Tax Evasion

Background
Theory usually assumes perfect tax enforcement, i..e zero tax evasion at zero cost. In practice,

  • Tax enforcement is costly (~10% of taxes collected in the US) for both the government and private agents
  • Tax evasion is substantial in most countries (~15% of income in US federal tax) and particularly large in developing countries

Allingham-Sandmo

Formulation:
Original AS assumption: penalty = theta*e

  • Penalty is a fraction of undeclared income

Under Yitzhaki (1974): penalty = theta*te

  • Penalty is a fraction of undeclared income and tax rate

Comparative Statics


  • Evasion is decreasing in probability of detection
  • Evasion is decreasing in penalty rate
  • If ARA is decreasing, evasion is increasing in true income
  • If ARA if decreasing, evasion is decreasing in tax rate
  • Under 4th result, as tax increases, evasion decreases
  • This may seem counterintuitive, however it is because under Yitzhaki model, penalty rate is expressed as a function of tax
  • Thus, with higher taxes, you have higher penalty rate which decreases evasion
  • In reality, evasion increases when taxes increase, even if penalties increase
  • Thus, the only way to remove this counter-intuitive result is to impose a non-linear penalty function

Compliance Puzzle

  • AS model states that, around e=0 (i.e. undeclared income = 0), and marginal expected utility of evasion >0, then people will evade
  • Compliance puzzle: AS model predicts that everybody evades taxes, but most people do not do so
  • 3 possible reasons: third-party information reporting, misperception, psychology and culture

Third-Party Information Reporting

  • AS model assumes pure self-reporting, but modern tax systems use extensive third-party reporting by firms
  • E.g., not only do you report your income to HMRC, your employer also reports your income to HMRC. Thus, it becomes more difficult to evade taxes.
  • If there is no collusion between taxpayer and the third party, matching of tax returns and third-party information reveal evasion. Thus, observed rate of audit is far below the rate of detection (which is close to 1).
  • When taxpayers have both third-party and self-reported income, detection probability is endogenous to under-reported income
  • According to IRS estimates, if there is substantial information reporting, tax evasion is 1%. If there is little or no information reporting, tax evasion is 63%.

Misperception

  • AS model assumes perfect knowledge of deterrence parameters, but taxpayers may overestimate deterrence
  • In other words, taxpayers think it is more difficult to evade taxes, or that there are higher penalties, than there actually is

Psychology and Culture

  • Morals, norms, reciprocity, guilt, shame

Kleven et al (2011): Third Party Reporting Model

Set-up: Now, detection probability is endogenous i.e. p(e), where detection probability is a function of undeclared income

  • Thus, evasion is now determined by the AS model*(1+ elasticity)
  • Elasticity = 1 extra dollar of evasion implies a higher detection probability on the infra-marginal units of evasion

Changes:

  • If elasticity is high (1 extra dollar leads to a much higher detection probability), this can continue to allow for low values of detection probability + penalty rate, and still be consistent with low evasion (i.e. compliance puzzle in AS model)
  • Third-party reporting causes p(e) to be very low for self-reported income and very high for third-party reported income

Predictions

Evasion rate is 0 for third-party reported income, large for self-reported income.

  • This implies that total evasion rate is small with extensive third-party reporting

Evasion effect of increased tax rates is 0 for third-party reported income, but possibly significant for self-reported income.

Evasion effect of increased enforcement (e.g. probability of detection, penalty rate) is 0 for third-party reported income, but significant for self-reported income.

  • Total enforcement effect is small with extensive third-party reporting.

International Tax Evasion

  • Policy interest in recent years on use of offshore accounts to evade taxes
  • Problem: No third party information from offshore banks
  • Goal: Improve third-party information
  • Requires international coordination
  • Most people hiding wealth are extremely wealthy people
  • More work is needed to understand and reduce the sometimes-large compliance costs of these information reporting regimes

Note that in some countries, share of financial wealth held offshore is extremely high. Some of this is to "protect" the individuals' wealth.