Countering Housing and Transportation Market Inefficiencies

Transportation Pricing Alternatives

As discussed by McMullen, Zhang & Nakahara (2010), marginal cost pricing results in efficient financing of road systems so that they operate at their optimal level of capacity. Thus, ideally, charges need to be variable to account for different levels of congestion.  Historically, the number of miles driven, road wear and gasoline consumption were relatively equitable and efficient. However, in recent years, large discrepancies in fuel efficiencies have emerged due to evolving technology. Vehicles operating on alternative fuels avoid gasoline taxes. Hence, a vehicle miles traveled (VMT) tax is superior to a fuel tax as it has the potential to more accurately price road use and damage for all vehicles while also addressing congestion costs on a variable basis.  The authors cite a number of studies relative to distributional impacts of a VMT tax. These other studies show that disproportionate impacts may affect certain populations based upon income level, urban or rural residence, extent of vehicle ownership and type of vehicle. However, the issue is complex, and consensus is not apparent. Although, it appears likely that a VMT tax could be somewhat regressive for low income populations without some type of mitigation. The authors find in their study for Oregon, assuming substitution of a fuel tax with a revenue neutral VMT tax of $0.012/mile, that it would be slightly more regressive for low-income households than the existing $0.24/gal. The VMT tax was not found to be regressive for rural households as they tend to have vehicles with lower mpg ratings that more than offsets the VMT charge (McMullen, Zhang & Nakahara, 2007).

Parry and Small (2005) calculate “second-best” optimal fuel taxes in both Britain and the U.S. to account for the externalities of congestion, accidents and emissions. Of note is that the social costs of road damage, noise, water pollution, vehicle/tire scrappage, and policing are not included but are relatively small in comparison to the costs included in the analysis. The findings are that the U.S. tax would need to be raised from $0.40/gal to $1.01/gal while the U.K. tax would need to be reduced from $2.80/gal to $1.34/gal. Associated welfare gains in such a scenario are 7.4 percent in the U.S. and 22.7 percent in the U.K. The higher U.K. tax is due to assumed larger marginal congestion costs. Fuel taxes are considered “second-best” because they do not account for the largest social cost of congestion with precision. The study accounts for this by also identifying an optimal VMT tax which would be about $0.15 per vehicle mile or comparable to $2.50/gal and $3.00/gal in the U.S. and U.K., respectively. Welfare gains in the U.S. with the VMT charges are almost 4 times that of the optimal fuel tax approach or 28.4 percent. In the U.K., the VMT tax scenario increases welfare by about 21 percent above the fuel tax method and 27.5 percent overall (Perry & Small, 2005).

Research using disaggregated 2009-2013 panel data for Ohio by Langer, Maheshri and Winston (2017) found that a VMT tax is much more economically and distributionally efficient than a fuel tax. The key advantage is that the VMT tax can be varied based on traffic (and pollution) levels on different streets and geographic areas. Overall, there would be a substantial loss in consumer surplus which is more than offset in increased government revenues. Assumptions included a gas tax rate of $0.54/gal vs. a differentiated VMT tax of $0.0575/rural mile and $0.2409/urban mile. The difference in rural and urban VMT charges is due to the larger costs of externalities in urban areas. As a result, a total surplus gain of $10.5 billion can be expected which is almost 20 percent higher than a second-best fuel tax increase.  (Langer, Maheshri and Winston, 2017).

Research by Delucchi (2007), revealed that total gasoline taxes in the U.S. are undercharged by a range of $0.20 - $0.70 per gallon to support road-related capital and operating/maintenance expenditures. If the uncovered social costs of roadway use are considered, about another $1.00/gal. ($1.25 in 2018 $) would need to be added to the price of fuel (Delucchi, 2007). Another study modeled a carbon tax on fuels to maintain CO2 emissions constant at the 1990 level (Jorgenson, D., Slesnick, and Wilcoxen, 1992). The finding is that the tax reduces money metric efficiency (social welfare) but only by about 0.187 percent of total wealth. Gains or losses to equity would range from <±1 percent depending upon the level of effort given to impartiality of the taxes (Jorgenson, D., Slesnick, and Wilcoxen, 1992).

According to a study by the Congressional Budget Office (CBO)(2011), passenger vehicles account for about 90 percent of all miles traveled on U.S. highways and a majority of costs. However, heavy trucks account for less than 10 percent of travel but cause the most pavement damage per mile. Currently, pricing of U.S. roads is not economically efficient as combined fuel taxes only produce about $0.02/mile in revenue from passenger vehicle owners while the costs of externalities or social costs caused by such travel is up to $0.1/mile and higher in urban areas. These costs include pavement damage, congestion, pollution, accidents, and dependence on foreign oil. Oil costs are an externality because users drive up oil demand and prices on one another which increases recession risks related to potential oil price shocks. These costs are most associated with VMT as opposed to consumption of fuel. The most optimal pricing network would be a combination of VMT taxes and fuel taxes. The main downside of such a mechanism is that they would raise transportation costs and the prices of goods while imposing a greater burden on low income households. This inequity would be offset to an extent with VMT taxes as low income households tend to drive less fuel-efficient vehicles (CBO, 2011).

Per the CBO (2011), the two primary problems with fuel taxes are that they cannot account for the variable costs of road congestion and the loss of accountability that occurs with more fuel-efficient vehicles.  VMT fees could rectify this but they impose costs to install and monitor metering devices in all vehicles which would require an extensive implementation period. VMT charges could vary based on the type of road used, time of day, and the extent of congestion. A combination of VMT and fuel taxes would result in greater/maximum economic efficiency or net benefits if several conditions hold true. First, more efficient use of roadway capacity occurs due to users only making trips that meet or exceed their costs. Second, collection costs are minimized and that they result in trivial effects on consumer decisions relative to employment, saving, and vehicle purchases. Finally, only those roadway projects are implemented in which benefits exceed costs. The combined fuel/VMT tax must reflect marginal costs to users or those incremental costs linked to benefits equal to or above the consumption costs for each trip (CBO, 2011). According to a report by the National Cooperative Highway Research Program (2016), estimated operating costs (administration, collection, enforcement) for fuel taxes and VMT fees are 0.92 percent and 6.5 percent of revenues, respectively. The VMT fee initial setup costs would be about 22 percent of revenues over the initiation period.

