help workers, communities and firms that stand to be affected adversely by the economy’s shift away from fossil fuels. Second, TABLE 5 : NET IMPACT OF CAP -AND -DIVIDEND BY STATE AND INCOME DECILE (PERCENTAGE OF MEDIAN INCOME ), CONTINUED g Pennsylvania 0.7% 6.3% 3.1% 2.0% 1.3% 0.8% 0.5% 0.2% 0.0% -0.2% -0.3% Rhode Island 0.7% 6.4% 3.2% 2.0% 1.3% 0.9% 0.5% 0.3% 0.0% -0.1% -0.3% South Carolina 0.9% 7.4% 3.8% 2.4% 1.6% 1.1% 0.7% 0.4% 0.1% -0.1% -0.3% South Dakota 0.8% 6.8% 3.5% 2.2% 1.5% 1.0% 0.6% 0.3% 0.1% -0.1% -0.3% Tennessee 0.7% 7.1% 3.4% 2.1% 1.4% 0.9% 0.5% 0.2% 0.0% -0.2% -0.4% Texas 0.6% 7.2% 3.4% 2.1% 1.3% 0.8% 0.4% 0.2% -0.1% -0.2% -0.4% Utah 0.5% 5.1% 2.6% 1.6% 1.0% 0.7% 0.4% 0.1% -0.1% -0.2% -0.4% Vermont 0.9% 5.9% 3.2% 2.1% 1.5% 1.0% 0.7% 0.4% 0.2% 0.0% -0.2% Virginia 0.3% 4.8% 2.3% 1.3% 0.8% 0.4% 0.2% 0.0% -0.2% -0.3% -0.4% Washington 0.8% 5.8% 3.0% 2.0% 1.4% 1.0% 0.6% 0.4% 0.2% 0.0% -0.2% West Virginia 0.8% 8.4% 4.0% 2.5% 1.6% 1.0% 0.6% 0.2% 0.0% -0.2% -0.4% Wisconsin 0.3% 4.2% 2.0% 1.2% 0.7% 0.4% 0.2% 0.0% -0.2% -0.3% -0.4% Wyoming 0.4% 5.2% 2.5% 1.5% 1.0% 0.6% 0.3% 0.1% -0.1% -0.3% -0.4% U.S. average 0.6% 6.5% 3.2% 2.0% 1.3% 0.8% 0.5% 0.2% 0.0% -0.2% -0.3% Decile medians State Median 1 2 3 4 5 6 7 8 9 10 FIGURE 6: TOP TEN AND BOTTOM TEN STATES : DEFENSE EXPENDITURE , FARM PROGRAMS , AND CAP -AND -DIVIDEND POLICY (DOLLARS PER CAPITA ) $0 $500 $1, 000 $1, 500 $2, 000 $2, 500 Defense expenditure Farm programs Net impacts of cap-and-dividend Me a n o f to p te n sta te s Me a n o f bo tto m te n sta te s CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 16 we compare the distributional impact of cap- and-dividend to that of H.R. 2454, the American Clean Energy and Security Act passed by the U.S. House of Representatives in June 2009, which proposes a variety of non-dividend uses.

Transitional adjustment assistance In addition to its impacts on consumers, a policy to curb carbon emissions will have impacts on businesses and workers. 13 In some sectors – coal mining is an important example – jobs will be lost. In others – for example, building retro- fits and the manufacture of clean energy tech- nologies – new jobs will be created.

Since production of renewables and energy effi- ciency are generally more labor-intensive than production of fossil fuels, job gains are likely to exceed job losses. 14 No automatic economic mechanism ensures, however, that job creation will occur in the same communities and for the same workers who are hit by job losses. To assist their transition to new livelihoods, a fraction of the carbon revenues initially could be allocated to the states as block grants for ad- justment assistance. In the first year of the policy, for example, 10% of permit auction revenues might be dedicated to this purpose, with the per- centage gradually phased out over time. Disbursement of transitional adjustment assis- tance funds in the form of block grants would allow the states to tailor policies to their own circumstances and priorities. In coal-mining states, for example, funds could be invested in the ecological restoration of landscapes that have been severely degraded by mountaintop removal, strip mining, and disposal of mine tail- ings and coal ash. In manufacturing-intensive states, funds could be invested in job training and support to “green” industries.

The American Clean Energy and Security Act of 2009 The American Clean Energy and Security Act (ACES) proposes to give away 85% of carbon permits in the initial years of the policy and to KEEPING GOVERNMENTS WHOLE Not only households will be impacted by the higher fossil fuel prices that result from a carbon cap. Government expenditure accounts for about 14% of U.S. carbon emissions. Of this to- tal, 3.6% comes from federal spending and 10.8% from state and local government spend- ing. To keep government whole – to avoid cuts in real government purchasing power – a compa- rable share of carbon revenues will need to flow to government coffers.

