Once Upon A Time...Tax Evasion

1
Once Upon A Time...Tax Evasion

• Mirco Tonin
• University of Southampton
• Warsaw, November 30th, 2007

2
Tax enforcement has always been an issue
the arrears which were due to the King from
the people who are in Egypt and all those who are
subject to his kingship, and (which) amounted to
a large total, he renounced
3
Many Perspectives

• Equity
• Efficiency
• Incidence

• Public Finance
• Law Enforcement
• Organizational Design
• Labour Supply
• Ethics

4
Economic Analysis

• Tax Evasion is choice under uncertainty
  (Allingham and Sandmo, 1972)
• BUT
• Some counterintuitive results (higher t, lower
  evasion)
• Issue becomes Why do people pay taxes?

5
Why do people pay taxes?

• No choice information reporting / withdrawal at
  source
• Desire to contribute to public good / honesty
• Overweight probability of detection

6
Data Sources

• Audit - Taxpayer Compliance Measurement Program
  by IRS complete detection?
• Survey
• Tax Amnesty
• Lab Experiments

7
Some Evidence

• US TCMP (Kleeper and Nagin, 1989)
• Look at declaration Huge differences in
  non-compliance
• 0.1 wages, salaries
• 74.4 rents, royalties, partnership income

8
EU survey on undeclared work

• What are the main reasons for under-declared
  work?
• General anwers in Hungary
• lower price of goods and services (59)
• helping someone in need of money (20)
• Answers from suppliers of undeclared work in
  Hungary
• Seasonal work not worth declaring (55)
• Both parties benefitted from it (51)
• High level of taxes and social security
  contributions
• (36)
• Person who acquired it insisted on
  non-declaration (28)
• Could not find a regular job (27)
• Undeclared work is common practice (21)
• Could get a higher fee for work (19)
• Unsatisfiesd with public services (15)

9
Lemieux, Fortin, Fréchette, The effect of taxes
on labor supply in the underground economy, AER
84(1)

• Use survey data for Quebec city, Canada, to study
  determinants of labour supply decisions in
  underground economy
• Information on hours of work and wages/earnings
  in both sectors
• Underground labor-market activity concentrated
  among people at the low end of income
  distribution
• Hours worked in underground sector depend
  negatively on wage rate in either regular or
  underground sector, with a large negative
  elasticity with respect to wage rate in regular
  sector
• Tax and transfer system does not seems to
  significantly distort allocation of hours of work
  from regular to underground sector.

10
Other Methods (1)

• Comparison of income and participation data from
  different sources
• Fiorio and DAmuri, 2005 survey-based data set
  and tax forms.
• Main assumption in survey something "close" to
  true income is declared.
• Focus on employment and self-employment.
• Analysis by deciles employment income evaded
  mainly in bottom deciles, with high rates of
  evasion low in the distribution and close to zero
  a median. Self-employment income also evaded
  mainly at lower levels of income, but evasion is
  present across the whole distribution

11
Other Methods (2)

• Comparison of income and consumption data from
  same source
• Pissarides and Weber, 1989
• Gorodnichenko and Sabirianova, 2006
• Feldman and Slemrod, 2007

12
What can we do?

• Optimal Auditing
• Design of Incentives for Tax Administration
• Presumptive Taxation

13
Presumptive Taxation

• Rebuttable vs Irrebuttable
• Minimum Tax vs Exclusive
• Mechanical vs Discretionary

14
The Bulgarian Case

• Problem
• massive underreporting of labour compensation

• Two changes in labour regulation in 2003
• compulsory registration of all concluded,
  amended, or terminated contracts
• introduction of minimum social security thresholds

15
Minimum Social Security Thresholds

• Varying according to
• occupational group (9 categories, e.g.
  administrative staff / service workers and sale
  workers)
• sector (48 in 2003, 73 at present, e.g.
  processing and preserving of fruit and vegetables
  / manufacture of dairy products)
• Potentially, 657 different MSIT

16
MINIMUM 180 leva per month (statutory minimum
wage) MAXIMUM 851 leva per month (management in
manufacture of coke, refined petroleum products
and nuclear fuel)
17
Minimum Social Security Thresholds (2)

• Negotiated each year with social partners (in
  2005 for 48 out of 68 sectors)

• If no agreement, fixed administratively

• MSITs negotiated with social partners usually
  become sectoral minimum wages through extension
  of collective agreements

• No systematic evaluation, but apparently
  successful in reducing underreporting

18
Conflict of Interest

• Recent Proposal by Industrial Association
• Workers liable to a fine in case of
  underreporting
• Exemption from the fine and additional protection
  against unfair dismissal if worker denounces
  illegal payments to authorities

19
The Italian Case

• Problem
• massive underreporting by small and medium
  enterprises, self-employed, and professionals

• Introduction in 1998 of Business Sector
  Analysis
• methodology to estimate revenues and
  compensations
• hybrid between auditing selection mechanism and
  presumptive taxation

20
Main Features

• Elaboration of sector-specific BSA
• 45 in 1998, 200 now, e.g. Retail sale of
  flowers, plants and seeds via permanent or mobile
  stalls/ Tour guide activities/ Dentistry
• participation of trade associations
• publication of results
• regular updating

21
Procedure
1. Estimation of revenues (based on sector/
organizational structure/ reference market/
business model/ geographical location/ structural
and accounting variables)

• 2. Declaration
• Accounting Revenues gt Estimate?
• Taxpayer voluntarily adjust to estimate?
• If yes, taxpayer is consistent, if no
  inconsistent

• 3. Auditing inconsistent taxpayer
• higher probability of an auditing
• reversal of burden of proof

22
Estimation of Revenues - Macro

1. Data Collection - survey of the population of
   interest to collect structural and accounting
   data (e.g. for retail sale of flowers approx. 100
   questions)

• Identification of Sectoral and Geographical
  Clusters (e.g. for retail sale of flowers, 8
  clusters - flower retailers with kiosk operating
  nearby cemeteries)

1. Estimation of Revenue Function regression
   analysis on selected sample relationship between
   revenues and structural and accounting variables

23
Example Retail sale of flowers, plants and
seeds via permanent or mobile stalls
6 sectoral clusters (4 displayed) 2 geographical
clusters 10 potential variables
24
Declaration

• Applied to self-employed, professionals, SME with
  revenues below EUR 5m other causes of exclusion
  (approx. 4m firms)

• Taxpayer fill in data and knows the amount of
  estimated revenues

• If accounting revenues are lower, can decide to
  adjust declaration to match estimate (a 3
  penalty applies for deviations above 10)

• If taxpayer decide NOT to do so, classified as
  inconsistent

25
Auditing

• Inconsistent taxpayers subject to
• higher probability of auditing
• reversal of burden of proof taxpayer has to
  document the reasons why revenues are below
  estimate
• Examples of acceptable reasons
• illness or maternity
• reduced business due to road reconstruction
• Economic marginality of business activity

26
Assessment

• Impact at declaration (2004)
• 69 naturally consistent
• 15 consistent by adjustment gt 3bn EUR
• 16 inconsistent
• evidence of BSA being more effective when
  recently updated and probability of auditing
  higher
• Impact on bookkeeping
• not systematically assessed, but most likely
  important once BSA established