The Atkinson–Stiglitz theorem is a theorem of public economics. It implies that no indirect taxes need to be employed where the utility function is separable between labor and all commodities. Non-linear income taxation can be used by the government and was in a seminal article by Joseph Stiglitz and Anthony Atkinson in 1976.[1] The Atkinson–Stiglitz theorem is an important theoretical result in public economics, spawning a broad literature that delimited the conditions under which the theorem holds. For example, Emmanuel Saez, a French-American professor and economist demonstrated that the Atkinson–Stiglitz theorem does not hold if households have heterogeneous preferences rather than homogeneous ones.[2] [3]
In practice, the Atkinson–Stiglitz theorem has often been invoked in the debate on optimal capital income taxation. As capital income taxation can be interpreted as the taxation of future consumption over the taxation of present consumption, the theorem implies that governments should abstain from capital income taxation if non-linear income taxation were an option since capital income taxation would not improve equity by comparison to the non-linear income tax, while additionally distorting savings.
For an individual whose wage is
w
\sumjqjxj=\sumj(xj+tj(xj))=wL-T(wL) ,
qi
xi
i
Uj=
(1+t'j)(-UL) | |
w(1-T') |
(j=1,2,...,N).
infty | |
\int | |
0 |
\left[wL-\sumjxj-\overline{R}\right]dF=0 .
f
H=\left[G(U)-λ\left\lbracewL-\sumjxj-\overline{R}\right\rbrace\right]f-\mu\thetaUL .
xj
-λ\left[\left(
\partialx1 | |
\partialxj |
\right)U+1\right]-
\mu\theta | |
f |
\left[
\partial2U | |
\partialx1\partialL |
\left(
\partialx1 | |
\partialxj |
\right)U+
\partial2U | |
\partialxj\partialL |
\right]=0 .
Then the following relation holds:
\left(
\partialx1 | |
\partialxj |
\right)U=-
Uj | |
U1 |
=-
1+t'j | |
1+t'1 |
.
λ\left[
1+t'j | |
1+t'1 |
-1\right]=
\mu\thetaUj | |
f |
\left[
\partial2U | |
\partialL\partialxj |
⋅
1 | |
Uj |
-
\partial2U | |
\partialL\partialx1 |
⋅
1 | |
U1 |
\right]=
\mu\thetaUj | |
f |
\partial | |
\partialL |
\left(ln{Uj
λ\left[
1+t'j | |
1+t'1 |
-1\right]=
\mu\thetaUj | |
f |
\partial | |
\partialL |
\left(ln{
Uj | |
U1 |
}\right) .
t'1
t'1=0
Uj=(1+t'j)\alpha
t'j | |
1+t'j |
=
\mu\theta\alpha | |
λf |
\partial | |
\partialL |
\left(ln{
Uj | |
U1 |
}\right) .
tj=0
Joseph Stiglitz explains why indirect taxation is unnecessary, viewing the Atkinson–Stiglitz theorem from a different perspective.[4]
Suppose that those who are in category 2 are the more able. Then, two conditions are imposed for Pareto efficient taxation at which a government aims. The first condition is that the utility of category 1 is equal to or more than a given level:
\overline{U}1\leV1(C1,Y1) .
R
\overline{R}
R=-(C1-Y1)N1-(C2-Y2)N2 ,
\overline{R}\leR ,
N1
N2
V2(C2,Y2)
l{L}=V2(C2,Y2)+\muV1(C1,Y1)+λ2(V2(C2,Y2)-V2(C1,Y1))+λ1(V1(C1,Y1)-V1(C2,Y2))+\gamma\left(-(C1-Y1)N1-(C2-Y2)N2-\overline{R}\right) ,
\mu
\partialV1 | |
\partialC1 |
-λ2
\partialV2 | |
\partialC1 |
+λ1
\partialV1 | |
\partialC1 |
-\gammaN1=0 ,
\mu
\partialV1 | |
\partialY1 |
-λ2
\partialV2 | |
\partialY1 |
+λ1
\partialV1 | |
\partialY1 |
+\gammaN1=0 ,
\partialV2 | |
\partialC2 |
+λ2
\partialV2 | |
\partialC2 |
-λ1
\partialV1 | |
\partialC2 |
-\gammaN2=0 ,
\partialV2 | |
\partialY2 |
+λ2
\partialV2 | |
\partialY2 |
-λ1
\partialV1 | |
\partialY2 |
+\gammaN2=0 .
For the case where
λ1=0
λ2=0
\partialVi/\partialYi | |
\partialVi/\partialCi |
+1=0 ,
i=1,2
λ1=0
λ2>0
\partialV2/\partialY2 | |
\partialV2/\partialC2 |
+1=0 ,
\partialV1/\partialY2 | |
\partialV1/\partialC1 |
=-
1-λ2(\partialV2/\partialY1)/N1\gamma | |
1+λ2(\partialV2/\partialC1)/N1\gamma |
.
\deltai=
\partialVi/\partialY1 | |
\partialVi/\partialC1 |
, (i=1,2)
\delta1+1
Also, note the following equation:
\delta1=-\left(
1-\nu\delta2 | |
1+\nu |
\right) ,
\nu
\nu=
λ2(\partialV2/\partialC1) | |
N1\gamma |
.
