Taxpayer Compliance Under Different Auditing Regimes

1
Taxpayer Compliance Under Different Auditing
Regimes

• Michael McKee
• Holly Professor of Economics, University of
  Tennessee
• Professor and Arthur Child Chair in Defence
  Economics, University of Calgary (as of January
  2006)

2
Introduction

• Tax evasion magnitude 200 Billion on the PIT
• Issues
• Government Revenues
• Efficiency Losses
• Social Norms
• Tax Administration
• Affects All Tax Sources (focus here on PIT)

3
Introduction

• Study of compliance
• Illegal activity observable?
• Field vs. Lab data
• Problems with field data
• Observable actions
• Low frequency of policy changes
• Issues for experimental design
• Parallelism

4
Topic 1 Tax Amnesties

• Issues for tax amnesties
• Revenue source do amnesties generate revenues?
• Post amnesty compliance expectations of future
  amnesties)?
• Future enforcement effort effects?
• Anticipation effect in absence of past amnesties
  in the present jurisdiction?
• Reference Alm, McKee, and Beck, 1990 NTJ

5
Amnesty Experiments

• Basic design individuals receive income,
  disclose income voluntarily, pay tax on disclosed
  income, face random audit with penalty for taxes
  not paid
• Experiment treatments amnesty (y/n),
  post-amnesty enforcement effort, announcement of
  no future amnesty (y/n), announcement of possible
  amnesty (y/n)
• Amnesty allows for payment of past due taxes
  without additional penalty

6
Amnesty Experiments

• Results
• Amnesty without increased enforcement lowers
  future compliance
• Anticipation effect lowers compliance
• Enforcement effort increase after an amnesty
  increases future compliance
• Revenue effects amount collected increases when
  higher post-amnesty enforcement effort is
  announced

7
Amnesty Experiments
8
Amnesty Experiments
9
Amnesty Experiments
10
Topic 2 Enforcement Effort

• Economics of crime approach to evasion
  increased enforcement effort leads to higher
  compliance
• Issues
• Effect of increased enforcement
• Uncertainty in enforcement

11
Enforcement Effort

• Field experiments (Slemrod et al)
• Minnesota experiment approximately 2000
  taxpayers in Minnesota get letter from Revenue
  Dept letter states return subject to close
  examination
• Analyze effect on compliance but do not have
  post audit data and infer compliance from income
  reporting levels
• Result low- and middle-income taxpayers
  increased reported levels of income but
  high-income taxpayers reduced their reported
  income

12
Enforcement Effort

• Lab Experiments
• Alm, Jackson and McKee (1992a and 1992b) test
  effects of increased enforcement
• Alm, Jackson and McKee (1992a) test effects of
  enforcement uncertainty
• Alm and McKee (2005) test the Slemrod et al
  setting in lab where we know the actual
  compliance rate

13
Enforcement Effort

• Summary of lab results
• AJM (1992b) report an increase of 10 percent in
  the audit rate will increase compliance by 2
  percent but that the elasticity on the fine rate
  is extremely small (though significant)
• AJM (1992a) report that uncertainty in the fine
  and audit rates significantly increases
  compliance
• Alm and McKee (2005) report that being audited is
  the focal event rather than the penalty

14
Enforcement Effort

• Results from uncertainty experiments

15
Topic 3 Endogenous Enforcement

• Use information provided in tax return to select
  for audit
• Increase audit productivity through selection
  process
• Unintended consequence is potential for
  coordination among taxpayers if the tax agency
  employs a rule based on comparing returns across
  individuals

16
Endogenous Enforcement

• Alm, Cronshaw and McKee (1993) compare
  performance of various endogenous enforcement
  strategies versus random audits
• Results summarized
• Cutoff rule generates highest tax revenues but
  requires many audits so revenue/audit ratio is
  1.86
• Conditional back audits have tax revenue/audit
  ratio of 6.77
• Random audits perform worst in terms of revenues
  but low probability settings yield highest
  revenue to audit ratio (8 for 5 audit
  probability)

17
Endogenous Enforcement

• Coordination game outcome?
• Alm and McKee (2004) look at the use of
  endogenous selection via a DIF-type rule as used
  by the IRS
• Results
• Cheap talk among taxpayers results in
  coordination outcome at low levels of compliance
• Combining DIF with random audit can overcome the
  coordination effect

18
Endogenous Enforcement
19
Topic 4 Indirect Effects

• Effects of audits reach beyond the taxpayers
  actually audited what is the causal effect?
• Field data (Dubin, Graetz and Wilde, 1990) report
  a ripple effect of 6
• Experiments (Alm, Jackson and McKee, 2005) yield
  a ripple effect of about 4
• Question how does this effect arise?
• Question can policy makers increase the effect?

20
Indirect Effects

• Experiment Setting and Features
• Parallelism Emphasis
• Tax Language screen refers to tax form, tax
  rate, audit and so on
• Earn income
• Individual audits
• Treatments are Communication and Official
  Announcements

21
Indirect Effects Experiment Details

• Instructions, practice rounds and questions
• Earn income figure 1
• Pay taxes (Disclose Income) figure 2
• Timed stage failure to file results in 10
  penalty plus automatic audit
• Undergo individual audit
• Bingo cage on screen after the filing decision
• Receive feedback information re own and others
  audit
• Round ends
• Treatments Information Official vs. Unofficial
  
• What we dont include here
• Public goods tax receipts in black hole
• Systematic (endogenous) audits
• Varying tax rates or penalty rates

22
Income Earning Image
23
Tax Form Image
24
Indirect Effects
25
Indirect Effects

• Official provision of previous audit information
  by the tax authority has negative effect on
  subsequent compliance
• Provision of unofficial information (allowed
  communication) by taxpayers themselves increases
  compliance
• Communication that reports evasion leads to lower
  compliance while communication that reports
  compliance leads to higher compliance norms
• Communication that reports not being audited has
  negative effect on compliance while communication
  that reports being audited has a positive effect
  effective enforcement

26
Indirect Effects
27
Conclusions

• What have we learned?
• Deterrence can work
• Amnesties are useful if accompanied by
  enforcement increases
• Endogenous audit selection is generally useful
  but must be accompanied by a random audit
• Audit information affects compliance
• What can we learn?
• Third party reporting
• Portfolio evasion
• Taxpayer-taxpayer communication