Content uploaded by Richard J. Cebula
Author content
All content in this area was uploaded by Richard J. Cebula on Mar 02, 2023
Content may be subject to copyright.
Content uploaded by Richard J. Cebula
Author content
All content in this area was uploaded by Richard J. Cebula on Apr 13, 2022
Content may be subject to copyright.
1
A Further Inquiry into Factors that Influence Federal Personal Income Tax Evasion in the United
States
Malissa L. Davis
North Carolina Agricultural and Technical State University
Richard Cebula
George Mason University
Robert Boylan
Jacksonville University
Abstract. The purpose of this study is to identify factors that may have influenced federal personal income
tax evasion in the United States during recent years. An established tax evasion model is expanded to
include four additional years of observations and five heretofore neglected explanatory variables. The
model examines the percentage of personal taxable income that is unreported to the IRS by using official
time series data for the years 1980 through 2016. It is found that the tax rate, unemployment rate, audit
rate, penalty rate, and income level, all influenced the degree of aggregate federal personal income tax
evasion, consistent with prior literature. However, this study finds that age, gender, the state tax rate and
the percentage of federal personal income tax returns that include Schedule C and/or Schedule A are
additional variables that, while having been overlooked in previous related studies, appear to have
impacted tax evasion.
Key Words: Tax evasion; federal income tax; age; gender; schedule C; schedule A; state tax rate
J.E.L Codes: H26, H2
1
A Further Inquiry into Factors that Influence Federal Personal Income Tax Evasion in the United
States
1. Introduction
Public services provided by the Federal Government are funded primarily from revenue obtained by
taxing income. The Federal Government imposes a form of voluntary self-reporting tax system that relies
significantly on the honesty and integrity of taxpayers. When taxpayers do not exercise these
characteristics, the government must increase its net borrowing (as reflected in the unified budget deficit)
and hence the national debt in order to continue the funding of public services. The Federal Government
imposes penalties to discourage tax evasion, and yet the United States (U.S.) will likely always have a tax
evasion issue because there are incentives to engage in tax evasion. While some taxpayers may freely
accept paying taxes as a manifestation of their civic duty, many taxpayers may likely think of other ways
to spend their earnings. Thus, taxpayers try to identify ways either to minimize or eliminate their tax
liability.
The personal income tax evasion literature clearly distinguishes between tax evasion and tax
avoidance (Alm, 1999; Franzoni, 1999; Slemrod & Yitzhaki, 2002; Sandmo, 2005; Slemrod, 2007). Tax
avoidance is the legal reduction of tax liabilities by taking advantage of the tax code (Alm, 1999). Tax
avoidance occurs when a taxpayer finds a loophole in the tax law and uses this ambiguity to minimize
taxes owed. There are many ways to reduce one’s tax liabilities, but ultimately it is the taxpayer’s
responsibility to avoid and not evade taxes. Tax evasion occurs when individuals deliberately ignore their
tax obligations in violation of the Internal Revenue Code (Alm, 1999; Franzoni, 1999; Slemrod, 2007).
While tax avoidance and tax evasion both reduce the amount of federal income taxes paid, tax evasion
uses methods that are illegal.
Since tax evasion involves illegal methods to reduce tax liabilities, it can be thought of as
noncompliance with tax laws. Noncompliance can occur when an individual underreports taxable income
or overstates allowable deductions or when taxpayers make an inherent error when interpreting tax law
2
(Alm, 1999; Slemrod, 2007). Compliance with tax laws occurs when a taxpayer reports to the correct tax
base, computes the correct tax liability and files his or her tax return with the correct payment before the
appropriate deadlines (Franzoni, 1999).
Since the Allingham and Sandmo (1972) publication, various theoretical studies have investigated
personal income tax evasion (Falkinger, 1988; Klepper, Nagin, and Spurr, 1991; Das-Gupta, 1994;
Caballe and Pandes, 1997; Sandmo, 2005; Gahramanov, 2009; Cebula and Feige, 2012). The tax evasion
literature has comprehensively been reviewed by Alm (1999), Andreoni, Erard, and Feinstein (1998),
Franzoni (1999), Slemrod and Yitzhaki (2002) and Slemrod (2007). However, there is little empirical
literature in recent years that focuses on the determinants of aggregate personal income tax evasion in the
U.S.
