THE SPREAD OF FLAT TAX IN EASTERN EUROPE: A COMPARATIVE STUDY
Anthony John Evans and Paul Dragos Aligica
Global Politician, NY
9/14/2007
ABSTRACT: The article is an exploratory comparative analysis of the
spread the "flat tax" in post communist Eastern Europe. The paper
overviews the flat tax as a public policy prescription, discusses the
various arguments that underpin its potential reception, introduces
and applies a synthetic comparative method to study its spread in nine
Eastern Europe countries, and draws several conclusions regarding the
conditions associated to the implementation of this economic policy
idea while suggesting further research directions.
1. Introduction
One of the most interesting economic developments in post communist
Central and Eastern Europe (CEE) is the spread of the flat tax. The
flat tax in CEE is swathed in an irony: unlike many other reforms it is
not just a response to the past, or an attempt to turnaround previous
errors. In fact, it is former communist countries that are leading
the way in liberal, free-market economic reforms. The remarkable
thing about the spread of the flat tax is that in less than fifteen
years it has gone from being deemed impossible to almost automatic,
as more and more CEE countries follow suit. The lesson is that the
implementation of radical economic policy is possible. But despite
increased attention from Western European countries and the US, the
diffusion of this economic policy idea has been somewhat contained to
former socialist economies. Therefore it’s timely for a systematic
study of how the policy has spread thus far throughout CEE (as of
January 2007), in order to understand how it may spread further.
2. The Flat Tax
The flat tax itself (or "proportional tax") is neither new nor radical
– it has been used throughout history and remains a common form of
taxation (most sales taxes are single-rate). The radicalism stems
from the specific policy of a flat tax on income, which replaces
sliding tax scales with a single fixed rate (alongside a tax-free
personal allowance)1. Although Hong Kong has had a flat tax for more
than half a century (Reynolds 1999) and both Jersey and Guernsey2 have
had a flat tax on personal income for many years, none of these cases
are independent or autonomous nations. For this reason the flat tax
imposed on Iraq by US forces3 doesn’t count as a genuine "adoption"
and therefore not part of this analysis. The criteria for this study
is the adoption of a flat-rate income tax by a self-governing nation,
and therefore flat-rate state income taxes in Illinois, Indiana,
Massachusetts, Michigan and Pennsylvania are also off the radar. But
these examples suggest that a flat tax is not totally alien to
the current economic policy practice, without detracting from the
trans-formative nature in those instances that are under investigation.
Before analyzing the processes through which the flat tax has spread,
it’s important to establish a clear chronology of who have adopted
a flat tax, and when4.
The first movers were Estonia, where a flat rate of 26% on personal
income and corporate profits was introduced in 19945. Their Baltic
neighbors soon followed suit – also in 1994 Lithuania flattened their
tax system (with a 33% rate on personal income)6, and in 1995 Latvia
launched a flat tax at 25%. In January 2001 Russia established a
flat rate of 13% on personal (but not corporate) income, followed by
Serbia’s introduction of a 14% rate on salaries in 2003.7 2004 saw
Ukraine adopt a flat tax of 13%, and 2004 also saw Slovakia launch
a comprehensive flat-tax system with Income Tax, Corporate Tax and
Value Added Tax all levied at 19%. Georgia’s parliament voted to
introduce a flat tax in December 2004 and a 12% rate was implemented
in January 2005. Similarly Romania unveiled a flat tax of 16% in
January 2005, following the election victory of the National Liberal
Party/Democratic Party alliance in late 2004. As of December 2006
nine separate countries have adopted a flat tax.
A chief advantage of a flat tax is the administrative simplicity that
results from a uniform rate. This factor is especially evident if tax
evasion is high, since an increase in compliance will reduce the dead
weight loss of collection. The same logic applies to tax avoidance
and although it’s hard to project hard numbers, the incentives clearly
favor economic activity and increased participation rates. The effect
of flatter taxes on tax revenue is controversial and depends on the
exact position on the infamous Laffer curve, however the rule of thumb
"what you lose by lowering rates, you gain by broadening the base"8
seems typical. Public revenues rose in Estonia by 0.2% of GDP9, by over
20% in each of the first two years following Russia’s adoption10,
rose in the Slovak Republic11, and rose in Romania12. Although
these boosts to government coffers may stem from alternative but
simultaneous economic reforms (or other factors entirely such as
surging oil prices), it’s clear that many pessimistic predictions
have not born fruit (Aligica and Terpe 2005). Legitimate debates will
discuss whether capital follows lower wages rather than lower taxes13;
whether higher Value Added Tax offsets a low flat tax14; and what if
the current "sprawling" complexity is simply an emergent and efficient
response to the complexity of income?15 However these results ignore
the deeper properties of the flat tax – the "theological" element
(Evans 2006a). Opponents point to a decreasing marginal utility of
wealth as justification for progressive taxation whereas advocates
focus on the efficiency (allowing resources to flow to their highest
return), equity (an even distribution of the tax burden), simplicity,
and incentives for political responsibility. These normative dimensions
allude to the certainty, convenience and fairness criteria set forth
by Smith (1991[1776]), which form the cornerstones of the classical
liberal tradition.
Indeed the twin champions of economic liberalism in the c20th –
Friedrich Hayek and Milton Friedman – both advocated a flat tax (Hayek
1956:265-284, Hayek 1960:315, Friedman 1962:175). These statements
formed the basis of the flat tax epistemic community that created and
supports the policy idea, embodied by Robert Hall and Alvin Rabushka’s
classic proposal (Hall and Rabushka 1995).
The flat tax has received academic attention but mainly in the form of
impact assessment or projections based on wider tax reform. Ventura
(1999) found that a Hall-Rabushka flat tax has a positive effect
on capital accumulation, aggregate labour in efficiency units’
increase, and the distribution of wealth becomes more concentrated;
and Stokey and Rebelo (1995) claim that tax reform will have little
effect on the US economy. However Heath (2006) provides a long list
of academic literature that predicts various beneficial effects. Due
to the freshness of the flat tax as a viable policy, most information
about its dissemination and adoption is in popular press and takes the
form of opinion editorials or other country-specific advocacy16. Deeper
analysis will take the form of an overview (Grecu 2005, Forbes 2006),
a specific proposal (Forman 1996, Armey 1996, Scott, Teather 2005), or
both (Heath 2006). However the literature that specifically looks at
how the flat tax has been spread, incorporating the rise of the flat
tax into a theoretical framework is still in an incipient stage. To
do this requires a systematic application of a comparative method.