The CBO (2011) estimates that, absent a VMT charge, combined fuel taxes would need to be about $1.30/gal (2009 $)($1.52 in 2018 $) to maximize efficiency of highway usage. In Illinois during 2018, the combined federal ($0.184) and state ($0.352) was $0.536/gal. Therefore, fuel taxes would need to be raised about $1.00/gal. If efficient VMT charges were employed, the fuel tax rate could be reduced to about $0.20/gal. (2011 $) or $0.22 (2018 $) plus a VMT rate of about $0.13/mile (2011 $) or $0.15/mile (2018 $)(CBO, 2011). Thus, in a VMT charge scenario, a passenger car averaging about 25 miles per gallon would pay an effective tax rate of about $3.95/gal [($0.20 + ($0.15 x 25)]. This would equate to a total fuel cost of about $5.95/gal or about a 138 percent increase (assuming existing gasoline cost of about $2.54/gal. including existing Illinois tax of $0.536/gal. in 2018). Given that the price elasticity of gasoline is at least about -0.3 in the long run (Litman, 2017), total passenger VMT could reasonably be expected to decrease by up to 40 percent or more. Of note is that for a non-VMT tax scenario, the CBO states that the maximum efficiency combined fuel tax rate of $1.30/gal (2009 $)($1.52 in 2018 $) would need to be increased to $2.10 (2009 $)($2.47 in 2018 $) assuming the 2016 MPG standards go into effect. This would equate to a total fuel cost of about $5.00/gal. in Illinois or a 100 percent increase and up to 30 percent or more decrease in VMT. The CBO (2011) estimated that a combination of fuel taxes and varying VMT taxes by time and location could result in combined annual net benefits of $60-$90 billion (2009 $)($70-$106 billion 2018 $) from reduced road maintenance and operations, congestion, and fuel consumption costs.

State Transport Pricing Pilots and Planning

The State of Oregon has operated the OReGO program since 2012 [Oregon Department of Transportation (ODOT), 2018]. OReGO is an initiative that allows a maximum of 5,000 volunteers to pay road usage fees on a per mile basis as opposed to the fuel tax method. In 2018, the fuel tax rate increased from $0.030 to $0.034/gal. and will rise $0.02 per year on the following even-numbered years through 2024. In 2018, the road use charge rate increased from 1.5 to 1.7 cents and will increase to 1.8 cents in 2020 and 1.9 cents in 2022. According to OReGO, the weight-mile tax used for heavy trucks is appropriate due to the significant damage they cause to roads. However, a weight-mile tax on personal passenger vehicles is not appropriate as there is no measurable difference in the damage they cause to roads. Volunteers are invoiced regularly for the miles they drive and receive credit for the fuel taxes they pay at the pump (ODOT, 2018).

In 2017, the State of Washington initiated the Washington Road Usage Charge (WA RUC) Pilot Project [Washington Department of Transportation (WDOT), 2018]. The WA RUC permits about 2,000 drivers to pay a mock $0.024/mile road usage fee for light-duty vehicles in lieu of the $0.494/gal. current state fuel tax. The charge is considered comparable to what the average driver in the state pays with a vehicle averaging 20.5 MPG. There are four mileage-reporting options ranging from manual to high-tech. The program concludes in 2019 with a report to public officials in 2020. Washington is one of 7 states that received federal grants to explore road usage fees (CA, DE, HI, OR, MN, MO)(WDOT, 2018).

The California State Transportation Agency in coordination with the California Department of Transportation (CDOT) concluded a Road Charge Pilot Program in 2017 (CDOT, 2017). The initiative consisted of about 5,000 participants which included urban, suburban and rural private vehicles, commercial vehicles, trucks and out-of-state drivers. Several reporting technologies for users were available ranging from manual to high-tech during a nine-month reporting period. Based on the mix of vehicles in the state, a revenue-neutral rate of $0.018/mile was used. 60 percent opted for the plug-in device of on-board diagnostics (OBD-II) to transmit VMT. Simulated invoices were generated, and mock road charge payments were made via an on-line purse. In surveys, the majority of users reported overall satisfaction with the road charge method and that it was fairer and more equitable than the gas tax framework. California is exploring a pay-at-the-pump road charge system with the hope of reducing administrative costs and perhaps achieving greater public acceptance (CDOT, 2017).

In 2014, the Governor of Washington State created an Executive Order requiring WDOT to compile a long-range plan for reducing the GHG emissions of travel [Federal Highway Administration (FHWA), 2015]. The overall state GHG reduction targets are: 1990 levels by 2020; 25 percent below 1990 levels by 2035; and 50 percent below 1990 levels by 2050. Washington State used FHWA’s Energy and Emissions Reduction Policy Analysis Tool (EERPAT) to estimate GHG reductions in several scenarios. A combined technology approach of vehicle fuel efficiency advances, substantial hybrid/electric vehicle market penetration, increased use of shorter trips by alternative modes, lower carbon content in fuels and electricity generation, and lower levels of light trucks would almost accomplish the 2050 goal. This combined approach could purportedly reach the year 2050 state goal. An alternative approach of VMT/carbon fees of $0.10-$0.25/mile, variable congestion pricing of $0.04-$0.10/mile, 100 percent increase in parking prices, and pay-as-you drive insurance could achieve the 2035 goal, but reductions would level-off and fall short of the 2050 goal. Other travel demand management (TDM) and system optimization policies also result in substantive GHG reductions. Combinations of all of these policies can potentially achieve reductions beyond the established targets (FHWA, 2015).

