I recently did a little research attempting to understand the effects of Citizens' Climate Lobby's carbon fee and dividend proposal on people in the lowest quintile of income earners. Below is the result of that research.
Above Average Carbon Emitters Among Low Income Households[*]
Alan Mattlage, Ph.D., M.L.S.
One of many strengths of Citizens’ Climate Lobby’s (CCL) carbon fee and dividend proposal is that it provides a net benefit to 86% of households in the lowest quintile of income earners (Ummel, 2016). The progressive distribution of benefits is driven by the fact that lower income households tend to have lower than average carbon footprints but will receive a modified equal per capita share of the revenues from the fee (one share for each adult and one-half share for each child up to two children per family). This makes the proposal one of the most economically progressive methods for transitioning to a de-carbonized economy. See Figure 1. Benefiting 86% still leaves 14% of the lowest quintile worse off – that’s about 2.8% of all American households or about 3.5 million households (U.S. Census Bureau, 2016). The following is an attempt to understand why these households are made worse off, how significant the burden is, and what can be done to relieve that burden.
The short answer is that we cannot know precisely who is burdened, but there are data that can give us a general understanding of the circumstances that will tend to make a household worse-off.
Figure 1 – Distribution of benefits and burdens across quintiles
Why are households made worse-off?
The obvious reason households are made worse off is that their dividend is less than the cost they incur from the carbon fee. This is true no matter which quintile the household is in. To answer the question why are 14% of the lowest income households made worse off, we need to understand why a household might have above-average emissions. Some of those reasons will be particularly common to low income households.
A net financial loss is probably the result of some combination of the following factors:
1. the household is composed of a single person,
2. the household’s electricity is provided disproportionately by a carbon-intensive fuel, like coal,
3. the household is located in a particularly harsh climate,
4. the household is poorly insulated and has inefficient HVAC equipment and appliances,
5. the household’s transportation costs are particularly high, and
6. the household buys an unusually large amount of carbon intensive consumer goods and services.
1. Household Composition
Perhaps the most important factor of all is the household’s size. Single person households will receive only one dividend share, while needing to cover the entire cost of maintaining the household, e.g., utilities, transportation, and necessary household consumer goods. Meanwhile, a two-adult household will receive twice the dividend, but they will share many of the costs of maintaining the household. Consequently, single person households are likely to be disproportionately represented in those households suffering a loss no matter what income quintile they are in.
This has been borne out by the household impact study conducted by Kevin Ummel for Citizens’ Climate Lobby. Ummel compared the net benefits going to two household types: “minority households” and “elderly households.” Minority households are composed of more members (sometimes more than two adults which compounds the dividend). Elderly households are, by definition, composed of only one or two adults. These households have similar incomes, but the mean benefit accruing to minority households would be significantly greater than the mean benefit accruing to elderly households: $148/year for minority households versus $2/year for elderly households considering all households (Ummel, 2016, p. 31; U.S. Census Bureau, 2016). For households in just the lowest quintile, 99% of families of four are benefitted, while 81% of elderly households are benefitted. The mean benefits for these households are $596/year and $138/year respectively. These disparities show that living in a small or single person household contributes to a financial burden under the fee and dividend proposal.
2. Local Carbon Intensity of Electricity
Ummel also broke down the consumer habits of each income quintile into nine categories and determined the carbon tax burden for each category. See Figure 2. The largest single category for all but the top quintile is “utilities.” Consequently, households that run on especially carbon intensive electricity would be disadvantaged by the carbon fee, while households running on clean energy would tend to benefit (Ummel, 2016, p. 34). The importance of this factor is suggested by comparing Figures 3 and 4. There is a rough correlation between the geographic distribution of the carbon intensity of the electricity supply and the financial impact on households in the lowest quintile.
3. Harsh Climates
Little need be said about this. Households in especially cold or hot climates certainly will have higher utility costs than households in milder climates; however, in light of Figure 4, this factor does not seem to be particularly significant as many regions of the country experiencing harsh climates tend to enjoy
Figure 2 – National tax burden by quintile
relatively greater benefits, e.g. New England, Florida, the Southwest, and Montana. With inevitable changes to the climate, many climates will become harsher, increasing the importance of this factor.