3. The Comparative Method
The standard methodological approaches have serious limits in
dealing with topics that combine issues related to the spread of
ideas and issues of comparative political economy. To analyse the
spread of economic policy ideas requires multiple cases, and at the
same time the retention of the holistic properties that put us in
the position to capture precisely the multi-causal paths toward the
adoption of the idea. Qualitative methods get bogged down in details
when comparing multiple cases involving intricate configurations
of factors, while quantitative methods oversimplify — sometimes
up to the point of irrelevance — complex and conjunctural causal
relationships. Consequently a synthetic "comparative method" is
needed. An adequate study of the spread of economic policy ideas
requires a combination of the two approaches, so that we can retain
the holistic value of each individual case, but also make comparisons
across multiple cases – to appreciate that the process of adoption
in Estonia was different from that in Serbia, but that they must
be combined somehow, in order to understand the implications for
Hungary. One methodological technique that provides such a combination
is "the comparative method", which is "a technique that uses Boolean
algebra to simplify complex data structures in a logical and holistic
manner" (Ragin 1987:14). Since this approach provides a frame into
which multiple case studies are analysed, it has the potential to
better uncover process rather than essence, and adjudicate between
alternative theories. Before the comparative method is performed,
however, we must define the degrees of analysis (the number of
cases and the number of conditions); define the binary conditions;
and state the theories being examined.
1. Degrees of analysis: Number of cases
The comparative method takes a number of cases, and ascertains
the presence (or absence) of various conditions for each case. The
intention is to uncover which conditions, and which combinations of
conditions lead to a pre-defined outcome – in our case that outcome
is whether the country has adopted a flat tax on personal income. As
of December 2006 there are nine countries that have adopted a flat
tax, and although there’s no objectively correct way to decide which
cases are of interest, it makes sense to keep the analysis close to
the environs of the adopters. In other words, since the adoption of
the flat tax thus far is a European phenomenon, this study will focus
on Europe.
To be more precise the flat tax is synonymous with post-Soviet
economic reforms, and therefore the initial cases of interest are
those European countries that were formerly communist. The traditional
distinction between Western Europe and Eastern Europe is becoming
increasingly irrelevant, but is more useful than a categorisation
along strict geographic (or even cultural) lines because shared
political institutions are of importance for studying the flat
tax. As compromise we can use a concept of "Central Europe", despite
the notoriously subjective way of defining it. For our purpose we’ve
focused on those countries of CEE from behind the Iron Curtain. This
excludes Germany and Austria on political grounds, and excludes Armenia
and Turkey on geographical and cultural grounds. Consequently this
study will focus on twenty case studies: Albania; Belarus Bosnia &
Herzegovina; Bulgaria; Croatia; Czech. Republic; Estonia; Georgia;
Hungary; Latvia; Lithuania; Macedonia; Moldova; Poland; Romania;
Russia; Serbia & Montenegro; Slovakia; Slovenia; and the Ukraine.
2. Degrees of Analysis: Number of Conditions
Having decided upon the number of cases, the second degree of analysis
that needs to be established is the number of conditions. There
are many factors that have legitimate claim to being relevant,
but since this is an exploratory study it makes sense to begin with
a relatively few conditions, that seem especially pertinent given
preliminary analysis. A strength of the comparative method approach
is that conditions can be expanded or neglected as appropriate, in
an iterative process. The first condition (A) focuses on the internal
fiscal situation, and whether the country has a monetary incentive to
alter it’s taxation revenue – either to bring more people into the tax
system (and reduce tax evasion), or to correct shortfalls in revenues
(due to budgetary pressures).
The second condition (B) is the external fiscal situation, and looks at
the threat of capital flight that a country faces. Levels of foreign
direct investment (FDI) and the ratio of imports and exports to GDP
help ascertain the degree to which an economy is interrelated with
its neighbours. These two conditions show how fiscal pressure for it’s
own sake, or fiscal pressure as a consequence of the action of others,
might make a flat tax a policy priority. The third condition (C) looks
at how much a country belongs to an international community. This
is similar to the second condition since it measures the external
relationships of a country, but focuses on social rather than economic
grounds. The fourth condition (D) looks at civic culture; to see what
role public debate might have during the policy adoption process. It
is important to know whether having a literate, educated and public
debate alters the propensity for the flat tax to spread. The fifth
condition (E) asks whether a policy champion promoted the flat tax,
and if so whether that person was a political opinion leader. In
other words to what extent is policy implementation dependent on it
having an individual espouser? And the sixth condition (F) is a crude
measure of contagion, by looking at whether there’d been a precedent.
This list is by no means exhaustive, and a number of additional
conditions come to mind, such as Did the flat tax possess media
support; Was it part of an electoral policy; What was it’s timing;
Was the proposed flat tax revenue neutral; and Had there been
rigorous impact assessment performed. However all of these issues
are somewhat represented in the six chosen conditions, and at this
stage of analysis a broad understanding is more important than a
crude list of preconditions since we are in the midst of flat tax
expansion. If we focus too much on the specific details of conditions
in the adopted countries, then there’s no scope for it to spread
further and we have created a closed system. Rather, we must look
at fundamental conditions, which can then be translated into the
climates of countries that are yet to adopt a flat tax. This is the
only way to straddle the middle ground and say that the flat tax can
spread farther, and there are common preconditions that can be used
to predict where. Broad conditions are required to prevent analysis
from being merely descriptive.
3. Binary conditions
The comparative method is based on Boolean algebra, "the algebra of
logic and sets" (Ragin 1987:85), a key feature of which is the use
of binary data. If a condition is present, we denote it with a "1",
if it is absent, it receives a "0". Therefore we must tighten the
definition of each condition, and define our assessment criteria. We
have already defined the outcome (X) as: "whether the country has a
flat tax on personal income", and so for countries that have adopted
a flat tax X=1, and for those who haven’t X=0.