Transport Pricing Alternatives for Illinois

According to U.S. Census Bureau Annual Survey of State Government Tax Collections data for Illinois, the inflation-adjusted gas tax revenue decreased by about 26 percent from 1994 to 2014 or about $1.75 to $1.29 billion (2014 $)(Vock, 2015). Illinois statewide annual VMT in these two years was 92.44 billion and 105.03 billion, respectively. Thus, the average charge per VMT has dropped from $0.0189 to $0.0123 over this time. Fuel tax pricing could be increased to adjust for inflation. In 2016, there were about 4.5 million vehicles registered in Illinois (Statista, 2018). Thus, the average annual VMT (AVMT) per vehicle is about 20,000. The Illinois population in 2017 was about 12.8 million. Therefore per capita VMT is about 7,222. State AVMT in 2016 was 108.16 billion. For 2018, assuming 108.16 billion and a return to the 1994 inflation adjusted equivalent VMT charge of $0.0189 equals about $2.04 billion in gas tax revenues (2014$) or $2.16 billion (2018 $). Assuming $2.50/gal cost, average of 20 mpg, and the $0.006/VMT state tax increase ($0.0189-$0.0123), this is a $0.12/gal. total fuel price increase or 4.8 percent. Assuming long-run price elasticity of demand at -0.3, this equates to a 1.44 percent decrease in VMT. Thus, the $2.16 billion in revenues could be adjusted downward to $2.093 billion. This is a substantial increase over the $1.35 billion in fuel tax revenue that Illinois collected in 2016 (U.S. Census Bureau, 2017).

Alternatively, a $1.37/gal. total state tax increase could be imposed pursuant to the aforementioned cited Delucchi research and above inflation-adjusted amount ($1.25/gal. + $0.12/gal). This could be considered a second-best pricing approach to a VMT tax but an improvement over the existing mechanism with an intent to cover externalities. Thus, total fuel prices would increase from about $2.50/gal. to $3.87/gal. This represents a 55 percent increase in price which equates to a 17 percent reduction in VMT [(-0.3 elasticity)(PCVMT from 7,222 to 5,994)]. Thus, 108.16 billion VMT in Illinois could be reduced to 89.77 billion VMT. Assuming 20 mpg average, this equates to fuel usage of about 4.5 billion/gals. Multiplying this by $1.722 in total state gas tax revenue ($1.37/gal. + $0.352 existing) equals $7.749 billion in state motor vehicle gas tax revenues. This is about 6 times the existing gas tax revenues and more than triples the inflation-adjusted scenario. Of note is that alternative fuel vehicles would need to be charged comparable rates based upon mileage which could be reported at emissions stations or other mechanisms. Apart from transportation maintenance and operations, the additional $6.5 billion in income could be used to pay down state debt, reduce both corporate and personal income taxes, and as rebates to low income individuals to offset equity concerns.

Upstream/Downstream Emissions/Carbon Pricing Techniques

Motor vehicle crash deaths per 100 million miles traveled in the U.S. declined relatively consistently from a high of about 3.36 in 1980 to a low of about 1.08 from 1975 to 2014. Rates in 2015 and 2016 increased slightly to 1.15 and 1.18, respectively. The annual estimated economic cost of motor vehicle crashes in the U.S. is $242 billion or about 1.3 percent of GDP $18.62 trillion (Insurance Institute for Highway Safety, 2017). In 2017, daily vehicle miles traveled (DVMT) in the Chicago MSA was about 199,366,797  [165,767,891 (IL) + 19,277,000 (Lake, IN) + 6,882,000 (Porter, IN) + 2,273,000 (Jasper, IN) + 675,000 (Newton, IN) + 4,491,906 (Kenosha, WI)] which equates to about 72,768,880,910 annually. Traffic deaths in the Chicago MSA were 625. Thus, the motor vehicle death crash rate was 1.16 per 100 million miles traveled or 6.54 per 100,000 population.  Assuming the aforementioned scenario of full-cost pricing based on the CBO study (2019), a 40-50 percent reduction in VMT for the Chicago MSA equates to 250-312 lives saved a year, an economic value of more than $2.8 billion annually just for the deaths avoided and not including injuries and property damage.

In 2008, British Columbia imposed a tax on fuel usage and carbon emissions at the rate of C$10 per ton of carbon dioxide which gradually increased to C$30 per ton ($24 US) by 2016 or about C$0.07 per liter of fuel (C$0.265/gal. or US$0.196/gal.)(Beaty, Lipsey and Elgie, 2016). The tax is required to be revenue-neutral and is offset by reductions in other taxes. As a result, the province has comparatively low personal and corporate income taxes. Since imposition of the tax, fuel usage has declined 16 percent in British Columbia but has risen by 3 percent throughout the rest of Canada. GDP growth has also been stronger in British Columbia than the rest of Canada over this period (Beaty, Lipsey and Elgie, 2016). Murray and Rivers (2015) also studied effects of British Columbia’s carbon tax and found that emissions have been reduced from 5-15 percent in the province with negligible impacts on the larger economy. Albeit, there are challenges for some emissions-intensive segments that is offset by improvements in other sectors. Many economists support the concept of revenue-neutrality as lowering of other taxes reduces economic distortions of taxation in general and has the potential to enhance GDP growth. While British Columbia’s revenue neutral policy started out in this fashion and provides some protection for low income populations, it has evolved to protect certain industrial sectors which reduces overall cost effectiveness of the charge. Other entities with similar programs are as follows ($US):  Alberta’s emission intensity target, exceedance fee ($12/ton); California cap/trade economywide ($12.21/ton); NE US electric power cap/trade ($6.06/ton); European Union economywide cap/trade ($7.34/ton); France carbon tax on transport/heating fuels ($15.66/ton). Ideally, such taxes are economically efficient if marginal costs of the tax on consumers equals the marginal benefits of emissions reductions (Murray and Rivers, 2015).

