Upload
vudan
View
220
Download
0
Embed Size (px)
Citation preview
The Occurence of Tax Amnesties: Theory and Evidence
Ralph-C. Bayer, Harald Oberhofer and Hans Winner
Journal of Public Economics (2015)
Presented byOnursal Bagırgan
Tax Amnesties: Theory and Evidence February 22, 2017 1 / 20
Introduction
Tax amnesty is a limited-time opportunity for a specified group oftaxpayers to pay a defined amount, in exchange for forgiveness of atax liability (including interest and penalties) relating to a previoustax period or periods and without fear of criminal prosecution.
A fiscal tool to deal with mounting public deficits
Frequently used, especially after the recent crisis
Not all amnesties raised considarable revenues
Tax Amnesties: Theory and Evidence February 22, 2017 2 / 20
Literature on Tax AmnestiesFiscal impacts of tax amnesties
Evidence from US: Mikesell (1986), Fisher at al. (1989), Alm andBeck (1991,1993), Luitel and Sobel (2007), Mikesell and Ross (2012)
Outside of US: Uchitelle (1989), Cassone and Marchese (1995), Almet al. (2000), Lopez-Labardia and Rodrigo (2003), Baer and LaBorgne (2008)
Tax Amnesties: Theory and Evidence February 22, 2017 3 / 20
Literature on Tax AmnestiesTaxpayer participation
Andreoni (1991): Consumption shock before amnesty
Malik and Schwab (1991): Learning risk preferences once an amnestyis offered
Andreoni (1992): Consumption smoothing
From criminology point of view
Gottfredson and Hirschi (1990), Nagin and Pogarsky (2004):Neglecting future bad consequences
Tax Amnesties: Theory and Evidence February 22, 2017 4 / 20
The paper
A theoretical model:
A new taxpayer motive: Signal of imminent detection
Example: Purchase of Swiss banking data by German authorities
Effect of taxpayers’ initial expectations and strategic deliquency
Ross and Buckwalter (2013): 4.3-16.5% of US State tax amnestyrevenues
Testing of the model:
A panel of US state amnesties 1981-2011
Tax Amnesties: Theory and Evidence February 22, 2017 5 / 20
Main Finding
The self-fulfilling character of tax amnesties:
Higher expectation of a tax amnesty=⇒ Higher tax evasion=⇒ Lower tax revenues=⇒ Higher necessity of tax amnesty
Tax Amnesties: Theory and Evidence February 22, 2017 6 / 20
The ModelTiming
1 Nature determines income y , the distribution of gross income is q(y).
2 The taxpayer learns y , declares income d∈ [0, y ] and pays tax td
3 Nature decides if some shock occurs that makes future detection acertainty, where s ∈ {0, 1}. Prob(s = 1) = σ.
4 The government decides if there is an amnesty. The probability is φ.
5 If there is an amnesty, then the taxpayer may declare additionalincome as as∈ [0, y − d ] and pays additional taxes tas .
6 Nature determines whether the true income is verifiable (withprobability p). The taxpayer pays fines F (y , d , as) if that happens.
Taxpayer’s expected utility:
EU(d , as) = U1(y , d) + δEU2(y , d , as , φ, σ)
Tax Amnesties: Theory and Evidence February 22, 2017 7 / 20
Taxpayer’s ProblemDecision to participate tax amnesty
Detection shock received:
EU2(as |α = 1, s = 1, y , d) =− F (y − d − a1)− ta1
=⇒ a∗1 = y − d
Detection shock not received:
EU2(as |α = 1, s = 0, y , d) =p[−F (y − d − a0)− ta0] + (1− p)[−ta0]=⇒ pF ′(y − d − a∗0) = t
Notice thatda∗1dd
=da∗0dd
= −1
Tax Amnesties: Theory and Evidence February 22, 2017 8 / 20
Taxpayer’s ProblemDecision to declare income
EU(d , a∗1, a∗0) =y − td
+ δ[σEU2(d , s = 1, a∗1(d)) + (1− σ)EU2(d , s = 0, a∗0(d))]
where
EU2(d , s = 1, a∗1) =− φ(y − d)t − (1− φ)(F (y − d))
EU2(d , s = 0, a∗0) =− φ(pF (y − d − a∗0) + a∗0t)− (1− φ)pF (y − d)
Tax Amnesties: Theory and Evidence February 22, 2017 9 / 20
Taxpayer’s ProblemDecision to declare income
The optimal decision is given by
F ′(y − d∗) =t(1− φδ)
δ(1− φ)(p(1− σ) + σ)
Lemma (1)
For δ < 1 at any interior solution,dd∗
dφ< 0 holds. Also notice that,
dd∗
dt< 0,
dd∗
dp> 0,
dd∗
dσ> 0,
dd∗
dy> 0
Tax Amnesties: Theory and Evidence February 22, 2017 10 / 20
Government’s ProblemGovernment’s benefit function
Government’s valuation is B(G , ρ)
G is government spending
ρ is the state of the economy
Assumptions on valuation
∂2B(G , ρ)
∂G 2≤ 0
∂2B(G , ρ)
∂G∂ρ> 0
Then, the benefit of an amnesty is
∆B = B(t(D + A), ρ)− B(tD, ρ)
Tax Amnesties: Theory and Evidence February 22, 2017 11 / 20
Government’s ProblemProperties of benefit function
∆B = B(t(D + A), ρ)− B(tD, ρ)
Lemma (2)
The benefit function has the following properties.
∂∆B
∂D= −t ∂B(tD, ρ)
∂G∂B(tD, ρ)
∂G≤ 0 =⇒ ∆B ≤ 0
∂∆B
∂ρ≥ 0
Tax Amnesties: Theory and Evidence February 22, 2017 12 / 20
Government’s ProblemAmnesty decision
The cost of amnesty is θ
θ is drawn from a distribution with cdf H(θ)
Government declares amnesty ⇐⇒ θ ≤ ∆B
The probability of an amnesty µ(∆B(D)) = H(∆B(D)).
Tax Amnesties: Theory and Evidence February 22, 2017 13 / 20
Perfect Bayesian Nash Equilibrium
Sequential Rationality: Already done
Consistency of beliefs:
Z (φ) = µ(∆B) ◦ ∆B(D) ◦D∗(φ)
Tax Amnesties: Theory and Evidence February 22, 2017 14 / 20
Perfect Bayesian Nash Equilibrium
Tax Amnesties: Theory and Evidence February 22, 2017 15 / 20
Testable Hypotheses from the Model
Tax amnesties have a self-fulfilling characteristic.
Higher ∆B, higher the occurrence probability of amnesty.
Effect of t is ambiguous.
Higher ρ, higher the occurrence probability of amnesty.
Lower θ, higher the occurrence probability of amnesty.
Tax Amnesties: Theory and Evidence February 22, 2017 16 / 20
Empirical Analysis
Tax Amnesties: Theory and Evidence February 22, 2017 17 / 20
Empirical Analysis
Tax Amnesties: Theory and Evidence February 22, 2017 18 / 20
Empirical Analysis
Tax Amnesties: Theory and Evidence February 22, 2017 19 / 20
Conclusion
Theoretical model:
Tax amnesties have self-fulfilling characteristic: Committing not toenact amnesties might be beneficial.
Model allows multiplicity of equilibria, i.e. different tax evasions arepossible for similar economies.
Testable hypotheses of theoretical model generally matches well withthe evidence from data.
Tax Amnesties: Theory and Evidence February 22, 2017 20 / 20