This page describes the calculation of the spending and revenue effects of the policy proposals in Fiscal Therapy, by William G. Gale[1].

Chapter 7: Reforming Health Care

The policies proposed in this chapter include:

  • Reinstating the individual mandate penalty
    • I begin with estimates for 2021-2027 in Congressional Budget Office (2017d) for eliminating the individual mandate. In 2021, for example, these estimates show that eliminating the mandate reduces spending by $30 billion and increases revenue by $3 billion. Multiplying these estimates by -1 provides a rough estimate for the cost of reinstating the mandate in each year. However, Congressional Budget Office (2018b) implies that the original estimates should be reduced by one-third, on average. To account for this change, I multiply each estimate by two-thirds. Each resulting dollar estimate is then converted to a share of GDP, using the baseline outlined in chapter 3. After 2027, revenue and cost estimates grow with major federal health costs through 2050 (Medicare, Medicaid, CHIP, and exchange subsidies), as outlined in chapter 3.
  • Appropriate CSR payments
    • Hall (2018) estimates that appropriating CSR payments from 2019-2021 would reduce the deficit by $32 billion from 2019-2027. I assume that the entire savings accrue through 2021, the years in which CSR payments are appropriated in the estimate. The $32 billion is split proportionally among years 2019-2021 according to projected spending on Medicaid, CHIP, and exchange subsidies, as outlined in chapter 3. After 2021, net deficit reduction grows through 2050 with Medicaid, CHIP, and exchange subsidy spending. The spending portion of the estimate comes from Congressional Budget Office (2017b), which provides estimates of cost sharing outlays through 2027. After 2027, the spending change associated with the policy grows with Medicaid, CHIP, and exchange subsidy costs. Revenue changes equal the difference between the spending increase and the net deficit reduction in each year. All estimates are converted to a share of GDP according to the baseline described in chapter 3.
  • Fix the family glitch
    • Buettgens, Dubay, and Kenney (2016) estimate this policy would require $6.5 billion of funding in 2016. I grow the nominal spending requirement at the same rate of major federal health costs through 2050 as outlined in chapter 3 and divide the amounts by baseline GDP to obtain the required increase in spending from the policy in 2021-2050.
  • Create a public option in the insurance marketplace
    • On the spending side, I begin with nominal savings estimates from Congressional Budget Office (2013a, Health Option 2) for 2016-2023 and calculate the associated share of major federal health costs (e.g. 0.5 percent), using Congressional Budget Office (2013b). I then shift the cost savings shares by five years so the implied CBO (2013a) share for 2016 is applied to 2021. After 2028, I assume the policy reduces major federal health spending by about 0.4 percent, which is the implied cost reduction in 2028. I then multiply the reduction share by projected major health costs in each year through 2050, following the baseline in chapter 3, and divide by GDP to obtain the policy’s effect as a percent of GDP.
    • On the revenue side, I begin with nominal revenue estimates from Congressional Budget Office (2013a, Health Option 2) for 2016-2023 and calculate the associated percent of GDP using Congressional Budget Office (2013b). I then shift the revenue estimates as a share of GDP by five years so the CBO (2013a) share for 2016 is applied to 2021. After 2028, I assume the policy raises revenue by 0.08 percent of GDP, which is the share in 2028.
  • Transition to a bundled payment system for Medicare
    • I begin with nominal savings estimates from Congressional Budget Office (2013a, Health Option 10) for 2017-2023 and calculate the associated share of Medicare costs (e.g. 0.5 percent), using Congressional Budget Office (2013b). I then shift the cost savings shares by four years so the implied CBO (2013a) share for 2017 is applied to 2021. After 2027, I assume the policy reduces Medicare spending by about 1 percent, which is the implied cost reduction in 2027. I then multiply the reduction share by projected Medicare costs in each year through 2050, following the baseline in chapter 3, and divide by GDP to obtain the policy’s effect as a percent of GDP.
  • Allow Medicare to have the same negotiation power for prescriptions as the Veterans Administration
    • Gagnon and Wolfe (2015) estimate that the policy would have saved $15.2 billion in 2010. Using Centers for Medicare and Medicaid Services (2018, Table 11), I calculated that the $15.2 billion is equivalent to about 26 percent of Medicare drug costs in 2010.
    • I apply the 26 percent savings share to projected Medicare drug costs through 2050. For 2021-2026, I take projected Medicare drug costs from Centers for Medicare and Medicaid Services (2018, Table 11). After 2026, the ratio of drug costs in (t) relative to (t-1) equals the ratio of costs in 2026 compared to 2025 in Center for Medicare and Medicaid Services (2018, Table 11).
    • I then convert the savings amounts to percentages of GDP using the baseline in chapter 3.
  • Introduce a premium support system for Medicare
    • Congressional Budget Office (2017c) provides nominal estimates of the policy from 2022-2026. Annual savings amounts are converted to shares of Medicare costs, using the baseline outlined in chapter 3. For 2021, I assume the policy reduces Medicare costs by 1.4 percent, equal to the implied share in 2022. After 2026, I assume that the policy reduces baseline Medicare costs by 3.9 percent, which is the implied savings in 2026. I then convert the spending reductions to shares of GDP, following the baseline in chapter 3.
  • Limit the tax expenditure for employer-sponsored insurance (ESI)
    • Congressional Budget Office (2016a, Health Option 18) provides spending and revenue estimates of the policy through 2026. After 2026, I allow the estimates to grow with major federal health expenditures, as outlined in chapter 3.
    • This policy would replace the Cadillac tax, which is not included in the chapter 3 baseline, but is included in the Congressional Budget Office (2016a) baseline. Thus, in order to calculate the true revenue effect of the ESI policy, I add estimates of the Cadillac tax through 2028 from Congressional Budget Office (2018a) to the revenue estimate in Congressional Budget Office (2016a). After 2028, for purposes of this calculation, I assume Cadillac tax revenue grows with major federal health expenditures, as outlined in chapter 3.
    • I then convert the policy’s adjusted spending and revenue estimates through 2050 to shares of GDP using the baseline in chapter 3.
  • Expand Medicaid
    • Dorn and Buettgens (2017, Table 2) estimate the policy would cost $427.5 billion from 2018-2027. I allocate annual amounts in proportion to major federal health costs in each year, using the baseline outlined in chapter 3. After 2027, I assume the policy’s cost grows with major federal health expenditures. I then divide nominal costs in 2021-2050 by GDP, following the baseline in chapter 3, to obtain the policy’s funding requirement as a share of GDP.
  • Increase and standardize the excise tax on alcoholic beverages
    • Congressional Budget Office (2016a, revenue option 38) provides nominal revenue estimates through 2026. Estimates from 2021-2026 are converted to shares of GDP using the baseline outlined in chapter 3. After 2026, revenue estimates for the policy as a share of GDP in (t) compared to (t-1) decline following the ratio of 2026 revenue to 2025 revenue.