This study seeks to contribute to the literature that examines factors that influence federal
personal income tax evasion by expanding Cebula’s model (2013). The most current adjusted gross
income data reported by the Department of the U.S. Treasury and the Bureau of Economic and Analysis
(BEA) is estimated through the year 2016 to empirically investigate the determinants of tax evasion in the
U.S. Using this data, key factors that potentially influence federal personal income tax evasion over time
are investigated for the years 1980 through 2016. Variables that have commonly been found to influence
tax evasion include average tax rate, the IRS audit rate and penalty rate, the unemployment rate, and the
real GDP per capita. However, five additional explanatory variables not previously analyzed in a formal
macro-empirical context are also examined, namely, the percentage of taxpayers that file a Schedule C
with Form 1040, the percentage of taxpayers that file a Schedule A with Form 1040, the state tax rate, the
percentage of the population age 65 and older, and a gender variable in the form of the ratio of the male
civilian labor force participation rate to the female civilian labor force participation rate.
2. Background: Measuring Noncompliance
3
The tax gap is defined as the difference between the federal income taxes that taxpayers are obligated to
pay based on tax code and the amount actually reported and paid to the Federal Treasury on a timely basis
(Mazur, Plumley, & Plumpley, 2007). It represents the difference between what taxpayers actually owe
and what they voluntarily report and ultimately pay to the government (Andreoni, Erard, & Feinstein,
1998; Slemrod, 2007). While there are some income sources that are reported to the IRS by a third party,
other income including many cash transactions are not reported to the IRS (Morse, 2009). Thus the gap
exists because some variables that often define the tax base, such as income and revenue, are often not
observable (Franzoni, 1999). It is difficult to empirically analyze the determinants of tax evasion over
time as it is challenging to find an appropriate measure of noncompliance. Noncompliance with the tax
code has historically been estimated by using “official data”, surveys or the currency model.
Historically, the IRS examined tax compliance through the IRS Taxpayer Compliance Measurement
Program (TCMP) which attempted to measure unreported income and the tax gap through intensive
audits. This program included line-by-line information reported by the taxpayer and compared that
information to amounts the examiner thought to be correct (Slemrod, 2007). The TCMP was criticized
because of the intrusiveness of the program, therefore the last TCMP audit conducted by the IRS was in
1988 (Cebula & Feige, 2012). The TCMP was also criticized because the audits were not able to identify
some unreported income items and often failed to detect non-filers and honest errors (Alm, 1999). Despite
the criticism that the TCMP failed to detect unreported income, it is still believed to be the best source of
measuring noncompliance with the tax law (Andreoni et al., 1998).
The IRS periodically provides estimates for the annual gross tax gap. The 2011-2013 annual average
gross tax gap is estimated at $441 billion, with underreporting of income representing $325 billion of the
reported amount. The IRS estimates do not include illegal activities such as gambling, prostitution and
drug trafficking (Mazur et al., 2007), however this estimate is thought to be one of the most
comprehensive estimates of noncompliance (Slemrod, 2007). Since the IRS gross tax estimates are only
4
released periodically and often group several years together, it is difficult to use this data to track time
series data.
The Bureau of Economic Analysis (BEA) has historically published a comparison of its personal
income measure and the IRS adjusted gross income (AGI) measure. Working with the BEA, Ledbetter
provided estimates of the aggregated unreported AGI for the years 1959-2005 (2004, 2007). Ledbetter
used the AGI approach, to calculate the AGI gap, defining the AGI gap as the “difference between the
BEA-derived adjusted gross income and the IRS adjusted gross income” (2007). The relative AGI can be
calculated by dividing this difference by the actual aggregate AGI obtained from the National Income and
Product Accounts (NIPA). The most recent estimate provided by the BEA is for the year 2005. However,
building upon Ledbetter’s model, the Treasury Department updated AGI gap estimates through the year
2012 (2016).
Feige’s (1994) General Currency Ratio (GCM) model is an alternative method to measure the size of
the AGI Gap. In subsequent research, Cebula & Feige (2012) present a refined and updated version of the
General Currency Ratio Model (GCM) for the period of 1960 – 2008. They incorporate estimates from
the 1988 TCMP and the 2001 NPR as benchmarks for the GCM computations. The Figures reported by
Cebula & Feige (2012) are significantly higher than most tax gap estimates, e.g., Ledbetter (2004; 2007)
and Tanzi (1982a; 1982b). Therefore, the Currency Model is not a broadly used model of noncompliance.
Surveys have also been used to estimate noncompliance; however, survey data may not be reliable as
taxpayers would not likely admit to committing fraudulent activity. Taxpayers normally report higher
compliance rates when surveyed (Andreoni et al., 1998) as evident by national study commissioned by
the IRS that reported 80 percent of respondents said that they had not underreported income or overstated
expenses on their tax returns in the last five years (Harris, 1987).