If the dividends paid to the public from carbon revenue are non-taxable, then policymakers will need to allocate a portion of the remaining car- bon revenue to this purpose. If they are taxable, we estimate that roughly 24 cents on the divi- dend dollar will flow back to government in the form of federal and state taxes (Boyce and Riddle 2008). With 80% of the total revenue dis- tributed as dividends, this means that taxes would recycle 19% of total carbon revenue to government, enough to offset fully the impact of higher fossil fuel prices on government purchas- ing power, with about 5% of total carbon reve- nues left over for other purposes. Taxable dividends are preferable to lower, non- taxable dividends from the standpoint of distri- butional equity. Taxation claims a bigger share of the dividends in upper-income brackets than it does from lower-income and middle-income households. Directly tapping the carbon revenue to obtain the same amount of money, by con- trast, reduces dividend payments equally to all, a result equivalent to a head tax, one of the most regressive forms of taxation. Whatever approach is used to keep government whole, some formula will be necessary to allo- cate carbon revenues amongst state and local governments. One way to do this, which would be consistent with the principles of cap-and- dividend, is to divide revenue among state and local governments in proportion to their popula- tions, with equal per capita amounts to each ju- risdiction. As in the case of dividends paid to individuals, this distribution would protect the governments’ purchasing power while giving them incentives to invest in energy efficiency and clean energy.

CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 17 auction the remaining 15%. It earmarks the al- lowance value (free permits and revenues) for a number of different uses. These include free permit allocations to electricity local distribution companies (LDCs), with the expectation that the allowance value will be “passed through” to in- dustrial, commercial and residential electricity customers; direct payments from auction reve- nues to low-income households; and allocations to oil refineries and to energy-intensive “trade vulnerable” industries.

The stated rationale for free allocations to LDCs is that this will protect consumers from the im- pact of higher electricity prices. Insofar as the value of allowances is passed through to con- sumers, rather than being captured by LDCs as higher profits, this is likely to mask the price signal to economize on electricity use. 15 If so, the burden of adjustment imposed by the car- bon cap will fall more heavily on other sectors of the economy, including transportation fuels, pushing up prices in those sectors even more and raising costs to consumers overall. 16 Starting in the 2020s, an increasing share of the permits would be auctioned and the reve- nues deposited in a “Climate Change Consumer Refund Account” for return to the public on an equal per capita basis. In this sense, ACES can be described as a cap-and-dividend policy with a very slow fuse.

A June 2009 analysis of the distributional im- pacts of the cap-and-trade provisions of ACES by the Congressional Budget Office (CBO) con- cludes that 79% of the allowance value would eventually find its way back to American house- holds. However, it would not flow to all house- holds in equal measure. For example, the CBO reckons (page 12) that “about 63 percent of the allowance value conveyed to businesses would ultimately flow to households in the highest in- come quintile,” as a result of higher profits paid out in proportion to corporate stock holdings. Combining the routes (in some cases rather cir- cuitous ones) by which auction revenues and the allowance value of free permits ultimately return to households, the CBO estimates that in the year 2020 nearly two-fifths of the total (37.5%) would go to the top quintile of house- holds in the nation’s income distribution. The middle quintile would receive the smallest share (14.6%), with the remaining quintiles getting 15.4-16.9% each. 17 In Figure 7, this outcome is contrasted with cap-and-dividend, in which each quintile re- ceives an amount equal to its share of the popu- lation: 20%.

Visibility of costs and benefits Leaving aside their distributional effects, a drawback of non-dividend uses of carbon reve- nues (and free permit allocations) is that their impact on households is less transparent than the cash-in-hand provided by dividends. From the standpoint of public support for the climate policy over the 40-year energy transition, what matters is not only the difference between costs from higher fuel prices and benefits from permit and revenue allocations, but also the visibility of these costs and benefits.

On the cost side of the scales, visibility is high indeed. Gasoline prices, for example, are per- haps the single most widely known price in FIGURE 7: DISTRIBUTION OF CARBON REVENUES TO HOUSEHOLDS : ACES V . CAP -AND -DIVIDEND (PERCENTAGE SHARE ) 15% 17% 15% 16% 38% 20% 20% 20% 20% 20% 0% 10% 20% 30% 40% Bottom 20% Next 20% Middle 20% Next 20% Top 20% ACES Cap-and-dividend CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 18 America: 165,000 filling stations across the country announce them in foot-high numbers.

Most consumers also are fairly well aware of the size of the numbers on the monthly checks they write to their electricity companies.

On the benefit side, visibility varies greatly amongst policy options. Most of the avenues by which ACES would transfer money to house- holds score low on visibility. Tax credits (al- though less visible than cash) to low-income households are perhaps the most readily visible avenue. Rebates from electricity local distribution companies (LDCs) may be gleaned from the fine print on monthly utility bills. Pay- backs via higher returns to stock ownership (in- cluding stocks held in pension plans) will be difficult, if not impossible, to distinguish from the many other economic factors that affect in- vestment returns.

Apart from its simplicity and fairness, an attrac- tion of cap-and-dividend is that the return of carbon revenue to the American people is highly visible: it comes back as cash in their wallets.

Cap-and-dividend clearly sends the carbon price signal, while at the same time maximizing public awareness that families can come out ahead no matter how high carbon prices rise. The policy’s underlying premise – that we are all equal co- owners of our nation’s share of the carbon stor- age capacity of the atmosphere – is likely to have wider public appeal than the premise that the air belongs to polluting corporations. The transition to a post-carbon economy cannot happen overnight. It will require decades of sus- tained policy, including steadily rising carbon prices, to drive it forward. Durable public back- ing for rising carbon prices is therefore essen- tial. The fact that dividends are highly visible, together with the fact that a majority of Ameri- can families come out ahead no matter what the carbon price, can provide the political foun- dation for long-term support for the policy.