As discussed by Murray and Rivers (2015), a primary concern is that the tax will raise prices on some goods while reducing consumer demand and firm output. Nevertheless, a number of studies demonstrate that a modest carbon tax has negligible impacts to the economy and may actually enhance the economy (double dividend hypothesis). One underlying study of the British Columbia carbon tax shows a reduction in household welfare of 0.08 percent that would be 0.13 percent without the revenue-neutral component. Yet, another study shows no statistically significant impacts which is logical given that the carbon tax is only about 5-6 percent of all tax revenue. The tax has not appeared to effect net growth, however, economic benefits from avoided climate change are not addressed in the pertinent studies. British Columbia households in the lowest income decile spend about 10 percent of total income on gasoline while those in the upper one-half spend only about 4 percent. In British Columbia, this was offset by the Low Income Climate Action Tax Credit which in 2011 returned up to about C$116 per adult and C$35 per child for households less than $31,700 and $37,000 for singles and couples, respectively. Further, reductions were available for the first two income brackets up to income of $75,000 which is larger than the rate for higher income households, although those below the first level experienced no reduction. A Northern and Rural Homeowner tax credit was later introduced to mitigate regressive impacts on non-urban populations. Equitable impacts can vary with changes in tax policies over time. Surveys show that public acceptance of the tax has improved over the years, but opposition continues to be highest with rural, male, older and low-middle income populations (Murray and Rivers, 2015).

As discussed by Shaffer (2016), a popular provision of carbon pricing and charging for externalities caused by the use of fossil fuels in transport is to make fees revenue-neutral. In other words, increased prices in fuel are offset by reductions in other government revenues such as sales, property and income taxes. The problem with this approach is that it serves to diminish the primary goal of reducing emissions as it lessens the financial impact on consumers. There will be some reductions in emissions due to the immediate hit to consumers when purchasing gasoline. However, savvy patrons will manage their money wisely, so they can continue driving at or near to previous levels. Another problem is the fluctuating price of oil. When oil prices plummet, the resulting reductions in gasoline prices are an incentive for consumers to drive more which compromises goals for carbon reduction. Rather than returning carbon pricing revenues to citizens, a more economically viable solution is to use funding from the pricing of externalities to help offset and mitigate the impacts of global warming on future generations (Shaffer, 2016). Public spending of these revenues has a larger impact on employment and GDP as opposed to tax cuts. However, some of the funding should be used for a tax credit to address the regressive nature of the charge on low and middle income households (Lee, 2015).

As discussed by Dervis and Foda (2016), oil prices have varied widely in recent decades from $10-$140 a barrel. This pattern provides an opportunity to introduce variable carbon taxes which gradually increases as oil prices decline and vice versa. Asymmetric adjustments weighted towards larger overall net increases in the tax over time would be consistent with methodologies called for to appropriately address climate change. Such a framework effectively weens society off fossil fuels while transferring oil company profits or producer surplus to government revenues while easing the burden on citizens. The key to political acceptance is to initiate the charge when oil prices are low (Dervis and Foda, 2016).

The Energy Innovation and Carbon Dividend Act (EICDA) was introduced in both houses of the U.S. Congress in late 2018. The EICDA is a carbon fee on fossil fuels of $15 per metric ton, increasing annually by $10/MT, on upstream suppliers of coal, oil and gas. Each upstream carbon charge of $10 per metric ton of carbon equates to about $0.11 per gallon of gasoline, $0.06 per therm of natural gas, and $0.009 to a kilowatt hour of coal energy. Energy-generating industries would be encouraged to lower carbon content due to increasing costs. Revenues would be collected by the U.S. Treasury and redistributed to all citizens. Annual redistributions would be about $200-$300 per adult in the first year and rise proportionately in the future along with the increased fees. Payments would be monthly and taxed as income. Administrative costs are capped at 2 percent of revenues. A border carbon adjustment is used on imported goods to protect U.S. manufacturers from unfair competition with firms in other countries that do not have a similar program. The prices of goods would increase in proportion to the amount of carbon generated from their manufacture and distribution with ranges from 0.2 percent for TVs to 1.1 percent for an airplane ticket for each $10/MT increase (Citizen’s Climate Lobby, 2018).

A meta-analysis by the Canada Department of Finance (2008) covered 18 studies of the price elasticity of demand for airline travel in six market segments. These segments are long-haul international business, long-haul international leisure, long-haul domestic business, long-haul domestic leisure, short-haul business and short-haul leisure. Elasticities in the studies range from a low of -0.198 in the first segment to a high of -1.743 in short-haul leisure. Elasticities tend to be much smaller in business than leisure classes. The middle one-half of the estimates for each of the six segments averages out to an elasticity of about -0.96 (Canada Department of Finance, 2008).

A study by Caron, Cohen, Brown, and Reilly (2018), found that a beginning $50/ton carbon tax that rises 5 percent annually would result in a 63 percent reduction in total U.S. GHGs by 2050 which exceeds the 50 percent reduction rate targeted globally. A $25/ton beginning charge increasing at the same rate is also adequate to meet the targeted rate. A 5 percent annual growth rate is necessary to account for economic growth. Almost 75 percent of the efficiency gains are in the power generation sector which will result initially in higher utility rates which then are lowered over time with additional investment in alternative fuels. The transportation sector is second at about 12 percent, the largest of any non-energy sector. They also find that the most efficient use of revenues is to reduce capital income taxes with transfers of 6-8 percent of CO2 tax revenue to low income populations to counter equity concerns. Rebates to households is the least efficient but the most progressive. Without accounting for the benefits of reduced CO2, overall welfare impacts are negligible, between -0.2 to -0.4 percent in the least advantageous recycling scheme of revenues and slightly positive in the most beneficial revenue recycling framework. If CO2 reduction benefits are considered, there is clearly a net welfare benefit in all scenarios. The primary reason for minimal negative welfare impacts is the reduced distortionary capital income taxes which spurs further investment (Caron, Cohen, Brown, and Reilly 2018).