The conditions are somewhat more complicated for two
reasons: timescale, and the trade-off between accuracy and
objectivity. Acknowledgement of such problems is the first step to
solving them, and we offset the issue of which time period we’re
concerned with the following rule: for an adoptee it’s "prior to
adopting", for a non-adoptee it’s "currently". Since we’re performing
process-driven analysis rather than taking a snapshot of an arbitrarily
defined moment, the timescale depends on whether adoption has taken
place. If it has, we’re concerned about the conditions prior to the
adoption. If it hasn’t, we’re interested in the present situation. The
second issue is whether we wish to be vaguely right or precisely
wrong, and we choose the former due to the intent to stimulate further
research, which enables us to refine the judgement. Therefore rather
than use a narrow question (with an objective answer), we shall use
the most accurate indicator and utilise a rigorous case study method to
answer it as crisply as possible. It might be less subjective to decide
condition A with "Was the government budget balanced?", but this fails
to fully reflect the fiscal situation in the country. Therefore we
will use the looser question of "Did/does the country have a problem
raising tax revenue" and use levels of tax evasion and the budget
deficit as evidence to answer it. If the country did/does have an
internal pressure to reform tax revenue A=1, if it didn’t/doesn’t then
A=0. Again, there are a number of objective statistics relevant for
condition B, but the binary condition will be the broader issue of
"Did/does the country fear capital flight?", when B=1 is the answer
if yes, and B=0 if not. C=1 if the country was/is an integrated part
of the international community, and C=0 if not. If the population
was/is engaged with political debate D=1, and D=0 if not.
Condition E depends on whether an opinion leader led/leads the flat tax
campaign, with E=1 if they did/do, and E=0 if not. Finally, condition
F=1 if a neighbouring country had/has previously adopted a flat tax,
and F=0 if they hadn’t/haven’t.
4. Theories being examined
Prior to collecting and evaluating the evidence, it is worth outlining
the theoretical frameworks under scrutiny, so that the results can
be seen in context.
The reason for this analysis is not simply to document what
happened and when, but to tease out potential causal relationships
that enlighten us as to why they happened then, and how they might
happen again.
A discussion of alternate theoretical frames will also reinforce the
choice of conditions, since they each posses attributes of broader
themes. For example, conditions A and B (the role of budgetary pressure
and the threat of capital flight) are part of an interest-explanation
that explains policy change by the pecuniary incentives faced by
actors17. A looks at whether it’s in a countries own domestic interests
to adopt a flat tax, regardless of the activity of neighbours. B asks
whether the policy in other countries affect the domestic interests,
demonstrating that policy reform can be strategic. Together they answer
both facets of interest-driven policy: whether is it domestically
rational, and whether it is strategically rational. Ideas-explanations
focus more on the role of institutions and individuals18, and in
this respect three conditions are relevant. Condition C looks at
international influence, to gage whether ideas are transmitted through
shared institutions.
Condition D looks at domestic institutions, and whether the media
and public debate/public opinion lead policymakers. And condition
E captures the individual element, and whether key individuals
and outstanding leaders ultimately dictate policy (Harberger
1993). Condition F is a crude measure of contagion, and confirms
whether it’s fruitful to study the adoption of the flat tax in
the first place. After all, if the spread is an automatic process
with a clearly predictable pattern, it might suggest that economic
explanations are irrelevant. If, however, condition F isn’t enough,
there must be something else at play.
The conditions reflect competing theories. A theory that emphasizes
the interests of the actors predicts the importance of conditions A
and B; a theory that emphasizes the ideas and beliefs focuses on C, D
and E; a "non-theory" can be captured by E; whilst a multi-causal path
dependent theory emphasises the role of E for some countries, and A,
B and F for others. In other words, the very fact that conjunctural
causation and multi-causal paths are a real issue, legitimises the
efforts to explore unorthodox methods and approaches.
4. Generating the Truth Table
1. A note on the data
In order to identify the presence or absence of the six conditions,
case studies of all twenty countries of interest must be performed,
focusing on the criteria outlined in the previous section. Aside
from formal flat-tax proposals; scholarly articles on the effects of
flat-taxes; official data about relevant variables; and other archival
records (such as survey data), these case studies make extensive use of
mainstream media reports and interviews. These sources are particularly
relevant because between them they manage the difficulties created by
the subjective nature of the analysis. For example, it is possible to
ascertain the official statistics on the size of imports and exports
relative to GDP, and it’s even possible to utilise projections and
formal estimations of the size of the informal economy (and hence the
level of tax evasion). However these objective facts are irrelevant
for condition A if the domestic policymakers are either unaware
of them, or do not see them as a problem. In short, we need to go
beyond establishing the retrospective evidence, and uncover whether
it’s an issue for the public and the policymakers. Mainstream media
is a crucial way to judge whether the conditions were relevant. But
the press is an imperfect measure of the local environment, and this
is explicitly recognised by condition D – in some countries civil
society isn’t developed so that policy debate mirrors public debate.
Therefore interviews were required to acquire information
first-hand. There were two types of information of interest. Firstly,
information to compensate for imperfect secondary sources (and fill in
the missing gaps needed to judge whether the conditions are present or
absent.) Secondly, interviews of key players were required to go beyond
the public debate, and shed light on the decision making of the policy
makers who actually implemented a flat tax19. The range of interviews
was large, from international experts such as Alvin Rabushka (Hall &
Rabushka 1995), Allister Heath (Heath 2006), David Storobin, Daniel
Mitchell, Howard Scott, and Corin Taylor to key players including
Mart Laar (former Prime Minister of Estonia), Martin Bruncko (former
economic advisor to the Slovakian finance minister); Dusko Stojkov
(advisor to the Serbian finance ministry), and other national experts
and decision makers who wished to remain anonymous.
2. Applying the conditions to the cases
Condition A: Tax Evasion and Budget Pressure.
