#1Modeling the Repeal of the Rhode Island Sales Tax Steven E. Collum, Revenue Policy Analyst Paul L. Dion, Ph.D., Chief Office of Revenue Analysis Rhode Island Department of Revenue FTA Revenue Estimation and Tax Research Conference 2014 Annual Meeting, San Antonio, TX September 27 - October 1, 2014#21. 2. Outline of the Presentation Introduction The 2013 Zero.Zero Sales Tax Report REMI PI+ Model Structure 3. Rhode Island and Its Neighbors 4. 5. Overview of REMI PI+ Model 6. 1. 8. Rhode Island Sales Tax Simulation Model ORA Modeling of Sales Tax Repeal RI PI+ Sales Tax Repeal Results Revenue Impacts of Sales Tax Repeal 10. Public Jobs Impact of Sales Tax Repeal 9. 11. Summary 12. Epilogue September 23, 2014 2014 Revenue Estimation and Tax Research Conference 2#3Introduction In the 2013 Session, the General Assembly considered two bills that proposed the broad repeal of the state's sales and use tax and the local meal and beverage tax State sales tax rate: 7.0 percent - - • ● Applies to most tangible property (64 exemptions) Some services (ex., telecommunications, pet grooming) - Local meal and beverage tax rate: 1.0 percent . Applies to the consumption of prepared foods and beverages In lieu of repealing the state sales and use tax, the General Assembly created the Joint Legislative Commission to Study the Sales Tax Repeal Act of 2013 September 23, 2014 2014 Revenue Estimation and Tax Research Conference 3#4The Zero.Zero Sales Tax Report The impetus for the formation of the Commission was a 2013 study produced by the Rhode Island Center for Freedom and Prosperity titled “Zero.Zero Sales Tax" • The centerpiece of the 2013 Zero.Zero Sales Tax report is the economic impact results derived from the STAMP model of the Rhode Island economy: - - - - The creation of approximately 25,000 private sector jobs Nearly $1.0 billion of Rhode Islanders' own money to be spent in the state's economy A $150 million revenue gain for cities and towns and more than $500 million in state tax receipts from increased economic activity Cost savings for businesses through eliminated compliance costs and reduced costs for supplies and services September 23, 2014 2014 Revenue Estimation and Tax Research Conference 4#5Rhode Island and Its Neighbors Troy Rensselaer Columbia Pittsfield Haverhill Nashua Methuen Lawrence Essex Lowell Gloucester Franklin Gardner Fitchburg Reading O Danvers O O Peabody Beverly Leominster Marblehead Middlesex Wobur • Lynn Somerville Waltham O Worcester Marlborough Suffolk Berkshire Hampshire Litchfield hess Holyoke Westfield Hampden Springfield Massachusetts 0 Ramingham Milford Pranklin Woonsocket Quincy Dedham Weymouth Norfolk Massachusetts Bay Brockton Plymouth Attleboro Providence Taunton Tolland Pawtucket Windham Bristol Providence Bristol West Warwick o Kent Warwick Fall River New Bedford Hartford Central O Manchester Hartford Bristol O Connecticut Newport Middletown Waterbury Menden Naugatuck New London Middlesex Washington Newport am Danbury New Haven North North Haven New London OJ Fairfield Sheltony New Haven Trumbu Milford ester Bridgeport Census 2010 - Population Density (per sq. mile) Norwalk Stamford 5 50 200 500 1,000 3,000 5,000 7,000 9,000 15,000 September 23, 2014 2014 Revenue Estimation and Tax Research Conference Dukes Barnstable Barnstable Town Nantucket 5#6REMI PI+ Model Structure • The REMI PI+ model used by the Office of Revenue Analysis is a one region model that is calibrated to Rhode Island's economy (RI PI+) ⚫RI PI* is a structural economic forecasting and policy analysis model that - Integrates input-output, computable general - equilibrium, econometric, and economic geography methodologies – Is dynamic, with forecasts and simulations generated on an annual basis and behavioral responses to compensation, price, and other economic factors September 23, 2014 2014 Revenue Estimation and Tax Research Conference 6#7Overview of REMI PI+ Model Commodity Access Index Intermediate Input Productivity (1) Output and Demand State and Local Government Spending Investment (3) Population and Labor Supply Migration Output Exports (2) Labor and Capital Demand Optimal Capital Population Employment Stock Intermediate Inputs Consumption Real Disposable Income (5) Market Shares Participation Rate Labor Access Labor Force Index Labor Productivity Domestic Market Share International Market Share Employment Opportunity (4) Compensation, Prices, and