Options for use of carbon pricing revenues include using the income to reduce other taxes in a revenue-neutral fashion, reducing the deficit, and refunds to lower income households to offset the regressive nature of the fees (Carbone, Morganstern, Williams and Burtraw, 2014). These are attractive as they offset the downside in terms of higher energy and bundled goods costs. The largest economic efficiency gains can be achieved by reducing capital taxes, namely corporate taxes and personal income levies on capital gains, dividends and interest. These tend to benefit high-income individuals. A slightly less economically efficient method is reducing labor taxes which has the effect of being most equitable amongst income levels. Rebates to low income households is not economically efficient but does provide equity benefits. Perhaps the best economically efficient framework is using externalities taxes as a down payment on the government deficit with smaller offsets in other tax reductions. While consumers are seemingly worse off in the short run, there are large efficiency gains to reducing the government debt in the near-term rather than later. This method is also more favorable to future generations (Carbone, Morganstern, Williams and Burtraw, 2014).

Marron and Morris (2016) categorize the options for the disposition of carbon taxes as follows:  counter the additional burden to persons, businesses and the overall economy; fund additional efforts to curb greenhouse gases; address the effects of climate change; and support unrelated public priorities. Research is clear that low income populations would experience a disproportionate burden from higher fuel and electricity costs. Reductions in taxes that impose the most distortive impacts consist of capital income, particularly corporate income taxes which are comparatively high in the U.S. Redundant efforts to a carbon tax such as incentives for home solar systems or power sector renewable fuels development may not provide additional benefits worth the costs. Alternatively, fill-in-the-gap policies such as funding for renewable energy/climate change research and promotion of environmentally friendly business practices not addressed by a carbon tax likely would provide more benefits.  Revenues could be used to assist victims of climate change and affected infrastructure due to flooding, wildfires, and sea level rise. Carbon tax revenues could be deposited in the government general fund which allows use for any public purpose. However, public support is thin when transparency is lacking beforehand on the specific use of a tax. The authors recommend a mix of tax offsets targeted primarily towards low income populations and severely burdened business sectors, judicious support of further climate change preventative/ mitigation efforts, and the avoidance of infrastructure projects inconsistent with the goal of reducing GHG’s. Strict revenue-neutrality may be too severe of an approach to achieve policy goals.  (Marron and Morris 2016).  Using a portion of carbon tax revenues to reduce public debt may garner public acceptance if promoted under the guise of reducing burdens to future generations which is consistent with addressing climate change.

An in-depth analysis by the ICMM (2013) shows how 16 governments utilize carbon revenues. These include the advancement of technology to reduce climate impacts and advancement of climate change initiatives in addition to offsets for economic sectors and vulnerable populations impacted by carbon taxes. These offsets can be in the form of refunds to counter higher energy costs and reductions in other income and labor taxes. However, in some cases governments treat carbon taxes as just another revenue source for general expenditures (ICMM, 2013).

Internalizing Crime Costs

Wang and Minor (2002) gauge job accessibility in Cleveland, Ohio census tracts using 1990 data and two GIS-based methods while controlling for spatial autocorrelation. One is a gravity-based job-accessibility index which acknowledges public transportation extended travel times and worker employment competition. The other is a ratio of jobs to resident workers based on sensible commuting times. The gravity-based model was found to be conceptually superior. The findings are consistent inverse associations between gravity-based job-accessibility and crime rates. The associations are stronger for economic crimes than for violent wrongdoings (Wang and Minor, 2002).

Ihlanfeldt (2006) studied the relationship between job opportunity and crime committed by young males in the Atlanta, Georgia region. Neighborhood employment and crime data are used while controlling for time/fixed effects and potential endogeneity of job opportunity. Crime data covers the years 1991 to 1994 in 205 census tracts and employment by place of work for the years 1986 to 1994 in 379 census tracts. The crime tracts studied are in the city of Atlanta (mostly in Fulton County) and the adjacent county of Dekalb, together the center of the region. The jobs data consists of these two counties and the ring counties of Clayton, Cobb, Douglas, Gwinnett, and Rockdale. Relationships are identified for job accessibility in relation to overall crime, violent crime, takings crime and property crime in addition to individual crime categories. Methodologies used consisted of ordinary least squares, random-trend, and an assortment of two-stage least squares models. Job opportunity varies by neighborhood and is increasingly constrained by distance. Spatial mismatch research identifies three mechanisms in which youth employment may be increased by job opportunity:  1) increasing wages net of commuting costs; 2) knowledge of available jobs; and 3) employers’ tendency to not employ workers making long commutes due to impacts on productivity/absenteeism. Economic models show that employment lowers crime by increasing the opportunity costs of criminal activity and potential jail time. Employment also reduces free time for criminal actions, lowering the stress of unemployment, and solidifying obedience to social norms. All of the models show statistically significant inverse relationships between job accessibility and the aforementioned crime categories except for murder. The elasticities are generally within a range of -2 to -4 (1 percent increase in job accessibility equates to a 2-4 percent decline in crime rate (per 1,000 people). The author cites the three primary methods that can be used to address this problem:  lower commuting costs to outlying employment; 2) disseminate information of job opportunities to disadvantaged populations; and 3) reimburse or incentivize employers to locate in poor neighborhoods (Ihlanfeldt, 2006).

               Job Accessibility - Low Estimate.

Based upon these job accessibility-crime studies, the third option above can be implemented by internalizing the total costs of crime to those responsible. The problem of job inaccessibility for residents of poor neighborhoods is caused by the migration of businesses and populations to outlying areas. Therefore, these entities can be monetarily charged for the proportion of the total costs of crime due to job inaccessibility. Research has monetized the total costs of crime by types of infraction (McCollister KE, French MT, & Ganf H. 2010). As an example, these monetary values are used with 2015 Uniform Crime Reporting Statistics for the Chicago Metropolitan Statistical Area (MSA). The total estimated crime costs for the Chicago MSA in 2015 is about $14.7 billion (Federal Bureau of Investigation, 2018). The number of existing low income and rent-assisted apartments are obtained for each of the 14 counties in the Chicago MSA from Affordable Housing Online (2018). Population and housing data are obtained from the U.S. Census (2018). Employment data is obtained from the Bureau of Labor Statistics (2018). Jobs-housing balances or ratios are then calculated for each county of the Chicago MSA. Affordable housing gaps and surpluses are then calculated for each county relative to both manufacturing jobs and total employment ratios.