Since all twenty cases are transition countries, we might expect
condition A – internal pressure to generate revenues from taxation
– to be prevalent throughout. However these transitions occurred
at different times, meaning that we can legitimately expect
deviations. Although to Western eyes the structural similarities are
blinding, this belies the differences that are of special interest in
trying to uncover why economic ideas spread to some countries and not
others. Rather than seek absolute values that reinforce the original
decision to group the cases under scrutiny, we are more concerned
with relative indicators that separate the countries (taking their
broader similarities as given). For example all twenty countries
possess large informal sectors, but our concern is which countries
have an especially large shadow economy. Whilst by definition this
issue is impossible to calculate, using figures provided by Schneider
(2003), we can estimate that the average shadow economy labour force
(for selected former Soviet Union and CEE countries) is 29.7% of the
working age population20. The only countries that have levels more
than one standard deviation higher than this average are Belarus,
Georgia, Russia and the Ukraine. We can also add Serbia to this list
since budgetary issues were important in the build up to the flat
tax – although reforms of VAT contributed more to the budget surplus,
there are estimates of 30-40% of national output avoiding consumption
tax.21 It is also important to add Czech Republic, Hungary and Poland
since their budget deficits are forecast to remain into 200722.
Condition B: Capital Flight.
With regard to tax regimes, a discrepancy between two neighbouring
countries provides an incentive for individuals or firms to choose
the more liberal rates, thereby causing capital flight away from
higher tax nations. According to Kevin Wadell "pressure comes from
competing economies"23. The two simplest proxies to assess the degree
of competitiveness between countries are the ratio of exports to gross
domestic product (GDP), and amounts of foreign direct investment
(FDI). The former reflects how open the economy is, and the latter
measures it’s intake of foreign capital. The level of exports of goods
and services (as a percentage of GDP) differs dramatically between
the countries concerned (ranging from 23.7% in Serbia to 79.9% in
Estonia), but the following have exports above 50% of GDP: Belarus;
Bulgaria; Czech Republic; Estonia; Hungary; Lithuania; Moldova;
Slovakia; Slovenia; Ukraine24.
Whilst exports can be positive-sum between competing nations
(for example countries benefit from reductions in trade barriers
with neighbours), FDI is a better measure of competitiveness since
funds are scarce and therefore one countries ability to entice new
investment is more likely to reduce that which is available to its
neighbours. CEE nations who received more than $0.5billion in 2000
and 2001 are Bulgaria; Croatia; Czech Republic; Hungary; Poland;
Romania; Russia; Slovakia; Ukraine25.
Combining these two proxies (to see which countries feature in both)
suggests that condition B is present in Bulgaria; Czech Republic;
Hungary; Slovakia; and Ukraine; but this fails to tell the whole
story. As discussed previously these objective statistics are
irrelevant unless they are present in the minds of the decision makers,
and therefore further sources are required to validate the initial
picture, and alter the list above. The case of the Czech Republic is
confirmed with the words of the shadow finance minister -Vlastimil
Tlusty – who believes that "if a neighbouring country [adopts a flat
tax] it’s necessary to follow suit"26. And similarly in Hungary, where
former President (and leader of the opposition party FIDESZ-MPP)
Viktor Orban says "Budapest will have "no choice" but to jump on
the "flat tax bandwagon" to retain the country’s share of foreign
investments that will otherwise flow to flat tax nations."27 However,
Latvia should be on the list; Heath claims that "At 25%, the rate
was chosen to undercut Estonia’s; the fact that Latvia felt forced
to follow Estonia’s lead confirms the power of tax competition and
the need for countries to attract capital."28 Georgia’s rate of 12%
was no coincidence if it was deliberately chosen to undercut Russia’s;
and Serbia also feared capital flight29.
Interestingly, although Slovakian politicians knew that a flat tax
would provide a competitive edge, it was not a key consideration and
therefore not a precondition of flat tax adoption.30
Condition C: Membership of International Community.
The countries that joined the World Trade Organisation in 1995/96
were Bulgaria; Czech Republic; Hungary; Poland; Romania; Slovakia
and Slovenia. Those who joined from 1999-2003 are Albania; Croatia;
Estonia; Georgia; Latvia; Lithuania; Macedonia; and Moldova.
Those who aren’t members are Belarus; Bosnia; Russia; Serbia and
Ukraine. Czech Republic; Estonia; Hungary; Latvia; Lithuania; Poland;
Slovakia and Slovenia are members of the EU, and Bulgaria and Romania
will join them in 2007.
Condition D: Civil Society.
There are various means to judge whether the population of a country is
engaged in political debate, none of which are perfect. It is possible
to use indicators that determine the capacity for civil society – such
as literacy rates or tertiary education, but these are crudely elitist.
Constitutional factors can be used such as the existence of a free
press, and whether that generates independent media. Institutions that
might indicate civil society – such as the Catholic Church – help as
well. Opinion polls help establish whether political issues are known
to the public, but not whether they’re understood, therefore voter
turnout (and other measures of voter apathy) are required to offset
a reliance on instrumental factors. And as before, expert judgement
provides the narrative that binds the story together – for example
Garton-Ash contrasts two of the cases under scrutiny: "Slovakia had a
vibrant civil society – or what Slovaks call "the third sector". There
was the Catholic Church. There were independent radio stations,
magazines, and the private television channel Markiza. And there were
numerous nongovernmental organizations [NGOs]… When I described
this civic campaign to opposition friends in Serbia a week later,
they threw up their hands in envious despair" (Garton-Ash 1999:358).
It’s important to distinguish between the presence of external networks
that disseminate ideas (i.e.
epistemic communities), external NGOs that influence domestic policy
(such as the Soros Fund), and genuine domestic organisations that
stimulate and facilitate debate. Whilst all three contribute to civil
society more generally, we are only interested in the last type. These
are apparent in larger countries (including Bulgaria) – especially
those with an established dissident culture (such as Czech Republic,
Hungary and Poland), but not in those most tightly controlled during
communism (e.g. Albania, Russia, Romania), or those governed by more
authoritarian means (Belarus). Since civil society is reliant on an
established middle-class and stable governance, the Balkan nations
afflicted by war (Bosnia, Croatia, Macedonia, and Serbia) lack the
requisite domestic institutions. Of those remaining the relatively
recent revolutions (Georgia and Ukraine) are too fresh to have created
a genuine civil society and too small (as is Moldova).