Expanding on Cebula’s model (2013), this study adopts the AGI gap approach as the measure of
aggregate personal federal income tax evasion. The AGI gap is the difference between the aggregate AGI
5
as computed from the National Income and Product Accounts (NIPA) and the AGI that is reported to the
IRS (Ledbetter, 2004, 2007; Foertsch, 2016). Ledbetter estimated the AGI gap for the years 1959 through
2005. In 2016, Foertsch on behalf of the US Department of the Treasury, updated the AGI gap estimates
through 2012. AGI gap estimates were provided by Cebula (2020) for the years 2013 through 2016.
3. The Framework: Determinants of Non-compliance
Building upon the theoretical model used in prior research (Cebula, 2013), this study empirically expands
a model of noncompliance to examine possible determinants of tax evasion. The relative probability that
the taxpayer will not report the correct taxable income to the IRS is treated as an increasing function of
the expected gross benefits the taxpayer will receive by not reporting the correct taxable income, egb, and
as a decreasing function of the expected gross costs to the taxpayer not reporting the correct taxable
income, egc. Probabilities are expressed in relative terms to reflect the non-compliance data available.
Thus, the ratio of the probability of not reporting the correct income to the IRS, pnr, to the probability of
reporting the correct income to the IRS (1-pnr) is reflected in equation (1).
(1) pnr/(1-pnr) = f(egb, egc), fegb > 0, fegc < 0
The expected gross benefits from not reporting taxable income to the IRS are directly related to the
federal personal income tax rate (Cebula & Feige, 2012; Feige, 1994; Jackson & Milliron, 1986;
Slemrod, 1985; Clotfelter, 1983). Previous studies have used either the maximum marginal tax rate
(MMT) or the average effective tax rate (AET) as an approximation for the federal income tax rate. The
MMT is thought to be a better representation of the progressivity of the federal personal income tax rate
schedule (Alm & Yunus, 2009; Gahramanov, 2009), while AET is considered a better representation of
the overall population of taxpayers (Cebula & Feige, 2012; Feige, 1994, Clotfelter, 1983). This study
utilizes the AET since it includes more of the taxpaying population. It is hypothesized, ceteris paribus,
that:
(2) egb = g(AET), gAET >0
6
The higher the unemployment rate, the higher the extent of tax evasion (Alm & Yunus, 2009;
Gahramanov, 2009). Unemployment rate may cause the underground economy to increase, as well as
create fear in employed workers. When the unemployment rates are higher, workers that are currently
employed may become nervous about possibly being laid-off in the future ultimately leading to the
underreporting of income (Alm & Yunus, 2009; Gahramanov, 2009; Cebula and Coombs, 2009). It is
hypothesized, ceteris paribus, that federal personal income tax evasion will increase as the unemployment
rate [UN] increases. Thus, equation (2) can be expanded by equation (3):
(3) egb = g(AET, UN), gAET > 0, gUN > 0
The higher the real income level (INC), the greater the extent of personal income tax evasion
(Clotfelter, 1983; Slemrod, 1985; Jackson & Milliron, 1986; Cebula, 2013). Higher income taxpayers
may also have greater access and knowledge of ways to reduce their tax obligations. (Cebula, 2013). This
study utilizes per capita real GDP [INC] as a measure of income. As a taxpayer’s income level rises, the
expected gross benefits from not reporting taxable income to the IRS also increases. Consistent with prior
literature, it is hypothesized, ceteris paribus, that federal personal income tax evasion will increase as
INC level increases. Thus, equation (3) can be expanded by equation (4):
(4) egb = g(AET, UN, INC), gAET > 0, gUN > 0, gINC > 0
Research suggests that women evade taxes less than men (Mason & Calvin, 1978; Baldry, 1987;
Brooks & Doob, 1990; Collins, Milliron and Toy; 1992; Richardson, 2006). This rationale is driven by
the belief that from a young age, females are conditioned to be more conservative than males (Tittle,
1980). There is limited research that uses official data to examine gender and the impact on tax evasion.
This study hypothesizes that as the ratio of men to women in the civilian labor force grows, the extent of
tax evasion will also grow. Thus, equation (4) can be expanded by equation (5):
(5) egb = g(AET, UN, INC, GEN), gAET > 0, gUN > 0, gINC > 0, gGEN > 0
7
The Schedule C reported with Form 1040 (SCHC) accounts for the business receipts and deductions
of sole proprietorships. It is easier for sole proprietorships to have unreported income, especially when
most of the business transactions are on a cash basis. Using microeconomic data, the literature suggests
that taxpayers that file a SCHC are more likely to evade taxes than the average taxpayer (Clotfelter, 1983;
Feinstein, 1991; Ali, Cecil, & Knoblett, 2001). To reflect the SCHC data, this study utilizes the
percentage of taxpayers that filed a SCHC as reported by the IRS. As can be best determined, this is the
first study that examines the SCHC variable by using macroeconomic time series data. Thus, it is
hypothesized that as the percentage of returns that include a SCHC with the 1040 tax return increases, the
degree of tax evasion will increase. Therefore, equation (5) can be expanded by equation (6).