As can be seen from Appendix Table 1 and Table 2, several of the Chicago collar counties have affordable housing deficiencies, based on the existing spread of those homes in the region. In total, respective 2.14 or 1.42 percent increases are needed in accessibility as measured by existing affordable housing units in these outer Chicago counties to match the proportion of the MSA’s manufacturing jobs or total employment in those areas. Thus, of the existing affordable housing stock, most of the units needed in those collar counties would need to come from the excess in Cook County. Based on the Ihlanfeldt (2006) study above, a conservative job accessibility (as measured by affordable housing availability) elasticity of crime (-2) is then used to calculate cost suballocations. Thus, respective 4.28 or 2.84 percent decreases in crime could be expected if the affordable housing gap is rectified in either of these “low” scenarios based on either manufacturing jobs or total employment ratios by county. This would be a reduction in the annual $14.7 billion Chicago MSA total costs of crime of either about $629.7 million or $417.2 million which is then suballocated to the counties by the amount of affordable housing needed. Those annual costs range up to respective highs of $146 million in Lake County, Illinois or $165 million in DuPage County, Illinois relative to manufacturing jobs and total employment ratios.

Collar counties would then have the option to build the necessary affordable housing or pay these amounts annually to account for the respective amounts of crime. Counties could charge these to residents and businesses via property taxes, sales taxes or other means. These could be made-revenue-neutral by reducing other taxes accordingly. If opting to make payments, the funding could be used to incentivize employers to locate in poor neighborhoods of Cook County to address the job accessibility and housing-employment mismatches. The per household annual costs in this “low” scenario range up to $797 in Kane County or $486 in in DuPage County, relative to manufacturing jobs and total employment ratios, respectively. Alternatively, these “low” estimated crime costs equate to annual costs per person + employee ranging up to $195 in Kane County, Illinois or $110 in DuPage County.

Job Accessibility – High Estimate.

Total crime cost collar county suballocations in a “high” estimate scenario can be computed if consideration is given to the demand of 264,442 affordable housing dwelling units in the Chicago MSA based on data from the National Low Income Housing Coalition (2017) as shown in Appendix Table 1 and Table 2. Consequently, this equates to an increase of about 7.7 percent in job accessibility and an expected 15.4 percent decrease in the crime rate when again assuming a conservative elasticity of -2. This estimate holds true in matching either the manufacturing jobs or total employment county ratios. This would be a reduction in the annual $14.7 billion Chicago MSA total costs of crime of about $2.3 billion which is then suballocated to the counties by the amount of affordable housing needed. Those annual costs range up to a collar county high of $488 million in DuPage County. However, note that Cook County would have an annual crime cost suballocation up to about $995 million due to its high proportional demand of affordable housing. Under the “high” cost scenario, this equates to an annual per household range for the collar counties up to $1,791 in DuPage County and, alternatively, costs per person + employee of up to $438 in Kane County.

               Job Accessibility – Validation/Dissimilarity Index.

Arkell (2017) used cross-sectional linear regression analysis to identify correlations of various independent variables with violent crime rates in 352 U.S. MSAs using most recently available data. One finding is that a one unit increase in the dissimilarity index (unevenness of white-black population spread) is associated with a 1.35 unit increase in the violent crime rate. The research also showed that the average MSA dissimilarity index is 45.1 while the metric was 75.9 in the Chicago MSA. The measure is interpreted as 75.9 of whites in the Chicago MSA would need to move for the two races to be evenly distributed with a dissimilarity index of 0. Changes in proportions of both existing affordable housing and its demand in the Chicago MSA to match employment proportions by county can be used as rough surrogates for the dissimilarity index to reach the 45.1 MSA average and 16.2 low (Prescott, Arizona MSA) from the Arkell (2017) violent crime model. Consequently, ceteris paribus, corresponding  decreases of 30.8 and 59.7 units in the Chicago MSA dissimilarity index may be expected to reduce violent crime rates by about 39 and 80 units or 9.0 and 17.4 percent, respectively. The estimated cost of Chicago MSA violent crime in 2016 was about $13.7 billion or 93 percent of total crime. Thus, the respective potential violent crime reductions cost savings are about $1.2 billion and $2.4 billion. The lower bound amount is about 2-3 times that of the -2 elasticity total crime method above and the upper amount is about the same.

These crime costs are suballocated to collar counties in “low” and “high” estimates for each under both the existing manufacturing job and total employment ratios. Under the manufacturing job ratio methodology, Chicago MSA annual collar county cost suballocations range up to $283 million and $242 million, both for DuPage County, in meeting the average dissimilarity index of 45.1 under the respective existing affordable housing and demand scenarios. Cook County has an annual suballocation of $241 million based on the demand for affordable housing.  To reach a dissimilarity index of 16.2 under the manufacturing jobs ratio methodology, the respective Chicago MSA annual collar county annual cost suballocations range up to about $548 million and $469 million, again, both in DuPage County. Cook County has an annual suballocation of $468 million based on the demand for affordable housing in this case.

Under the total employment ratio methodology, collar county cost suballocations range up to $484 million and up to $264 million, both in DuPage County, in meeting the average dissimilarity index of 45.1 under the respective existing affordable housing and demand scenarios. Cook County has an annual suballocation of $538 million based on the demand for affordable housing. To reach a dissimilarity index of 16.2 under the total employment ratio methodology, the respective Chicago MSA annual collar county annual cost suballocations range up to about $938 million and $512 million, again, both in DuPage County. Cook County has an annual suballocation of about $1 billion based on the demand for affordable housing in this case.