This public support will make it possible to tighten the carbon cap and further raise fossil fuel prices to higher levels, bringing billions of dollars in private investment in clean energy and energy efficiency. In this sense, returning carbon revenue directly to the public not only protects family incomes but also is a highly lev- eraged use of carbon revenue.

VI. CONCLUSIONS Cap-and-dividend is a policy to manage a scarce resource: our planet’s carbon-absorptive capac- ity. A consequence of any policy to limit use of a resource – to manage scarcity – is the creation of property rights. A cap-and-permit system will raise the prices of fossil fuels and all other goods and services that use these fuels in their production and distribution. Each consumer’s carbon footprint will now come with a price. The money that is paid by consumers does not dis- appear from the nation’s economy: it is trans- ferred to owners of the newly created property.

The premise of cap-and-dividend is that this property should belong equally and in common to all. By auctioning permits – rather than giving them free-of-charge to corporations or other po- litically favored entities – and by returning most of the auction revenue to the public, cap-and- dividend combines price incentives to reduce carbon emissions with protection for consumers from the impact of higher fuel prices on their real incomes. The majority of families come out ahead, receiving dividends that more than off- set the price increases. In this paper we have shown that this positive outcome holds not only at the national level but also within each of the 50 states.

The differences across states in the household impacts of cap-and-dividend are small com- pared to differences across income brackets, and also compared to inter-state differences in CAP-AND-DIVIDEND CLEARLY SENDS THE CARBON PRICE SIGNAL, WHILE AT THE SAME TIME MAXIMIZING PUBLIC AWARENESS THAT FAMILIES CAN COME OUT AHEAD NO MATTER HOW HIGH CARBON PRICES RISE. CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 19 defense spending and federal farm programs.

Because dividends are distributed equally, variations in the net impact of cap-and-dividend arise solely from differences in carbon foot- prints. Households who consume more fossil fuels (and more of the things made and distrib- uted using them) will pay more; those who con- sume less will pay less. Residents of states that have moved more aggressively to promote en- ergy efficiency, such as California, will do better than average. But the differences in carbon footprints between the top 10% and bottom 10% of the income distribution are far greater than the differences between median house- holds across the states.

Whereas cap-and-dividend returns carbon reve- nue equally to each person, the American Clean Energy and Security (ACES) Act would allocate revenues and free permits in a variety of ways that would impact different households differ- ently. The Congressional Budget Office (2009) estimates that roughly two-fifths of the resulting income would flow to households in the top 20% of the nation’s income distribution – an outcome that would disproportionately benefit upper-income states as well as upper- income individuals.

Perhaps even more politically salient than net distributional effects is the visibility of transfers of carbon revenue (or allowance value) to the public. Dividends in the form of checks in the mail or deposits into bank accounts will provide highly tangible benefits to consumers, against which they can weigh the impacts of higher prices. The transfers in the ACES policy mix, such as rebates in electricity bills and capital gains for corporate shareholders, would be less apparent. For reasons of both economic fairness and transparency, therefore, cap-and-dividend offers a way to secure durable public support for an effective policy to wean the economy from de- pendence on fossil fuels. A proactive U.S. policy, in turn, will be a crucial condition for an effec- tive international agreement to confront the global challenge of climate change. CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 20 METHODOLOGICAL APPENDIX Our state-level estimates of the distributional incidence of higher fossil fuel prices on households are based on a car- bon charge (i.e., permit price) of $25/ton CO 2 ($92/ton C). 18 We include both direct effects via household energy con- sumption (i.e., via increases in the prices of heating oil, gasoline, natural gas, and electricity) and indirect effects via impacts on the prices of other goods and services (e.g., food and manufactured goods) that use fossil fuels in their pro- duction and distribution. 19 Following the usual practice, we assume that 100% of the permit cost is passed through to the final consumer. If coal is mined in West Virginia, and used to produce steel in Ohio, that is used to manufacture an automobile in Michigan, that is sold to a consumer in Connecticut, it is the Connecticut consumer who pays the associated carbon charge.

To estimate impacts at the state level, we adjust national- level estimates to account for three variables: 1. interstate differences in income; 2. interstate differences in the carbon intensity of electricity consumed by households; and 3. regional differences in consumption patterns, arising from differences in energy use for heating and cooling, driv- ing behavior, etc.

The first adjustment – for interstate differences in income – is based on data from the 2000 U.S. Census that allow us to measure income inequality within states and to construct state-specific per capita income deciles. For comparability with our expenditure data, we convert these to 2003 figures by adjusting for nominal income growth.

The second adjustment – for interstate differences in the carbon intensity of electricity consumption – is based on the carbon intensity of electricity generated in each state, with adjustments to account for imports of electricity across state lines within interconnected power grids.