(6) egb = g(AET, UN, INC, GEN, SCHC), gAET > 0, gUN > 0, gINC > 0, gGEN > 0, gSCHC > 0
The state income tax rate has virtually been overlooked in terms of the impact on tax evasion,
although one study combined the impacts of federal and state tax rates (Clotfelter, 1983). It is
hypothesized that as the state tax rate increases, the taxpayer will have a greater incentive to evade taxes.
As best can be determined, this is the first study that examines the state tax rate [STATE] and the impact
on tax evasion on a time-series macroeconomic level.
Therefore, equation (6) can be expanded by equation (7).
(7) egb = g(AET, UN, INC, GEN, SCHC, STATE), gAET > 0, gUN > 0, gINC > 0, gGEN > 0, gSCHC > 0,
gSTATE > 0
The expected gross costs of not reporting the correct income to the IRS are hypothesized to be an
increasing function of the expected risks/costs (Feinstein, 1994; Caballe and Pandes, 1997). Only a small
percentage of returns are audited by the IRS, therefore the perception that the chance of being audited is
low may convince taxpayers to evade taxes. It is believed that the higher the audit rate [AUDIT], the
lower the extent of federal personal income tax evasion (Allingham & Sandmo, 1972; Alm, Jackson, et
al., 1992; Erard & Feinstein, 1994; Cebula & Feige, 2012). There are a variety of penalties imposed on
8
persons found to have engaged in tax evasion, but at a minimum, the taxpayer will have to pay interest on
the unreported income. Penalty rates [PEN] have been tested extensively in theoretical and empirical tax
evasion models (Allingham & Sandmo, 1972; Alm, Jackson, et al., 1992; Cebula & Feige, 2012; Erard &
Feinstein, 1994), suggesting that as the costs associated with tax evasion increase, the extent of the
evasion decreases. Thus, it is hypothesized, ceteris paribus, that federal personal income tax evasion will
decrease as audit rates and penalties increase. Thus, we have:
(8) egc = g(AUDIT, PEN), gAUDIT > 0, gPEN > 0
There is limited research on the impact that form A has on tax evasion. Form A is the form used on
the 1040 where taxpayers can itemize deductions instead of taking the standard deduction. Many of the
items reported by taxpayers on form A can be verified by a third-party. For example, mortgage interest
rates and state and local taxes can be verified by a third party, therefore taxpayers may be likely to report
the correct amount for the deductions. This suggests that the cost associated with tax evasion will
increase, ultimately decreasing the extent of tax evasion. Thus, it is hypothesized, ceteris paribus, that as
the percentage of taxpayers that file a form A increases, the extent of federal personal income tax evasion
will decrease. Thus, we have:
(9) egc = g (AUDIT, PEN, SCHA), gAUDIT > 0, gPEN > 0, gSCHA > 0
It is hypothesized that individuals near retirement or who are retired are least likely to engage in a
risky behavior such as tax evasion. The individual likely has the perception that he/she has worked hard
during his/her lifetime and is not willing to sacrifice the income received from retirement. The costs of
not reporting taxable income outweigh the benefits of not reporting taxable income. Literature appears to
suggest that income tax evasion is negatively impacted by age (Baldry, 1987; Dubin & Wilde, 1988;
Feinstein, 1991; Friedland et al., 1978; Witte & Woodbury, 1985). As best can be determined, this is the
first study that examines the explanatory variable, AGE, on a time-series macroeconomic level. This study
9
hypothesizes that there is an indirect relationship between the age of the taxpayer and the extent of tax
evasion. Thus equation (9) can be expanded by equation (10).
