Transcript Toward a Theory of Tax Systems

Tax Administration and
Tax Systems
Joel Slemrod
University of Michigan
Tax Administration Research Centre
University of Exeter
March 10, 2014
The Modern Theory of Taxation
• The modern theory of taxation, beginning circa 1970,
represented a major breakthrough in how economics
addressed the evaluation of taxes.
• It provided a rigorous basis for analysis, which
enables intellectual progress.
• Like all models, the modern—now standard—model
of taxation is stylized, making simplifying
assumptions and emphasizing certain aspects of
taxation at the expense of others.
• But it hampers addressing many key issues.
6 Limitations of the Standard Toolkit
1. Little attention paid to the administrative and
compliance costs of taxation.
2. A focus on tax rates and, to a lesser extent, tax
bases, to the relative exclusion of all other tax
system instruments, such as enforcement
tools (audits, evasion penalties, public
disclosure) and administrative choices.
6 Limitations of the Standard Toolkit
3. A focus on what might be called real behavioral
responses to taxation, to the relative exclusion of
avoidance and evasion responses (e.g., Mirrlees
considers only labor supply choices; Allingham &
Sandmo introduce evasion).
4. A recognition of the central role of asymmetric
information between the government and private
citizens, but extreme assumptions regarding what is
measurable without cost and what is not measurable
at any cost. (e.g., in Mirrlees the tax authority can
measure income, but cannot measure ability or effort.)
6 Limitations of the Standard Toolkit
5. No meaningful role for firms. With the
standard assumption of constant returns to
scale, firm size is indeterminate and
irrelevant. The implications of firm
heterogeneity cannot be pursued.
6. No concern with the details of tax
remittance. In spite of folk theorems about
its irrelevance, it does not hold in the real
world.
A Tax-Systems Approach
I will argue two points today:
1. A tax-systems perspective can provide insight
into many important issues relevant for
taxation that the standard model misses.
2. Tax administration, broadly defined, is central
to a tax-systems perspective.
Tax Systems, co-authored with Christian
Gillitzer, explores these issues in depth.
Some PR Tips
• To economists, “administration” is not sexy.
• I suggest using “implementation” when
possible.
• I also recommend playing up the “economics
of information” aspects of the field of inquiry.
• If it’s not too late, re-title the new journal the
Journal of Tax Systems (just kidding).
What Is a Tax System?
A set of rules, regulations, and procedures that:
1. Defines what events or states of the world trigger tax
liability (tax bases and rates).
2. Specifies who or what entity must remit that tax and
when (remittance rules).
3. Details procedures for ensuring compliance, including
information-reporting requirements and the
consequences (including penalties) of not remitting the
legal liability (enforcement rules).
Standard analysis presumes that tax liability can be
ascertained and collected costlessly, in which case
statement 2 is irrelevant and statement 3—and any tax
administration—is unnecessary.
Outline of Today’s Talk
• Building Blocks of Tax Systems
– Multiple sources of cost—not just distortion of
choices
– Multiple behavioral responses—not just “real”
– Multiple tax instruments—not just rates and bases
• Optimal Tax Systems
– New analysis of standard tax instruments
– Analysis of new tax instruments
– Implications of the information revolution
Administrative Costs
• Collecting taxes, especially non-capriciously,
requires a costly bureaucracy.
• One reason is the need for a payment and
collection system.
• For any given objective, there are more and less
effective ways for a tax administration to operate.
• For example, should a tax administration be
organized by tax levy (e.g., corporate tax versus
value-added tax), or by taxpayer segment (e.g.,
corporations versus high-income individuals)?
Administrative Costs
There are important procedural differences across
tax systems, including:
• the degree of self-assessment
• the extent to which income withholding at source
is used
• the amount of arms-length information reporting
to the tax authority.
Administrative Costs
• Market (non-cash) transactions facilitate
administration.
• AC is a function of the physical size, tangibility,
visibility and mobility of the tax base (e.g., it is
harder to tax diamonds than windows),
whether there is a registration of the tax base
(e.g., owners of cars), and the number of
taxpayer units.
• AC is an increasing function of the complexity
and lack of clarity of the tax law.
• AC tends to be discontinuous and to have
decreasing average costs with respect to the tax
rate.
Compliance Costs
• Compliance costs, borne in the first instance by
taxpayers, tend to dwarf AC. For the U.S. income
tax, CC =~10% of revenues, versus 0.6% for AC.
• Burden may be shifted, as for tax liability.