These policies interact with each other in numerous ways that are not addressed in Fiscal Therapy. Some interactions would reduce cost savings; others would increase cost savings.  For example, the fiscal cost of creating a public option and expanding Medicaid would be either smaller or larger than the fiscal costs of pursuing each policy separately, depending on which of two interactions dominates. First, because some people who would qualify for expanded Medicaid would not enroll in the public option, the net fiscal benefits of the public option are overestimated in my original calculations. Second, the original calculations overstate the costs of Medicaid expansions, which would be lower than estimated if some beneficiaries who qualify for expanded Medicaid instead enroll in the public option.

Chapter 8: Saving Social Security

This chapter proposes several changes to Social Security, including:

  • Raise the retirement age
  • Change the benefit formula
  • Raise payroll tax cap
  • Add new bend point
  • Raise minimum benefit level
  • Raise survivor benefits
  • Reinstate benefits for college-aged children
  • Limit spousal benefits for high income households
  • Increase taxation of benefits for high income households
  • Use chained CPI-U
  • Raise payroll tax rate
  • Change treatment of government workers

The revenue and spending changes are estimated as a package. Goss (2016, Table 1c) provides annual estimates from 2016-2090 of the proposal’s effect on expenditures and non-interest income as a share of GDP. I shift the estimates by three years so the first year with deficit reduction is 2021. For example, as estimated by Goss (2016), if started in 2018, the package would reduce program expenditures in 2030 by 0.27 percent of GDP and increase program revenues in 2030 by 0.57 percent of GDP. The estimates in chapter 8 assume that those estimates are delayed three years to 2033.