Costs Compensation Rate Composite Compensation Rate Production Costs Housing Price Consumer Prices Real Compensation Rate Composite Prices September 23, 2014 2014 Revenue Estimation and Tax Research Conference 7#8. Overview of REMI PI+ Model The RI PI+ model is comprised of: - - - - - - Seventy industrial sectors, generally delineated at the three digit NAICS code level Seventy-five consumer sectors, with individual commodities delineated by BEA NIPA tables Two factor markets, Capital and Labor - Three government sectors ● • Federal government military and non-military expenditures (2 total) State and local government pecuniary and non-pecuniary expenditures (2 or 3 total) Three investment sectors ● Residential and nonresidential buildings (1 each) Producer's durable equipment One rest of world sector September 23, 2014 2014 Revenue Estimation and Tax Research Conference 8#9RI Sales Tax Simulation Model • The Rhode Island Sales Tax Simulation Model, developed by Chainbridge Software, allocates the sales tax across 230 consumer product categories - 64.9 percent of the state sales tax is paid by households • ● ● 32.2 percent is paid on durable goods purchases 31.4 percent is paid on nondurable goods purchases 36.4 percent is paid on services September 23, 2014 2014 Revenue Estimation and Tax Research Conference 6#10RI Sales Tax Simulation Model • The Rhode Island Sales Tax Simulation Model allocates the sales tax across 433 business input and 180 business investment categories - 35.1 percent of the state sales tax is paid by businesses • • 26.5 percent is paid on investment goods purchases 73.5 percent is paid on intermediate goods purchases September 23, 2014 2014 Revenue Estimation and Tax Research Conference 10 10#11• . RI Sales Tax Simulation Model ORA calculated the sales tax allocation percentage for each of the 230 consumer product categories in the RI Sales Tax Model The sales tax allocation percentage is the ratio of the total sales tax paid for a given consumer product category divided by the total sales tax paid for all consumer product categories ORA mapped the 230 consumer product categories from the sales tax simulation model to the 75 consumer sectors in RI PI* model using NAICS Codes The sales tax allocation percentages were summed in accordance with the above mapping A similar process was used to map the sales tax allocation percentages for the 433 business input and 180 business investment categories contained in the RI Sales Tax Model to the 70 industry sectors contained in the RI PI+ model September 23, 2014 2014 Revenue Estimation and Tax Research Conference 11 14#12ORA Modeling of Sales Tax Repeal The 2013 Zero.Zero Sales Tax report used the sales tax estimate of $904.3 million contained in the Governor's FY 2014 Recommended Budget ORA broke down the $904.3 million into sales tax paid by consumers and sales tax paid by businesses with 64.9% allocated to consumers and 35.1% to businesses The $586.5 million of consumer sales taxes was then allocated to the 75 RI PI+ consumer categories using the relevant sales tax allocation percentages The $317.8 million of business sales taxes paid was allocated to the 70 RI PI+ industries in a similar manner September 23, 2014 2014 Revenue Estimation and Tax Research Conference 12#13• • ORA Modeling of Sales Tax Repeal In the 75 consumer sectors, RI PI+ reduced relative prices of goods and services subject to the sales tax by $586.5 million In the 70 industry sectors, RI PI* reduced production costs of businesses that made input and investment purchases subject to the sales tax by $317.8 million ORA found that on average over the past five fiscal years state general revenues were spent as follows: - - - - - Purchases of goods and services: 41.9 percent Local government wages and salaries: 31.6 percent State government wages and salaries: 19.3 percent Purchases of investment goods: 0.01 percent Grants to nonprofits serving households: 1.4 percent State payroll costs (retirement, FICA, etc.): 5.8 percent September 23, 2014 2014 Revenue Estimation and Tax Research Conference 33 13#14• ORA Modeling of Sales Tax Repeal ORA modeled the $904.3 million in lost revenue from the sales tax repeal as follows: - - - - - Demand for goods and services: $379.2 million ↓ Local government wages and salaries: $285.5 million ↓ State government wages and salaries: $174.1 million ↓ Grants to nonprofits serving