(10) egc = g(AUDIT, PEN, SCHA, AGE), gAUDIT > 0, gPEN > 0, gSCHA > 0, gAGE > 0
Substituting from Equations (7) and (10) into Equation (1) yields:
(11) pnr/(1-pnr) = egb =g(AET, UN, INC, GEN, SCHC, STATE, AUDIT, PEN, SCHA, AGE)
gAET >0, gUN >0, gINC >0, gGEN >0, gSCHC >0, gSTATE >0, gAUDIT <0, gPEN <0, gSCHA <0, gAGE <0
Let AGI represent the actual total value of the aggregate federal adjusted gross income in the
economy, i.e., AGI = UAGI +RAGI, where UAGI is the dollar size of the unreported aggregate
federal adjusted gross income in the economy, and RAGI is the dollar size of the reported aggregate
federal adjusted gross income in the economy. Thus:
(12) UAGI = (pnr)*AGI
and
(13) RAGI = (1-pnr)*AGI
It then follows that:
(14) UAG/RAGI = (pnr)*AGI/(1-pnr)*AGI=(pnr)/(1-pnr)
From (11) and 14, substitution for pnr/(1-pnr) in (1) yields:
(15) UAG/RAGI = g(AET, UN, INC, GEN, SCHC, STATE, AUDIT, PEN, SCHA, AGE)
gAET >0, gUN >0, gINC >0, gGEN >0, gSCHC >0, gSTATE >0, gAUDIT <0, gPEN <0, gSCHA <0, gAGE <0
10
4. Estimation Results
Based on the framework in equation (15), the equation estimate is as follows:
(16) Log(UAGI/RAGIt) = a0 + a1AETt-1 + a2UNt-1 + a3INCt-1 + a4GENt + a5SCHCt + a6STATEt-1 +
a7AUDITt-1 + a8PENt-1 + a9SCHAt-1 + a10 AGEt-1 + u
where
(UAGI/RAGI)t is the ratio of the aggregate unreported federal adjusted gross income in year t to the
aggregate reported federal adjusted gross income in year t, expressed as a percent;
AETt-1 is the average effective federal personal income tax rate expressed as a percentage in year t-1;
UNt-1 is the seasonally adjusted civilian unemployment rate expressed as a percent in year t-1;
INCt-1 is the real per capita GDP in year t-1;
GENt is the ratio of the male labor force participation rate to the female labor force participation rate;
SCHCt is the percentage of filed federal personal income tax returns that contained a Schedule C in year t;
STATEt-1 is the maximum marginal state income tax rates across all of the states by year t;
AUDITt-1 is the percentage of filed federal personal income tax returns that were audited by the IRS in
year t-1;
PENt-1 is the interest rate used by the IRS in year t-1 to impose interest penalties on the detected
underpayment of tax liability;
SCHAt-1 is the percentage of filed federal personal income tax returns that contained a Schedule A in year
t;
AGEt-1 is the percentage of the U.S. population greater than the age of 65;
and u is the stochastic error term.
The time series data for this study runs from the years 1980 to 2016. All data are annual. The
choice of the year 1980 reflects the limited available data for the percentage of taxpayers that file a
schedule C with their 1040 tax return. The AGIGAP data were obtained from the BEA (2005, 2007) and
11
the U.S. Treasury (2016). The IRS Statistics of Income Tax Stats (SOI) was used to collect data for the
average tax rate and the percent of taxpayers that file a Schedule C (SCHC) and Schedule A (SCHA)
(Form 1040). AGE data were obtained from the U.S. Census Bureau, Current Population Survey, Table
HH-3 (2017). AUDIT rate data for the years 1979 through 2005 were obtained from the Government
Accounting Office (1996, Table I.1) and the U.S. Census Bureau (1994: Table 519, 1998: Table 550,
1999: Table 556, 2001: Table 546, 2010: Table 469). The audit rate data for the years 2008-2016 were
obtained from the IRS SOI data Table 9B (2017). Data from the Federal Reserve Bank of St. Louis
(FRED) were collected for PENALTY, GEN, INC, UN, and STATE.
To allow for the fact that the dependent variable is contemporaneous with two of the explanatory
variables, SCHCt and GENt, there is a possibility of simultaneity bias. Accordingly, the model is to be
estimated by two-stage least squares (TSLS). Estimation by TSLS requires the adoption of IVs
(instrumental variables), which are as follows: LABOR (Labor force participation rate in the U.S. in year
t-2) for the variable GEN and the variable INT (the 10-year interest rate in year t-2) for the variable
SCHEDC. These instruments are highly correlated with the unlagged explanatory variable with which
associated but by virtue of being lagged two years are uncorrelated with the error term in the system. The
TSLS estimation of equation is corrected for heteroskedasticity by adopting the Newey-West (1987)
correction and is corrected for autocorrelation by adopting a first order autoregression. The results of the
TSLS equation are provided in Table 1.
The estimated coefficients on all ten explanatory variables are consistent with the hypothesized
signs. Eight of the explanatory variables are statistically significant at the 1% level, while two are
statistically significant at the 5% level. The inverted AR Root is -.35; therefore, there is no concern with
non-stationarity.
The results show that the coefficient on the average effective federal personal income tax rate
variable (AET) is positive and statistically significant at the 1% level. Thus, the higher the average
effective tax rate, the greater the percentage of taxable income that is unreported. It can be concluded that
12
taxpayers derive a greater benefit from evading taxes as the average effective tax rate goes up. This is
consistent with the findings in previous literature that suggests that as income tax rates increase, the
extent of tax evasion will increase (Clotfelter, 1983; E. L. Feige, 1994; Gahramanov, 2009; Cebula &
Feige, 2012).