               Land Development.

Research by Voith (2001) quantified the extent of demand for residential land based upon changes in price. The U.S. has experienced tremendous population disbursement in the post-World War II period due to extensive investment in the transportation system, namely construction of the interstate highway system. Consequently, the amount of residential land opened up for development at low prices has increased allowing for larger lots. Zoning regulations, mortgage interest tax deductions, and the sheer volume of vacant land have also facilitated the trend. Voith (2001) cites research that the federal tax mortgage reduction effectively lowers the price of land by about 12 percent, increasing land consumption by about 19.2 percent, which lowers residential density by about 16.1 percent. One study cited by Voith (2001) finds that changes in income do not have a large impact on the demand for land. A 1 percent increase in income equates to a 0.4 percent increase in land consumption expenditures. Another estimates that a 1 percent increase in land price reduces the amount of land used by 0.75 percent. Thus, Voith (2001) surmises that increases in the price of land have the potential to significantly alter urban land use patterns while minimizing any counter effects on consumer satisfaction. The reason is that consumers will buy smaller lots/homes while using any money saved for other equally attractive uses at lower relative prices when the price elasticity of demand for land is high. If the elasticity was low, consumer benefits would suffer much more. In his study, Voith (2001) applies the Bartik-Epple procedure to estimate the price elasticity of demand for residential land. The data set consisted of home sales in Montgomery, Pennsylvania covering transactions of almost 100,000 single-family detached houses during the period of 1972-1997. Statistical models were then constructed to estimate the prices of only the land while controlling for characteristics of the homes and neighborhoods. The finding is that a 1 percent increase in the price of land reduces land consumption by 1.6 percent. Voith (2001) concludes with the prevailing thought that the American taste for large lots and open space is probably much less pronounced while the low price of residential land is the predominant factor.

Song and Zenou (2006) analyze property tax rates for 448 U.S. urbanized areas and find that there are two countervailing factors. An increase in property taxes has a negative impact on developer profits causing more investment in land which lowers structural improvements. This tends to reduce population density and increase geographic expansion (building height effect). Yet, increased property taxes  leads to higher prices, smaller homes and higher population densities (dwelling size effect). The latter is more important as consumption of housing and composite goods are very interchangeable with elasticities greater than 1.  Therefore, increases in property taxes lowers the utility or usefulness of housing  population which increases population density. The finding is that a 1 percent increase in property taxes reduces urbanized areas by 0.4 percent (Song & Zenou, 2006). The study provides further confirmation on how reduced land prices, facilitated by either lower real estate taxes or the federal mortgage deduction, increases lower density land consumption in exurban areas.  

​

According to Demographia (2018), the Chicago urbanized area (UA) expanded from 1,585 square miles (1,014,400 acres) in 1990 to 2,123 in 2000 and, 2,443 (1,563,520 acres) in 2010. Thus, the Chicago UA increased by 549,120 acres during the 20-year period. Assuming a conventional market rate of urban land at $500,000 per acre this equates to total land value of about $274 billion for the 549,120 new urbanized acres (Albouy, Ehrlich & Shin, 2018). Assuming that 15 percent of total crime can be attributed to jobs-housing mismatch (or similarly job inaccessibility or dissimilarity index), this equates in Chicago to about $14.7 billion annual cost of crime X 20 years = $294 billion X 15 percent = $44.1 billion. $44.1 billion/549,120 acres = $80,310 charge per acre. This is about a 16 percent increase in the price of land which conservatively could have reduced the number of acres developed by about 16 percent or 87,859 acres (assuming price elasticity of demand for land -1), or mid-range by about 25.6 percent or 140,575 acres (-1.6 elasticity), or liberally by 32 percent or 175,718 acres (-2 elasticity).

Analysis by Litman (2015) quantifies the economic costs of low density land development in the U.S. at more than $1 trillion annually, consisting of $400 billion external and $625 billion internal. Internal costs are experienced by those residents living in newer dispersed communities while external costs are imposed by those same residents on others. More specifically, the study provides economically optimal urban form for geographically un-constrained, semi-constrained and constrained localities.  Optimal ranges across these types are as follows:  regional population density per acre of 8 to 32+; auto mode shares of 50 to <10 percent; vehicle ownership rates per 1,000 residents of 400 to <200. As a comparison, based on Census data the Chicago UA population density in 2010 was 5.5 (8,608,208 residents/1,563,250 acres); personal vehicle commute mode share is about 80 percent; and vehicle ownership rates per 1,000 residents is about 750 (Census, 2018).  Litman (2015) also analyzes rates of traffic fatalities per 100,000 residents for major metropolitan areas. The results are that rates tend to be 10-20 in affluent/ auto-dependent cities, 5-10 in affluent/compact cities, and 1.5-3 in affluent/compact cities with viable transport demand management programs. According to the Centers for Disease Control and Prevention (2012), the Chicago MSA traffic fatality rate was 5.9 per 100,000 population. Litman (2015) further calculates the annual average cost of externalities per vehicle at about $4,700 in the U.S. during 2014. Considering the total U.S. 260 million registered vehicles, this equates to about $1.22 trillion in total vehicle-related externalities. Given further that total U.S. VMT in 2014 was about 3 trillion, this translates to about $0.40 in externalities per mile or about $8.00/gal. assuming a 20 mpg average (Litman, 2015).

