Eco 3630 Public Economics

Spring 2025
References
• Chapter 19 Gruber
• Rothstein, Jesse. 2010. "Is the EITC as Good as an NIT? Condi?onal
Cash Transfers and Tax Incidence." American Economic Journal:
Economic Policy, 2(1):177-208.
• CheJy, Raj, Adam Looney, and Kory KroQ. "Salience and taxa?on:
Theory and evidence." American economic review 99, no. 4 (2009):
1145-77.
Summary of the previous lecture
• Equilibrium is independent of who nominally pays the tax (i.e.,
statutory incidence ≠ economic incidence)
• Less elas2c factors bear more of the tax
• More efficient to tax rela@vely inelas2c goods
• Efficiency loss increases more than propor2onally as tax rate rises
• Efficiency costs are higher when there are pre-exis2ng distor@ons in
the same direc@on as the effect of the tax rate
General Equilibrium Tax Incidence
• Examples so far have focused on partial equilibrium incidence which
considers impact of a tax on one market in isolation
• General equilibrium models consider the effects on related markets of
a tax imposed on one market
• E.g. imposition of a tax on cars may reduce demand for steel;
additional effects on prices in equilibrium beyond car market.
Example : Soda tax
• Consider the market for Soda beverages in Berkeley
• Berkeley imposes a Soda tax since 2015: $.01 per ounce (=$.12/can)
• Goal was to reduce soda consumpHon for beIer health (people
overdrink).
• Here narrower quesHon: Who bears the incidence?
• If soda demand in Berkeley is inelasHc, consumers bear burden
• Demand for Soda in Berkeley is likely to be elasHc: if price of Soda in
Berkeley goes up, you consume less Soda [intenHon of the tax] or you
buy Soda elsewhere
• Consider extreme case of perfectly elasHc demand
Producer bears the burden
If Soda demand perfectly elas2c then:
• Berkeley Soda sellers (supermarkets, restaurants) bear the full burden
of the tax.
• But Soda sellers are not self-contained entities
• Companies are just a technology for combining capital and labor to produce
an output.
• Capital: land, physical inputs like building, kitchen equipment, etc. Labor:
cashier staff, cooks, waitstaff, etc.
• Ultimately, these two factors (capital or labor) must bear the loss in
profits due to the tax [if consumer demand is perfectly elastic]
• Incidence is “shi+ed backward” to capital and labor.
• Assume that labor supply is perfectly elas<c because Berkeley
restaurant/supermarket workers can always go and work in Oakland if they
get paid less in Berkeley
• Capital, in contrast, is perfectly inelas<c in short-run: you cannot pick up
the shop and move it in the short run.
• In short run, capital bears tax because it is completely inelas<c . Soda
business owners lose (not consumers or workers)
• In the longer-run, the supply of capital is also likely to be highly elas<c:
Investors can close or sell the shop, take their money, and invest it
elsewhere.
• If both labor and capital are highly elas4c in the long run, who bears
the tax?
• The one addi4onal inelas4c factor is land.
• The supply is clearly fixed.
• When both labor and capital can avoid the tax, the only way Soda
sellers will remain in Berkeley is if they pay a lower rent on their land.
• Soda tax ends up hur4ng Berkeley landowners in general equilibrium
[if Soda demand, labor and capital are fully elas4c]
• This is of course an idealized example, in prac4ce, demand, labor, and
capital are not fully elas4c so that incidence is shared
Another example – EITC and it’s incidence
• Rothstein, Jesse. 2010. "Is the EITC as Good as an NIT? Conditional
Cash Transfers and Tax Incidence." American Economic Journal:
Economic Policy, 2(1):177-208.
• The EITC provides a subsidy of up to 45% for earned income up to a
cap, then levels off, then it phases out at a rate of up to 21.06%.
• With three or more qualifying children, the EITC can be as high as
$6,660 per year.
• Might its benefits to low-income workers be “shifted” via reduced
pre-tax wage rates?
• Jesse Rothstein argues that the EITC induces an increase in labor
supply among low-income workers, which in turn drives wages down.
• To the extent that wages fall, the benefit to the intended
beneficiaries—low-income workers—is reduced, and some of the
intended transfer redounds to employers.
• Moreover, it hurts low-skill workers who do not receive the EITC,
because their wages fall and they get no credit.
• His analysis suggests that, of every dollar they receive from the EITC,
low-skill single workers lose $0.30 through lower wages.
• Employers of low-skill labor capture a whopping $0.73: $0.30 from
single mothers plus $0.43 from ineligible workers whose wages fall
when the EITC is expanded.
• The net transfer to all low-skill workers, considering both those who
get the EITC and those who don’t, is less than 28 cents per dollar
spent.
Tax Salience: A New Theory

• Traditional model assumes that all individuals are fully aware of taxes
that they pay
• Is this true in practice? May be not be because many taxes are not
fully salient.
• Do you know your exact marginal income tax rate? Do you think
about it when choosing a job?
• Chetty, Looney, Kroft AER ’09: test this assumption in the context of
commodity taxes and develop a theory of taxation with inattentive
consumers
• Sales tax is paid at the cash register and not displayed on price tags in
stores. They use both RCT and quasi-experimental methods:
1) Randomized field experiment with supermarket stores.
• In one treatment store: they display new price tags showing the level of sales
tax and total price on a subset of products
• Compare shopping behavior for treated products vs. control products in
treated store, before and aCer new tags are implemented (DD strategy)
• Repeat the analysis in control stores as a placebo DD strategy
• Concern – Hawthrone Effect:
• The decrease in the demand is not because of the salience of the tax but
because of new price tags
• No way to distentangle how new price tags worked. Thus, supplement results
with natural experiment.

2. Use changes in beer excise and sales taxes across states

• Excise tax is salient because built into posted price while sales tax is not
salient because it is not included in posted price.
• We expect larger changes in demand due to excise tax changes than sales tax
changes.
• Concerns with the policy experiment : Changes in the excise tax might
be correlated with demand for beer – business cycles for instance.
• Try to address this concerns using various robustness checks.
Other implication
• The standard model says that the statutory tax incidence doesn’t
ma4er for economic incidence.
• However, tax on producers are included in the posted price and thus,
more salient. Example is excise tax. Producers will bear a larger
burden of these taxes.
• Taxes on consumers are not included in the posted price (at least not
in the US) and thus, they are not salient. Consumers will bear a larger
burden of these taxes.
Notes