Chapter 9: Investing in People

The policy proposed in this chapter simply increases social spending by 1 percent of GDP per year. I add 1 percent of GDP of non-interest spending to the baseline each year from 2021-2050.

Chapter 10: Investing for Growth and Security

The policy proposal in this chapter has three main components:

  • Increase infrastructure spending
  • Increase R&D spending
  • Increase defense spending

Increase infrastructure spending:

  • I start with the total needs for each type of infrastructure in constant 2015 dollars from 2016-2040 according to American Society of Civil Engineers (2016) and assume that total is the need through 2040. Thus, I shift the starting date by five years to 2021 and account for spending that would have occurred before 2021. Baseline spending projections from 2016-2020 across all levels of government are subtracted from each infrastructure category’s funding need, where state and local infrastructure spending is assumed to be three times federal spending (Congressional Budget Office 2015). I assume that federal infrastructure spending in the baseline represents about 17 percent of non-defense discretionary spending, equal to its historical average share (Congressional Budget Office 2015). In order to assign half of the funding responsibilities to the federal government, 50 percent of the remaining need is distributed such that the resulting shares of GDP from 2021-2040 are equal. However, since the funding gap is published in real 2015 dollars and baseline GDP is in nominal dollars, I convert GDP to real dollars using a mix of the employment cost index and GDP price index (Congressional Budget Office 2018a). The resulting percent of GDP spending goal is then continued through 2050.
  • The proposal provides enough funding for the federal government to meet the aforementioned constant share of GDP spending goal. It represents the difference between the spending goal and baseline infrastructure spending, where baseline infrastructure spending is roughly 17 percent of non-defense discretionary spending, as outlined in chapter 3.

Increase R&D spending:

  • The R&D proposal doubles non-defense R&D spending as a share of GDP relative to the baseline in 2021 and keeps funding at that share of GDP through 2050 (about 0.7 percent of GDP). Baseline non-defense R&D spending is calculated by multiplying baseline non-defense discretionary spending, as outlined in chapter 3, by 10.7 percent, which equals the average share of non-defense R&D spending relative to total non-defense discretionary spending from 1985-2014 (Office of Management and Budget 2018, Table 9.7). The effect of the policy in each year is the 0.7 percent of GDP spending goal less the R&D spending estimate in the baseline for that year.
  • Additionally, the R&D proposal eliminates the provision in the Tax Cuts and Jobs Act that requires firms to amortize their R&D expenditures starting in 2022. For 2022-2027, nominal revenue estimates equal the estimates of the amortization provision from Joint Committee on Taxation (2017). The estimates are then converted to percentages of GDP following the baseline in chapter 3. Since the policy’s effect shrinks as a share of GDP over time, the nominal annual revenue estimate is held constant after 2026 at $6.3 billion, which is equal to the policy’s effect in 2026.

Increase defense spending:

  • The proposal funds the base defense budget from President Obama’s final Future Year Defense Program (FYDP), as extended by the Congressional Budget Office (2017a). I begin with annual budget authority estimates under the FYDP in real dollars from 2021-2032 and multiply them by 1.03 since the Congressional Budget Office (2017a) projects that costs would be about 3 percent higher on average through 2032 if several broad areas were to experience growth similar to that in the past. After 2032, I allow base spending to grow 1 percent faster than inflation.
  • I then add overseas contingency operations and non-DOD defense funding (such as nuclear programs in the Department of Energy) through 2050. In real 2017 dollars, OCOs are assumed to be constant at about $68 billion per year, while non-DOD-related funding is assumed to be $27 billion per year, consistent with recent spending (Congressional Budget Office 2017a; Daniels and Harrison 2018; Harrison 2016; Office of Management and Budget 2018).
  • However, the policy effects in Fiscal Therapy require estimates of outlays, not budget authority. To account for the difference between the two measures of spending, the aforementioned total budget authority is adjusted downward. Through 2028, the adjustment follows Congressional Budget Office (2018a); afterward, I hold the adjustment constant at the 2028 level of roughly 3 percent.
  • I then convert the real defense outlay goals to percentages of GDP, using a mix of the employment cost index and the GDP price index to measure inflation (Congressional Budget Office 2018a).
  • The final defense proposal represents the difference between the outlay goals as a percent of GDP less baseline defense spending through 2050, as outlined in chapter 3.

