EC 351 LN 6 1 Optimal Taxation Commodity-H

Download as pdf or txt
Download as pdf or txt
You are on page 1of 19

Optimal Taxation: Commodity Taxation

EC 351 Public Finance

Yeliz Kaçamak

Boğaziçi University

1
Optimal Taxation

How can there be optimal features of anything so awful as


taxation?
Think of going to the dentist. An optimal dentist causes the least
possible pain and harm while doing the job that needs to be done.
It is the same with taxation. Optimal taxation has the feature of
raising the money that the government needs, in a fair and
equitable way, while causing the least possible associated problems
(deadweight loss) for the economy.
The study of optimal taxation helps identify features of desirable
tax systems.
Note: actual tax systems often differ from optimal tax systems.

2
Efficient tax systems

Optimal taxation typically entails trading off efficiency and


distributional considerations.
It is helpful, however, to start with the features of efficient tax
systems.
Consider the following situation:

• A government needs to raise a given amount of revenue to


pay for public expenditures.
• All consumers are identical.
• The government is limited to taxing sales of different
commodities.

Then, should the government tax all commodity sales at the same
rate?
3
Optimal Commodity Taxes

Answer: No! Commodity taxes should vary with supply and


demand conditions.
If there is one commodity that is inelastically demanded or
supplied, then it should be taxed extremely heavily – if possible, all
the needed tax revenue should come from taxing it.
Such a tax would entail zero deadweight loss.
Can this insight be generalized? Is there a general rule that
illustrates how taxes should be related to supply and demand
conditions, for taxes that are chosen to minimize deadweight loss?

4
The Ramsey Rule

Frank P. Ramsey (1903-1930) proved that the popular intuition


that all commodities should be taxed at the same rate was wrong.
Consider the impact of increasing the tax on a single commodity.
There are two effects:

• tax revenue rises.


• deadweight loss rises.

The tax rates are optimal ⇐⇒ for a given change in tax revenue
(associated with higher tax rates), the tax change produces the same
increment to deadweight loss, regardless of which tax rate is increased.
If not : deadweight loss could be reduced by lowering one tax rate and
raising another, while leaving tax revenue unchanged.
This can be summarized as: 4DWL = m4(TaxRevenue), with m a
constant. This equation must hold for all taxed commodities.
5
Formal Problem: Assumptions

A1 There is a revenue target, denoted by R


A2 All consumers are identical (representative consumer)
A3 Gov’t can only levy commodity taxes. There are k number of
goods.
A4 All commodity markets are perfectly competitive, i.e., εs = ∞
(⇒ all burden is on consumers)
A5 εD of each commodity is known
A6 Cross-price elasticities are 0

6
Government’s Problem

Consider one (representative) consumer who consumes k different


goods. Government chooses the tax rates across goods, {τi }ki=1 , to
minimize deadweight loss for a given government revenue
requirement.
Total DWL:
k
X k
X
DWL = DWL(τi ) = DWLi
i=1 i=1

Total Tax Revenue:


k
X k
X
TR = R(τi ) = Ri
i=1 i=1

7
Government’s Problem

k
X
min DWLi
{τi }ki=1
i=1
Xk
s.t. Ri = R (λ)
i=1

λ is the value of additional government revenues: The value of having


another dollar in the government’s hands relative to its next best use in
the private sector.
If λ is large, it implies that additional government revenues are quite
valuable relative to keeping the money in private hands;
if λ is small, then additional government revenues have little value
relative to the value

8
Government’s Problem

Lagrangian and the FOC

k
X k
X
L: DWLi + λ(R − Ri ) (1)
i=1 i=1
∂L ∂DWLi ∂Ri
: −λ =0 (2)
∂τi ∂τi ∂τi
Then, at the optimum it must be the case that:

∂DWLi
∂τi
∂Ri
= λ (3)
∂τi
MDWLi
⇒ = λ (4)
MRi
9
Remark

MDWLi
Before moving on note that the equation MRi = λ has to hold
for all markets. Which in turn implies
MDWL1 MDWL2 MDWLk
= = ... =
MR1 MR2 MRk

MDWL1 MDWL2
Why can’t we have MR1 > MR2 at the optimal?

