Provincial Taxation of High Incomes: What are the Impacts on Equity and Tax Revenue?

Size: px

Start display at page:

Download "Provincial Taxation of High Incomes: What are the Impacts on Equity and Tax Revenue?"

Susanna Casey
6 years ago
Views:

1 Provincial Taxation of High Incomes: What are the Impacts on Equity and Tax Revenue? Kevin Milligan Vancouver School of Economics University of British Columbia Michael Smart Department of Economics University of Toronto IRPP-CLSRN Conference Inequality in Canada Ottawa: February, 2014 Comments welcome Milligan and Smart: Provincial Taxation of High Incomes Page 1 of 22

2 Should redistribution happen centrally or locally? Traditional public finance / fiscal federalism advice: (E.g. Oates CJE 1968; Musgrave 1971) With mobile factors, horizontal tax competition limits ability of local governments to redistribute. Less mobility of factors out of whole country, so central redistribution more efficient. But: Recent Canadian experience shows 4 (or 5) provinces increasing top rates; not federal government. No sign any federal party is going to. Milligan and Smart: Provincial Taxation of High Incomes Page 2 of 22

3 Recent Canadian experience: Provinces take the lead Notes: Shown are the highest personal marginal tax rate by province in The five provinces with recent increases in their top rates are shown, with the year of the increase indicated to the right of the bar. Milligan and Smart: Provincial Taxation of High Incomes Page 3 of 22

4 Our question: What are the implications of provincial taxation of higher incomes? Implications for equity Implications for tax revenues Plan for talk: Estimates of responsiveness of taxable income to provincial tax rates Simulations of a counterfactual: each province adds a high-income tax bracket o Present impact on equity, revenues. Milligan and Smart: Provincial Taxation of High Incomes Page 4 of 22

5 Provincial income taxation in Canada: Under the Constitution Act, Section 91 / 92 allocate power for direct taxation to both federal and provincial levels. Base is co-occupied. Tax Collection Agreements: Since 1962, in all provinces except Quebec. Agreeing provinces must use federal tax base. Since 2000/2001, provinces free to set own brackets and rates. Milligan and Smart: Provincial Taxation of High Incomes Page 5 of 22

6 Figure 1: Provincial Top Tax Rates, 2000 to 2014 Notes: Shown are the highest personal marginal tax rate by province by year from 2000 to The source is the Canadian Tax and Credit Simulator. Milligan and Smart: Provincial Taxation of High Incomes Page 6 of 22

7 What impact do taxes have on reported income? See recent survey by Saez, Slemrod, and Giertz (2012 J. Economic Literature) 1980s focus: real responses Adjustment of labour supply, consumption / savings / investment, residency. Q: Do hyper-competitive top lawyers and hedge fund workers really cut hours worked in response to taxes? More recent focus: Shifting / avoidance Look at reported income, not hours etc. Responses: timing, accounting, form o Assisted by expensive advice, specialized financial structures. Milligan and Smart: Provincial Taxation of High Incomes Page 7 of 22

8 Figure 2: Provincial Top Income Shares vs. Top Tax Rates, 1988 to 2011 Notes: The graph plots the top one percent income share against the top combined federal-provincial marginal tax rate by province for the years The income shares use total incomes (excluding capital gains). The source is CANSIM table The tax rates are from the Canadian Tax and Credit Simulator. Milligan and Smart: Provincial Taxation of High Incomes Page 8 of 22

9 Estimation: How do taxes affect reported income of top 1%? Use CANSIM high income database to estimate: ( ) Share of income going to the top 1% Top tax rate in province p in year t. Total income in province p in year t. Provincial fixed effects. Year fixed effects. Object of interest: Elasticity of taxable income Milligan and Smart: Provincial Taxation of High Incomes Page 9 of 22

10 Table 1: Regression Results (2) (3) (4) Sample P99 P95-P99 P99.9 Observations R-squared Tax elasticity 0.664** * [0.270] [0.111] [0.615] Log Income 0.633*** ** 0.843*** [0.0785] [0.0336] [0.183] Notes: Reported are coefficients from regressions of the log share of top percentile income on the net of tax rate and controls on provincial level data from 1988 to Each column contains the results of a different regression. All specifications include year and province fixed effects. The income fractile used for each column is listed in the sample row. Other details can be found in the main text. (See Milligan and Smart 2014 for more) Milligan and Smart: Provincial Taxation of High Incomes Page 10 of 22

11 How to interpret the elasticity: The elasticity of means that a 10% decrease in ( ) leads to a 6.64% decrease in the top 1% income share. How does this compare to literature? Milligan and Smart (2014) use same data; find similar robust result. Department of Finance (2010) finds e=0.62 with similar approach. SSG (2012) conclude that 0.25 is middle of range of estimates for US. Why different from SSG (2012)? Existing evidence may be confounded by time series trends. e is a function of particular tax base. Canada US Using provincial variation, some income may leak horizontally to other provinces, so parameter may be different federally/provincially. Milligan and Smart: Provincial Taxation of High Incomes Page 11 of 22

12 A Counterfactual: Each province introduces top tax bracket Consider the following exercise: Take the tax system in Institute a top tax bracket in every province at that province s 1% threshold. Rate for new bracket is +5% over the current top rate. Milligan and Smart: Provincial Taxation of High Incomes Page 12 of 22