The third adjustment - for regional differences in consump- tion patterns – is based on the region-specific Consumer Expenditure Survey (CEX) measures reported by Burtraw et al. (2009) for household consumption of electricity, gaso- line, natural gas and heating oil for 11 regions (4 of which are single states: CA, TX, FL, and NY). Regional consumption patterns, adjusted for intra-regional income differences, are used because the CEX sample size does not allow state- level disaggregation. TABLE A .1: INCOME BY STATE AND DECILE (ANNUAL MEDIAN INCOME PER CAPITA ) Decile medians State State mean State median 1 2 3 4 5 6 7 8 9 10 Alabama 19933 13308 3033 5242 7257 9412 11886 14899 18816 24402 33786 58381 Alaska 24833 18806 5516 8682 11373 14109 17124 20653 25065 31097 40732 64117 Arizona 22220 15544 3870 6472 8789 11222 13977 17286 21529 27491 37331 62436 Arkansas 18525 12772 3092 5225 7139 9161 11461 14234 17807 22850 31221 52761 California 24889 16616 3788 6545 9062 11752 14841 18603 23494 30469 42186 72895 Colorado 26356 18829 4887 8048 10830 13728 16986 20873 25827 32738 44052 72553 Connecticut 31525 20964 4745 8220 11399 14802 18714 23484 29692 38554 53464 92628 Delaware 25540 18527 4959 8075 10792 13606 16753 20490 25229 31807 42508 69214 D.C 31408 17795 3082 5895 8671 11801 15564 20346 26833 36521 53716 102747 Florida 23624 15925 3695 6343 8748 11310 14243 17805 22423 28989 39980 68631 Georgia 23183 15895 3808 6460 8847 11373 14251 17729 22215 28560 39113 66357 Hawaii 23589 16969 4465 7316 9815 12411 15323 18791 23200 29337 39356 64489 Idaho 19552 14231 3835 6229 8312 10467 12874 15730 19347 24362 32510 52800 Illinois 25320 17521 4271 7200 9822 12588 15730 19516 24387 31256 42640 71871 Indiana 22353 16350 4452 7203 9591 12055 14803 18058 22175 27873 37111 60046 Iowa 21561 15925 4426 7107 9420 11798 14441 17561 21495 26922 35684 57303 Kansas 22473 16138 4232 6943 9321 11794 14569 17876 22082 27940 37510 61543 Kentucky 19828 13417 3135 5368 7392 9545 12007 14993 18861 24353 33534 57414 Louisiana 18534 12179 2698 4711 6564 8556 10855 13666 17337 22597 31485 54984 Maine 21406 15398 4052 6639 8907 11263 13905 17052 21053 26622 35714 58521 CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 21 We implemented these adjustments by the following steps:

1. Estimate median income by decile in each state: We obtain state-level data on mean income and the Gini index of income distribution from the US census. 20 From these data, we estimate median incomes for each decile by as- suming that income distribution has a log-normal distribu- tion – the distribution most commonly assumed in the literature (Kemp-Benedict, 2001). The means and Ginis pro- vide sufficient information to determine a unique log-normal distribution. We take estimated incomes at the 5th, 15th, 25th, etc. percentiles of this distribution as the medians for each decile. The results are shown in Table A.1.

2. Calculate national expenditure on consumption of five categories of goods: electricity, gasoline, natural gas, fuel Decile medians State State mean State median 1 2 3 4 5 6 7 8 9 10 Maryland 28071 20192 5313 8706 11679 14769 18234 22361 27607 34910 46832 76739 Massachusetts 28441 19428 4621 7860 10781 13878 17409 21681 27197 35008 48018 81678 Michigan 24294 17297 4458 7361 9920 12590 15595 19185 23763 30159 40643 67109 Minnesota 25423 18534 5012 8130 10841 13644 16772 20481 25178 31685 42250 68533 Mississippi 17373 11531 2600 4511 6261 8134 10290 12920 16345 21237 29473 51130 Missouri 21848 15311 3825 6389 8670 11064 13772 17023 21189 27041 36692 61288 Montana 18796 13475 3521 5785 7772 9840 12162 14929 18452 23362 31388 51563 Nebraska 21494 15722 4281 6926 9222 11592 14234 17364 21323 26802 35685 57738 Nevada 24098 17276 4515 7416 9964 12616 15592 19141 23657 29952 40242 66108 New Hampshire 26131 19423 5471 8742 11553 14436 17632 21397 26135 32655 43154 68952 New Jersey 29596 20330 4887 8280 11331 14558 18232 22669 28390 36476 49915 84573 New Mexico 18916 12994 3124 5292 7242 9305 11653 14489 18146 23314 31904 54055 New York 25632 16298 3407 6078 8578 11295 14461 18368 23517 30967 43700 77972 North Carolina 22255 15512 3835 6431 8746 11182 13942 17260 21521 27514 37420 62747 North Dakota 19473 14126 3781 6157 8228 10374 12773 15623 19236 24251 32411 52772 Ohio 23017 16360 4202 6947 9369 11898 14746 18150 22493 28565 38524 63692 Oklahoma 19338 13407 3280 5521 7526 9640 12039 14929 18645 23881 32554 54801 Oregon 22948 16395 4255 7008 9430 11953 14790 18175 22488 28506 38357 63173 Pennsylvania 22883 15950 3943 6612 8993 11497 14335 17747 22128 28290 38475 64517 Rhode Island 23768 16417 3988 6731 9190 11785 14735 18291 22869 29328 40042 67580 South Carolina 20598 14305 3512 5904 8042 10295 12850 15926 19879 25446 34661 58271 South Dakota 19246 13845 3643 5969 8008 10126 12502 15331 18928 23936 32110 52616 Tennessee 21253 14463 3416 5826 8003 10314 12953 16149 20281 26138 35907 61237 Texas 21498 14492 3363 5772 7961 10292 12962 16203 20406 26381 36382 62455 Utah 19929 14907 4256 6767 8916 11114 13546 16405 19995 24924 32839 52209 Vermont 22603 16560 4524 7311 9727 12220 14997 18285 22442 28192 37507 60610 Virginia 26274 18413 4600 7684 10426 13305 16562 20471 25482 32519 44125 73705 Washington 25176 18049 4717 7748 10410 13180 16290 19997 24716 31292 42042 69067 West Virginia 18057 12219 2855 4889 6732 8692 10934 13654 17176 22178 30539 52286 Wisconsin 23311 17355 4905 7828 10337 12908 15758 19114 23333 29137 38476 61401 Wyoming 20969 15237 4093 6655 8888 11198 13781 16846 20731 26121 34883 56727 U.S. average 23657 16160 3844 6538 8968 11544 14481 18034 22623 29120 39942 67940 TABLE A .1: INCOME BY STATE AND DECILE (ANNUAL MEDIAN INCOME PER CAPITA ), CONTINUED CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 22 oil and other: We use national consumption data from the 2003 Consumer Expenditure Survey to calculate the carbon charge for each household, using the methodology de- scribed in Boyce and Riddle (2007), with two further ad- justments: (i) we include home ownership expenses as expenditures; and (ii) we use corrected survey weights (which affects the magnitude of expenditure but has little effect on its distribution).