households: $12.3 million ↓ Investment: $0.4 million ↓ Savings on state payroll costs: $52.8 million ↑ Due to the savings on state payroll costs, ORA only modeled $851.5 million in reduced spending by state government in Year 1 The sales tax revenue repeal does cause other revenues to increase as economic activity increases. Thus, in Year 2, state government spending is reduced by $798.5 million. Similarly, for Year 3 it is reduced by $744.1 million, for Year 4 it is reduced by $688.4 million and for Year 5 it is reduced by $632.6 million September 23, 2014 2014 Revenue Estimation and Tax Research Conference 14 14#15RI PI+ Sales Tax Repeal Results The RI PI+ model starts with a baseline scenario of the state's economic performance - - The baseline scenario shows the economy's performance in the absence of a tax policy change The baseline scenario projects the RI economy's growth path in the absence of a tax policy change • ● Compounded Annual Growth Rate (CAGR) for Private Employment: 2.7% CAGR for Investment: 12.6% CAGR for Gross State Product: 6.2% • CAGR for Population: 0.6% RI PI+ simulates the path of the state's economy after the sales tax repeal and the reduction in state government spending and compares the simulation value after the tax policy change to the baseline scenario September 23, 2014 2014 Revenue Estimation and Tax Research Conference 15#16RI PI+ Sales Tax Repeal Results RI PI+ Simulation Value vs. RI PI+ Baseline Scenario Sales Tax Repeal and State Spending Reduction Year 1 Year 2 Year 3 Year 4 Year 5 Private Jobs 926 2,385 3,606 4.588 5,383 Investment (millions) $75 $117 $145 $165 $179 RI GSP (millions) $(602) $(434) $(283) $(149) $(29) Population 3,969 7,147 9,909 12,288 14,326 September 23, 2014 2014 Revenue Estimation and Tax Research Conference 16#17• RI PI+ Sales Tax Repeal Results As the Sales Tax Repeal and State Spending Reduction table shows, the repeal of the sales tax and the reduction in state government spending results in a substantive impact on the state's economy - - - Private employment increases by 5,383 jobs in Year 5 Private investment increases by $179 million in Year 5 RI Gross State Product decreases by $29 million in Year 5 RI population increases by 14,326 in Year 5 To isolate the effect of the reduction in state government spending on the economic impact of the repeal of the sales tax, ORA ran the RI PI+ model with only the repeal of the sales tax RI PI+ compared the simulation value after the sales tax repeal without the reduction in state government spending to the same baseline scenario September 23, 2014 2014 Revenue Estimation and Tax Research Conference 17#18RI PI+ Sales Tax Repeal Results RI PI+ Simulation Value vs. RI PI* Baseline Scenario Sales Tax Repeal Only Private Jobs Year 1 Year 2 Year 3 Year 4 Year 5 8,993 9,963 10,518 10,692 10,649 Investment (millions) $153 $235 $282 $308 $319 RI GSP (millions) $701 $817 $901 $950 $978 Population 7,847 13,769 18,572 22,371 25,322 September 23, 2014 2014 Revenue Estimation and Tax Research Conference 18#19RI PI+ Sales Tax Repeal Results The Sales Tax Repeal Only table shows that eliminating the sales tax w/out reducing State spending has a bigger impact on the RI economy - Private employment increases by 10,649 jobs in Year 5 Private investment increases by $319 million in Year 5 - RI GSP increases by $978 million in Year 5 - RI population increases by 25,322 in Year 5 By Year 5, nearly twice as many private jobs are created and RI Gross State Product is $1.0 billion more than under the Sales Tax Repeal and State Spending Reduction simulation The difference in the two simulation values vis-à-vis the baseline is due to the direct, indirect and induced effects of State government spending September 23, 2014 2014 Revenue Estimation and Tax Research Conference 19 19#20RI PI+ Sales Tax Repeal Results Under the Sales Tax Repeal and State Spending Reduction simulation, the decrease in state government spending: - ↓ demand for goods and services and labor by government (direct effect) - ↓ demand for goods and services and labor by suppliers of goods and services and labor to government (indirect effect) - ↓ demand for goods and services and labor by suppliers of goods and services and labor to the suppliers of goods and services and labor to government (induced effect) Part of the rebound in the economy by Year 5 is the result of increased government spending