Chapter 11: Personal Taxes

This chapter proposes several changes to personal taxation, including:

  • Tax capital gains at death and increase the top capital gains rate to 24.2 percent
    • This policy is based off the proposal in Office of Management and Budget (2016) to “reform the taxation of capital income.” Nominal revenue estimates for 2021-2026 come from Office and Management and Budget (2016) and are converted to shares of GDP using the baseline in chapter 3. After 2026, revenue estimates as a share of GDP grow with baseline income tax revenue, as outlined in chapter 3.
  • Raise the lower capital gains tax rate to 20 percent
    • McClelland (2017) estimates this policy would increase revenue by $50 billion from 2018-2027. From the $50 billion, I allocate annual revenue estimates proportionally to each year’s share of cumulative GDP over the decade. I then convert the revenue estimates to shares of GDP and shift each estimate by three years to make the policy start in 2021. After 2030, revenue estimates as a share of GDP grow with baseline income tax revenue, as outlined in chapter 3.
  • Tax like-kind exchanges
    • The policy assumes complete elimination of the individual income tax expenditure for like-kind exchanges. I start with tax expenditure estimates from Joint Committee on Taxation (2018) from 2018-2021. I then divide by GDP and shift the estimates by three years so the starting year of the policy is 2021. After 2024, policy estimates as a share of GDP grow with baseline income tax revenues, as outlined in chapter 3.
  • Treat carried interest as ordinary income
    • Congressional Budget Office (2016a, revenue option 12) provides nominal revenue estimates from 2021-2026. Since the resulting estimates decline as a share of GDP over time, after 2026, the ratio between the revenue estimate in (t) relative to (t-1) as a share of GDP equals the ratio of the revenue estimate in 2026 versus 2025.
  • Eliminate $10 billion of various loopholes
    • I convert $10 billion to a share of GDP in 2021 using the baseline in chapter 3 and then hold the revenue estimate constant at that percent of GDP through 2050.
  • Phase-out the mortgage interest deduction
    • This policy eliminates 10 percentage points of baseline mortgage interest deduction (MID) costs each year until the deduction is completely eliminated. After the deduction is eliminated, the annual revenue estimate is the baseline cost of the deduction in that year.
    • To estimate baseline MID costs, I begin with estimates of the deduction from Joint Committee on Taxation (2018) from 2018-2021. I then reduce the estimates by 10 percent, since investors would change their portfolios if the deduction were eliminated. Drukker, Gayer, and Rosen (2017) estimate that the rebalancing-adjusted revenue loss is about 90 percent of the conventionally estimated revenue loss. The estimates are then converted to shares of GDP and shifted by three years to make the policy start in 2021. After 2024, MID costs grow with baseline income tax revenue, as outlined in chapter 3.
  • Create a homeowner credit
    • I multiply the proposed $10,000 credit by the estimated two million first-time homeowners in 2016 (Genworth Mortgage Insurance 2017) and divide the result by 2021 GDP, following the baseline in chapter 3. After 2021, I hold the annual cost constant at that percent of GDP.
  • Restore pre-TCJA income tax rates and brackets
    • For 2021-2028, the annual revenue estimate equals the sum of (a) the estimate of the income tax rate and bracket provisions in the Tax Cut and Jobs Act as written from Joint Committee on Taxation (2017) and (b) the estimate of extending the provisions past 2025 from Congressional Budget Office (2018a). After 2028, the revenue estimate as a share of GDP grows with baseline individual income taxes as outlined in chapter 3.
  • Increase IRS funding to bolster enforcement
    • This policy is the IRS tax enforcement proposal from Office of Management and Budget (2016). For 2021-2026, nominal outlay and revenue estimates come from Table 11-2 in the report. Estimates are then converted to shares of GDP following the baseline in chapter 3. After 2026, revenue and outlay estimates are held constant as a share of GDP at their respective 2026 level.
  • Convert the estate tax to an inheritance tax
    • Batchelder (2016) estimates that the policy would increase revenue by $199 billion from 2016-2026. I allocate annual revenue estimates in proportion to estate and gift tax projections from 2016-2026 in Congressional Budget Office (2016b). I then convert the estimates to percentages of GDP and shift the results by five years to make the policy begin in 2021.
    • However, the Batchelder (2016) estimates were based on a pre-TCJA baseline. Thus, from 2021-2028, I add to the shifted Batchelder (2016) estimates the sum of (a) the estimate of the estate tax provision in the Tax Cut and Jobs Act as written from Joint Committee on Taxation (2017) and (b) the estimate of extending the provision from Congressional Budget Office (2018a) since the baseline in chapter 3 permanently extends TCJA provisions.
    • After 2028, the revenue estimate as a share of GDP is held constant at its 2028 level.