The estimated coefficient for the unemployment variable (UN) is positive and statistically
significant at the 1% level. As hypothesized, the higher the unemployment rate, the higher the AGI gap.
This is consistent with prior research that argues that the higher the unemployment rate, the greater the
extent of tax evasion (Alm & Yunus, 2009; Gahramanov, 2009). This is likely driven by households
entering the underground economy to obtain some type of income until they are employed.
The estimated coefficient on the income variable (INC) is positive and significantly significant at
the 1.0 percent level. This suggests that the higher the real income levels, the higher the tax liability, thus
the greater the incentives to evade taxes. The findings are consistent with the literature on this variable
that suggests that the higher the income, the greater the level of tax evasion because higher income
taxpayers have greater and more sophisticated knowledge regarding ways to avoid taxes (Cebula & Feige,
2012).
The GEN explanatory variable is negative and statistically significant at the 1% level. It can be
concluded that when the ratio of men to women in the labor force is higher, the AGI gap will be higher.
Previous research suggests that women evade taxes less than men (Baldry, 1987; Brooks & Doob, 1989;
Collins et al., 1992; Mason & Calvin, 1978; Vogel, 1974). As best can be determined, no research has
looked at the ratio of men to women in the labor force.
The estimated coefficient for SCHC is positive and statistically significant at the 1% level. On a
microeconomic cross-sectional analysis level, the literature has shown that taxpayers that file a Schedule
C are more likely to evade taxes than the average taxpayer (Clotfelter, 1983; Feinstein, 1991). This is
13
accomplished by both understating revenues and inflating costs. As can be best determined, this is the
first study that examines the Schedule C variable by using macroeconomic time series data.
As hypothesized, the estimated coefficients on the AUDIT and PEN variables are both negative.
The penalty variable is statistically significant beyond the 1% level. This suggests that taxpayers are
inclined to evade taxes less when penalty interest rates are higher. The audit variable is statistically
significant at the 5% level. Based on this analysis, higher audit rates are a deterrent to federal personal
income tax evasion. Consistent with prior literature, higher audit rates and penalty rates can lead to a
decline in tax evasion (Allingham & Sandmo, 1972; Alm, Jackson, et al., 1992; Erard & Feinstein, 1994,
Cebula & Feige, 2012).
The explanatory variable AGE is negative and statistically significant at the 1% level. As
hypothesized, the older the population, the less the extent of tax evasion. The costs and risks associated
with tax evasion are too high for older taxpayers. There is literature that suggests that age is negatively
related to the extent of income tax evasion (Baldry, 1987; Dubin & Wilde, 1988; Feinstein, 1991;
Friedland et al., 1978; Witte & Woodbury, 1985). However, these studies were conducted using
microeconomic cross-sectional data or survey data. As best can be determined, this is the first study that
examines the explanatory variable, age65, on a time-series macroeconomic level.
5. Conclusion
This study extends the tax evasion model by Cebula (2013). In this study, the tax evasion model is
updated to include several additional years of observations for the dependent variable UAGI/RAGI. The
study is unique because (a) it uses official data that is more current than any previous study; and (b) it
investigates three non-economic explanatory variables that have been largely neglected by previous
studies with time-series macroeconomic data.
The study uses the two-stage least squares estimation to identify factors that influence federal
personal income tax evasion in the U.S. for the study period of 1980 through 2016. The estimates
14
provided by the model show the magnitude of federal personal income tax evasion (as measured by the
AGI Gap) is positively impacted by the average tax rate, the unemployment rate, income level, gender,
percentage of schedule C filers, and the state tax rate. The extent of tax evasion is negatively impacted by
the audit rate, IRS interest penalties, the percentage of taxpayers that file a Schedule A and AGE.
Many of the results of the study are consistent with prior literature. When audit and penalty rates are
higher, tax evasion is likely to decline because taxpayers are fearful of getting caught and would be worse
off having to pay the additional penalty. As with prior literature, the higher the tax rate and the higher the
income level, the higher the extent of tax evasion. This is logical since taxpayers would have more to gain
from tax evasion if their higher income places them in a higher marginal tax bracket. This logic can also
be applied to the state tax rate.
Lastly, the study shows that demographic/noneconomic variables play a vital role in the extent of the
adjusted gross income gap. As more of the population is over the age of 65, the level of tax evasion will
likely decline since it would be too risky to evade taxes. Finally, if more taxpayers file a form Schedule
C, which might occur if self-employment becomes more prevalent, tax evasion would be expected to
increase. This makes sense because Schedule C afford multiple opportunities to underreport income and
exaggerate expenses. As more taxpayers file a form A on the 1040, the extent of tax evasion should also
decline due to the fear of getting caught by third party confirmation. Moreover, since women are most
risk averse than men, if the female labor force participation rate of women grows more (less) rapidly than
that of men, the degree of tax evasion would tend to decline (increase), ceteris paribus.