\delta1<\delta2
-1<\delta1<\delta2
For the case where
λ1>0
λ2=0
Consider a case where income level and several commodities are observable. Each individual's consumption function is expressed in a vector form as:
bf{C}1=\sumjC1jbf{e}j
bf{C}2=\sumjC2jbf{e}j .
R\leq
2 | |
\sum | |
k=1 |
(YkNk)-N1\sumjC1j-N2\sumjC2j .
\mu
\partialV1 | |
\partialC1j |
-λ2
\partialV2 | |
\partialC1j |
+λ1
\partialV1 | |
\partialC1j |
-\gammaN1=0 ,
\mu
\partialV1 | |
\partialY1 |
-λ2
\partialV2 | |
\partialY1 |
+λ1
\partialV1 | |
\partialY1 |
+\gammaN1=0 ,
\partialV2 | |
\partialC2j |
+λ2
\partialV2 | |
\partialC2j |
-λ1
\partialV1 | |
\partialC2j |
-\gammaN2=0 ,
\partialV2 | |
\partialY2 |
+λ2
\partialV2 | |
\partialY2 |
-λ1
\partialV1 | |
\partialY2 |
+\gammaN2=0 .
λ1=0
λ2>0
| ||||||
|
=1 ,
| ||||||
|
=1 .
\partial2Uk | |
\partialCkj\partialLk |
=0 ,
\partialV1 | |
\partialC1j |
=
\partialV2 | |
\partialC1j |
.
| ||||||
|
=1 .
Consider a scenario in which individuals with high abilities, who typically earn higher incomes as a reflection of their skills, downplay their abilities. In this case, it could be argued that the government needs to randomize the taxes imposed on the low ability individuals, to increase the effectiveness of screening. It is possible that under certain conditions the taxes can be randomized without damaging the low-ability individuals. For the case where an individual chooses to show their ability, a tax schedule is related to
\lbrace
* | |
C | |
2 |
,
* | |
Y | |
2 |
\rbrace
\lbrace
* | |
C | |
1 |
,
* | |
Y | |
1 |
\rbrace
\lbrace
** | |
C | |
1 |
,
** | |
Y | |
1 |
\rbrace
To avoid hitting the low ability group, the mean consumption must be shifted upwards at each
Y
\overline{C}1
\overline{Y}1
* | |
C | |
1 |
=\overline{C}1+h ,
* | |
Y | |
1 |
=\overline{Y}1+λh
** | |
C | |
1 |
=\overline{C}1-h ,
** | |
Y | |
1 |
=\overline{Y}1-λh .
The utility function is
V2(
* | |
C | |
1 |
,
* | |
Y | |
1 |
)
V2
** | |
(C | |
1 |
,
** | |
Y | |
1 |
)
V | |
2C* |
(d\overline{C}1+dh)+
V | |
2Y* |
(d\overline{Y}1+λdh)+
V | |
2C** |
(d\overline{C}1-dh)+
V | |
2Y** |
(d\overline{Y}1-λdh)=0 ,
V | |
1C* |
(d\overline{C}1+dh)+
V | |
1Y* |
(d\overline{Y}1+λdh)+
V | |
1C** |
(d\overline{C}1-dh)+
V | |
1Y** |
(d\overline{Y}1-λdh)=0 .
\begin{bmatrix} SV2&SV2\\ SV1&SV1\end{bmatrix} \begin{bmatrix} d\overline{C}\\ d\overline{Y} \end{bmatrix}=-\begin{bmatrix} DV2C+λDV2\\ DV1C+λDV1\end{bmatrix}dh ,
SVk=
V | |
kC* |
+
V | |
kC** |
SVk=
V | |
kY* |
+
V | |
kY** |
k=1,2
DVk=
V | |
kC* |
-
V | |
kC** |
DVk=
V | |
kY* |
-
V | |
kY** |
Then:
\limh
d(\overline{Y | |
- |
\overline{C})}{dh}=
F1-F2 | |
(-2)(MRS1-MRS2) |
,
MRSk=-(
\partialVk | |
\partialC1 |
)-1
\partialVk | |
\partialY1 |
F1,F2
F1=(
\partialV2 | |
\partialC1 |
)-1M2(1-MRS1)
F2=(
\partialV1 | |
\partialC1 |
)-1M1(1-MRS2)
Mk
Mk=DVk+λDVk
\overline{Y}-\overline{C}
h
h=0
Mk=0
d2(\overline{Y | |
- |
\overline{C})}{dh2
H1=
d(F1-F2) | |
dh |
1 | |
-2(MRS1-MRS2) |
H2=(-1)
d(\overline{Y | |
- |
\overline{C})}{dh}
dln{(-2)(MRS1-MRS2) | |
}{d |
h}
H2
h=0
d2(\overline{Y | |
- |
\overline{C})}{dh2
I1=(V2CC+2λV2+λ2V2)(
\partialV2 | |
\partialC1 |
)-1(1-MRS1)
I2=(-1)(V1CC+2λV1+λ2V1)(
\partialV1 | |
\partialC1 |
)-1(1-MRS2)
Since
MRS2<MRS1<1
(V2CC+2λV2+λ2V2)(
V | |
1C1 |
+
V | |
2Y1 |
)-(V1CC+2λV1+λ2V2)(
V | |
2C1 |
+
V | |
2Y1 |
)<0 .