Condition E: Policy Champion.
Establishing the presence of a policy champion runs the danger of
appealing to hindsight, since it would be easy to identify a key
proponent of a flat tax in countries where it’s been adopted. Therefore
the criteria must be strict, and reveal someone who believes in
the flat tax’s "theological" properties (Evans 2006a), as well as
being in a position of political influence – either as President,
Prime Minister, Minister of Finance, or a chief advisor to either
position. Despite the degree of influence people such as Alvin Rabushka
have regarding the flat tax, he wouldn’t qualify as a policy champion
on the grounds that he doesn’t occupy a domestic position.
Such international advocates are the fuel that spreads the ideas, but
not the ones who implement them. The advantage of defining the policy
champion relatively precisely is that the burden of proof shifts from
the motivations of the many actors involved in the adoption process,
to assessing the merits of those claiming to steer events. Rather than
trawl through the evidence searching for a policy champion, we assume
that if that person exists, they’d make themselves known. Therefore
each country is assigned a "0", unless they can prove otherwise. Policy
champions are Laar (Estonia); Illarionov (Russia); Niklos (Slovakia);
Stolojan (Romania). An interesting case is Hungary where Vickor Orban –
currently in opposition – is a known advocate. In this country a new
election – or new coalition government – would be the constitutional
moment required to thrust Mr. Orban into becoming a policy champion.
Condition F: Precedent.
Since this is a crude measure, it is easy to determine: for flat tax
countries, had a neighbouring country (i.e. shares a border) already
adopted a flat tax; for non-flat tax countries, has a neighbouring
country already adopted a flat tax? Estonia, Russia and Serbia are
the only countries that fail to satisfy the first condition, and
only Slovenia fail to satisfy the second. This condition highlights
the few instances where geographical proximity isn’t a precondition
of flat tax adoption, and is of interest for two reasons. Firstly,
it indicates cases where ideas might dominate interests, since
adoption cannot be explained by a domino effect. Secondly, it helps us
identify "isolated" countries where ideas channels (such as epistemic
communities) might be required to offset the lack of interest channels
(such as structural similarities with neighbouring states).
5. Reduction and Boolean analysis
Table 3a shows the truth table constructed from the analysis in
section four, following the technique laid out by C. Ragin (1987). To
maintain close dialogue with the cases the name of each country is
retained, and equivalent cases (such as Albania and Bosnia) have not
been combined. Although six conditions imply a far greater number of
logically possible combinations of conditions (i.e. 26 = 64 rows), for
simplicity only observed cases are reported. Note that this doesn’t
affect the final analysis since only cases that lead to a flat tax
(indicated in bold face) are used in the data reduction stage of
the analysis.
Table 1: Truth Table for Flat Tax Adoption in Central & Eastern Europe
(1994-2006) Country Conditions Outcome A B C D E F X Albania 0 0 0
0 0 1 0 Belarus 1 0 0 0 0 1 0 Bosnia&H 0 0 0 0 0 1 0 Bulgaria 0 1 0
1 0 1 0 Croatia 0 0 0 0 0 1 0 Czech Rep. 1 1 1 1 0 1 0 Estonia 0 0
0 0 1 0 1 Georgia 1 1 0 0 0 1 1 Hungary 1 1 1 1 0 1 0 Latvia 0 1 0
0 0 1 1 Lithuania 0 1 0 0 0 1 1 Macedonia 0 0 0 0 0 1 0 Moldova 0 0
0 0 0 1 0 Poland 1 0 1 1 0 1 0 Romania 0 0 0 0 1 1 1 Russia 1 0 0 0
1 0 1 Serbia&M 1 1 0 0 0 0 1 Slovakia 0 0 1 1 1 1 1 Slovenia 0 0 1
0 0 0 0 Ukraine 1 1 0 0 0 1 1 N=20(9)
Prior to reducing the truth table into a set of results, it is worth
reiterating the tentative nature of the comparative method. There are
two separate matters of judgement that are open to debate; the first
is the choice of conditions (A, B, C, D, and E); and the second is
the establishment of presence or absence (for each of the cases). The
plausibility of the results can be used to re-assess both matters of
judgement, but we should not confuse validity ("0" or "1" given A)
and soundness (the choice of A). Since this is an exploratory study
the results are intended to lead toward a refined version of the
initial conditions, and insights into which ones require elaboration
and further enquiry.
1. Results
Using uppercase letters to denote presence, and lowercase to denote
absence of each condition, the "primitive" form of the data is
as follows:
X = abcdEf + ABcdeF + aBcdeF + aBcdeF + abcdEF + AbcdEf + ABcdef +
abCDEF + ABcdeF
Notice that there are two pairs of similar results, and the nine
results can thus be condensed into seven:
X = abcdEf + ABcdeF + aBcdeF + abcdEF + AbcdEf + ABcdef + abCDEF
At this point we can utilise "Boolean minimization", "If two Boolean
expressions differ in only one causal condition yet produce the
dame outcome, then the causal condition that distinguishes the two
expressions can be considered irrelevant and can be removed to create
a simpler, combined expression" (Ragin 1987:93).
Hence
1. abcdEf combines with abcdEF to produce abcdE 2. ABcdeF combines with
ABcdef to produce Abcde 3. ABcdeF combines with aBcdeF to produce BcdeF
4. abcdEf combines with AbcdEf to produce bcdEf 5. abCDEf remains.
The first minimization shows that geographical proximity is an
irrelevant condition if a policy champion exists, and there are
no budgetary or trade pressures, nor influence in international
institutions or a developed civil society. The second shows that
precedent is also immaterial if both measures of financial pressure
exist, and there is no civil society, institutional influence, or
policy champion.
The third says that budgetary pressures make no difference if there
are external financial pressures, a precedent, and the absence of the
remaining conditions. The fourth shows another situation in which
condition A is irrelevant, this time when a policy champion is the
only present condition.