As documented by Tanguay and Gingras (2011), economics research has identified the following variables that explain low-density land development:  population, income, transport costs, agricultural rents, industrial locations, climate and topography. The authors use multivariate analysis to identify the effect of gasoline prices on urban sprawl in the 12 largest Canadian metropolitan areas covering 1986-2006. A primary finding is that a 1 percent increase in gasoline prices causes a 60 percent decrease in low-density housing and a 0.32 percent increase of population living in the inner city. (Tanguay and Gingras, 2011). McGibany (2004) analyzed data covering the latter one-half of the 1980’s in 40 small to mid-sized U.S. metropolitan locations to determine the association between gasoline prices/taxes on the geographic size of urbanized areas. The main finding is that higher gasoline prices, mainly driven by a $0.01 increase in the state gasoline tax, resulted in an average reduction in the amount of urbanized land by about 4.7 square miles (McGibany, 2004). McGrath (2005) analyzes the 33 largest U.S. metropolitan areas using data from the decennial U.S. censuses from 1950 through 1990 to identify independent variables associated with the dependent variable of urbanized area geographic size. Substantive findings are that a one percent increase in transportation costs reduces urbanized land area by an average of 0.28 percent. Further, optimal transport pricing would have reduced the sizes of urbanized areas by up to an average of 12 percent (McGrath, 2005).

Ewing, Schieber and Zegeer (2003) analyze county-level data in 48 U.S. counties for 101 of the largest metropolitan areas using a four component sprawl index:  low-density residential development; highly segregated homes and businesses; a lack of strong downtowns; and road systems characterized by large blocks and poor access from place to place. The findings are that a 1 percent increase in the index (more compactness) equates to declines in traffic and pedestrian fatality rates of 1.49 percent and 1.47-3.56 percent, respectively (Ewing, Schieber and Zegeer 2003).

Conclusion

U.S. housing and transportation market inefficiencies and subsidies have contributed substantially to racial/class segregation, excessive transportation demand, and crime in addition to unsustainable/ premature land development and environmental degradation. Scientific evidence is clear that human activities are more likely than not a significant factor in climate change consisting of rising average temperatures, ecosystem alterations, and extreme weather events such as flooding, drought, and wildfires. Accordingly, worldwide greenhouse gas emission reductions of up to 50 percent lower than 1990 levels by 2050 have been targeted through multi-national agreement. Upstream carbon-related charges to coal, gasoline and oil producers offers perhaps the easiest method for capturing environmental costs in a market-based fashion throughout the economy while minimizing administration expenses.

CO2 charges starting at $15/MT and rising by $10/MT annually, $50/MT or $25/MT and rising by 5 percent annually all have good potential in meeting global carbon reduction goals. It is critical that these fees are offset with reductions elsewhere in the tax system to minimize impacts on consumers. Upstream energy charges have been modeled to show that initial increases in energy rates for natural gas and electricity will be followed by reductions as the market adjusts through expansion of renewable energy such as solar, wind and hydro power. There will be at least some initial and long term increased costs to businesses and for consumer goods in sum, particularly those that are energy-intensive. However, these will largely be countered as new economywide efficiencies gain a foothold, consumers make adjustments in spending patterns, and distortionary taxes are reduced. It is most efficient if carbon pricing revenues are offset by capital investment/dividends tax reductions both at corporate and personal levels. Other appropriate use of revenues includes reducing government debt and income taxes. Such offsets should be proportionately more beneficial to lower income populations as they experience greater burdens from rising energy costs. Tax credits are more efficient than rebates and have less administrative costs. The revenues should also be used to mitigate natural disasters due to climate change and initiatives to further the renewable energy industry.

Transport infrastructure planning suffers perpetually from the “theory of the second best” because decision-making occurs primarily based on the demand for travel which is derived from an improperly priced market. Consequently, gasoline prices would need to rise between about a low of $1.25/gal. to a high of $8.00/gal just to cover social costs. Each annual $10/MT upstream CO2 charge would increase gasoline by about $0.11/gal. or a total increase of about $2.50/gal. over 24 years, effectively doubling the price of gasoline which would reduce average PCVMT from about 10,000 to 7,000 in the last year based on long run fuel price elasticity of transport demand of -0.3. Transportation emissions, accidents, deaths, and injuries would be by about the same proportions. This would meet the CBO’s estimate for coverage of all externalities, absent the use of a VMT charge methodology. The $10/MT upstream charge would increase the cost of a typical airline ticket by about 1.1 percent annually resulting in about a 27 percent price increase and comparable travel decrease after the 24-year period given average airline travel price elasticity of demand of about -1. Absent a universal upstream CO2 charge or a VMT fee, it would be appropriate to implement a comparable charge at the pump. However, a VMT tax is up to four times more efficient than a per gallon tax due to its ability to account for congestion. Per the CBO, a charge of $0.22/gal plus a VMT rate of $0.15 would be appropriate. Implementation costs would be about 22 percent of revenues at the outset and 6.5 percent thereafter. Expectations for a typical vehicle averaging 25 mpg are a total increase of about $3.95/gal, raising the price of gasoline by about 138 percent. Comparable offsets in the overall tax system similar to the upstream CO2 pricing scenario would also be necessary. In all cases, variable rates are necessary to establish a floor when market prices are low to ensure appropriate CO2 reductions.

Upstream carbon pricing does not address housing-employment mismatches which exacerbate crime, nor does it counter the costs of traffic congestion. One study demonstrates that the job accessibility elasticity of crime is up to -4. Balanced affordable housing and employment by county within urban regions is a proxy for accessibility. Using a conservative elasticity of -2, it is reasonable to impose charges to suburban/exurban areas lacking sufficient affordable housing to account for up to about 15 percent of crime costs to be used as economic development incentives in the urban core. Such transfers in the Chicago region would total up to $2.4 billion annually which would range up to a high in DuPage County of about $1,500 per household or $340 per resident and employee. Alternatively, based on price elasticity of land development at -1.6, placing the annual charge in the Chicago region on exurban development likely would have prevented up to 26 percent or about 220 square miles of newly urbanized land from 1990-2010. The size of U.S. urbanized areas could be as much as 47 percent smaller geographically if the following policies had been implemented to address economic distortions:  optimal transport prices (0.12), crime cost internalization (0.16), and elimination of the mortgage interest deduction (0.19).  

​

​