• Both AC and CC ultimately burden citizens,
although only AC shows up as government
expenditures.
• CCs have been generally measured using surveys.
Biases going both ways: a survey is like a tax
form, and some may want to express frustration.
Compliance Costs
• How to value time spent?
• Voluntary versus involuntary CCs—does it matter
in measuring social cost?
• Which costs are truly marginal is tricky, especially
for businesses. What value, if any, does
addressing one’s financial affairs provide?
• Trade-offs in administrative versus compliance
costs: existing expertise, transparency, differing
shadow cost (i.e., AC must be funded).
Multiple Behavioral Responses:
Evasion
• Positive theoretical analysis has taken off since
the deterrence model of Allingham and Sandmo
(1972).
• Non-deterrence theories (e.g., duty, altruism,
process) have proliferated, but with mixed
empirical support to date, while there is
overwhelming support for deterrence that I’ll
discuss later.
Evasion
• Empirical analysis is challenging, due to the
nature of evasion. (“You can’t measure the
right-hand-side variables, and you can’t
measure the left-hand-side variable!”)
• A similar, but not identical, problem is faced
by tax administrations.
• Let’s work together!
Evasion
There are several promising developments:
1. Traces-of-income approach
– Pissarides-Weber, Feldman-Slemrod, etc.
2. Analysis of administrative data (in the past
mostly in Scandinavia, but now in Canada, the
UK, and the US)
Evasion
3. We may never be able to do randomized field
experiments on tax rates or bases, leaving us out of the
“credibility revolution” in empirical economics.
But we can do them for other tax-system instruments.
Examples include:
• Slemrod et al. on income tax in Minnesota
(USA)
• Kleven et al. on income tax in Denmark
• Pomeranz on VAT in Chile
• Fellner et al. on TV fees in Austria
Multiple Behavioral Responses:
Avoidance
• What is it? “Taxpayer efforts to reduce their
tax liability that do not alter their
consumption basket other than due to income
effects.”
• Examples: paying a tax professional to search
for deductions, tax arbitrage, slightly re-timing
a transaction, slightly re-engineering a vehicle.
Avoidance
• Much income tax avoidance arises because
what triggers tax liability is income per se, but
rather surrogate tax bases that may be
justified on administrative or compliance cost
grounds.
• Examples: taxes on realized capital gains;
income shifting; arbitrary, incoherent
treatment of financial transactions such as
debt versus equity.
Interactions among Responses
• Example: Income from real U.S. investment in Puerto
Rico was subject to a low rate, and was not subject
to any U.S. residual tax, thus making income shifting
very attractive.
• Result: Much U.S. investment in high-margin
activities in P.R. (electronics, pharma, high-fashion)
• The income-shifting was facilitated by real
investment, thus providing an avoidance-facilitation
implicit subsidy to real investment there.
• Lesson: Tax-systems issues matter for real behavior.
Interactions among Responses
• The notion of avoidance facilitation suggests that
the response to a pre-tax price and the tax rate
may not be identical; rethink Rosen (1976), who
interpreted the differential response of labor
supply to the pre-tax wage rate and the tax rate
as tax illusion (“lack of salience” in modern
terminology).
• But, the effective price of a real decision depends
on the avoidance and evasion technology: the tax
can be “finessed,” but the pre-tax price cannot
be.
Multiple Tax Instruments
• The vast majority of tax analysis addresses tax
rates and bases.
• Thus it ignores critical aspects of tax systems such
as audits, evasion penalties, information
reporting, withholding regimes, taxpayer
education, self-assessment regimes, public
disclosure.
• The good news is that modern analytical tools,
econometric and theoretical, can be applied to
these policies.
Withholding and
Information Reporting
In the United States, the income tax noncompliance
rate is:
• 56% when “little or no” information reporting
• 11% when “some” information reporting
• 8% when “substantial” information reporting
• 1% when both withholding and substantial
information reporting
This suggests a central role of firms, which I’ll discuss
later, and provides support for the deterrence theory.
Market Transactions
• Basing tax liability on market transactions provides a
natural check on accuracy, relies on betterdocumented data, and establishes arm’s-length
prices.
• But cash transactions are much more problematic.
Compare Gordon-Li argument for subsidizing
relationships with financial institutions and the
recent IRS 1099-K initiative.
• Issues also arise with family firms, as examined by
Kopczuk and Slemrod (2011).
Public Disclosure
• Policy in the United States during the Civil War, and
again in the 1920s and 1930s.