While this study gives some insight to the determinants of tax evasion, it is impossible to predict the
behavior of every single taxpayer. Taxpayers evade taxes for many different reasons, and it is difficult to
model the factors that determine tax evasion since taxpayers are very complex. There are many cultural
and political factors that should also be included in the model. One of the purposes of this study is to
extend the tax evasion model to include noneconomic explanatory variables and test these variables by
analyzing actual data from official sources. This study reiterates that demographic variables influence tax
15
evasion. As more time-series official data becomes available, other demographic variables can be added
to the model. There is also an opportunity to add political and cultural related explanatory variables to the
model.
There are tens of millions of taxpayers in the U.S., and each taxpayer has his/her own cost/benefit
calculus process. There is no tax evasion model that can identify every explanatory variable that has an
influence on tax evasion. However, this study shows that identifiable real-world variables influence the
extent of personal income tax evasion. As more and more noneconomic variables are added to the model,
the complexity of the taxpayer and his/her decision to evade taxes could possibly be better explained.
This study attempts to lay some additional groundwork for adding more noneconomic variables to the tax
evasion model and testing these variables on a macroeconomic level.
16
References
Ali, M. M., Cecil, H. W., & Knoblett, J. A. (2001). The effects of tax rates and enforcement
policies on taxpayer compliance: A study of self-employed taxpayers. Atlantic economic
journal, 29(2), 186-202.
Allingham, M. G., & Sandmo, A. (1972). Income tax evasion: A theoretical analysis. Journal of
public economics, 1(3-4), 323-338.
Alm, J. (1999). Tax evasion. Encyclopedia of taxation and tax policy, 2, 401-404.
Alm, J., Jackson, B. R., & McKee, M. (1992). Estimating the determinants of taxpayer
compliance with experimental data. National Tax Journal, 107-114.
Alm, J., & Yunus, M. (2009). Spatiality and persistence in US individual income tax compliance.
National Tax Journal, 101-124.
Andreoni, J., Erard, B., & Feinstein, J. (1998). Tax compliance. Journal of economic literature,
36(2), 818-860.
Baldry, J. C. (1987). Income tax evasion and the tax schedule: Some experimental results. Public
Finance= Finances publiques, 42(3), 357-383.
Board of Governors of the Federal Reserve System (US), Effective Federal Funds Rate
[FEDFUNDS], retrieved from FRED, Federal Reserve Bank of St. Louis;
https://fred.stlouisfed.org/series/FEDFUNDS, July 1, 2019.
Brooks, N., & Doob, A. N. (1989). Tax evasion: Searching for a theory of compliant behavior:
Canadian Institute for Advanced Research.
Cebula, R. J. (2004). Income tax evasion revisited: the impact of interest rate yields on tax-free
municipal bonds. Southern Economic Journal, 418-423.
Cebula, R. J., & Feige, E. L. (2012). America’s unreported economy: measuring the size, growth
and determinants of income tax evasion in the US. Crime, Law and Social Change, 57(3),
265-285.
Clotfelter, C. T. (1983). Tax evasion and tax rates: An analysis of individual returns. The review
of economics and statistics, 363-373.
Collins, J. H., Milliron, V. C., & Toy, D. R. (1992). Determinants of tax compliance: A
contingency approach. The Journal of the American Taxation Association, 14(2), 1.
Erard, B., & Feinstein, J. S. (1994). Honesty and evasion in the tax compliance game. The RAND
Journal of Economics, 1-19.
Feige, E. L. (1994). The underground economy and the currency enigma. Public Finance=
Finances publiques, 49(Supplement), 119-136.
Feinstein, J. S. (1991). An econometric analysis of income tax evasion and its detection. The
RAND Journal of Economics, 14-35.
Foertsch, T. (2016). Using a Reconciliation of NIPA Personal Income and IRS AGI to Analyze
Tax Expenditures. Office of Tax Analysis
Franzoni, L. A. (1999). TAX EVASION AND TAX COMPLIANCE.
Friedland, N., Maital, S., & Rutenberg, A. (1978). A simulation study of income tax evasion.
Journal of public economics, 10(1), 107-116.
Gahramanov, E. (2009). The Theoretical Analysis of Income Tax Evasion Revisited. Economic
Issues, 14(1).
Government Accounting Office. (1996). “Individual Audit Rates.” Retrieved from
http://www.gao.gov/archive/1996/gg96091.pdf.