3. Adjust expenditures in response to price increases and dividends: We adjust consumption expenditures to respond to the new price structure, using short-run price elasticities drawn from the literature (see Boyce and Riddle 2007), and to the increase in income in response to dividend payments.

4. Estimate relationship between category-specific expendi- ture and total expenditure: We use a log-quadratic func- tional form to estimate the relationship between each category of expenditure and total expenditure for each household. 21 5. Calculate predicted expenditures in each of five catego- ries for each state and income decile: Incomes (from the Census) do not match perfectly with CEX expenditure data.

There are several reasons for this: (i) expenditure differs from income due to saving (or borrowing); (ii) household expenditure does not include tax payments, whereas Cen- sus income is pre-tax; and (iii) the CEX data on expenditure may be subject to under-reporting. To apply the relationship between carbon charges and expenditures estimated from the CEX, we must first match the appropriate expenditure level to the census income level for median households in each decile. To do this, we calculate means and Gini in- dexes for the expenditures from the CEX data, and find the transformation that converts the national Census data on income into a log-normal distribution with mean and Gini that matches the CEX data. We apply this transformation to the decile median income for each state to obtain median total expenditures for each decile in each state. We then apply the relationships from step (4) to estimate the cate- gory-specific expenditures for each state. 6. Adjust for regional differences in consumption patterns:

We begin with the data presented in Appendices A-D of Burtraw et al. (2009) which report region-wise data on household electricity, gasoline, natural gas, and fuel oil con- sumption in physical units (kWh, gallons, cubic feet) per household. The ratios of regional to national averages from these data are then applied to our national estimates of expenditure in dollars from step (3). These regional expendi- ture levels on the four fuels are then compared to predicted regional expenditures based on weighted averages of the results by state from step (5). This ratio gives an adjustment factor for each region, which is then applied to all states in the region.

22 Expenditures on other goods are adjusted to make the total expenditure on all five categories for each region remain the same as it was before the regional ad- justments.

7. Find carbon intensity of electricity generation by state:

Carbon intensities of electricity consumption for each state were calculated by Jesse Jenkins of the Breakthrough Insti- tute. These are based on the USEPA’s e-Grid data for the year 2005, combining data on the carbon intensity of elec- tricity generated in each state with adjustments to account for imports of electricity across state lines within intercon- nected power grids. 23 8. Apply carbon loading factors to expenditures on each of the five expenditure categories: The loading factors for each fuel, in units of carbon per dollar, are calculated using Input- Output (IO) accounts. We use the 2003 IO tables, 24 with ad- justments using the 2002 benchmark IO tables which pro- vided more detailed breakdowns. 25 We assign carbon emissions from coal, oil, and natural gas using emissions data from the U.S. Energy Information Administration (EIA). 26 Using a methodology similar to that described in Metcalf (1999), we trace this carbon through the economy to de- termine the final carbon content of each commodity cate- gory from the IO accounts, including indirect uses. To assign these loading factors to the CEX expenditure categories, we first convert the commodity categories from the IO accounts into Personal Consumption Expenditure (PCE) categories using bridge tables produced by U.S. Bureau of Economic Analysis, 27 and then from PCE categories into CEX catego- ries using the documentation for the National Bureau of Economic Research (NBER) CEX family-level extracts. 28 In the case of electricity, the loading factor is adjusted in each state in proportion to the carbon intensity of electricity gen- eration from step (7). In the case of the “other goods” cate- gory of expenditure, the loading factor is derived from the loading factors of the different goods and services that make up this category, which can vary across deciles. We therefore estimate the relationship between this loading factor and total expenditure, and use this to construct load- ing factors for each decile in each state. 29 Finally, the load- ing factor for each expenditure category is multiplied by the corresponding expenditures to obtain the carbon footprint.

9. Adjust for consistency with National Accounts data: The carbon content for all categories of expenditure is scaled by a constant factor so that the total carbon content of house- hold consumption is correct in proportion to total U.S. car- bon emissions (see Boyce and Riddle 2008).

10. Calculate increased spending on each category of goods: A permit price of $25 per ton CO 2 is multiplied by the carbon content of each expenditure category from step (9) to obtain the impact of carbon pricing on expenditure in each category. The total increase in expenditure is the sum of the increases for each category. The results are shown in Table A.2.

CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 23 Decile medians State State mean State Median 1 2 3 4 5 6 7 8 9 10 Alabama 316 283 136 178 209 238 267 299 336 384 453 608 Alaska 321 295 155 196 226 253 280 310 343 386 448 581 Arizona 289 259 124 163 191 218 244 274 308 351 416 561 Arkansas 301 272 134 174 203 230 257 287 321 365 428 567 California 288 251 108 149 179 207 236 268 305 354 426 593 Colorado 359 325 166 213 247 278 309 343 382 433 507 673 Connecticut 345 302 137 184 218 251 284 321 364 420 506 707 Delaware 370 339 179 226 260 291 322 356 395 444 517 675 D.C 397 337 143 196 236 275 315 360 413 484 594 861 Florida 304 268 126 166 196 224 253 285 322 370 442 607 Georgia 349 315 156 202 236 267 298 332 372 422 497 661 Hawaii 333 301 155 198 229 257 286 318 354 401 469 620 Idaho 270 245 121 158 184 208 232 259 289 327 384 505 Illinois 340 305 148 193 227 258 289 323 362 412 486 652 Indiana 378 349 188 236 271 302 333 366 404 452 521 669 Iowa 350 323 172 217 250 279 308 339 375 420 484 621 Kansas 354 324 168 214 247 278 308 341 378 426 495 644 Kentucky 346 314 157 202 236 266 297 331 369 418 490 647 Louisiana 314 281 135 177 208 236 265 297 334 381 451 604 Maine 286 259 128 166 194 219 245 273 305 346 406 538 Maryland 359 325 166 212 246 277 309 343 382 433 508 674 Massachusetts 344 306 146 192 226 257 289 324 365 418 498 679 Michigan 347 316 159 205 239 269 300 333 371 419 490 645 Minnesota 363 332 172 220 254 285 316 349 388 437 508 662 Mississippi 289 258 123 162 190 217 244 274 308 351 416 557 Missouri 354 323 165 211 245 276 306 340 378 427 498 652 Montana 296 270 137 176 204 230 256 284 316 357 417 546 Nebraska 333 306 159 203 234 263 291 322 357 401 464 601 Nevada 320 289 145 187 217 246 274 305 341 387 454 603 New Hampshire 316 288 147 188 218 246 273 303 338 381 446 586 New Jersey 342 303 144 189 223 254 286 321 362 415 496 679 New Mexico 302 272 135 175 204 231 258 287 322 365 430 572 New York 292 251 108 148 178 206 236 268 307 357 435 617 North Carolina 330 298 148 191 223 253 282 314 352 399 469 621 North Dakota 351 324 173 218 250 280 309 340 376 421 486 623 Ohio 359 328 169 216 250 281 312 345 384 432 504 658 Oklahoma 313 283 140 182 212 240 269 299 334 379 444 586 Oregon 268 239 111 148 175 200 226 253 286 327 388 524 Pennsylvania 313 281 138 179 210 238 266 297 333 379 447 598 TABLE A .2: CARBON PRICE IMPACT BY STATE AND INCOME DECILE ($ PER CAPITA ) CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 24 s TABLE A .2: CARBON PRICE IMPACT BY STATE AND INCOME DECILE ($ PER CAPITA ), CONTINUED Decile medians State State mean State Median 1 2 3 4 5 6 7 8 9 10 Pennsylvania 313 281 138 179 210 238 266 297 333 379 447 598 Rhode Island 307 274 131 172 202 230 259 290 326 372 441 596 South Carolina 291 261 125 165 193 220 247 277 311 354 418 558 South Dakota 299 273 137 177 206 232 259 287 320 361 421 550 Tennessee 324 291 142 185 217 246 276 308 345 393 463 619 Texas 332 298 143 187 220 251 282 315 354 404 477 638 Utah 337 312 170 213 243 270 298 327 360 402 462 590 Vermont 269 243 118 155 181 206 230 257 288 327 385 511 Virginia 364 329 165 212 247 279 312 347 388 440 517 688 Washington 275 245 113 151 179 205 231 260 293 336 400 542 West Virginia 325 294 146 189 221 250 279 310 347 393 460 607 Wisconsin 364 336 181 228 261 291 321 353 389 435 502 643 Wyoming 351 323 172 217 249 278 307 339 375 421 487 629 U.S. average 317 283 135 177 209 238 267 299 337 385 457 618 CAP & DIVIDEND: A STATE-BY-STATE ANALYSIS / BOYCE & RIDDLE / PAGE 25 NOTES 1 In addition to carbon pricin g, the climate policy package may include regulatory standards and public investment in energy efficiency and renewable energy (Boyce 2009a). 2 The extent to which offsets reduce atmospheric carbon is often difficult to ascertain. It is hard to say, for example, whether a forest would have been replanted (or cut down) in the absence of an offset deal, or whether a coal-burning power plant in Asia would have been built to different en- ergy-efficiency specifications. Concerns about “additionality” have already surfaced in the voluntary offset market; see, for example, Elgin (2007).

3 A tax-and-dividend policy is advocated, for example, by nsen (2009).

James Ha 2 blank 3 blank 4 For discussion, see Boyce (2009b). On concerns about the potential for speculative bubble s in carbon derivatives mar- kets, see Chan (2009).

