EconPort - Public Finance - Effects of a Per Unit Tax

Summary: Students participate in one or more double auction experiments that include introduction of a per unit tax. Graphical displays of experiment data demonstrate key concepts in the microeconomic analysis of taxes.

Motivation: It can be difficult to incorporate the economic analysis of the effects of taxes into a person's way of thinking about public policy. For example, seeing the amount of retail sales tax added to the prices of goods purchased seems to provide evidence that the burden of the tax is entirely borne by consumers. It is difficult to understand that part of the burden of the tax is borne by sellers because the prices of goods would be higher in the absence of the tax. The double auction experiments with and without a per unit tax make it possible for students to experience the full effects of introducing a tax into a competitive market.

Concepts Covered

Liability to pay a tax
Incidence of a tax
Deadweight loss or excess burden of a tax

Recommended Procedure: The recommended order is to first run the experiments on EconPort either in class (if computers are available) or out of class. Next, cover the lecture material on EconPort. Finally, use the graphical displays of data on EconPort to demonstrate: (a) the deadweight loss from the tax; (b) the independence of incidence of the tax from liability to pay the tax; and (c) the dependence of incidence of the tax on price elasticities of demand and supply.

Experiments

General Considerations

Although hypothetical payoffs can be used, better results are usually achieved by using salient payoffs in a medium valued by the students. These payoffs can be in cash or in extra credit points. Cash payoffs can be made from funds provided by a lab fee if this is admissible at the college. If extra credit points are used then it is recommended that any curve grading be done before extra credit is factored in (so as to avoid creating tournament incentives).
Several experiments should be run with different treatments to provide data to illustrate several features of tax theory.
All experiments should be run with equal numbers of buyers and sellers so as to preserve essential features of the experimental designs. This is important; do not disregard this advice.
The total number of students in each experiment, (half buyers and half sellers) can vary between 6 and 30.

Parameterized Treatment

Treatment 1. Introduction of a tax on buyers into a market with symmetric demand and supply.

Treatment 2. Introduction of a tax on sellers into a market with symmetric demand and supply.

Treatment 3. Introduction of a tax on buyers into a market with perfectly inelastic supply.

Treatment 4. Introduction of a tax on sellers into a market with perfectly elastic supply.

Treatment 5. Introduction of a tax on sellers into a market with perfectly inelastic demand.

Expected Outcomes

Each of the treatments has 10 trading periods with no tax followed by 10 periods with the indicated per unit tax.
The market should converge to near competitive equilibrium outcomes by period 10.
Introduction of the tax should disequilibrate the market
The market should converge to near the tax equilibrium outcomes by period 20.

Pedagogic Uses of Data

Treatment 1 data Comparison of period 10 data with period 20 data demonstrates:

The deadweight loss that is evidenced by buyer and seller payoffs that decrease from period 10 to period 20 by an amount greater than the tax revenue.
The shared incidence of the tax evidenced by payoffs that decrease from period 10 to period 20 for both buyers and sellers.

Treatment 2 data Same interpretation as Treatment 1.

Treatments 1 and 2 data combined Comparison of period 20 data from Treatment 1 with period 20 data from Treatment 2 demonstrates:

The incidence of a tax on buyers and sellers is essentially the same regardless of whether liability to pay the tax is on buyers or sellers.

Treatment 3 data Comparison of period 10 data with period 20 data demonstrates:

With perfectly inelastic supply the incidence of a tax on buyers falls almost entirely on sellers.
Thus, although buyers are liable to pay the tax, the incidence (or burden) of the tax is mainly shifted to the sellers.

Treatment 4 data Comparison of period 10 data with period 20 data demonstrates:

With perfectly elastic supply the incidence of a tax on sellers falls almost entirely on buyers.
Thus, although sellers are liable to pay the tax, the incidence (or burden) of the tax is mainly shifted to the buyers.

Treatment 5 data Comparison of period 10 data with period 20 data demonstrates:

With perfectly inelastic demand the incidence of a tax on sellers falls almost entirely on buyers.
Thus, although sellers are liable to pay the tax, the incidence (or burden) of the tax is mainly shifted to the buyers.

Lecture

The lecture material is developed on the assumption that students have previously been introduced to step function market supply and demand functions that correspond to goods traded in discrete units (as in everyday life).

Textbook derivations of key results on the theory of taxation are developed using linear supply and demand graphs. Parallel derivations of some of the results are developed using demand and supply step functions.

Lecture