Chapter 12: Taxing Business

This chapter proposes several changes to business taxation, including:

  • Tax like-kind exchanges
    • The policy assumes complete elimination of the corporate tax expenditure for like-kind exchanges. I start with tax expenditure estimates from Joint Committee on Taxation (2018) from 2018-2021. I then divide by GDP and shift the estimates by three years so the starting year of the policy is 2021. After 2024, policy estimates as a share of GDP grow with baseline corporate tax revenues, as outlined in chapter 3.
  • Close the Gingrich-Edwards loophole
    • The policy is based on Revenue Option 23 in Congressional Budget Office (2016a) to “tax all pass-through business owners under SECA and impose a material participation standard.” For 2021-2026, nominal revenue estimates come from CBO (2016a). I then calculate the corresponding shares of GDP using the baseline in chapter 3. After 2026, revenue estimates as a share of GDP grow with Social Security revenue as a share of GDP as outlined in chapter 8 and reported in Goss (2016).
  • Eliminate the pass-through deduction
    • For 2021-2028, the annual revenue estimate equals the sum of (a) the estimate of the pass-through provision as written from Joint Committee on Taxation (2017) and (b) the estimate of extending the provision past 2025 from Congressional Budget Office (2018a). After 2028, the revenue estimate as a share of GDP grows with baseline individual income taxes as outlined in chapter 3.
  • Eliminate FDII provisions
    • For 2021-2027, the revenue estimate equals the size of the foreign-derived intangible income provision in each year from Joint Committee on Taxation (2017), converted to a share of GDP according to the baseline in chapter 3. After 2027, the revenue estimate grows with baseline corporate tax revenues as a share of GDP, as outlined in chapter 3.
  • Increase the corporate tax rate to 25 percent, allow expensing of all investment, and eliminate interest deductions
    • The revenue estimate of the policy is the sum of (a) the estimate of raising the rate of the current system to 25 percent and (b) the estimate of converting to a cash flow tax system.[2]
      • For part (a), I simply assume the increase in revenue as 0.04/0.21 of baseline corporate tax revenues as a share of GDP as outlined in chapter 3, where 0.04 equals the difference between the proposed 25 percent rate and the current 21 percent rate.
      • The approach in part (b) proceeds in several steps.
        • First, I calculate the revenue change of moving from pre-TCJA law to a 21 percent cash flow tax. I start with revenue estimates of expensing all investment and disallowing a deduction for net interest expense on new loans under a 20 percent rate from 2021-2036, relative to a pre-TCJA baseline, in accordance with Nunns et al. (2016). I then multiply the revenue change by 21/20 and convert the estimates to shares of GDP using the baseline in Nunns et al. (2016). After 2036, the ratio of the revenue estimate as a share of GDP between (t) and (t-1) equals the ratio between the 2035 and 2036 revenue estimate.
        • I then calculate the effect of expensing equipment and limiting interest deductions in the Tax Cuts and Jobs Act, using the sum of these provisions’ revenue estimates from 2021-2027 as reported in Joint Committee on Taxation (2017) and extended in Congressional Budget Office (2018a). After 2027, the revenue estimates as a share of GDP grow with baseline corporate income taxes as outlined in chapter 3.
        • The difference between these two calculations is scaled up by a ratio of 25/21 to simulate the effect of switching systems with a 25 percent rate.