Finally, it’s important to retain the only primitive form that
cannot be reduced, namely the lack of internal and external financial
pressure, but the presence of civil society, international influence,
a policy champion, and precedent.
This process sharpens the thinking on flat tax adoption, and applying
the results back to the original cases yields several insights. A
number of cases validate the interest-explanation that predicts that
internal budget pressures, external capital flight pressures, and
precedent will lead to the adoption of the flat tax. This holds for
Ukraine and Georgia – neither of which had institutional influences,
civil society, or a real policy champion.
The idea-explanation is relevant for two reasons: firstly, the
absence of a policy champion is offset by precedent; and secondly,
both countries had recent revolutions. These genuine "constitutional
moments" (Ackerman, 1992) are an occurrence of circumstance, and are
represented by the lack of a developed civil society. To understand the
relationship between the two interest-explanations (A and B) compare
Latvia and Lithuania with Belarus. All three have low institutional
influence, low civil society, no policy champion but they all have
precedent. Also whilst Belarus has internal pressure to raise tax
revenue, Latvia and Lithuania have external pressures over capital
flight. Again this supports the interest-explanation, but also shows
that external factors are more telling than the internal factor.
Were Belarus to feel pressure from competing countries it would take
the same expression as Latvia and Lithuania, and therefore be expected
to adopt a flat tax – since just one condition is missing me might
call Belarus "ripe". Note also that there are no cases with AbcdEF,
which is also one condition away from Belarus. Although this case
lies in counterfactual space we’d expect the addition of a policy
champion to make flat tax adoption in Belarus more likely, showing
another possible avenue that would lead to a flat tax: the former
being an interest-channel, the latter an ideas-channel.
Intuitively we might expect all conditions to independently increase
the likelihood of flat tax adoption, and the only way to test this is
to find a case with all conditions present. Such a case doesn’t exist,
however, suggesting two things. Firstly, the two cases that come
closest to a "full house" are Czech Republic and Hungary: both lack
a policy champion (E). Since we know that the Hungarian opposition
party advocate a flat tax, this is another "ripe" country, should
they come to power. But secondly, the fact that no current adoptee
has all conditions present suggests that they might offset each other,
and we have the evidence to speculate how.
Slovakia is the only flat tax country that we’ve judged to have
international influence and a developed civil society. One option is
to reappraise the initial judgement, and question whether Slovakia
warrants a "1" for each condition. The alternative is to expand those
two conditions, to see what they’re really representing. Condition C
might merely be a proxy for "bigness" or "development", and demonstrate
why the flat tax is associated with small, poorer countries (and
therefore explain why it has thus far failed to spread to Western
Europe). But if this is the case, Russia should be adapted to have this
condition present. What this implies is that the condition as we’ve
defined it is inadequate (either being too imprecise to succinctly
capture the current situation, or too tight and therefore letting
Russia through the net), but the reason for the original definition
was to establish the legitimacy of the traditional ideas-explanation:
that ideas transmit by becoming embedded within institutions. This
analysis contradicts that assertion, since the flat tax is more
consistent with separate, isolated countries rather than members
of the same club, sharing institutional structures. Therefore our
understanding isn’t satisfied, but we’ve adequately dealt with one
of the competing theories: if ideas do matter we need a more refined
concept than as captured by condition C.
So countries that appear to be only lacking a policy champion
(such as Bulgaria, Czech Republic, Hungary, Poland), might be an
illusion: we have no cases with either of their condition but with
a policy champion, therefore we do not know if that would make
the difference. On the contrary, as discussed above, the presence
of some conditions may offset an ideologically committed opinion
leader. Whereas condition C might be a reflection of general size and
influence, condition D might represent the lack of something else: a
constitutional moment. Since constitutional moments are characterised
by "procedural irregularities" (Ackerman 1992) these are harder to
achieve under the watchful eye of the public. Whilst it’s true that
constitutional moments possess an "energized, proactive public taking
an active and maybe even directive part in the extended deliberations
of a constitutional moment" (Burnham 1999:2249) it’s not clear that
mass media is a necessary condition. On the contrary, we might see
the people shouting on the streets, politicians seizing the moment,
and a silent press. For this reason we might expect D and E to offset
each other: if there’s no real civil society and a constitutional
moment appears, radical reforms can be made without a policy champion
(e.g. Georgia, Ukraine, and to a lesser extent Latvia, Lithuania
and Serbia); however if the policy is going to come up against the
gatekeepers of debate it takes a strong politician, with convictions,
to pass it through (Slovakia).
Finally, one case appears to contradict the theory that either
interest-explanations (captured within conditions A and B), combined
with precedent (F) matters; or an ideas-driver/policy champion (E) has
to lead and define the reforms. This is the case of Serbia (ABcdef)
but two important caveats must be applied. Firstly the measure of
precedent is crude and only looks at geographical proximity. When
Serbia adopted a flat tax Estonia, Latvia, Lithuania and Russia had
all already done so, and although this represents the crossing from
Eastern Europe to Central Europe, the Serbian debate was influenced by
the knowledge that it had already been done. Secondly, it’s debateable
whether Serbia’s flat tax should count as being analytically equivalent
to the others. As already mentioned, Alvin Rabushka contests whether
Lithuania’s flat tax truly qualifies, and in the case of Serbia
only salaries are taxed at a flat rate – other forms of personal
income are taxed at different rates. Also when Serbia adopted the
flat tax it was part of broader reforms where the centrepiece was
the introduction of VAT. It was VAT that was seen as the weapon to
solve the fiscal imbalances, and the flat tax involved significantly
less political capital than in other adopting countries.31 Finally,
other reports fail to include Serbia as a flat tax adoptee.32
2. Factoring: how and when do interests-explanations matter?
Since the prevailing theory of concern is the interest-explanation,
it would be useful to perform a technique called "factoring" to
focus in more detail at the interest conditions. We have already
established that condition B, rather than A is the most important
factor and indeed it is an appeal to competitiveness over FDI that
is most evident in reports. Therefore we can separate the minimised
expressions into two sets: those consistent with the presence of B,
and those consistent with the absence (b). Thus:
X = b(aCDEF +acdE + cdEf) + B(Acde + cdeF)
The striking result is that in each case where there was no real
consideration over the threat of capital flight, condition E is
present. And in each case where there was financial pressure from
external sources, condition E is absent. In other words a policy
champion was not necessary in countries that had an interest to adopt
a flat tax anyway, but if that interest is nonexistent, ideas-drivers
can still get the policy implemented. Therefore both are necessary
conditions to explain the spread of the flat tax.