4. Poor Insulation and Inefficient HVAC Equipment and Appliances
Again, it is obvious that households which live in poorly insulated housing and have inefficient HVAC equipment and appliances will bear a greater burden from a carbon tax as more energy will be required to achieve the same results. This is a particularly significant problem for poor, low income households. Such households often live in rental property, the owner of which is unconcerned about energy efficiency since the utilities are paid by the tenants. Low income housing stock is significantly less efficient than average. If low income housing were brought up to the national average, low income households would see a 35% reduction in their utility costs (Drehbol and Ross, 2016).
The second largest category is “gasoline,” but while our common acquaintance with gasoline as a fossil fuel might lead us to think this would be the most significant factor in establishing a household’s carbon footprint, it actually makes up less than 25% of the total carbon footprint of U.S. households. Still, if a household uses more gasoline than the national average, this factor will contribute to reducing the benefit they would receive from the carbon dividend and might, like other carbon-intensive factors, push them into a net loss.
Figure 3 – Carbon intensity of energy supply
Figure 4 – Lowest quintile households benefited by fee and dividend
However, simply comparing gasoline costs between the lowest quintile and other quintiles might not be very revealing, since a disproportionate number of households in the lowest quintile likely will not have cars. This means the gasoline costs borne by households in the lowest quintile will fall mainly on the car-owning households.
6. Indirect Costs
Utility costs and gasoline costs are “direct” energy purchases. For the lowest income quintile, they make up 50% of the total costs. The other categories are “indirect” energy purchases that appear as embodied energy in the goods and services that a household purchases. Any household which purchases goods and services which are in total more carbon intensive than those purchased by the average household will find their net financial benefits lowered by these categories of expenditures, possibly pushing them into a net loss. Because indirect purchases make up 50% of the lowest quintile’s energy purchases and even larger percentages of higher quintiles, these purchases are especially important in determining a household’s net benefit. Indirect costs are the primary factor that burdens the higher quintiles and benefits the lowest income quintile, but lowest quintile households that purchase consumer products with an above average carbon footprint, nevertheless, will be penalized by these categories of purchases.
If a household remains well-enough below average in enough of the consumption categories, it will likely benefit from the fee and dividend mechanism, despite a high carbon footprint in other categories; however, if their consumption is in total above average across these categories, they will suffer a loss. The worst-case scenario would likely be a single-person household, reliant upon carbon-based electricity, in a harsh environment, living in a poorly insulated house, with a long commute, purchasing unusually carbon-intensive goods and services. Surely not every one of these conditions need hold to push someone into a loss, but various combinations of these factors could do so. No single profile of a low-income household can tell the story for all 3.5 million households.
How significant is the burden?
Forty-seven percent of all households suffer a loss and their median loss amounts to $195/year. Among the lowest income quintile, only 14% of households suffer a loss and their median loss is only $96/year or $8/month. Consequently, the absolute burden is least for households in the lowest income quintile, but their median loss amounts to 0.79% of their income while the median loss nationally is only 0.25% of income. So even as the lowest quintile is much more carbon virtuous than higher quintiles, their low income could make managing their loss nevertheless more difficult.
There are additional considerations that are aggravating factors:
1. steadily rising fee,
2. limited conservation options for households in the lowest income quintile, and
3. the ability of higher income quintiles to achieve conservation savings.
CCL’s fees are proposed to begin at $15 in the first year and rise $10 in each subsequent year. This means that the households (in all quintiles) that benefit will gain greater benefits as more revenue is collected, but burdened households will be made increasingly worse-off. This could become a significant problem over time for households in the lowest quintile, unless those households are able to reduce their carbon footprint. In many cases, we cannot reasonably expect this. Furthermore, if households in higher quintiles are able to reduce their emission faster than households in the lowest quintile, less revenue will be collected without a comparable reduction of costs for the lowest quintile. Households in the lowest quintile might be left behind in the race to reduce carbon emissions.
There are several mitigating factors, however. Ummel’s study does not seek to predict how the quintiles will fair in future years, but another study by Regional Economic Modeling, Inc. has projected the consequences of the fee and dividend nationally and regionally. It provides some consolation that the fee and dividend will boost economic growth and create 2.1 million jobs in the first ten years. These potentially could help reduce the harms across the board and soften the burden for the lowest quintile. There are, however, more immediately recognizable mitigating factors:
1. the pass-through assumption,
2. under-reported high quintile expenditures,
3. price insensitivity among high income quintiles,
4. home ownership,
5. family wealth, and
6. temporary loss of income.