Chapter 13: Taxing Consumption

This chapter proposes a broad-based 5 percent tax-exclusive VAT with a rebate starting in 2021, with the rate increasing by one percentage point each year until it reaches 10 percent in 2026. After 2026, the VAT rate remains at 10 percent.

Toder, Nunns, and Rosenberg (2012, Table 6) provides nominal revenue estimates for a 7.7 percent tax-exclusive broad-based VAT with a rebate in 2015. Estimates include gross VAT revenue, revenue offsets, rebate cost, associated reductions in federal spending, and the net change in the federal deficit. Toder, Nunns, and Rosenberg (2012, Table 4) shows that GDP in 2015 in the analysis is estimated to be $18.662 trillion.

I first convert each revenue and spending estimate for the 7.7 percent VAT to shares of GDP. I then scale each estimate up or down according to the 5-10 percent rate schedule under my proposal. I scale each estimate relative to its “tax-inclusive rate.” For example, a 7.7 percent tax-exclusive rate corresponds to a 7.15 percent tax-inclusive rate, and a 10 percent tax-exclusive rate corresponds to a 9.09 percent tax-inclusive rate.

Using that procedure, a 10 percent VAT yields gross revenue equal to 3.66 percent of GDP. Revenue offsets equal 1.34 percent of GDP. Rebates equal 1.29 percent of GDP. The net change in revenue gross revenue less offsets and rebates (1.03 = 3.66-1.34-1.29).  Reductions in spending equal 1.55 percent of GDP. The net change in the primary deficit is 2.57 percent of GDP (= 1.03 + 1.55, with rounding).

Chapter 14: Improving the Environment

This chapter proposes a carbon tax beginning in 2021 at $30 per ton of carbon emissions. In each subsequent year through 2050, the rate rises by 5 percent above inflation.

I start with nominal gross revenue estimates from McKibbin et al. (2018) for an identical policy that runs from 2020-2040. I then convert the revenue estimates to shares of GDP, according to the baseline in chapter 3. To simulate the effect of the policy starting in 2021 instead of 2020, I shift each gross revenue estimate as a share of GDP downward by one year. For example, I shift the McKibbin et al. (2018) gross revenue estimate for 2040 by one year to 2041.

After 2041, gross carbon tax revenue as a share of GDP grows at the rate equal to the average increase in gross carbon tax revenue as a share of GDP from 2036-2041. Thus, gross carbon tax revenue as a share of GDP increases by roughly 2.2 percent (not percentage points) each year from 2041-2050.

To calculate net carbon tax revenue, I multiply gross revenue by 0.6375 (=0.75*0.85). The first part of the adjustment accounts for the fact that carbon tax payments would be a deductible business expense and so would reduce revenues from the individual and corporate income tax.  The Congressional Budget Office (2009) estimates the size of the offset to be about 25 percent of gross tax revenue. The second part of the adjustment accounts for the income support payments for low-income households, equal to 15 percent of gross carbon tax revenue.

Endnotes

[1] Brookings Institution and Urban-Brookings Tax Policy Center.  I am greatly indebted to Aaron Krupkin for his work generating the estimates described here.

[2] In equation form, current revenue = 0.21 * TCJA base.   Revenue under a 25 percent cash flow tax = 0.25 * cash flow base = 0.25 (TCJA base + change in base).  The revenue change is = (0.25*TCJA base) – (0.21*TCJA base) + (0.25 * change in base) = (0.04 * TCJA base) + (0.25 * change in base).

 

References

American Society of Civil Engineers. 2016. “Failure to Act: Closing the Infrastructure Investment Gap for America’s Economic Future.” American Society of Civil Engineers: Reston, Virginia.

Batchelder, Lily. 2016. “The “Silver Spoon” Tax: How to Strengthen Wealth Transfer Taxation.” Washington Center for Equitable Growth.

Buettgens, Matthew, Lisa Dubay and Genevieve Kenney. 2016. “Marketplace Subsidies: Changing the Family Glitch Reduces Family Health Spending but Increases Government Costs.” Health Affairs. 35(7): 1167-1175.

