An Experimental Evaluation of Strategies to Increase Property Tax Compliance: Free-Riding in the City of Brotherly Love
An Experimental Evaluation of Strategies to
Increase Property Tax Compliance:
Free-Riding in the City of Brotherly Love
Michael Chirico, Robert Inman, Charles Loeﬄer,
John MacDonald, and Holger Sieg
University of Pennsylvania
September 21, 2015
The lack of tax compliance has become a policy issue of
central importance to all levels of government in developed
and developing economies.
In 2009, the developed economies of the OECD reported an
average tax non-compliance rate of 14.2.
The rates range from a low of 2-3 percent in Austria,
Denmark, Germany, Korea, and Norway to 25 percent or more
in Belgium, Iceland, Portugal, the Slovak Republic, to a high
of 73 percent in Greece.
Why is Noncompliance a Signiﬁcant Concern?
1. Governments are denied the revenues needed to provide basic
public services essential for ensuring the safety, health, and
minimal well-being of all citizens.
2. If there is signiﬁcant non-compliance and basic services are to
be provided, then tax rates will need to rise on those who pay
3. Non-compliance undermines the principle that everyone has to
pay their “fair share” of taxes.
What do I owe?
Taxpayers have the ability to inﬂuence what is owed on any
tax that requires self-reporting of income or assets, such as
self-reported consulting or business income.
The need for self-reporting is reduced as the formal economy
and the use of audited business records expands.
Taxes which can take advantage of those records maximize
tax compliance and are preferred for just this reason.
Examples are the value added tax and the property tax.
What happens if I don’t pay?
The taxing jurisdiction can also control compliance by
inﬂuencing the decision to evade, once the tax liability has
The most common strategy is the economic stick - ﬁnes and
Liens and sheriﬀ sales are more extreme and costly measures
to enforce property tax compliance.
The empirical evidence regarding the eﬀectiveness of
economic sanctions to impact aggregate collections is mixed.
Are there Alternatives to Fines and Penalties?
Researchers have turned attention to stressing other
motivations for increasing tax compliance.
Such motives are grounded in the value taxpayers place in
helping to achieve positive outcomes such as the provision of
Individuals may also derive satisfaction from knowing, or from
having others know, that they have done their civic duty.
Researchers have started to explore the eﬀectiveness of softer,
nudge approaches to reinforce these motivations.
Our agenda here is to extend our understanding of tax
compliance to include the payment of local property taxation.
In the fall of 2014, we assisted the City of Philadelphia’s
Department of Revenue in designing an experiment to
evaluate potential nudge strategies to improve local property
The City’s historical performance in tax compliance has not
been good, collecting only 90 percent of assessed property tax
revenues compared to an average compliance rate among
large U.S. cities of nearly 95 percent. (See Table 1.)
We conducted a controlled experiment redrafting letters
reminding taxpayers that their 2014 property tax payments
Figure 1: Standard Due Letter
PO BOX 148
PHILA PA 19105-0148
I...III.I ........ II..II....IIII...I
5107 DUNLAP ST
CITY OF PHILADELPHIA
DEPARTMENT OF REVENUE
AUGUST 12, 2014
Phone: (215) 686-6442
Property: 5707 DUNLAP ST
REAL ESTATE TAX BILL
Includes payments posted through AUGUST 07, 2014
This bill represents the real estate tax liabilities for this account You must pay these liabilities immediately.
Interest, penalty, and/or additions have been calculated to the due date. Additional amounts will accrue after
that date. Only payment in full or a payment agreement will prevent enforcement action.
THIS BILL MAY NOT REPRESENT YOUR TOTAL TAX LIABILITY
Period Tax Due
Penalty Charges Total Amount Due
023459700 2014 755.76 68.02 0.00 0.00 823.78
Total 755.76 68.02 0.00 0.00 823.78
----------------- DETACH HERE ----------------------------
RETURN THIS PORTION WITH PAYMENT
PAYMENT DUE: $823.78
ON OR BEFORE SEPTEMBER 06, 2014
MAKE CHECKS PAYABLE TO:
CITY OF PHILADELPHIA
NOTICE #: 5518914149812
DEPARTMENT OF REVENUE
P.O. BOX 148
PHILA, PA 19105-0148
3 3 3 b l l 2 3 1 T T Q 0 D D 5 f l l f l T l M 1 4 0 f l l 2 0 0 0 0 0 0 0 D 0 0 D 0 0 0 D 0 0 0 D D 0 0 0 0 D 0 0 D 0 0 D l T f i l ] i 0 5 0 0
Our Notiﬁcation Strategies
We proposed three additional formats for DoR’s reminder
letter. (See the appendix of the paper.)