1. The Pass-Through Assumption
Perhaps most significant mitigating factor is that Ummel has assumed that 100% of the fee is passed on to consumers. This undoubtedly will not happen. For excise taxes generally, businesses commonly assume roughly 25% of the cost by reducing returns to capital and restricting or reducing wages (Tax Policy Center, 2017). To some extent, as wages are reduced across all wage levels and as stockholders tend to be in higher income quintiles, the costs borne by the business will largely burden households in the higher quintiles. So, the 14% of low-income households or 3.5 million households that are projected to lose financially is an overestimate. Other analyses use different pass-through assumptions. For example, the Department of Treasury’s Office of Tax Analysis assumes no costs are passed on to the consumer (Horowitz, 2017).
2. Under-Reported Expenditures
Ummel also noted that there is a discrepancy between expenditures reported in the Bureau of Labor Statistics’s Consumer Expenditure Survey and the Personal Consumption Expenditures component of the U.S. Bureau of Economic Analysis’s national accounts. The consequence of this is that reported expenditures are known to be underestimates of actual expenditures. Utilities, gasoline, and a few other expenditures are reported more or less accurately, but other categories were under-reported by 47%. Importantly, expenditures made by households in higher income quintiles are disproportionately under-reported. Nonetheless, Ummel assumes that under-reporting is uniform across all income levels. In a footnote, he recognizes that this will underestimate spending among the rich (Ummel, 2016, p. 4-6). Underestimating these expenditures means that the aggregate revenue will be larger than estimated, and the dividend to the lowest income quintile also will be larger. This is especially important as the underestimated expenditures are indirect expenditures and make up 60% of the expenditures of the highest income quintile. Again, this will not only increase the dividend to all households, it will reduce the percent of lowest income households that are expected to be made worse off by the fee and dividend.
3. Price Insensitivity
Perhaps one of the reasons that expenditures are under-reported by households in the higher quintiles is that they are less concerned about their expenditures than less well-off households. Consequently, they do not notice those expenditures. They also are likely to be less price sensitive than the less well-off. They will be willing to make the same purchases even as prices rise somewhat. According to Randy Schnepf of the Congressional Research Service, this is true for consumers making food purchases, “low-income consumers who spend a significant share of their household budget on food are likely to be … more responsive to price changes … than high-income consumers with lower food budget shares” (Schnepf, 2013, p. 28). In the long run, this is very good news for households in the lowest income quintile. That households in the higher quintiles will be relatively insensitive to price changes will mean that the lowest income households will be better able to keep pace with any conservation efforts made by those in higher quintiles.
4. Home Ownership
Not everyone who is in the lowest income quintile is poor. Many people have substantial wealth in their houses. Retirees who have paid off their mortgage can live comfortably with a relatively low income. Others may have substantial savings conservatively invested. While their income is low, they nonetheless could enjoy a lifestyle that involves higher than average expenditures and large carbon footprints. Nationally, there are 6,896,000 homeowners with incomes less than $30,000 who are 65 years old or older (U.S. Census Bureau, 2015). Some of these still might be carrying a mortgage, but the number is likely small. While they may not have the security of people in the highest quintiles, their ability to afford the increased costs of a carbon fee might be similar to middle class households. The existence of these households reduces the percent of households about which we should be acutely concerned.
5. Family Wealth
Like homeowners with low incomes, other households in the lowest quintile are not poor in the concerning sense. They may be nominally distinct households, but benefit from their wealthy families. Some college students would be good examples of these “households.” While officially falling within the lowest quintile, they could have resources that allow them to make above average expenditures and thereby be among the households that are made worse off by the fee and dividend. Roughly 20.5 million students attended college in fall of 2016. A plurality of them came from the highest income quartile of families, i.e., families with an income of at least $116,000. 87% of high school students in the highest quartile went on to college (National Center for Education Statistics, 2017; Pell Institute, 2016). Again, many “households” composed of college students from this stratum of society might be nominally in the lowest income quintile, but well-able to afford an increase in the cost of carbon intensive goods and services.