• Current policy in Norway, Sweden, and Finland; it was
policy in Japan from 1949-2004.
• A serious proposal in Australia recently.
• Supporters say it reduces tax evasion/avoidance and
improves policy transparency, opponents decry the loss
of privacy and the negative attention brought to the
affluent.
• What do we know about its consequences?
• New evidence from Japan in Hasegawa et al. (2012) and
from Norway in Slemrod et al. (2013).
Optimal Tax Systems
• Changes answers to classic OT questions, such as
optimal income tax progressivity and optimal
commodity taxation.
• Raises new questions, such as:
– How many resources to devote to enforcement?
• What is an optimal auditing structure?
– How to collect information (e.g., through thirdparty information reports)?
– What is the proper role of firms in remittance?
– If higher top tax rates would induce taxable
income flight offshore, should we abandon the
attempt, or crack down on the flight itself?
Bringing Firms into Tax Theory:
Remittance
• Public finance textbooks assert a remittance
irrelevance proposition: it doesn’t matter which
side of a taxed transaction must remit.
• This is wrong, though, notwithstanding the cupat-the-counter metaphor.
• Examples: predominance of firm remittance, VAT
vs. RST, employer withholding, small business
exemptions.
• Evidence: see Kopczuk et al. (2013), where the
amount of diesel tax pass-through is shown to
depend on the point of collection.
Why Exempt Small Businesses?
• Exemption violates the standard (DiamondMirrlees) injunction against violating production
inefficiency.
• But it economizes on collection costs.
• This trade-off has not been closely addressed by
the optimal tax literature, although see
Dharmapala et al. (2011) on optimal taxation
with per-firm administrative costs.
• One reason is that heterogeneous firms are
absent from the theory of taxation, although they
are ubiquitous in other fields of economics.
Taxes and the Missing Middle
• Dharmapala, Slemrod, and Wilson (2011) identify
conditions under which it is optimal to exempt small
firms from taxation, and consider per-firm
administrative costs.
• These inefficiencies occur in part because some firms
obtain the tax exemption by reducing their outputs
to inefficiently low levels—below the tax net cutoff—
creating a “missing middle” of intermediate-sized
firms.
• This production inefficiency is balanced against the
cost savings from collecting revenue from, on
average, larger firms.
Line Drawing
Much real-world scuffling about taxation involves
drawing and interpreting lines, yet analysis of this
topic is almost completely absent from economic
analysis. Why?
• One reason is that optimal commodity tax
theory allows an unlimited number of tax rates,
but this is infeasible.
• Differentiating tax classes is often done by
referring to characteristics of goods, types of
income, or transactions.
Tax-Driven Product Innovation
• Drawing lines (e.g., in characteristics space) creates
notches in choice sets—a small change in some
aspects of the taxed good creates a large change in
tax liability.
• This causes tax-driven product innovation, goods just
slightly on the low-tax side of the line.
• Examples: car-like motorcycles in Indonesia and carlike panel trucks in Chile; 17.9” arrows in the United
States, outrageous costumes worn by ABBA.
• A very important example: equity-like finance that
qualifies for (debt-like) interest deductibility.
Future Directions: Information
Revolution
• Computerization of the tax collection process is the
most visible example of the effects.
• Can base tax liability on a wider range of information.
– Finnish income-based speeding fines
– Use of tags (even genome?)
– Smart (tax) cards
– But it works both ways. Note the existence of
automatic sales suppression devices, known as
“zappers,” that skim cash sales by excluding
random transactions from the apparent electronic
record.
Insights
• The British economist Frank Hahn once wrote that
“optimal tax formulas are either guides to action or
nothing at all.”
• As of now, these formulas mostly refer to a stylized
world far from the reality of withholding,
information reports, audits, tax havens and
evasion, and where line drawing and notches lurk
everywhere.
• Tax-systems analysis applies rigorous economic
analysis of taxation to issues that are prominent in
the formulation and administration (say
implementation!) of real-world tax policy.
Policy Contributions
• Policymakers should recognize the interrelationship
among tax rates, tax bases, enforcement, and
administration.
• There are many alternative ways to raise revenue,
and many types of costs, some that show up in
government budgets but most of which do not.
• The costs of using one tax instrument often depends
on the setting of the others.
• Recognizing that tax policy is really tax-system policy
can ward off substantial policy errors, such as
foregoing tax increases because the existing base is
too narrow or too poorly enforced.
Thank you!