In addition to the listing of tax payments, interest, and
penalties, the alternative letters contained a sentence that
1. threatened the potential loss of the taxpayer’s home or
property if taxes were not paid, or
2. appealed to the positive community beneﬁts in provided public
services that the taxpayer’s dollars provide, or
3. appealed to the positive beneﬁts of fulﬁlling your civic duty to
yourself and others by paying your taxes.
For legal reasons we were not allowed to modify the reminder
letter directly. Hence, we used a second letter that was added
to the standard letter in the mailing cycles.
Our Experimental Design
Our approach to randomization was constrained by the
logistics of DoR’s enforcement faculties.
We concluded after several discussions with our collaborators
at DoR that it would be logistically impossible to assign
properties at random to diﬀerent treatments.
Instead, we chose to exploit the pseudo-random assignment of
properties to billing cycles and randomized treatments across
We conducted a variety of balance checks to make sure that
this as pseudo-random assignment of properties to billing
cycles was eﬀective. (See Table 3 for details.)
Descriptive Statistics of Our Sample
Full Sample Analysis Sample
Number Observations 134887 4927
Amount Due 4409 3465
Assessed Property Value 138867 186691
Value of Tax 1586 2405
Length of Debt 6 4
% Residential 80 80
% w/ Philadelphia Mailing Address 88 83
% Owner-Occupied 24 22
Diﬀerences between the full sample and our analysis sample are
largely due to the fact that 2/3 of tax payers are assigned to
collection agencies in September of each year.
Eﬀectiveness of Treatments
Group Treated Number Total Taxes Percent
Days Treated Owed Ever Paid
Threat 1 499 $1,839,826 14
Service 4 2,211 $8,003,148 15
Civic Duty 2 1,142 $3,794,900 18
Control 2 1,075 $3,294,516 16
Percent Dollars Per Dollars Total Generated
Paid in Full Day Treated above Control All Days
Threat 8 $71,176 $10,883 $ 10,883
Service 7 $111,932 $51,639 $206,557
Civic Duty 12 $76,109 $15,816 $ 31,632
Control 10 $60,293 $ 0 $ 0
Formal Statistical Analysis
The mean payment per letter was $174 for all three treatment
groups combined and $112 for the control group.
The diﬀerence in means between the treatment and control
groups is thus $62 per letter.
There is a fair bit of heterogeneity among treatments. The
diﬀerences in means are $90 for the service treatment, $21 for
the civic duty treatment, and $30 for the threat treatment.
The eﬀects are similar in the sole owner sample and smaller if
we exclude commercial properties.
The only signiﬁcant positive coeﬃcient is the moral appeal
treatment in the sample of sole owners. (See Table 5 for
Statistical Analysis of Discrete Outcomes
To gain more insights into the causal impact of the diﬀerent
treatment letters on taxpayer compliance we turn the
statistical analysis of discrete outcomes.
We focus on two discrete outcomes:
1. “ever paid”
2. “paid in full”
We estimate logistic regressions.
Empirical results are summarized in Tables 6-11 of the paper.
The ﬁndings are mixed, but we ﬁnd some signiﬁcant eﬀects.
Summary of Findings
We ﬁnd compelling statistical evidence that the appeal to
civic duty is an eﬀective strategy for encouraging at least
some tax payment and often payment in full.
Its most signiﬁcant impact are on residents with relatively low
levels of tax debt ($0 to $300).
Stressing the beneﬁts of payment for the provision of city
services may also have improved tax compliance.
Its most signiﬁcant impact appears to be on taxpayers with
the highest levels of outstanding tax debts (owing more than
We have shown that diﬀerent notiﬁcation strategies have the
potential to improve Philadelphia tax compliance and
collections among tardy city taxpayers.
Our ﬁndings imply that it may be beneﬁcial to target a
message to speciﬁc group in the population.
Our research provides ample scope for future research.
One potential way of improving threat letters may involve
providing more speciﬁc and relevant, local information to tax
payers such as the foreclosure rate with a given neighborhood.
We would appreciate any suggestions on how to improve our