• This implies that market 1 creates more DWL per an


additional dollar of tax revenue
• Then, the government decrease taxes in market 1 and increase
taxes in market 2 to the point marginal ratios are equal. They
will still raise the same tax revenue but will cause less DWL in
total.

10
Deriving the Ramsey Rule

Recall that:
1 εS · εD Q
DWL = − · · (τ )2
2 εS − εD p
and
TR = Qi · τi
Then:
MDWLi = − εD · τ · Qi
MRi = Qi
Equation (2) then implies:
−εDi · τ · Qi
= λ
Qi
(5)
1
τi∗ = − λ
εDi
This is also called the inverse elasticity rule. (A4 is crucial)
11
Inverse Elasticity Rule

1
τi∗ = − λ
εDi

τi∗ is the optimal tax, εDi is the elasticity of demand, and λ is the
shadow value.
Optimal taxation therefore balances two rules:

• Elasticity rule: Lower taxes on goods with more elastic demand.


• Broad base rule: Better to tax a wide variety of goods moderately
than few goods heavily.

Also note that as λ increases the taxes on each market has to increase.
Why?

12
Interpreting the Ramsey Rule

The Ramsey Rule with these simplifications is known as the inverse


elasticity rule. (The more general rule does not rule out cross-price
elasticities)
The more elastic is demand or supply, the greater a change in Q
will accompany a given change in τ . If this effect is large, the tax
rate should be small.

• Take two commodities with (identical) infinitely elastic supply


elasticities, and differing demand elasticities.
• Optimal tax rates on the two commodities are inversely proportional
to their demand elasticities.
• (For example, if one commodity has three times the demand
elasticity, it should have a tax rate one-third as high.)

13
Equity Implications of the Ramsey Model

The Ramsey Rule is derived in a setting in which all consumers are


identical, so everyone has an interest in efficient pricing.
In life, consumers are not identical, and governments design tax
systems to do a form of redistribution, by imposing heavier tax
burdens on wealthier consumers.

14
Equity Implications of the Ramsey Model

Imagine that the government had only two goods it could tax:
bread and caviar:
Elasticity of demand for caviar is much higher than that for bread.
The inverse elasticity rule would suggest that the government tax
bread much more highly than caviar.
This means taxing more heavily the good consumed by poor
people.
This might hurt vertical equality.

15
Modified Ramsey Rule

It is very simple to modify the Ramsey Rule to accommodate


distributional concerns.
The resulting “Many-person Ramsey Rule” then indicates that tax
rates should be highest on commodities that are inelastically
demanded and/or purchased (or sold) by people with the highest
incomes.

16
Modified Ramsey Rule

The theory of optimal taxation teaches us that you have to care


about efficiency even if your primary goal is distribution, and you
have to care about distribution even if your primary goal is
efficiency.

• Perfectly efficient taxes (for example, head taxes) usually have bad
distributional features.
• Maximally redistributive taxes (for example, income taxes with very
high tax rates on the rich and negative tax rates on the poor)
usually have bad efficiency properties.

17
Many Person Ramsey

Supose there are two people with utilities Uh (x1 , x2 ) and


Ul (x1 , x2 ). Suppose also the government’s SWF is denoted by W .
min W (Uh , Ul )
{ti }ki=1
2
X 2
X
s.t. ti x1i + ti x1i = R (λ)
i=1 i=2

At the end, rather than all goods having their compensated


demand reduced in the same proportion, equity results in the
goods consumed primarily by the poor facing less of a reduction.
In simple terms, this should translate into lower rates of tax on the
goods consumed by the poor relative to those determined solely by
efficiency. Equity therefore succeeds in moderating the hard edge
of the efficient tax structure.
18
References

Jonathan Gruber, Public Finance and Public Policy, Fifth Edition, 2019
Worth Publishers, Chapter 20
Ramsey, Frank P. “A Contribution to the Theory of Taxation.” The
Economic Journal 37.145 (1927): 47-61.(web)

These lecture notes are heavily influenced by Jim Hines’s undergraduate


PF course.

19

You might also like