17
Harris, L. (1987). Taxpayer opinion survey. Conducted for the US Internal Revenue Service,
Internal Revenue Service Document(7292).
Internal Revenue Service, Statistics of Income, retrieved from https://www.irs.gov/statistics/soi-
tax-stats-integrated-business-data, July 1, 2019.
Jackson, B. R., & Milliron, V. C. (1986). Tax compliance research: Findings, problems, and
prospects. Journal of accounting literature, 5(1), 125-165.
Ledbetter, M. (2004). Comparison of BEA Estimates of Personal Income and IRS Estimates of
Adjusted Gross Income. Survey of Current Business, 84(4), 8-22.
Ledbetter, M. (2007). Comparison of BEA Estimates of Personal Income and IRS Estimates of
Adjusted Pre-tax Income. Survey of Current Business, 87(11), 35-41.
Mason, R., & Calvin, L. D. (1978). A study of admitted income tax evasion. Law & Soc'y Rev.,
13, 73.
Mazur, M. J., Plumley, A. H., & Plumpley, A. H. (2007). Understanding the tax gap. National
Tax Journal, 569-576.
Richardson, G. (2006). Determinants of tax evasion: A cross-country investigation. Journal of
International Accounting, Auditing and Taxation, 15(2), 150-169.
Sandmo, A. (2005). The theory of tax evasion: A retrospective view. National Tax Journal, 643-
663.
Slemrod, J. (2007). Cheating ourselves: The economics of tax evasion. The journal of economic
perspectives, 21(1), 25-48.
Slemrod, J. B. (1992). Did the Tax Reform Act of 1986 Simplify Tax Matters. The journal of
economic perspectives, 6(1), 45-47.
Slemrod, J., & Yitzhaki, S. (2002). Tax avoidance, evasion, and administration. Handbook of
public economics, 3, 1423-1470.
Tanzi, V. (1982a). A second (and more skeptical) look at the underground economy in the U.S..
The underground economy in the U.S. and abroad, Lexington (Mass.), Lexington, 38-56.
Tanzi, V. (1982b). The underground economy in the U.S. and abroad: Free Press.
Tittle, C. R. (1980). Sanctions and social deviance: The question of deterrence.
U.S. Bureau of Labor Statistics, Civilian Unemployment Rate [UNRATE], retrieved from FRED,
Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/UNRATE, July 1,
2019.
U.S. Bureau of Labor Statistics, All Employees: Total Nonfarm Payrolls [PAYEMS], retrieved
from FRED, Federal Reserve Bank of St. Louis;
https://fred.stlouisfed.org/series/PAYEMS, July 1, 2019.
U.S. Bureau of Labor Statistics, Civilian Labor Force Participation Rate: Men [LNS11300001],
retrieved from FRED, Federal Reserve Bank of St. Louis;
https://fred.stlouisfed.org/series/LNS11300001, July 1, 2019.
U.S. Bureau of Labor Statistics, Civilian Labor Force Participation Rate: Women
[LNS11300002], retrieved from FRED, Federal Reserve Bank of St. Louis;
https://fred.stlouisfed.org/series/LNS11300002, July 1, 2019.
U.S. Census Bureau (2017). Table HH3. Households by Age of Householder, 1960-2012. Retrieved
from https://www.census.gov/data/tables/time-series/demo/families/households.html, July
1, 2019.
U.S. Census Bureau, (1994), Statistical Abstract of the U.S., 1994. Washington, D.C.,: U.S
Government Printing Office.
18
U.S. Census Bureau, (1998), Statistical Abstract of the U.S., 1998. Washington, D.C.,: U.S
Government Printing Office.
U.S. Census Bureau, (1999), Statistical Abstract of the U.S., 1999. Washington, D.C.,: U.S
Government Printing Office.
U.S. Census Bureau, (2001), Statistical Abstract of the U.S., 2001. Washington, D.C.,: U.S
Government Printing Office.
U.S. Census Bureau, (2010), Statistical Abstract of the U.S., 2010. Washington, D.C.,: U.S
Government Printing Office.
U.S. Department of the Treasury, Internal Revenue Service. Tax Gap Estimates for Tax Years
2008-2010. Press release. Washington, D.C., April 2016.
Vogel, J. (1974). Taxation and public opinion in Sweden: An interpretation of recent survey data.
National Tax Journal, 499-513.
Witte, A. D., & Woodbury, D. F. (1985). The effect of tax laws and tax administration on tax
compliance: The case of the US individual income tax. National Tax Journal, 1-13.
19
Table 1 - AGI Estimates (Do we need this, or can we site the paper where you used the estimates)
20
Table 2 – Regression