6. Conclusion
This paper explored the case of adoption of the flat tax in CEE while
at the same time it used the CEE case to get a set of exploratory
insights into the ways in which ideas and interests affect the
adoption of a specific economic policy. With the mono-causal
deterministic assumption discarded, one could explore multi-causal
paths and patterns through the comparative method illustrated by the
paper. Although theories based on the driving force of interests could
explain some cases, they cannot explain all. Therefore even with six
simple conditions, an epistemic element has to be integrated. We have
seen that ideas, interests and consequences are all requisites for
a complete explanation, but for some cases, the ideas-based approach
is sufficient to explain adoption – and it is these cases which act
as precedent for others to follow. Finally, it is worth commenting on
the predictive power of this analysis and the implications for other
countries in other regions. The truth table synthesising our data
appears deterministic since it is purely descriptive. However the
table is liable to change over time since it only reflects current
information and this is subject to change. Conflicting data and the
emergence of new evidence might highlight a mis-categorisation;
or refined theory and deeper knowledge of the cases might alter
the judgement. But assuming this table is accurate, what prediction
insights have we gained? One of the strongest conjectures is that if
a policy champion emerges for instance in China, the flat tax could
begin a fresh wave of adoption in the Far East.
>From there new avenues for its diffusion may open up.
Of course, we are in the realm of speculations. But at least this is
the beginning of conversation based not on pure intuitions but on an
attempt to methodically analyze the existing cases of the phenomenon
of interest.
————————————— —————————————–
CITATIO NS
1 This personal allowance creates a degree of progressiveness into
the flat tax but a debate as to whether this makes it truly flat is
largely semantic.
If the allowance is defined as a percentage of income earned below
a particular threshold, then we have a progressive system with two
rates (one of which is zero). If instead the allowance is defined as
a non-taxable exemption it’s a single flat rate system.
The distinction is largely irrelevant since it doesn’t alter the
substance of the policy or the novelty of its implementation. Our
view is that the "flat taxes" under scrutiny are genuine because all
taxable income is taxed at a flat rate.
2 These "Channel Islands" lie just off the northern French coast,
and possess a special constitutional status from Great Britain.
3 See "U.S. Administrator Imposes Flat Tax System on Iraq" by Milbank
and Pincus, Washington Post 2nd November 2003
4 The reason why different studies provide different years for the
adoption of a flat tax is down to whether the author is referring to
the passing of the legislation, or when it actually comes into effect.
The trouble with focusing on the legislative position is having
to define for each country the precise stage at which it becomes
"adopted". For clarity, and to fit into existing studies the year of
adoption refers to when the flat tax hits the people.
5"Estonia kicked off the trend in 1991 with a flat-rate of 26pc on
personal income, later cut to 20pc", in "Poland’s single tax rate ‘is
wake-up call for the Chancellor’" by Ambrose Evans-Pritchard, Daily
Telegraph 16th March 2005. Estonia "has just cut its flat income tax
by two percentage points to 24% and has promised that the tax will
fall to 20% in two years’ time" (Heath 2006:83).
6 Some commentators don’t consider that Lithuania’s 33% rate is
equivalent to its neighbours, since it is too high to be competitive,
and fails to account for other forms of income (see "A Competitive Flat
Tax May Spread to Lithuania" by Rabushka, A. The Russian Economy24th
March 2005).
7 Like Lithuania, Serbia levies different flat rates on alternative
sources of income.
8 "Eastern Europe Embraces Flat Tax: Former Communist Nations See
Revenue Rise, Cheating Decline" The Boston Globe 22nd February
2005. Alternatively, "The point of the flat tax is that it broadens
the tax base. People avoid less, evade less and declare more. Then
they earn more", "Missing the point" by Madsen Pirie The Adam Smith
Institute
9 "The Case for Flat Taxes – Simplifying Tax Systems" The Economist
16th April 2005
10 "Flat Approach Seems to Generate an Upward Curve" by Lynn, Liquid
Africa, Comtex 30th November 2004
11 "High Taxes Wither Away" by Fund, New York Sun 1st March 2005
12 "The Truth about the Flat Tax" by Sorin Ionita, SAR Policy Brief
No. 18, May 2006
13 "Flat Is Beautiful", The Economist 5th March 2005
14 "The Case for Flat Taxes – Simplifying Tax Systems" The Economist
16th April 2005
15 "Why tax cannot be express on a postcard" by John Kay, The Financial
Times 8th February 2005; Kay and King (1986); or Hettich and Winer 1999
16 For examples see "The Case for Flat Taxes" New Zealand Herald 31st
January 2005; "Why the case for a flat-tax system is irresistible"
by Madsen Pirie, Financial Times 17th February 2005; "The Beauty of
the Flat Tax" National Review 2nd March 2005; Should Ireland Join the
Flat Tax Club"? Business & Finance Magazine 7th April 2005; "The Case
for Flat Taxes – Simplifying Tax Systems" The Economist 16th April 2005
17 The predominant method of rational-choice political science, which
stems from the interest groups theories of Olson (1965) and Becker
(1983).
18 See Hall (1989); Goldstein (1993); Goldstein and Keohane (1993),
Braun and Busch (1999), Legro (2000) and Campbell (2001).
19 We deliberately chose broad conditions, and therefore cannot
fulfil the binary conditions by simple recourse to objective
facts. Interviews are used to show that subjective criteria do not
mean arbitrary criteria; it simply acknowledges that the analysis
is tied to personal judgement. Subjectivism can both be help and a
hindrance, but the crucial point is that it exists, and therefore
must be countered. And the potential pitfalls caused by a reliance on
judgement have been checked by ethnographic fieldwork in a key case
(Romania) and formal interviews.