Centers for Medicare and Medicaid Services. 2018. “NHE Projections 2017-2026.” Department of Health and Human Services: Washington, DC.

Congressional Budget Office. 2009. “The Role of the 25 Percent Revenue Offset in Estimating the Budgetary Effects of Legislation.”  Economic and Budget Issue Brief, Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2013a. “Options for Reducing the Deficit: 2014 to 2023.” Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2013b. “Updated Budget Projections: Fiscal Years 2013 to 2023.” Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2015. “Public Spending on Transportation and Water Infrastructure, 1958-2014.” Congressional Budget Office: Washington, DC.

Congressional Budget Office 2016a. “Options for Reducing the Deficit: 2017 to 2026.” Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2016b. “An Update to the Budget and Economic Outlook: 2016 to 2026.” Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2017a. “An Analysis of the Obama Administration’s Final Future Years Defense Program.” Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2017b. “Federal Subsidies for Health Insurance Coverage for People Under Age 65: 2017 to 2027.” Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2017c. “A Premium Support System for Medicare: Updated Analysis of Illustrative Options.” Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2017d. “Repealing the Individual Health Insurance Mandate: An Updated Estimate.” Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2018a. “The Budget and Economic Outlook: 2018 to 2018.” Congressional Budget Office: Washington, DC.

Congressional Budget Office. 2018b. “Federal Subsidies for Health Insurance Coverage for People Under Age 65: 2018 to 2028.” Congressional Budget Office: Washington, DC.

Daniels, Seamus P., and Todd Harrison. 2018. “Making Sense of the Bipartisan Budget Act of 2018 and What it Means for Defense.” Center for Strategic and International Studies: Washington, DC.

Dorn, Stan and Matthew Buettgens. 2017. “The Cost of Not Expanding Medicaid: An Updated Analysis.” The Urban Institute and the Robert Wood Johnson Foundation: Washington, DC.

Drukker, Austin, Ted Gayer and Harvey Rosen. 2017. “The Mortgage Interest Deduction: Revenue and Distributional Effects.” Griswold Center for Economic Policy Studies Working Paper No. 251.

Genworth Mortgage Insurance. 2017. “First Time Homebuyer Market Report.” Genworth Mortgage Insurance Corporation, Richmond VA. June.

Goss, Stephen C. 2016. Memorandum to Kent Conrad and James B. Lockhart III. Social Security Administration: October 11. https://www.ssa.gov/oact/solvency/ BPCCRSPS_20161011.pdf.

Hall, Keith. 2018. “Letter to the Honorable Lamar Alexander, Re: Appropriation of Cost-Sharing Reduction Subsidies.” Congressional Budget Office: Washington, DC.

Harrison, Todd. 2016. “Analysis of the FY2017 Defense Budget.” A Report of the CSIS International Security Program’s Defense Outlook Series: Washington, DC

Joint Committee on Taxation. 2017. “Estimated Budget Effects of the Conference Agreement for H.R. 1, the “Tax Cuts and Jobs Act.” JCX-67-17.

Joint Committee on Taxation. 2018. “Estimates of Federal Tax Expenditures for Fiscal Year 2017-2021.” JCX-34-18.

McClelland, Robert. 2017. “Raise the 15 percent Tax Rate on Capital Gains to Boost Revenue in a Progressive Way.” Urban-Brookings Tax Policy Center: Washington, DC.

McKibbin, Warwick, Adele Morris, Peter Wilcoxen and Weifing Liu. 2018. “The Role of Border Carbon Adjustments in a US Carbon Tax.” Climate Change Economics. 9(1): 1840011-1-1840011-41.

Nunns, James R., Leonard E. Burman, Jeffrey Rohaly, Joseph Rosenberg, and Benjamin R. Page 2016. “An Analysis of the House GOP Tax Plan.” Tax Policy Center: Washington, DC.

Office of Management and Budget. 2016. “Budget of the United States Government, Fiscal Year 2017.” The White House: Washington, DC.

Office of Management and Budget. 2018. “Historical Tables.” The White House: Washington, DC.

Toder, Eric, Jim Nunns, and Joseph Rosenberg. 2012. “Implications of Different Bases for a VAT.” Urban-Brookings Tax Policy Center: Washington, DC.