5 Office of Management and Budget (2009) “Summary Table S-2: Effect of Budget Proposals on Projected Deficits.” The budget put the amounts over the decade at $525.7 billion and $120 billion, respectively.

6 Details of our methods are given in the Appendix.

7 The ratio between the carbon charges to the highest and lowest deciles is somewhat lo wer than the ratio of the car- bon footprints shown in Figure 2, because the figures in Table 1 incorporate changes in demand due to higher fossil fuel prices (with demand for necessities being less price- elastic than demand for luxuries) and after receipt of the dividend.

8 For details on the data sour ces used to calculate these shares, see Boyce and Riddle (2008).

9 We assume in our calculations that the price elasticity of demand is constant across inco me deciles. There is some evidence, however, that demand elasticity is greater in the lower-income deciles, in which case the progressivity of the cap and dividend policy would be somewhat stronger than shown in these results. For di scussion on this point, see Boyce and Riddle (2007, p. 13).

10 For this reason, low-income states tend to fare somewhat better under the cap-and-divide nd policy than high-income states. In West Virginia, for example, the effect of lower- than-average income outweighs the effect of the state’s more carbon-intensive electricity supply: the median house- hold sees a net benefit equivalent to 0.8% of its income, above the national median of 0.6%.

11 The data for these calculations on military expenditure are from www.statemaster.com/graph/mil_def_con_exp_ percap-defense-contracts-expenditures-per-capita. The data on farm programs are from the Environmental Working Group’s database, farm.ewg.org/farm/progdetail.php?fips =00000&yr=2006&progcode=total&page=states. We are grateful to Elizabeth Stanton et al. (2009) for suggesting these comparisons. 12 Apart from being less transparent, a drawback of free allocations is that they make permit trading a necessary element of the policy, since those who get the free permits are not identical to those who need them.

13 A carbon pricing policy will also have impacts on the pur- chasing power of local, state, and federal governments. For discussion, see the si debar on page 17.

14 For discussion of these employment effect s, see Pollin et al. (2008).

15 Provisions to separate allo wance-value rebates from kilo- watt hour-based charges in electricity bills, so as to maintain incentives for electricity use reduction at the margin, will dampen this effect only insofar as consumers read and are able to make sense of the fine print in their monthly bills.

16 For discussion of these and other problems associated with provision of free allowances to LDCs, see Sweeney et al. (2009) and Stone and Shaw (2009).

17 Calculated from Table 2 in CBO (2009, p. 16). The CBO’s results hinge, among other things, on the possibly optimistic assumption that the state public utility commissions will ensure that the full value of free allocations to LDCs is passed to their customers. If no t, the distributional impact of ACES could be more inequitable.

18 The assumed carbon price affects the magnitude of the dollar amounts reported, but not the distributional pattern.

The incidence of higher (or lower) carbon prices can be cal- culated simply by multiplying our numbers by the ratio of the assumed price to ours. For example, a more ambitious tar- get resulting in a permit price of $50/ton CO 2 would double the dollar values we report. 19 For details, see Boyce & Riddle (2007) where we report estimates by expenditure decile at the national level.

20 These census data are av ailable at: www.census.gov/ hhes/www/income/histinc /state/statetoc.html.

21 We obtain the following estimates:

ln(electricity expenditure) = 2.297 + 0.333*ln(expenditure) + 0.003*ln(expenditure-squared). ln(gasoline expenditure) = -11.786 + 3.265*ln(expenditure) - 0.145*ln(expenditure-squared). ln(natural gas expenditure) = -5.097 + 1.752*ln(expenditure) - 0.073*ln(expenditure-squared). ln(fuel oil expenditure) = -3.117 + 1.156*ln(expenditure) - 0.043*ln(expenditure-squared). ln(other goods expenditure) = 3.123 + 0.323*ln(expenditure) + 0.036*ln(expenditure-squared). 22 We assigned the seven states that are not in any of the regions in Burtraw et al. (2009) as follows: Northeast for Vermont, Northwest for Wyomin g and Alaska, Mountains for New Mexico, Plains for Iowa and North Dakota, and Florida for Hawaii.

23 Stanton et al. (2009) report a similar state-level measure of the carbon intensity of electr icity, using the national aver- age instead of regional power grids to estimate the carbon content of electricity imported across state lines. The corre- lation between their state measure and ours is 0.98. 24 US Bureau of Economic Analysis, “1998-2007 Supple- mentary Make and Use Tables after redefinitions at the summary level,” available at www.bea.gov/industry/io_ an- nual.htm.

25 US Bureau of Economic Analysis, "2002 Standard Make and Use Tables at the Summary Level,” available at www.bea.gov/industry/io_benchmark.htm.

26 EIA, “International Energy Annual 2006”, available at www.eia.doe.gov/iea/carbon.htm l. Additional data on the small amount of crude oil that does not go to refineries are taken from: EIA, “Petroleum Na vigator, US Crude Oil Supply and Deposition” (available at tonto.eia.doe.gov/ dnav/pet/pet_sum_crdsnd_adc_mbbl_a.htm), and EIA, “Pe- troleum Navigator; Refining & Processing; Weekly Inputs, Utilization & Production” (available at tonto.eia.doe.gov/ dnav/pet/pet_pnp_wiup_dcu_nus_w.htm). 27 US Bureau of Economic Analysis, “PCEBridge_2002- 2007,” available at www.bea.gov/industry/more.htm.