20 See Table 2 (Schneider 2003), figures for 1998/99, not including
Albania, Bosnia, Serbia.
21 Anonymous interview, September 2006
22 See Table 4: Fiscal Balances, "Regional Overview: Central and
Eastern Europe", EBRD Annual Meeting, Institute of International
Finance 2006
23 "Flat-tax movement stirs Europe" by Tzortzis, Christian Science
Monitor 8th March 2005
24 Eurostat (figures are an average of 2000-2004)
25 UNCTAD, World Investment Report 2002
26 "Look east, say flat tax advocates" by Reynolds, Prague Post 9th
December 2005
27 "The Flat Tax Revolution in Europe" by David Storobin Global
Politician 8th May 2006
28 Heath 2006:82
29 Anonymous interview, September 2006
30 Anonymous interview, September 2006
31 Anonymous interview, September 2006
32 See Table 1. of Keen, Kim and Varsano (2006)
SOURCES
Ackerman, B., (1992) The Future of Liberal Revolution, Yale University
Press
Aligica, P.D., and Terpe, H., (2005) "The Flat Tax Experiment: The
Romanian Case", Mercatus Center Working Paper
Aligica, P.D., and Evans, A.J., (2008) "Thought Experiments,
Counterfactuals and Comparative Analysis", Review of Austrian
Economics, forthcoming
Armey, R., (1996) Flat Tax, Ballantine Books
Becker, G.S., (1983) "A Theory of Competition among Pressure Groups
for Political Influence", Quarterly Journal of Economics 98: 371-400
Braun, D., and Busch, A., Eds. (1999) Public Policy and Political Idea,
Edward Elgar
Burnham, W.D., (1999) "Constitutional Moments and Punctuated
Equilibria: A Political Scientist Confronts Bruce Ackerman’s We The
People", The Yale Law Journal 108(8): 2237-2277
Campbell, J.L., (2001) "Institutional Analysis and the Role of
Ideas in Political Economy" in Campbell, J.L., and Pedersen, O.K.,
Eds. The Rise of Neoliberalism and Institutional Analysis, Princeton:
Princeton University Press
Campbell, J.L., and Pedersen, O.K., Eds. (2001) The Rise of
Neoliberalism and Institutional Analysis, Princeton: Princeton
University Press
Evans, A.J., (2005) "Ideas and Interests: The Flat Tax", Open
Republic 1(1)
Evans, A.J., (2006a) "The Spread of Economic Theology: The Flat Tax
in Romania", Romania Economics and Business Review 1(1)
Evans, A.J., (2006b) "Constitutional Moments and Subjectivist Public
Choice: From Chandeliers to Fireworks", Mercatus Center Working Paper
Forbes, S., (2006) Flat Tax Revolution, Regnery Publishing
Foreman, N., (1996) "Single Rate Tax: The path to real simplicity",
Demos
Friedman, M., (1962) Capitalism and Freedom, University of Chicago
Press
Garton-Ash, T., (1999) History of the Present, Vintage Books
Goldstein, J., (1993) Ideas, Interests and American Trade Policy,
N.Y.:Cornell University Press
Goldstein, J., and Keohane, R.O., Eds. (1993) Ideas and Foreign
Policy: Beliefs, Institutions and Political Change, R Ithaca: Cornell
University Press
Grecu, A., (2005) "Flat Tax – The British Case", Adam Smith Institute
Hall, P., Ed. (1989), The Political Power of Economic Ideas:
Keynesianism across Nations, Princeton University Press
Hall, P., and Rabushka, A., (1995) The Flat Tax, 2nd Ed. Hoover
Harberger, A., (1993) "Secrets of Success: A Handful of Heroes",
American Economic Review Papers and Proceedings83: 343-50
Hayek, F.A., (1956) "Progressive Taxation Reconsidered" in Sennholz,
H., Ed. On Freedom and Free Enterprise, Princeton, NJ: Van Nostrand
Hayek, F.A., (1960) The Constitution of Liberty, University of
Chicago Press
Heath, A., (2006) The Flat Tax: Towards a British Model, The TaxPayers’
Alliance
Hettich, W., and Winer, S.L., (1999) Democratic Choice and Taxation:
A theoretical and Empirical Analysis University of Cambridge Press
Kay, J.A., and King, M.A., (1986) The British Tax System, Oxford
University Press
Keen, M., Kim, Y., and Varsano, R., (2006) "The "Flat
Tax(es)":Principles and Evidence", IMF Working Paper
Legro, J.W., (2000) "The Transformation of Policy Ideas", American
Journal of Political Science 44(3): 419-432
Olson, M., (1965) The Logic of Collective Action, Cambridge, MA:
Harvard University Press
Ragin, C.C., (1987) The Comparative Method: Moving Beyond Qualitative
and Quantitative Strategies, University of California Press: Berkeley
Reynolds, A., (1999) The Hong Kong Tax System The Hudson Institute.
Schneider, F., (2003) "The Development of the Shadow Economies
and Shadow Labor Force of 22 Transition and 21 OECD Countries",
Unpublished manuscript
Scott, H., "Could a 10% Flat Rate Tax Work in the UK?", Unpublished
Manuscript
Smith, A., (1991[1766]) An Enquiry into the Nature and Causes of the
Wealth of Nations, Prometheus Books
Stokey, N.L., and Rebelo, S., (1995) "Growth Effects of Flat-Rate
Taxes", Journal of Political Economy 103(3)
Teather, R., (2005) "A Flat Tax for the UK – A Practical Reality",
Adam Smith Institute
Ventura, G., (1999) "Flat tax reform: A quantitative exploration",
Journal of Economic Dynamics and Control 23: 1425-1458
Anthony Evans is Lecturer at ESCP-EAP European School of Management
London and Paul Aligica is Senior Fellow at Mercatus Center – George
Mason University, USA and Associate Professor at National School –
SNSPA, Bucharest, Romania.
http://www.adamsmith.org/blog/archives/001319.php