13 Table 2: Simulated Base and Counterfactual Tax Systems Base System +5 System Bracket Bracket Threshold Rate Threshold Rate NL $ 63, % $ 168, % PE $ 98, % $ 139, % NS $ 150, % $ 154, % NB $ 120, % $ 147, % QC $ 78, % $ 169, % ON $ 78, % $ 215, % MB $ 67, % $ 161, % SK $ 116, % $ 180, % AB $ % $ 281, % BC $ 100, % $ 190, % Fed $ 128, % $ 128, % Fed-QC $ 128, % $ 128, % Notes: shown are the bracket threshold for the base and +5 tax systems used in our simulations. The base system is from The federal rate at the bottom of the table is different for Quebec because of the 16.5% provincial abatement. Milligan and Smart: Provincial Taxation of High Incomes Page 13 of 22

14 Measuring impact of +5 system on equity How to measure equity? Looking at ratios for top 1% not useful. Gini coefficient isn t going to pick up much for a change at top 1% Our alterative: Use the Average Tax Rate: Change in ATR tells us the change in proportion of income individuals have available to consume. Milligan and Smart: Provincial Taxation of High Incomes Page 14 of 22

15 Figure 3: Average Provincial Tax Rates under Existing and Counterfactual Tax Policies Notes: Plotted are the average tax rates for Quebec, New Brunswick, and Alberta under the 2011 tax system and under a new tax system featuring a high tax bracket with a rate 5 percent higher. The x-axis indicates the level of earned income and the y-axis the average tax rate. The tax rates are calculated using the Canadian Tax and Credit Simulator. Milligan and Smart: Provincial Taxation of High Incomes Page 15 of 22

16 Results of +5 system on equity Three ways to view results: Has no impact for those up to top 1% threshold. Varies by province. Impact for those over threshold rises slowly At average income level for top 1%, only a 2.3% change in ATR. (Alberta) Implication: Top 1% share in AB has about doubled. Taking back 2.3% isn t much. To undo changes in inequality would require much larger tax changes. Milligan and Smart: Provincial Taxation of High Incomes Page 16 of 22

17 Measuring impact of +5% system on provincial revenue Decompose into two conceptual chunks: Net Revenue = Mechanical effect - Behavioural effect Mechanical effect: How much does revenue go up by applying new rates to existing taxable income? Assumes no behavioural response. In +5% system, varies by province because of strong differences in provincial income distributions. Behavioural effect: Using elasticity of 0.664, how much taxable income vanishes? Varies by province because of size of existing top tax rate; differing income distributions. Milligan and Smart: Provincial Taxation of High Incomes Page 17 of 22

18 Figure 4: Per Tax Filer Mechanical Effect and Net Revenue Notes: Plotted are the mechanical effect and net revenue per-tax filer of a new five percent tax bracket on incomes over the top one percent threshold in each province. Milligan and Smart: Provincial Taxation of High Incomes Page 18 of 22

19 Table 3: Explaining Differences in Revenue Across Provinces PE AB Average income of top 1% 203, ,475 Threshold to be in top 1% 139, ,096 Ratio (inverted Pareto) Mechanical revenue per capita Provincial revenue max'ing tax rate 22.8% 32.7% Actual 2011 top tax rate 18.4% 10.0% How much of mechanical effect remains 6% 71% Two drivers of vast PEI-Alberta difference: Income distribution in AB much more skewed. More income above 1% threshold. PEI s 2011 tax rate already near the revenue-maximizing level. Not AB. Milligan and Smart: Provincial Taxation of High Incomes Page 19 of 22

20 Other impact: Federal tax revenues shrink Because of co-occupied tax base, there is vertical fiscal externality: If reported income goes down in response to provincial tax rate, federal taxable income also affected. Apply federal tax rates to tax base shrinkage to gauge size of vertical externality. Not accounted for here: Other tax bases could be positively affected. o Corporate revenues might go up. o Other province revenues could go up. Milligan and Smart: Provincial Taxation of High Incomes Page 20 of 22

21 Table 4: Assessing the Vertical Fiscal Externality NL PE NS NB QC ON MB SK AB BC Mechanical revenue Behavioural revenue Net provincial revenue Change in federal revenue Gains in other tax bases +? +? +? +? +? +? +? +? +? +? Notes: Reported are results for each province of the simulated revenue impact of a new tax bracket starting at each province s top one percent income threshold with a tax rate 5 percent higher than the existing top rate. The simulations are based on 2011 data. Change in federal revenue is potentially very large Feds don t get any mechanical effect since they didn t raise rates. Federal top tax rate is 29%, so loss of taxable income hurts feds more Milligan and Smart: Provincial Taxation of High Incomes Page 21 of 22

22 Summary Four main findings: Provincial taxable income elasticity is high. Equity: +5% system wouldn t do much to reverse top 1% income surge. Revenue: Vast provincial differences in revenue potential of new top 1% bracket. Negative impact on federal revenues as large as provincial revenue gains. In future work: Try to sort out magnitudes of tax avoidance vs. horizontal shifting Calculate optimal rates that fully account for all vertical and horizontal externalities. Milligan and Smart: Provincial Taxation of High Incomes Page 22 of 22

Results are preliminary. Comments welcome.

Results are preliminary. Comments welcome. Estimating high-income tax elasticities using sub-national variation in tax rates Kevin Milligan Vancouver School of Economics University of British Columbia Michael Smart Department of Economics University