6. Temporary Loss of Income
The unemployment rate is currently 4.3% or 6.86 million people and the average length of unemployment is about 26 weeks. Only a quarter of the unemployed remain unemployed for longer than 26 weeks (U.S. Bureau of Labor Statistics, 2017). Loss of employment will surely send some number of households into the lowest income quintile for a time. All other things being equal, they will likely not moderate their consumption habits significantly, if they have some savings to tide them over the period of temporary unemployment. As household consumption correlates strongly with income, some number of them likely will be above the national average. What this means is that some number of households with large carbon footprints will be in the lowest quintile only temporarily and will have greater resources in future years to recoup whatever burdens they experience during their weeks or months of unemployment. These households will, in essence, be more likely to have high consumption habits and be more financially secure than those with chronically low incomes. Recent research has revealed that in addition to people losing employment, income is often temporarily lost because of reduced hours during a business downturn (Morduch and Schneider, 2017). The same situation faces small business owners suffering temporary losses. We, of course, should be concerned about all of these groups, but their plight can be distinguished from households with chronically low incomes and for which we should be acutely concerned.
These six mitigating factors or categories of households all reduce the number of households and the burden they would suffer from the fee and dividend. That is, each will reduce the percent of households in the lowest income quintile that are made worse off by the fee and dividend, such that 14% is an over estimate of households that will be significantly or chronically disadvantaged. Additionally, the first three factors (the pass-through assumption and under-reported expenditures and price insensitivity by the higher income households) should increase the dividend for everyone and the mean net financial returns for the lowest income quintile.
Finally, data indicate that the burden among the lowest quintile will fall less on the lowest decile (Ummel, 2016 and Horowitz, 2017).
What can be done to relieve the burdens that remain?
Despite the expected reduction in the number of households that will suffer a net financial loss and the expected reduction in the size of that loss, some households in the lowest income quintile will continue to suffer losses. Consequently, it will be important to find ways to relieve those burdens. One simple way would be to modify the formula for distributing revenue. CCL is promoting a modified per capita distribution (one share for each adult and one-half share for each child up to two children per family). CCL believes the simplicity of this formula will be a political selling point, but it tends to burden single person households. Other distribution formulas, nearly as simple, could be implemented that would protect those households.
Another significant method to protect people in the lowest quintile would be to reduce their carbon intensive expenditures. As the largest single category of expenditures is utilities, significant relief can be achieved by de-carbonizing electricity generation. This is precisely what the carbon fee and dividend is most likely to accomplish. With rising costs for generating coal and gas fueled electricity, wind, solar, and other clean energy sources will gain greater and greater market share. Already, coal-fired power plants are closing around the country and are announced to be retired sooner than initially planned. If, ideally, all electricity was generated from clean sources, the carbon-based expenditures of the lowest quintile would be reduced significantly. As this expenditure is proportionately larger than expenditures by higher income households, the net benefits to the lowest income households would further improve. As electricity becomes cleaner, the carbon fee burden associated with consumer goods would also fall. Businesses will seek to reduce their production costs by making use of low-carbon inputs, thereby reducing the cost of the fee to consumers. So, the success of the fee and dividend would reduce both the fee and the dividend, essentially making it less and less relevant to household finances. The problem of the steadily rising fee would be counter-acted by the reduced number of goods and services to which the fee applies.
The increased cost of gasoline production would stimulate more fuel-efficient transportation. Already, hybrids and electric vehicles are gaining market share. Countries such as Norway, India, China, and France have already announced intentions to promote or move entirely to electric vehicles. These are, of course, foreign markets, but they indicate the direction that the automobile industry is heading. Virtually all major auto manufacturers are now producing an all-electric car. While the upfront cost of new hybrid or electric vehicles will be beyond the reach of most all low-income households, more efficient vehicles in the used car market will become within reach. As these vehicles appear on the market, the carbon-intensity of transportation expenditures by the lowest quintile will fall. Here, too, the fee and dividend is designed to make itself obsolete.
