Revising the conventional tax‐effort principle
| Author | John Leventides,Basil Dalamagas,Panagiotis Palaios,Stefanos Tantos |
| Date | 01 July 2020 |
| Published date | 01 July 2020 |
| DOI | http://doi.org/10.1111/sjpe.12239 |
272
|
wileyonlinelibrary.com/journal/sjpe Scott J Polit Econ. 2020;67:272–299.
© 2019 Scottish Econ omic Society
Accepted: 5 Dece mber 2019
DOI: 10 .1111/sjpe.1 2239
ORIGINAL ARTICLE
Revising the conventional tax-effort principle
BasilDalamagas |JohnLeventides |PanagiotisPalaios |
Stefanos Tantos
Department of Economics, National and
Kapodist rian University of A thens Faculty
of Economics, A thens, Greece
Correspondence
Basil Dalamag as, Departme nt of Economics,
National and K apodistrian Un iversity of
Athens Facult y of Economics, Athe ns,
Greece.
Email: dalamaga@econ.uoa.gr
Abstract
The standard me thodology on tax-effort (i.e ., the ratio of ac-
tual tax revenue to it s optimal level) is to run a regr ession
of actual tax revenu e on countries’ specific (macroeconomi c,
demographic, geographical, political, social, and institutional)
variables. The re sulting predicted (f itted) values are t hen
taken to represent th e optimal (desired or maximum) level of
tax revenue. The cr ucial issue of tracing out how the optimal
tax revenue should b e allocated to the fisc al objectives (eq-
uity, efficiency) d oes not seem to be of any interes t to the
researchers on t ax-effort. The present paper a rgues that the
standard meth odology is not without faults an d needs revis-
ing. We demonstrate tha t an optimal tax system can b e safely
derived from maxim izing a utility function with r espect to (in)
direct tax rate s. The manipulation of th e first-order condi-
tions, using a novel mathe matical module, leads to an infinite
number of optimal dire ct–indirect tax rate s. The selection of
the optimal mix of thes e tax rates is dependent on the coun-
try-specific households’ preferences over equity/efficiency,
as they are formulated by voter s’ volition in election periods.
A simulation procedur e helps understanding how t he optimal
tax revenue is chosen a nd how it can be optimall y allocated
to fiscal object ives, in the context of a panel data set inclu d-
ing a large number of develo ped and developing count ries.
Throughout our tex t, the optimal ta x revenue is defined as
the sum of the produc ts of the optimal (in)direct t ax rates and
their correspond ing tax bases. In th e simple Arrow–Debreu
economy, the above sum is shown to be equ al to the differ-
ence between incom e and consumption.
|
273
DALAMAGA S et AL.
1 | INTRODUCTION
Typically, optimal tax r evenue is studied using cro ss-country data on a lar ge sample of developed and/or devel op-
ing economies . The determination of opt imal tax revenue is an ess ential prerequisite for e stimating the tax-ef fort,
which is define d as the ratio of the act ual tax revenue to its o ptimal level. The opt imal tax revenue in tur n coincides
with the predic ted tax collection that is d etermined from a regression re lating the actual tax revenu e to various
explanatory (control) variables, such as per capita income, shares of imports, exports, agriculture, mining etc. in
GDP, corruption, bure aucracy, quality of instit utions, political, dem ographic, geographic al and social factors, a nd
so on. If the tax-ef fort index is less than or gr eater than one, the countr y is identified as undert axed (low tax-ef-
fort) or overt axed (high tax-effor t), respectively.
In estimating t he tax-effort regress ion equation, a number of app roaches are employed to fin d an observable
proxy for the un observable potentia l (optimal) revenue variable .
In existing lite rature, tax potenti al (ratio of actual ta x revenue to GDP) and optima l tax revenue (i.e., the r atio of
actual tax r evenue to the predicted or f itted one) are used intercha ngeably to define tax-ef fort (see, for example,
Jean-Francois Br un & Diakité, 2016; Mawej je & Sebudde, 2019). In the p resent study, we focus o n the ratio of
actual (or obser ved) to predicted reven ue to estimate the tax-effo rt.
In sharp contras t to the above conventio nal approache s that predict t he optimal tax reve nue by regressing
the actual t ax revenue on a vector of many ad ho c explanatory variabl es—most of which are latent varia bles—our
proposed tech nique builds upon the ex plicitly revealed prefe rences of private and pub lic agents during the voti ng
political pro cess. In all demo cratic countrie s, the politic al party (or coa lition of partie s) tend to seize power on
the basis of a multi-goa l program (including fiscal a nd, especiall y, tax policy) that has b een previously a pproved
by the elector ate. The primary objec tive of each program revolves ar ound the optimal allocati on of state-owned
resources (mainl y tax revenue) to the objec tives of equity (fai r distribution of incom e) and efficiency (i ncentives to
work and invest, growth), via properly scheduled adjustments in direct/indirect tax rates.
Thus, the prefe rences to the most prefe rred allocation of res ources are unveiled by the ma jority of voters-tax-
payers. Then, t he political party t hat have announced a fiscal po licy program (meeting the c riteria of the elector-
ate) seize power and have res ponsibility for im plementing the rel evant policy. To this end, the gover nment chooses
the suitable co mbination of (in)di rect tax rate s from the set of the op timization proce sses, on the basi s of two
accepted social n orms, which will be descr ibed in Section 2.
To contrast our propos ed methodolog y with the convent ional approach es used in previou s literature, it
would be constr uctive to acquaint the reader w ith the main characterist ics of the prevailing theoreti cal struc-
tures on tax-ef fort, and this a nalysis is under taken in Appen dix S1. If we would set off ou r new approach
against exist ing tax-effor t research, it wou ld become clear t hat our explicit “ voters’ authoriz ation technique”
for tracing out t axpayers’ preference s is a substitute to, rather t han being complimenta ry with the conventiona l
econometric techniques, which are based directly or indirectly on regression analyses with a large number of
determining fa ctors. This happens be cause our methodolog y has the merit of incorporat ing all the explanatory
variables of opt imal tax revenue for eac h household into a single “ vote”, and nobody can seriousl y cast doubt on
the validity of t he institutional stru ctures within democra tic societies. Therefore , there is no room to unify the
KEYWORDS
optimal alloc ation of tax liabilities, o ptimal tax revenue, ta x
capacity, tax-effort, utility maximization
JEL CLASSIFICATION
H30; H61; E62
To continue reading
Request your trial