Your Biggest Refund, Guaranteed? Tax Filing Method and Reported Tax Liability. Samara Gunter Economics

Transcription

1 Your Biggest Refund, Guaranteed? Tax Filing Method and Reported Tax Liability Samara Gunter Economics

2 Tax preparation service providers promise to maximize your tax refund H&R Block: Get your billions back, America. TurboTax: Get your maximum refund, guaranteed. Premise of these advertisements: Taxes are so complex that, without the use of software or paid preparers, filers will miss credits and deductions and over-report tax liability. If true, then even in the absence of changes in tax laws the shift away from self-prepared paper returns to tax software could reduce tax collections.

3 The Tax Filing Experience Changes in filing method change the experience of paying taxes May change useof tax deductions and credits Has revenue impacts even in absence of changes in tax law or economic conditions May change understanding and awarenessof tax deductions and tax credits Can affect behavioral responses Limits effectiveness of tax incentives

4

5

6

7

8 Americans have changed how they pay their taxes Between 1993 and 2003, the number of tax filers who used software to prepare their return rose from 8% to 25% Between 2002 and 2009, the number of Americans who filed self-prepared paper returns fell from 23.4 percent to 8.9 percent. The fraction of self-prepared online returns rose from 9.3 percent to 25.1 percent. Has the shift in filing method impacted filers reported tax liability?

9 Tax Filing Methods, e-file Demographic Data (Median Rate in ZIP Code) e-file Paper Online V-Code Paid Preparer

10 Changes in Household Computer and Internet Use (CPS supplement) Households with a computer Households with internet access

11 Selection Bias in Choice of Filing Method Tax filing method is partly determined by the complexity of tax return and tax filers circumstances Filers are more likely to use software or preparers if they are eligible for more credits, can itemize, etc Consequence: regressing tax filing method on tax variables such as average tax rate, credit use, credit value, probability of itemizing, value of itemized deductions, etc, will be biased

12 Past Literature Guyton, Korobow, Lee, and Toder(NTJ 2005) Selection of filing type taxpayers are about equally more likely to increase use of paid preparers or software as their returns become more complex more likely to use paid preparers than software if they are self-employed and have a high ratio of tax due to tax liability more likely to choose software than paid preparers if they own a home computer and are highly educated Compliance cost of tax filing, by filing type estimates show savings of $223 for switching to self-preparation and $212 for software users of paid preparers would spend on average an additional 23.4 hours if they prepared their own returns without software and an additional 22.8 hours if they used software

13 Kopczukand Pop-Eleches(2007) introduction of state electronic filing had a significant effect on participation in the EITC likely driven by expansion of tax preparation services into low-income neighborhoods. Chetty and Saez(2009) experiments in H&R Block offices to demonstrate that understanding of, and response to, the Earned Income Tax Credit rose after tax preparers spent two minutes explaining the credit to EITC-eligible households. Ebenstein and Stange(2010) no effects of changing from in-person applications for unemployment insurance to phone or internet-based applications on UI take-up or other changes in UI benefits.

14 Identification Strategy: (Panel) Instrumental Variables

15 IRS ZIP Code Tax Data IRS Statistics of Income: Individual Income Tax Statistics ZIP Code Data Tax return data aggregated to ZIP code level (number of returns, adjusted gross income, wages and salaries, credit use, deductions, other items from tax return) Available 1998, 2001, IRS: e-file Demographic Data Tax filing method and demographic data Available , but 2011 is problematic JUST got 1998 and 2001 Matched file (with many thanks to Dan Meyer `16) ZIP-code data spanning (7 years)

16 Decision Map: Tax Filing Method Use Paid Preparer e-file Paper Tax Filer Paper e-file Self-Prepare Software CD Online Paper Online e-file Paper

17 Tax Filing in the e-file Demographic Data 1998, 2001, Percent paper + Percent e-file 100% Percent paid preparer Percent paper + Percent e-file + Percent online = 100% Percent paid preparer Percent v-coded (computer generated, paper filed)

18 Considerations: Interpreting Results Paid preparers should be familiar with the tax code changes in paper filing vs. e-filing driven by preparers should not affect deductions, credit use, and overall tax liability Also: changes in paper vs. e-filing for preparers needn t indicate a change in method Changes to e-filing driven by self-preparers indicate a switch in tax preparation method they cannot e-file a self-prepared paper form Using e-filing rate to proxy for shift away from self-prepared paper will downward-bias results The increase in e-filing driven by preparers should not affect deductions & credit use

19

20

21

22 Possible Instruments for Filing Method Computer use/internet access FCC data on presence/number of high-speed service providers: (1999-June 2008 zip-code level, December 2008 onward census tract or county level) Current Population Survey data (individual/household data on computer ownership and internet use, smallest viable geographic aggregation is state level) Cost constraints on spread of technology Housing density (zip-code level, doesn t vary over time, effects may vary over time) Elevation (county level, doesn t vary over time)

23 Regulation of Tax Preparers Generally unregulated 2011: IRS Return Preparer Initiative Required that paid preparers pass a certification exam, pay annual fees and complete at least 15 hours of education courses annually. Struck down by US District Court for District of Columbia in 2013 IRS does not have the authority to regulate tax preparers Lower court decision upheld by US Court of Appeals-DC Circuit in February 2014

24

25

26 The Latest Results OLS results: long differences IV results (long differences and panel IV): instrument for e-filing rates using FCC data on holding companies providing high-speed internet IV results: instrument for e-filing rates using ZIP code housing density

27 e-filing and Average Tax Rates (tax/income)

28 OLS Long Differences: Itemized Deductions

29 OLS-Long Differences Results Increases in e-filing are associated with lower average tax rates higher rates of itemizing (mixed) higher average itemized deductions But what if ZIPs with larger e-filing increases are ZIPS in which people had more changes in tax circumstances?

30 Population Density IV: Long Differences

31 Population Density IV Increases in e-filing decrease average tax rates decrease(?!) percent itemizing Increase itemized deductions conditional on itemizing is different Some coefficient magnitudes (too) large Percent itemizers result is surprising Control variables matter Results are comparable using panel FE instead of long differences

32 FCC Internet Holding Company IV: Panel FE

33 FCC Internet Company IV Same results as population density IV! decrease average tax rates decrease(?!) percent itemizing increase itemized deductions conditional on itemizing Magnitudes are in the same ballpark

34 Conclusions Results suggest that e-filing reduces tax liability and increases use of tax credits and deductions Probable mechanism: shift toward tax software Policy implications Changes in tax filing method do have revenue effects Changes in tax filing method may have behavioral effects Estimates of behavioral responses to tax changes using data from the 1990s and 2000s may be partially influenced by filing method effects