There are a number of ways to reduce the cost of the fee to the lowest quintile that lie outside the fee and dividend proposal: local, state, and federal support for energy efficiency upgrades for affordable housing, support for “Energy Star” appliances, “cash for clunkers” programs, and progressive utility rate structures are a few obvious tools. Money for these programs could be raised by allowing households to voluntarily decline their dividend. Checks would come to them with a notice that if they chose not to cash the check in a specific time period, the money would be used to support energy efficiency programs or perhaps research and development into alternative fuels. Given that most of the revenue generated would come from the highest quintiles and that these checks would make up typically only about 0.2% of their annual income, many well-off households likely would be willing to make this contribution.
Finally, the costs and benefits examined here are only financial. De-carbonizing our economy will benefit vulnerable communities located near polluting power plants. This is a significant concern among environmental justice advocates. Cap-and-trade carbon pricing plans do not address this problem. Under a cap-and-trade system, polluting power plants can purchase off-sets and continue to operate polluting plants usually affecting poorer neighborhoods. In contrast, a fee and dividend plan puts market pressures on all fossil fuel plants equally, thereby promoting healthier conditions for vulnerable communities. The study conducted by Regional Economic Modeling, Inc. for CCL found that CCL’s fee and dividend plan would prevent 230,000 premature deaths in the courses of 20 years.
While we cannot expect that all households within the lowest quintile will be benefitted by CCL’s carbon fee and dividend, it is clear that the proposal is extremely progressive and that the estimate that 14% of the lowest quintile will suffer a burden is an overestimation. There are clearly ways in which the households in the lowest quintile can be protected. Among these is simply the fact the fee and dividend plan is designed to make itself obsolete, particularly in those expenditure categories that are most damaging to the lowest quintile.
Drehbol, Ariel and Lauren Ross, “Lifting the High Energy Burden in America’s Largest Cities: How Energy Efficiency Can Improve Low Income and Underserved Communities,” American Council for an Energy-Efficient Economy, April 2016.
Horowitz, John, et al., “Methodology for Analyzing a Carbon Tax,” Working Paper 115, Office of Tax Analysis, U.S. Department of Treasury, January 2017.
Morduch, Jonathan and Rachel Schneider, The Financial Diaries: How American Families Cope in a World of Uncertainty, Princeton: Princeton University Press, 2017.
National Center for Educational Statistics, “Fast Facts,” U.S. Department of Education, 2017 https://nces.ed.gov/fastfacts/display.asp?id=372, retrieved 6/26/2017.
Nystrom, Scott and Patrick Luckowk, “The economic, climate, fiscal, power, and demographic impact of a national fee-and-dividend carbon tax,” Washington, D.C., Cambridge, Mass.: Regional Economic Modeling, Inc. and Synapse Energy Economics, Inc., June 9, 2014.
Pell Institute, “Indicators of higher education equity in the United States: 2016 historical trend report,” 2016, http://www.pellinstitute.org/downloads/publications-Indicators_of_Higher_Education_Equity_in_the_US_2016_Historical_Trend_Report.pdf, retrieved 6/26/2017.
Schnepf, Randy, “Consumers and food price inflation,” Congressional Research Service, Sept. 13, 2013.
Tax Policy Center, “Briefing Book,” Urban Institute and Brookings Institution, 2017, http://www.taxpolicycenter.org/briefing-book/who-bears-burden-federal-excise-taxes, retrieved 6/26/2017.
Ummel, Kevin, “Impact of CCL’s proposed carbon fee and dividend policy: A high-resolution analysis of the financial effect of U.S. households,” prepared for Citizens’ Climate Lobby, Working Paper v1.4, April, 2016.
U.S. Bureau of Labor Statistics, 2017, https://www.bls.gov/home.htm, retrieved 7/6/2017.
U.S. Census Bureau, “America’s Families Living Arrangements,” 2016, https://www.census.gov/data/tables/2016/demo/families/cps-2016.html, retrieved 6/26/2017.
U.S. Census Bureau, “American Housing Survey,” 2016, https://www.census.gov/programs-surveys/ahs/data/interactive/ahstablecreator.html#?s_areas=a00000&s_year=n2015&s_tableName=Table1&s_byGroup1=a1&s_byGroup2=a1&s_filterGroup1=t1&s_filterGroup2=g1, retrieved 6/26/2017