28 NBER, Documentation for “Consumer Expenditure Survey Family Level Extracts,” available at www.nber.org/data/ ces_cbo.html.

29 We again use a log-quadratic function, and obtain the following estimate:

ln(carbon intensity of other goods expenditure) = -6.665 - 0.541*ln(expenditure) + 0.030*ln(expenditure-squared).

REFERENCES Boyce, James K. (2009a) ‘Smart Climate Policy,’ E3: Eco- nomics for Equity and the Environment, Policy Brief No. 1, April. Online at www.e3network.org/briefs/Boyce_Smart _Climate_Policy.pdf.

Boyce, James K. (2009b) ‘Cap and No Trade,’ Commentary posted on capanddividend.org, January. Online at http://www.capanddividend.org/?q=node/233.

Boyce, James K. and Matthew E. Riddle (2007) ‘Cap and Dividend: How to Cur b Global Warming While Protecting the Incomes of American Families,’ Amherst, MA: Political Econ- omy Research Institute, Working Paper No. 150, November.

Online at www.peri.umass.edu/fileadmin/pdf/working_pa- pers/working_paper s_101-150/WP150.pdf.

Boyce, James K. and Matthew E. Riddle (2008) ‘Keeping the Government Whole: The Impact of a Cap-and-Dividend Policy for Curbing Global Warming on Government Revenue and Expenditure,’ Amherst, MA: Political Economy Research In- stitute, Working Paper No. 188, November. Online at http://www.peri.umass.edu/fileadmin/pdf/working_papers/ working_papers_151-200/WP188.pdf.

Boyce, James K. and Matthew E. Riddle (2009) ‘Cap and Dividend: How to Curb Global Warming While Promoting Income Equity,’ in Jonathan Harris and Neva Goodwin, eds., Twenty-First Century Macroeconomics: Responding to the Climate Challenge. Cheltenham and Northampton: Ed- ward Elgar.

Broder, John (2009) ‘Adding Something for Everyone, House Leaders Won Climate Bill,’ The New York Times, 1 July, pp.

A1, A17.

Burtraw, Dallas, Richard Sw eeney and Margaret Walls (2009) ‘The Incidence of U.S. Climate Policy: Alternative Uses of Revenues from Cap-and-Trade Auctions,’ Washing- ton, DC: Resources for the Futu re, Discussion Paper, April.

Online at www.rff.org/RFF/Documents/RFF-DP-09-17.pdf.

Chan, Michelle (2009) ‘Subprime Carbon? Re-thinking the World’s Largest New Derivative s Market,’ Washington, DC:

Friends of the Earth, Marc h. Online at www.foe.org /pdf/SubprimeCarbonReport.pdf.

Congressional Budget Office (CBO) (2009) ‘The Estimated Costs to Households From the Cap-and-Trade Provisions of H.R. 2454,’ June 19. Online at www.cbo.gov/ftpdocs/ 103xx/doc10327/06-19-CapAndTradeCosts.pdf. Elgin, Ben (2007) ‘Another Inconvenient Truth,’ Business Week, 26 March. Online at www.businessweek.com/maga- zine/content/07_13/b4027057.htm.

Hansen, James E. (2009) ‘Carbon Tax & 100% Dividend vs.

Tax & Trade.’ Testimony submitted to the Committee on Ways and Means, U.S. House of Representative, 25 Febru- ary. Online at www.columb ia.edu/~jeh1/mailings/2009/ 20090226_WaysAndMeans.pdf.

Kemp-Benedict, Eric (2001) "Income Distribution and Pov- erty: Methods for Using Available Data in Global Analysis," Polestar Technical Note #4 Metcalf, Gilbert E. (1999) "A Di stributional Analysis of an Environmental Tax Shift," National Bureau of Economic Re- search Working Paper #6546.

Office of Management and Bu dget (2009) A New Era of Re- sponsibility: Renewing Amer ica’s Promise. Online at www.whitehouse.gov/omb/assets/fy2010_new_era/a_new _era_of_responsibility2.pdf.

Pollin, Robert, Heidi Garrett-Peltier, James Heintz and Helen Scharber (2008) Green Recovery: A Program to Create Good Jobs & Start Building a Low-Carbon Economy. Amherst, MA:

Political Economy Research In stitute, and Washington, DC:

Center for American Progre ss, September. Online at www.peri.umass.edu/fileadmin/ pdf/other_publication_type s/peri_report.pdf.

Stanton, Elizabeth A., Frank Ackerman and Kristen Sheeran (2009) ‘Greenhouse Gases and the American Lifestyle: Un- derstanding Interstate Differences in Emissions,’ Portland, OR: Economics for Equity and the Environment, May. Online at www.e3network.org/pape rs/NRDC_state_emissions_ report.pdf.

Stone, Chad and Hannah Shaw (2009) ‘Senate Can Strengthen Climate Legislatio n by Reducing Corporate Wel- fare and Boosting True Consumer Relief,’ Washington, DC:

Center for Budget and Policy Priorities, July 10. Online at http://www.cbpp.org/file s/7-10-09climate.pdf.

Sweeney, Rich, Josh Blonz an d Dallas Burtraw (2009) ‘The Effects on Households of Alloca tions to Electricity Local Dis- tribution Companies,’ Washin gton, DC: Resources for the Future, June 5. Online at www.rff.org/wv/Documents/LDC_ Allocation_090605.pdf.