as other revenue sources increase with increased economic activity September 23, 2014 2014 Revenue Estimation and Tax Research Conference 20 20#21Revenue Impacts of Sales Tax Repeal ORA Revenue Impact Based on RI PI* Economic Impact Revenue Source (millions) Baseline Year 1 Year 2 Year 3 Year 4 Year 5 Sales Tax $904 $ - $- $ - $- $ - Business Taxes $238 $244 $250 $257 $263 $270 Motor Fuel Tax $137 $143 $148 $154 $160 $165 Motor Vehicle Fees $49 $51 $53 $55 $57 $59 Estate Tax $35 $36 $36 $37 $38 $38 Personal Income Tax $1,131 $1,151 $1,171 $1,192 $1,215 $1,238 Cigarette and Alcohol Taxes $149 $155 $161 $167 $174 $180 Other Taxes $153 $156 $159 $162 $165 $168 Other Sources $766 $780 $794 $809 $824 $840 Total $3,562 $2,716 $2,772 $2,833 $2,896 $2,958 Difference from Baseline $(846) $(790) $(729) $(666) $(604) Cumulative Difference $(846) $(1,636) $(2,365) $(3,031) $(3,635) September 23, 2014 2014 Revenue Estimation and Tax Research Conference 21#22Public Jobs Impact of Sales Tax Repeal RI PI+ State and Local Government Jobs Impact Results Direct Effect Year 1 Year 2 Year 3 Year 4 Year 5 State Government (2,991) (2,817) (2,638) (2,455) (2,271) Local Government (5.526) (5,200) (4.866) (4,524) (4,181) Direct Subtotal (8,517) (8,017) (7,504) (6,979) (6,452) Change (8,517) 500 513 525 527 Indirect and Induced Effects Year 1 Year 2 Year 3 Year 4 Year 5 State Government 33 85 126 157 180 Local Government 61 143 209 260 299 Indirect and Induced Subtotal 94 228 335 417 479 Change 94 134 107 82 62 Direct, Indirect, Induced Total (8,423) (7,789) (7,169) (6,562) (5,973) Change (8,423) 634 620 607 589 September 23, 2014 2014 Revenue Estimation and Tax Research Conference 22#23Summary Modeling a tax policy change using REMI PI* requires the ability to translate aggregate revenue changes into sector specific impacts - REMI PI+ allows the modeler to use a predetermined spreader, usually based off of output or industry demand but this will spread impact to all sectors . • As noted earlier, Rhode Island has 64 exemptions from the state sales and use tax Manufacturing inputs and equipment are exempt but using the spreader function may result in an allocation to the manufacturing sector when none is warranted The Rhode Island Sales Tax Simulation Model provided a mapping between the aggregate revenue changes and the sector specific impacts of a sales tax repeal September 23, 2014 2014 Revenue Estimation and Tax Research Conference 23#24. Summary By using NAICS codes, ORA was able to map the impact of the sales tax repeal from the Sales Tax Simulation Model's household and business sectors to RI PI +'s industry and consumer sectors The impact of the sales tax repeal on the business sector was modeled as a reduction in the production costs of impacted industries The impact of the sales tax repeal on the household sector was modeled as a reduction in the final price paid by consumers September 23, 2014 2014 Revenue Estimation and Tax Research Conference 24 24#25Summary The government sector in PI+ did not align with the state's general revenues so ORA created a custom government sector that allocated state general revenue expenditures to PI + based on an in-house mapping of expenditures to NAICS codes • This gave ORA the flexibility to hold some expenditure items harmless (i.e., debt service) and more narrowly target others (i.e., pass through local aid) One important lesson we learned is that over a multi-year horizon, the impact of a revenue reduction on government spending diminishes as private economic activity increases and fixed employment costs are avoided September 23, 2014 2014 Revenue Estimation and Tax Research Conference 25#26• Epilogue The proposed sales tax repeal in Rhode Island became a "Battle of the Alamo" between the State Tax Analysis Modeling Program (STAMP) and REMI's PI+ model ORA's original presentation to the Joint Legislative Commission to Study the Sales Tax Repeal Act of 2013 contained numerous slides on the assumptions and differences between the two models • This presentation can be obtained by emailing Paul Dion at [email protected] The Institute on Taxation and Economic Policy (ITEP) also wrote a piece based on this "Mexican Standoff" which is available at http://itep.org/itep_reports/ September 23, 2014 2014 Revenue Estimation and Tax Research Conference 26#27Thank You September 23, 2014 2014 Revenue Estimation and Tax Research Conference 27